WATER RESEARCH

Editor-in-Chief 

Mogens Henze 

Institute of Environment & Resources 

Technical University of Denmark 

Bygningstorvet 

DK-2800 KGS Lyngby 

Denmark 

Tel: 45 4525 1477 

Fax: 45 4593 2850 

E-mail: moh@er.dtu.dk 

Editors 

J. Block 

Université H. Poincaré, Nancy I 

France 

David Dixon 

University of Melbourne 

Australia 

Hiroaki Furumai 

The University of Tokyo 

Japan 

Gregory Korshin 

University of Washington 

USA 

Anna Ledin 

Technical University of Denmark 

Denmark 

Eberhard Morgenroth 

University of Illinois Urbana-Champaign 

USA 

W. Rauch 

University Innsbruck 

Austria 

Maria Reis 

Universidade Nova de Lisboa/FCT 

Portugal 

Hang-Shik Shin 

Korea Advanced Institute of Science 

and Technology 

Korea 

Mark van Loosdrecht 

Delft University of Technology 

The Netherlands 

Thomas Ternes 

Bundesanstalt für Gewässerkunde 

Germany 

Stefan Wuertz 

Univ. of California, Davis 

USA 

Hanqing Yu 

University of Science & Technology of China 

China 

Associate Editors 

Andrew Baker 

The University of Birmingham 

UK 

WATER RESEARCH 

A Journal of the International Water Association 

Damien Batstone 

The University of Queensland 

Australia 

G-H. Chen 

The Hong Kong University of Science 

& Technology 

Hong Kong 

China 

Tom Curtis 

Univ. of Newcastle upon Tyne 

UK 

Ana Deletic 

Monash University 

USA 

Francis de los Reyes III 

North Carolina State University 

USA 

Rob Eldridge 

The University of Melbourne 

Australia 

Rosina Girones 

University of Barcelona 

Spain 

Stephen Gray 

Victoria University 

Australia 

Kate Grudpan 

Chiang Mai University 

Thailand 

Xiaodi Hao 

Beijing University of Civil Engineering 

and Architecture 

China 

E.E. Herricks 

University of Illinois - Urbana 

USA 

H-Y. Hu 

Tsinghua University 

China 

P.M. Huck 

University of Waterloo 

Canada 

Bruce Jefferson 

Cranfi eld University 

UK 

Sergey Kalyuzhnyi 

Moscow State University 

Russian Federation 

Jaehong Kim 

Georgia Institute of Technology 

USA 

Jes La Cour Jansen 

Lund Institute of Technology 

Sweden 

G. Langergraber 

BOKU/Univ. of Natural Res. and 

Applied Life Scs. 

Austria 

S-L. Lo 

National Taiwan University 

Taiwan 

Y. Matsui 

Hokkaido University 

Japan 

Max Maurer 

EAWAG 

Switzerland 

How Yong Ng 

National University of Singapore 

Singapore 

Satoshi Okabe 


Japan 

S.L. Ong 

National University of Singapore 

Singapore 

Jong M. Park 

Pohang University of Science & Technology 

Korea 

Miguel Salgot 

University of Barcelona 

Spain 

David Sedlak 

University of California, Berkeley 

USA 

Jean-Philippe Steyer 

LBE-INRA 

France 

M. Takahashi 


Japan 

Kai Udert 

EAWAG 

Switzerland 

V.P. Venugopalan 

BARC Facilities 

India 

E. von Sperling 

Federal University of Minas Gerais 

Brazil 

A.I. Zouboulis 

Aristotle University 

Greece 

Editorial Offi ce 

Mailing address 

Water Research 

Radarweg 29 

1043 NX Amsterdam 

The Netherlands 

Tel: 31 20 485 2701 

Fax: 31 20 485 2521; 31 20 485 3264 

E-mail: wr-eo@elsevier.com 

Publication information: Water Research (ISSN 0043-1354). For 2010, volume 44 is scheduled for publication. Subscription prices are available upon request from the 

publisher or from the Elsevier Customer Service Department nearest you or from this journal’s website (http://www.elsevier.com/locate/watres). Further information 

is available on this journal and other Elsevier products through Elsevier’s website (http://www.elsevier.com). Subscriptions are accepted on a prepaid basis 

only and are entered on a calendar year basis. Issues are sent by standard mail (surface within Europe, air delivery outside Europe). Priority rates are available upon 

request. Claims for missing issues should be made within six months of the date of dispatch. 

Author Enquiries: Please visit this journal’s homepage at http://www.elsevier.com/locate/watres. You can track accepted articles at http://www.elsevier.com/ 

trackarticle and set up e-mail alerts to inform you of when an article’s status has changed. Also accessible from here is information on copyright, frequently asked 

questions and more. Contact details for questions arising after acceptance of an article, especially those relating to proofs, will be provided by the publisher. 

Orders, claims, and journal enquiries: Please contact the Elsevier Customer Service Department nearest you: 

St. Louis: Elsevier Customer Service Department, 11830 Westline Industrial Drive, St. Louis, MO 63146, USA; phone: (877) 8397126 [toll free within the USA]; 

(+1) (314) 4537076 [outside the USA]; fax: (+1) (314) 5235153; e-mail: JournalCustomerService-usa@elsevier.com 

Oxford: Elsevier Customer Service Department, The Boulevard, Langford Lane, Kidlington OX5 1GB, UK; phone: (+44) (1865) 843434; fax: (+44) (1865) 

843970; e-mail: JournalsCustomerServiceEMEA@elsevier.com 

Tokyo: Elsevier Customer Service Department, 4F Higashi-Azabu, 1-Chome Bldg, 1-9-15 Higashi-Azabu, Minato-ku, Tokyo 106-0044, Japan; phone: (+81) (3) 

5561 5037; fax: (+81) (3) 5561 5047; e-mail: JournalsCustomerServiceJapan@elsevier.com 

Singapore: Elsevier Customer Service Department, 3 Killiney Road, #08-01 Winsland House I, Singapore 239519; phone: (+65) 63490222; fax: (+65) 67331510; 

e-mail: JournalsCustomerServiceAPAC@elsevier.com 

Application for membership of International Water Association should be made to: Executive Director, IWA, Alliance House, 12 Caxton Street, 

London SW1H 0QS, U.K. (Tel.: +44 207 654 5500; Fax: +44 207 654 5555; e-mail: members@IWAhq.org.uk; website: http://www.IWAhq.org.uk). 

Registered Charity (England) No. 289269. Individual membership is available from £30 upwards. 

For details contact IWA.

Editorial 

Editorial to special issue in Water Research 

Emerging contaminants in water 

The chemical pollution of natural waters is one of the big 

challenges of the 21st century. Based on the rapid evolution in 

analytical chemistry, with the possibility to detect more polar 

compounds a whole new suite of ‘‘emerging contaminants’’ 

such as pharmaceuticals, hormones and perfluorinated 

compounds has been identified in various compartments of 

the water cycle including both natural and technical aquatic 

systems. The discovery of these micropollutants in the aquatic 

environment has triggered research efforts to investigate 

sources and mitigation strategies with conventional and novel 

treatment processes and assess their importance with regard 

to ecotoxicology and human health. In this special issue on 

emerging contaminants, we have compiled 20 research articles 

and 2 reviews, which deal with these timely issues. 

As mentioned above, quantification of micropollutants in 

aquatic systems is a key requirement to assess their fate. 

Several studies in this issue address the determination of 

pharmaceuticals in archived biosolids (Halden et al.), in 

wastewater treatment (Lindberg et al.) and lagoon treatment 

(Wong et al.). Another study uses analytical data to evaluate 

the fraction of pharmaceutical residues in wastewater originating 

from hospitals (Ort et al.). Finally, a review covers the 

occurrence and fate of phytoestrogens in the environment 

(Liu et al.). 

To further elucidate the relevance of the micropollutants 

detected in various aquatic compartments, their (eco)toxicological 

potential has to be assessed. Several papers address 

related issues for individual compounds (lipid regulators, 

Fernandez-Pinas et al.) or classes of compounds (ionic liquids, 

Yun et al.) and for pharmaceuticals in advanced wastewater 

treatment systems such as powdered activated carbon and 

ozonation with in vivo and in vitro tests (Escher et al., Stalter 

et al.). Furthermore, one study is focused on toxicity nanotube 

suspensions (Tarabara et al.), which have been on the radar of 

emerging contaminants recently. Finally, one paper focuses 

on the toxicological relevance of emerging contaminants for 

drinking water (Schriks et al.). 

In recent years, municipal wastewater has been recognized 

as an important source of micropollutants to the 

water research 44 (2010) 351 

Available at www.sciencedirect.com 

journal homepage: www.elsevier.com/locate/watres 

receiving water bodies. Therefore, mitigation strategies for 

the minimization of the discharge of these compounds play 

an increasingly important role in the urban water management. 

In this special issue, the removal of benzotriazoles 

(Reemtsma et al.) and pharmaceuticals, caffeine and DEET (Yu 

et al.) and other emerging contaminants (Rosal et al.) has been 

investigated for conventional wastewater treatment. 

Numerous papers address the oxidative removal of micropollutants 

from wastewater with chlorine, chlorine dioxide, 

ferrate, ozone, advanced oxidation processes, such as UV/ 

H2O2, (solar) photo-Fenton and non-thermal plasma (Lee 

et al., Malato et al., Reungoat et al., Gerrity et al., Mendez- 

Arriaga). Other options for removal of micropollutants 

include membrane processes such as reverse osmosis (Hu 

et al.) and nanofiltration (Yangali-Quintanilla et al.) as well as 

sorption on sludge (Carrere et al.). The challenges caused by 

harmful algae producing toxins for desalination operations 

were reviewed by David Caron and co-workers. Finally, mitigation 

of micropollutants may also occur during managed 

aquifer recharge (Drewes et al.) or in biological Fenton-like 

processes (Vicent et al.). 

In short, we are very happy to provide you this Theme Issue 

on Emerging Contaminants. We appreciate the contributions 

by the authors, reviewers and editorial staff of Water Research 

to this project. 

Thomas Ternes* 

Federal Institute of Hydrology (BFG), 

Am Mainzer Tor 1, 56068 Koblenz, Germany 

*Corresponding author. Tel.: þ49 261 1306 5560; 

fax: þ49 261 1306 5363. 

Urs von Gunten 

EAWAG, Ueberlandstrasse 133, Duebendorf CH-8600, 

Switzerland 

0043-1354/$ – see front matter 

ª 2010 Published by Elsevier Ltd. 

doi:10.1016/j.watres.2010.01.015

Review 

Environmental fate and toxicity of ionic liquids: A review 

Thi Phuong Thuy Pham a , Chul-Woong Cho a , Yeoung-Sang Yun a,b, * 

a Department of Bioprocess Engineering, Chonbuk National University, Jeonju, Chonbuk 561-756, Republic of Korea 

b Division of Semiconductor and Chemical Engineering and Research Institute of Industrial Technology, Chonbuk National University, 

Chonbuk 561-756, Republic of Korea 

article info 

Article history: 

Received 31 May 2009 

Received in revised form 

27 August 2009 

Accepted 12 September 2009 

Available online 24 September 2009 

Keywords: 

Ionic liquids 

Toxicity 

Degradation 

Biodegradation 

Environmental fate 

Sorption 

Contents 

water research 44 (2010) 352–372 



abstract 

Ionic liquids (ILs) are organic salts with low melting point that are being considered as 

green replacements for industrial volatile organic compounds. The reputation of these 

solvents as ‘‘environmental friendly’’ chemicals is based primarily on their negligible vapor 

pressure. Nonetheless, the solubility of ILs in water and a number of literature 

documenting toxicity of ILs to aquatic organisms highlight a real cause for concern. The 

knowledge of ILs behavior in the terrestrial environment, which includes microbial 

degradation, sorption and desorption, is equally important since both soil and aquatic 

milieu are possible recipients of IL contamination. This article reviews the achievements 

and current status of environmental risk assessment of ILs, and hopefully provides 

insights into this research frontier. 

ª 2009 Elsevier Ltd. All rights reserved. 

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 353 

2. Toxicological aspect of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354 

2.1. Effects of ILs in an enzyme level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354 

2.2. Antibacterial activity of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 356 

2.3. Toxicity of ILs to algae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 

2.4. Cytotoxicity of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 361 

2.5. Phytotoxicity of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363 

2.6. Toxicity of ILs to invertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363 

2.7. Inhibitory effects of ILs on vertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364 

3. Environmental fate of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364 

3.1. Chemical degradation of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364 

3.2. Biodegradability of ILs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365 

3.3. Sorption of ILs in environmental systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367 

* Corresponding author. Division of Semiconductor and Chemical Engineering, Chonbuk National University, Jeonju, Chonbuk 561-756, 

Republic of Korea. Tel.: þ82 63 270 2308; fax: þ82 63 270 2306. 

E-mail address: ysyun@chonbuk.ac.kr (Y.-S. Yun). 

0043-1354/$ – see front matter ª 2009 Elsevier Ltd. All rights reserved. 

doi:10.1016/j.watres.2009.09.030

4. Concluding remarks and future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 368 

Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369 

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 369 

1. Introduction 

Most of volatile organic compounds (VOCs) commonly used in 

industrial applications cause a major concern in the current 

chemical processing industry. The main problems are the 

toxicity of the organic solvents to both the process operators 

and the environment as well as the volatile and flammable 

nature of these solvents which make them a potential explosion 

hazard (Schmid et al., 1998). Recently, the deleterious 

effects of many solvents combined with serious 

environmental issues, such as atmospheric emissions and 

contamination of aqueous effluents are making their use 

prohibitive. Thus, many researchers have focused on the 

development of ‘‘green engineering’’ which represents 

research aimed at finding environmentally benign alternatives 

to harmful chemicals. Among the neoteric solvents applicable 

in ‘‘green technologies’’ ionic liquids (ILs) have garnered 

increasing attention over the others such as supercritical CO 2 

(Blanchard et al., 1999; Blanchard and Brennecke, 2001; 

Kazarian et al., 2000) and aqueous biphasic systems (Myasoedov 

et al., 1995; Rogers et al., 1995; Willauer et al., 1999). 

Ionic liquids, formerly known as molten salts, constitute 

one of the hottest areas in chemistry these days. Basically, 

they have melting points below 100 C, which can be achieved 

by incorporating a bulky asymmetric cation into the structure 

together with a weakly-coordinating anion (Ranke et al., 2004). 

The unique, highly solvating, yet non-coordinating environment 

of ILs provides an attractive medium for various types of 

chemical processes. Also, the physical properties of ILs can be 

tailored by a judicious variation in the length and branching of 

the alkyl chain and the anionic precursor (Fuler et al., 1997; 

Huddleston et al., 2001). In this way, ILs can be made taskspecific 

for a certain application. The almost limitless structural 

possibilities of ILs, as opposed to limited structural 

variations within molecular solvents, make them ‘‘designer 

water research 44 (2010) 352–372 353 

Fig. 1 – Applications of ionic liquids. 

solvents’’ (Marsh et al., 2004; McFarlane et al., 2005; Sheldon, 

2005). Some independent reports (Hagiwara and Ito, 2000; 

Olivier, 1999; Welton, 1999) and many reviews (Earle and 

Seddon, 2000; Rooney and Seddon, 2001) have highlighted ILs 

as representing a state-of-the-art, innovative approach to 

sustainable chemistry, with the argument that their vapor 

pressure is immeasurably low and they are not flammable. 

Recently, the application of these liquids as reaction media for 

organic synthesis, catalysis, or biocatalysis has been well 

documented (Earle and Seddon, 2000; Wasserscheid and 

Keim, 2000; Welton, 1999) (Fig. 1). Gordon (2001) pointed out 

that there is an obvious advantage in performing many reactions 

in ILs due to the improvement in process economics, 

reaction activity, selectivity and yield. 

Although ILs can lessen the risk of air pollution due to their 

insignificant vapor pressure, they do have significant solubility 

in water (Anthony et al., 2001; McFarlane et al., 2005; 

Wong et al., 2002). As a result, this is the most likely medium 

through which ILs will be released into the environment. Ionic 

liquids currently are not widely used in industrial applications; 

nonetheless, continued development and further use of 

these solvents may lead to accidental discharge and 

contamination. The properties that make them be the target 

of industrial interest (i.e. high chemical, thermal stability and 

non-volatility) suggest potential problems with degradation or 

persistence in the environment. In general, the deficiency of 

information and uncertainty surrounding the environmental 

impact of ILs is a major barrier to the utilization of these 

compounds by industry. Initial efforts have been made to 

overcome this drawback and offer a preliminary insight into 

the behavior of ILs in the aqueous environments. These 

studies provided extensive data sets, e.g. on (eco)toxicity, 

biodegradability, bioaccumulation and distribution of ILs in 

different environmental compartments. Therefore, it is 

necessary to consolidate all the available data in a single

354 

review to lay the groundwork for more comprehensive 

community and ecosystem investigations. The overall objective 

of this review is to systematically gather and interpret 

existing information about the fate, removal options and 

(eco)toxicological assessment strategies of ILs. 

2. Toxicological aspect of ILs 

The current literature represents a number of studies addressing 

the biological effects of ILs evaluated on the basis of 

toxicological test systems. The ILs toxicities towards these 

systems of different levels of biological complexity as well as 

several environmental compartments (Fig. 2) aresuccessively 

discussed in the following subsections. All the structures of 

IL compounds discussed in this review were listed in Table 1. 

The acronyms used for these substances were adapted 

from Ranke et al. (2007a). In this way, the cation head groups 

were abbreviated as ‘‘IM’’ for imidazolium, ‘‘Py’’ for pyridinium, 

‘‘Pyr’’ for pyrrolidinium, ‘‘Mor’’ for morpholinium, ‘‘Pip’’ for 

piperidinium, ‘‘Quin’’ for quinolinium, ‘‘N’’ for quaternary 

ammonium and ‘‘P’’ for quaternary phosphonium. The alkyl 

chains attached to the head group were given as numbers 

corresponding to the number of carbon in the alkyl residues. For 

example, the 1-butyl-3-methylimidazolium moiety was denoted 

as IM14. In case the carbon chain length equals or exceeds 10, 

the numbers were separated by a hyphen (e.g. IM1-10 indicated 

1-decyl-3-methylimidazolium). Particularly, for pyridinium 


entities, the carbon-bound alkyl chains were appended to the 

head group at different positions and the abbreviation was 

made by noting the position of attachment and a symbol for the 

attached group (e.g. Py4-2Me for 1-butyl-2-methylpyridinium). 

The anionic components were shortened as they are in the 

periodic table for the halides. For tetrafluoroborate, hexafluorophosphate, 

bis(trifluoromethylsulfonyl)imide, dicyanamide 

and hydrogen sulfate the abbreviations were BF4, PF6, 

(CF 3SO 2) 2N, CN(N) 2 and HSO 4 in respective to their structural 

formula. 

2.1. Effects of ILs in an enzyme level 

Enzyme inhibition data by ILs include those of the acetylcholinesterase 

from electric eel (Electrophorus electricus) 

(Arning et al., 2008; Jastorff et al., 2005; Matzke et al., 2007; 

Ranke et al., 2007b; Stasiewicz et al., 2008; Stock et al., 2004; 

Torrecilla et al., 2009; Zhang and Malhotra, 2005), the AMP 

deaminase (Sk1adanowski et al., 2005) and the antioxidant 

enzyme system of mouse liver (Yu et al., 2009a). The enzyme 

acetylcholinesterase plays the most important role in nerve 

response and function. Also, acetylcholinesterase catalyzes 

the hydrolysis of acetylcholinesters with a relative specificity 

for acetylcholine, which is a neurotransmitter common to 

many synapses throughout mammalian nervous systems 

(Fulton and Key, 2001; Massoulié et al., 1993). Thus, an inhibition 

of acetylcholinesterase leads to various adverse effects 

in neuronal processes, such as heart diseases or myasthenia 

Fig. 2 – The flexible (eco)toxicological test battery considering aquatic and terrestrial compartments as well as different 

trophic levels including enzymes, luminescent marine bacteria, freshwater green algae, mammalian cells, duckweed, 

freshwater crustacean and zebrafish (Adapted from Matzke et al. (2007) by permission of the Royal Society of Chemistry).

Table 1 – Selection of cationic and anionic structures of commonly used ionic liquids. 

Cation 


CH 3 

CH 3 

R 2 

Head group Side chain 

N N + 

R 1 

Imidazolium (IM) 

CH 3 

N + 

Pyridinium (Py) 

N + 

CH 3 

R 1 

Pyrrolidinium (Pyr) 

O 

N + 

CH 3 

R 1 

Morpholinium (Mor) 

N + 

CH3 Piperidinium (Pip) 

N + 

R 1 

R 1 

Quinolinium (Quin) 

N + 

R1 R2 

R 4 

R3 

Quaternary ammonium (N) 

P + 

R1 R2 

R 4 

R3 

Quaternary phosphonium (P ( 

R 1 

R1 ¼ -C2H5, -C3H7, -C4H9, 

-C 5H 11, -C 6H 13, -C 7H 15, 

-C 8H 17, -C 9H 19, -C 10H 21, 

-C14H29, -C16H33, -C18H37, 

-C19H39 

R 2 ¼ -CH 3,-C 2H 5 

R1 ¼ -C2H5, -C3H7, -C4H9, 

-C 5H 11, -C 6H 13, -C 8H 17 

R1 ¼ -C4H9, -C6H13, -C8H17 

R 1 ¼ -C 4H 9 

R1 ¼ -C4H9 

R1 ¼ -C4H9, -C6H13, -C8H17 

R 1-4 ¼ -CH 3,-C 2H 5,-C 3H 7, 

-C4H9, -C6H13 

R 1-4 ¼ -C 4H 9,-C 6H 13, -C 14H 29 

(continued on next page)

356 

Table 1 (continued) 

in humans (Chemnitius et al., 1999; Pope et al., 2005). Ranke 

et al. (2007b) published a comprehensive collection of acetylcholinesterase 

inhibition values for 292 compounds covering 

a large variety of ILs and closely related salts. Among these 

data, only those of the commonly tested ILs are summarized 

in Table 2 for the ease of comparison of ILs toxicity from 

molecular up to organism levels of biological complexity. 

It was found that all observed inhibitory effects on the enzyme 

could be exclusively accounted for the cationic moiety (Arning 

et al., 2008). In particular, the IL with pyridinium as cationic 

core structure inhibited the enzyme slightly stronger than the 

imidazolium analogue whereas the compounds based on 

phosphonium was less inhibitory. All anion species exerted 

no effect on the enzyme activity with only exception of the 

fluoride anion and the fluoride containing [SbF6] and [PF6] 

species. Both species are known to readily undergo hydrolysis 

in contact with moisture and thus the fluoride seems to be the 

active compound. The non-inhibiting effects of anion might 

be explained by their limited interactions with the active site 

of this enzyme (Matzke et al., 2007). In addition, a correlation 

between an increasing chain length of the side chains 

connected to the cationic head groups and an enhanced 

inhibitory potential of the ILs was found. It is believed that the 

mechanism involves the similarity of the positively charged 

imidazolium or pyridinium to the choline part that binds to 

the anionic site of the enzyme, such that the longer alkyl 

chain results in an improved fit (Stock et al., 2004). 

Sk1adanowski et al. (2005) discussed the usefulness of in 

vitro AMP deaminase inhibition assay as a potential molecular 

method in prospective risk analysis of imidazolium-based ILs. 

The results revealed that IM14 salts associated with [PF 6] , 

[BF 4] , p-tosylate and [Cl] demonstrated a dose-dependent 

inhibition of AMP deaminase activity. The IC50 values 

(concentration of ILs inhibiting 50% of enzyme activity) for 

those containing a fluorine compartment [PF6] and [BF4] are 

lower (5 mM) than those for [Cl] and p-tosylate (10 mM), which 

indicated the adverse effect of these fluoride-containing 

anions. The other study on enzyme inhibition assay dealt with 

Head group Side chain 

Anion Chloride Cl 

Bromide Br 

Tetrafluoroborate [BF4] 

Hexafluorophosphate [PF6] 

Bis(trifluoromethylsulfonyl)imide [(CF 3SO 2) 2N] 

Dicyanamide [(CN)2N] 


the effects of acute exposure of intraperitoneal injection of 

aqueous IM18 Br on the antioxidant enzymes of the treated 

mouse liver (Yu et al., 2009a). The antioxidant enzymes tested 

included superoxide dismutase, catalase, glutathione peroxidase 

and glutathione-S-transferase. The results showed that 

administration of IM18 Br modified activities of these defense 

enzymes in mouse liver, and caused damage to livers of 

treated mice at median lethal dose (LD 50) of 35.7 mg/kg. 

Though data published by these authors did not cover 

a large variety of ILs, the enzyme inhibition assays suggest the 

trend in which cationic moiety is the dominating factor 

influencing the toxicity of ILs, especially when substituted 

with a long alkyl side chain. Regarding the anion types, 

perfluoronated ions are of toxicological interest due to 

hydrolysis resulting in HF formation, while the others cause 

less prominent effect. 

2.2. Antibacterial activity of ILs 

Bacteria serve as an ideal starting point for ILs toxicity 

estimations as they have short generation times. Preliminary 

toxicological investigations have shown quaternary ammonium 

and pyridinium compounds have critical inhibitory 

effects on a variety of bacteria and fungi (Babalola, 1998; 

Kelman et al., 2001; Li et al., 1998). In the studies of Pernak’s 

group (Cieniecka-Ros1onkiewicz et al., 2005; Pernak et al., 

2001a; Pernak et al., 2001b; Pernak and Chwa1a, 2003; Pernak 

et al., 2003; Pernak et al., 2004a), they observed a trend of 

increasing toxicity with an increase in the alkyl chain length 

substituent in the pyridinium, imidazolium and quaternary 

ammonium salts to various bacteria including rods, cocci 

and fungi. As a measure of microbial activity of imidazolium 

and pyridinium ILs with varying alkyl chain lengths, Docherty 

and Kulpa (2005) also used a group of microorganisms 

possessing a variety of physiological and respiratory activities. 

It was found that imidazolium and pyridinium bromides 

incorporated hexyl- and octyl-chain had considerable antimicrobial 

effect to pure cultures of Escherichia coli, 

F 

F 

F 

B - 

F 

F 

P - 

F 

F 

F 

F 

F 

N - 

O O 

F3C CF3 S S 

O 

O 

N - 

N N

Table 2 – Toxicity of ILs to different levels of biological complexity including enzyme, bacteria, algae, rat cell line, human cell lines, duckweed and invertebrate. 

Compound Log10EC50 (mM) a 

IM12 Cl 2.06 14 

IM12 BF4 IM12 PF6 

2.05 14 

IM12 (CF3SO2)2N 2.03 14 

IM13 Cl 2.27 14 

IM13 BF4 2.28 0.03 18 

IM13 PF6 

2.22 14 

IM14 Cl 1.91 0.04 11 

Acetylcholin esterase Vibrio 

fischeri 

2.05 14 

IM14 Br 1.90 0.02 18 

IM14 BF 4 1.98 0.018 11 

Escherichia 

coli 

Pseudo kirchneriella 

subcapitata 

Scenedesmus 

vacuolatus 

IPC-81 HeLa MCF7 b 

Lemna 

minor 

4.55 10 

4.33 0.11 19 

N.A. N.A. 2.78 0.06 19 

N.A. N.A. N.A. N.A. N.A. 

N.A. 5.25 0.06 9 

N.A. N.A. 3.44 14 

4.00 0.04 20 

N.A. N.A. N.A. 

N.A. N.A. N.A. N.A. 3.92 14 

N.A. N.A. N.A. N.A. 

N.A. N.A. N.A. N.A. N.A. 3.26 0.04 20 


N.A. N.A. N.A. N.A. >4.30 14 


3.94 0.06 13 

N.A. N.A. N.A. 3.47 14 


N.A. N.A. N.A. N.A. >3.00 14 


3.71 0.14 4 

2.95 5 

3.34 0.13 6 

3.47 0.04 19 

4.01 0.05 4 

3.07 0.03 13 

3.35 5 

3.27 0.09 6 

3.55 0.04 13 

3.10 0.17 6 

3.12 0.35 16 

3.07 0.29 6 

N.A. 2.34 0.01 21 

N.A. 3.46 0.062 3 

4.60 0.02 9 

N.A. 2.11 11 

2.26 0.08 11 

3.55 14 

N.A. 3.43 14 

3.12 14 

N.A. N.A. 2.82 11 

3.44 0.11 20 

3.72 0.05 17 

3.66 0.08 20 

Daphnia 

magna c 

1.93 4 

1.93 0.06 1 

N.A. N.A. 1.57 4 

1.56 0.07 1 

1.85 0.06 22 

N.A. 2.49 7 

1.68 4 

1.67 0.11 1 

IM14 PF6 2.15 0.05 18 

4.15 0.06 9 

2.20 0.04 21 

N.A. 3.10 14 

4.14 0.22 17 

N.A. N.A. 1.85 4 

1.85 0.10 1 

IM14 (CF3SO2)2N 1.96 0.021 11 

3.39 0.08 4 

2.55 0.15 9 

1.80 0.07 12 

1.81 0.15 11 

2.68 14 

3.07 0.08 20 

N.A. 2.45 0.08 11 

N.A. 

IM14 (CN)2N 1.95 0.07 18 

3.67 0.10 4 

2.99 5 

N.A. N.A. N.A. 3.15 14 


IM15 Cl 1.96 14 

N.A. N.A. N.A. N.A. >3.00 14 


IM15 BF4 

1.86 14 

3.14 0.02 13 

N.A. N.A. N.A. >3.00 14 


IM15 PF6 

1.85 14 

N.A. N.A. N.A. N.A. >3.00 14 


IM16 Cl 1.92 14 

1.94 10 

2.32 0.16 6 

2.91 0.09 13 

N.A. 1.92 0.09 21 

0.08 19 

2.85 14 


IM16 Br N.A. 1.42 0.12 4 

0.81 5 

N.A. 2.57 0.15 2 

N.A. N.A. N.A. N.A. N.A. 0.78 4 

1.06 0.04 22 

IM16 BF4 1.88 14 

3.18 0.03 13 

N.A. N.A. N.A. 2.98 14 


IM16 PF6 

1.88 14 

2.17 0.06 6 

3.25 0.67 9 

N.A. N.A. 2.91 14 


IM16 (CF3SO2) 2N 2.15 14 

N.A. 2.53 0.15 9 

N.A. N.A. 2.24 14 

N.A. 2.81 8 

N.A. N.A. 

IM17 Cl 2.07 14 

N.A. N.A. N.A. N.A. 2.53 14 


IM17 BF4 

2.12 14 

2.44 0.06 13 

N.A. N.A. N.A. 2.58 14 


IM17 PF6 

1.91 14 

N.A. N.A. N.A. N.A. 2.30 14 


IM18 Cl 1.60 14 

1.19 0.11 6 

1.01 0.06 19 

N.A. 1.46 21 

2.67 0.37 19 

2.01 14 


IM18 Br N.A. 0.63 0.07 4 

0.07 5 

N.A. 1.65 0.25 2 

N.A. N.A. 2.48 0.04 20 

N.A. N.A. 1.33 4 

0.54 0.12 22 

(continued on next page) 

water research 44 (2010) 352–372 357


Compound Log 10EC 50 (mM) a 


fischeri 

Escherichia 

coli 


subcapitata 

Scenedesmus 

vacuolatus 


IM18 BF4 1.53 0.025 11 

1.41 0.07 13 

N.A. N.A. 2.30 11 

1.59 14 

2.48 0.02 20 

2.84 8 

0.90 7 

N.A. 

IM18 PF6 2.03 14 

0.95 0.12 6 

2.64 0.15 9 

N.A. N.A. 1.96 14 


IM18 (CF3SO2)2N 2.03 14 

N.A. N.A. N.A. N.A. 1.64 14 

2.28 0.02 20 


IM19 BF4 N.A. 0.72 0.04 13 

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 

IM1-10 Cl 1.09 14 

0.50 0.07 13 

0.23 0.06 19 

N.A. N.A. 3.57 0.06 19 

1.34 14 


IM1-10 BF4 1.10 0.04 18 

0.18 0.06 13 

N.A. N.A. N.A. 0.77 14 


IM1-10 PF6 1.68 14 

N.A. N.A. N.A. N.A. 1.50 14 


IM1-14 Cl 0.54 14 

0.15 0.07 19 

N.A. N.A. 2.48 0.2 19 

0.42 14 


IM1-16 Cl 0.68 14 

0.23 0.08 19 

N.A. N.A. > 2.00 19 

0.19 14 


IM1-18 Cl 0.96 14 

1.45 0.05 19 

N.A. N.A. > 2.00 19 

0.01 14 


IM1-19 Cl 1.36 14 

N.A. N.A. N.A. N.A. 1.40 14 


IM1-19 BF4 1.43 14 

N.A. N.A. N.A. N.A. 1.65 14 


IM1-19 PF6 

1.62 14 

N.A. N.A. N.A. N.A. 1.85 14 


IM22 Br 2.08 14 

N.A. N.A. N.A. N.A. >3.00 14 


IM23 Br 2.21 14 

N.A. N.A. N.A. N.A. >3.30 14 


IM24 BF4 2.03 0.01 18 

2.8 0.04 13 

N.A. N.A. N.A. 3.26 14 

4.36 0.09 17 


IM25 BF4 N.A. 3.14 13 


IM26 Br 1.77 14 

N.A. N.A. N.A. N.A. 2.01 14 


IM26 BF4 1.84 14 

2.15 0.05 13 

N.A. N.A. N.A. 2.26 14 


IM2-10 Br 0.92 14 

N.A. N.A. N.A. N.A. 0.53 14 


Py Cl >3.00 14 

N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 

Py2 Cl 2.10 14 


Py3 Br 2.22 14 


Py3 (CF3SO2)2N 2.21 14 


Py4 Cl 1.70 14 

3.41 0.08 4 

2.64 5 

3.18 0.06 19 

N.A. 2.57 0.06 21 

2.59 0.11 19 

N.A. N.A. N.A. 2.32 0.18 19 

N.A. 

Py4 Br 1.77 14 

3.40 0.01 4 

2.73 5 

N.A. N.A. N.A. 3.90 14 

3.50 0.07 20 


Py4 BF4 1.80 14 

N.A. N.A. N.A. N.A. 3.18 14 


Py4 PF6 

1.84 14 


Py4 (CN)2N N.A. 3.31 0.10 4 


2.61 5 

Py5 Br 1.52 14 


Py5 (CF3SO2)2N 1.55 14 


Py6 Cl 1.72 14 


Py6 Br N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. N.A. 1.07 4 

Py6 PF6 1.76 14 


Py6 (CF3SO2)2N 1.85 14 


Py8 Cl 1.60 14 

N.A. N.A. N.A. N.A. 1.27 14 


Py8 (CF3SO2) 2N 1.40 14 


Py4-2Me Cl 0.70 14 


Py4-2Me BF4 

0.82 14 

N.A. N.A. N.A. N.A. 3.25 14 


Lemna 

minor 

Daphnia 

magna c 

358 

water research 44 (2010) 352–372

Py4-3Me Cl 1.15 14 


Py4-3Me Br N.A. 2.75 0.13 4 

N.A. 3.46 0.062 3 

N.A. N.A. N.A. N.A. N.A. 1.76 4 

2.12 5 

Py4-3Me BF4 1.53 0.02 18 

N.A. N.A. N.A. N.A. 3.30 14 


Py4-3Me PF6 1.45 0.02 18 


Py4-3Me (CN) 2N 1.22 14 

2.66 0.05 4 

1.99 5 


Py6-3Me Cl 1.06 14 

1.44 10 


Py6-3Me Br N.A. 2.06 0.16 4 

N.A. N.A. N.A. N.A. N.A. N.A. N.A. 0.59 4 

1.48 5 

Py6-4Me Cl 1.44 14 


Py6-4Me BF4 

1.48 14 

N.A. N.A. N.A. N.A. 2.17 14 


Py8-3Me Cl 0.64 14 


Py8-3Me Br N.A. 0.79 0.05 4 

N.A. N.A. N.A. N.A. N.A. 1.00 8 

N.A. 0.40 4 

Py8-4Me Cl 1.11 14 

Py8-4Me BF4 1.22 14 

Pyr14 Cl 1.92 14 

Pyr14 Br 1.93 14 

0.25 5 

N.A. N.A. N.A. N.A. 1.63 14 

N.A. N.A. N.A. N.A. 1.49 14 

>4.30 19 

N.A. N.A. 3.37 0.10 19 

>4.30 14 

N.A. N.A. 3.67 0.28 3 

N.A. 3.77 14 



N.A. N.A. 2.16 0.25 19 

N.A. 

N.A. 4.64 0.02 15 

N.A. N.A. 

Pyr14 BF4 1.91 14 

N.A. N.A. N.A. N.A. 2.90 14 


Pyr14 (CF3SO2)2N 2.13 14 

N.A. N.A. >2.38 12 

2.53 19 

3.01 14 

N.A. 3.14 8 

2.98 0.32 19 

N.A. 

Pyr14 (CN)2N 1.98 14 

N.A. N.A. N.A. N.A. 4.23 14 


Pyr16 Cl 2.48 14 

2.99 10 

N.A. N.A. N.A. 2.91 14 


Pyr16 (CF3SO2) 2N 2.60 14 


Pyr18 Cl 2.36 14 

N.A. N.A. N.A. N.A. 2.59 14 


Pyr18 BF4 

2.02 14 

N.A. N.A. N.A. N.A. 1.82 14 


Pyr66 2.08 14 

N.A. N.A. N.A. N.A. 1.23 14 


Mor14 Cl N.A. >4.30 19 


Mor14 Br 2.71 14 

N.A. N.A. N.A. >4.00 19 

>4.30 14 

N.A. N.A. 3.11 0.13 19 

N.A. 

Mor14 (CF3SO2) 2N 2.78 14 

N.A. N.A. N.A. 2.00 19 

3.43 14 

N.A. N.A. 3.15 0.13 19 

N.A. 

Pip14 Br 1.83 14 

4.27 0.09 19 

N.A. N.A. 3.27 0.12 19 

4.03 14 

N.A. 4.15 0.04 15 

0.47 19 

N.A. 

Pip14 (CF3SO2)2N 1.78 14 

N.A. N.A. N.A. 2.08 19 

3.41 14 

N.A. 2.93 8 

2.85 0.07 19 

N.A. 

Quin4 Br 0.79 14 

N.A. N.A. N.A. N.A. 2.32 14 


Quin4 BF4 0.62 14 

N.A. N.A. N.A. N.A. 2.16 14 


Quin6 BF4 

0.48 14 

N.A. N.A. N.A. N.A. 1.07 14 


Quin8 Br N.A N.A. N.A. N.A. N.A. 0.03 14 


Quin8 BF4 0.30 14 

N.A. N.A. N.A. N.A. 0.17 14 


N1111 Br N.A. >5.00 4 


N1114 (CF3SO2)2N 2.60 14 

N.A. N.A. N.A. N.A. 3.61 14 


N1123 (CF3SO2) 2N 2.34 14 


N1124 Cl 2.06 14 

N.A. N.A. N.A. >4.00 19 

>4.30 14 

N.A. N.A. 0.83 0.67 19 

N.A. 

N1124 (CF3SO2)2N 2.03 14 

N.A. N.A. N.A. 1.78 0.17 19 

3.43 13 


N2222 Cl 2.80 14 

N.A. N.A. N.A. N.A. >3.48 14 


N2222 Br N.A. >5.00 4 

N.A. N.A. N.A. N.A. 4.26 0.04 20 


N2226 Br N.A. 2.46 0.16 4 


N4444 Br 2.30 14 

3.27 0.07 4 

N.A. N.A. N.A. 2.25 14 

N.A. N.A. N.A. 1.47 4 

P4444 Br 2.61 14 

2.71 4 

N.A. N.A. N.A. 1.66 14 

N.A. N.A. N.A. 0.95 4 

P666-14 Br 2.85 14 

3.41 0.02 4 


4.25 8 

4.15 8 


water research 44 (2010) 352–372 359

360 


Compound Log10EC50 (mM) a 

Daphnia 

magna c 

Lemna 

minor 


Scenedesmus 

vacuolatus 


subcapitata 

Escherichia 

coli 


fischeri 

N.A. N.A. N.A. N.A. 0.48 14 



N.A. N.A. N.A. N.A. 0.24 14 

1.90 0.05 20 



P666-14 BF4 3.47 0.08 18 

P666-14 PF6 

>3.30 17 

P666-14 (CF3SO2)2N >3.48 14 

P666-14 (CN) 2N 3.40 0.2 18 

References: 1 Bernot et al. (2005a); 2 Cho et al. (2007); 3 Cho et al. (2008a,b); 4 Couling et al. (2006); 5 Docherty and Kulpa (2005); 6 Garcia et al. (2005); 7 Jastorff et al. (2005); 8 Kumar et al. (2009); 9 Lee et al. (2005); 

10 11 12 13 14 15 16 17 18 

Luis et al. (2007); Matzke et al. (2007); Pretti et al. (2008); Ranke et al. (2004); Ranke et al. (2007b); Salminen et al. (2007); Samorì et al. (2007); Stepnowski et al. (2004); Stock et al. (2004); 

19 Stolte et al. (2007a); 20 Wang et al. (2007); 21 Wells and Coombe (2006); 22 Yu et al. (2009). 

a N.A. means not available (not determined). 

b Toxicity of ILs is expressed as log10IC50 (mM) in case of MCF7 cell line. 

c Toxicity of ILs is expressed as log10LC50 (mM) in case of D. magna. 


Staphylococcus aureus, Bacillus subtilis, Pseudomonas fluorescens 

and Saccharomyces cerevisiae. The anion performed nearly no 

effect on antimicrobial activity in the case of imidazolium 

analogues (Docherty and Kulpa, 2005; Garcia et al., 2005; Lee 

et al., 2005; Pernak et al., 2003; Pernak et al., 2004a) whereas 

this was not the case for phosphonium salts. Within the group 

of alkyltrihexylphosphonium ILs in the study of Cieniecka- 

Ros1onkiewicz et al. (2005), both cation structure and the type 

of anion had effects on the biological activity. 

The antibacterial activity of ILs not only involves in 

hampering the growth rate of microbes but also interferes 

with their productivity. Matsumoto et al. (2004a) tested the 

toxicity of imidazolium-based ILs to lactic acid producing 

bacterium Lactobacillus rhamnosus to examine whether these 

compounds can replace conventional organic solvents in the 

extractive fermentation of lactate. The results showed that 

the bacterium L. rhamnosus grew, consumed glucose, and 

produced lactate in the presence of imidazolium-based ILs. 

A change of alkyl length in the imidazolium cation had little 

difference on the survival of the cells. In a similar study 

(Matsumoto et al., 2004b), they focused on hiochii bacteria, 

Lactobacillus homochiochii and Lactobacillus fructivorans and also 

found that the bacteria could produce lactic acid in the presence 

of ILs. Nonetheless, the lactic acid producing activities of 

these bacteria generally decreased with the extension of alkyl 

chain length in the imidazolium cation moiety. 

Water miscible ILs had various effects on the physiology of 

Clostridium sporogenes when tested as additives in culture 

media or reaction media for reduction of nitrobenzene 

(Dipeolu et al., 2008). In their study, 2-hydroxyethyltrimethylammonium 

dimethylphosphate and N,N-dimethylethanolammonium 

acetate increased the growth rate of C. sporogenes; 

by contrast, IM14 BF4 and IM12 EtSO 4 inhibited growth. 

Although IM12 EtSO 4 inhibited growth, it was sufficiently 

non-toxic to allow efficient reduction of nitrobenzene using 

harvested cells. Thus, it is recommended that both noninhibitory 

and partially inhibitory ILs should be screened for 

use in biotransformation. Nonetheless, Ganske and Bornscheuer 

(2006) referred that ILs could have substantial inhibitory 

effects on the growth of microorganisms when they 

explored the effects of the two most commonly used ILs IM14 

BF 4 and IM14 PF 6 on the growth of E. coli, Pichia pastoris and 

Bacillus cereus. 

Regarding inhibition assays used in assessment of environmental 

potential risk of a compound in aquatic milieu, the 

bioluminescence assay using Vibrio fischeri (formerly known as 

Photobacterium phosphoreum) is one of the most applied (Kaiser 

and Palabrica, 1991; Steinberg et al., 1995). This is a rapid, costeffective, 

and well-established method for toxicity determination 

focusing on environmental issues, and also a standard 

ecotoxicological bioassay in Europe (DIN EN ISO 11348). The 

published data on ILs toxicity towards V. fischeri were listed on 

Table 2 and were comprehensively interpreted in the study of 

Peraccini et al. (2007). Although it has been claimed that 

modifications of the anion lead to changes in chemical and 

physical properties of ILs (Sheldon, 2001), no clear increase in 

toxicity caused by the anion could be observed, and toxicity 

seemed to be determined mainly by the cationic component 

(Ranke et al., 2004). This is likely explained by the fact that 

lipophilic part of the molecules can be intercalated into the

membrane, whereas their ionic head group is at least partially 

solvated in the aqueous solution, as suggested by Austin et al. 

(1998). The ILs toxicity was also observed to correlate directly 

with the length of the n-alkyl residues in the methylimidazolium 

cation (Romero et al., 2008). Interestingly, Ranke 

et al. (2004) noted a slight hormetic effect at concentrations 

below inhibitory concentrations. Concerning the anionic 

influence, compounds with [PF6] were found to be slightly 

more toxic than compounds with other anions in their study 

(Ranke et al., 2004). The anion [(CF 3SO 2) 2N] showed no 

intrinsic toxicity to V. fischeri in the report of Matzke et al. 

(2007); in contrast, an increased in toxicity was found for all 

tested compounds combined with [(CF3SO2)2N] for V. fischeri 

(Stolte et al., 2007a). Couling et al. (2006) extended the bioluminescence 

inhibition assay to pyridinium derivatives and it 

was noted that the quaternary ammonium compounds 

seemed to be less toxic to V. fischeri than the pyridinium 

and imidazolium analogues. Also, the quantitative structureproperty 

relationship (QSPR) modeling suggested that imidazolium 

cations, with two nitrogen atoms, are predicted to 

be more toxic than pyridinium moieties, which only have 

one nitrogen atom in the structure. In addition, the QSPR 

correlation predicted that quaternary ammonium cations are 

less toxic than those with cations containing nitrogen-bearing 

rings, which was in agreement with the experimental results 

(Couling et al., 2006). However, in contrast to the cases of 

aromatic ILs and ammonium compounds, the authors were 

unsuccessful in modeling the behavior of phosphonium salts 

using the developed correlation. 

2.3. Toxicity of ILs to algae 

As algae are primary producers, either directly or indirectly, of 

organic matter required by animals in freshwater food chains, 

their ecology is crucial in providing the energy for sustaining 

other higher trophic levels. The ubiquity of algae makes these 

organisms ideal for toxicological studies and, because they 

have a short life cycle they can respond quickly to environmental 

change (Blaise, 1993; Lewis, 1995). To date, several 

groups have focused their attention on the use of algal 

primary producers to assess the effects of ILs to aquatic 

environments (Cho et al., 2007; Cho et al., 2008a,b,c; 

Grabinska-Sota and Kalka, 2006; Kulacki and Lamberti, 2008; 

Lata1a et al., 2005; Matzke et al., 2007; Matzke et al., 2008; Pham 

et al., 2008a,b; Pretti et al., 2009; Stolte et al., 2007a; Wells and 

Coombe, 2006). Cho and co-workers used Pseudokirchneriella 

subcapitata (formerly known as Selenastrum capricornutum) to 

study the effect of different head groups, side chains and 

anions of ILs on algal growth rate and photosynthetic activity. 

The data revealed that the toxic influence of ILs on growth 

rates were more significant than those of photosynthetic 

performance (Pham et al., 2008b). Once again, the trend of 

increasing toxicity with increasing alkyl chain length was 

observed in their reports (Cho et al., 2007; Pham et al., 2008b). 

Regarding the anionic effects, P. subcapitata was sensitive 

to the anion moieties in the order: [SbF6] > [PF6] > 

[BF4] > [CF3SO3] > [C8H17OSO3] > [Br] z [Cl] . In particularly, 

it was found that with respect to IL incorporating perfluorinated 

anion (i.e. IM14 BF4), EC50 values (concentrations 

which lead to a 50% reduction of the exposed organisms 


relative to control) of the previously prepared stock solution 

(6 months prior to experiment) were significantly lower 

compared to those of the freshly made one (Pham et al., 

2008a). This might be due to hydrolytic effects of IM14 BF4 

leading to fluoride formation, as confirmed by ion chromatography 

analysis. This implies that after ILs are released into 

the aqueous system; they can become more hazardous than 

expected by laboratory data with fresh ILs. In a detailed study 

on hydrolysis of fluoride-containing anions, Cho et al. (2008a) 

showed that IM14 SbF 6 generated a greater amount of fluoride 

compared to IM14 BF 4, but no fluoride formation occurred 

with the hexafluorophosphate. When only small amounts of 

fluoride ions were formed from IM14 SbF6 and IM14 BF4 within 

96 h, the formed fluoride ion did not affect the algal growth 

rate. Nevertheless, the fluoride ion formation from IM14 BF4 

increased with incubating time of the stock solution; thus, the 

toxicity might significantly increase according to the further 

formed fluoride ions. In view of cationic effect, Pyr14 Br was 

found to be the least toxic of all the ILs tested to P. subcapitata 

(Cho et al., 2008b). For the limnic green alga Scenedesmus 

vacuolatus, a severe toxicity was found for 1-butyl-4-(dimethylamino)pyridinium, 

whereas the quaternary ammonium 

and morpholinium compounds exhibited no toxicity (Stolte 

et al., 2007a). Despite the extensive studies on the toxicological 

impact of ILs towards freshwater phytoplankton, inhibition 

mechanism of both the growth rate and photosynthetic 

activity by ILs has not been described by the authors. 

Lata1a et al. (2005), who selected two marine algae Oocystis 

submarina (green algae) and Cyclotella meneghiniana (diatom) as 

testing organisms, found that the two species differed 

dramatically in their ability to recover from IL exposure. 

Additionally, it was discovered that IL toxicity declined with 

increasing salinity. The lower toxicity of IL in this case is 

probably due to the reduced permeability of IL cations through 

the algal cell walls. High amounts of chloride provide a good 

ion-pairing environment for imidazolium cations, which 

consequently compete with hydroxyl or silanol functional 

groups in the cell-wall structure of green alga and diatom, 

respectively. Though no information on EC50 values was 

described, the facts emerged from this work provide useful 

information in the further fate assessment of ILs in marine 

environments. 

2.4. Cytotoxicity of ILs 

As a cellular test system, promyelotic leukemia rat cell line 

IPC-81 has been frequently used in cytotoxicity assays of ILs, 

with the reduction of the WST-1 dye as an indicator of cell 

viability (Matzke et al., 2007; Ranke et al., 2004; Ranke et al., 

2007a; Stasiewicz et al., 2008; Stolte et al., 2006; Stolte et al., 

2007b; Torrecilla et al., 2009). It was observed that ILs with 

polar ether, hydroxyl and nitrile functional groups within the 

side chains exhibited low cytotoxicity compared to those 

incorporated with ‘‘simple’’ alkyl side chains (Kumar et al., 

2009; Stasiewicz et al., 2008; Stolte et al., 2007b). Those functional 

groups were thought to impede cellular uptake by 

membrane diffusion and reduce lipophilicity based interactions 

with the cell membrane (Stolte et al., 2007b). Taking 

a closer look at the effects of sub-structural elements of ILs, 

[(CF3SO2)2N] anion and 4-(dimethylamino)pyridinium cation

362 

were described to have intrinsic effects of anion and head 

group on cytotoxicity, respectively (Stolte et al., 2007b). The 

well known side chain length effect (decrease in EC50 values 

with elongation of the alkyl side chain) could also be 

confirmed in these studies. 

To date many studies have analyzed the toxicity of ILs on 

human cell lines (Frade et al., 2007; García-Lorenzo et al., 

2008; Hassoun et al., 2002; Kumar et al., 2009; Salminen et al., 

2007; Stepnowski et al., 2004; Wang et al., 2007). These in vitro 

systems have been extremely beneficial in studying the 

molecular basis of chemical’s biological activity, including its 

toxic mode of action (Blaauboer et al., 1998) and could facilitate 

extrapolation of in vitro data with regard to possible 

effects on humans (Malich et al., 1997). Most of studies dealt 

with HeLa cells exemplifying prototypical cells of the human 

epithelium which is normally the site of first contact of an 

organism with toxicants. According to Stepnowski et al. 

(2004), the cytotoxicity data implied that effects of IM14 

cation coupled with chloride, tetrafluoroborate or hexafluorophosphate 

were probably dependent on the anionic 

moieties. The lowest effect concentrations for tetrafluoroborate 

species were found to be 0.63 mM, whereas 

hexafluorophosphate and chloride inhibited HeLa cell growth 

at comparably high concentrations of >10 mM. Surprisingly, 

when the anion effect was compared, the strongest inhibition 

was found for [PF6] . This might be due to hydrolysis 

affecting fluoride formation, thus causing serious toxicological 

consequences through the decomposition product. 

A similar phenomenon was observed by Ranke et al. (2004) in 

IPC-81 leukemia cells, where the lower toxicity of 1-n-butyl-3methylimidazolium 

hexafluorophosphate in comparison to 

the hexafluorophosphate anion alone was explained by 

reduced anion uptake due to the formation of an ion pair. 

The anion in this ion pair can, however, also be partially 

decomposed. This was shown in recent work by Swatloski 

et al. (2003), who identified traces of 1-n-butyl-3-methylimidazolium 

fluoride hydrate as a decomposition product 

formed during the purification of the 1-n-butyl-3-methylimidazolium 

hexafluorophosphate. 

As shown by Wang et al. (2007) the phosphonium 

bis(trifluoromethylsulfonyl)imide salts performed the highest 

inhibitory to HeLa cells, followed by alkylimidazolium, 

alkylpyridinium, alkyltriethylammonium and N-alkyl-N,Ndimethyl-N-(2-hydroxylethyl)ammonium 

salts, in decreasing 

order. For each cation class the toxicity increased with 

increasing chain length of the alkyl substituent for a given anion: 

1-ethyl-3-methylimidazolium bromide yielded an EC50 of 

8.4 mM, substituting the ethyl moiety for a butyl group led to an 

EC50 of 2.8 mM, and for an octyl moiety an EC50 of 0.3 mM. This 

result was consistent with what has been observed in other 

studies. Salts containing the tetrafluoroborate anion showed the 

highest EC50, followed closely by bromide and chloride. Bis(trifluoromethylsulfonyl)imide 

salts were significantly more toxic 

than their halide counterparts. However, the effect of changing 

the anion was smaller than that of changing the alkyl substituent, 

e.g. while 1-ethyl-3-methylimidazolium tetrafluoroborate 

was observed to have an EC 50 of 9.9 mM, the corresponding 

bromide and bis(trifluoromethylsulfonyl)imide salts had EC50 of 

8.4 and 1.8 mM, respectively – these all considerably less 

toxic than 1-octyl-3-methylimidazolium bromide. 


The CaCo-2 cells were used in the study of García-Lorenzo 

et al. (2008) with the aim of a convenient screening method 

for obtaining first rough estimates for the toxic potential of 

ILs. The obtained data showed that in general, ILs with longer 

alkyl chains were more lipophilic than those with shorter 

alkyl chains. The former can be presumed to have a tendency 

to be incorporated into the phospholipid bilayers of biological 

membranes. In this respect, some authors have indicated 

that the increased toxicity of longer ILs can be accounted for 

enhanced membrane permeability altering the physical 

properties of the lipid bilayer (Lata1a et al., 2005; Ranke et al., 

2004; Stepnowski et al., 2004). Additionally, it has been 

proposed that the mode of toxic action for ILs takes place 

through membrane disruption because of the structural 

similarity of imidazolium-based ILs to detergent, pesticides 

and antibiotics able to cause membrane-bound protein 

disturbance (Docherty and Kulpa, 2005). Recently, Ranke et al. 

(2007a,b) have demonstrated that lipophilicity of ILs dominates 

their in vitro cytotoxicity over a wide range of structural 

variations. The contribution of the anionic part of the ILs to 

the observed biological effect was evaluated by comparing 

the EC 50 values obtained for the cations IM16 and IM18, 

combined with two different anions [Cl] and [PF 6] . For both 

cations, a stronger toxic effect was found for chloride derivatives, 

but not for fluoride containing hexafluorophosphate. 

A similar result was reported by Stock et al. (2004) where the 

inhibitory effects of IM14 Cl and IM14 PF6 on the acetylcholinesterase 

activity were compared. In addition, slightly 

higher cytotoxicity for the chloride derivative has also been 

observed when the cytotoxicity of IM14 Cl and IM14 PF6 on 

HeLa cells was tested (Stepnowski et al., 2004). This implies 

the effect of perfluorinated ions is not drastic to all but vary 

according to species of organisms tested. Several authors 

have pointed out that altering the anion has only minimal 

effects on the toxicity of several imidazolium compounds 

(Bernot et al., 2005a; Garcia et al., 2005; Ranke et al., 2004). 

This indicates that ILs toxicity seems to be related to the alkyl 

chain branching and to the hydrophobicity of the imidazolium 

cation but not to the various anions. In this respect, 

a recent study using the IPC-81 rat leukemia cell line with 

a large pool of anions demonstrated that most of the 

commercially available anions showed no or only marginal 

cytotoxic effects. However, anionic compartments with 

lipophilic and hydrolysable structural elements are likely to 

be of considerable relevance with respect to the toxicity of ILs 

(Stolte et al., 2006). 

In a recent study (Frade et al., 2007), the human cell lines 

such as HT-29 and CaCo-2 cells were utilized to estimate the 

inhibitory effect of ILs with several types of cations and 

anions. In both cells, IM14, IM12OH (1-(2-hydroxyethyl)-3methylimidazolium), 

IM12O2O1 (1-(2-(2-methoxyethoxy)ethyl)-3-methylimidazolium) 

and cholines were the least 

toxic cations independently of the anion. Within the studied 

combinations, it can be noted that IM14 PF 6, IM14 acesulfame, 

IM12OH BF 4/PF 6, IM12O2O1 BF 4/PF 6, IM12OH acesulfame and 

IM12OH saccharine are not toxic and present good alternatives 

to organic solvents. Meanwhile, increasing the length of 

the substituent chain may contribute to a significant 

increasing of imidazolium toxicity. It was also noted that 

[(CF3SO2)2N] anion decreased the toxicity to a large extent,

independently of the cation and for both cell types, which was 

in accordance with Salminen et al. (2007). 

2.5. Phytotoxicity of ILs 

The studies on phytotoxic activity of ILs were conducted 

mostly on the duckweed, Lemna minor, a common aquatic 

vascular plant (Jastorff et al., 2005; Larson et al., 2008; Matzke 

et al., 2007; Stolte et al., 2007a). In general, 1-alkyl-3methylimidazolium 

compounds with longer alkyl chains were 

more toxic to L. minor than those with short alkyl chain 

lengths. Imidazolium and pyridinium cations with butyl 

groups had similar EC50s (the concentrations that produced 

a 50% reduction in root growth) (39.07 and 32.54 mM, respectively); 

while the equivalent ammonium cation had a much 

higher EC50 (101.48 mM; i.e., less toxic) (Larson et al., 2008). In 

consideration of anionic effect, [(CF 3SO 2) 2N] was found to 

cause moderate toxicity to this duckweed (EC 50 ¼ 6300 mM) 

(Matzke et al., 2007). On the other hand, this anion had no or 

even a positive influence on the observed effects on L. minor 

(Stolte et al., 2007a). 

Focusing on the terrestrial environment, Matzke et al. 

(2009a) investigated the influence of differently composed 

soils, with varying contents of the clay minerals smectite and 

kaolinite, on the toxicity of different anion species of imidazolium-based 

ILs towards the wheat Triticum aestivum. The 

data showed that IM14 (CF3SO2)2N appeared the most toxic, 

independently of the type and concentration of added clay. 

This is totally in contrast to the findings of Stolte et al. (2007a), 

who reported that [(CF 3SO 2) 2N] caused no harm to L. minor, 

indicating the toxic effect of this anion is different between 

certain plants. The toxicity of 1-butyl-3-methylimidazolium 

incorporated chloride, tetrafluoroborate and hydrogen sulfate 

was mainly controlled by the cationic moiety. The observed 

effects varied according to the added clay type and clay 

concentration. An increase of clay content resulted in less 

inhibitory effects of these substances. On the contrary, for 

IM14 combined with bis(trifluoromethylsulfonyl)imide the 

addition of clay minerals led to higher toxicity compared to 

the reference soil. Since results are contradictious further 

study is necessary to unravel the underlying mechanism. 

Moreover, a detailed study on the effect of IM14 BF 4 on the 

wheat T. aestivum seedlings (Wang et al., 2009) showed that 

IM14 BF4 was hazardous to the early development of wheat 

and had varying effects on different organs. At low concentrations, 

IM14 BF4 did not inhibit, and even promoted, wheat 

seedling growth. Nonetheless, at high concentrations, this IL 

inhibited wheat seedling growth significantly and decreased 

chlorophyll content, thereby reducing photosynthesis and 

plant growth. Therefore, the authors suggested that dilution 

could decrease the toxicity of IM14 BF 4 to plants and would be 

a good method for remediating IL-polluted environments. 

In another research, the phytotoxicity tests of chiral ILs 

containing (-)-nopyl derivatives were carried out in a plant 

house using spring barley (Hordeum vulgare) which is a monocotyledonous 

plant, and a common radish (Raphanus sativus L. 

subvar. radicula Pers.) which is a dicotyledonous plant (Ba1czewski 

et al., 2007). According to the data obtained, increasing 

the concentration of ILs resulted in a systematic decrease in 

the crop fresh weight of total sprouts and the crop fresh 


weight per plant, both for spring barley and for common 

radish. It could also be noted that common barley was a more 

resistant plant which fairly well tolerates test IL concentrations 

up to 200 mg kg 1 of soil; whereas, for radish, the growth 

and development inhibiting concentration is 100 mg kg 1 of 

soil. Using the same target plant (H. vulgare), Pernak et al. 

(2004b) reported that the 1,3-dialkoxymethylimidazolium 

tetrafluoroborate salts introduced to the soil at concentration 

of 1,000 mg kg 1 , or 100 mg kg 1 dry mass of soil, were found 

to exert a phytotoxic effect on monocotyledonous plants. 

On the other hand, at a concentration of 10 mg kg 1 no such 

effect on the growth of the roots was notified. Concerning 

phytotoxicity of ILs to garden cress (Lepidium sativum L.) in soil 

environment, Studzińska and Buszewski (2009) have proved 

that hazardous effects of imidazolium ILs are closely 

connected with organic matter content in soil. Soil with more 

organic carbon was observed to sorb IL cations more extensively 

than soil with little or no organic matter; hence, the 

more fertile in soil, the lower probability of hazardous effect of 

ILs to plants. On the other hand, the hazardous character of 

analyzed ILs was strongly connected with their hydrophobicity, 

indicating that the more hydrophobic IL, the higher 

decrease of seed germination. 

Although intensive work has not been conducted on 

phytotoxic influence of ILs, the available data offer initial 

hints for environmental scientists dealing with the potential 

impact of ILs towards aqueous and terrestrial plants. 

2.6. Toxicity of ILs to invertebrates 

Ecotoxicological literature of ILs to invertebrates mainly focus 

on the use of Daphnia magna as a test organism (Bernot et al., 

2005a; Couling et al., 2006; Garcia et al., 2005; Grabinska-Sota 

and Kalka, 2006; Luo et al., 2008; Nockemann et al., 2007; Pretti 

et al., 2009; Samorì et al., 2007; Wells and Coombe, 2006; Yu 

et al., 2009b). Daphnia is an important link between microbial 

and higher trophic levels (McQueen et al., 1986), and has been 

the subject of hundreds of intensive ecological studies. The 

results of all studies again observed the well-established link 

between toxicity and alkyl chain length of the tested ILs 

containing imidazolium, pyridinium or quaternary ammonium 

as counter cations. The most toxic compound towards 

D. magna was found to be IM18 Br whereas the least toxic one 

was IM14 Cl with log10EC50 values of 1.33 and 1.93, respectively 

(Table 2). Also, the nature of the anion was suggested to 

have smaller effects compared to those of the cation. In a 

recent study, Luo et al. (2008) investigated the developmental 

toxicity of IM18 Br on D. magna. It was found that this 

compound exhibited toxicity on the development of three 

generation of D. magna with the decrease of number of 

offspring and average brood size correlated to increasing IM18 

Br concentrations. This indicated that IM18 Br could cause 

deleterious effect to the population of Daphnia and indirectly 

disturb freshwater food webs. Couling et al. (2006) used 

experimental data to determine which part of the IL molecule 

is responsible for the observed toxic effects through a quantitative 

structure-property relationship (QSPR) modeling. In this 

respect, correlative and predictive equations were generated 

and proved that there was a distinct influence of the length of 

alkyl residues attached to the aromatic nitrogen atoms

364 

towards D. magna. Moreover, the models predicted that the 

toxicity increased slightly with increasing number of aromatic 

nitrogen atoms in cation ring. This implies that ammonium 

salts are less toxic than pyridinium salts, which in turn are 

less toxic than imidazolium moieties. Interestingly, it was 

noted that methylating the aromatic carbons could be 

effective in reducing toxicity to D. magna, indicating that 1-nbutylpyridinium 

bromide can be more toxic than 1-n-butyl-3methylpyridinium 

bromide, which is more toxic than 1-nbutyl-3,5-dimethylpyridinium 

analogue. The QSPR, though is 

still in its infancy, has contributed initial guidelines for 

a rationale design of a new category of ILs with an acceptable 

environmental profile. 

Other studies include data on the snail Physa acuta (Bernot 

et al., 2005b), the spring tail Folsomia candida (a soil invertebrate) 

(Matzke et al., 2007), Caenorhabditis elegans (a soil 

roundworm) (Swatloski et al., 2004) and Dreissena polymorpha 

(zebra mussel) (Costello et al., 2009). It was also demonstrated 

a positive relationship between alkyl chain length and toxicity 

in these reports. In the research of Bernot et al. (2005b), the 

estimated LC 50 (median lethal concentration) ranged from 

3.50 to 1799.8 mM (0.54 to 3.26 in the logarithmic form), which 

implied that P. acuta are less sensitive to ILs than are D. magna 

(log10LC50 (mM) ranging from 1.33 to 1.93 (Table 2)). Also, the 

authors observed that at low concentrations, the IL may 

suppress snail movement, but concentrations above this 

threshold level trigger an escape response, causing the 

organism to move faster. Grazing patterns, nonetheless, 

showed that snails grazed less at higher IL concentrations. 

Physa spp. are key components of freshwater food webs, 

because they graze algae and are themselves important prey 

for fish and invertebrate predators (Bernot and Turner, 2001; 

Osenberg and Mittelbach, 1989). Thus, nonlethal IL concentrations 

affected P. acuta behaviors, potentially influencing 

individual fitness and good web interactions. 

2.7. Inhibitory effects of ILs on vertebrates 

Zebrafish (Danio rerio) plays an important role in ecotoxicology 

as a prominent model vertebrate. Concerning toxicity of ILs to 

the zebrafish, Pretti et al. (2006) revealed that ILs may cause 

a completely different effect on fish according to their chemical 

structures. As imidazolium, pyridinium and pyrrolidinium 

showed a LC50 (lethal effect) >100 mg L 1 , they could be 

regarded as non-highly lethal towards zebrafish. On the other 

hand, the ammonium salts showed LC50 remarkably lower 

than that reported for organic solvents and tertiary amines. 

In general, these data referred that fish are less sensitive to ILs 

toxicity compared to other species belonging to lower trophic 

levels. 

In a recent report, Li et al. (2009) used the frog Rana nigromaculata 

as an amphibian model for toxicity testing. 

Amphibians are often the main vertebrate group prone to 

contaminant exposure in aquatic systems mostly because 

their larvae live in water (Lahr, 1997; Mann and Bidwell, 2000). 

In their study, they evaluated the toxic effects of IM18 Br on 

the early embryonic development of the frog R. nigromaculata. 

The results demonstrated that the highest embryonic 

mortality occurred in the neural plate stage, followed by the 

early gastrula and early cleavage stages with the LC50 values 


being 42.4, 43.4 and 85.1 mg/L, respectively, indicating that the 

developmental toxicity of IM18 Br in the frog was stagesensitive. 

The number of dead embryos was also found to 

increase with the increasing concentrations of the IL IM18 Br. 

The developmental impact of IM18 Br was claimed not only in 

this finding but also in the work of Luo et al. (2008), who 

investigated on D. magna. 

Other work in the literature has focused on the acute 

toxicity of ILs on rats and mice (Bailey et al., 2008; Cheng et al., 

2009; Landry et al., 2005; Pernak and Czepukowicz, 2001; Sipes 

et al., 2008). The values of acute toxicity of 3-hexyloxymethyl- 

1-methylimidazolium tetrafluoroborate were found to be 

LC50 ¼ 1400 and 1370 mg kg 1 for female and male Wistar rats, 

respectively (Pernak and Czepukowicz, 2001). Bailey et al. 

(2008) studied the effects of prenatal exposure of mice to IM14 

Cl due to the potential for human exposure as a result of water 

or soil contamination from industrial effluent or accidental 

spills. As shown in the experimental data, after being contacted 

to the IL, fetal weight was considerably reduced at the 

two highest concentrations (169 and 225 mg kg 1 d 1 ). 

Malformations were also somewhat more numerous at the 

highest dosage, suggesting that IM14 Cl may be teratogenic. 

Maternal toxicity was also present, indicating that IM14 Cl 

appeared to be developmentally toxic at maternally toxic 

dosages. Also, IM14 Cl has been shown to cause thermal irritation 

when applied topically to rats, but produced only 

minimal contact sensitization when evaluated in the mouse 

local lymph node assay (Landry et al., 2005). Additionally, in 

this report, it is worth noting that the transdermal toxicity of 

IM14 Cl was influenced by the vehicle of administration. Use 

of the organic solvent, dimethylformamide, accentuated the 

acute toxicity. Very high concentrations of IM14 Cl (up to 95% 

IM14 Cl in water) applied to the rat skin were markedly less 

acutely toxic. This result may have a practical guideline that to 

reduce the acute toxicity, ILs can be handled in pure form with 

water as a co-solvent. 

3. Environmental fate of ILs 

3.1. Chemical degradation of ILs 

Ionic liquids possess excellent chemical and thermal stability, 

which gives, unfortunately, a negative aspect for their treatment 

after usage prior to disposal. To assess the persistence of 

ILs in the environment as well as verify possibilities of their 

cleanup by chemical methods, several groups have focused 

their attention on oxidative and thermal degradation of ILs in 

aqueous media (Awad et al., 2004; Baranyai et al., 2004; 

Berthon et al., 2006; Itakura et al., 2008; Li et al., 2007; Morawski 

et al., 2005; Siedlecka and Stepnowski, 2009; Siedlecka 

et al., 2008a,b; Stepnowski and Zaleska, 2005). Pioneering work 

in the field of oxidative degradation was done by Stepnowski 

and Zaleska (2005) and Morawski et al. (2005) who showed that 

the greatest degradation efficiency for imidazolium ILs was 

achieved with a combination of UV light and a catalytic 

oxidant such as hydrogen peroxide or titanium dioxide. 

Subsequently, Li et al. (2007) studied the oxidative degradation 

of 1,3-dialkylimidazolium ILs in hydrogen peroxide/acetic acid 

medium assisted by ultrasonic chemical irradiation. It was

observed that 99% of tested compounds was degraded after 

72 h. In addition, advanced oxidative degradation in the 

presence of reactive peroxides generated by Fenton reagent 

has been applied for the removal of ILs from water (Siedlecka 

and Stepnowski, 2009; Siedlecka et al., 2008a,b). According to 

the results, in a Fenton system with 1 mM of Fe(III) and 

100 mM of H2O2, more than 97% of IM14 Cl was observed to 

degrade after 90 min. For Pyr14 Cl, IM16 Cl and IM18 Cl, the 

levels of degradation were 92%, 88% and 68%, respectively. 

Investigations of the degradation mechanisms indicated IM18 

Cl was more resistant to oxidation by OH radicals cleaved 

from H 2O 2, suggesting that the oxidation rates of imidazolium 

ILs by OH are structure-dependent (Siedlecka and Stepnowski, 

2009). The level of degradation was dependent on the alkyl 

chain length, consistent with Stepnowski and Zaleska (2005), 

who indicated that lengthening the alkyl chain lowered the 

rate of IL degradation. On contrast, the different length of the 

side chains and the type of anions did not affect the degradation 

process (Li et al., 2007). Regarding the thermal degradation 

studies of alkylimidazolium salts (Awad et al., 2004), 

extension of the alkyl chain enhanced the thermo-oxidative 

degradation of imidazolium salts. Interestingly, methyl 

substitution in the 2-position (i.e. between the two N atoms) 

was observed to decrease the oxidative decomposition of 

imidazolium ILs. The longer alkyl chain was also observed to 

induce an enhancement in photocatalytic decomposition of 

ILs (Morawski et al., 2005), which was not in the case of 

oxidative degradation (Stepnowski and Zaleska, 2005; Siedlecka 

and Stepnowski, 2009). Nonetheless, detailed account on 

the degradation of ILs by photocatalysis is required to verify 

this phenomenon. 

3.2. Biodegradability of ILs 

In contrast to chemical degradation, which requires the 

assistance of a certain oxidant for catalysis, biodegradation is 

the microbial breakdown of chemical compounds. Biodegradation 

seems to be more environmentally friendly compared 

to chemical decomposition process. The initial attempt to 

examine the degradation potential of different IM14 cations 

combined with [Br] , [BF 4] , [PF 6] , [N(CN) 2] , [(CF 3SO 2) 2N] 

and octylsulfate as the counter ion was done using the Sturm 

and Closed-Bottle test protocols by the group of Scammells 

(Garcia et al., 2005; Gathergood and Scammells, 2002; Gathergood 

et al., 2004; Gathergood et al., 2006). Nonetheless, no 

compound showed significant degree of biodegradation with 

the exception of the octylsulfate-containing IL. The next step 

study on the biodegradation of ILs involved the design of ILs 

containing biodegradable side chains (Gathergood and 

Scammells, 2002). The design was done according to the 

principles of Boethling (Boethling, 1994, 1996; Howard et al., 

1991) who identified three important parameters including 

the potential sites of enzymatic hydrolysis (for example, 

esters and amides) and oxygen in form of hydroxyl, aldehyde 

or carboxylic acid groups as well as unsubstituted linear alkyl 

chains (especially 4 carbons) and phenyl rings, which 

represent possible sites for attack by oxygenases. However, 

for a balance between chemical properties and biodegradability, 

not all of these factors were suitable for ILs. The 


addition of oxygen containing functional groups such as 

alcohols, aldehyde and carboxylic acids was reported to 

restrict the ILs performance as reaction media whereas the 

incorporation of phenyl rings was known to increase the 

melting points of IL solvents (McGuinness and Cavell, 2000). 

Therefore, ester or amide group was selected to be coupled in 

alkyl side chain of ILs. The introduction of ester groups 

derived from a C2 acid and C4 or higher alcohol in the 

3-N-substitutent was demonstrated to increase the biodegradation 

of imidazolium-based ILs (Gathergood et al., 2004). This 

can be explained by the fact that introduction of ester moiety 

probably provides a site susceptible to enzymatic attack 

(Gathergood et al., 2004; Gathergood et al., 2006) and hence, 

improves the biodegradation level. Though the addition of 

amide group is informed to improve the biodegradation of 

organic compounds (Boethling, 1994, 1996; Howard et al., 

1991), no critical enhancement of biological degradation was 

noted when this group was appended into the imidazoliumbased 

ILs (Gathergood et al., 2004). However, no compound 

could be classified as ‘‘readily biodegradable’’ corresponding 

to Organization for Economic Cooperation and Development 

(OECD) standards (U.S. EPA, 1998), for which 60–70% or greater 

biodegradation by activated sludge microbial inoculate is 

required within a 10-day window in a 28-day period. Finally, 

the combination of the octylsulfate anion and imidazolium 

cation containing ester side chains resulted in readily biodegradable 

IL (Gathergood et al., 2006). Recently, Stolte et al. 

(2008) also paid their attention on investigation of functional 

groups incorporating alkyl chain ILs. Nonetheless, the introductions 

of terminal hydroxyl, carboxyl, ether and nitrile 

groups did not improve the biological degradation as 

expected. 

Kumar et al. (2006) investigated the fate of IM14 BF 4 when 

in contact with soil-microorganisms, wastewater microorganisms, 

Pseudomonas putida and E. coli. Although IM14 BF4 

was indicated to be recalcitrant in Sturm and Closed-Bottle 

test assays as mentioned above, it was observed in this study 

that P. putida was able to break down IM14 BF4 after 15 days of 

incubation. The breakdown products were monitored 

using GC-MS and identified to be 1-H-methylimidazole and 

1-H-butylimidazole, which were in consistent with the 

theoretical metabolism scheme proposed by Jastorff et al. 

(2003). In case of bacteria from soil and wastewater, the 

metabolic intermediates appeared on the 12th day. It was 

also noted that different intermediate peaks were observed at 

different retention time with different microbes, indicating 

that the degradation mechanism of IM14 BF4 may vary in 

correspondence to certain microbes and metabolic pathways. 

In another study, the biodegradation pathway of IM18 moiety 

(Fig. 3) was proposed based on intermediate products via 

HPLC-MS analysis after 24-day period of incubation with 

activated sludge (Stolte et al., 2008). The metabolism of IM18 

cation appeared to undergo oxidation reactions catalyzed 

probably by mono-oxygenases, e.g. the cytochrome P 450 

system on the terminal methyl group (u-oxidation). The 

alcohol formed was subsequently oxidized and converted 

into aldehydes, and then into carboxylic acids by dehydrogenases. 

The resulting carboxylic acids then might undergo 

b-oxidation and finally generated two carbon fragments that 

can enter the tricarboxylic acid cycle as acetyl Co-A (Fig. 3).

366 

retention time 

in min. 

8.9 

13.5 / 14.4 

12.2 / 12.7 

10.0 – 12.5 

19.5 

16.2 

26.5 

24.2 

26.7 

m/z + 

195 

211 

209 

225 

183 

197 

155 

169 

141 

intensity 

4*10 5 

3*10 5 / 2*10 5 

1*10 6 / 2*10 6 

2*10 4 – 6*10 4 

1*10 5 

3*10 5 

0.5*10 5 

4*10 6 

1*10 5 

N N + 

N N + 

N N + 

N N + 

N N + 

N N + 

N N + 

N N + 

O 

OH 

The proposed pathways provide basic information to both 

environmental scientists and chemical engineers; however, 

no studies have sought to examine the toxicity of metabolic 

products after degradation of ILs. This issue is of paramount 

importance since metabolism might not always end in less 

toxic products. 

Subsequently, Wells and Coombe (2006) extended the 

microbial degradation study with ammonium, imidazolium, 

phosphonium and pyridinium compounds by measuring the 

biological oxygen demand. The authors observed no biodegradability 

of cations incorporated short chains (C 4) within 

this test series, which was in agreement with Docherty et al. 

(2007) and Stolte et al. (2008). For longer alkyl chains (C 12, C 16 

and C 18) containing ILs, a strong inhibitory effect of these 

compounds on the inoculum used was found, indicating the 

active microbial consortium was significantly impacted by ILs 

toxicity. In recent studies (Docherty et al., 2007; Grabinska- 

Sota and Kalka, 2004; Harjani et al., 2008; Stasiewicz et al., 

2008), pyridinium-based ILs were reported to be fully catabolized 

by microbial community in activated sludge. This can be 

inferred from the fact that degradation pathways for pyridine – 


OH 

OH 

O 

OH 

O 

OH 

H 

N N + 

O 

O 

OH 

OH 

N N + 

N N + 

N N + 

the precursor of pyridinium-based compounds – under 

aerobic and anaerobic conditions were intensively investigated 

in the work of Kaiser et al. (1996). With respect to the 

common 1,3-dialkylpyridinium ILs, Pham et al. (2009) reported 

that after 21 days of incubation, microorganisms from activated 

sludge were able to break down Py4-3Me Br. Analyses of 

HPLC and MS/MS demonstrated that this biodegradation 

led to the formation of 1-hydroxybutyl-3-methylpyridinium, 

1-(2-hydroxybutal)-3-methylpyridinium, 1-(2-hydroxyethyl)- 

3-methylpyridinium and methylpyridine. Based on these 

intermediate products, biodegradation pathways were also 

suggested (Fig. 4), thereby providing the basic information 

which might be useful for assessing the factors related to the 

environmental fate and behavior of this commonly used 

pyridinium IL. Although this is the first report on biodegradation 

intermediates and pathway of pyridinium ILs, the 

authors have failed to systematically screen a single microorganism 

or a microbial consortium responsible for biodegradation 

of ILs. Therefore, it is needed to further investigate 

which type of microorganism is adaptable to ILs and which is 

responsible for degradation process. Also, the broken 

OH 

O 

OH O 

+ 

Fig. 3 – Biodegradation pathways of 1-octyl-3-methylimidazolium by activated sludge microbial community (Reproduced 

from Stolte et al. (2008) by permission of the Royal Society of Chemistry). 

m/z + 

211 

209 

225

C 

H 3 

C 

H 3 

N + 

N + 

C 

H 3 

structures much further than methylpyridinium were not 

measured in the study. In particular, the possibility of 

cleavage of heterocyclic ring in ILs molecular structure (both 

pyridinium in this work and imidazolium in the study of Stolte 

et al. (2008)) has not been indentified. This issue should be 

clarified through further studies. Interestingly, Py4-3Me Br 

was not found to be metabolized by the activated sludge 

community (Docherty et al., 2007). This was attributed likely 

to the high IL concentration used in the study of Docherty’s 

group, which consequently inhibited the microbial consortium. 

Nonetheless, it was demonstrated that the structural 

manipulation of the pyridinium skeleton may lead to ILs with 

greater biodegradable extent compared to imidazolium-based 

compouds (Harjani et al., 2008). 

Concerning the anionic effect, ILs with halide counter ions 

were postulated to be more stable to degradation than 

perfluoronated ions (Awad et al., 2004; Gathergood and 

Scammells, 2002). In a preliminary study, Gathergood and 

Scammells (2002) confirmed this assumption and showed that 

the biodegradation efficiency decreased in the order 

[PF 6] > [BF 4] > [Br] with 60%, 59% and 48% of CO 2 evolution 

values, respectively. In a later study (Gathergood et al., 2006), 

it was found that the octylsulfate anion is considerably more 

biodegradable than the other commonly used anions. 

The alkyl chain with C4, C6 or C8 was found to increase the 

rate of degradation (Docherty et al., 2007; Stolte et al., 2008); 

nonetheless, further increasing the chain length to C12, C16 or 

C18 was noted to cause toxic effect towards inoculum (Wells 

and Coombe, 2006). However, it was stated that the long octyl 

side chain was not a compulsory factor for biodegradation, but 

more important is a certain overall lipophilicity of the 

compound (Stolte et al., 2008). 

I 

C 

H 3 

C 

H 3 

OH 

H 

+ 

+ 

N + 

N + 

N + 


OH 

C 

H 3 

C 

H 3 

CH 3 

OH 

O 

O 

OH 

II 

C 

H 3 

N + C H3 CH3 OH 

N + 

OH 

+ C 

H 3 

Fig. 4 – Biodegradation pathways of 1-butyl-3-methylpyridinium entity by microorganisms from activated sludge (Reprinted 

with permission from Pham et al. (2009). Copyright 2009 American Chemical Society). 

C 

H 3 

N + 

H 

CH 3 

3.3. Sorption of ILs in environmental systems 

+ 

C 

H 3 

Because the aquatic and terrestrial environments are possible 

recipients for contaminants, the distribution and behavior of 

ILs in soil are also extremely important. The retention and 

mobility of ILs in soils and sediments are strongly influenced by 

its tendency to be sorbed onto the various components of the 

soil matrix (Stepnowski, 2005). Since imidazolium-based ILs 

possess high electron acceptor potential of delocalized 

aromatic systems and hydrophobic components (e.g. the alkyl 

chain) (Stepnowski, 2005), they could be sorbed onto soils and 

sediments via several mechanisms. In a preliminary study, 

Stepnowski (2005) proved that electrostatic interactions 

contributed to the sorption of the imidazolium cations. Moreover, 

totally contrast trends were also observed demonstrating 

an extremely strong and practically irreversible sorption onto 

fine-textured marine sediments and a relatively weakly and 

reversibly binding to peaty soil (with the highest organic 

carbon content) (Stepnowski, 2005). This indicates the importance 

of the mineral component of the soil (sediment) in the 

sorption mechanism of ILs. Also, in this work, the author 

pointed out that compounds with longer alkyl chains were 

irreversibly bound to the soil component, which was in 

agreement with Stepnowski et al. (2007) and Matzke et al. 

(2009b). Interestingly, Beaulieu et al. (2008) did not find a positive 

effect of alkyl chain length on the sorption of alkylimidazolium-based 

ILs to aquatic sediments and suggested that 

hydrophobic interactions were not the most important sorption 

mechanism. The contrast results between these groups 

imply that sorption mechanisms of ILs may vary according to 

properties and composition of the environmental systems. The 

studies suggest that ILs may be retained by aquatic sediments; 

OH

368 

nonetheless, the toxic action of these sorbed materials towards 

aquatic organisms has not been addressed. Also, further 

efforts should be continued to elucidate the reversibility of IL 

sorption on these sediments to give a better understanding of 

the ILs fate after being released into aqueous environment. 

Matzke et al. (2009b) investigated the influences of the two 

different clay minerals kaolinite and smectite as well as of 

organic matter on the cation sorption and desorption behaviors 

of three imidazolium based ILs including IM14 BF 4, IM18 

BF 4 and IM14 (CF 3SO 2) 2N in soil. The addition of organic matter 

and clays was observed to increase the sorption/decrease the 

desorption of all ILs tested, and in particular smectite had 

striking effects on the sorption efficiency of all substances. 

It is worth noting that not only the cationic moiety with 

different alkyl side chain lengths but the anionic compartments 

can also play an importance role in the sorption/ 

desorption processes. Imidazolium compounds with BF4 as 

a counter anion showed higher sorption capacity compared to 

that of [(CF3SO2)2N] , indicating the high potential of this type 

of IL to form ionic pairs in the soil matrix (Matzke et al., 2009b). 

However, further work with a variety of ILs incorporated 

different cationic and anionic moieties should be carried out 

to verify these phenomena. 

Gorman-Lewis and Fein (2004) examined the sorption 

behavior of IM14 Cl onto a range of surfaces which are 

commonly found in the near-surface environment. The results 

suggested that IM14 Cl could be minimally retarded by noninterlayer 

clay system and might lead to unimpeded transport 

through subsurface groundwater. Also, the adsorption 

capacity of this IL onto bacterial surfaces was not high, which 

might be due to the low hydrophobicity of IM14 Cl. Additionally, 

investigations of the adsorption of IM14 Cl towards 

different media carried out by our group (Vijayaraghavan et al., 

2009) have shown that retardation of this compound was 

possible only by an ion-exchange resin and activated carbon, 

which was in consistency with the work of Anthony et al. 

(2001). However, no significant adsorption of IM14 Cl in the 

media of a fermentation waste (Corynebacterium glutamicum) 

and dried activated sludge was observed in our study. 

Conclusively, the data currently available demonstrated 

that ILs incorporated imidazolium cation can be sorbed to 

organic matter, whether found in aquatic sediments or 

terrestrial soils, and the presence of clays significantly 

enhanced the sorption capacity. 

4. Concluding remarks and future directions 

Ionic liquids, of which the most often cited attribute is their 

negligible vapor pressure, have been suggested as a green 

alternative to traditional organic solvents with the desire to 

minimize diffusion to the atmosphere. Low volatility, however, 

does not completely eliminate potential environmental 

hazards and might pose serious threats to aquatic and terrestrial 

ecosystems. The studies of environmental fate and 

toxicity of ILs have shown that the ILs commonly used to date 

are toxic in nature and their toxicities vary considerably across 

organisms and trophic levels. In general, the effect of anionic 

moieties is not drastic as the alkyl length effect except for the 

case of [(CF3SO2)2N] , which shows a clear (eco)toxicological 


hazard potential. The other perfluorinated anions have been 

also proved to be hazardous due to hydrolytically unstable 

properties. In addition, the introduction of functional polar 

groups to the alkyl chain has been shown to reduce the toxicity 

of ILs and increase the biodegradation efficiency to some 

extent. This indicates the possibility of tailoring ILs by coupling 

suitable functional groups to their structure, which in turn 

leads to a more environmental friendly compound. The side 

chain length effect has been found to be consistent in all levels 

of biological complexity as well as different environmental 

compartments. Also, an increase in alkyl-chain length, or lipophilicity, 

was observed to be related to an increase in the rate 

of degradation as well as an increase in toxicity. This indicates 

a conflict of aims between minimizing the toxicity and maximizing 

the biodegradability of these neoteric solvents. 

Regarding the cationic compartment, pyridinium has been 

found to be more environmental friendly than imidazolium 

from both viewpoints of toxicology and microbial degradation. 

It can therefore be suggested that the structural manipulation 

of the pyridinium skeleton should be considered in design of 

a sustainable IL. From the currently available data, it is clear 

that some commonly used ILs are very far away from the image 

of green chemicals that are often cited in the literature. The 

uncertainties in their sustainable development hinder the 

applications of ILs under real conditions. Although some 

attempts have been made to give important hints in the 

prospective design and synthesis of inherently safer ILs, 

comprehensive studies dealing with the behaviors of ILs in 

aqueous media still await to be conducted. The important 

features required for the thorough insight into environmental 

fate of ILs include, but are not limited to: 

- Providing more fundamental understanding into the 

mechanism for IL-induced toxicity to different levels of 

biological complexity. The underlying mechanisms of IL 

toxicity have rarely been studied. 

- Assessing the biodegradability of cationic and anionic 

compartments and toxicity of their degradation intermediates. 

This may provide useful information in consciously 

designing safer chemicals. 

- Investigating the aerobic and anaerobic biodegradation of 

ILs, which would suggest initial guidelines for the treatment 

of ILs waste by using the existing aerobic and anaerobic 

wastewater treatment facilities. Especially, anaerobic 

degradation awaits to be investigated. 

- Defining which organisms or enzymes may promote 

degradation pathways and determining specific microbial 

consortium or cultivatable communities capable of 

biotransformation of ILs. 

- Performing the ecotoxicity and biodegradation tests in real 

environmental conditions instead of controlled conditions 

of laboratory experiments, which would be advantageous in 

understanding the fate and behavior of ILs under real 

conditions. For this, the potential toxicological effects at 

population level and community level should be addressed. 

It must be encouraged to use tools such as experimental 

mesocosms to study the effects of ILs at higher levels of 

organization. 

- Creating database of environmentally benign structure 

moieties of ILs based upon their toxicological and

iodegradation information, which would be practically 

useful as a reference for manufacturers and regulators to 

properly develop and regulate the use of ILs. 

Acknowledgements 

This work was supported by NRF Grant funded by the Korean 

Government (KRF-2007-521-D00106, NRL 2009-0083194, and in 

part WCU R31-2008-000-20029-0). 

references 

Anthony, J.L., Maginn, E.J., Brennecke, J.F., 2001. Solution 

thermodynamics of imidazolium-based ionic liquids and 

water. J. Phys. Chem. B 105, 10942–10949. 

Arning, J., Stolte, S., Böschen, A., Stock, F., Pitner, W.-R., Welz- 

Biermann, U., Jastorff, B., Ranke, J., 2008. Qualitative and 

quantitative structure activity relationships for the inhibitory 

effects of cationic head groups, functionalized side chains and 

anions of ionic liquids on acetylcholinesterase. Green Chem. 

10, 47–58. 

Austin, R.P., Barton, P., Davis, A.M., Manners, C.N., Stansfield, M.C., 

1998. The effect of ionic strength on liposome-buffer and 

1-octanol-buffer distribution coefficients. J. Pharm. Sci. 87, 

599–607. 

Awad, W.H., Gilman, J.W., Nyden, M.,Harris,R.H.,Sutto,T.E., 

Callahan, J., Trulove, P.C., DeLong, H.C., Fox, D.M., 2004. 

Thermal degradation studies of alkyl-imidazolium salts and 

theirapplicationinnanocomposites. Thermochim. Acta 

409, 3–11. 

Babalola, G.O., 1998. Anti-bacterial activity of synthetic 

N-heterocyclic oxidizing compounds. Lett. Appl. Microbiol. 26, 

43–46. 

Bailey, M.M., Townsend, M.B., Jernigan, P.L., Sturdivant, J., Hough- 

Troutman, W.L., Rasco, J.F., Swatloski, R.P., Rogers, R.D., 

Hood, R.D., 2008. Developmental toxicity assessment of the 

ionic liquid 1-butyl-3-methylimidazolium chloride in CD-1 

mice. Green Chem. 10, 1213–1217. 

Ba1czewski, P., Bachowska, B., Bia1as, T., Biczak, R., Wieczorek, W. 

M., Balińska, A., 2007. Synthesis and phytotoxicity of new 

ionic liquids incorporating chiral cations and/or chiral anions. 

J. Agric. Food Chem. 55, 1881–1892. 

Baranyai, K.J., Deacon, G.B., MacFarlane, D.R., Pringle, J.M., Scott, J.L. 

, 2004. Thermal degradation of ionic liquids at elevated temperatures. 

Aust. J. Chem. 57, 145–147. 

Beaulieu, J.J., Tank, J.L., Kopacz, M., 2008. Sorption of imidazoliumbased 

ionic liquids to aquatic sediments. Chemosphere 70, 

1320–1328. 

Bernot, R.J., Turner, A.M., 2001. Predator identity and traitmediated 

indirect effects in a littoral food web. Oecologia 129, 

139–146. 

Bernot, R.J., Brueseke, M.A., Evans-White, M.A., Lamberti, G.A., 

2005a. Acute and chronic toxicity of imidazolium-based 

ionic liquids on Daphnia magna. Environ. Toxicol. Chem. 24, 

87–92. 

Bernot, R.J., Kennedy, E.E., Lamberti, G.A., 2005b. Effects of ionic 

liquids on the survival, movement, and feeding behavior of 

the freshwater snail, Physa acuta. Environ. Toxicol. Chem. 24, 

1759–1765. 

Berthon, L., Nikitenko, S.I., Bisel, I., Berthon, C., Faucon, M., 

Saucerotte, B., Zorz, N., Moisy, Ph, 2006. Influence of gamma 


irradiation on hydrophobic room-temperature ionic liquids 

[BuMeIm] PF 6 and [BuMeIm] (CF 3SO 2) 2N. Dalton Trans, 2526–2534. 

Blaauboer, B.J., Balls, M., Barratt, M., Casati, S., Coecke, S., 

Mohamed, M.K., Moore, J., Rall, D., Smith, K.R., Tennant, R., 

Schwetz, B.A., Stokes, W.S., Younes, M., 1998. Alternative 

testing methodologies and conceptual issues. Environ. Health 

Perspect 106, 413–418. 

Blaise, C.R., 1993. In: Richardson, M.L. (Ed.), Ecotoxicology 

Monitoring. VCH, Weinheim, pp. 83–108. 

Blanchard, L.A., Hancu, D., Beckman, E.J., Brennecke, J.F., 1999. 

Green processing using ionic liquids and CO 2. Nature 399, 28–29. 

Blanchard, L.A., Brennecke, J.F., 2001. Recovery of organic 

products from ionic liquids using supercritical carbon dioxide. 

Ind. Eng. Chem. Res. 40, 287–292. 

Boethling, R.S., 1994. Cationic Surfactants, Surfactant Science 

Series, vol. 53. Marcel Dekker, New York, pp. 95–135. 

Boethling, R.S., 1996. Designing safer chemicals. ACS Symp. Ser. 

640, 156. 

Chemnitius, J.-M., Sadowski, R., Winkel, H., Zech, R., 1999. 

Organophosphate inhibition of human heart muscle 

cholinesterase isoenzymes. Chem. Biol. Interact 120, 183–192. 

Cheng, Y., Wright, S.H., Hooth, M.J., Sipes, I.G., 2009. 

Characterization of the disposition and toxicokinetics of 

N-butylpyridinium chloride in male F-344 rats and female 

B6C3F1 mice and its transport by organic cation transporter 2. 

Drug Metab. Dispos. 37, 909–916. 

Cho, C.-W., Pham, T.P.T., Jeon, Y.-C., Vijayaraghavan, K., Choe, W.-S., 

Yun, Y.-S., 2007. Toxicity of imidazolium salt with anion bromide 

to a phytoplankton Selenastrum capricornutum:effectofalkylchain 

length. Chemosphere 69, 1003–1007. 

Cho, C.-W., Pham, T.P.T., Jeon, Y.-C., Yun, Y.-S., 2008a. Influence of 

anions on the toxic effects of ionic liquids to a phytoplankton 

Selenastrum capricornutum. Green Chem. 10, 67–72. 

Cho, C.-W., Jeon, Y.-C., Pham, T.P.T., Vijayaraghavan, K., Yun, Y.-S., 

2008b. The ecotoxicity of ionic liquids and traditional organic 

solvents on microalga Selenastrum capricornutum. Ecotoxicol. 

Environ. Saf. 71, 166–171. 

Cho, C.-W., Pham, T.P.T., Jeon, Y.-C., Min, J., Jung, H.Y., Lee, D.S., 

Yun, Y.-S., 2008c. Microalgal photosynthetic activity 

measurement system for rapid toxicity assessment. 

Ecotoxicology 17, 455–463. 

Cieniecka-Ros1onkiewicz, A., Pernak, J., Kubis-Feder, J., Ramani, A., 

Robertson, A.J., Seddon, K.R., 2005. Synthesis, anti-microbial 

activities and anti-electrostatic properties of phosphoniumbased 

ionic liquids. Green Chem. 7, 855–862. 

Costello, D.M., Brown, L.M., Lamberti, G.A., 2009. Acute toxic effects 

of ionic liquids on zebra mussel (Dreissena polymorpha) surviving 

and feeding. Green Chem. 11, 548–553. 

Couling, D.J., Bernot, R.J., Docherty, K.M., Dixon, J.K., Maginn, E.J., 

2006. Assessing the factors responsible for ionic liquid toxicity 

to aquatic organisms via quantitative structure-property 

relationship modeling. Green Chem. 8, 82–90. 

Dipeolu, O., Green, E., Stephens, G., 2008. Effects of watermiscible 

ionic liquids on cell growth and nitro reduction using 

Clostridium sporogenes. Green Chem. 11, 397–401. 

Docherty, K.M., Kulpa, C.F., 2005. Toxicity and antimicrobial 

activity of imidazolium and pyridinium ionic liquids. Green 

Chem. 7, 185–189. 

Docherty, K.M., Dixon, J.K., Kulpa, C.F., 2007. Biodegradability of 

imidazolium and pyridinium ionic liquids by an activated 

sludge microbial community. Biodegradation 18, 481–493. 

Earle, M.J., Seddon, K.R., 2000. Ionic liquids. Green solvents for the 

future. Pure Appl. Chem. 72, 1391–1398. 

Frade, R.F.M., Matias, A., Branco, L.C., Afonso, C.A.M., Duarte, C. 

M.M., 2007. Effect of ionic liquids on human colon carcinoma 

HT-29 and CaCo-2 cell lines. Green Chem. 9, 873–877. 

Fuler, J., Carlin, R.T., Osteryoung, R.A., 1997. The room temperature 

ionic liquids 1-ethyl-3-methylimidazolium tetrafluoroborate:

370 

electrochemical couples and physical properties. J. 

Electrochem. Soc. 144, 3881–3886. 

Fulton, M.H., Key, P.B., 2001. Acetylcholinesterase inhibition in 

estuarine fish and invertebrates as an indicator of 

organophosphorus insecticide exposure and effects. Environ. 

Toxicol. Chem. 20, 37–45. 

Ganske, F., Bornscheuer, U.T., 2006. Growth of Escherichia coli, 

Pichia pastoris and Bacillus cereus in the presence of the ionic 

liquids [BMIM] [BF 4] and [BMIM] [PF 6] and organic solvents. 

Biotechnol. Lett. 28, 465–469. 

Garcia, M.T., Gathergood, N., Scammells, P.J., 2005. Biodegradable 

ionic liquids. part II: effect of the anion and toxicology. Green 

Chem. 7, 9–14. 

García-Lorenzo, A., Tojo, E., Tojo, J., Teijeira, M., Rodríguez- 

Berrocal, F.J., González, M.P., Martínez-Zorzano, V.S., 2008. 

Cytotoxicity of selected imidazolium-derived ionic liquids in 

the human Caco-2 cell line. Sub-structural toxicological 

interpretation through a QSAR study. Green Chem. 10, 508–516. 

Gathergood, N., Scammells, P.J., 2002. Design and preparation of 

room-temperature ionic liquids containing biodegradable side 

chains. Aust. J. Chem. 55, 557–560. 

Gathergood, N., Garcia, M.T., Scammells, P.J., 2004. Biodegradable 

ionic liquids: part I. Concept, preliminary targets and 

evaluation. Green Chem. 6, 166–175. 

Gathergood, N., Scammells, P.J., Garcia, M.T., 2006. Biodegradable 

ionic liquids. Part III: the first readily biodegradable ionic 

liquids. Green Chem. 8, 156–160. 

Gordon, C.M., 2001. New developments in catalysis using ionic 

liquids. Appl. Catal. A 222, 101–117. 

Gorman-Lewis, D.J., Fein, J.B., 2004. Experimental study of the 

adsorption of an ionic liquid onto bacterial and mineral 

surfaces. Environ. Sci. Technol 38, 2491–2495. 

Grabinska-Sota, E., Kalka, J., 2004. An assessment of the toxicity 

of pyridinium chlorides and their biodegradation 

intermediates. Environ. Int 28, 687–690. 

Grabinska-Sota, E., Kalka, J., 2006. Toxicity of imidazolium chlorides 

to aquatic organisms. Polish J. Environ. Stud. 15, 405–409. 

Hagiwara, R., Ito, Y., 2000. Room temperature ionic liquids of 

alkylimidazolium cations and fluoroanions. J. Fluorine Chem. 

105, 221–227. 

Harjani, J.R., Singer, R.D., Garcia, M.T., Scammells, P.J., 2008. The 

design and synthesis of biodegradable pyridinium ionic 

liquids. Green Chem. 10, 436–438. 

Hassoun, E.A., Abraham, M., Kini, V., Al-Ghafri, M., Abushaban, A., 

2002. Cytotoxicity of the ionic liquid, 1- N-butyl-3methylimidazolium 

chloride. Res. Commun, Pharmacol. 

Toxicol 7, 23–31. 

Howard, P.H., Boethling, R.S., Stiteler, W., Meylan, W., Beauman, J., 

1991. Development of a predictive model for biodegradability 

based on BIODEG, the evaluated biodegradation data base. Sci. 

Total Environ. 110, 635–641. 

Huddleston, J.G., Visser, A.E., Reichert, W.M., Willauer, H.D., 

Broker, G.A., Rogers, R.D., 2001. Characterization and 

comparison of hydrophilic and hydrophobic room 

temperature ionic liquids incorporating the imidazolium 

cation. Green Chem. 3, 156–164. 

Itakura, T., Hirata, K., Aoki, M., Sasai, R., Yoshida, H., Itoh, H., 

2008. Decompostition and removal of ionic liquid in aqueous 

solution by hydrothermal and photocatalytic treatment. 

Environ. Chem. Lett. doi:10.1007/s10311-008-0177-7. 

Jastorff, B., Störmann, R., Ranke, J., Mölter, K., Stock, F., 

Oberheitmann, B., Hoffmann, W., Hoffmann, J., Nüchter, M., 

Ondruschka, B., Filser, J., 2003. How hazardous are ionic 

liquids? Structure-activity relationships and biological testing 

as important elements for sustainability evaluation. Green 

Chem. 5, 136–142. 

Jastorff, B., Mölter, K., Behrend, P., Bottin-Weber, U., Filser, J., 

Heimers, A., Ondruschka, B., Ranke, J., Schaefer, M., 


Schröder, H., Stark, A., Stepnowski, P., Stock, F., Störmann, R., 

Stolte, S., Welz-Biermann, U., Ziegert, S., Thöming, J., 2005. 

Progress in evaluation of risk potential of ionic liquids-basis 

for an eco-design of sustainable products. Green Chem. 7, 

362–372. 

Kaiser, K.L.E., Palabrica, V.S., 1991. Photobacterium phosphoreum, 

toxicity data index. Water Poll. Res. J. Can. 26, 361–431. 

Kaiser, J.P., Feng, Y.C., Bollag, J.M., 1996. Microbial metabolism of 

pyridine, quinoline, acridine, and their derivatives under 

aerobic and anaerobic conditions. Microbiol. Rev. 60, 483–498. 

Kazarian, S.G., Briscoe, B.J., Welton, T., 2000. Combining ionic 

liquids and supercritical fluids: in situ ATR-IR study of CO 2 

dissolved in two ionic liquids at high pressures. Chem. 

Commun., 2047–2048. 

Kelman, D., Kashman, Y., Rosenberg, E., Ilan, M., Ifrach, I., 

Loya, Y., 2001. Antimicrobial activity of the reef sponge 

Amphimedon viridis from the Red Sea: evidence for selective 

toxicity. Aquat. Microb. Ecol 24, 9–16. 

Kulacki, K.J., Lamberti, G.A., 2008. Toxicity of imidazolium ionic 

liquids to freshwater algae. Green Chem. 10, 104–110. 

Kumar, R.A., Papaïconomou, N., Lee, J.-M., Salminen, J., Clark, D.S., 

Prausnitz, J.M., 2009. In vitro cytotoxicities of ionic liquids: effect 

of cation rings, functional groups, and anions. Environ. Toxicol. 

24, 388–395. 

Kumar, S., Ruth, W., Sprenger, B., Kragl, U., 2006. On the 

biodegradation of ionic liquid 1-butyl-3-methylimidazolium 

tetrafluoroborate. Chim. Oggi 24, 24–26. 

Lahr, J., 1997. Ecotoxicology of organisms adapted to life in 

temporary freshwater ponds in arid and semi-arid regions. 

Arch. Environ. Contam. Toxicol 32, 50–57. 

Landry, T.D., Brooks, K., Poche, D., Woolhiser, M., 2005. Acute 

toxicity profile of 1-butyl-3-methylimidazolium chloride. Bull. 

Environ. Contam. Toxicol 74, 559–565. 

Larson, J.H., Frost, P.C., Lamberti, G.A., 2008. Variable toxicity of 

ionic liquid-forming chemicals to Lemna minor and the 

influence of dissolved organic matter. Environ. Toxicol. Chem. 

27, 676–681. 

Lata1a, A., Stepnowski, P., N˛edzi, M., Mrozik, W., 2005. Marine 

toxicity assessment of imidazolium ionic liquids: acute effects 

on the Baltic algae Oocystis submarina and Cyclotella 

meneghiniana. Aqua. Toxicol 73, 91–98. 

Lee, S.-M., Chang, W.-J., Choi, A.-R., Koo, Y.-M., 2005. Influence of 

ionic liquids on the growth of Escherichia coli. Korean. J. Chem. 

Eng 22, 687–690. 

Lewis, M.A., 1995. In: Rand, G.M. (Ed.), Fundamentals of 

Aquatic Toxicology: Effects, Environment Fate, and Risk 

Assessment, second ed. Taylor and Francis, Washington, 

DC, USA, pp. 135–170. 

Li, G., Shen, J., Zhu, Y., 1998. Study of pyridinium-type functional 

polymers. II. Antibacterial activity of soluble pyridinium-type 

polymers. J. Appl. Polym. Sci. 67, 1761–1768. 

Li, X., Zhao, J., Li, Q., Wang, L., Tsang, S.C., 2007. Ultrasonic 

chemical oxidative degradations of 1,3-dialkylimidazolium 

ionic liquids and their mechanistic elucidations. Dalton Trans, 

1875–1880. 

Li, X.-Y., Zhou, J., Yu, M., Wang, J.-J., Pei, Y.C., 2009. Toxic effects of 

1-methyl-3-octylimidazolium bromide on the early embryonic 

development of the frog Rana nigromaculata. Ecotox. Environ. 

Saf. 72, 552–556. 

Luis, P., Ortiz, I., Alcado, R., Irabien, A., 2007. A novel group 

contribution method in the development of a QSAR for 

predicting the toxicity (Vibrio fischeri EC 50) of ionic liquids. 

Ecotoxicol. Environ. Saf 67, 423–429. 

Luo, Y.-R., Li, X.-Y., Chen, X.-X., Zhang, B.-J., Sun, Z.-J., Wang, J.-J., 

2008. The developmental toxicity of 1-butyl-3-octylimidazolium 

bromide on Daphnia magna. Environ. Toxicol 23, 736–744. 

Malich, G., Markovic, B., Winder, C., 1997. The sensitivity and 

specificity of the MTS tetrazolium assay for detecting the in

vitro cytotoxicity of 20 chemicals using human cell lines. 

Toxicology 124, 179–192. 

Mann, R.M., Bidwell, J.R., 2000. Application of the FETAX protocol 

to assess the development toxicity of nonylphenol ethoxylate 

to Xenopus laevis and two Australian frogs. Aquat. Toxicol 51, 

19–29. 

Marsh, K.N., Boxall, J.A., Lichtenthaler, R., 2004. Room 

temperature ionic liquids and their mixtures–a review. Fluid 

Phase Equilibr 219, 93–98. 

Massoulié, J., Pezzementi, L., Bon, S., Krejci, E., Vallette, F.-M., 

1993. Molecular and cellular biology of cholinesterases. Prog. 

Neurobiol. 41, 31–91. 

Matsumoto, M., Mochiduki, K., Fukunishi, K., Kondo, K., 2004a. 

Extraction of organic acids using imidazolium-based ionic 

liquids and their toxicity to Lactobacillus rhamnosus. Sep.Purif. 

Technol. 40, 97–101. 

Matsumoto, M., Mochiduki, K., Kondo, K., 2004b. Toxicity of ionic 

liquids and organic solvents to lactic acid-producing bacteria. 

J. Biosci. Bioeng 98, 344–347. 

Matzke, M., Stolte, S., Thiele, K., Juffernholz, T., Arning, J., Ranke, J., 

Welz-Biermann, U., Jastorff, B., 2007. The influence of anion 

species on the toxicity of 1-alkyl-3-methylimidazolium ionic 

liquids observed in an (eco)toxicological test battery. Green 

Chem. 9, 1198–1207. 

Matzke, M., Stolte, S., Böschen, A., Filser, J., 2008. Mixture effects 

and predictability of combination effects of imidazolium 

based ionic liquids as well as imidazolium based ionic liquids 

and cadmium on terrestrial plants (Triticum aestivum) and 

limnic green algae (Scenedesmus vacuolatus). Green Chem. 10, 

784–792. 

Matzke, M., Stolte, S., Arning, J., Uebers, U., Filser, J., 2009a. Ionic 

liquids in soils: effects of different anion species of 

imidazolium based ionic liquids on wheat (Triticum aestivum) 

as affected by different clay minerals and clay concentrations. 

Ecotoxicology 18, 197–203. 

Matzke, M., Thiele, K., Müller, A., Filser, J., 2009b. Sorption and 

desorption of imidazolium based ILs in different soil types. 

Chemosphere 74, 568–574. 

McFarlane, J., Ridenour, W.B., Luo, H., Hunt, R.D., Depaoli, D.W., 

Ren, R.X., 2005. Room temperature ionic liquids for 

separating organics from produced water. Sep. Sci. Technol 

40, 1245–1265. 

McGuinness, D.S., Cavell, K.J., 2000. Donor-functionalized 

heterocyclic carbine complexes of palladium(II): efficient 

catalysts for C-C coupling reactions. Organometallics 19, 

741–748. 

McQueen, D.J., Post, J.R., Mills, E.L., 1986. Trophic relationships in 

freshwater pelagic ecosystems. Can. J. Fish. Aqua. Sci. 43, 

1571–1581. 

Morawski, A.W., Janus, M., Goc-Maciejewska, I., Syguda, A., 

Pernak, J., 2005. Decomposition of ionic liquids by 

photocatalysis. Pol. J. Chem. 79, 1929–1935. 

Myasoedov, B.F., Molochnikova, N.P., Shkinev, V.M., Spivakov, B.Y., 

1995. The Behavior of Actinides in Two-Phase Aqueous Systems 

Based on Polyethylene Glycol. Plenum, New York, pp. 91. 

Nockemann, P., Thijs, B., Driesen, K., Janssen, C.R., Hecke, K.V., 

Meervelt, L.V., Kossmann, S., Kirchner, B., Binnemans, K., 

2007. Choline saccharinate and choline acesulfamate: ionic 

liquids with low toxicities. J. Phys. Chem. B 111, 5254–5263. 

Olivier, H., 1999. Recent developments in the use of non-aqueous 

ionic liquids for two-phase catalysis. J. Mol. Cat. A: Chem. 146, 

285–289. 

Osenberg, C.W., Mittelbach, G.G., 1989. Effects of body size on the 

predator–prey interaction between pumpkinseed sunfish and 

gastropods. Ecol. Monogr. 59, 405–432. 

Pernak, J., Czepukowicz, A., 2001. New ionic liquids and their 

antielectrostatic properties. Ind. Eng. Chem. Res. 40, 2379– 

2383. 


Pernak, J., Rogo _za, J., Mirska, I., 2001a. Synthesis and 

antimicrobial activities of new pyridinium and 

benzimidazolium chlorides. Eur. J. Med. Chem. 36, 313–320. 

Pernak, J., Kalewska, J., Ksycińska, H., Cybulski, J., 2001b. 

Synthesis and anti-microbial activities of some pyridinium 

salts with alkoxymethyl hydrophobic group. Eur. J. Med. 

Chem. 36, 899–907. 

Pernak, J., Chwa1a, P., 2003. Synthesis and anti-microbial activities of 

choline-like quaternary ammonium chlorides. Eur. J. Med. Chem. 

38, 1035–1042. 

Pernak, J., Sobaszkiewicz, K., Mirska, I., 2003. Anti-microbial 

activities of ionic liquids. Green Chem. 5, 52–56. 

Pernak, J., Goc, I., Mirska, I., 2004a. Anti-microbial activities of 

protic ionic liquids with lactate anion. Green Chem. 6, 323–329. 

Pernak, J., Sobaszkiewicz, K., Foksowicz-Flaczyk, J., 2004b. Ionic 

liquids with symmetrical dialkoxymethyl-substituted 

imidazolium cations. Chem. Eur. J 10, 3479–3485. 

Pham, T.P.T., Cho, C.-W., Vijayaraghavan, K., Min, J., Yun, Y.-S., 

2008a. Effect of imidazolium-based ionic liquids on the 

photosynthetic activity and growth rate of Selenastrum 

capricornutum. Environ. Toxicol. Chem. 27, 1583–1589. 

Pham, T.P.T., Cho, C.-W., Min, J., Yun, Y.-S., 2008b. Alkyl-chain 

length effects of imidazolium and pyridinium ionic liquids on 

photosynthetic response of Pseudokirchneriella subcapitata. 

J. Biosci. Bioeng 105, 425–428. 

Pham, T.P.T., Cho, C.-W., Jeon, C.-O., Chung, Y.-J., Lee, M.-W., 

Yun, Y.-S., 2009. Identification of metabolites involved in the 

biodegradation of the ionic liquid 1-butyl-3-methylpyridinium 

bromide by activated sludge microorganisms. Environ. Sci. 

Technol 43, 516–521. 

Peraccini, D., Chiappe, C., Intorre, L., Pretti, C., 2007. In: Letcher, T. 

M. (Ed.), Thermodynamics, Solubility and Environmental 

Issues. Elsevier, UK, pp. 259–278. 

Pope, C., Karanth, S., Liu, J., 2005. Pharmacology and toxicology of 

cholinesterase inhibitors: uses and misuses of a common 

mechanism of action. Environ. Toxicol. Pharmacol 19, 433–446. 

Pretti, C., Chiappe, C., Pieraccini, D., Gregori, M., Abramo, F., 

Monni, G., Intorre, L., 2006. Acute toxicity of ionic liquids to 

the zebrafish (Danio rerio). Green Chem. 8, 238–240. 

Pretti, C., Chiappe, C., Baldetti, I., Brunini, S., Monni, G., Intorre, L., 

2009. Acute toxicity of ionic liquids for three freshwater 

organisms: Pseudokirchneriella subcapitata, Daphnia magna and 

Danio rerio. Ecotoxicol. Environ. Saf. 72, 1170–1176. 

Ranke, J., Mölter, K., Stock, F., Bottin-Weber, U., Poczobutt, J., 

Hoffmann, J., Ondruschka, B., Filser, J., Jastorff, B., 2004. 

Biological effects of imidazolium ionic liquids with varying 

chain length in acute Vibrio fischeri and WST-1 cell viability 

assays. Ecotox. Environ. Saf 58, 396–404. 

Ranke, J., Müller, A., Bottin-Weber, U., Stock, F., Stolte, S., Arning, J., 

Störmann, R., Jastorff, B., 2007a. Lipophilicity parameters for 

ionic liquid cations and their correlation to in vitro cytotoxicity. 

Ecotox. Environ. Saf 67, 430–438. 

Ranke, J., Stolte, S., Störmann, R., Arning, J., Jastorff, B., 2007b. 

Design of sustainable chemical products – the example of 

ionic liquids. Chem. Rev. 107, 2183–2206. 

Rogers, R.D., Bond, A.H., Bauer, C.B., Zhang, J., Rein, S.D., Chomko, R.R., 

Roden, D.M., 1995. Partitioning behavior of 99 Tc and, 129 I from 

simulated Handford tank wastes using polyethylene-glycol based 

aqueous biphasic systems. Solvent Extr. Ion Exch. 13, 689–713. 

Romero, A., Santos, A., Tojo, J., Rodríguez, A., 2008. Toxicity and 

biodegradability of imidazolium ionic liquids. J. Hazard. Mater 

151, 268–273. 

Rooney, D.W., Seddon, K.R., 2001. In: Wypych, G. (Ed.), Handbook 

of Solvents. ChemTech Publishing, Toronto, pp. 1459–1484. 

Salminen, J., Papaiconomou, N., Kumar, R.A., Lee, J.-M., Kerr, J., 

Newman, J., Prausnitz, J.M., 2007. Physicochemical properties 

and toxicities of hydrophobic piperidinium and pyrrolidinium 

ionic liquids. Fluid Phase Equilibr 261, 421–426.

372 

Samorì, C., Pasteris, A., Galletti, P., Tagliavini, E., 2007. Acute 

toxicity of oxygenated and nonoxygenated imidazoliumbased 

ionic liquids to Daphnia magna and Vibrio fischeri. 

Environ. Toxicol. Chem. 26, 2379–2382. 

Schmid, A., Kollmer, A., Mathys, R.G., Withot, B., 1998. Development 

toward large-scale bacterial bioprocesses in the presence of bulk 

amounts of organic solvents. Extremophiles 2, 249–256. 

Sheldon, R., 2001. Catalytic reactions in ionic liquids. Chem. 

Commun, 2399–2407. 

Sheldon, R.A., 2005. Green solvents for sustainable organic 

synthesis: state of the art. Green Chem. 7, 267–278. 

Siedlecka, E.M., Stepnowski, P., 2009. The effect of alkyl chain 

length on the degradation of alkylimidazolium- and 

pyridinium-type ionic liquids in a Fenton-like system. 

Environ. Sci. Pollut. Res. 16, 453–458. 

Siedlecka, E.M., Mrozik, W., Kaczyński, Z., Stepnowski, P., 

2008a. Degradation of 1-butyl-3-methylimidazolium chloride 

ionic liquid in a Fenton-like system. J. Hazard. Mater 154, 

893–900. 

Siedlecka, E.M., Go1˛ebiowski, M., Kumirska, J., Stepnowski, P., 

2008b. Identification of 1-butyl-3-methylimidazolium chloride 

degradation products formed in Fe(III)/H2O2 oxidation system. 

Chem. Anal. (Warsaw) 53, 943–951. 

Sipes, I.G., Knudsen, G.A., Kuester, R.K., 2008. The effects of dose 

and route on the toxicokinetics and disposition of 1-butyl- 

3-methylimidazolium chloride in male F-344 rats and female 

B6C3F1 mice. Drug Metab. Dispos. 36, 284–293. 

Sk1adanowski, A.C., Stepnowski, P., Kleszczyński, K., 

Dmochowska, B., 2005. AMP deaminase in vitro inhibition by 

xenobiotics. A potential molecular method for risk assessment of 

synthetic nitro- and polycyclic musks, imidazolium ionic liquids 

and N-glucopyranosylammoniumsalts.Environ.Toxicol.Phar. 

19, 291–296. 

Stasiewicz, M., Mulkiewicz, E., Tomczak-Wandzel, R., 

Kurmirska, J., Siedlecka, E.M., Go1ebiowski, M., Gajdus, J., 

Czerwicka, M., Stepnowski, P., 2008. Assessing toxicity and 

biodegradation of novel, environmentally benign ionic liquids 

(1-alkoxymethyl-3-hydroxypyridinium chloride, saccharine 

and acesulfamates) on cellular and molecular level. Ecotox. 

Environ. Saf 71, 157–165. 

Steinberg, S.M., Poziomek, E.J., Engelmann, W.H., Rogers, K.R., 1995. 

A review of environmental applications of bioluminescence 

measurements. Chemosphere 30, 2155–2197. 

Stepnowski, P., Sk1adanowski, A.C., Ludwiczak, A., qaczyńska, E., 

2004. Evaluating the cytotoxicity of ionic liquids using human 

cell line HeLa. Hum. Exp. Toxicol 23, 513–517. 

Stepnowski, P., 2005. Preliminary assessment of the sorption of 

some alkyl imidazolium cations as used in ionic liquids to 

soils and sediments. Aust. J. Chem. 58, 170–173. 

Stepnowski, P., Zaleska, A., 2005. Comparison of different 

advanced oxidation processes for the degradation of room 

temperatureionicliquids.J.Photoch.Photobio.A170,45–50. 

Stepnowski, P., Mrozik, W., Nichthauser, J., 2007. Adsorption of 

alkylimidazolium and alkylpyridinium ionic liquids onto 

natural soils. Environ. Sci. Technol 41, 511–516. 

Stock, F., Hoffmann, J., Ranke, J., Störmann, R., Ondruschka, B., 

Jastorff, B., 2004. Effects of ionic liquids on the 

acetycholinesterase–a structure-activity relationship 

consideration. Green Chem. 6, 286–290. 

Stolte, S., Arning, J., Bottin-Weber, U.,Matzke,M.,Stock,F.,Thiele,K., 

Uerdingen, M., Welz-Biermann, U., Jastorff, B., Ranke, J., 2006. 

Anion effects on the cytotoxicity of ionic liquids. Green Chem. 8, 

621–629. 

Stolte, S., Matzke, M., Arning, J., Böschen, A., Pitner, W.-R., 

Welz-Biermann, U., Jastorff, B., Ranke, J., 2007a. Effects of 


different head groups and functionalized side chains on 

the aquatic toxicity of ionic liquids. Green Chem. 9, 1170–1179. 

Stolte, S., Arning, J., Bottin-Weber, U., Müller, A., Pitner, W.-R., 

Welz-Biermann, U., Jastorff, B., Ranke, J., 2007b. Effects of 

different head groups and functionalized side chains on the 

cytotoxicity of ionic liquids. Green Chem. 9, 760–767. 

Stolte, S., Abdulkarim, S., Arning, J., Blomeyer-Nienstedt, A.-K., 

Bottin-Weber,U.,Matzke,M.,Ranke,J.,Jastorff,B.,Thöming, J., 

2008. Primary biodegradation of ionic liquid cations, identification 

of degradation products of 1-methyl-3-octylimidazolium chloride 

and electrochemical wastewater treatment of poorly 

biodegradable compounds. Green Chem. 10, 214–224. 

Studzińska, S., Buszewski, B., 2009. Study of toxicity of imidazolium 

ionic liquids to watercress (Lepidium sativum L.). Anal. Bioanal. 

Chem. 393, 983–990. 

Swatloski, R.P., Holbrey, J.D., Rogers, R.D., 2003. Ionic liquids are 

not always green: hydrolysis of 1-butyl-3-methylimidazolium 

hexafluorophosphate. Green Chem. 5, 361–363. 

Swatloski, R.P., Holbrey, J.D., Memon, S.B., Caldwell, G.A., 

Caldwell, K.A., Rogers, R.D., 2004. Using Caenorhabditis elegans 

to probe toxicity of 1-alkyl-3-methylimidazolium chloride 

based ionic liquids. Chem. Commun, 668–669. 

Torrecilla, J.S., García, J., Rojo, E., Rodríguez, R., 2009. Estimation 

of toxicity of ionic liquids in leukemia rat cell line and 

acetylcholinesterase enzyme by principal component 

analysis, neural networks and multiple linear regressions. J. 

Hazard. Mater 164, 182–194. 

U.S. Environmental Protection Agency 1998. Fate, transport, and 

transformation test guidelines. OPPTS 835.3110. Readily 

Biodegradability. 

Vijayaraghavan, K., Pham, T.P.T., Cho, C.-W., Won, S.W., Choi, S. 

B., Mao, J., Kim, S., Kim, Y.-R., Chung, B.W., Yun, Y.-S., 2009. 

An assessment on the interaction of a hydrophilic ionic liquid 

with different sorbents. Ind. Eng. Chem. Res. 48, 7283–7288. 

Wang, L.-S., Wang, L., Wang, L., Wang, G., Li, Z.-H., Wang, J.-J., 

2009. Effect of 1-butyl-3-methylimidazolium tetrafluoroborate 

on the wheat (Triticum aestivum L.) seedlings. Environ. Toxicol 

24, 296–303. 

Wang, X., Ohlin, C.A., Lu, Q., Fei, Z., Hu, J., Dyson, P.J., 2007. 

Cytotoxicity of ionic liquids and precursor compounds 

towards human cell line HeLa. Green Chem. 9, 1191–1197. 

Wasserscheid, P., Keim, W., 2000. Ionic liquids–New ‘‘solutions’’ for 

transition metal catalysis. Angew. Chem. Int. Ed. 39, 3772–3789. 

Wells, A.S., Coombe, V.T., 2006. On the freshwater ecotoxicity and 

biodegradation properties of some common ionic liquids. Org. 

Pro. Res. Dev. 10, 794–798. 

Welton, T., 1999. Room-temperature ionic liquids, solvents for 

synthesis and catalysis. Chem. Rev. 99, 2071–2083. 

Willauer, H.D., Huddleston, J.G., Griffin, S.T., Rogers, R.D., 1999. 

Partitioning of aromatic molecules in aqueous biphasic 

systems. Sep. Sci. Technol 34, 1069–1090. 

Wong, D.S.H., Chen, J.P., Chang, J.M., Chou, C.H., 2002. Phase 

equilibria of water and ionic liquids [emim][PF 6] and [bmim][PF 6]. 

Fluid Phase Equilibr 194–197, 1089–1095. 

Yu, M., Li, S.-M., Li, X.-Y., Zhang, B.-J., Wang, J.-J., 2009a. Acute 

effects of 1-octyl-3-methylimidazolium bromide ionic liquid 

on the antioxidant enzyme system of mouse liver. Ecotox. 

Environ. Saf. 71, 903–906. 

Yu, M., Wang, S.-H., Luo, Y.-R., Han, Y.-W., Li, X.-Y., Zhang, B.-J., 

Wang, J.-J., 2009b. Effects of the 1-alkyl-3-methylimidazolium 

bromide ionic liquids on the antioxidant defense system of 

Daphnia magna. Ecotox. Environ. Saf. 72, 1798–1804. 

Zhang, C., Malhotra, S.V., 2005. Increased paraoxon detection by 

acetylcholinesterase inactivation with ionic liquid additives. 

Tatanta 67, 560–563.

A QSAR model for predicting rejection of emerging 

contaminants (pharmaceuticals, endocrine disruptors) 

by nanofiltration membranes 

Victor Yangali-Quintanilla a,b, *, Anwar Sadmani a , Megan McConville a,b , 

Maria Kennedy a , Gary Amy a,b 

a 

UNESCO-IHE, Institute for Water Education, Westvest 7, 2611AX Delft, The Netherlands 

b 

Delft University of Technology, Stevinweg 1, 2628CN Delft, The Netherlands 

article info 


Received 25 February 2009 


4 June 2009 

Accepted 26 June 2009 

Available online 3 July 2009 

Keywords: 

Pharmaceuticals 

Endocrine disruptors 

Nanofiltration 

Modelling 

QSAR 




abstract 

A quantitative structure activity relationship (QSAR) model has been produced for predicting 

rejection of emerging contaminants (pharmaceuticals, endocrine disruptors, 

pesticides and other organic compounds) by polyamide nanofiltration (NF) membranes. 

Principal component analysis, partial least square regression and multiple linear regressions 

were used to find a general QSAR equation that combines interactions between 

membrane characteristics, filtration operating conditions and compound properties for 

predicting rejection. Membrane characteristics related to hydrophobicity (contact angle), 

salt rejection, and surface charge (zeta potential); compound properties describing 

hydrophobicity (log Kow, log D), polarity (dipole moment), and size (molar volume, molecular 

length, molecular depth, equivalent width, molecular weight); and operating conditions 

namely flux, pressure, cross flow velocity, back diffusion mass transfer coefficient, 

hydrodynamic ratio (Jo/k), and recovery were identified as candidate variables for rejection 

prediction. An experimental database produced by the authors that accounts for 106 

rejection cases of emerging contaminants by NF membranes as result of eight experiments 

with clean and fouled membranes (NF-90, NF-200) was used to produce the QSAR model. 

Subsequently, using the QSAR model, rejection predictions were made for external 

experimental databases. Actual rejections were compared against predicted rejections and 

acceptable R 2 correlation coefficients were found (0.75 and 0.84) for the best models. 

Additionally, leave-one-out cross-validation of the models achieved a Q 2 of 0.72 for internal 

validation. In conclusion, a unified general QSAR equation was able to predict rejections of 

emerging contaminants during nanofiltration; moreover the present approach is a basis to 

continue investigation using multivariate analysis techniques for understanding 

membrane rejection of organic compounds. 


* Corresponding author: UNESCO-IHE, Institute for Water Education, Westvest 7, 2611AX Delft, The Netherlands. Tel.: þ31 15 215 1745; 

fax: þ31 15 215 2921. 

E-mail addresses: v.a.yangaliquintanilla@tudelft.nl, v.yangaliquintanilla@unesco-ihe.org (V. Yangali-Quintanilla). 


doi:10.1016/j.watres.2009.06.054

374 


Nanofiltration (NF) and reverse osmosis (RO) are technologies 

that provide medium to high rejections of organic 

compounds present as emerging contaminants (micropollutants) 

in water, namely endocrine disrupting 

compounds (EDCs), pharmaceutically active compounds 

(PhACs), and pesticides (Kiso et al., 2001; Schäfer et al., 2003; 

Kimura et al., 2003, 2004; Nghiem et al., 2004). The presence 

of micropollutants has been identified in surface water 

bodies, sewage treatment plant effluents, and stages of 

drinking water treatment plants and even at trace-levels in 

finished drinking water (Kolpin et al., 2002; Heberer, 2002; 

Castiglioni et al., 2006). The possible effects on aquatic 

organisms and human health, associated with the 

consumption of water containing low concentrations of 

single compounds, have been presented in toxicology studies 

(Pomati et al., 2006; Escher et al., 2005; Vosges et al., 2008). 

The studies demonstrate that researchers do not yet 

understand the exact risks from decades of persistent 

exposure to random combinations of low levels of pharmaceuticals, 

EDCs, and other organic compounds; hence, the 

long-term effects of consumption of water containing low 

concentrations of micropollutants will remain as an unanswered 

question for the foreseeable future. Meanwhile, 

water treatment facilities are implementing monitoring 

programs, research organizations dealing with water reuse 

have published reports, and studies have addressed the topic 

(Drewes et al., 2006; Verliefde et al., 2007). An important 

aspect to deal with the problem has been the identification 

of compound physicochemical properties and membrane 

characteristics to explain transport, adsorption and removal 

of micropollutants by different mechanisms, explicitly size/ 

steric exclusion, hydrophobic adsorption and partitioning, 

and electrostatic repulsion (Kiso et al., 2001; Schäfer et al., 

2003; Kimura et al., 2003; Nghiem et al., 2004, Ozaki and Li, 

2002; Van der Bruggen and Vandecasteele, 2002; Bellona and 

Drewes, 2005; Xu et al., 2005). A number of articles have 

proposed a mechanistic understanding of the interaction 

between membranes and organic compounds; others have 

tried to apply fitting parameter models to model rejection 

(Cornelissen et al., 2005; Kim et al., 2007; Verliefde et al., 

2008). However, there have been few models to ‘‘predict’’ the 

rejection of compounds. To overcome that status, our 

objective was to create a general QSAR model to predict 

rejections based on an integral approach that considers 

membrane characteristics, filtration operating conditions 

and physicochemical compound properties. A quantitative 

structure activity relationship (QSAR) is a method that 

relates an activity of a set of compounds quantitatively to 

chemicals descriptors (structure or property) of those 

compounds (Sawyer et al., 2003). QSAR has the objective of 

prediction but maintaining a relationship to mechanistic 

interpretation. Applications of QSAR for the development of 

models to find relationships between membranes and 

organic compounds have been presented in journals related 

to drug discovery and medicinal chemistry for analysis of 

permeability of membranes to organic compounds (Ren 

et al., 1996; Fujikawa et al., 2007). The study of reverse 


osmosis membranes has also experienced the application of 

QSAR principles. Campbell et al. (1999) performed a QSAR 

analysis of surfactants influencing attachment of a mycobacterium 

to cellulose and aromatic polyamide reverse 

osmosis membranes; their objective was to understand the 

relationship between surfactant molecular properties and 

activity on the membrane surface to inhibit bacterial 

attachment to the membrane in order to reduce biofilm 

formation and to increase permeate production. More 

recently, Libotean et al. (2008) developed an artificial neural 

network model based on quantitative structure-property 

relations, the model claim to predict organic solute passage 

through reverse osmosis membranes considering simultaneous 

correlation of organic solutes (molecular descriptors) 

and membrane properties. Alike, our study uses the concept 

of QSAR analysis to quantify an activity, compound rejection 

by a membrane, in terms of organic compound physicochemical 

properties, membrane characteristics (salt rejection, 

pure water permeability, molecular weight cut-off, 

charge, hydrophobicity) and operating conditions (pressure, 

flux, cross flow velocity, back diffusion mass transfer coefficient, 

recovery). In this work a QSAR model was constructed 

with internal experimental data used for training. 

The model was internally validated using measures of 

goodness of fit and prediction. Subsequently, after identification 

of a relationship in form of an equation, estimation of 

rejections for an external dataset for different compounds 

and membranes were used to externally validate the model. 

Similarly, rejections of more emerging organic contaminants 

can be predicted in advance, before nanofiltration or reverse 

osmosis applications. Nevertheless, the QSAR model is 

applicable in the range of boundary experimental conditions 

that will be defined in the experimental section of this 

publication (Section 3). 

2. Theory 

2.1. Principal component analysis 

Principal component analysis (PCA) is a method that allows 

simplification of many variables into a group of a few variables 

that might be measuring the same principles of a system. It 

may occur that a system considers an abundance of variables 

to explain a process; in this case principal component analysis 

reduces the redundancy of information. The general objectives 

of PCA are data reduction and interpretation. Although p 

variables (components) are required to reproduce the total 

system variability, often much of this variability can be 

accounted for by a small number k of principal components. 

The k is the number of components (reduced) that represent 

the initial p variables. In general, PCA is concerned with 

whether the covariances or correlations between a set of 

observed variables x 1, x 2, ., x p can be explained in terms of 

a smaller number of unobservable components, c1, c2, ., ck, 

where k < p. Thus, there is as much information in the k 

components as there is in the original p variables. Comprehensive 

details about the theory of PCA can be found elsewhere 

(Jolliffe, 2002; Johnson and Wichern, 2007; Everitt and 

Dunn, 2001).

2.2. Multiple linear regression 

Multiple linear regression is a method of analysis for assessing 

the strength of the relationship between a set of explanatory 

variables known as independent variables, and a single 

response or dependent variable. Applying multiple regression 

analysis to a set of data results in what are known as regression 

coefficients, one for each explanatory variable. The 

multiple regression model for a response variable, y, with 

observed values, y1, y2, ., yn (where n is the sample size) and q 

explanatory variables, x1, x2, ., xq with observed values, x1i, 

x2i, ., xqi for i ¼ 1, ., n, is 

yi ¼ b 0 þ b 1x1i þ b 2x2i þ / þ b qxqi þ 3i 

The regression coefficients, b0, b1, ., bq, are generally estimated 

by least squares. The term 3i is the residual or error for 

individual i and represents the deviation of the observed value 

of the response for this individual from that expected by the 

model. These error terms are assumed to have a normal 

distribution with variance s 2 . The fit of a multiple regression 

model can be judged with calculation of the multiple correlation 

coefficient, R, defined as the correlation between the 

observed values of the response variable and the values predicted 

by the model. The squared value of R (R 2 ) gives the 

proportion of the variability of the response variable accounted 

for by the explanatory variables. Analysis of variance 

(ANOVA) will provide an F-test of the null hypothesis that 

each of b0, b1, ., bq, is equal to zero, or in other words that R 2 is 

zero (Landau and Everitt, 2004). 

2.3. Principal component regression and partial least 

squares regression 

Principal component regression (PCR) is a method in which 

the components from the principal component method are 

used for regression. Hence, the principal components of 

the matrix X are used as regressors of a dependent Y. The 

orthogonality of the principal components eliminates the 

multicollinearity problem. But, the problem of choosing an 

optimum subset of predictors remains. A possible strategy is 

to keep only a few of the first components. But they are chosen 

to explain X rather than Y, and therefore, nothing guarantees 

that the principal components, which ‘‘explain’’ X, are relevant 

for Y. Problems may arise, however, if there is a lot of 

variation in X. PCR finds, somewhat uncritically, those latent 

variables that describe as much as possible of the variation in 

X. But sometimes the variable itself gives rise to only small 

variations in X, and if the interferences vary a lot, then the 

latent variables found by PCR may not be particularly good at 

describing Y. In the worst case important information may be 

hidden in directions in the X-space that PCR interprets as 

disturbance, and therefore leaves out. Partial least squares 

regression (PLS) is able to cope better with this problem, by 

forming variables that are relevant for describing Y. By 

contrast, PLS regression finds components from X that are 

also relevant for Y. Specifically, PLS regression searches for 

a set of components (called latent vectors) that performs 

a simultaneous decomposition of X and Y with the constraint 

that these components explain as much as possible of the 


(1) 

covariance between X and Y. This step generalizes PCA. The 

goal of PLS regression is to predict Y from X and to describe 

their common structure (Abdi, 2003; Jørgensen and Goegebeur, 

2006). 

3. Experimental 

3.1. Chemicals, membranes, materials and 

experimental conditions 

The list of organic compounds representing emerging 

contaminants is presented in Table 1. The compounds were 

selected considering: their occurrence in surface water and 

drinking water, their identification as priority emerging 

contaminants, the availability of analytical methods, their 

quantitative and qualitative representation of physicochemical 

properties. The pharmaceutical compounds (caffeine, 

sulfamethoxazole, acetaminophen, phenacetin, phenazone, 

carbamazepine, naproxen, ibuprofen, metronidazole), endocrine 

disrupting compounds (17b-estradiol, estrone, bisphenol 

A, nonylphenol, atrazine) and sodium alginate were 

purchased from Sigma–Aldrich (Sigma–Aldrich, Schnelldorf, 

Germany). Potassium chloride, sodium hydroxide, hydrochloric 

acid and magnesium sulphate anhydrous were 

purchased from J.T.Baker (J.T.Baker, Deventer, Netherlands). 

Two thin film composite NF membranes were selected for 

this study (NF-200 and NF-90, Dow-Filmtec, Dow Chemical 

Co., Midland, MI). The experimental setup consisted of two 

filtration SEPA CF II (GE Osmonics, Minnetonka, MN) cells and 

cell holders in parallel, in order to increase permeate 

production and achieve hydrodynamic conditions, two 

hydraulic pumps (Power Team, Bega Int. BV, Netherlands), 

a 60 litres stainless steel tank (Tummers, Netherlands), 

a positive displacement pump (Hydra-Cell pump, Wanner 

Eng. Inc., Minneapolis, MN), a frequency converter (VLT 

microdrive, Danfoss, SA), a chiller/heater (Julabo, Germany), 

control needle valves, pressure gauges, flow meters, 

a proportional pressure relief valve and stainless steel 

tubings (Swagelok BV, Netherlands), a digital balance 

(Sartorius, Germany) and, a computer for flow rate data 

acquisition. A piece of membrane was compacted with 

deionised water for 6 h at a pressure of 276 kPa before performing 

an experiment with the membrane. The experiments 

were conducted in a recycle mode in which permeate 

and concentrate were recirculated into the feed tank for the 

first 72 h (a pre-equilibration period); then, permeate was 

collected within the next 24 h. The feed solution of all the 

experiments contained a cocktail of 14 compounds (concentration 

ranging from 6.5 to 65 mg/L). The main reason for 

conducting experiments at concentrations of mg/L was to 

accelerate steady state (after membrane adsorption) conditions 

in a limited time (3 days). At very low concentrations 

more time would be needed to achieve steady state conditions 

and at short time tests low concentrations may lead to 

over-estimation of rejections. A specific flux decline of 15% of 

initial flux was targeted to foul the membranes using a feed 

solution containing w10 mg/L DOC of sodium alginate. The 

study of Lee et al. (2004) concluded that polysaccharides were 

important membrane foulants. Sodium alginate was used as

376 

Table 1 – List of compounds and physicochemical properties. 

Name Molecular Acid 

weight 

(g/mol) 

pKa 

20 C a 

b 

Log Kow a 

Log D Dipole 

(pH 7) moment 

(debye) c 

surrogate of polysaccharides. Alginate is frequently used as 

a model for organic matter of algae origin (Henderson et al., 

2008). Experiments were carried out for both clean and fouled 

NF-90 and NF-200 membranes. The selection of membranes 

was based on a qualitative rejection assessment of emerging 

contaminants with molecular weight of more than 150 Da by 

membranes with a MWCO between 200 and 300 Da. A total of 

eight experiments were performed; at hydrodynamic ratio of 

pure water permeation flux to back diffusion mass transfer 

coefficient (J0/k) ofw1 and recovery 3% (NF-90 clean, NF-200 

clean), at J0/k of w1 and recovery 8% (NF-200 clean), at J0/k of 

w2 and recovery 3 (NF-90 clean), and at J 0/k of w2 and 

recovery 8 (NF-90 clean and fouled, NF-200 clean and fouled). 

The calculation method of k (back diffusion mass transfer 

coefficient) was presented in a previous publication (Yangali- 

Quintanilla et al., 2009). All the experiments were carried out 

at a controlled temperature of 20 C, an ionic strength of 

10 mM as KCl and a pH of 7. The transmembrane pressures 

were in the range of 276–483 kPa. The fluxes were between 

4.3 and 30.2 L/m 2 per day; cross flow velocities were between 

0.5 and 4.5 cm/s. The experiments produced a total internal 

dataset of 106 rejection cases; the dataset can be accessed as 

supplementary data. The boundary experimental conditions 

of the internal dataset are presented in Table 2. The internal 

dataset was used to develop the model. An external dataset 

that gathered three different datasets was used for validation 

of the model. The external dataset is presented as supplementary 

data. Experimental conditions for the first part of 

the external dataset can be obtained from Kim et al. (2007) 

and Yangali-Quintanilla et al. (2008). Experimental conditions 

for the second and third parts can be found in Verliefde et al. 

(2008); the data correspond to filtration experiments using 

synthetic water solutions. 

Molar 

volume d 

(cm 3 / 

mol) 

Molec. 

length 

(nm) e 

3.2. Analytical equipment, analyses of compounds 

and membranes 

The pharmaceuticals and endocrine disruptors (with the 

exception of atrazine) were analyzed by Technologiezentrum 

Wasser (TZW, Karlsruhe, Germany). The detection limit was 

10 ng/L per compound. The uncertainty of estimates was of 

15% according to a validation method of the analysis 

protocol and due to high concentrations of the samples (mg/L). 

The analyses of pharmaceuticals were performed according to 

Table 2 – Data range of membrane characteristics, 

operating conditions and rejections. 

Variable Unit Min. value Max. value 

Molecular weight cut-off 

(MWCO) 

Pure water permeability 

(PWP) 

Molec. 

width 

(nm) e 

Da 200 300 

L/m 2 per 

day/kPa 

Molec. 

depth 

(nm) e 

Equiv. 

width 

(nm) e 

Classification f 

Acetaminophen 151 10.2 0.46 0.23 4.55 120.90 1.14 0.68 0.42 0.53 HL-neutral 

Phenacetine 179 N/A 1.58 1.68 4.05 163.00 1.35 0.69 0.42 0.54 HL-neutral 

Caffeine 194 N/A 0.07 0.45 3.71 133.30 0.98 0.87 0.56 0.70 HL-neutral 

Metronidazole 171 N/A 0.02 0.27 6.30 117.80 0.93 0.90 0.48 0.66 HL-neutral 

Phenazone 188 N/A 0.38 0.54 4.44 162.70 1.17 0.78 0.56 0.66 HL-neutral 

Sulfamethoxazole 253 5.7 0.89 0.45 7.34 173.10 1.33 0.71 0.58 0.64 HL-ionic 

Naproxen 230 4.3 3.18 0.34 2.55 192.20 1.37 0.78 0.75 0.76 HB-ionic 

Ibuprofen 206 4.3 3.97 0.77 4.95 200.30 1.39 0.73 0.55 0.64 HB-ionic 

Carbamazepine 236 N/A 2.45 2.58 3.66 186.50 1.20 0.92 0.58 0.73 HB-neutral 

Atrazine 216 N/A 2.61 2.52 3.43 169.80 1.26 1.00 0.55 0.74 HB-neutral 

17b-estradiol 272 10.3 4.01 3.94 1.56 232.60 1.39 0.85 0.65 0.74 HB-neutral 

Estrone 270 10.3 3.13 3.46 3.45 232.10 1.39 0.85 0.67 0.76 HB-neutral 

Nonylphenol 220 10.3 5.71 5.88 1.02 236.20 1.79 0.75 0.59 0.66 HB-neutral 

Bisphenol A 228 10.3 3.32 3.86 2.13 199.50 1.25 0.83 0.75 0.79 HB-neutral 

a ADME/Tox Web Software. 

b Experimental database: SRC PhysProp Database. 

c Chem3D Ultra 7.0. 

d ACD/ChemSketch Properties Batch. 

e Molecular Modeling Pro. 

f HL ¼ hydrophilic, HB ¼ hydrophobic. 


0.86 2.23 

Salt rejection (SR) a 

– 0.96 0.98 

Zeta potential (ZP) MV 48.04 10.78 

Contact angle (CA) 39.3 58.0 

Pressure (P) kPa 276 483 

Cross flow velocity (v) cm/s 0.5 4.5 

Back diffusion mass 

transfer coefficient (k) 

cm/s 2.70E-04 5.99E-04 

Flux (J ) L/m 2 per day 4.3 30.2 

Hydrodynamic ratio 

(J0/k) 

– 1 2 

Recovery (recov) % 3 8 

Rejection (rejection) % 17.7 99.0 

a 2000 mg/L MgSO4, 25 C, recovery 15%, pressure 1034 kPa, pH 8.

the protocols described by Sacher et al. (2001, 2008). Analyses 

of 17b-estradiol, estrone, nonylphenol and bisphenol A were 

done by gas chromatography/mass spectrometry (GC/MS) 

after automated solid-phase extraction onto a polymeric 

material and subsequent silylation of the analytes. First, 10 ml 

of a 50 ng/ml solution of 4-n-nonylphenol in acetone which 

was used as internal standard for the overall procedure were 

added to an aliquot of the water sample (1000 ml). Automated 

solid-phase extraction (Tekmar AutoTrace, Germany) was 

done on plastic cartridges filled with 200 mg of bondelut 

material (Fa. Varian, Darmstadt, Germany). After the enrichment 

step the solid-phase material was dried in a gentle 

stream of nitrogen. Elution was done with 4 ml of acetone. The 

acetone was evaporated to 100 ml in a stream of nitrogen and 

to dryness in a drying oven at 80 C. The dry residue was 

reconstituted with 100 ml of a silylation reagent mixture 

(N-methyl-N-trimethylsilyltrifluoro acetamide (MSTFA)/2% 

trimethyliodo silane). After a reaction time of 20 min at 80 C 

(drying oven), determination of the derivatives was done by 

GC/MS using a 6890 GC/MS system from Agilent Technologies 

(Waldbronn, Germany). 

Concentrations of atrazine were determined using microplate 

enzyme-linked immunoabsorbent assay (ELISA) kits 

(Abraxis LLC, Warminster, PA). Atrazine was determined with 

a detection limit of 0.04 mg/L, and uncertainty of 15%. To 

determine the hydrophobicity of membranes, contact angles 

of clean and fouled membrane surfaces were measured with 

CAM200 optical contact angle and surface tension meter (KSV 

Instruments, Finland) at Delft University of Technology; to 

measure contact angle, the sessile drop method was used. 

Surface charge, in terms of zeta potential, of clean and fouled 

membranes was quantified using ELS-8000 zeta potential 

analyzer (Otsuka Electronics, Japan). The zeta potential analyses 

were determined using a Milli-Q water solution at pH 7 

and ionic strength of 10 mM KCl. The zeta potential was 

determined using the electrophoresis method using a cell 

consisting of membrane and quartz cells. The zeta potential 

was calculated from the electrophoretic mobility using the 

Smoluchowski formula, a detailed explanation of calculation 

was provided in a previous publication (Shim et al., 2002). The 

pH of the solutions was measured using a calibrated Metrohm 

691 pH-meter (Metrohm AG, Herisau, Switzerland); the electrical 

conductivity and temperature were measured with 

a WTW Cond 330i (WTW GmbH, Weilheim, Germany) portable 

conductivity meter. Clean and fouled membranes were characterized 

to determine magnesium sulphate salt rejection at 

standard conditions specified by manufacturers, a pure water 

solution containing 2000 mg/L of magnesium sulphate at 25 C 

and pH 8 was filtrated at pressure of 1034 kPa and recovery of 

15%. 

3.3. Characterization and classification of compounds 

The acid dissociation constant as log K a (pK a) was used to 

determine the speciation of the organic compound in ionic 

species at pH 7. For hydrophobicity determination, log Kow and 

log D were used; log Kow is the octanol–water partition coefficient 

and log D is the ratio of the equilibrium concentrations 

of all species (unionized and ionized) of a molecule in octanol 

to the same species in the water phase. Values of pKa and log D 


were calculated by ADME/Tox web software. Solubility and 

log Kow values were obtained from SRC Physprop experimental 

database. The value of the molecular dipole moment 

was equal to the vector sum of the individual bond dipole 

moments. Dipole moment was calculated by Chem3D Ultra 

7.0, Cambridgesoft. Size descriptors included molar volume 

(MV), molecular length, molecular width, molecular depth and 

equivalent molecular width. The molecular length is defined 

as the distance between the two most distant atoms. The 

molecular width and molecular depth (width > depth) are 

measured by projecting the molecule on the plane perpendicular 

to the length axis and the equivalent molecular width 

is defined as the geometric mean of width and depth (Santos 

et al., 2006). Molar volumes were calculated using the program 

ACD/ChemSketch Properties Batch, ACD/Labs; and Molecular 

Modeling Pro, ChemSW, was used to compute size descriptors 

after optimization geometry of a molecule from the interaction 

of conformational analysis and energy minimization with 

a semi-empiric method MOPAC-PM3. Based on pKa and log - 

Kow values, the compounds were classified as hydrophilic 

neutral, hydrophilic ionic, hydrophobic ionic and hydrophobic 

neutral (see Table 1). Compounds with log Kow 2 were 

referred to as hydrophobic; therefore those with log Kow < 2 

were hydrophilic. The classification was based on an early 

reference (Connell, 1990). Although the value of 2 may seem 

low to consider hydrophobicity, the classification was not 

used in constructing the models and therefore the magnitude 

of log Kow or log D became more important. Table 1 shows the 

calculated values of molecular weight, pKa, log Kow, log D, 

dipole moment, molar volume, molecular length, molecular 

width, molecular depth and equivalent width. 

4. Results and discussion 

4.1. QSAR methodology 

The procedure to find a general QSAR equation to describe 

rejection was performed in four phases. The first phase was 

the organization of data from the experimental part. The data 

comprised of 106 rejection cases. The database showing the 

rejection cases is presented as supplementary data. A total of 

21 initial variables were used. The variables considered as 

compound descriptors were molecular weight (MW), solubility, 

log Kow, log D, dipole moment, molar volume, molecular 

length, molecular width, molecular depth and equivalent 

width; variables describing membrane characteristics were 

molecular weight cut-off (MWCO), pure water permeability 

(PWP), magnesium sulphate salt rejection (SR), charge of the 

membrane as zeta potential (ZP), and hydrophobicity as 

contact angle (CA); variables describing operating conditions 

were operating pressure (P), cross flow velocity (v), back 

diffusion mass transfer coefficient (k), flux (J ), ratio of pure 

water permeation flux J 0 and back diffusion mass transfer 

coefficient (J0/k) and recovery. The range of values for 

membrane characteristics, operating conditions and rejections 

is presented in Table 2. The second phase was dedicated 

to the process of variables reduction using a correlation 

matrix and factor analysis with principal component analysis. 

The third phase corresponded to the regression analysis. In

378 

the third phase three methodologies were implemented; the 

first was principal component analysis (PCA), with sequential 

application of multiple linear regressions (MLR). The second 

method was the use of partial least squares (PLS) regression 

and MLR; and the third method was the use of MLR only. The 

last phase was the validation process. The model was internally 

validated using measures of goodness of fit (regression 

coefficients) and prediction (leave-one-out cross-validation); 

Section 4.5 gives details about the validation process. External 

validation of the general QSAR model was implemented by 

predicting rejections for an external dataset of experiments 

performed with different compounds and membranes, and 

with comparable operating conditions. PCA and PLS were 

performed using the research and statistical package SPSS 

Statistics 16.0. Leave-one-out cross-validations of the models 

were performed with MobyDigs (Talete, Milano, Italy). 

4.2. Variables reduction with principal component 

analysis and QSAR model 

The correlation matrix of the initial 21 variables was scrutinized 

in order to obtain a not positive definite matrix as 

a requisite of PCA; the matrix is accompanied as supplementary 

data. A matrix is called not positive definite when there 

are both positive and negative eigenvalues. In the case of 

symmetric matrices, such as a correlation matrix, positive 

definiteness will only hold if the matrix and every principal 

submatrix have a positive determinant. A non-positive definite 

input matrix may signal a perfect linear dependency of 

one variable on another, known as collinearity. This was the 

case for MWCO and salt rejection (SR) that were perfectly 

linearly correlated. Therefore application of PCA considering 

independently one variable or the other will give the same 

results of variables reduction and number of components. In 

other words, MWCO will not be excluded with PCA, the 

variable will be separated in advance and the results obtained 

for SR may be replaced by the variable MWCO, or vice versa. 

Once an appropriate matrix was defined, the variables were 

analyzed in terms of how significant their correlations with 

rejection were; those correlations are also shown as an additional 

row and column of the 21 21 variables matrix. Rejection 

is only a reference variable to evaluate correlation with 

the rest of variables. After a sequential implementation of 

PCA, three components were extracted; they defined the 

initial database of 21 variables with 11 variables describing 

three relations namely membrane/operating-conditions 

(comp. 1: flux, pure water permeability, salt rejection, zeta 

potential, mass transfer coefficient, cross flow velocity), 

hydrophobicity/size (comp. 2: length, log Kow, log D) and size 

(comp. 3: equivalent width, depth). The final three components 

accounted for 89.3% explanation of total variance. It is 

important to mention that these results were produced for the 

experimental dataset. 

The next step was the implementation of multiple linear 

regression (MLR) using the new set of variables. The use of 

MLR after PCA presents the advantage of a more simplified 

modelling approach. Moreover, the analysis of data before 

MLR may help to identify variables that are similar in 

response, which was the case of SR and MWCO. The dependent 

variable for all regression analyses was rejection. Two 


methods of linear regression were used, the first method is 

called enter (forced) method; which performs a regression 

with the contribution of all variables entered to model the 

dependent variable. The second method is stepwise regression; 

which is a more sophisticated method. Each variable is 

entered in sequence and its contribution is assessed according 

to an F-test. In the present study an F-test with a statistical 

significance >0.10 implied removal of the variable, and F-test 

with a significance

Elimelech, 2003). Thus, SR is ultimately serving as a comparison 

parameter between membranes of the same type 

(aromatic polyamide) but with narrow differences in pore size, 

and possibly with differences in charge. The QSAR equation 

merges information about interaction of membrane characteristics, 

filtration operating conditions and organic 

compounds properties to predict rejections during nanofiltration. 

According to Eq. (2), contact angle and zeta potential 

as measurements of hydrophilicity and membrane surface 

charge, respectively, were not part of the equation and 

therefore did not contribute quantitatively to model rejection. 

However, size/steric hindrance effects related to salt rejection, 

and hydrophobicity of the solutes were part of the model 

equation. In conclusion, rejection increased by size/steric 

hindrance effects; but hydrophobicity decrease rejection due 

to adsorption and partitioning mechanisms. 

The results of PCA and the model cannot be generalized to 

experimental conditions outside of the boundary experimental 

conditions defined in Section 3. For instance, excessive 

changes of pH affect the ionic speciation of charged 

compounds, obviously pK a values of solutes and pH of feed 

waters will determine boundary conditions for applicability of 

the model. Changes of membrane properties such as charge 

and pore size due to swelling will also influence the model. 

Other considerations for application of the model are the type 

of membrane used (aromatic polyamide), fluxes, pressures 

and cross flow velocities. Cellulose acetate or even different 

membrane chemistry than NF-90 or NF-200 will influence the 

PCA and the model. Nonetheless, the approach is valid and 

can be generalized under certain conditions in upscale NF 

applications. 

4.3. QSAR model after partial least squares 

regression and MLR 

The following variables: MW, solubility, MV, MWCO, ZP, CA, P, 

v, J, Jo/k, recovery and width, were removed after partial least 

squares (PLS) regression. Therefore, the final PLS model is 

defined by variables named log Kow, log D, dipole, length, 

depth, eqwidth, PWP, SR and k. The main advantage of PLS is 

the ability to handle collinearity among the independent 

variables. In contrast, principal component analysis can not 

control collinearity. Another advantage of PLS was that the 

calculation process was simpler. However, PLS is used more as 

a predictive technique and not as an interpretive technique 

such as MLR. Therefore, in this exploratory analysis, PLS 

regression serves as a variable selection process and as 

prelude for implementation of MLR. Once again as occurred 

with PCA, a reduced number of variables simplified the 

implementation of MLR. After applying stepwise regression to 

the reduced number of variables, the model result obtained 

was also Eq. (2). 

4.4. QSAR model after multiple linear regression 

To finalize the model development under various statistical 

application scenarios, the implementation of direct multiple 

linear regression was performed. The R 2 is calculated for all 

possible subset models. Using this technique, the model with 

the largest R 2 is declared the best linear model. However, this 


technique has some disadvantages. First, the R 2 increases 

with each variable included in the model. Therefore, this 

approach encourages including all variables in the best model 

although some variables may not significantly contribute to 

the model. This approach also contradicts the principal of 

parsimony that encourages as few parameters in a model as 

possible. Thus, the application of MLR without previous data 

analysis is a possibility when the number of variables is 

limited. Another disadvantage during the present study was 

that MLR was not able to distinguish collinearity between 

variables. This was the case for variables MWCO and SR; 

MWCO can replace the role of SR in Eq. (2). The new equation 

presented an R 2 of 0.75; the F-test was 52.5 with a significance 

of w0%. The equation was the following 

265:150 eqwidth 117:356 depth þ 81:662 length 

rejection ¼ 

5:229 log D 0:272 MWCO 62:565 

(3) 

It is important to mention that Eq. (3) may have been equally 

defined in Section 4.2 if the variable selected before implementing 

PCA was MWCO and not SR as explained before; 

therefore it is not surprising that Eqs. (2) and (3) only show 

differences in two coefficients. However, considering practical 

or operational facts, it may be more difficult to determine 

changes of MWCO when fouling occurs; on the other hand, 

salt rejections tests are part of monitoring practices. In addition, 

the variable of MWCO may be difficult to define when 

a range of MWCO exists for a membrane rather than an 

approximate single MWCO value if compared to salt rejection, 

although that fact is not desirable for a membrane because 

will greatly influence rejection of solutes with sizes close to or 

in the range of MWCO. 

It may be argued that a simple stepwise MLR will produce 

the same results of a PCA or PLS followed by MLR, however, 

the application of MLR without PCA or PLS has the disadvantage 

of removing or adding appropriate variables during the 

iterative process of stepwise MLR. The stepwise MLR is only 

dependent on the fulfillment of a statistical condition; it even 

may happen that different combination of variables defines 

a good equation. However, the advantage of PCA or PLS is that 

only important variables are considered, and only those will 

be part of the final MLR implementation. 

4.5. Internal and external validation of the QSAR model 

Actual (measured) rejection values (106 rejection cases) versus 

modelled (fitted) rejections of the data used to generate the 

model are shown in Fig. 1, the dataset is provided as supplementary 

data; a 95% confidence interval shows that very few 

modelled rejections were out of that interval. Besides of a good 

fit of a model, it is necessary an assessment of the predictive 

power of the model, i.e. appropriate robustness (Eriksson 

et al., 2003). The R 2 is the most widely used measure of the 

ability of a QSAR model to reproduce the internal data in the 

training (goodness of fit), but does not explain its robustness 

and prediction power. One technique to evaluate prediction is 

the leave-one-out cross-validation technique, in which one 

case at a time is iteratively held-out from the training set and 

the rest is used for model development and the excluded case 

is predicted by the developed model (Gramatica, 2007).

380 

According to Gramatica, the predictive power of a model may 

be estimated by the goodness of prediction parameter Q 2 

leave-one-out (1-PRESS/TSS, where PRESS is the predictive 

error sum of squares and TSS is the total sum of squares). In 

general, a Q 2 > 0.5 is regarded as good and Q 2 > 0.9 as excellent 

(Eriksson et al., 2003). For the developed QSAR models, the 

model with SR (Eq. (2)) presented a Q 2 leave-one-out of 0.72, 

and the model with MWCO (Eq. (3)) presented a Q 2 leave-oneout 

of 0.72. After internal cross-validation it was demonstrated 

that Eqs. (2) and (3) were valid to model rejection, 

however, an adjustment must be made to the equation before 

using it to compare measured vs. predicted rejections for 

external databases. This adjustment was necessary to overcome 

the mathematical structure of the equation. Using 

a physical interpretation, it was evident that size parameters 

referring to variables length and equivalent width may be 

large enough to cause rejection predictions over 100%, which 

can be explained after observing positive coefficients for 

equivalent width and length. This situation may also be 

detrimental for rejection predictions of ionic compounds of 

medium to large size (0.6–1.2 nm as equivalent width) that 

mostly are rejected due to electrostatic repulsion and less 

steric hindrance. Therefore Eqs. (2) and (3) can be transformed 

to the following conditional equation 

rejection ¼ 

a b 

measured rejection (%) 

HB-ion 

HB-neu 

HL-ion 

HL-neu 

95% 

confidence 

interval 

100 if QSAR model 100 

QSAR model 

modeled rejection (%) 

length, eqwidth, 

depth, log D, SR 

R² = 0.75 

An external dataset (that gathered three different datasets) 

was selected for external validation of the QSAR model. The 

external dataset is presented as supplementary data. The first 

part of the external dataset corresponds to membrane Filmtec 

NF-90. The second part corresponds to NF membrane Trisep 

TS-80 and the third part corresponds to NF membrane Desal 

HL. Desal HL membrane has a main difference with NF-90 and 

Trisep TS-80 membranes, viz. the MWCO of Desal HL is in the 

range of 150–300 Da, while NF-90 and Trisep TS-80 were 

reported to have a MWCO of 200 Da. Therefore an average 

MWCO of 225 Da was assumed for Desal HL during the 

application of Eq. (3). It is worthwhile to mention that the 


(4) 


HB-ion 

HB-neu 

HL-ion 

HL-neu 

95% 

confidence 

interval 

modeled rejection (%) 

Fig. 1 – QSAR model of experimental database: (a) Eq. (2); and (b) Eq. (3). 

length, eqwidth, 

depth, log D, 

MWCO 

R² = 0.75 

second and third parts of the external dataset were generated 

using spiral wound membrane elements instead of flat sheet 

membranes. Fig. 2a show plotted results of measured rejections 

vs. predicted rejections after calculations with Eq. (4), for 

QSAR model with Eq. (2) (SR). According to Fig. 2aanR 2 of 0.75 

was obtained after considering all compounds of the external 

dataset. However, after observing the rejection cases by NF-90 

for bromoform (BF) and trichloroethene (TCE), it appeared that 

BF and TCE may be considered atypical results because their 

rejections did not correspond to their size and hydrophobicity 

when compared to other compounds with comparable 

molecular descriptors. According to Table 3 BF has approximately 

the same hydrophobicity and polarity as CF, but BF is 

bigger than CF; therefore, measured rejection for BF was 

expected to be higher than rejection of CF (0%) due to size 

exclusion. The measured rejection of TCE (3%) was not 

comparable to the rejection of perchloroethene (39%) 

although they have the same length but a very small difference 

in equivalent width. In an experiment conducted by Kim 

et al. (2007) rejections of BF and TCE were of 50 and 33%, 

respectively, for a low pressure reverse osmosis membrane. It 

was also observed that the measured rejection (53%) of 

linuron (LNU) may be considered as atypical observation 

because a higher rejection was expected. According to Table 3, 

monolinuron is close to LNU in size and polarity; thus 

measured rejection of monolinuron was of 77%, higher than 

53%. Also according to Table 3, carbamazepine is close to LNU 

in size and hydrophobicity; thus measured rejection of 

carbamazepine was of 94%, higher than 53%. A similar 

explanation was given for the low rejection (70%) of N-acetyl- 

L-tyrosine (NAT) by Desal HL compared to 94% rejection by 

TS-80 as can be seen in Table 3. Other non-expected rejection 

was that of 2-(1H)-quinoline (QNL), with a rejection of 22% 

when compared to 2-methoxyethanol and perchloroethene 

with rejections of 32 and 39%, respectively; even though the 

length of QNL is greater than the length of 2-methoxyethanol 

and perchloroethene, a lower rejection of QNL was observed. 

However, rejections of methacetin (MTC) and NDPA may be 

influenced by their small equivalent width as can be seen in

a b 


100 

80 

60 

40 

ETH 

MTBE 

R 2 20 

QNL 

NDPA 

0 

BF TCE 

MTC 

= 0.75 

0 20 40 60 80 100 

predicted rejection (%) 

Table 3. The response of the model for rejection of those 

particular compounds (with 2*equivalent width < or w length) 

was of over prediction, but the model presented better 

response for rejections of metribuzin, atrazin and N-acetyl-Ltyrosine 

(all with length >2*equivalent width) as observed in 

Table 3. Ethanol (ETH) and MTBE with rejections of 38 and 60%, 

respectively, were expected to be lower for Desal HL 

membrane because it was observed that rejection of ETH was 

of 9% for TS-80; and MTBE was expected to have a rejection 

compared to that of 2-methoxyethanol (32%) or perchloroethene 

(39%) due to proximities in size. After selection 


NDPA 

MTC 

NAT 

LNU 

0 

0 20 40 60 80 100 


R 2 = 0.84 

and justified separation of the mentioned rejection cases, 

Fig. 2b presents the predictions of the external dataset with an 

R 2 of 0.84. In a similar explanatory scenario, Figs. 3a,b show 

measured rejections vs. predicted rejections after calculations 

with Eq. (4), for QSAR model with equation 3 (MWCO). The 

main difference between Figs. 2b and 3b was that the model 

with MWCO (Fig. 3b) showed a lower R 2 (0.80) than the model 

with SR (R 2 ¼ 0.84), meaning that the latter had a better 

goodness of fit for external prediction response. Moreover, the 

characterization of magnesium sulphate salt rejection for 

a membrane may be preferred instead of MWCO, particularly 

Table 3 – Partial list of rejections and compound properties of external dataset. 

Compound Abb. Length Eqwidth Depth Log D Dipole Measured rejection Predicted rejection Membrane 

Chloroform 0.53 0.42 0.35 1.97 1.12 0 0 NF-90 

Ethanol 0.64 0.52 0.51 0.31 1.55 9 14 TS-80 

Ethanol ETH 0.64 0.52 0.51 0.31 1.55 38 0 Desal HL 

Carbontetrachloride 0.64 0.6 0.57 2.83 0.30 35 26 NF-90 

Bromoform BF 0.69 0.56 0.48 2.40 1.00 0 33 NF-90 

MTBE MTBE 0.77 0.63 0.59 0.94 1.37 60 25 Desal HL 

Trichloroethene TCE 0.78 0.49 0.36 2.29 0.95 3 36 NF-90 

Perchloroethene 0.78 0.59 0.45 3.40 0.11 39 46 NF-90 

2-methoxyethanol 0.87 0.52 0.51 0.77 0.25 32 34 TS-80 

2-(1H)-quinoline QNL 1.00 0.52 0.36 1.26 3.38 22 53 TS-80 

NDPA NDPA 1.16 0.60 0.53 1.36 3.40 45 69 TS-80 

NDPA NDPA 1.16 0.60 0.53 1.36 3.40 19 55 Desal HL 

Metribuzin 1.17 0.74 0.64 0.47 0.52 97 99 TS-80 

Carbamazepine 1.20 0.73 0.58 2.45 3.66 88 94 TS-80 

Linuron LNU 1.21 0.69 0.53 3.20 2.11 53 84 TS-80 

Monolinuron 1.22 0.69 0.65 2.30 2.02 79 77 TS-80 

Atrazin 1.26 0.74 0.55 2.61 3.43 91 100 TS-80 

Methacetin MTC 1.28 0.52 0.42 1.03 2.20 38 72 TS-80 

Methacetin MTC 1.28 0.52 0.42 1.03 2.20 5 58 Desal HL 

N-acetyl-L-tyrosine NAT 1.33 0.71 0.60 2.18 3.45 94 100 TS-80 

N-acetyl-L-tyrosine NAT 1.33 0.71 0.60 2.18 3.45 70 100 Desal HL 


NF-90 TS-80 HL Linear ( ) NF-90 TS-80 HL Linear ( ) 

Fig. 2 – Predicted rejections for external dataset using magnesium sulphate SR: (a) all external data; and (b) selected external 

data. 

100 

80 

60 

40 

20

382 


100 

80 

60 

40 

20 

for nanofiltration and low pressure reverse osmosis 

membranes; besides, the effect of fouling in membranes can 

also be quantified by salt rejection experiments. We can state 

that the QSAR model with SR demonstrated to be acceptable 

for the external dataset of NF-90, Trisep TS-80 and Desal HL 

with an R 2 of 0.75 and 0.84 for the total external dataset and 

justified selected external dataset, respectively. Although the 

model can be valid with limitations related to boundary 

experimental conditions mentioned in Section 3, its applicability 

and approach can be of value for the construction of 

a model with combined datasets organized in training and 

testing groups. 

5. Conclusions 

0 

ETH 

BF 

MTBE 

TCE 

QNL 

NDPA 

MTC 

R 2 = 0.74 

0 20 40 60 80 100 


– A general QSAR model equation was developed to merge 

information about interaction of membrane characteristics, 

filtration operating conditions and solute properties 

to predict rejections of emerging contaminants during 

nanofiltration. 

– The QSAR model identified that the most important 

variables that influence rejection of organic solutes were 

log D, salt rejection, equivalent width, depth and length. 

– Rejection increased by size/steric hindrance effects, 

solute hydrophobicity decreased rejection due to adsorption 

and partitioning mechanisms. 

– Salt rejection incorporated steric hindrance and electrostatic 

repulsion effects that were related to the membrane 

structure and operating conditions. 

– The use of MWCO was acceptable for modelling purposes; 

however NF membranes with a broad range of MWCO 

(pore size and distribution) may difficult estimation of 

rejection of contaminants, thus magnesium sulphate salt 

rejection may be more appropriate. 


a b 

NDPA 

MTC 

NAT 

LNU 


100 

80 

60 

40 

20 

0 

0 20 40 60 80 100 



The authors acknowledge Delft Cluster and EU Techneau 

Project for funding this project. The authors also acknowledge 

Tae-Uk Kim and Arne Verliefde for providing data and details 

of their research. The authors thank Filmtec (Dow Chemical 

Co.) for donating the membranes, Dr. Jaeweon Cho of GIST 

(Korea), Dr. Frank Sacher of TZW (Germany) and Steven 

Mookhoek of TU Delft (Netherlands) for contributing with 

analytical results and facilities. 

Appendix. 

Supplementary data 

Supplementary data associated with this article can be found, 

in the online version, at doi:10.1016/j.watres.2009.06.054 

references 

R 2 = 0.80 

NF-90 TS-80 HL Linear ( ) NF-90 TS-80 HL Linear ( ) 

Fig. 3 – Predicted rejections for external dataset using MWCO: (a) all external data; and (b) selected external data. 

Abdi, H., 2003. Partial least squares (PLS) regression. In: Lewis- 

Beck, M., Bryman, A., Futing, T. (Eds.), Encyclopaedia of Social 

Sciences Research Methods. Sage, Thousand Oaks, CA. 

Bellona, C., Drewes, J.E., 2005. The role of membrane surface 

charge and solute physico-chemical properties in the rejection 

of organic acids by NF. Journal of Membrane Science 249 (1–2), 

227–234. 

Campbell, P., Srinivasan, R., Knoell, T., Phipps, D., Ishida, K., 

Safarik, J., Cormack, T., Ridgway, H., 1999. Quantitative 

structure-activity relationship (QSAR) analysis of surfactants 

influencing attachment of a Mycobacterium sp. to cellulose

acetate and aromatic polyamide reverse osmosis membranes. 

Biotechnology and Bioengineering 64 (5), 527–544. 

Castiglioni, S., Bagnati, R., Fanelli, R., Pomati, F., Calamari, D., 

Zuccato, E., 2006. Removal of pharmaceuticals in sewage 

treatment plants in Italy. Environmental Science & 

Technology 40 (1), 357–363. 

Connell, D.W., 1990. Bioaccumulation of Xenobiotic Compounds. 

CRC Press, Boca Raton, FL, pp. 75-110. 

Cornelissen, E.R., Verdouw, J., Gijsbertsen-Abrahamse, A.J., 

Hofman, J.A.M.H., 2005. A nanofiltration retention model for 

trace contaminants in drinking water sources. Desalination 

178 (1–3), 179–192. 

Drewes, J.E., Amy, G., Kim, T.-U., Xu, P., Bellona, C., 

Oedekoven, M., Macalady, D., 2006. Rejection of Wastewater- 

Derived Micropollutants in High-Pressure Membrane 

Applications Leading to Indirect Potable Reuse Effect of 

Membranes and Micropollutant Properties. WateReuse, 

Alexandria, VA. 

Eriksson, L., Jaworska, J., Worth, A.P., Cronin, M.T.D., 

McDowell, R.M., Gramatica, P., 2003. Methods for reliability 

and uncertainty assessment and for applicability evaluations 

of classification and regression-based QSARs. Environmental 

Health Perspectives 111 (10), 1361–1375. 

Escher, B.I., Bramaz, N., Eggen, R.I.L., Richter, M., 2005. In vitro 

assessment of modes of toxic action of pharmaceuticals in 

aquatic life. Environmental Science & Technology 39 (9), 

3090–3100. 

Everitt, B.S., Dunn, G., 2001. Applied Multivariate Data Analysis. 

Arnold, London. 

Fujikawa, M., Nakao, K., Shimizub, R., Akamatsu, M., 2007. QSAR 

study on permeability of hydrophobic compounds with 

artificial membranes. Bioorganic & Medicinal Chemistry 15 

(11), 3753–3767. 

Gramatica, P., 2007. Principles of QSAR models validation: 

internal and external. QSAR & Combinatorial Science 26 (5), 

694–701. 

Heberer, T., 2002. Occurrence, fate, and removal of 

pharmaceutical residues in the aquatic environment: a review 

of recent research data. Toxicology Letters 131 (1–2), 5–17. 

Henderson, R.K., Baker, A., Parsons, S.A., Jefferson, B., 2008. 

Characterisation of algogenic organic matter extracted from 

cyanobacteria, green algae and diatoms. Water Research 42 

(13), 3435–3445. 

Hoek, E.M.V., Elimelech, M., 2003. Cake-enhanced concentration 

polarization: a new fouling mechanism for salt-rejecting 

membranes. Environmental Science & Technology 37 (24), 

5581–5588. 

Johnson, R.A., Wichern, D.W., 2007. Applied Multivariate 

Statistical Analysis. Pearson Prentice Hall, Upper Saddle 

River, NJ. 

Jolliffe, I.T., 2002. Principal Component Analysis. Springer-Verlag, 

New York. 

Jørgensen, B., Goegebeur, Y., 2006. Multivariate Data Analysis and 

Chemometrics. Department of Statistics, University of 

Southern Denmark. 

Kim, T.-U., Drewes, J.E., Summer, R.S., Amy, G., 2007. Solute 

transport model for trace organic neutral and charged 

compounds through nanofiltration and reverse osmosis 

membranes. Water Research 41 (17), 3977–3988. 

Kimura, K., Amy, G., Drewes, J., Heberer, T., Kim, T., Watanabe, Y., 

2003. Rejection of organic micropollutants (disinfection byproducts, 

endocrine disrupting compounds, and 

pharmaceutically active compounds) by NF/RO membranes. 

Journal of Membrane Science 227 (1–2), 113–121. 

Kimura, K., Toshima, S., Amy, G., Watanabe, Y., 2004. Rejection of 

neutral endocrine disrupting compounds (EDCs) and 

pharmaceutical active compounds (PhACs) by RO membranes. 



Kiso, Y., Sugiura, T., Kitao, T., Nishimura, K., 2001. Effects of 

hydrophobicity and molecular size on rejection of aromatic 

pesticides with nanofiltration membranes. Journal of 

Membrane Science 192 (1–2), 1–10. 

Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S.D., 

Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, hormones, 

and other organic wastewater contaminants in U.S. Streams, 

1999–2000: a National Reconnaissance. Environmental Science 

& Technology 36 (6), 1202–1211. 

Landau, S., Everitt, B., 2004. A Handbook of Statistical Analysis 

using SPSS. Chapman & Hall/CRC, Boca Raton, FL. 

Lee, N., Amy, G., Croue, J.-P., Buisson, H., 2004. Identification and 

understanding of fouling in low-pressure membrane (MF/UF) 

filtration by natural organic matter (NOM). Water Research 38 

(20), 4511–4523. 

Libotean, D., Giralt, J., Rallo, R., Cohen, Y., Giralt, F., Ridgway, H., 

Rodriguez, G., Phipps, D., 2008. Organic compounds passage 

through RO membranes. Journal of Membrane Science 313 

(1-2), 23–43. 

Nghiem, L.D., Schäfer, A.I., Elimelech, M., 2004. Removal of 

natural hormones by nanofiltration membranes: 

measurement, modelling, and mechanisms. Environmental 

Science & Technology 38 (6), 1888–1896. 

Nghiem, L.D., Schäfer, A.I., Elimelech, M., 2005. Pharmaceutical 

retention mechanisms by nanofiltration membranes. 

Environmental Science & Technology 39 (19), 7698–7705. 

Ozaki, H., Li, H., 2002. Rejection of organic compounds by ultralow 

pressure reverse osmosis membrane. Water Research 36 

(1), 123–130. 

Pomati, F., Castiglioni, S., Zuccato, E., Fanelli, R., Vigetti, D., 

Rossetti, C., Calamari, C., 2006. Effects of a complex mixture of 

therapeutic drugs at environmental levels on human 

embryonic cells. Environmental Science & Technology 40 (7), 

2442–2447. 

Ren, S., Das, A., Lien, E.J., 1996. QSAR analysis of membrane 

permeability to organic compounds. Journal of Drug Targeting 

4 (2), 103–107. 

Sacher, F., Lange, F.T., Brauch, H.-J., Blankenhorn, I., 2001. 

Pharmaceuticals in groundwaters: analytical methods and 

results of a monitoring program in Baden-Württemberg, 

Germany. Journal of Chromatography A 938 (1–2), 199–210. 

Sacher, F., Ehmann, M., Gabriel, S., Graf, C., Brauch, H.-J., 2008. 

Pharmaceutical residues in the river Rhinedresults of a onedecade 

monitoring programme. Journal of Environmental 

Monitoring 10 (5), 664–670. 

Santos, J.L.C., de Beukelaar, P., Vankelecom, I.F.J., Velizarov, S., 

Crespo, J.G., 2006. Effect of solute geometry and orientation on 

the rejection of uncharged compounds by nanofiltration. 

Separation and Purification Technology 50 (1), 122–131. 

Sawyer, C.N., McCarty, P.L., Parkin, G.F., 2003. Chemistry for 

Environmental Engineering and Science. McGraw-Hill, New 

York, pp. 479-520. 

Schäfer, A.I., Nghiem, L.D., Waite, T.D., 2003. Removal of the 

natural hormone estrone from aqueous solutions using 

nanofiltration and reverse osmosis. Environmental Science & 

Technology 37 (1), 182–188. 

Shim, Y., Lee, H.-J., Lee, S., Moon, S.-H., Cho, J., 2002. Effects of 

natural organic matter and ionic species on membrane surface 

charge. Environmental Science & Technology 36 (17), 3864–3871. 

Van der Bruggen, B., Vandecasteele, C., 2002. Modeling of the 

retention of uncharged molecules with nanofiltration. Water 

Research 36 (5), 1360–1368. 

Van der Bruggen, B., Schaep, J., Wilms, D., Vandecasteele, C., 

1999. Influence of molecular size, polarity and charge on the 

retention of organic molecules by nanofiltration. Journal of 

Membrane Science 156 (1), 29–41. 

Verliefde, A., Cornelissen, E., Amy, G., Van der Bruggen, B., van 

Dijk, H., 2007. Priority organic micropollutants in water

384 

sources in Flanders and the Netherlands and assessment of 

removal possibilities with nanofiltration. Environmental 

Pollution 146 (1), 281–289. 

Verliefde, A.R.D., Cornelissen, E.R., Heijman, S.G.J., Verberk, J.Q.J.C., 

Amy, G.L., Van der Bruggen, B., van Dijk, J.C., 2008. The role of 

electrostatic interactions on the rejection of organic solutes in 

aqueous solutions with nanofiltration. Journal of Membrane 

Science 322 (1), 52–66. 

Vosges, M., Braguer, J.-C., Combarnous, Y., 2008. Long-term 

exposure of male rats to low-dose ethinylestradiol (EE2) 

in drinking water: effects on ponderal growth and on 

litter size of their progeny. Reproductive Toxicology 25 

(2), 161–168. 


Xu, P., Drewes, J.E., Bellona, C., Amy, G., Kim, T.-U., Adam, M., 

Heberer, T., 2005. Rejection of emerging organic 

micropollutants in nanofiltration-reverse osmosis membrane 

applications. Water Environment Research 77 (1), 40–48. 

Yangali-Quintanilla, V., Kim, T.-U., Kennedy, M., Amy, G., 2008. 

Modeling of RO/NF membrane rejections of PhACs and organic 

compounds: a statistical analysis. Drinking Water Engineering 

and Science 1 (1), 7–15. 

Yangali-Quintanilla, V., Sadmani, A., McConville, M., 

Kennedy, M., Amy, G., 2009. Rejection of pharmaceutically 

active compounds and endocrine disrupting compounds by 

clean and fouled nanofiltration membranes. Water Research 

43 (9), 2349–2362.

Harmful algae and their potential impacts on desalination 

operations off southern California 

David A. Caron a, *, Marie-Ève Garneau a , Erica Seubert a , Meredith D.A. Howard a,b , 

Lindsay Darjany a , Astrid Schnetzer a , Ivona Cetinić a , Gerry Filteau c , Phil Lauri d , 

Burton Jones a , Shane Trussell e 

a 

Department of Biological Sciences, University of Southern California, 3616 Trousdale Parkway, Los Angeles, CA 90089-0371, USA 

b 

Southern California Coastal Water Research Project, 3535 Harbor Blvd., Suite 110, Costa Mesa, CA 92626, USA 

c 

Separation Processes, Inc., 3156 Lionshead Avenue, Suite 2, Carlsbad, CA 92010, USA 

d 

West Basin Municipal Water District, 17140 Avalon Blvd., Suite 210, Carson, CA 90746, USA 

e 

Trussell Technologies, Inc., 6540 Lusk Boulevard, Suite C175, San Diego, CA 92121, USA 

article info 


Received 2 April 2009 


12 June 2009 

Accepted 23 June 2009 

Available online 30 June 2009 

Keywords: 

Harmful algal blooms 

Desalination 

Red tides 

Phytoplankton 

Phytotoxins 




abstract 

* Corresponding author. Tel.: þ1 213 740 0203; fax: þ1 213 740 8123. 

E-mail address: dcaron@usc.edu (D.A. Caron). 


doi:10.1016/j.watres.2009.06.051 

Seawater desalination by reverse osmosis (RO) is a reliable method for augmenting 

drinking water supplies. In recent years, the number and size of these water projects have 

increased dramatically. As freshwater resources become limited due to global climate 

change, rising demand, and exhausted local water supplies, seawater desalination will play 

an important role in the world’s future water supply, reaching far beyond its deep roots in 

the Middle East. Emerging contaminants have been widely discussed with respect to 

wastewater and freshwater sources, but also must be considered for seawater desalination 

facilities to ensure the long-term safety and suitability of this emerging water supply. 

Harmful algal blooms, frequently referred to as ‘red tides’ due to their vibrant colors, are 

a concern for desalination plants due to the high biomass of microalgae present in ocean 

waters during these events, and a variety of substances that some of these algae produce. 

These compounds range from noxious substances to powerful neurotoxins that constitute 

significant public health risks if they are not effectively and completely removed by the RO 

membranes. Algal blooms can cause significant operational issues that result in increased 

chemical consumption, increased membrane fouling rates, and in extreme cases, a plant to 

be taken off-line. Early algal bloom detection by desalination facilities is essential so that 

operational adjustments can be made to ensure that production capacity remains unaffected. 

This review identifies the toxic substances, their known producers, and our present 

state of knowledge regarding the causes of toxic episodes, with a special focus on the 

Southern California Bight. 

ª 2009 Elsevier Ltd. All rights reserved.

386 


1.1. General overview of harmful algal blooms: 

a growing global concern 

Microscopic algae constitute an essential component of all 

aquatic food webs. Photosynthetic production of organic 

material by this diverse group of species comprises the 

primary source of nutrition for all heterotrophic forms of life 

in much of the world’s ocean and freshwater ecosystems. 

Microalgae can reach high abundances in the plankton during 

periods of optimal growth and reduced grazing pressure by 

herbivores. Such localized mass proliferations are known as 

algal (or phytoplankton) blooms. In addition, a small proportion 

of microalgal species are capable of producing a number 

of noxious or toxic compounds that cause a variety of adverse 

effects on ecosystem structure and function. These 

substances pose the potential for ecosystem damage, food 

web disruption and marine animal mortality, and present 

a significant human health risk through the consumption of 

contaminated seafood and, in at least one case, direct exposure 

to water or aerosols containing these toxic compounds. 

Additionally, the algal biomass and the associated organic 

load cause significant desalination operational issues, 

impacting the pretreatment system and possibly forcing the 

treatment plant to be taken off-line (Petry et al., 2007). 

Countless human deaths resulting from the consumption 

of seafood contaminated with algal toxins have been avoided 

through rigorous monitoring programs, but sea life has not 

been so fortunate. Approximately one half of all unusual 

marine mammal mortality incidents are now attributable to 

the ingestion of food or prey contaminated by harmful algal 

blooms (Ramsdell et al., 2005). Losses in revenue due to the 

direct contamination of seafood products and indirect effects 

on tourism and other uses of coastal areas have been estimated 

in the tens of millions of dollars annually in the U.S. 

states along the Pacific coast (Trainer et al., 2002). 

There is now convincing evidence that harmful algal 

bloom (HAB) events are increasing at local, regional and global 

scales worldwide (Smayda, 1990; Hallegraeff, 1993, 2003; 

Anderson et al., 2002; Glibert et al., 2005a) and along the North 

American west coast in particular (Horner et al., 1997; Trainer 

et al., 2003). This increased occurrence may be due in part to 

better detection of HAB episodes in recent years or the global 

dispersal of toxic algal species via the transport of resting 

spores in ships’ ballast waters (Hallegraeff and Bolch, 1992; 

Burkholder et al., 2007), but another very likely cause is the 

increasing impact of anthropogenic activities on coastal 

ecosystems (Smayda, 1990; Anderson et al., 2002; Glibert et al., 

2005b, 2006; Howard et al., 2007; Cochlan et al., 2008; Kudela 

et al., 2008a). Recent reports reveal extensive and, in some 

cases, newly emerging occurrences of HABs along the coasts 

of the U.S. (Fig. 1). These incidents engender a variety of 

noxious impacts on ecosystems and public health, including 

direct effects on organisms due to the production of acutely 

toxic substances, and indirect effects such as reduced availability 

of dissolved oxygen in the water column resulting from 

the decomposition of the extensive amounts of organic 

substances usually produced during such blooms. The 


Fig. 1 – Distribution of some well-known regional HAB 

issues along U.S. shores, including (a) Alaska and (b) 

Hawaii. Causes and impacts of these poisoning events are 

defined in Tables 1–3. Summarized from information 

presented on the Harmful Algae webpage (http://www. 

whoi.edu/redtide/). 

dramatic increases in biomass and organic load that accompany 

these events pose a significant threat to seawater desalination 

facilities (Gaid and Treal, 2007). 

1.2. Regional HAB issues along U.S. coastlines 

Harmful algae are present throughout U.S. coastal waters, but 

not all species are of equal concern in all regions (Fig. 1). For 

example, toxic species of the dinoflagellate genus Alexandrium 

are common over vast stretches of the U.S. coastline, but 

coastal regions of the northeastern and northwestern U.S. 

appear to experience particularly high rates of occurrence of 

toxic ‘red tides’ caused by these species. The neurotoxins 

produced by Alexandrium, called saxitoxins, cause paralytic 

shellfish poisoning (PSP) in humans when ingested through 

contaminated seafood (particularly filter-feeding shellfish). 

Similarly, several toxic species of the diatom genus Pseudonitzschia 

occur along the entire U.S. coastline but significant 

concentrations of the neurotoxin, domoic acid, produced by 

these species have historically constituted a health threat 

primarily in the northeastern and northwestern U.S. (Bates 

et al., 1989) where it has been documented as the cause of 

amnesic shellfish poisoning (ASP) in humans. However, high 

concentrations of domoic acid in the plankton and in diverse 

planktivorous organisms have been recently documented 

along the entire Pacific coast of the U.S. (Scholin et al., 2000; 

Trainer et al., 2002; Schnetzer et al., 2007), as well as in the 

Gulf of Mexico (Pan et al., 1998). Domoic acid has been 

attributed to numerous marine animal mortalities along the 

U.S. west coast. In the Gulf of Mexico, primarily along the 

west coast of Florida, extensive and recurrent blooms of the 

dinoflagellate Karenia brevis produce a suite of toxins, known 

as brevetoxins, that can be aerosolized by breaking waves and

induce neurotoxic shellfish poisoning (NSP) in people inhaling 

the aerosols (see review of Kirkpatrick et al., 2004). The 

Tampa Bay seawater desalination facility is the only operating 

seawater desalination treatment plant of significant 

size in the United States. It is located along the west coast of 

Florida and is likely to encounter algal blooms that contain 

brevetoxin. 

Less toxic blooms also take place with regional specificity. 

The pelagophyte Aureococcus anophagefferens causes ‘brown 

tides’ in coastal waters of Rhode Island, near Long Island (NY) 

and southward along the mid-Atlantic coast of the U.S. since 

1985. No specific toxins have been identified from A. anophagefferens, 

and no human fatalities have been directly attributed 

to these blooms. Nevertheless, this species appears to be 

unpalatable or inhibitory to many filter-feeding mollusks and 

has caused substantial mortality among these populations, 

including commercially valuable species (Bricelj and Lonsdale, 

1997). 

Other microalgal species can disrupt food webs or cause 

reductions in water quality without producing acutely toxic 

conditions. Among these are the ‘colorful’ red tides of the 

dinoflagellate Lingulodinium polyedrum, a yessotoxin producer, 

that have occurred periodically throughout several decades 

along the south and central Californian coasts (Horner et al., 

1997; Gregorio and Pieper, 2000). These blooms have so far 

been found to be relatively innocuous in these waters but 

massive accumulations of these cells could have significant 

impact on desalination plants because of increased turbidity, 

high suspended solids and organic loading of influent water. 

Furthermore, accumulations of cells in protected harbors can 

cause fish mortality by depleting oxygen dissolved in the 

water, further challenging influent screening and pretreatment 

systems at desalination plants. Other taxa, such as 

species of the prymnesiophyte genus Phaeocystis, produce 

substances that can lead to enormous buildups of sea foam 

along coasts (Armonies, 1989). 

1.3. Desalination, plankton and water quality issues 

Large research programs have developed within different 

geographic areas throughout the U.S. to address regional HAB 

issues. These programs are designed to study the species, 

toxins and environmental causes of HAB outbreaks. These 

efforts, as well as local, county, state and federal monitoring 

programs provide basic information for marine resource use 

and have focused almost exclusively on threats to human 

health via the consumption of contaminated seafood. Unfortunately, 

few if any of these programs provide sufficient 

information on appropriate temporal and spatial resolution 

for thoroughly assessing the potential impact of HAB events 

on reverse osmosis desalination operations. Moreover, toxin 

analyses have primarily examined the presence of these 

substances in particulate material (plankton or animal tissue, 

particularly shellfish and finfish), and therefore may be poor 

predictors for the amount of toxins that might occur in 

seawater in the dissolved state during algal blooms, which 

would be most likely to be loaded onto reverse osmosis 

membranes during desalination. 

There are two potential impacts that HABs may have on 

seawater desalination facilities: (1) algal toxins in ocean water 


pose a significant treatment challenge for the reverse osmosis 

system to ensure that these molecules are effectively removed 

and (2) increased turbidity, total suspended solids and total 

organic content resulting from algal biomass and growth 

challenge the entire desalination facility’s treatment train. 

The significance of these issues will depend on the specific 

algae forming a bloom and the toxin(s) or other substances 

that they produce, the magnitude and duration of the bloom, 

and the specific desalination process conducted. For example, 

multi-effect distillation and multi-flash distillation might be 

susceptible to (2) but would be much less affected by toxins in 

the water (1). Desalination using reverse osmosis presumably 

would be vulnerable to both issues. Therefore, for the latter 

desalination approach, a thorough understanding of HAB 

episodes in terms of incidence and seasonality, vertical and 

horizontal spatial distribution, as well as biological aspects 

such as algal composition within a geographical region could 

help optimize the design and operational efficiency of desalination 

plants employing reverse osmosis. 

This paper provides an overview of HABs occurring along 

the continental U.S. coastline with special emphasis on the 

southwestern U.S., and provides some insight on the potential 

impacts that these events may have on the seawater desalination 

process. In recent years, this geographical area has 

become a focal point of discussions regarding desalination 

(Cooley et al., 2006) because of its sizable population and the 

particularly tenuous nature of the water supply to this region. 

Although numerous issues involving the desalination process 

are now being examined (Separation Processes Inc., 2005; Gaid 

and Treal, 2007; Pankratz, 2008, 2009), very limited information 

exists on the risks that algal blooms pose to seawater 

desalination facilities. A review of the major species 

producing harmful blooms, the substances they produce, and 

information on the spatial and temporal distributions of 

blooms are presented along with some conclusions on their 

potential impacts. This paper also provides some general 

guidelines on how early detection may help prevent or minimize 

the impact of HABs on a desalination facility’s production 

capacity or its water quality. 

2. Toxin producers and toxin concentrations 

of the west coast 

A variety of toxins including several powerful neurotoxins are 

produced by microalgae, and a number of these toxins and 

potentially toxic algal species have been detected on the U.S. 

west coast (Table 1). The ability to rapidly detect and quantify 

toxic algae in natural water samples is problematic at this 

time. Many of these species are difficult to identify using light 

microscopy. For this reason, new genetic and immunological 

methods for species identification and enumeration have 

been appearing rapidly in the literature (Miller and Scholin, 

1998; Bowers et al., 2000, 2006; Coyne et al., 2001; Caron et al., 

2003; Galluzzi et al., 2004; Anderson et al., 2005; Mikulski et al., 

2005, 2008; Ahn et al., 2006; Handy et al., 2006; Moorthi et al., 

2006; Iwataki et al., 2007, 2008; Demir et al., 2008; Matsuoka 

et al., 2008). Moreover, many toxin-producing algal species 

exhibit variable toxin production in response to environmental 

conditions, and among different strains of the same species

388 

Table 1 – Planktonic species occurring along the west coast of the U.S. that are potential concerns for reverse osmosis 

operations. 

Microalgae Toxin(s) Poisoning Event References 

Diatoms 

Pseudo-nitzschia spp. 

P. australis b 

P. cuspidata b 

P. delicatissima b 

P. fraudulenta b 

P. multiseries b 

P. pungens b 

P. pseudodelicatissima b 

P. seriata a 

Dinoflagellates 

Alexandrium spp. 

A. acatenella a 

A. catenella b 

A. fundyense a 

A. hiranoi a 

A. ostenfeldii a 

A. tamarense a 


Lingulodinium polyedrum b 

Gonyaulax spinifera a 

Protoceratium reticulatum a,c 


Dinophysis spp. 

D. acuminata a 

D. acuta a 

D. caudate 

D. fortii a 

D. norvegica a 

D. rotundata a 

D. tripos a 

Prorocentrum spp. 

P. micans 

P. minimum a,d 

Raphidophytes 

Chattonella marina a 

Fibrocapsa japonica a 

Heterosigma akashiwo a 

Domoic acid (DA) Amnesic Shellfish Poisoning (ASP) 

even when isolated from the same geographic region (Smith 

et al., 2001; Trainer et al., 2001; Kudela et al., 2004). 

Laboratory experiments have revealed a wide range of 

physico-chemical factors that increase or decrease toxin 

Human effects 

Gastro-intestinal symptoms 

Neurologic symptoms 

Death 

Ecosystem effects 

Marine mammal mortalities 

Bird mortalities 

Saxitoxins (STXs) Paralytic Shellfish Poisoning (PSP) 

Human effects 


Paralysis 

Death 


Marine mammal mortalities 

Yessotoxins (YTXs) Human and ecosystem 

effects 

None reported 

Okadaic acid (OA) 

Dinophysistoxins (DTXs) 

Pectenotoxins (PTXs) 


Diarrhetic Shellfish 

Poisoning (DSP) 

Human effects 



None reported 

Brevetoxins (PbTxs) Neurotoxic Shellfish 

Poisonining (NSP) 

Human effects 

Gastroenteritis 

Neurologic symptoms 

Respiratory irritation 

and/or failure 


Marine mammal 

mortalities 

Fish mortality events 

a Reported to produce toxin. 

b Reported to produce toxin on the west coast of the United States. 

c Conflicting reports on toxicity of P. reticulatum cultures isolated from California, Washington and Florida. 

d Reported to be present on the west coast of Mexico. 

Subba Rao et al. (1988), Bates et al. (1989), 

Martin et al. (1990), Buck et al. (1992), 

Garrison et al. (1992), Rhodes et al. (1996), 

Horner et al. (1997), Lundholm et al. (1997), 

Rhodes et al. (1998), Trainer et al. (2000, 2001), 

Baugh et al. (2006) 

Sommer and Meyer (1937), 

Gaines and Taylor (1985), 

Steidinger (1993), 

Scholin et al. (1994), 

Taylor and Horner (1994), 

Jester (2008) 

Holmes et al. (1967), 

Satake et al. (1997, 1999), 

Draisci et al. (1999a), 

Paz et al. (2004, 2007), 

Armstrong and Kudela (2006), 

Rhodes et al. (2006), Howard et al. (2007) 

Holmes et al. (1967), 

Yasumoto et al. (1980), 

Murata et al. (1982), 

Yasumoto et al., (1985), Cembella (1989), 

Lee et al. (1989), Horner et al. (1997), 

Cembella (2003), Miles et al. (2004), 

Shipe et al. (2008), Sutherland (2008) 

Loeblich and Fine (1977), 

Hershberger et al. (1997), 

Gregorio and Connell (2000), 

Hard et al. (2000), 

Tyrell et al. (2002), O’Halloran et al. (2006) 

production by harmful species of algae, and which appear to be 

species-specific (see review of Granéli and Flynn, 2006). Reports 

of factors inducing toxin production have sometimes been 

conflicting, presumably indicating that multiple factors, or

perhaps generally stressful conditions, may stimulate toxin 

production. Factors affecting toxin production include: (1) 

temperature (Ono et al., 2000); (2) light intensity (Ono et al., 2000); 

(3) salinity (Haque and Onoue, 2002a,b); (4) trace metal availability, 

especially iron (Ladizinsky and Smith, 2000; Rue and 

Bruland, 2001; Maldonado et al., 2002; Wells et al., 2005; Sunda, 

2006) but also copper (Maldonado et al., 2002) andselenium 

(Mitrovic et al., 2004, 2005); (5) macronutrient availability 

including silicate (Pan et al., 1996b; Fehling et al., 2004; Kudela 

et al., 2004), phosphate (Pan et al., 1996a, 1998; Fehling et al., 

2004), nitrogen (Bates et al., 1991; Pan et al., 1998; Kudela 

et al., 2004) and combinations of nutrient limitation (Anderson 

et al., 1990; Flynn et al., 1994; John and Flynn, 2000); (6) cellular 

elemental ratios of nutrients and physiological stress (Granéli 

and Flynn, 2006; Schnetzer et al., 2007); (7) growth phase 

(Anderson et al., 1990; Bates et al., 1991; Flynn et al., 1994; 

Johansson et al., 1996; Maldonado et al., 2002; Mitrovic et al., 

2004). The precise combination(s) of environmental factors that 

select for population growth of particular algal species within 

diverse natural assemblages, and the specific conditions that 

induce toxin production, are poorly understood for most 

harmful algae. This present state of knowledge makes it difficult 

to predict the timing, duration or spatial extent of the vast 

majority of HAB events and the toxic events resulting from 

them. 

Our ability to thoroughly characterize HABs is also 

complicated by the complex array of toxins produced by algae. 

Marine algal species produce a suite of toxic components 

(Yasumoto and Murata, 1993), and unidentified toxins 

undoubtedly remain to be described. Additionally, most toxins 

are actually composed of families of closely related 

compounds. Slightly different forms of a toxin can exhibit very 

different levels of toxicity, or may be characterized differently 

by some detection methods and analytical approaches 

(Garthwaite et al., 2001; Lefebvre et al., 2008). Such complexity 

and variability can sometimes yield vague or contradictory 

conclusions regarding the exact source of toxicity in a natural 

sample (Bates et al., 1978; Paz et al., 2004, 2007). 

Finally, characterization of HAB events is complicated by 

inherent difficulties associated with linking specific toxins 

measured in natural water samples to a specific algal species 

in a complex, natural phytoplankton assemblage and, as 

noted above, the presence of toxic species in a water sample 

does not necessarily indicate the presence of toxins. Despite 

these shortcomings, there is considerable knowledge of many 

of the major algal toxins and their producers in U.S. coastal 

waters that constitute the most important potential concerns 

for desalination activities because they are the most likely to 

be encountered in ocean water intakes. 

2.1. Domoic acid 

2.1.1. Toxin description and activity 

Domoic acid (Fig. 2; Table 3) is an amino acid derivative 

belonging to the kainoid class of compounds containing three 

carboxyl groups and one secondary amino group (Wright 

et al., 1990; Jeffery et al., 2004). All four groups are charged at 

neutral pH, and the carboxyl groups become successively 

protonated as pH decreases, yielding five possible protonated 

forms of domoic acid (Quilliam, 2003; Jeffery et al., 2004). There 


are currently ten known isomers of domoic acid, including the 

isodomoic acids A through H and the domoic acid 5’ diestereomer 

(Jeffery et al., 2004). 

Domoic acid and other members of the kainoid class are 

glutamate analogues that interfere with neurochemical 

pathways by binding to glutamate receptors of brain neurons 

(Wright et al., 1990; Quilliam, 2003). The resulting effect of 

these neuroexcitants, or excitotoxins, is a continuous stimulation 

of the neurons, which can lead to rupture and/or 

eventual formation of lesions (Wright et al., 1990). Depolarized 

neurons result in short-term memory loss (Clayden et al., 

2005), which has led to the common name for the illness 

related to the consumption of seafood contaminated with 

domoic acid: amnesic shellfish poisoning (ASP). Symptoms of 

ASP include gastroenteritis (vomiting, diarrhea, abdominal 

cramps) that can be experienced in humans within 24 h after 

ingestion, and neurological symptoms of confusion, memory 

loss, disorientation, seizures, coma and/or cranial nerve 

palsies that are typically experienced within 48 h (Perl et al., 

1990; Wright et al., 1990). The number of human illnesses 

resulting from domoic acid poisoning has been few (Horner 

et al., 1997), likely due to active monitoring of fisheries. 

However, cultured blue mussels (Mytilus edilus) contaminated 

with domoic acid poisoned 107 people and killed three during 

the first major documented ASP outbreak in 1987 on Prince 

Edward Island, Canada (Perl et al., 1990). 

ASP poses a serious threat to marine wildlife, and the 

deaths of thousands of marine mammals and sea birds have 

been attributed to domoic acid intoxication (Bates et al., 1989; 

Scholin et al., 2000; Gulland et al., 2002; Caron et al., unpublished 

data). The first documented poisoning episode of 

marine animals related to domoic acid on the U.S. west coast 

was attributed to Pseudo-nitzschia australis and occurred in 

September 1991 off central California (Table 2; Buck et al., 

1992; Fritz et al., 1992). High concentrations of domoic acid 

were also detected in Washington and Oregon in the 1990s 

(Wekell et al., 1994; Adams et al., 2000; Trainer et al., 2002), and 

a decade later in coastal waters off southern California 

(Schnetzer et al., 2007). The frequency and severity of these 

toxic events appears to be increasing (Trainer et al., 2007). 

2.1.2. Producers 

The production of domoic acid and its isomers is confined to 

approximately a dozen chain-forming marine pennate diatom 

species within the genus Pseudo-nitzschia (Bates and Trainer, 

2006), a genus containing species that form long chains of 

cells attached at their ends (Fig. 3a and b). The main toxin 

producing species that have been documented on the U.S. 

west coast include: P. australis, P. delicatissima, P. fraudulenta, P. 

multiseries, P. pungens, P. pseudodelicatissima, P. seriata and 

P. cuspidata (Tables 1 and 2). These species are distinguished 

based on fine morphological features of their silica frustules 

(Fig. 3a and b). These distinctions are subtle and require 

careful electron microscopical analysis and elaborate taxonomic 

training. As a consequence, historical misidentifications 

are not unusual and debates regarding some species 

descriptions are still unresolved. 

It is surprising that the first reports of ASP on the west coast 

of the U.S. were not recorded until the 1990s, even though 

Pseudo-nitzschia species have been recorded in surveys of

390 

phytoplankton species in the Southern California Bight since 

1917 (Allen, 1922, 1924, 1928, 1936, 1940, 1941; Reid et al., 1970, 

1985; Lange et al., 1994; Fryxell et al., 1997; Thomas et al., 2001). 

Given that these species generally comprise a significant 

portion of the total diatom assemblage in these waters, it can 

be surmised that either toxin production has increased in these 

west coast species, or that poisoning events prior to the 1990s 

have occurred but have not been attributed to these diatoms. 

Historical accounts of ‘unusual animal mortality events’ along 

the U.S. west coast tend to support the latter hypothesis. 

There have been increasing numbers of toxic events 

recorded along the U.S. west coast (Table 2), notably in Puget 

Sound (Trainer et al., 2003, 2007), Monterey Bay (Vigilant and 

Silver, 2007; R. Kudela, unpubl. data), Santa Barbara Channel 

(Trainer et al., 2000; Anderson et al., 2006; Mengelt, 2006), San 

Pedro Channel (Busse et al., 2006; Schnetzer et al., 2007), 

Newport Beach (Busse et al., 2006) and San Diego (Lange et al., 

1994; Busse et al., 2006). Most recently, toxic blooms of Pseudonitzschia 

in the Long Beach-Los Angeles Harbor and San Pedro 

Channel have been particularly toxic, with some of the highest 

domoic acid concentrations recorded for the U.S. west 

coast (Caron et al., unpublished data). The increased incidence 

and severity of these toxic episodes off the western U.S. coast 

parallels the increase in frequency and intensity of harmful 


Fig. 2 – Chemical structures of commonly encountered toxins produced by microalgae in U.S. coastal waters. 

algal blooms observed globally (Smayda, 1990; Hallegraeff, 

1993, 2003; Anderson et al., 2002; Glibert et al., 2005b). 

2.2. Saxitoxins 


Saxitoxin is a complex guanidine-based alkaloid that exists as 

more than 30 identified analogues in nature (Llewellyn, 2006). 

It is the most powerful marine toxin currently known and 

among the most dangerous poisons on Earth, except for some 

venoms and bacterial toxins (Schantz et al., 1957). Due to its 

acute toxicity, saxitoxin is currently listed as a chemical 

weapon in Schedule 1 of the Chemical Weapons Convention 

(Llewellyn, 2006). Saxitoxins display a rigid tricyclic core 

(Fig. 2; Table 3) and are very stable in biological and physiological 

solutions (Rogers and Rapoport, 1980). This nitrogenrich 

molecule and its chemical relatives are polar and have 

a positive charge at pH 7.7 (Shimizu et al., 1981). Consequently, 

they are soluble in water and alcohols, and insoluble 

in organic solvents (Schantz et al., 1957). 

Saxitoxins are known to disrupt the flow of ions through 

voltage gate sodium channels (Catterall, 1992; Cestele and 

Catterall, 2000). It has also been recently discovered that they 

have the ability to bind to calcium (Su et al., 2004) and

Table 2 – Distribution and concentrations of marine toxins in plankton of confirmed toxin producers in U.S. west coast waters. 

Toxin(s) Location and year Causative species Particulate mgL 1 

(nmol L 1 Cellular 

) pg cell 1 

Dissolved pg mL 1 

(nmol L 1 ) 

Domoic 

acid 

Washington coast and Juan de 

Fuca Eddy, WA (1997, 1998) 

P. pseudodelicatissima b.d.–2.7 


3.6–8.7 

References 

b.d.–4.6 Adams et al. (2000), 

Trainer 

et al. (2001, 2002) 

Penn Cove, WA (1997) P. pungens b.d.–0.8 Trainer et al. (1998) 

P. multiseries 

P. australis 

P. pseudodelicatissima 

Washington coast, WA (2001) P. australis b.d.–0.03 Marchetti et al. (2004) 

Washington coast, WA (2003) Pseudo-nitzschia spp. (0.4–15) 2 10 4 –0.3 a 

(b.d.–4.3) a 

0.1–94.4 (1–5) 

Baugh et al. (2006) 

Puget Sound, WA (2005) P. pseudodelicatissima b.d.–14 Trainer et al. (2007) 


Central Oregon coast, OR (1998) P. australis 0.5 35 Trainer et al. (2001) 

Pt. Año Nuevo, 

San Francisco, CA (1998) 

Bolinas Bay, 

San Francisco, CA (2003) 

P. pungens 0.1–0.7 0.3–6.3 Trainer et al. (2000) 


P. australis 0.15–9.4 Howard et al. (2007) 

Monterey Bay, CA (1991, 1998) P. australis b.d.–12.3 3–37 Buck et al. (1992), 

0.1–6.7 7.2–75 

Garrison 

et al. (1992), Walz 

et al. (1994), Scholin 

et al. (2000) 

Monterey Bay, CA (1998) P. pseudodelicatissima 0.1–0.4 0.8–1.2 Trainer et al. (2000, 

P. multiseries 0.67 6 

2001) 

Monterey Bay, CA (2000) Pseudo-nitzschia spp. b.d.–24 b.d.–8491 Bargu et al. (2002, 2008) 

P. australis 

Monterey Bay, CA (2002–2003) Pseudo-nitzschia spp. 24 Vigilant and Silver 

(2007) 

Morro Bay, CA (1998) P. australis 1.3–7.4 37–78 Trainer et al. (2000, 

2001) 

San Luis Obispo, CA (2003–2005) P. australis 1.5–7.6 9–38 Mengelt (2006) 


Point Conception, CA (1998) P. australis 2.2–6.3 15–22 Trainer et al. (2000) 

Santa Barbara, CA (1998) P. australis 0.5–1.2 0.1–0.9 Trainer et al. (2000) 

P. pungens 

P. pseudodelicatissima 


water research 44 (2010) 385–416 391


Toxin(s) Location and year Causative species Particulate mgL 1 

(nmol L 1 ) 

Cellular 

pg cell 1 

Dissolved pg mL 1 

(nmol L 1 ) 

References 

Santa Barbara Channel, CA (2003) P. australis 0.03–1.7 0.14–2.1 Anderson et al. (2006) 

Santa Barbara (Santa Rosa Island 

and north San Miguel) (2004) 

Southern California Bight, 

CA (2003, 2004) 

San Diego and Orange 

counties, CA (2004) 

P. australis 6–12 b.d.–80 Mengelt (2006) 


Pseudo-nitzschia spp. 5.6–12.7 b.d.–117 Schnetzer et al. (2007) 

P. australis 

P. cuspidata 

P. australis b.d.–2.33 Busse et al. (2006) 


Saxitoxins Sequim Bay, WA (2004–2007) Alexandrium spp. 0.02–0.5 150–800 Lefebvre et al. (2008) 

Oregon coast, OR (2004) Alexandrium spp. 0.004–0.028 Jester et al. 

(unpublished data) 

Humboldt Bay, CA (2004) A. catenella 1.6–19 a 

San Mateo County coast, CA (2004) A. catenella 2.1–62.6 a 

Monterey Bay, CA (2004) A. catenella 0.6–31.3 a 

Jester (2008) 

Jester (2008) 

Jester (2008) 

Monterey Bay, CA (2003–2005) A. catenella b.d.–0.962 Jester et al. (2009b) 

Morro Bay, CA (2004) A. catenella 1.4–16.6 a 

Yessotoxin La Jolla, CA (1993) Lingulodinium polyedrum 0.002–0.02 a 

Brevetoxins Indian Inlet, Bald Eagle Creek 

and Torque Canal, DE (2000) 

b.d.: Below detection limit. 

a Toxin concentration from cells in culture. 

0–0.05 a 

Jester (2008) 

Armstrong and Kudela 

(2006) 

Howard et al. (2008) 

Chattonella cf. verruculosa 0.008–

potassium channels (Wang et al., 2003) and to be a weak 

inhibitor of neuronal nitric oxide synthase (reviewed in Llewellyn, 

2006). These activities directly affect the nervous 

system, and the consumption of seafood containing saxitoxin 

can result in serious human illness and death, commonly 

referred to as paralytic shellfish poisoning (PSP). 

Minor symptoms of PSP, such as burning or tingling sensation 

of the lips and face, dizziness, headache, salivation, intense 

thirst and perspiration, vomiting, diarrhea and stomach 

cramps, can be experienced within 30 min after the consumption 

of contaminated seafood (Llewellyn, 2006). The consumption 

of a lethal dose can result in death within hours due to 

muscular paralysis and respiratory difficulty followed by 

complete respiratory arrest. PSP outbreaks result in more than 

2000 illnesses worldwide each year, with a 5–10% mortality rate 

(Hallegraeff, 2003). PSP toxins also have adverse effects on 

marine wildlife that can cause mortalities among fish, marine 

mammal and seabird populations (Geraci et al., 1989; Montoya 

et al., 1996; Shumway et al., 2003). There are presently no records 

of unusual animal mortality events along the Californian coast 

that are attributable to saxitoxin poisoning (Jester et al., 2009b), 

but occurrence of the toxins in species consumed by humans is 

sufficient to warrant year-round monitoring. 


Saxitoxins are biosynthesized by dinoflagellates in marine 

ecosystems, most notably species within the genus Alexandrium 

(Fig. 3c), as well as Gymnodinium catenatum, Pyrodinium 

bahamense var. copressum, and by some cyanobacteria in 

freshwater ecosystems (Hallegraeff, 2003). Blooms of toxic and 

noxious dinoflagellates are often referred to as ‘red tides’ 

because of the red discoloration of water created by the 

accessory pigments of the cells. However, toxic levels of saxitoxins 

can be attained at dinoflagellate abundances that do not 

significantly discolor the water because of the exceptionally 

high potency of saxitoxins (Burkholder et al., 2006). This situation 

exists for Alexandrium in that it does not typically reach 

‘bloom’ abundances on the U.S. west coast, and constant toxin 

monitoring is necessary to identify toxic conditions (Langlois, 

2007; IOC HAB Programme, 2008; Jester et al., 2009a). 

Alexandrium species and measurable saxitoxin concentrations 

are common along the U.S. west coast (Table 1), although 

concentrations reported for this toxin typically have not been 

as high as noted along the U.S. east coast (Table 2). Thus, few 

west coast studies have contributed to our understanding the 

dynamics of Alexandrium abundances and saxitoxin production 

while ongoing research in the Gulf of Maine represents 

the most comprehensive regional study of this dinoflagellate 

(Anderson et al., 2005; McGillicuddy et al., 2005). Combined 

field observations, laboratory studies and modeling efforts 

have led to a scenario for toxic events along the northeastern 

coast of the U.S. that involve an interplay between river 

runoff, resuspension of dinoflagellate cysts from coastal 

sediments, favorable offshore growth conditions, and winds 

that generate onshore flow into coastal shellfish areas. A 

monitoring study of PSP in Puget Sound (WA) from 1993 to 

2007 underscores that the timing and location of PSP 

outbreaks and high Alexandrium abundances are highly variable 

and not easily predicted from local or large-scale climate 

data (Moore et al., 2009). However, the study points out that 


periods of warm air and low stream flow may favor saxitoxin 

accumulation in sentinel mussels (Moore et al., 2009). 

2.3. Brevetoxins 


Brevetoxins are polyether, non-polar compounds that depolarize 

cell membranes by opening voltage gate sodium ion 

channels and induce enhanced inward flux of ions into cells 

(Lin et al., 1981; Baden, 1983, 1989; Purkerson et al., 1999). 

Brevetoxins exist as two structural types and multiple analogs 

possessing various levels of potency (Baden, 1989; Cembella, 

2003; Kirkpatrick et al., 2004). The types differ in their ladderframe 

polycyclic ether structural backbones and are designated 

type A and type B (Fig. 2). The brevetoxin derivatives 

found in the marine environment (PbTx-2, PbTx-3 and PbTx-9; 

Table 3) are produced most commonly by dinoflagellate and 

raphidophyte algae and are of the structural type B (Baden, 

1989; Baden et al., 2005). 

Brevetoxins bind to site 5 of the voltage-sensitive sodium 

channel in neurons, causing these channels to remain open 

and fire repeatedly (Catterall, 1992; Cestele and Catterall, 

2000). Brevetoxin poisoning in humans is referred to as 

neurotoxic shellfish poisoning (NSP), and includes gastrointestinal 

symptoms of nausea, diarrhea and abdominal pain, 

neurologic symptoms of paresthesia, and respiratory irritation 

and/or failure (Kirkpatrick et al., 2004). 

The effects of brevetoxins on human health are well 

documented along the western coast of Florida where severe, 

nearly annual red tides caused by the dinoflagellate Karenia 

brevis release large amounts of brevetoxins into the air when 

the fragile cells are broken in breaking waves at the water’s 

edge (Kusek et al., 1999; Kirkpatrick et al., 2004). The aerosolized 

toxins constitute a significant health risk when they 

are inhaled and, as a result, K. brevis blooms are one of the 

most intensively studied and best-understood regional HABs. 

Red tides caused by K. brevis have been implicated in marine 

mammal fatalities, fish kills and human illnesses. Brevetoxins 

have not been reported from the U.S. west coast, and therefore 

no known human fatalities or health issues have yet been 

attributed to brevetoxins from that region. 


Several dinoflagellate species and a few raphidophyte species 

produce a suite of brevetoxins (Baden, 1989). K. brevis, the 

most notorious brevetoxin producer within the Gulf of 

Mexico, has not been observed on the west coast of the U.S., 

but several species of raphidophytes that are potential brevetoxin 

producers have been documented (Tables 1 and 2). 

Heterosigma akashiwo (Fig. 3e), Chattonella marina (Fig. 3f) and 

Fibrocapsa japonica have been isolated and cultured from 

coastal waters off southern California. In general there are few 

reports of significant blooms of these species on the west 

coast, although blooms of raphidophytes in San Francisco Bay 

and Delaware Inland Bays have been observed with cell 

abundances in excess of 10 8 cells L 1 (Herndon et al., 2003; 

Coyne et al., 2005). In part, this lack of information is 

a consequence of the fact that raphidophyte species are 

notoriously difficult to identify using traditional microscopical 

techniques because they preserve poorly (Hallegraeff and

394 


Hara, 1995; Throndsen, 1997). Recently developed genetic 

approaches for the identification and quantification of some 

raphidophytes are beginning to provide much-needed tools 

for studying the distributions and ecology of these HAB 

species (Handy et al., 2006; Demir et al., 2008). Despite the 

difficulties of characterizing these blooms, fish kills have been 

attributed to raphidophyte blooms on the west coast of the 

U.S. although these studies have not quantified brevetoxins 

(Hershberger et al., 1997; Hard et al., 2000). 

2.4. Diarrhetic shellfish toxins 


Toxins that cause diarrhetic shellfish poisoning (DSP) include 

okadaic acid, dinophysistoxins and pectenotoxins (Ramsdell 

et al., 2005). Okadaic acid is a monocarboxylic acid named for 

the marine sponge Halichondria okadai from which it was first 

isolated (Tachibana et al., 1981). Okadaic acid can also be 

found in natural water samples in polar and non-polar esteric 

forms (Prassopoulou et al., 2009). The first dinophysistoxin 

described was isolated from the mussel M. edilus and was 

found to be a methyl form of okadaic acid (Murata et al., 1982). 

Okadaic acid and dinophysistoxins are linear polyethers 

(Fig. 2) and the mode of action is the inhibition of protein 

phosphatases (Takai et al., 1987; Bialojan and Takai, 1988; 

Haystead et al., 1989), enzymes that play a key role in 

dephosphorylation in many biological processes including cell 

cycle regulation. The pectenotoxins are lipid soluble and differ 

structurally from other diarrhetic toxins in that they possess 

a lactone ring (Fig. 2), and not considered to be protein phosphatase 

inhibitors, but have a high actin-depolarizing action 

(Hori et al., 1999). There is speculation that pectenotoxins may 

not produce diarrhetic effects (Cembella, 2003). 

DSP toxins (Table 3) were named for the human symptoms 

resulting from the ingestion of contaminated shellfish, 

including inflammation of the intestinal tract, diarrhea, 

abdominal cramps, vomiting and nausea beginning 30 min to 

a few hours after ingestion (Hallegraeff, 2003). In addition to the 

symptoms listed above, okadaic acid is known to be a strong 

tumor promoter (Suganuma et al., 1988), although the potential 

health implications of this activity due to the ingestion of 

contaminated seafood is unknown. There are presently no 

documented cases of DSP resulting from okadaic acid, dinophysistoxins 

or pectenotoxins on the U.S. west coast, and thus 

these toxins are not regularly monitored in the marine environment. 

DSP toxins have been detected in mussels and water 

samples from California (Sutherland, 2008), so it is possible 

that DSP has occurred on the U.S. west coast but has been 

attributed to other sources of contamination. 



Okadaic acid and dinophysistoxins are produced by a few 

species of the dinoflagellate genus Prorocentrum (Cembella, 

2003) and most commonly by species of the genus Dinophysis. 

Dinophysis species present on the western U.S coast include 

D. acuminata, D. acuta, D. caudata, D. fortii, D. norvegica, D. 

rotundata, and D. tripos (Table 2). D. acuminata and D. fortii have 

been documented in Californian waters for many years 

(Bigelow and Leslie, 1930). D. acuminata produces okadaic acid 

(Yasumoto et al., 1985), D. fortii produces okadaic acid, dinophysistoxins 

and pectenotoxins (Yasumoto et al., 1980; Murata 

et al., 1982) while D. rotundata and D. tripos produce 

dinophysistoxin-1 (Lee et al., 1989). 

Dinophysis species are technically not phytoplankton, but 

heterotrophic protists that retain chloroplasts acquired from 

their prey. Dinophysis acquires its chloroplasts by preying on 

ciliates, which in turn prey on cryptophyte algae. This 

complex trophic relationship has made the culture of these 

species unsuccessful until recently (Park et al., 2006), and 

therefore no information on the environmental conditions 

that influence toxin production by these species presently 

exists. However, genus members are easily distinguished by 

light microscopy because they possess a pronounced ‘keel’ 

and other unique morphological aspects (Fig. 3d). These 

species are often encountered in plankton samples. Abundances 

of 10 3 cells L 1 are commonly encountered (Nishitani 

et al., 2005), and occasionally they may reach abundances in 

excess of 10 5 cells L 1 (Carpenter et al., 1995). 

2.5. Yessotoxin 

2.5.1. Toxin Description 

Yessotoxin (Fig. 2; Table 3) is a chiral molecule of high polarity 

due to the presence of two sulfate groups. The molecule 

consists of fused polyether rings organized into a ladder 

shaped skeleton (Yasumoto and Murata, 1993; Wright and 

Cembella, 1998), a structure similar to other ladder-like polyether 

toxins such as the ciguatera toxin complex (ciguatoxins 

and maitotoxin), gambieric acids and brevetoxins (Yasumoto 

and Murata, 1993; Wright and Cembella, 1998). There are 

nearly 100 analogs of yessotoxin that have been identified to 

date (Satake et al., 1997, 1999; Ciminiello et al., 1998, 2000, 

2001; Daiguji et al., 1998; Miles et al., 2004, 2005a,b, 2006; Paz 

et al., 2006). 

The yessotoxin class was named for the species of 

scallop, Patinopecten yessoensis, in which it was initially 

detected (Murata et al., 1987). Yessotoxin was originally 

classified in the DSP-toxin class because it was detected with 

other DSP toxins, but it appears that it does not induce 

Fig. 3 – Major algal toxin producers occurring along the U.S. west coast. (a) Pseudo-nitzschia australis, a producer of domoic 

acid; (b) Scanning electron micrograph of Pseudo-nitzschia australis; (c) Alexandrium catenella, a producer of saxitoxin; (d) 

Dinophysis sp., a producer of okadaic acid; (e) Heterosigma akashiwo, a producer of brevetoxins; (f) Chattonella marina, 

a producer of brevetoxins; (g) Cochlodinium sp.; (h) Lingulodinium polyedrum, a producer of yessotoxin; (i) Phaeocystis globosa 

colony; (j) foam produced by the prymnesiophyte, Phaeocystis accumulating along the shore; (k) Unconcentrated seawater 

from King Harbor, City of Redondo Beach, with significant discoloration due to an algal bloom; (l) Prorocentrum sp., the 

dominant organism in (k); and (m) higher magnification of Prorocentrum sp. from (l). Scale bars [ 10 mm. Photo (b) courtesy of 

Peter Miller, (c) courtesy of Carmelo Tomas, (j) courtesy of Cindi Heslin.

Table 3 – Summary of toxins that can be present in Southern California waters. MW: molecular weight. 

Toxin Properties Formula MW Mode of action References 

Domoic acid (DA) Hydrosoluble 

At pH 7: DA 3 

Saxitoxins (STXs) Hydrosoluble 

pH 7: Stable 

C 15H 21NO 6 311.14 Binds to glutamate receptors in the brain 

disrupting normal neurochemical 

transmission 

C 10H 17N 7O 4 299.3 Bind to site 1 of voltage-sensitive sodium 

channels and block sodium conductance; 

bind to calcium and potassium channels 

Brevetoxins (PbTxs) Liposoluble Bind to site 5 of voltage-sensitive sodium 

Brevetoxin 2 (PbTx 2) C 50H 70O 14 895.1 

Brevetoxin 3 (PbTx 3) C50H72O15 897.1 

Brevetoxin 9 (PbTx 9) C 50H 74O 14 899.1 

Diarrhetic shellfish toxins 

channels, shifting activation to more 

negative membrane potentials and block 

channel activation 

Okadaic acid (OA) 

Dinophysistoxins (DTXs) 

C44H68O13 805 Inhibits protein phosphatases, inhibits 

dephosphorylation of proteins 

Pectenotoxins (PTXs) Liposoluble High actin-depolarizaing action 

Yessotoxins (YTXs) Hydrosoluble C55H80O21S2Na2 1187.3 Activation of phosphodiesterase in 

the presence of external Ca 2þ ; 

Disruption of the E-cadherin–catenin 

system in epithelial cells and potentially 

disrupting its tumour suppressive functions 

Wright et al. (1990), Quilliam (2003) 

Wong et al. (1971), Wang et al. (2003), Su et al. (2004), 

Catterall (1992), Cestele and Catterall (2000) 

Lin et al. (1981), Baden (1983, 1989), 

Purkerson et al. (1999) 

Tachibana et al. (1981), Murata et al. (1982), 

Yasumoto et al. (1985), Takai et al. (1987), 

Bialojan and Takai (1988), Haystead et al. (1989), 

Hori et al. (1999) 

Murata et al. (1987), Takahashi et al. (1996), 

Alfonso et al. (2003), Ronzitti et al. (2004) 

396 


diarrhetic effects (Ogino et al., 1997). Accordingly, the regulation 

of the European Commission on marine biotoxins now 

considers yessotoxins separately from DSP toxins (European 

Commission, 2002). The great number of yessotoxin analogs 

complicates toxicity studies, and may explain the sometimes 

contradictory reports of its mode of action. Studies have 

shown that lysosomes, the immune system and the thymus 

(with tumorigenic implications) are the biological targets of 

yessotoxin (Franchini et al., 2004; Malagoli et al., 2006), while 

other reports have indicated cardiotoxic effects (Terao et al., 

1990; Ogino et al., 1997; Aune et al., 2002). The cardiotoxicity 

of yessotoxin might be attributed to phosphodiesterase 

activation in the presence of external calcium ions (Alfonso 

et al., 2003). Unlike the other marine toxins mentioned 

above, there have been no reported human health issues or 

marine mammal deaths associated with yessotoxins. 


There are three known yessotoxin-producing dinoflagellates, 

Protoceratium reticulatum, Lingulodinium polyedrum, and 

Gonyaulax spinifera, and they have all been observed in coastal 

waters off the western U.S. (Table 1). According to phylogenetic 

analyses of available rRNA gene sequences, the capacity 

for yessotoxin production appears to be restricted to the order 

Gonyaulacales (Howard et al., 2009). However, toxin production 

among strains within each species appears to be highly 

variable. 

Yessotoxin has been detected in L. polyedrum isolates 

cultured from around the globe (Tubaro et al., 1998; Draisci 

et al., 1999a; Strom et al., 2003; Paz et al., 2004), including 

isolates from Californian coastal waters (Armstrong and 

Kudela, 2006). The reported cellular concentrations in the 

latter cells ranged from below detection to 1.5 pg cell 1 , indicating 

that L. polyedrum is significantly less toxic than 

P. reticulatum or G. spinifera. Yessotoxin has been recorded in 

blue mussels at low concentrations along the U.S. west coast 

(Table 2) during red tides caused by L. polyedrum, as well as 

during non-bloom conditions, but yessotoxin production by 

this dinoflagellate appears to be less than toxin levels 

produced by isolates from other geographical regions (see 

Table 2 in Howard et al., 2008). Expansive and dense blooms of 

L. polyedrum (Fig. 3 h) have been reported in California since 

1901, but there have only been anecdotal reports of health 

problems associated with the red tides caused by L. polyedrum 

(Kudela and Cochlan, 2000). 

Yessotoxin production by isolates of P. reticulatum has 

been confirmed (Satake et al., 1997; Boni et al., 2002; Miles 

et al., 2002; Riobo et al., 2002; Stobo et al., 2003; Samdal et al., 

2004; Eiki et al., 2005), but concentrations ranging from 

below detection to 79 pg cell 1 have been reported for 

isolates from Washington, California and Florida (Paz et al., 

2004, 2007). Isolates of G. spinifera appear to be the most 

prolific yessotoxin producers. Concentrations in New Zealand 

isolates ranged from below detection to 200 pg cell 1 

(Rhodes et al., 2006), more than 20-fold higher than the 

per-cell toxicity of P. reticulatum and 600-fold higher than 

L. polyedrum. G. spinifera does not generally bloom in high 

densities on the U.S. west coast, but it has been frequently 

observed at low abundances (Howard et al., 2008; M. Silver, 

pers. comm.) and has reached bloom concentrations in 


Tomales Bay, north of San Francisco (G. Langlois pers. 

comm.). Yessotoxins are monitored in New Zealand, Europe 

and Japan but they are not routinely measured on the U.S. 

west coast. Howard et al. (2008) was the first study to confirm 

yessotoxins in California and Washington coastal waters, 

albeit at very low concentrations. 

2.6. Toxin detection and quantification 

A wide variety of methodologies and technologies exist to 

detect, characterize and quantify the major toxins produced 

by microalgae. These approaches can be broadly divided into 

those used to characterize biological activity (toxicity assays) 

and those used to identify specific chemical structure(s) 

(immunological, various analytical techniques). Because of 

the highly variable approaches employed, and in most 

cases the highly diverse set of compounds comprising a toxin 

class, the methods provide somewhat different estimates of 

absolute toxin concentrations or presumed toxicity. It is also 

noteworthy that the majority of the protocols used to measure 

algal toxins have focused on the analysis of tissue samples 

that may be a source of human contamination (e.g. shellfish or 

finfish tissue) or plankton material filtered from seawater 

samples. Relatively few studies have examined the concentrations 

of toxins dissolved in seawater (Table 2). For this 

reason, limits of detection for toxins in seawater are poorly 

known for most approaches. However, based on analytical 

approaches presently available, a practical limit of detection 

for a few of the major concerns (domoic acid and saxitoxins) is 

in the range of 0.1 mg per liter of seawater for immunological 

approaches, but a detection limit of approximately 0.01 mg per 

liter for dissolved domoic acid in seawater using highperformance 

liquid chromatography has been reported 

(Pocklington et al., 1990). Detection limits for toxins in 

particulate material in the water can be significantly lower 

because particles can be concentrated by filtration prior to 

extraction. Knowledge of the concentrations of dissolved 

toxins would be preferable from the perspective of reverse 

osmosis desalination operations because dissolved toxins 

(rather than cell-bound toxins) would most likely impact 

these membranes. 

Domoic acid has been detected and quantified in seawater, 

plankton, shellfish extract and homogenate, as well as sea bird 

and mammalian body fluid (e.g. blood, urine, amniotic 

fluid, cerebral spinal fluid). Analytical approaches for these 

measurements include commercially available immunological 

techniques (enzyme-linked immunosorbent assay; ELISA) 

(Garthwaite et al., 1998, 2001), high pressure liquid chromatography 

(HPLC) with ultraviolet (UV) diode array detection 

(DAD) (Quilliam et al., 1989; Quilliam, 2003), receptor binding 

assay (RBA) (Van Dolah et al., 1994), mouse bioassay (MBA) or 

liquid chromatography–mass spectrometry (LC–MS). The 

results obtained by these various approaches are not yet 

completely compatible or comparable, and therefore comparisons 

across studies using different analytical methods can 

be problematic. In general, the choice of an approach is 

a compromise between cost of analysis (or access to costly 

equipment), sample throughput, sensitivity and analytical goal 

(e.g. thorough chemical characterization versus overall

398 

toxicity, or human health risk versus scientific understanding 

of bloom dynamics). 

Saxitoxins can be rapidly detected and quantified with 

commercial ELISA kits, but the high specificity of these tests 

precludes the recognition of certain members of the saxitoxin 

family, especially the neo-saxitoxin (Garthwaite et al., 2001). 

Saxitoxin detection and quantification is often accomplished 

by HPLC, RBA, LC–MS, and MBA. Regulatory programs for 

seafood consumption are still based on ‘lethal mouse dosage’. 

Similarly, detection and quantification of brevetoxin and its 

derivatives in seawater, shellfish, and mammalian body fluids 

can be accomplished using commercially available ELISA kits 

(Naar et al., 2002), by HPLC, RBA (Van Dolah et al., 1994), LC–MS 

and MBA. A comparative study also quantified brevetoxins by 

radioimmunoassay (RIA) and a neuroblastoma (N2A) cytotoxicity 

assay (Twiner et al., 2007). 

Quantification of the DSP toxin suite can be underestimated 

by ELISA because commercial ELISA assays are 

usually optimized to detect okadaic acid and not the dinophysistoxins 

(Garthwaite et al., 2001). HPLC with fluorimetric 

detection (HPLC-FLD) has been commonly used to detect and 

quantify okadaic acid, its polar and non-polar esters, as well 

as dinophysistoxins (Lee et al., 1987). The MBA method has 

also been routinely used for the detection of DSP toxins. 

The detection and quantification of yessotoxin is problematic 

because of the extensive suite of derivatives that may 

exist. HPLC-FLD analysis (Yasumoto and Takizawa, 1997), 

MBA, LC-MS (Draisci et al., 1999b; Paz et al., 2006, 2008) and 

ELISA (Samdal et al., 2004, 2005) have been employed. 

2.7. Other potentially toxic, noxious 

and nuisance organisms 

Reports of newly occurring HAB species, or recognition of 

extant issues that have gone previously undocumented, are 


increasing our awareness of other potentially harmful bloomforming 

algae along the U.S. west coast. For example, blooms 

of an emerging potentially toxic organism, Cochlodinium sp. 

(Fig. 3g) off central California have recently been reported 

(Curtiss et al., 2008; Iwataki et al., 2008; Kudela et al., 2008b). 

This species is difficult to identify using light microscopy, and 

therefore researchers have begun to use gene sequences to 

provide accurate identification (Iwataki et al., 2007, 2008; 

Matsuoka et al., 2008). While this organism has only recently 

reached sufficient abundances to discolor Californian waters, 

there has already been one reported abalone loss in central 

California that appears to be associated with a bloom of 

Cochlodinium (R. Kudela pers. comm). 

Species of the dinoflagellate Prorocentrum (Fig. 3k–m) occasionally 

bloom in Californian coastal waters where they can 

attain very high abundances periodically, causing discolorations 

of the water and nuisance accumulations of algae 

(Holmes et al., 1967; Shipe et al., 2008). The Prorocentrum 

species known to occur in Californian waters have not 

yet demonstrated DSP toxin production, but species from 

elsewhere in the world are known to produce theses toxins 

(Table 1). Similarly, massive blooms of the dinoflagellate 

Akashiwo sanguinea are common in coastal waters of southern 

and central California and have recently been the cause of 

seabird mortality due to surfactant-like proteins (Jessup et al., 

2009). Although they may not be overtly toxic, these blooms 

can cause animal mortalities, deplete oxygen, and result in an 

increased organic and biomass loading to a seawater desalination 

facility. 

The prymnesiophyte Phaeocystis globosa infrequently 

attains high abundances off the Californian coast (Armonies, 

1989). This species produces single cells that are

Phaeocystis are consumed by many zooplankton species but 

the colonies are typically a poor food source. Selective feeding 

on single cells appears to favor colony formation and the 

accumulation of colonies in the water column (Netjstgaard 

et al., 2007). When released via colony destruction or algal life 

cycle events, the colony matrix material is easily worked into 

a ‘sea foam’ that can form layers many centimeters (even 

meters) thick at the ocean surface or along the coastline over 

fairly extensive regions (Fig. 3j). 

3. Spatial and temporal patterns 

of harmful algae 

A fundamental aspect of the biology of harmful algal blooms, 

and of vital importance for desalination operations, is the 

tendency for rapid and dramatic changes in the spatial and 

temporal distributions of these species. These changes occur 

rapidly across a wide range of scales, and pose challenges for 

documenting and predicting these distributions. Numerous 

approaches and instruments have been developed to characterize 

the dynamics of phytoplankton communities. These 

approaches presently have significant limitations on their 

abilities to identify species composition of a bloom, but they 

provide crucial information on the emergence and longevity 

of bloom events as well as their vertical and geographical 

extent. This information can help seawater desalination 

facilities adjust operations to ensure reliable production for 

the duration of the bloom. 

3.1. Temporal variability 

Significant temporal variations in the abundances of phytoplankton 

take place on time scales ranging from hours to 

decades. Short-term temporal variability (hours to a few weeks) 

can be a consequence of rapid population growth or consumption 

by herbivores, sinking of senescent populations, diel 

vertical migration, tidal movement, and aggregation or 

dispersal by physical processes such as water mass convergences 

or divergences. 

Diel vertical migration of several dinoflagellates has been 

attributed to geotaxis, phototaxis and nutrient status (Eppley 

et al., 1968; Blasco, 1978; Cullen and Horrigan, 1981; Levandowsky 

and Kaneta, 1987). Classic responses involve nighttime 

sinking out of surface waters to deeper water where 

nutrient concentrations are greater, and rising into surface 

waters for photosynthesis during daytime. Shifts in nutrient 

cell quotas that accompany these migrations may have 

significant implications for toxin production because cell 

toxicity can be related to nutrient status of the cells (Anderson 

et al., 1990; Flynn et al., 1994; John and Flynn, 2000; Flynn, 

2002). 

Diel-to-weekly variations in phytoplankton abundance can 

be characterized using self-contained or wirelessly networked 

sensor packages. Data collected in King Harbor of the City of 

Redondo Beach, CA (Fig. 4) demonstrate the efficacy of these 

instruments for providing high temporal resolution of chlorophyll 

a fluorescence (which approximates phytoplankton 

biomass) and pertinent environmental factors (e.g. dissolved 

oxygen and temperature) that provide insight into the factors 


controlling the pattern. These data reveal a 4-fold variation in 

phytoplankton standing stock over a two-day period. In 

addition, changes in water quality criteria were easily and 

rapidly identified (e.g. decrease in dissolved oxygen concentration 

near noontime on September 13). Daily variations in 

the latter parameter can be extreme at night during algal 

blooms. High resolution, short-term monitoring approaches 

allow rapid detection of sentinel parameters, and in turn 

provide information for the development of predictive models 

of HAB events. 

Chlorophyll a fluorescence sensors provide valuable 

information on the short-term temporal patterns of total algal 

biomass, but these instruments cannot identify noxious algal 

species within a phytoplankton assemblage. More sophisticated 

instruments now coming online offer that possibility. 

For example, the Environmental Sample Processor (http:// 

www.mbari.org/ESP) isanin situ instrument capable of performing 

real-time identifications of HAB and other microbial 

species, as well as toxin analyses such as domoic acid using 

on-board molecular analyses (Greenfield et al., 2006). Similarly, 

handheld devices now exist for the detection of some 

HAB species and toxins in the field (Casper et al., 2007). These 

rapid and highly specific analyses are becoming valuable tools 

for quick determinations of toxin presence resulting from 

algal blooms. These more costly instruments can be used to 

improve the information available on bloom composition 

once the cheaper and more readily available sensors identify 

an emerging bloom event. 

Seasonal variability of HAB species and their toxins along 

the Californian coast can be gleaned from the records of the 

Marine Biotoxin Monitoring Program (MBMP) of the California 

Department of Public Health (CDPH), a program started after 

a major domoic acid outbreak in the fall of 1991. At present, 

the annual effort involves the analysis of approximately 300 

shellfish samples for domoic acid and >1000 samples for PSP 

toxins from all fifteen Californian coastal counties (CDPH, 

2007). Shellfish toxin information provides a reasonable 

representation of toxins in the upper water column over 

seasonal and annual scales such as demonstrated in studies 

on toxic dinoflagellates (Montojo et al., 2006; Jester et al., 

2009a; Moore et al., 2009). Detailed information on the 

sampling effort is provided in the MBMP annual reports 

(Langlois, 2007). 

Distributions of domoic acid and PSP toxins along the 

Californian coast during the period 2002–2007 based on the 

MBMP data reveal both seasonal and geographical trends (Figs. 

5 and 6). Monthly averages for the 6-year period (histograms) 

and maximal concentrations (triangles and lines) showed 

detectable concentrations of PSP toxins and domoic acid in 

shellfish in nearly all months for all Californian coastal 

counties (Fig. 5). Domoic acid concentrations showed 

pronounced seasonality, with very high peaks during spring 

and much smaller peaks during fall. This temporal trend is 

concordant with previous observations along the southern 

Californian coast (Lange et al., 1994; Walz et al., 1994; Schnetzer 

et al., 2007; Shipe et al., 2008). Fall domoic acid peaks correspond 

to minor blooms of toxic Pseudo-nitzschia occasionally 

noted on the west coast of the U.S. (Bolin and Abbott, 1960; 

Buck et al., 1992; Walz et al., 1994; Trainer et al., 2002; 

Schnetzer et al., 2007).

400 

Domoic acid 

(µg/100g) 

PSP toxins 

(µg/100g) 

70000 

60000 

50000 

40000 

30000 

20000 

15000 

10000 

5000 

Seasonal peaks in domoic acid along the U.S. west coast 

differ along a latitudinal gradient. Highest domoic acid 

concentrations in Washington have been observed during the 

fall (Trainer, 2002; Office of Shellfish and Water Protection, 2008; 

Moore et al., 2009) compared to spring blooms that are common 

in southern California (Schnetzer et al., 2007). This north-south 

trend in seasonality is also evident on a smaller latitudinal scale. 

Blooms along the Californian coast have been more frequently 

observed later in the year in northern counties (Humboldt 

versus Santa Barbara counties in Fig. 6). The time lag between 

bloom periods observed from south to north may be related to 

the timing of the California Current System (CCS) upwelling 

maximum, which brings nutrients into surface waters and 

promotes phytoplankton growth. The CCS upwelling occurs in 

early spring in southern California, in June off Washington, 

0 

1600 

1400 

1200 

1000 

800 

600 

400 

300 

200 

100 

0 


Mean of monthly averages 

Absolute maximum 

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 

Fig. 5 – Seasonal variability of domoic acid (upper panel) and PSP toxin (lower panel) concentrations along the Californian 

coast. Monthly averages of data collected from 2002 and 2007 summarized for fifteen Californian coastal counties are shown 

as histograms. Also shown are the maximal values recorded during each month over the entire study period (triangles and 

solid lines). Toxin concentrations were derived from shellfish tissue. Data from CDPH (Langlois, 2007). 

and throughout summer in northern California and Oregon 

(Reid et al., 1956; Landry, 1989). 

Monthly averages as well as maximal PSP toxin concentrations 

showed less pronounced seasonal variability than 

domoic acid during the period 2002–2007, although the highest 

concentrations were recorded from July to September (Fig. 5). 

This seasonal pattern of maximal PSP is in agreement with the 

last 25 years of monitoring results (Langlois, 2007). It is also 

consistent with the 1927–1989 observations on the Californian 

coast indicating that most significant concentrations of the 

toxin take place between May and October (Price et al., 1991). 

The highest PSP toxin concentrations in shellfish have also 

been observed in the summer and the fall periods off the 

Washington and Oregon coasts (Trainer et al., 2002; Determan, 

2003).

North 

South 

Domoic acid 

(µg/100 g) 

Domoic acid 

(µg/100 g) 

14000 

12000 

10000 

8000 

6000 

4000 

2000 

0 

14000 

12000 

10000 

8000 

6000 

4000 

2000 

0 

Domoic acid 

Santa Barbara County 

Diel to seasonal temporal patterns in ASP and PSP 

outbreaks off the U.S. west coast are augmented by large 

interannual variability in the intensity and frequency of 

HABs. Interannual variations in HABs presumably are 

related, at least in part, to changes in atmospheric and 

hydrographic features modulated by the 3–7 year cycles of El 

Niño-Southern Oscillations (ENSOs) (Price et al., 1991; Horner 

et al., 1997), but the exact relationship between HABs and 

these climatic events is not clear because there are still 

relatively few observations spanning these temporal 

regimes. Warming off California during El Niño episodes 

reduces seasonal upwelling, enhances physical stratification 

in the CCS and lowers the nutricline in the water column 

(the depth at which nutrient concentrations increase 

rapidly). It is clear that these changes in water stability and 

nutrient availability have significant impacts on plankton 

productivity and community structure, but the specific 

responses of the phytoplankton communities vis-à-vis HAB 

events are not yet predictable (Barber and Chavez, 1983). 

ASP outbreaks occurred during El Niño episodes of 1991 and 

1997–98 along the coasts of Oregon, Washington and California 

(Table 2; Moore et al., 2008), but the specific factors contributing 

to toxic Pseudo-nitzschia blooms during these events 

could not be clearly identified (Horner and Postel, 1993; Trainer 

et al., 2000). Interannual variability in ASP and PSP concentrations 

in coastal waters along the Californian coast was 

evident and substantial in the MBMP dataset during the period 

2002–2007 (Fig. 6). ASP and PSP outbreaks were frequent, but 


Humboldt County 

**** * ** ** ** ** 

April May June 

** * * ** 

April May June 

PSP toxins 

(µg/100 g) 

PSP toxins 

(µg/100 g) 

140 

120 

100 

80 

60 

40 

20 

0 

140 

120 

100 

80 

60 

40 

20 

0 

PSP toxins 

Marin County 

July August September 

San Luis Obispo County 

2002 

2003 

2004 

2005 

2006 

2007 

* not detected 

** * * * * * * ** * 

July August September 

Fig. 6 – Interannual variation in toxin concentration in four Californian coastal counties from 2002 to 2007 during three 

months exhibiting high concentrations of domoic acid (April, May, June) and saxitoxin (July, August, September). Humboldt 

and Marin counties are located north of Santa Barbara and San Luis Obispo counties. Toxin concentrations were measured 

from shellfish tissues. Data from CDPH (Langlois, 2007). 

they varied in intensity and the timing of peak concentrations 

between the years within a single geographical location, and 

between southern and northern Californian counties in the 

same year and season. Notably, the magnitude of this variability 

is on the same order of magnitude as the variability 

observed on short-term (daily) or seasonal time scales. 

Little is known regarding the longer time scale fluctuations 

in HABs along the U.S. west coast. Multi-decadal fluctuations 

in ocean temperature are known to provoke shifts in the 

biological regime (Chavez et al., 2003), and it is anticipated 

that climatic shifts would affect the timing, intensity or 

frequency of phytoplankton blooms. Long-term regime shifts 

may affect the occurrence of blooms of L. polyedrum, 

a producer of yessotoxin, along the coast of southern 

California (Tables 1 and 2). The Pacific Ocean was cooler in the 

years preceding 1976, and red tides dominated by L. polyedrum 

commonly developed along the Southern California Bight 

during fall (Gregorio and Pieper, 2000). During a recent warm 

regime (1976-mid 1990s), red tides occurred during winter and 

spring and persisted until summer in the region of the Los 

Angeles River mouth (Gregorio and Pieper, 2000). Recent 

massive blooms of L. polyedrum during fall may indicate 

a return to the pre-1976 conditions (Moorthi et al., 2006). 

The generality surmised from data depicting short- to longterm 

temporal variability in phytotoxin dynamics is that 

variability can be high at all scales. Given our present state of 

understanding regarding the specific combination of forcing 

factors that give rise to this high variability, it is difficult to

402 

Fig. 7 – Two-dimensional presentations of chemical and physical data collected by an autonomous vehicle (Webb Slocum 

glider, Teledyne Webb Research, East Falmouth, MA) along a nearshore–offshore transect at Newport Beach, CA shown in (f) 

(indicated on map by the blue line). Contour plots are shown for (a) temperature, (b) chlorophyll a fluorescence, (c) salinity, 

(d) backscattered light and (e) water density. The Webb Slocum glider is an autonomous vehicle commonly employed in 

coastal ecosystems. This buoyancy-driven underwater vehicle generates horizontal motion by ascending and descending 

with pitched wings (Schofield et al., 2007). A rudder directs heading while buoyancy is controlled by pumping seawater into 

and out of the nose of the vehicle. This long-lived, low-power glider achieves horizontal velocities of approximately 25– 

30 cm s L1 with vertical velocities of 10–15 cm s L1 . 

accurately predict the timing and magnitude of toxic blooms. 

For these reasons, monitoring at multiple temporal scales is 

necessary to adequately characterize plankton dynamics. 

3.2. Spatial variability 

Spatial variability of HABs is considerable at multiple scales, 

analogous (and strongly related) to the temporal variability 

described above. Blooms can be highly localized (10s of 

meters) or expansive (100s of kilometers), and distributions 

vertically within the water column are heterogeneous over 

scales of centimeters to meters. The geographical extent and 

heterogeneous nature of the U.S. west coast results from 

differences in local hydrography that are manifested in smalland 

large-scale differences in spatial patterns of toxic blooms. 

Regional-scale variations in HAB distributions are illustrated 

by MBMP data during 2002–2007 for ASP and PSP 

concentrations observed in counties along northern and 

central California (Fig. 6). ASP events (frequency and level of 


toxicity) were generally lower in northerly Humboldt County 

during this period relative to Santa Barbara County nearly 

1000 km to the south. More recently, high domoic acid 

concentrations have been observed within the Southern California 

Bight, presumably indicating a continuing southward 

movement of the Pseudo-nitzschia spring blooms (Langlois, 

2007; Schnetzer et al., 2007). On the other hand, Marin County 

(north of San Francisco) exhibited higher monthly PSP values 

during most months than San Luis Obispo County located 

nearly 500 km to the south. This general latitudinal trend in 

PSP events is consistent with findings that the three southernmost 

counties (Los Angeles, Orange and San Diego) 

generally experience low concentrations of PSP relative to 

northern California (Price et al., 1991; Langlois, 2007). 

Regional-scale and geographical differences in ASP and PSP 

events have also been reported along the coastline of Oregon 

(Trainer et al., 2002). 

Small-scale spatial variability (horizontally and vertically) 

can also be dramatic. HAB events can be highly restricted

geographically (e.g. relegated to a protected embayment). 

Even within spatially extensive blooms, phytoplankton 

biomass is often highly discontinuous over very small spatial 

scales because of differences in local circulation, and wind or 

wave forcing. Considerable spatial variability in the abundance 

of P. australis within Monterey Bay has been observed, 

a phenomenon that has been attributed to advective forces 

(Buck et al., 1992). Discontinuities within a water column, such 

as thermoclines, haloclines, nutriclines and light absorption 

often leads to the establishment of subsurface microlayers 

where phytoplankton biomass can be many-fold elevated 

relative to algal standing stocks only centimeters above or 

below (Dekshenieks et al., 2001; Rines et al., 2002). 

Characterization of phytoplankton spatial distributions 

include approaches ranging from shore-based sampling, to 

the use of oceanographic ships, to remote sensing of largescale 

patterns using satellite imagery. Vertical profiling of 

phytoplankton assemblages can be accomplished using 

over-the-side, ship-based instrument packages and more 

recently autonomous vehicles equipped with a variety of 

sensor packages. The use of autonomous vehicles to provide 

synoptic measurements of phytoplankton biomass from 

chlorophyll a fluorescence and pertinent chemical/physical 

parameters is state-of-the-art for obtaining two-dimensional 

cross-sections or three-dimensional patterns in the water 

column (Fig. 7). 

Autonomous vehicles provide time- and depth-stamped 

measurements of a variety of parameters that can be optimized 

for a specific mission. Such an instrument deployed off Newport 

Beach, CA (blue line in Fig. 7f) during May–June 2008 yielded 

detailed patterns of chemical/biological parameters that 

provided information on the extent and vertical distribution of 

phytoplankton biomass (Fig. 7a–e). Evidence of upwelling in this 

118°10’W 

Particulate domoic acid 

(µg L-1) 

118°00’W 

0.01-0.50 

0.51-1.00 

1.01-1.50 

1.51-2.00 

2.01-2.50 

2.51-3.00 

3.01-3.50 

3.51-4.00 

4.01-4.50 

4.51-5.00 

33°45’N 

33°40’N 

33°35’N 

Fig. 8 – Spatial variability in domoic acid concentrations 

contained in the total particulate material (algal cells, other 

microbes and detritus) in surface waters within the San 

Pedro Bay area including the Long Beach-Los Angeles 

harbor area. Data were collected at 20 sampling stations 

(locations indicated by filled circles). 


spring deployment was apparent from the upward-pointing 

isotherms (temperature) and isopycnals (density) on the 

shoreward end of the transect (left sides of Fig. 7a, e). This was 

particularly evident in the temperature plot 5–10 km from shore 

between the surface and 20 m (Fig. 7a, red circle). Nutrients (not 

shown) were generally depleted in surface waters, and 

increased with decreasing temperature. This physico-chemical 

structure is concordant with a significant subsurface maximum 

in chlorophyll a concentration, indicative of a response of the 

phytoplankton assemblage to elevated nutrient concentrations 

at this depth (red circle on left in Fig. 7b), as previously observed 

(Jones et al., 2002). The cross-sectional picture provided by the 

autonomous vehicle indicated that the phytoplankton 

community had a patchy structure on the scale of meters 

(vertically) and kilometers (horizontally). Two major horizontal 

patches were observed at 6–12 km and at 15–21 km from shore 

(red circles in Fig. 7b). 

Measurements in addition to chlorophyll a fluorescence can 

add information on the observed general patterns such as 

particle size distributions derived from the optical backscatter 

spectrum. The backscatter profile obtained at a wavelength of 

532 nm indicated a patch of particles that were not of algal 

nature (Fig. 7d, red circle). The wavelength-dependent slope of 

the backscatter, which is dependent on the particle size distribution, 

can also be mapped to indicate the size-class of phytoplankton 

particles that dominate the chlorophyll a maxima. 

This information is particularly important for a seawater 

desalination facility, where the incoming particle size distribution 

is known to impact the source water filterability. 

Autonomous vehicles allow nearly synoptic measurements 

of the spatial distribution of phytoplankton and ancillary 

parameters. Many of these instruments operate for 

significant periods of time (weeks) and thereby supply 

temporal as well as spatial coverage. Knowledge of the 

temporal evolution and spatial organization of coastal marine 

systems enables a better understanding of the linkages 

between physical processes and the biological responses that 

contribute to the formation of algal blooms. Moreover, data 

from these instruments can be telemetered to the laboratory 

in near-real time and used to direct costly efforts such as 

shipboard sampling, or plan operations for land-based activities 

and measurements. 

Broad-scale, horizontal distributions can also be acquired 

via shipboard sampling programs (Fig. 8). Shipboard work is 

time and labor intensive, but onboard sample processing 

enables more sophisticated analyses than autonomous vehicles 

are presently capable of providing. Moreover, ships and 

other manned platforms permit time series studies at a single 

study site. Shipboard sampling conducted in April 2008 in the 

Long Beach-Los Angeles harbor area and the adjacent San 

Pedro Channel demonstrated considerable spatial variability 

in the distribution of HAB species and their toxins within this 

relatively small area (approximately 500 km 2 ). Results indicated 

a patchy distribution of domoic acid in particulate 

material (i.e. within phytoplankton cells) collected near the 

surface with highest concentrations in the vicinity of the 

harbor breakwater and at several offshore locations (Fig. 8). 

Intermediate regions exhibited toxin concentrations that 

were more than an order of magnitude less than these 

maxima.

404 

3.3. Environmental driving factors 

Characterizing the factors that lead to the stimulation of 

harmful algae and the production of toxins by these algae has 

been an area of very active research for decades. These studies 

involve field observations to document the spatiotemporal 

extent of blooms and toxin concentrations in plankton and 

marine life, and laboratory experiments aimed at understanding 

the key environmental factors leading to HAB events 

and toxin production. The overall results gleaned from many 

years of work group into three basic categories: (1) factors and 

conditions leading to phytoplankton blooms in general, (2) 

factors leading specifically to the growth of HAB species, and 

(3) factors leading to toxin production. 

Numerous factors have been implicated as contributors to 

the observed global expansion of HABs (Smayda, 1990; Hallegraeff, 

1993, 2003; Anderson et al., 2002; Glibert et al., 2005b). 

Phytoplankton blooms occur naturally as a consequence of 

the vertical mixing of deep, nutrient-rich waters into lighted 

surface waters. This process occurs seasonally in temperate 

environments due to winter storm events, and due to coastal 

upwelling events caused by appropriate regional wind conditions. 

There is no a priori reason why these ‘natural’ sources 

of nutrients cannot lead to HAB events, but the global increase 

in frequency and severity of HABs implies that human activities 

may be an underlying reason for this escalation. 

Eutrophication of coastal ecosystems is a growing global 

concern that has clear consequences for blooms of nearshore 

algal populations (Anderson et al., 2008; Heisler et al., 2008; 

Howarth, 2008). Nutrient enrichment has been implicated in 

harmful blooms occurring in some protected bays, but the 

linkage between nutrient discharges mediated by human 

activities and many HAB events is still unconfirmed. For 

example, field studies have shown that coastal upwelling of 

nitrate-rich waters can be a driving factor leading to toxigenic 

Pseudo-nitzschia blooms along the U.S. west coast (Horner 

et al., 2000; Scholin et al., 2000; Anderson et al., 2006) but the 

specific role of river/coastal runoff in domoic acid production 

is unclear (Scholin et al., 2000; Schnetzer et al., 2007). The 

importance of nutrient discharge into coastal waters is, of 

course, dependent on the amount of nutrients available for 

phytoplankton growth from natural sources but the latter 

term is poorly defined in most situations. Constructing 

nutrient budgets for coastal ecosystems is an area ripe for 

future work. In the meantime, it has been speculated that 

anthropogenic nutrient sources, such as elevated nutrient 

concentrations in river discharge, coastal runoff from agricultural 

land, and sewage discharge may significantly 

increase the total amounts of nutrients available for the 

growth of coastal phytoplankton (Scholin et al., 2000; Glibert 

et al., 2005a,b, 2006; Howard et al., 2007; Kudela et al., 2008a). 

There now exists a basic understanding of the general 

conditions that favor the growth of phytoplankton per se 

(Allen et al., 2008). Despite this basic understanding, there is 

still only limited information on the specific conditions that 

selectively stimulate the growth of harmful algal bloomforming 

species of phytoplankton. As a result, mathematical 

models that attempt to predict HAB events tend to be more 

correlative than deterministic (i.e., they identify the 


conditions that may promote a HAB, rather than the conditions 

that will promote a bloom). One generality is that rarely 

can one identify a ‘silver bullet’, a single parameter or set of 

circumstances that provide an accurate prediction of the 

occurrence of a particular HAB species. The environmental 

circumstances leading to the dominance of a HAB population 

over all other species of algae in a given locale are composed of 

a complex set of physical, chemical and biological conditions 

with poorly known variances, and these conditions appear to 

be species-specific. 

Biological factors contributing to the success or demise of 

individual HAB taxa include allelopathy among competing 

phytoplankton, mixotrophy by HAB species, and the deterrence 

of potential consumers via the production of noxious or 

toxic compounds (Strom et al., 2003; Burkholder et al., 2008; 

Buskey, 2008; Flynn, 2008; Smayda, 2008). These biological 

interactions presuppose a period of stable environmental 

conditions in order that the scenarios of allelopathy, grazer 

deterrence or phagotrophic activity of HAB species can play 

themselves out. This requirement may explain why the 

formation of a stable water mass appears to play a role in the 

development of some HAB events (Scholin et al., 2000). 

Models predicting the population growth of potentially 

toxic algae are necessary for understanding bloom dynamics, 

but these models must also integrate information on toxin 

induction. Many toxins do not appear to be constitutively 

produced by algae, but are induced by a variety of specific 

environmental conditions that are not completely understood. 

Silica, phosphorus, nitrogen and trace metal limitations, 

and nutrient or elemental ratios (in addition to or 

instead of absolute concentrations) have all been implicated 

in toxin induction (Pan et al., 1996a, 1998; Rue and Bruland, 

2001; Fehling et al., 2004; Wells et al., 2005; Granéli and Flynn, 

2006; Schnetzer et al., 2007). Again, species-specific differences 

(and perhaps strain-specific differences) may exist in 

the factors promoting toxin production. Physical aspects such 

as temperature and light intensity may stimulate toxin 

production by some harmful algae (Ono et al., 2000). 

Relative fluorescence 

7 

45 

6 

Relative fluorescence 

TMP 

40 

35 

5 

30 

4 

25 

3 

20 

2 

15 

10 

1 

5 

0 

0 

0 5 10 15 20 25 30 35 40 45 

Days 

Transmembrane pressure 

(TMP, psi) 

Fig. 9 – Correlation between chlorophyll a fluorescence 

(open squares, in relative fluorescence units) and 

transmembrane pressure (filled circles) in a pilot-scale 

reverse osmosis desalination system. Increased loading of 

phytoplankton biomass resulted in greater 

transmembrane pressure.


(cells L-1 ) 

Particulate domoic acid 

(µg L-1 ) 

Dissolved domoic acid 

(µg L-1 ) 

2.0x10 5 

1.5x10 5 

1.0x10 5 

5.0x10 4 

4. Desalination operations and HAB events 

A thorough understanding of the specific factors and conditions 

giving rise to harmful algal blooms and toxin production 

in coastal waters will require a great deal of additional 

research before accurate models for predicting these toxic 

events will be readily available. Until then, appropriate 

monitoring strategies to detect imminent bloom events and 

the ability to track the evolution of an active bloom, coupled 

with an understanding of the potential toxins being produced, 

0 

4.0 

3.0 

2.0 

1.0 

0.5 

0.0 

3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

0.0 

01-Mar-05 

01-May-05 


01-Jul-05 

01-Sep-05 

01-Nov-05 

Fig. 10 – Time series of abundances of Pseudo-nitzschia spp. cells (top), domoic acid concentrations in particulate material 

(middle) and dissolved in seawater (bottom) at a coastal monitoring site in El Segundo, CA. 

01-Jan-06 

01-Mar-06 

01-May-06 

01-Jul-06 

01-Sep-06 

01-Nov-06 

the toxin chemistry, and their rejection by seawater reverse 

osmosis membranes, provide a seawater desalination facility 

with the best strategy for making operational adjustments to 

ensure that the treatment plant capacity or product water 

quality remains unaffected. 

4.1. Concern with harmful algal blooms and their 

toxin production 

As seawater desalination has continued to become more costeffective 

and less energy intensive (Al-Sahlawi, 1999), many

406 

communities are planning or implementing seawater desalination 

facilities (Al-Sahlawi, 1999; Burbano et al., 2007). The 

selected pretreatment procedures and the process engineering 

that determines the ultimate facility design is entirely 

dependent upon the source water quality (e.g. suspended 

solids, turbidity, organic material content, algal cell content, 

etc.) and its variations, particularly for facilities incorporating 

open intakes (Bonnelye et al., 2004b; Gaid and Treal, 2007). 

When the seawater desalination process is performed by 

reverse osmosis membranes, the selection and proper operation 

of a pretreatment system is paramount to the success of 

the downstream desalination process (Tenzer et al., 1999; 

Bonnelye et al., 2004b; Separation Processes Inc., 2005; Gaid 

and Treal, 2007; Petry et al., 2007). 

Algal blooms are known to have significant negative 

impacts on reverse osmosis desalination facilities. A variety of 

pretreatment trains have been considered to address the 

difficult source water quality associated with algal blooms, 

where the organic and biomass load increase dramatically 

(Adin and Klein-Banay, 1986; Al Arrayedhy, 1987; Hasan Al- 

Sheikh, 1997; Watson, 1997; Abdul Azis et al., 2000; Bonnelye 

et al., 2004a,b; Burbano et al., 2007; Gaid and Treal, 2007; Petry 

et al., 2007; Peleka and Matis, 2008). Recently, microfiltration 

and ultrafiltration membrane pretreatment has been identified 

as a component of a preferred pretreatment train due to the 

consistent, high quality water produced by membrane filtration, 

especially when compared to conventional processes 

(Wilf and Schierach, 2001). However, one significant drawback 

to implementation of these modern pretreatment technologies 

is that they are as susceptible, or possibly more so, to significant 

algal blooms (Bonnelye et al., 2004a). 

An early warning system can provide information to 

a seawater desalination facility so that functional changes can 

be made to efficiently maintain operations even as source 

water quality deteriorates. Turbidity sensors offer a rapid 

measurement of the total amount of suspended particles in 

the intake water for making these decisions. On in situ 

instruments such as the Slocum glider, backscatter 

measurements yield this type of information (Fig. 7d). 

Measurements of chlorophyll a fluorescence can augment this 

surveillance approach by estimating the degree to which algal 

bimoass contributes to the total load of suspended particles. 

Responses to deteriorating water quality may include chemical 

additions, use of additional pretreatment equipment, or 

additional staff preparations (e.g. maintenance activities, 

guaranteeing all membranes and filters are clean in preparation 

for an event) to continuously deliver a high quality 

feedwater to the reverse osmosis system that will produce the 

desalinated drinking water. 

4.2. Addressing spatiotemporal variability in HAB 

abundance and early detection 

As detailed above, it is essential that a desalination facility 

incorporate a means of rapid algal bloom detection so that, 

when necessary, proper process changes can be made to 

maintain the production capacity. Sensors for detecting an 

eminent algal bloom can be located at the desalination facility 

to inform personnel regarding changes in water quality that 

are directly observed on the source water. Fig. 9 presents the 


transmembrane pressure (TMP) of a microfiltration system 

that serves as pretreatment to a pilot-scale reverse osmosis 

desalination system along with the levels of chlorophyll 

a fluorescence observed in the feedwater. It is clear from this 

figure that increased membrane fouling rates (e.g. faster daily 

rise in the TMP) were associated with increasing chlorophyll 

a fluorescence (i.e. increased algal biomass) in the source 

water. 

It is well known that higher concentrations of algae cause 

increased membrane fouling rates in microfiltration systems 

that are frequently incorporated, or considered, in today’s 

desalination facilities (Gijsbertsen-Abrahamse et al., 2006; Lee 

and Walker, 2006; Reiss et al., 2006). A more complete approach 

might include a monitoring system located offshore that 

measures some of the primary factors influencing algal 

blooms, such as nutrient monitoring in near-real time using 

new in situ sensor technology (Glibert et al., 2008). Such information 

would be useful to both the desalination facility and 

HAB researchers who are continually improving their understanding 

of the causative factors that produce HABs and their 

associated toxins. Using the information provided by offshore 

sensors, the desalination facility personnel could note trends 

and shifts in driving factors that generate algal blooms and 

make any chemical orders or perform maintenance procedures 

that have significant lead times. The same offshore sensor 

might also incorporate real-time monitors of sentinel parameters 

for changes in algal biomass, such as turbidity and chlorophyll 

a, allowing the facility to prepare for changes in 

chemical additions and redundant equipment service. 

Monitoring of basic chemical parameters of seawater (e.g. 

chlorophyll a concentrations) will provide valuable information 

for facility operations, but this activity is not sufficient to 

fully assess potentially toxic conditions that might arise from 

algae that do not require high standing stocks to constitute 

a significant toxic threat, such as Alexandrium species. 

Species-specific approaches, such as automated in situ 

instruments or laboratory-based methods, as well as chemical/immunological 

analyses that identify and quantify 

specific algal toxins are necessary to more thoroughly characterize 

the potential hazards posed by HAB species. The 

consistent removal of these potentially toxic substances 

through the reverse osmosis process is both a function of 

size (e.g. molecular weight) and charge (e.g. zeta potential) 

(Amy et al., 2005). Depending on the size and charge of 

the contaminant of concern, the rejection, or removal, by the 

reverse osmosis process will differ. It is important that we 

continue to broaden our knowledge on potentially toxic 

substances excreted by algal stock and their associated 

blooms. 

The approach for obtaining this information would be best 

complemented with knowledge of the species that are present 

regionally, the potential problems they pose (e.g. specific toxins 

and the amounts of soluble microbial products and extracellular 

polymeric substances excreted), the spatial extent of HAB 

episodes, and their seasonality. The seasonality of Pseudo-nitzschia 

spp. and domoic acid near the intake of a pilot desalination 

plant in El Segundo, CA, exemplifies the usefulness of routine 

monitoring for identifying potentially toxic conditions in coastal 

waters adjacent to a plant (Fig. 10). Abundances of Pseudo-nitzschia 

spp. and concentrations of domoic acid contained in the

algal cells or dissolved in the intake water exhibited a springtime 

peak. Knowledge of the seasonality of this toxic bloom-forming 

species allows intensive sampling of coastal waters during 

spring when toxic events are more common, improving the 

overall effectiveness of the monitoring effort and making it 

more cost effective. 

Historical and real-time information on the spatial distribution 

of HABs can provide information vital for optimizing 

design and performance of desalination operations. Local/ 

regional hydrography, and resultant algal blooms can differ 

dramatically. When constructing a new intake pipeline, the 

selection of its location (e.g. depth and distance from shore) can 

be greatly enhanced through the use of offshore monitoring 

devices and efforts to take into account the presence of any 

local accumulations of algal biomass due to currents, water 

mass convergences/divergences or internal waves, and also 

subsurface maxima in algal abundance. Properly locating 

offshore monitoring can provide significant information that 

will allow optimal location of a new intake pipeline or identification 

of issues that might affect an existing one, thereby 

significantly reducing the organic and suspended solid loads 

present in the feedwater during algal bloom events. These 

considerations will ease pretreatment operations, reduce the 

cost of water production, and help improve the facility’s 

longevity. 


5.1. Potential impacts, unresolved issues 

and research prospectus 

The presence of harmful algae in coastal waters that might be 

employed in reverse osmosis desalination pose potential 

problems for these operations that have been known to even 

cause desalination facilities to temporarily cease production 

(Tenzer et al., 1999; Pankratz, 2008). As the number of seawater 

desalination facilities continues to grow with lower costs and 

increasing demand, it is essential that these operating facilities 

develop the tools necessary to allow process changes and 

ensure capacity objectives continue to be met. Regardless of 

the pretreatment configuration, changes in source water 

quality require adjustments and these changes need to carefully 

coordinate to ensure that the reverse osmosis membranes 

are not irreversibly fouled or damaged in the process. 

Benchmark work is required to establish the effectiveness 

of the seawater reverse osmosis process in dealing with HAB 

toxins and other phytoplankton-derived substances. Even if 

advanced pretreatment technologies such as microfiltration 

are implemented upstream of the reverse osmosis process, 

passage of transparent extracellular material produced by the 

algal bloom (Alldredge et al., 1993; Hong et al., 1997) may affect 

reverse osmosis membrane performance. Additionally, the 

physical durability of phytoplankton varies greatly and the 

pretreatment process might disrupt cells and create significantly 

higher concentrations of dissolved organic substances, 

including toxins, than were originally present in the source 

water. For example, dissolved domoic acid has been observed 

in the seawater passing through the prefiltration process 

(Fig. 10, bottom panel) but it is unclear if these values are 


higher due to cell breakage as a result of the prefiltration 

process. Therefore, it is important that the international 

desalination community carefully characterize these potential 

contaminants and their removal to improve treatment 

approaches in seawater desalination. 

To our knowledge, there are no published reports on the 

effectiveness of reverse osmosis for removing dissolved algal 

toxins from seawater. Some of these toxin molecules (e.g. 

domoic acid) are near the theoretical molecular size of molecules 

rejected by reverse osmosis membranes, but experimental 

studies are required to validate the effective of this 

process on toxin removal. In addition, more information will 

be needed to understand the potential impact of discharged 

brine and pretreatment backwash water resulting from the 

reverse osmosis desalination process on the ecology of coastal 

ecosystems. The use of ferric sulfate or ferric chloride as 

a pretreatment coagulant would concentrate toxic algae and 

their associated toxins if they are present in the intake water. 

Similarly, the discharge of brine resulting from the reverse 

osmosis process would contain elevated concentrations of 

dissolved algal toxins relative to unfiltered seawater. The 

degree of concentration of these toxins would not be expected 

to be large, but the significance of these processes will depend 

on the starting concentrations in the raw intake or prefiltered 

water and the degree of concentration due to treatment. There 

is presently no information on algal toxins in these 

discharges. 

HABs on the U.S. west coast exhibit significant generalities 

across geographical and temporal scales (e.g. many of the 

same species occur throughout the region), but the details of 

bloom dynamics differ with geographic location, depth and 

season (and perhaps on interannual and decadal scales). The 

high degree of variability associated with these events makes 

constant monitoring of HABs in intake water for desalination 

a vital issue. Regional HAB programs and regulatory agencies 

along the U.S. west coast presently provide useful information 

for some known potential problems (e.g. ASP and PSP toxins) 

for end users that need information on coastal water quality. 

Awareness (and augmentation) of this information could 

improve planning and safe operation of desalination facilities. 

Monitoring of newly emerging HAB concerns (e.g. Cochlodinium 

spp.), or HABs and toxins that are presently poorly characterized 

(e.g. NSP, DSP, and yessotoxin poisoning) should also 

be implemented in the future to allow evaluation of their 

potential impacts on desalination processes. 

New technologies for toxin detection and quantification, 

and in situ monitoring of biological and chemical parameters 

are rapidly improving our ability to monitor coastal ecosystems 

and identify potentially problematic situations involving 

HABs. Advances in in situ observing technologies (sensor 

networks, autonomous sensor-equipped vehicles) provide the 

capability for obtaining unprecedented resolution in the 

spatial and temporal distributions of chemical and physical 

parameters, and some biologically important features 

(Sukhatme et al., 2007). New approaches and instruments for 

toxin detection help identify contaminated seafood products, 

and constitute sentinels for the threats of HABs to marine 

animal populations. Future uses of coastal waters for desalination 

will also benefit from, and contribute to, these 

activities.

408 

Acknowlegements 

The authors are grateful to Mr. Mark Donovan at Separation 

Processes, Inc. for providing the data for Fig. 9. The preparation 

of this manuscript was supported in part by funding from 

a contract between the West Basin Municipal Water District, 

Department of Water Resources and the University of Southern 

California, National Oceanic and Atmospheric Administration 

grants NA05NOS4781228 and NA07OAR4170008, Sea Grant 

NA07OAR4170008, National Science Foundation grants CCR- 

0120778 (Center for Embedded Networked Sensing; CENS), 

DDDAS-0540420, MCB-0703159, and a NASA Earth and Space 

Science Fellowship Grant NNX06AF88H. M.-È. Garneau was 

supported by a fellowship from the Fonds québécois de 

recherche sur la nature et les technologies (FQRNT). 

references 

Abdul Azis, P.K., Al-Tisan, I., Al-Daili, M., Green, T.N., Dalvi, A.G.I., 

Javeed, M.A., 2000. Effects of environment on source water for 

desalination plants on the eastern coast of Saudi Arabia. 

Desalination 132 (1–3), 29–40. 

Adams, N.A., Lesoing, M., Trainer, L., 2000. Environmental 

influences on domoic acid accumulation in razor clams on the 

Washington coast. Journal of Shellfish Research 19 (2), 

1007–1015. 

Adin, A., Klein-Banay, C., 1986. Pretreatment of seawater by 

flocculation and settling for particulates removal. 

Desalination 58 (3), 227–242. 

Ahn, S., Kulis, D.M., Erdner, D.L., Anderson, D.M., Walt, D.R., 2006. 

Fiber-optic microarray for simultaneous detection of multiple 

harmful algal bloom species. Applied and Environmental 

Microbiology 72 (9), 5742–5749. 

Al Arrayedhy, M., 1987. Pre- and post-treatment at the RO plant at 

RA’s Abu Jarjur, Bahrain. Desalination 63 (C), 81–94. 

Al-Sahlawi, M.A., 1999. Seawater desalination in Saudi Arabia: 

economic review and demand projections. Desalination 123 

(2–3), 143–147. 

Alfonso, A., de la Rosa, L., Vieytes, M.R., Yasumoto, T., Botana, L.M., 

2003. Yessotoxin, a novel phycotoxin, activates 

phosphodiesterase activity: effect of yessotoxin on cAMP 

levels in human lymphocytes. Biochemical Pharmacology 65 

(2), 193–208. 

Alldredge, A.L., Passow, U., Logan, B.E., 1993. The abundance and 

significance of a class of large, transparent organic particles in 

the ocean. Deep-Sea Research Part I Oceanographic Research 

Papers 40 (6), 1131–1140. 

Allen, W.E., 1922. Observations on surface distribution of marine 

diatoms between San Diego and Seattle. Ecology 3 (2), 140–145. 

Allen, W.E., 1924. Surface catches of marine diatoms and 

dinoflagellates made by U.S.S. Pioneer between San Diego and 

Seattle in 1923. In: University of California Publications in 

Zoology, vol. 26, pp. 243–248. 

Allen, W.E., 1928. Review of five years of studies on phytoplankton at 

Southern California piers, 1920–1924, inclusive. Bulletin of the 

Scripps Institution of Oceanography Technical Services 1, 

357–401. 

Allen, W.E., 1936. Occurrence of marine plankton diatoms in a tenyear 

series of daily catches in Southern California. American 

Journal of Botany 23 (2), 60–63. 

Allen, W.E., 1940. Summary of results of twenty years of 

researches on marine phytoplankton. Proceedings of the 6th 

Pacific Scientific Congress 3, 576–583. 


Allen, W.E., 1941. Twenty years’ statistical studies of marine 

plankton dinoflagellates of Southern California. American 

Midland Naturalist 26, 603–635. 

Allen, J.I., Smyth, T.J., Siddom, J.R., Holt, M., 2008. How well can 

we forecast high biomass algal biomass events in a eutrophic 

coastal sea? Harmful Algae 8 (1), 70–76. 

Amy, G., Kim, T.U., Yoon, J., Bellona, C., Drewes, J., Pellegrino, J., 

Heberer, T., 2005. Removal of micropollutants by NF/RO 

membranes. Water Science and Technology: Water Supply 5 (5), 

25–33. 

Anderson, C.R., Brzezinski, M.A., Washburn, L., Kudela, R., 2006. 

Circulation and environmental conditions during a toxigenic 

Pseudo-nitzschia australis bloom in the Santa Barbara Channel, 

California. Marine Ecology Progress Series 327, 119–133. 

Anderson, D.M., Kulis, D.M., Sullivan, J.J., Hall, S., Lee, C., 1990. 

Dynamics and physiology of saxitoxin production by the 

dinoflagellates Alexandrium spp. Marine Biology 104 (3), 511–524. 

Anderson, D.M., Glibert, P.M., Burkholder, J.M., 2002. Harmful algal 

blooms and eutrophication: nutrient sources, composition, and 

consequences. Estuaries 25 (4B), 704–726. 

Anderson, D.M., Kulis, D.M., Keafer, B.A., Gribble, K.E., Marin, R., 

Scholin, C.A., 2005. Identification and enumeration of 

Alexandrium spp. from the Gulf of Maine using molecular 

probes. Deep-Sea Research Part II 52 (19–21), 2467–2490. 

Anderson, D.M., Burkholder, J.M., Cochlan, W.P., Glibert, P.M., 

Gobler, C.J., Heil, C.A., Kudela, R.M., Parsons, M.L., Rensel, J.E.J., 

Townsend, D.W., Trainer, V.L., Vargo, G.A., 2008. Harmful algal 

blooms and eutrophication: examining linkages from selected 

coastal regions of the United States. Harmful Algae 8 (1), 39–53. 

Armonies, W., 1989. Occurrence of meiofauna in Phaeocystis 

seafoam. Marine Ecology Progress Series 53, 305–309. 

Armstrong, M., Kudela, R., 2006. Evaluation of California isolates 

of Lingulodinium polyedrum for the production of yessotoxin. 

African Journal of Marine Science 28 (2), 399–401. 

Aune, T., Sørby, R., Yasumoto, T., Ramstad, H., Landsverk, T., 2002. 

Comparison of oral and intraperitoneal toxicity of yessotoxin 

towards mice. Toxicon 40 (1), 77–82. 

Baden, D.G., 1983. Marine food-borne dinoflagellate toxins. 

International Review of Cytology 83, 99–150. 

Baden, D.G., 1989. Brevetoxins: unique polyether dinoflagellate 

toxins. The FASEB Journal 3 (7), 1807–1817. 

Baden, D.G., Bourdelais, A.J., Jacocks, H., Michelliza, S., Naar, J., 

2005. Natural and derivative brevetoxins: historical 

background, multiplicity and effects. Environmental Health 

Perspectives 113 (5), 621–625. 

Barber, R.T., Chavez, F.P., 1983. Biological consequences of El 

Niño. Science 222 (4629), 1203–1210. 

Bargu, S., Powell, C.L., Coale, S.L., Busman, M., Doucette, G.J., 

Silver, M.W., 2002. Krill: a potential vector for domoic acid in 

marine food webs. Marine Ecology Progres Series 237, 209–216. 

Bargu, S., Powell, C.L., Wang, Z., Doucette, G.J., Silver, M.W., 2008. 

Note on the occurrence of Pseudo-nitzschia australis and 

domoic acid in squid from Monterey Bay, CA (USA). Harmful 

Algae 7 (1), 45–51. 

Bates, H.A., Kostriken, R., Rapoport, A., 1978. The occurrence of 

saxitoxin and other toxins in various dinoflagellates. Toxicon 

16 (6), 595–601. 

Bates, S.S., Trainer, V.L., 2006. The ecology of harmful diatoms. In: 

Granéli, E., Turner, J.T. (Eds.), Ecology of Harmful Algae. 

Springer, Berlin, pp. 81–93. 

Bates, S.S., Bird, C.J., Defreitas, A.S.W., Foxall, R., Gilgan, M., 

Hanic, L.A., Johnson, G.R., McCulloch, A.W., Dodense, P., 

Pocklington, R., Quilliam, M.A., Sim, P.G., Smith, J.C., Rao, D.V. 

S., Todd, C.D., Walter, J.A., Wrigh, J.L.C., 1989. Pennate diatom 

Nitzschia pungens as the primary source of domoic acid, a toxin 

in shellfish from eastern Prince Edwards Island, Canada. 

Canadian Journal of Fisheries and Aquatic Sciences 46 (7), 

1203–1215.

Bates, S.S., DeFreitas, A.S.W., Milley, J.E., Pocklington, R., 

Quilliam, M.A., Smith, J.C., Worms, J., 1991. Controls on 

domoic acid production by the diatom Nitzschia pungens f. 

multiseries in cultures: nutrients and irradiance. Canadian 

Journal of Fisheries and Aquatic Sciences 48 (7), 1136–1144. 

Baugh, K.A., Bush, J.M., Bill, B.D., Lefebvre, K.A., Trainer, V.L., 

2006. Estimates of specific toxicity in several Pseudo-nitzschia 

species from the Washington coast, based on culture and field 

studies. African Journal of Marine Science 28 (2), 403–407. 

Bialojan, C., Takai, A., 1988. Inhibitory effect of a marine-sponge 

toxin, okadaic acid, on protein phosphatases. Specificity and 

kinetics. Biochemical Journal 256 (1), 283–290. 

Bigelow, H.B., Leslie, M., 1930. Reconnaissance of the waters and 

plankton of Monterey Bay July 1928. Bulletin of the Museum of 

Comparative Zoology 70 (5), 430–581. 

Blasco, D., 1978. Observations on the diel migration of marine 

dinoflagellates off the Baja Californian coast. Marine Biology 

46 (1), 41–47. 

Bolin, R.L., Abbott, D.P., 1960. Studies on the marine climate and 

phytoplankton of the central coastal area of California, 

1954–1960. California Cooperative Oceanic Fisheries 

Investigations 9, 23–45. 

Boni, L., Ceredi, A., Guerrini, F., Milandri, A., Pistocchi, R., Poletti, 

R., Pompei, M., 2002. Toxic Gonyaulax Grindley Reinecke in the 

North-Western Adriatic Sea (Italy). In: Tenth International 

Conference on Harmful Algal Blooms, St. Petersburg, FL 

(abstract). 

Bonnelye, V., Snaz, M.A., Comte, C., Colas, F., Plasse, L., Gueguen, F., 

2004a. Ultrafiltration versus conventional pre-treatment 

upstream RO: advantages and limits, Seoul, Korea. 

Bonnelye, V., Sanz, M.A., Durand, J.P., Plasse, L., Gueguen, F., 

Mazounie, P., 2004b. Reverse osmosis on open intake 

seawater: pre-treatment strategy. Desalination 167 (1–3), 

191–200. 

Bourdelais, A.J., Tomas, C.R., Naar, J., Kubanek, J., Baden, D.G., 2002. 

New fish-killing alga in coastal Delaware produces neurotoxins. 

Environmental Health Perspectives 110 (5), 465–470. 

Bowers, H.A., Tengs, T., Glasgow Jr., H.B., Burkholder, J.M., 

Rublee, P.A., Oldach, D.W., 2000. Development of real-time 

PCR assays for rapid detection of Pfiesteria piscicida and related 

dinoflagellates. Applied and Environmental Microbiology 66 

(11), 4641–4648. 

Bowers, H.A., Tomas, C., Tengs, T., Kempton, J.W., Lewitus, A.J., 

Oldach, D.W., 2006. Raphidophyceae (Chadefaud ex Silva) 

systematics and rapid identification: sequence analyses and 

real-time PCR assays. Journal of Phycology 42 (6), 1333–1348. 

Bricelj, V.M., Lonsdale, D.J., 1997. Aureococcus anophagefferens: 

causes and ecological consequences of brown tides in U.S. 

mid-Atlantic coastal waters. Limnology and Oceanography 42 

(5), 1023–1038. 

Buck, K.R., Uttal-Cooke, L., Pilskaln, C.H., Roelke, D.L., Villac, M.C., 

Fryxell, G.A., Cifuentes, L., Chavez, F.P., 1992. Autecology of 

the diatom Pseudo-nitzschia australis, a domoic acid producer, 

from Monterey Bay, California. Marine Ecology Progress Series 

84 (3), 293–302. 

Burbano, A.A., Adham, S.S., Pearce, W.R., 2007. The state of fullscale 

RO/NF desalination - Results from a worldwide survey. 

Journal of American Water Works Association 99 (4), 116–127. 

Burkholder, J.M., Azanza, R.V., Sako, Y., 2006. The ecology of 

harmful dinoflagellates. In: Granéli, E., Turner, J.T. (Eds.), 

Ecology of Harmful Algae. Springer, Berlin, pp. 53–66. 

Burkholder, J.M., Hallegraeff, G.M., Melia, G., Cohen, A., 

Bowers, H.A., Oldach, D.W., Parrow, M.W., Sullivan, M.J., 

Zimba, P.V., Allen, E.H., Kinder, C.A., Mallin, M.A., 2007. 

Phytoplankton and bacterial assemblages in ballast water of 

U.S. military ships as a function of port of origin, voyage time, 

and ocean exchange practices. Harmful Algae 6 (4), 486–518. 


Burkholder, J.M., Glibert, P.M., Skelton, H.M., 2008. Mixotrophy, 

a major mode of nutrition for harmful algal species in eutrophic 

waters. Harmful Algae 8 (1), 77–93. 

Buskey, E.J., 2008. How does eutrophication affect the role of 

grazers in harmful algal bloom dynamics? Harmful Algae 8 (1), 

152–157. 

Busse, L.B., Venrick, E.L., Antrobus, R., Miller, P.E., Vigilant, V., 

Silver, M.W., Mengelt, C., Mydlarz, L., Prezelin, B.B., 2006. 

Domoic acid in phytoplankton and fish in San Diego, CA, USA. 

Harmful Algae 5 (1), 91–101. 

Caron, D.A., Dennett, M.R., Moran, D.M., Schaffner, R.A., Lonsdale, D. 

J., Gobler, C.J., Nuzzi, R., McLean, T.I., 2003. Development and 

application of a monoclonal-antibody technique for counting 

Aureococcus anophagefferens, an alga causing recurrent brown 

tides in the Mid-Atlantic United States. Applied and 

Environmental Microbiology 69 (9), 5492–5502. 

Carpenter, E.J., Janson, S., Boje, R., Pollehne, F., Chang, J., 1995. 

The dinoflagellate Dinophysis norvegica: biological and 

ecological observations in the Baltic Sea. European Journal of 

Phycology 109. 

Casper, E.T., Patterson, S.S., Bhanushali, P., Farmer, A., Smith, M., 

Fries, D.P., Paul, J.H., 2007. A handheld NASBA analyzer for the 

field detection and quantification of Karenia brevis. Harmful 

Algae 6 (1), 112–118. 

Catterall, W.A., 1992. Cellular and molecular biology of voltagegated 

sodium channels. Physiological Reviews 72 (Suppl. 4.), 

S15–S48. 

CDPH, 2007. Preharvest Shellfish Protection and Marine Biotoxin 

Monitoring Program. California Department of Public Health. 

Cembella, A.D., 1989. Occurrence of okadaic acid, a major 

diarrheic shellfish toxin, in natural populations of Dinophysis 

spp. from the eastern coast of North America. Journal of 

Applied Phycology 1 (4), 307–310. 

Cembella, A.D., 2003. Chemical ecology of eukaryotic microalgae 

in marine ecosystems. Phycologica 42 (4), 420–447. 

Cestele, S., Catterall, W.A., 2000. Molecular mechanisms of 

neurotoxin action on voltage-gated sodium channels. 

Biochimie 82 (9–10), 883–892. 

Chavez, F.P., Ryan, J., Lluch-Cota, S.E., Niquen, M., 2003. From 

anchovies to sardines and back: multidecadal change in the 

Pacific ocean. Science 299 (5604), 217–221. 

Ciminiello, P., Fattorusso, E., Forino, M., Poletti, R., Viviani, R., 2000. 

Structure determination of carboxyhomoyessotoxin, a new 

yessotoxin analogue isolated from Adriatic mussels. Chemistry 

Research in Toxicology 13 (8), 770–774. 

Ciminiello, P., Fattorusso, E., Forino, M., Magno, S., Poletti, R., 2001. 

42,43,44,45,46,47,55-Heptanor-41-oxohomoyessotoxin, a new 

biotoxin from mussels of the Northern Adriatic Sea. Chemical 

Research in Toxicology 14 (5), 596–599. 

Ciminiello, P., Fattorusso, E., Forino, M., Magno, S., Poletti, R., 

Viviani, R., 1998. Isolation of adriatoxin a new analogue of 

yessotoxin from mussels of the Adriatic Sea. Tetrahedron Letters 

39 (48), 8897–8900. 

Clayden, J., Read, B., Hebditch, K., 2005. Chemistry of domoic acid, 

isodomoic acids, and their analogues. Tetrahedron 61 (24), 

5713–5724. 

Cochlan, W.P., Herndon, J., Kudela, R.M., 2008. Inorganic and organic 

nitrogen uptake by the toxigenic diatom Pseudo-nitzschia australis 

(Bacillariophyceae). Harmful Algae 8 (1), 111–118. 

Cooley, H., Gleick, P.H., Wolff, G., 2006. Desalination, with a grain 

of salt: a California perspective. In: Hart, I. (Ed.). Pacific 

Institute for Studies in Development, Environment, and 

Security, Hayward, CA, p. 88. 

Coyne, K.J., Hutchins, D.A., Hare, C.E., Cary, S.C., 2001. Assessing 

temporal and spatial variability in Pfiesteria piscicida 

distributions using molecular probing techniques. Aquatic 

Microbial Ecology 24 (3), 275–285.

410 

Coyne, K.J., Handy, S.M., Demir, E., Whereat, E.B., Hutchins, D.A., 

Portune, K.J., Doblin, M.A., Cary, S.C., 2005. Improved 

quantitative real-time PCR assays for enumeration of harmful 

algal species in field samples using an exogenous DNA reference 

standard. Limnology and Oceanography Methods 3, 381–391. 

Cullen, J.J., Horrigan, S.G., 1981. Effects of nitrate on the diurnal 

vertical migration, carbon to nitrogen ratio, and the 

photosynthetic capacity of the dinoflagellate Gymnodinium 

splendens. Marine Biology 62 (2–3), 81–89. 

Curtiss, C.C., Langlois, G.W., Busse, L.B., Mazzillo, F., Silve, M.W., 

2008. The emergence of Cochlodinium along the California 

Coast (USA). Harmful Algae 7 (3), 337–346. 

Daiguji, M., Satake, M., Ramstad, H., Aune, T., Naoki, H., 

Yasumoto, T., 1998. Structure and fluorometric HPLC 

determination of 1-desulfoyessotoxin, a new yessotoxin 

analog isolated from mussels from Norway. Natural Toxins 6 

(6), 235–239. 

Dekshenieks, M.M., Donaghay, P.L., Sullivan, J.M., Rines, J.E.B., 

Osborn, T.R., Twardowski, M.S., 2001. Temporal and spatial 

occurrence of thin phytoplankton layers in relation to physical 

processes. Marine Ecology Progress Series 223, 61–71. 

Demir, E., Coyne, K.J., Doblin, M.A., Handy, S.M., Hutchins, D.A., 

2008. Assessment of microzooplankton grazing on Heterosigma 

akashiwo using a species-specific approach combining 

quantitative real-time PCR (QPCR) and dilution methods. 

Microbial Ecology 55 (4), 583–594. 

Determan, T., 2003. Paralytic shellfish poisoning (PSP) patterns in 

Puget Sound shellfish in 2001-A report for the Puget Sound 

Ambient Monitoring Program Office of Food Safety and 

Shellfish Programs. Washington State Department of Health, 

Olympia. 

Draisci, R., Ferretti, E., Palleschi, L., Marchiafava, C., Poletti, R., 

Milandri, A., Ceredi, A., Pompei, M., 1999a. High levels of 

yessotoxin in mussels and presence of yessotoxin and 

homoyessotoxin in dinoflagellates of the Adriatic Sea. Toxicon 

37, 1187–1193. 

Draisci, R., Palleschi, L., Giannetti, L., Lucentini, L., James, K.J., 

Bishop, A.G., Satake, M., Yasumoto, T., 1999b. New approach to 

the direct detection of known and new diarrhetic shellfish toxins 

in mussels and phytoplankton by liquid chromatography-mass 

spectrometry. Journal of Chromatrography 847 (1–2), 213–221. 

Eiki, K., Satake, M., Koike, K., Ogata, T., Mitsuya, T., Oshima, Y., 

2005. Confirmation of yessotoxin production by the 

dinoflagellate Protoceratium reticulatum in Mutsu Bay. Fisheries 

Science 71 (3), 633–638. 

Eppley, R.W., Holm-Harisen, O., Strickland, J.D.H., 1968. Some 

observations on the vertical migration of dinoflagellates. 

Journal of Phycology 4 (4), 333–340. 

European Commission, 2002. Directive Commission Decision, No. 

C2002 1001. Official Journal of the European Communities 

March 15, 2002, 1.75/63, 16.3. 

Fehling, J., Davidson, K., Bolch, C.J., Bates, S.S., 2004. Growth and 

domoic acid production by Pseudo-nitzschia seriata 

(Bacillariophyceae) under phosphate and silicate limitation. 


Flynn, K.J., 2002. Toxin production in migrating dinoflagellates: 

a modelling study of PSP producing Alexandrium. Harmful 

Algae 1 (2), 147–155. 

Flynn, K.J., 2008. Attack is not the best form of defense: lessons from 

harmful algal bloom dynamics. Harmful Algae 8 (1), 129–139. 

Flynn, K., Franco, J.M., Fernandez, P., Reguera, B., Zapata, M., 

Wood, G., Flynn, K.J., 1994. Changes in toxin content, biomass 

and pigments of the dinoflagellate Alexandrium minutum during 

nitrogen refeeding and growth into nitrogen or phosphorus 

stress. Marine Ecology Progres Series 111, 99–109. 

Franchini, A., Marchesini, E., Poletti, R., Ottaviani, E., 2004. Lethal 

and sub-lethal yessotoxin dose-induced morpho-functional 


alterations in intraperitoneal injected Swiss CD1 mice. Toxicon 

44 (1), 83–90. 

Fritz, L., Quilliam, M.A., Wright, J.L.C., Beale, A.M., Work, T.M., 

1992. An outbreak of domoic acid poisoning attributed to the 

pennate diatom Pseudo-nitzschia australis. Journal of Phycology 

28 (4), 439–442. 

Fryxell, G.A., Villac, M.C., Shapiro, L.P., 1997. The occurrence of 

the toxic diatom genus Pseudo-nitzschia (Bacillariophyceae) on 

the West Coast of the USA, 1920–1996: a review. Phycologia 36 

(6), 419–437. 

Gaid, K., Treal, Y., 2007. Le dessalement des eaux par osmose 

inverse: l’expÈrience de VÈolia Water. Desalination 203 (1–3), 

1–14. 

Gaines, G., Taylor, F.J.R., 1985. An exploratory analysis of PSP 

patterns in British Columbia: 1942–1984. In: Anderson, D.M., 

White, A.W., Baden, D.G. (Eds.), Toxic dinoflagellates. Elsevier 

Science, New York, pp. 439–444. 

Galluzzi, L., Penna, A., Bertozzini, E., Vila, M., Garces, E., 

Magnani, M., 2004. Development of a real-time PCR assay for 

rapid detection and quantification of Alexandrium minutum 

(a dinoflagellate). Applied and Environmental Microbiology 70 

(2), 1199–1206. 

Garrison, D.L., Conrad, S.M., Eilers, P.P., Waldron, E.M., 1992. 

Confirmation of domoic acid production by Pseudo-nitzschia 

australis (Bacillariophyceae) cultures. Journal of Phycology 28 

(5), 604–607. 

Garthwaite, I., Ross, K.M., Miles, C.O., Hansen, R.P., Foster, D., 

Wilkins, A.L., Towers, N.R., 1998. Polyclonal antibodies to 

domoic acid, and their use in immunoassays for domoic acid 

in sea water and shellfish. Natural Toxins 6 (3–4), 93–104. 

Garthwaite, I., Ross, K.M., Miles, C.O., Briggs, L.R., Towers, N.R., 

Borrell, T., Busby, P., 2001. Integrated Enzyme-Linked 

Immunosorbent Assay screening system for amnesic, 

neurotoxic, diarrhetic, and paralytic shellfish poisoning toxins 

found in New Zealand. Journal of AOAC International 84 (5), 

1643–1648. 

Geraci, J.R., Anderson, D.M., Timperi, R.J., St. Aubin, D.J., Early, G.A., 

Prescott, J.H., Mayo, C., 1989. Humpback whales (Megaptera 

novaeangliae) fatally poisoned by dinoflagellate toxin. Canadian 

Journal of Fisheries and Aquatic Sciences 46 (11), 1895–1898. 

Gijsbertsen-Abrahamse, A.J., Schmidt, W., Chorus, I., Heijman, S.G.J., 

2006. Removal of cyanotoxins by ultrafiltration and 

nanofiltration. Journal of Membrane Science 276 (1–2), 252–259. 

Glibert, P.M., Anderson, D.M., Gentien, P., Granéli, E., Sellner, K.G., 

2005a. The global, complex phenomena of harmful algal 

blooms. Oceanography 18 (2), 137–147. 

Glibert, P.M., Seitzinger, S., Heil, C.A., Burkholder, J.M., Parrow, M.W., 

Codispoti, L.A., Kelly, V., 2005b. The role of eutrophication in 

the global proliferation of harmful algal blooms. Oceanography 

18 (2), 198–209. 

Glibert, P.M., Harrison, J., Heil, C., Seitzinger, S., 2006. Escalating 

worldwide use of urea – a global change contributing to 

coastal eutrophication. Biogeochemistry 77 (3), 441–463. 

Glibert, P.M., Kelly, V., Alexander, J., Cosdispoti, L.A., Boicourt, W.C., 

Trice, T.M., Michael, B., 2008. In situ nutrient monitoring: 

a tool for capturing nutrient variability and the antecedent 

conditions that support algal blooms. Harmful Algae 8 (1), 

175–181. 

Granéli, E., Flynn, K., 2006. Chemical and physical factors 

influencing toxin content. In: Granéli, E., Turner, J.T. (Eds.), 

Ecology of Harmful Algae. Springer, Berlin, pp. 229–241. 

Greenfield, D.I., Marin, R., Jensen, S., Massion, E., Roman, B., 

Feldman, J., Scholin, C., 2006. Application of the Environmental 

Sample Processor (ESP) methodology for quantifying Pseudonitzschia 

australis using ribosomal RNA-targeted probes in 

sandwich and fluorescent in situ hybridization. Limnology and 

Oceanography Methods 4, 426–435.

Gregorio, D.E., Connell, L., 2000. Range of Heterosigma akashiwo 

(Raphidophyceae) expanded to include California, USA. In: 

Ninth International Conference on Harmful Algal Blooms, 

Tasmania, Australia (abstract). 

Gregorio, D.E., Pieper, R.E., 2000. Investigations of red tides along 

the southern California. Bulletin of Southern California 

Academy of Sciences 99 (8), 147–160. 

Gulland, F.M.D., Haulena, M., Fauquier, D., Langlois, G., Lander, M.E., 

Zabka, T., Duerr, R., 2002. Domoic acid toxicity in California sea 

lions (Zalophus californianus): clinical signs, treatment and 

survival. Veterinary Record 150 (15), 475–480. 

Hallegraeff, G., Hara, Y., 1995. Taxonomy of harmful marine 

raphidophytes. In: Hallegraef, G., Anderson, D.M., Cembella, A.D. 

(Eds.), Manual on Harmful Marine Microalgae. IOC UNCESCO, 

Paris, pp. 365–371. 

Hallegraeff, G.M., 1993. A review of harmful algal blooms and 

their apparent global increase. Phycologia 32 (2), 79–99. 

Hallegraeff, G.M., 2003. Harmful algal blooms: a global overview. 

In: Hallegraeff, G.M., Anderson, D.M., Cembella, A.D. (Eds.), 

Manual on Harmful Marine Microalgae. Monographs on 

Oceanographic Methodology. UNESCO Publishing, Paris 

pp. 25–49. 

Hallegraeff, G.M., Bolch, C.J., 1992. Transport of diatom and 

dinoflagellate resting spores in ships’ ballast water: 

implications for plankton biogeography and aquaculture. 

Journal of Plankton Research 14 (8), 1067–1084. 

Handy, S.M., Hutchins, D.A., Cary, S.C., Coyne, K.J., 2006. 

Simultaneous enumeration of multiple raphidophyte species 

by multiprobing and mutiplexing: capabilities and limitations. 

Limnology and Oceanography: Methods 4, 193–204. 

Haque, S.M., Onoue, Y., 2002a. Effects of salinity on growth and 

toxin production of a noxious phytoflagellate, Heterosigma 

akashiwo (Raphidophyceae). Botanica Marina 45 (4), 356–363. 

Haque, S.M., Onoue, Y., 2002b. Variation in toxin compositions of 

two harmful raphidophytes, Chattonella antiqua and Chattonella 

marina, at different salinities. Environmental Toxicology 17 (2), 

113–118. 

Hard, J.J., Connell, L., Hershberger, W.K., Harrell, L.W., 2000. 

Genetic variation in mortality of chinook salmon during 

a bloom of the marine alga Heterosigma akashiwo. Journal of 

Fish Biology 56 (6), 1387–1397. 

Hasan Al-Sheikh, A.H., 1997. Seawater reverse osmosis 

pretreatment with an emphasis on the Jeddah Plant operation 

experience. Desalination 110 (1–2), 183–192. 

Haystead, T.A.J., Sim, A.T.R., Carling, D., Honnor, R.C., Tsukitani, Y., 

Cohen, P., Hardie, D.G., 1989. Effects of the tumour promoter 

okadaic acid on intracellular protein phosphorylation and 

metabolism. Nature 337, 78–81. 

Heisler, J., Glibert, P.M., Burkholder, J.M., Anderson, D.M., 

Cochlan, W., Dennison, W.C., Dortch, Q., Gobler, C.J., Heil, C.A., 

Humphries, E., Lewitus, A., Magnien, R., Marshall, H.G., 

Sellner, K., Stockwell, D.A., Stoecker, D.K., Suddleson, M., 2008. 

Eutrophication and harmful algal blooms: a scientific 

consensus. Harmful Algae 8 (1), 3–13. 

Herndon, J., Cochlan, W.P., Horner, R., 2003. Heterosigma akashiwo 

blooms in San Francisco Bay. Interagency Ecological Program 

for the San Francisco Estuary Newsletter 16, 46–48. 

Hershberger, P.K., Rensel, J.E., Postel, J.R., Taub, F.B., 1997. 

Heterosigma bloom and associated fish kill. Harmful Algae 

News 16 (1), 4. 

Holmes, R.W., Williams, P.M., Eppley, R.W., 1967. Red water in 

La Jolla, 1964–1966. Limnology and Oceanography 12 (3), 

503–512. 

Hong, Y., Smith Jr., W.O., White, A.M., 1997. Studies on 

transparent exopolymer particles (TEP) produced in the Ross 

Sea (Antarctica) and by Phaeocystis antarctica 

(Prymnesiophyceae). Journal Of Phycology 33 (3), 368–376. 


Hori, M., Matsuura, Y., Yoshimoto, R., Ozaki, H., Yasumoto, T., 

Karaki, H., 1999. Actin depolymerizing action by marine toxin, 

pectenotoxin-2. Folia Pharmacologica Japonica 114 (Suppl. 1), 

225–229. 

Horner, R.A., Postel, J.R., 1993. Toxic diatoms in western 

Washington waters (U.S. west coast). Hydrobiologia 269/270 

(1), 197–205. 

Horner, R.A., Garrison, D.L., Plumley, F.G., 1997. Harmful algal 

blooms and red tide problems on the U.S. west coast. 

Limnology and Oceanography 42 (2), 1076–1088. 

Horner, R.A., Hickey, B.M., Postel, J.R., 2000. Pseudo-nitzschia 

blooms and physical oceanography off Washington state. 

South African Journal of Marine Science 22 (1), 299–308. 

Howard, M.D.A., Cochlan, W.P., Ladizinsky, N., Kudela, R.M., 2007. 

Nitrogenous preference of toxigenic Pseudo-nitzschia australis 

(Bacillariophyceae) from field and laboratory experiments. 


Howard, M.D.A., Silvera, M., Kudela, R.M., 2008. Yessotoxin 

detected in mussel (Mytilus californicus) and phytoplankton 

samples from the U.S. west coast. Harmful Algae 7 (5), 

646–652. 

Howard, M.D., Smith, G.J., Kudela, R.M., 2009. Phylogenetic 

relationships of yessotoxin-producing dinoflagellates, based 

on the large subunit and internal transcribed spacer ribosomal 

DNA domains. Applied and Environmental Microbiology 75 (1), 

54–63. 

Howarth, R.W., 2008. Coastal nitrogen pollution: a review of 

sources and trends globally and regionally. Harmful Algae 8 

(1), 14–20. 

IOC HAB Programme, 2008. IOC Harmful Algal Bloom Website. 

Intergovernmental Oceanic Commission of UNESCO. 

Iwataki, M., Kawami, H., Matsuoka, K., 2007. Cochlodinium 

fulvescens sp nov (Gymnodiniales, Dinophyceae), a new chainforming 

unarmored dinoflagellate from Asian coasts. 

Phycological Research 55 (3), 231–239. 

Iwataki, M., Kawami, H., Mizushima, K., Mikulski, C.M., 

Doucette, G.J., Relox, J.R., Anton, A., Fukuyo, Y., Matsuoka, K., 

2008. Phylogenetic relationships in the harmful dinoflagellate 

Cochlodinium polykrikoides (Gymnodiniales, Dinophyceae) 

inferred from LSU rDNA sequences. Harmful Algae 7 (3), 

271–277. 

Jeffery, B., Barlow, T., Moizer, K., Paul, S., Boyle, C., 2004. Amnesic 

shellfish poison. Food and Chemical Toxicology 42 (4), 

545–557. 

Jessup, D.A., Miller, M.A., Ryan, J.P., Nevins, H.M., Kerkering, H.A., 

Mekebri, A., Crane, D.B., Johnson, T.A., Kudela, R.M., 2009. 

Mass stranding of marine birds caused by a surfactantproducing 

red tide. PLOS One 4, E4550. 

Jester, R., 2008. An investigation into the prevalence of 

Alexandrium derived toxins in marine food webs. Ph.D., 

University of California, Santa Cruz, Santa Cruz, pp. 116. 

Jester, R., Lefebvre, K.A., Langlois, G., Vigilant, V., Baugh, K.A., 

Silver, M.W., 2009a. A shift in the dominant toxin-producing 

algal species in central California alters phycotoxins in food 

webs. Harmful Algae 8 (2), 291–298. 

Jester, R.J., Baugh, K.A., Lefebvre, K.A., 2009b. Presence of 

Alexandrium catenella and paralytic shellfish toxins in finfish, 

shellfish and rock crabs in Monterey Bay, California, USA. 

Marine Biology 156 (3), 493–504. 

Johansson, N., Granéli, E., Yasumoto, T., Carlsson, P., Legrand, C., 

1996. Toxin production by Dinophysis acuminata and D. acuta 

cells grown under nutrient sufficient and deficient conditions. 

In: Yasumoto, T., Oshima, Y., Fukuyo, Y. (Eds.), Harmful and 

Toxic Algal Blooms. Intergovernmental Oceanographic 

Commission of UNESCO, Paris, pp. 277–279. 

John, E.H., Flynn, K.J., 2000. Growth dynamics and toxicity of 

Alexandrium fundyense (Dinophyceae): the effect of changing N:

412 

P supply ratios on internal toxin and nutrient levels. European 


Jones, B.H., Noble, M.A., Dickey, T.D., 2002. Hydrographic and 

particle distributions over the Palos Verdes continental shelf: 

spatial, seasonal and daily variability. Continental Shelf 

Research 22 (6–7), 945–965. 

Kirkpatrick, B., Fleming, L.E., Squicciarini, D., Backer, L.C., 

Clark, R., Abraham, W., Benson, J., Cheng, Y.S., Johnson, D., 

Pierce, R., Zaias, J., Bossart, G.D., Baden, D.G., 2004. Literature 

review of Florida red tide: implications for human health 

effects. Harmful Algae 3 (2), 99–115. 

Kudela, R.M., Cochlan, W.P., 2000. Nitrogen and carbon uptake 

kinetics and the influence of irradiance for a red tide bloom off 

southern California. Aquatic Microbial Ecology 21 (1), 31–47. 

Kudela, R.M., Roberts, A., Armstrong, M., 2004. Laboratory 

analyses of nutrient stress and toxin production in Pseudonitzschia 

spp. from Monterey Bay, California. In: Steidinger, K., 

Landsberg, J., Tomas, C., Vargo, G. (Eds.). Florida 

Environmental Research Institute and UNESCO, St. Petersberg, 

pp. 136–138. 

Kudela, R.M., Lane, J.Q., Cochlan, W.P., 2008a. The potential role 

of anthropogenically derived nitrogen in the growth 

of harmful algae in California, USA. Harmful Algae 8 (1), 

103–110. 

Kudela, R.M., Ryan, J.P., Blakely, M.D., Lane, J.Q., Peterson, T.D., 

2008b. Linking the physiology and ecology of Cochlodinium to 

better understand harmful algal bloom events: a comparative 

approach. Harmful Algae 7 (3), 278–292. 

Kusek, K.M., Vargo, G., Steidinger, K., 1999. Gymnodinium breve in 

the field in the lab and in the newspaper - scientific and 

journalistic analysis of Florida red tides. Contributions in 

Marine Science 34, 1–229. 

Ladizinsky, N., Smith, G.J., 2000. Accumulation of domoic acid by 

the coastal diatom Pseudo-nitzschia multiseries: a possible 

copper complexation strategy. Journal of Phycology 36, 41. 

Landry, M.R., 1989. Broad-scale distributional patterns of 

hydrographic variables on the Washington/Oregon shelf. In: 

Landry, M.R., Hickey, B.M. (Eds.), Coastal Oceanography of 

Washington and Oregon. Elsevier, New York, pp. 1–40. 

Lange, C.B., Reid, F.M.H., Vernet, M., 1994. Temporal distribution 

of the potentially toxic diatom Pseudo-nitzschia australis at 

a coastal site in Southern California. Marine Ecology Progress 

Series 104 (3), 309–312. 

Langlois, G., 2007. Marine Biotoxin Monitoring Program Annual 

Report. California Department of Public Health for California 

Department of Fish and Game, Sacramento. 

Lee, J., Walker, H.W., 2006. Effect of process variables and natural 

organic matter on removal of microcystin-LR by PAC-UF. 

Environmental Science and Technology 40 (23), 7336–7342. 

Lee, J.S., Yanagi, T., Kenma, R., Yasumoto, T., 1987. Fluorometric 

determination of diarrhetic shellfish toxins by highperformance 

liquid chromatography. Agricultural and 

Biological Chemistry 51 (3), 877–881. 

Lee, J.S., Igarashi, T., Fraga, S., Dah, E., Hovgaard, P., Yasumoto, T., 

1989. Determination of diarrhetic toxins in various 

dinoflagellate species. Journal of Applied Phycology 1 (2), 

147–152. 

Lefebvre, K.A., Bill, B.D., Erickson, A., Baugh, K.A., O’Rourke, L., 

Costa, P.R., Nance, S., Trainer, V.L., 2008. Characterization of 

intracellular and extracellular saxitoxin levels in both field 

and cultured Alexandruim spp. samples from Sequim Bay, 

Washington. Marine Drugs 6, 103–116. 

Levandowsky, M., Kaneta, P.J., 1987. Behaviour in dinoflagellates. 

In: Taylor, F.J.R. (Ed.), The Biology of Dinoflagellates. Blackwell, 

Oxford, pp. 360–397. 

Lin, Y.Y., Risk, M., Ray, S.M., Van Engen, D., Clardy, J., Golik, J., 

James, J.C., Nakanishi, K., 1981. Isolation and structure of 

brevetoxin B from the ‘red tide’ dinoflagellate Ptychodiscus 


brevis (Gymnodinium breve). Journal of American Chemical 

Society 103 (22), 6773–6775. 

Llewellyn, L.E., 2006. Saxitoxin, a toxic marine natural product 

that targets a multitude of receptors. Natural Product Reports 

23 (2), 200–222. 

Loeblich III, A.R., Fine, K.E., 1977. Marine chloromonads: More 

widely distributed in neritic environments than previously 

thought. Proceedings of the Biological Society of Washington 

90 (2), 388–399. 

Lundholm, N., Skov, J., Pocklington, R., Moestrup, Ã., 1997. Studies 

on the marine planktonic diatom Pseudo-nitzschia. 2. 

Autecology of P. pseudodelicatissima based on isolates from 

Danish coastal waters. Phycologia 36, 381–388. 

Malagoli, D., Marchesini, E., Ottaviani, E., 2006. Lysosomes as the 

target of yessotoxin invertebrate and vertebrate cell lines. 

Toxicology Letters 167 (1), 75–83. 

Maldonado, M.T., Hughes, M.P., Rue, E.L., Wells, M.L., 2002. The 

effect of Fe and Cu on growth and domoic acid production by 

Pseudo-nitzschia multiseries and Pseudo-nitzschia australis. 


Marchetti, A., Trainer, V.L., Harrison, P.J., 2004. Environmental 

conditions and phytoplankton dynamics associated 

with Pseudo-nitzschia abundance and domoic acid in the 

Juan de Fuca eddy. Marine Ecology Progress Series 281, 

1–12. 

Martin, J.L., Haya, K., Burridge, L.E., Wildish, D.J., 1990. Nitzschia 

pseudodelicatissima - a source of domoic acid in the Bay of 

Fundy, eastern Canada. Marine Ecology Progress Series 67 (2), 

177–182. 

Matsuoka, K., Iwataki, M., Kawami, H., 2008. Morphology and 

taxonomy of chain-forming species of the genus Cochlodinium 

(Dinophyceae). Harmful Algae 7 (3), 261–270. 

McGillicuddy Jr., D.J., Signell, R.P., Stock, C.A., Keafer, B.A., 

Keller, M.D., Ehtland, R.D., Anderson, D.M., 2005. A 

mechanisms for offshore initiations of harmful algal blooms 

in the coastal Gulf of Maine. Journal of Plankton Research 25, 

1131–1138. 

Mengelt, C., 2006. Ultraviolet Photoecology, Dark Survival, and 

Seasonal Abundance of Pseudo-nitzschia australis and P. 

multiseries in Coastal Waters of Central California. 

Dissertation, University of California, Santa Barbara. 

Mikulski, C.M., Morton, S.L., Doucette, G.J., 2005. Development 

and application of LSU rRNA probes for Karenia brevis in the 

Gulf of Mexico, USA. Harmful Algae 4 (1), 49–60. 

Mikulski, C.M., Park, Y.T., Jones, K.L., Lee, C.K., Lim, W.A., Lee, Y., 

Scholin, C.A., Doucette, G.J., 2008. Development and field 

application of rRNA-targeted probes for the detection of 

Cochlodinium polykrikoides Margalef in Korean coastal waters 

using whole cell and sandwich hybridization formats. 


Miles, C.O., Aasen, J., Dahl, E., Samdal, I., Aune, T., Briggs, L.R., 

Olseng, C.D., Edvardsen, B., Naustvoll, L.J., 2002. Yessotoxin in 

Norwegian Blue Mussels caused by Protoceratium reticulatum. 

In: Tenth International Conference on Harmful Algal Blooms, 

St. Petersburg, FL (abstract). 

Miles, C.O., Wilkins, A.L., Samdal, I.A., Sandvik, M., Petersen, D., 

Quilliam, M.A., Nausvoll, L.J., Rundberget, T., Torgersen, T., 

Hovgaard, P., Jensen, D.J., Cooney, J.M., 2004. A novel 

pectenotoxin, PTX-12, in Dinophysis spp. and shellfish from 

Norway. Chemical Research and Toxicology 17 (11), 

1423–1433. 

Miles, C.O., Wilkins, A.L., Hawkes, A.D., Selwood, A.I., Jensen, D.J., 

Munday, R., Cooney, J.M., Beuzenberg, V., 2005a. 

Polyhydroxylated amide analogs of yessotoxin from 

Protoceratium reticulatum. Toxicon 45 (1), 61–71. 

Miles, C.O., Samdal, I.A., Aasen, J.A.G., Jensen, D.J., Quilliam, M.A., 

Petersen, D., Briggs, L.M., Wilkins, A.L., Rise, F., Cooney, J.M., 

MacKenzie, A.L., 2005b. Evidence for numerous analogs of

yessotoxin in Protoceratium reticulatum. Harmful Algae 4 (6), 

1075–1091. 

Miles, C.O., Wilkins, A.L., Hawkes, A.D., Selwood, A.I., Jensen, D.J., 

Cooney, J.M., Beuzenberg, V., MacKenzie, L., 2006. 

Identification of 45-hydroxy-46,47-dinoryessotoxin, 44-oxo- 

45,46,47-trinoryessotoxin, and 9-methyl-42,43,44,45,46,47,55hepta-nor-38-en-41-oxoyessotoxin, 

and partial 

characterization of some minor yessotoxins, from 


Miller, P.E., Scholin, C.A., 1998. Identification and enumeration of 

cultured and wild Pseudo-nitzschia (Bacillariophyceae) using 

species-specific LSU rRNA-targeted fluorescent probes and 

filter-based whole cell hybridization. Journal of Phycology 34 

(2), 371–382. 

Mitrovic, S.M., Fernández Amandi, M., MacKenzie, L., Furey, A., 

James, K.J., 2004. Effects of selenium, iron and cobalt 

addition to growth and yessotoxin production of the toxic 

marine dinoflagellate Protoceratium reticulatum in culture. 

Journal of Experimental Marine Biology and Ecology 313 (2), 

337–351. 

Mitrovic, S.M., Hamilton, B., MacKenzie, L., Furey, A., James, K.J., 

2005. Persistence of yessotoxin under light and dark 

conditions. Marine Environmental Research 60 (3), 397–401. 

Montojo, U.M., Sakamoto, S., Cayme, M.F., Gatdula, N.C., Furio, E. 

F., Relox, J.R.J., Sato, S., Fukuyo, Y., Kodam, M., 2006. 

Remarkable difference in accumulation of paralytic shellfish 

poisoning toxins among bivalve species exposed to Pyrodinium 

bahamense var. compressum bloom in Masinloc bay. Philippines 

Toxicon 48 (1), 85–92. 

Montoya, N.G., Akselman, R., Franco, J.M., Carreto, J.I., 1996. 

Paralytic shellfish toxins and mackerel (Scomber japonicus) 

mortality in the Argentine sea. In: Yasumoto, T., Oshima, Y., 

Fukuyo, Y. (Eds.), Harmful and Toxic Algal Blooms. 

Intergovernmental Oceanographic Community, UNESCO, 

Paris, pp. 417–420. 

Moore, S.K., Mantua, N.J., Kellogg, J.P., Newton, J.A., 2008. Local 

and large-scale climate forcing of Puget Sound oceanographic 

properties on seasonal to interdecadal timescales. Limnology 

and Oceanography 53 (5), 1746–1758. 

Moore, S.K., Mantua, N.J., Hickey, B.M., Trainer, V.L., 2009. Recent 

trends in paralytic shellfish toxins in Puget Sound, 

relationships to climate, and capacity for prediction of toxic 

events. Harmful Algae 8 (3), 463–477. 

Moorthi, S.D., Countway, P.D., Stauffer, B.A., Caron, D.A., 2006. 

Use of quantitative real-time PCR to investigate the dynamics 

of the red tide dinoflagellate Lingulodinium polyedrum. 

Microbial Ecology 52 (1), 136–150. 

Murata, M., Shimatani, M., Sugitani, Y., Oshima, Y., 

Yasumoto, T., 1982. Isolation and structural elucidation of 

the causative toxin of the diarrhetic shellfish poisoning. 

Bulletin of the Japanese Society of Scientific Fisheries 48 (4), 

549–552. 

Murata, M., Kumagai, M., Lee, J.S., Yasumoto, T., 1987. Isolation 

and structure of Yessotoxin, a novel polyether compound 

implicated in diarrhetic shellfish poisoning. Tetrahedron 

Letters 28 (47), 5869–5872. 

Naar, J., Bourdelais, A., Tomas, C., Kubanek, J., Whitney, P.L., 

Flewelling, L., Steidinger, K., Lancaster, J., Baden, D.G., 2002. 

A competitive ELISA to detect brevetoxins from Karenia brevis 

(formerly Gymnodinium breve) in seawater, shellfish, and 

mammalian body fluid. Environmental Health Perspectives 

110 (2), 179–185. 

Netjstgaard, J.C., Tang, K.W., Steinke, M., Dutz, J., Koski, M., 

Antajan, E., Long, J.D., 2007. Zooplankton grazing on 

Phaeocystis: a quantitative review and future challenges. 

Biogeochemistry 83 (1–3), 147–172. 

Nishitani, G., Yamaguchi, M., Ishikawa, A., Yanagiya, S., 

Mitsuya, T., Imai, M., 2005. Relationships between occurrences 


of toxic Dinophysis species (Dinophyceae) and small 

phytoplanktons in Japanese coastal waters. Harmful Algae 4 

(4), 755–762. 

O’Halloran, C., Silver, M., Holman, T., Scholin, C., 2006. 

Heterosigma akashiwo in central California waters. Harmful 

Algae 5 (2), 124–132. 

Office of Shellfish and Water Protection, 2008. 2007 Annual 

Report: Commercial and Recreational Shellfish Areas in 

Washington State Olympia. Office of Shellfish and Water 

Protection. 

Ogino, H., Kumagai, M., Yasumoto, T., 1997. Toxicologic 

evaluation of Yessotoxin. Natural Toxins 5 (6), 255–259. 

Ono, K., Khan, S., Onoue, Y., 2000. Effects of temperature and light 

intensity on the growth and toxicity of Heterosigma akashiwo 

(Raphidophyceae). Aquaculture Research 31 (5), 427–433. 

Pan, Y., Subba Rao, D.V., Mann, K.H., 1996a. Changes in domoic 

acid production and cellular chemical composition of the 

toxigenic diatom Pseudo-nitzschia muliseries under 

phosphate limitation. Journal of Phycology 32 (3), 371–381. 

Pan, Y., Subba Rao, D.V., Mann, K.H., Li, K.W., Harrison, W.G., 

1996b. Effects of silicate limitation on production of domoic 

acid, a neurotoxin, by the diatom Pseudo-nitzschia multiseries. I. 

Batch culture studies. Marine Ecology Progress Series 131 (1–3), 

225–233. 

Pan, Y., Bates, S.S., Cembella, A.D., 1998. Environmental stress 

and domoic acid production by Pseudo-nitzschia: 

a physiological perspective. Natural Toxins 6 (3–4), 127–135. 

Pankratz, T., 2008. Red tides close desal plants. Water 

Desalination Report 44, 1. 

Pankratz, T., 2009. Speaking of red tides. Water Desalination 

Report 45, 2. 

Park, M.G., Kim, S., Kim, H.S., Myung, G., Kang, Y.G., Yih, W., 2006. 

First successful culture of the marine dinoflagellate Dinophysis. 

Aquatic Microbial Ecology 45 (2), 101–106. 

Paz, B., Riobo, P., Fernandez, A.L., Fraga, S., Franco, J.M., 2004. 

Production and release of yessotoxins by the dinoflagellates 

Protoceratium reticulatum and Lingulodinium polyedrum in 

culture. Toxicon 44 (3), 251–258. 

Paz, B., Riobo, P., Souto, M.L., Gil, L.V., Norte, M., Fernandez, J.J., 

Franco, J.M., 2006. Detection and identification of 

glycoyessotoxin A in a culture of the dinoflagellate 


Paz, B., Daranas, A.H., Cruz, P.G., Franco, J.M., Pizarro, G., 

Souto, M.L., Norte, M., Fernández, J.J., 2007. Characterization of 

okadaic acid related toxins by liquid chromatography coupled 

with mass spectrometry. Toxicon 50 (2), 225–235. 

Paz, B., Daranas, A.H., Norte, M., Riobo, P., Franco, J.M., 

Fernández, J.J., 2008. Yessotoxins, a group of marine polyether 

toxins: an overview. Marine Drugs 6 (6), 73–102. 

Peleka, E.N., Matis, K.A., 2008. Application of flotation as 

a pretreatment process during desalination. Desalination 222 

(1–3), 1–8. 

Perl, T.M., Bedard, L., Kosatsky, T., Hockin, J.C., Todd, E.C.D., 

McNutt, L.A., Remis, R.S., 1990. Amnesic shellfish poisoning: 

a new clinical syndrome due to domoic acid. Canadian 

Diseases Weekly Report 16 (Suppl 1E), 7–8. 

Petry, M., Sanz, M.A., Langlais, C., Bonnelye, V., Durand, J.P., 

Guevara, D., Nardes, W.M., Saemi, C.H., 2007. The El Coloso 

(Chile) reverse osmosis plant. Desalination 203 (1–3), 

141–152. 

Pocklington, R., Milley, J.E., Bates, S.S., Bird, C.J., deFreitas, A.S.W., 

Quilliam, M.A., 1990. Trace determination of domoic acid and 

phytoplankton by high-performance liquid chromatography 

of the fluorenylmethoxycarbonyl (FMOC) derivative. 

International Journal of Environmental Analytical Chemistry 

38, 351–358. 

Prassopoulou, E., Katikou, P., Georgantelis, D., Kyritsakis, A., 

2009. Detection of okadaic acid and related esters in mussels

414 

during diarrhetic shellfish poisoning (DSP) episodes in 

Greece using the mouse bioassay, the PP2A inhibition assay 

and HPLC with fluorimetric detection. Toxicon 53 (2), 

214–227. 

Price, D.W., Kizer, K.W., Hansgen, K.H., 1991. California’s paralytic 

shellfish poisoning prevention program, 1927–89. Journal of 

Shellfish Research 10 (11), 119–145. 

Purkerson, S.L., Baden, D.G., Fieber, L.A., 1999. Brevetoxin 

modulates neuronal sodium channels in two cell lines derived 

from rat brain. Neurotoxicology 20 (6), 909–920. 

Quilliam, M., 2003. Chemical methods for domoic acid, the 

amnesic shellfish poisoning (ASP) toxin. In: Hallegraeff, G., 

Anderson, D., Cembella, A. (Eds.), Manual on Harmful Marine 

Microalgae, Monographs on Oceanographic Methodology. 

Intergovernmental Oceanographic Commission (UNESCO), 

Paris, pp. 247–266. 

Quilliam, M.A., Sim, P.G., McCulloch, A.W., McInnes, A.G., 1989. 

High-performance liquid chromatography of domoic acid, 

a marine neurotoxin, with application to shellfish and 

plankton. International Journal of Environmental Analytical 

Chemistry 36 (3), 139–154. 

Ramsdell, J.S., Anderson, D.M., Glibert, P.M., 2005. HARRNESS, 

Harmful Algal Research and Response: a National 

Environmental Science Strategy 2005–2015. Ecological Society 

of America, Washington, DC, 96 pp. 

Reid, J.L., Roden, G.I., Wyllie, J.G., 1958. Studies of the California 

Current system. California Cooperative Oceanic Fisheries 

Investigations Progress Report, 1 July 1956–1 January 1958, 

pp. 27–56. 

Reid, F.M.H., Fuglister, E., Jordan, J.B., 1970. The ecology of the 

plankton off La Jolla, California, in the period April through 

September, 1967. Part 5: phytoplankton taxonomy and 

standing crop. Bulletin of the Scripps Institute of 

Oceanography Techical Series 17, 51–66. 

Reid, F.M.H., Lange, C.B., White, M.M., 1985. Microplankton 

species assemblages at Scripps Pier from March to November 

1983 during the 1982–1983 El Niño Event. Botanica Marina 28 

(10), 443–452. 

Reiss, C.R., Robert, C., Owen, C., Taylor, J.S., 2006. Control of MIB, 

geosmin and TON by membrane systems. Journal of Water 

Supply: Research and Technology – AQUA 55 (2), 95–108. 

Rhodes, L., White, D., Syhre, M., Atkinson, M., 1996. Pseudonitzschia 

species isolated from New Zealand coastal waters: 

domoic acid production in vitro and links with shellfish 

toxicity. In: Yasumoto, T., Oshima, Y., Fukuyo, Y. (Eds.), 

Harmful and Toxic Algal Blooms. IOC of UNESCO, Paris 

pp. 155–158. 

Rhodes, L., Scholin, C., Garthwaite, I., Haywood, A., Thomas, A., 

1998. Domoic acid producing Pseudo-nitzschia species deduced 

with whole cell DNA probe-based and immunochemical 

assays. In: Reguera, B., Blanco, J., Fernández, M.L., Wyatt, T. 

(Eds.), Harmful Algae. Xunta de Galicia and 

Intergovernmental Oceanographie Commission of UNESCO, 

pp. 274–277. 

Rhodes, L., McNabb, P., de Salas, M., Briggs, L., Beuzenberg, V., 

Gladstone, M., 2006. Yessotoxin production by Gonyaulax 

spinifera. Harmful Algae 5 (2), 148–155. 

Rines, J.E.B., Donaghay, P.L., Dekshenieks, M.M., Sullivan, J.M., 

Twardowski, M.S., 2002. Thin layers and camouflage: hidden 

Pseudo-nitzschia spp. (Bacillariophyceae) populations in a fjord 

in the San Juan Island, Washington, USA. Marine Ecology 

Progress Series 225, 123–137. 

Riobo, P., Paz, B., Fernandez, M., Fraga, S., Franco, J.M., 2002. 

Lipophylic toxins of different strains of Ostreopsidaceae and 

Gonyaulacaceae. In: Tenth International Conference on 

Harmful Algal Blooms, St. Petersburg, FL (abstract). 

Rogers, R.S., Rapoport, H., 1980. The pKas of saxitoxin. Journal of 

American Chemical Society 102 (24), 7335–7339. 


Ronzitti, G., Callegari, F., Malaguti, C., Rossini, G.P., 2004. Selective 

disruption of the E-cadherin–catenin system by an algal toxin. 

British Journal of Cancer 90 (5), 1100–1107. 

Rue, E.L., Bruland, K.W., 2001. Domoic acid binds iron and copper: 

a possible role for the toxin produced by the marine diatom 

Pseudo-nitzschia. Marine Chemistry 76 (1–2), 127–134. 

Samdal, I.A., Naustvoll, L.J., Olseng, C.D., Briggs, L.R., Miles, C.O., 

2004. Use of ELISA to identify Protoceratium reticulatum as 

a source of yessotoxin in Norway. Toxicon 44 (1), 

75–82. 

Samdal, I.A., Aesen, J.A.G., Briggs, L.R., Dahl, E., Miles, C.O., 2005. 

Comparison of ELISA and LC–MS analyses for yessotoxins in 

blue mussels (Mytilus edulis). Toxicon 46, 7–15. 

Satake, M., Mackenzie, L., Yasumoto, T., 1997. Identification of 

Protoceratium reticulatum as the biogenetic origin of yessotoxin. 

Natural Toxins 5 (4), 164–167. 

Satake, M., Ichimura, T., Sekiguchi, K., Yoshimatsu, S., Oshima, Y., 

1999. Confirmation of yessotoxin and 45, 46, 47trinoryessotoxin 

production by Protoceratium reticulatum 

collected in Japan. Natural Toxins 7 (4), 147–150. 

Schantz, E.J., Mold, J.D., Stanger, D.W., Shavel, J., Riel, F.J., 

Bowden, J.P., Lynch, J.M., Wyler, R.S., Riegel, B., Sommer, H., 

1957. Paralytic shellfish poison. VI. A Procedure for the 

isolation and purification of the poison from toxic clam and 

mussel tissues. Journal of American Chemical Society 79 (19), 

5230–5235. 

Schnetzer, A., Miller, P.E., Schaffner, R.A., Stauffer, B.A., Jones, B.H., 

Weisberg, S.B., DiGiacomo, P.M., Berelson, W.M., Caron, D.A., 

2007. Blooms of Pseudo-nitzschia and domoic acid in the San 

Pedro Channel and Los Angeles harbor areas of the Southern 

California Bight, 2003-2004. Harmful Algae 6 (3), 372–387. 

Schofield, O., Kohut, J., Aragon, D., Creed, L., Graver, J., 

Haldeman, C., Kerfoot, J., Roarty, H., Jones, C., Webb, D., 

Glenn, S., 2007. Slocum gliders: robust and ready. Journal of 

Field Robotics 24 (6), 473–485. 

Scholin, C.A., Herzog, M., Sogin, M., Anderson, D.M., 1994. 

Identification of group-and strain-specific genetic markers for 

globally distributed Alexandrium (Dinophyceae). Part II. 

Sequence analysis of a fragment of the LSU rRNA gene. Journal 

of Phycology 30 (6), 999–1011. 

Scholin, C.A., Gulland, F., Doucette, G.J., Benson, S., Busman, M., 

Chavez, F.P., Cordaro, J., DeLong, R., De Vogelaere, A., 

Harvey, J., Haulena, M., Lefebvre, K.A., Lipscomb, T., 

Loscutoff, S., Lowenstine, L.J., Marin 3rd, R., Miller, P.E., 

McLellan, W.A., Moeller, P.D., Powell, C.L., Rowles, T., 

Silvagni, P., Silver, M., Spraker, T., Trainer, V.L., Van 

Dolah, F.M., 2000. Mortality of sea lions along the central 

California coast linked to a toxic diatom bloom. Nature 403 

(6765), 80–84. 

Separation Processes Inc., 2005. Investigation of Microfiltration 

and Reverse Osmosis Seawater Desalination. National Water 

Research Institute. 

Shimizu, Y., Hsu, C.P., Genenah, A.A., 1981. Structure of saxitoxin 

in solutions and stereochemistry of dihydrosaxitoxins. Journal 

of American Chemical Society 103, 605–609. 

Shipe, R.F., Leinweber, A., Gruber, N., 2008. Abiotic controls of 

potentially harmful algal blooms in Santa Monica Bay, 

California. Continental Shelf Research 28 (18), 2584–2593. 

Shumway, S.E., Allen, S.M., Boersma, P.D., 2003. Marine birds and 

harmful algal blooms: sporadic victims or under-reported 

events? Harmful Algae 2 (1), 1–17. 

Smayda, T.J., 1990. Novel and nuisance phytoplankton blooms in 

the sea: evidence for a global epidemic. In: Granéli, E., 

Gundström, B., Edler, L., Anderson, D.M. (Eds.), Toxic Marine 

Phytoplankton. Elsevier, New York, pp. 29–40. 

Smayda, T.J., 2008. Complexity in the eutrophication–harmful 

algal bloom relationship, with comment on the importance of 

grazing. Harmful Algae 8 (1), 140–151.

Smith, G., Ladizinsky, N., Miller, P., 2001. Amino acid profiles in 

species and strains of Pseudo-nitzschia from Monterey Bay, 

California: insights into the metabolic role(s) of domoic acid. 

In: Hallegraef, G., Blackburn, S., Bolch, C., Lewis, R. (Eds.). IOC, 

UNESCO, Hobart, Tasmania, pp. 324–327. 

Sommer, H., Meyer, K.F., 1937. Paralytic shellfish poisoning. 

Archives of Pathology 24, 560–598. 

Steidinger, K.A., 1993. Some taxonomic and biologic aspects of 

toxic dinoflagellates. In: Falconer, I. (Ed.), Algal Toxins in 

Seafood and Drinking Water. Academic Press, London, pp. 1–28. 

Stobo, L.A., Lewis, J., Quilliam, M.A., Hardstaff, W.R., Gallacher, S., 

Webster, L., Smith, E.A., McKenzie, M., 2003. Detection of 

yessotoxin in UK and Canadian isolates of phytoplankton and 

optimization and validation of LC–MS methods. Canadian 

Technical Report of Fisheries and Aquatic Sciences 2498, 8–14. 

Strom, S., Wolfe, G.V., Slajer, A., Lambert, S., Clough, J., 2003. 

Chemical defenses in the microplankton II: inhibition of 

protist feeding by B-dimethylsulfoniopropionate (DMSP). 


Su, Z., Sheets, M., Ishida, H., Li, F., Barry, W.H., 2004. Saxitoxin 

blocks L-type ICa. Journal of Pharmacology and Experimental 

Therapeutics 308 (1), 324–329. 

Subba Rao, D.V., Quilliam, M.A., Pocklington, R., 1988. Domoic 

acid – a neurotoxic amino acid produced by the marine diatom 

Nitzschia pungens in culture. Canadian Journal of Fisheries and 

Aquatic Sciences 45 (12), 2076–2079. 

Suganuma, M., Fujiki, H., Suguri, H., Yoshizawa, S., Hirota, M., 

Nakayasu, M., Ojika, M., Wakamatsu, K., Yamada, K., 

Sugimura, T., 1988. Okadaic acid: an additional non-phorbol- 

12-tetradecanoate-13-acetate-type tumor promoter. 

Proceedings of the National Academy of Sciences, United 

States of America 85 (6), 1768–1771. 

Sukhatme, G.A.D., Zhang, B., Oberg, C., Stauffer, B.A., Caron, D.A., 

2007. The design and development of a wireless robotic 

networked aquatic microbial observing system. 

Environmental Engineering Science 24, 205–215. 

Sunda, W., 2006. Trace metals and harmful algal blooms. In: 

Granéli, E., Turner, J.T. (Eds.), Ecology of Harmful Algae. 

Springer, Berlin, pp. 203–214. 

Sutherland, C.M., 2008. Diarrhetic Shellfish Toxins Linked to Local 

Dinophysis Populations in the California Coastal Waters of 

Monterey Bay. Dissertation, University of California, Santa 

Cruz, Santa Cruz. 

Tachibana, K., Scheuer, P.J., Tsukitani, Y., Kikuchi, H., Engen, D.V., 

Clardy, J., Gopichand, Y., Schmitz, F.J., 1981. Okadaic acid, 

a cytotoxic polyether from two marine sponges of the genus 

Halichondria. Journal of American Chemical Society 103 (9), 

2469–2471. 

Takahashi, H., Kusumi, T., Kan, Y., Satake, M., Yasumoto, T., 1996. 

Determination of the absolute configuration of yessotoxin, 

a polyether compound implicated in diarrhetic shellfish 

poisoning, by NMR spectroscopic method using a chiral 

anisotropic reagent, methoxy-(2-naphthyl) acetic acid. 

Tetrahedron Letters 37 (39), 7087–7090. 

Takai, A., Bialojan, C., Troschka, M., Rüegg, J.C., 1987. Smooth 

muscle myosin phosphatase inhibition and force 

enhancement by black sponge toxin. FEBS Letters 217 (1), 

81–84. 

Taylor, F.J.R., Horner, R.A., 1994. Red tides and other problems 

with harmful algal blooms in Pacific Northwest coastal waters. 

In: Review of the Marine Environment and Biota of Strait of 

Georgia, Puget Sound and Juan de Fuca Strait. Canadian 

Fisheries and Aquatic Science Technical Report 1948, 

pp. 175–186. 

Tenzer, B., Adin, A., Priel, M., 1999. Seawater filtration for fouling 

prevention under stormy conditions. Desalination 125 (1–3), 

77–88. 


Terao, K., Ito, E., Oarada, M., Murata, M., Yasumoto, T., 1990. 

Histopathological studies on experimental marine toxin 

poisoning: the effects in mice of yessotoxin isolated from 

Patinopecten yessoensis and of a desulfated derivative. Toxicon 

28 (9), 1095–1104. 

Thomas, W.H., Zavoico, T., Hewes, C., 2001. Historical 

Phytoplankton Species Time-Series Data (1917–1939) from the 

North American Pacific Coast. Scripps Institution of 

Oceanography. Ref. no. 01-12. 

Throndsen, J., 1997. The plankton marine flagellates. In: 

Tomas, C.R. (Ed.), Identifying marine phytoplankton. 

Academic Press, San Diego, pp. 591–729. 

Trainer, V.L., 2002. Harmful algal blooms on the U.S. west coast. 

In: Taylor, F.J., Trainer, V.L. (Eds.), Harmful Algal Blooms in the 

PICES Region of the North Pacific, pp. 89–118. PICES Scientific 

Report 23. 

Trainer, V.L., Adams, N.G., Bill, B.D., Anulacion, F., Wekell, J.C., 

1998. Concentration and dispersal of a Pseudo-nitzschia bloom 

in Penn Cove, Washington, USA. Natural Toxins 6 (3–4), 

113–126. 

Trainer, V.L., Adams, N.G., Bill, B.D., Stehr, C.M., Wekell, J.C., 

Moeller, P., Busman, M., Woodruff, D., 2000. Domoic acid 

production near California coastal upwelling zones, June 1998. 


Trainer, V.L., Adams, N.G., Wekell, J.C., 2001. Domoic acid 

producing Pseudo-nitzschia species off the U.S. west coast 

associated with toxification events. In: Hallegraeff, G.M., 

Blackburn, S.I., Bolch, C.J., Lewis, R.J. (Eds.), Harmful Algal 

Blooms 2000. Intergovernmental Oceanographic Commission 

of UNESCO, Paris, pp. 46–49. 

Trainer, V.L., Hickey, B.M., Horner, R.A., 2002. Biological and 

physical dynamics of domoic acid production off the 

Washington Coast. Limnology and Oceanography 47 (5), 

1438–1446. 

Trainer, V.L., Eberhart, B.T.L., Wekell, J.C., Adams, N.G., 

Hanson, L., Cox, F., Dowell, J., 2003. Paralytic shellfish toxins in 

Puget Sound, Washington State. Journal of Shellfish Research 

22 (1), 213–223. 

Trainer, V.L., Cochlan, W.P., Erickson, A., Bill, B.D., Cox, F.H., 

Borchert, J.A., Lefebvre, K., 2007. Recent domoic acid closures 

of shellfish harvest areas in Washington State inland 

waterways. Harmful Algae 6 (3), 449–459. 

Tubaro, A., Sidari, L., Della Loggia, R., Yasumoto, T., 1998. 

Occurrence of yessotoxin in phytoplankton and mussels from 

Northern Adriatic Sea. In: Proceedings of the VIII International 

Conference on Harmful Algae, Vigo, Spain (abstract). 

Twiner, M.J., Dechraoui, M.Y.B., Wang, Z., Mikulski, C.M., Henry, M. 

S., Pierce, R.H., Doucette, G.J., 2007. Extraction and analysis of 

lipophilic brevetoxins from the red tide dinoflagellate Karenia 

brevis. Analytical Biochemistry 369 (1), 128–135. 

Tyrell, J.V., Connell, L.B., Scholin, C.A., 2002. Monitoring for 

Heterosigma akashiwo using a sandwich hybridization assay. 


Van Dolah, F.M., Finley, E.L., Haynes, B.L., Doucette, G.J., 

Moeller, P.D., Ramsdell, J.S., 1994. Development of rapid and 

sensitive high throughput pharmacologic assays for marine 

phycotoxins. Natural Toxins 2 (4), 189–196. 

Vigilant, V.L., Silver, M., 2007. Domoic acid in benthic flatfish on 

the continental shelf of Monterey Bay, California, USA. Marine 

Biology 151 (6), 2053–2062. 

Walz, P.M., Garrison, D.L., Graham, W.M., Cattey, M.A., 

Tjeerdema, R.S., Silver, M.W., 1994. Domoic acid-producing 

diatom blooms in Monterey Bay, California: 1991–1993. 

Natural Toxins 2 (5), 271–279. 

Wang, J., Salata, J.J., Bennett, P.B., 2003. Saxitoxin is a gating 

modifier of hERG Kþ channels. Journal of General Physiology 

121 (6), 583–598.

416 

Watson, M., 1997. Control of Membrane Fouling Due to Algae. 

Applied Research Department. Technology Transfer Note. 

No.12. 

Wekell, J.C., Gauglitz, E.J.J., Barnett, H.J., Hatfield, C.L., Simons, D., 

Ayres, D., 1994. Occurrence of domoic acid in Washington 

state razor clams (Siliqua patula) during 1991–1993. Natural 

Toxins 2 (4), 197–205. 

Wells, M.L., Trick, C.G., Cochlan, W.P., Hughes, M.P., Trainer, V.L., 

2005. Domoic acid: the synergy of iron, copper, and the 

toxicity of diatoms. Limnology and Oceanography 50 (6), 

1908–1917. 

Wilf, M., Schierach, M.K., 2001. Improved performance and cost 

reduction of RO seawater systems using UF pretreatment. 

Desalination 135 (1–3), 61–68. 

Wong, J.L., Oesterlin, R., Rapoport, H., 1971. Struture of 

saxitoxin. Journal of American Chemical Society 93 (26), 

7344–7345. 

Wright, J.L.C., Cembella, A.D., 1998. Ecophysiology and 

biosynthesis of polyether marine biotoxins. In: Anderson, D.M., 

Cembella, A.D., Hallegraeff, G.M. (Eds.), The Physiological 


Ecology of Harmful Algal Blooms. Springer-Verlag, Heidelberg, 

pp. 427–451. 

Wright, J.L.C., Bird, C., de Freitas, A., Hampson, D., McDonald, J., 

Quilliam, M., 1990. Chemistry, biology, and toxicology of 

domoic acid and its isomers. Canadian Diseases Weekly 

Report 16 (Suppl. 1E), 21–26. 

Yasumoto, T., Murata, M., 1993. Marine toxins. Chemical Reviews 

93, 1867–1909. 

Yasumoto, T., Takizawa, A., 1997. Fluorometric measurement of 

yessotoxins in shellfish by high-pressure liquid 

chromatography. Bioscience, Biotechnology and Biochemistry 

61 (10), 1775–1777. 

Yasumoto, T., Oshima, Y., Sugawara, W., Fukuyo, Y., Oguri, H., 

Igarashi, T., Fujita, N., 1980. Identification of Dinophysis fortii as 

the causative organism of diarrhetic shellfish poisoning. 

Bulletin of the Japanese Society of Scientific Fisheries 46 (11), 

1405–1411. 

Yasumoto, T., Murata, M., Oshima, Y., Sano, M., Matsumoto, G.K., 

Clardy, J., 1985. Diarrhetic shellfish toxins. Tetrahedron 41 (6), 

1019–1025.

Occurrence and removal of pharmaceuticals, caffeine and 

DEET in wastewater treatment plants of Beijing, China 

Qian Sui, Jun Huang, Shubo Deng, Gang Yu*, Qing Fan 

POPs Research Centre, Department of Environmental Science & Engineering, Tsinghua University, Beijing 10084, China 

article info 


Received 14 January 2009 


5 July 2009 

Accepted 8 July 2009 


Keywords: 


Wastewater 

Removal efficiency 

Advanced treatment 

Risk assessment 

China 


abstract 

With the progress of sensitive analytical techniques, the 

frequent detection of various pharmaceuticals in the aquatic 

environment has received global concerns of both the 

academic community and the public (Daughton and Ternes, 

1999; Jones et al., 2005). After intake by humans or animals, 

the pharmaceuticals will be partially converted to metabolites, 

however, partially excreted unchanged or as conjugates, 

and finally delivered to the wastewater treatment plants 

(WWTPs). As there is no unit specifically designed to remove 

these compounds, the elimination by most WWTPs seems to 

be inefficient (Ternes, 1998; Castiglioni et al., 2006; Lishman 

et al., 2006; Nakada et al., 2006; Santos et al., 2007; Vieno et al., 

2007b; Xu et al., 2007; Gulkowska et al., 2008; Paxeus, 2004). 

Together with treated wastewater, these compounds are 




* Corresponding author. Tel.: þ86 10 62787137; fax: þ86 10 62794006. 

E-mail address: yg-den@tsinghua.edu.cn (G. Yu). 


doi:10.1016/j.watres.2009.07.010 

The occurrence and removal of 13 pharmaceuticals and 2 consumer products, including 

antibiotic, antilipidemic, anti-inflammatory, anti-hypertensive, anticonvulsant, stimulant, 

insect repellent and antipsychotic, were investigated in four wastewater treatment plants 

(WWTPs) of Beijing, China. The compounds were extracted from wastewater samples by 

solid-phase extraction (SPE) and analyzed by ultra-performance liquid chromatography 

coupled with tandem mass spectrometry (UPLC–MS/MS). Most of the target compounds were 

detected, with the concentrations of 4.4 ng L 1 –6.6 mgL 1 and 2.2–320 ng L 1 in the influents 

and secondary effluents, respectively. These concentrations were consistent with their 

consumptions in China, and much lower than those reported in the USA and Europe. Most 

compounds were hardly removed in the primary treatment, while their removal rates ranging 

from 12% to 100% were achieved during the secondary treatment. In the tertiary treatment, 

different processes showed discrepant performances. The target compounds could not be 

eliminated by sand filtration, but the ozonation and microfiltration/reverse osmosis (MF/RO) 

processes employed in two WWTPs were very effective to remove them, showing their main 

contributions to the removal of such micro-pollutants in wastewater treatment. 


released to the aquatic environment, and consequently found 

to contaminate the receiving water bodies (Lindqvist et al., 

2005; Kasprzyk-Hordern et al., 2009), or even raw water sources 

of drinking water treatment plant (Ternes et al., 2002; 

Vieno et al., 2007a; Radjenovic et al., 2008). Meanwhile, results 

of toxicology studies have revealed that some pharmaceuticals 

are suspected to have direct toxicity to certain aquatic 

organisms (Ferrari et al., 2003; Jjemba, 2006; Grung et al., 2008; 

Quinn et al., 2008). Besides, their continual but undetectable 

effects could accumulate slowly, and finally lead to irreversible 

change on wildlife and human beings (Daughton and 

Ternes, 1999). Therefore, the occurrence and behavior of 

pharmaceuticals in the WWTPs, which are both the sink and 

source of the compounds, should be focused on. So far, 

concentrations of pharmaceuticals from various therapeutic 

classes in the WWTPs have been well documented in the

418 

North America (Thomas and Foster, 2005; Lishman et al., 

2006), Japan (Nakada et al., 2006) and some European countries 

(Ternes, 1998; Castiglioni et al., 2006; Santos et al., 2007; Vieno 

et al., 2007b; Jones et al., 2007; Paxeus, 2004). Reported species 

and concentrations of pharmaceuticals varied from country to 

country, and plant to plant, owing to the different usage 

patterns. Meanwhile, the removal efficiencies of pharmaceuticals 

also varied much (Nakada et al., 2006; Gulkowska et al., 

2008), indicating that the removal could be affected by both 

the compound-specific properties, and the factors concerning 

specific WWTPs, such as types of treatment processes, solids 

retention time (SRT), hydraulic retention time (HRT), 

temperature, etc. In recent years, very few studies about the 

situation in China have been reported. Only one specific 

therapeutic class, antibiotics, has been investigated by limited 

previous studies (Xu et al., 2007; Gulkowska et al., 2008; Chen 

et al., 2008). Therefore, it is necessary and important to 

investigate the occurrence and removal of pharmaceuticals 

from different therapeutic classes in the WWTPs of China. 

Due to the low efficiency of conventional wastewater 

treatment processes, some advanced treatment technologies 

have been evaluated. Ozonation was found to be effective to 

remove pharmaceuticals in real municipal WWTPs of Japan 

(Nakada et al., 2007; Okuda et al., 2008) and Germany (Ternes 

et al., 2003). Nanofiltration (NF) and reverse osmosis (RO) 

membrane filtration, the well-proven technologies to remove 

pharmaceuticals from different kinds of waters, have also been 

applied at bench, pilot and full scale (Khan et al., 2004; Nghiem 

et al., 2005; Drewes et al., 2005; Al-Rifai et al., 2007; Watkinson 

et al., 2007; Comerton et al., 2008; Radjenovic et al., 2008). 

Retention behavior of pharmaceuticals during the processes 

associated with physicochemical properties of pharmaceuticals, 

membranes as well as the solution chemistry, and 

mechanisms of pharmaceutical rejection have been discussed 

in Kimura et al. (2004), Nghiem et al. (2005), Nghiem and 

Coleman (2008) and Comerton et al. (2008). Recently, considering 

the requirement of reclaimed water, several advanced 

treatment facilities have been installed in the WWTPs of 

Beijing. However, the removal efficiency of micro-pollutants, 

such as pharmaceuticals, has not been evaluated yet. 

In the present study, we investigated the contamination 

levels of 13 pharmaceuticals and 2 consumer products from 8 

classes (i.e. antibiotic, antilipidemic, anti-inflammatory, antihypertensive, 

anticonvulsant, stimulant, insect repellent and 

antipsychotic) in four WWTPs of Beijing, China, which have 

different advanced treatment units, and evaluated the elimination 

efficiencies of the target pharmaceuticals. To the best 

of our knowledge, this is the first report on the occurrence and 

removal of pharmaceuticals and consumer products from 

multiple classes in the WWTPs of China, especially for the 

situation during the advanced treatment processes. 

2. Materials and methods 

2.1. Chemicals 

All the standards including chloramphenicol (CP), nalidixic 

acid (NA), trimethoprim (TP), bezafibrate (BF), clofibric acid 

(CA), gemfibrozil (GF), diclofenac (DF), indometacin (IM), 


ketoprofen (KP), mefenamic acid (MA), metoprolol (MTP), 

carbamazepine (CBZ), caffeine (CF), N,N-diethyl-meta-toluamide 

(DEET) and sulpiride (SP) (Appendix) were of analytical 

grade (>90%), and purchased from Sigma–Aldrich (Steinheim, 

Germany). Isotopically labeled compounds, used as internal 

standards, were 13 C-phenacetin obtained from Sigma– 

Aldrich, and 3 D-mecoprop from Dr. Ehrenstorfer (Augsburg, 

Germany). HPLC grade methanol, acetone, dichloromethane, 

hexane, as well as formic acid were provided by Dikma (USA), 

and ultra-pure water was produced by a Milli-Q unit (Millipore, 

USA). Stock solutions of individual compound were prepared in 

methanol and mixture standards with different concentrations 

were prepared by diluting the stock solutions before each 

analytical run. All the solutions were stored at 4 C in the dark. 

2.2. Sample collection 

Four full-scale municipal WWTPs, referred as A, B, C and D, 

were selected in our study. These WWTPs employ similar 

conventional treatment processes: primary treatment to 

remove particles coupled with secondary biological treatment. 

For the secondary biological treatment processes, 

WWTPs A and D employ anaerobic/anoxic/oxic (A 2 /O) activated 

sludge process, anoxic/oxic (A/O) activated sludge 

process is adopted in WWTP B, and WWTP C employs oxidation 

ditch (OD). Other detailed information on each WWTP, 

such as inhabitants served, daily flow, HRT and SRT are shown 

in Table 1. Part of the secondary effluents was further treated 

in WWTPs A, B and D, by the processes of ultrafiltration 

(UF)/ozone, sand filtration (SF) and microfiltration/reverse 

osmosis (MF/RO), respectively. In WWTP A, a dead-end 

ultrafiltration system (Zenon GE) is used. The whole system 

has 6 trains of Zee-Weed 1000 membrane. Each train contains 

9 cassettes of 57–60 modules per cassette. The membrane, 

with the pore size of 0.02 mm, is made by PVDF. The module is 

operated in an outside/in configuration at a constant flow of 

23 L (m 2 h) 1 and the total treatment capacity reaches 

80,000 m 3 d 1 . The membrane is hydraulically backwashed at 

a constant flow rate of 34 (m 2 h) 1 , and 29 times per day. The 

backwash phase lasts for 1 min. Maintenance cleaning is 

conducted once per day. Membranes are soaked in the sodium 

hypochlorite solution (50 mg L 1 ) for 25 min. For the ozonation 

process, gaseous ozone is generated from an ozone generator 

(Mitsubishi Electric). The ozone dosage and contact time in the 

reaction tank is 5 mg L 1 and 15 min, respectively. The pH of 

the wastewater before ozonation ranges 6.5–8.0 and shows no 

significant change after ozonation. As the heart of the 

advanced treatment in WWTP D, a spiral-wound crossflow 

module is employed for the reverse osmosis (RO) membrane 

filtration. The RO membrane (Filmtec, DOW) is made from 

a thin-film composite polyamide material. Each module is 

designed to operate at a water flux of 1.3 m 3 h 1 , and a product 

water recovery of 75–80%. The trans-membrane pressure is 

between 0.04 and 0.06 MPa, and the salt rejection remained at 

the level of 99%. Every 3–6 months, normally when the transmembrane 

pressure reaches above 0.06 MPa, the membrane is 

cleaned with 0.1%(w) sodium hydroxide solution (for organic 

foulants), 2%(w) citric acid (for inorganic foulants) and 0.5%(w) 

formaldehyde (as biocide). Schematic diagram of treatment 

processes in the four WWTPs is shown in Fig. 1.

The samples were collected once from the four WWTPs 

during June and July 2008, with no compensation for HRT. All 

of them were collected as grab samples in duplicate (500 mL 

for influents and 1000 mL for the others) in prewashed amber 

glass bottles, kept in the cooler and transported to the laboratory. 

Immediately after delivery to the laboratory, they were 

filtered through prebaked (400 C, >4 h) glass microfiber filters 

(GF/F, Whatman) to remove particles and stored at 4 C before 

extraction. 

2.3. Sample extraction and analysis 

The method for the extraction and analysis of pharmaceuticals 

and consumer products is presented elsewhere (Sui et al., 

in press) and briefly described here. After the solid-phase 

extraction (SPE) cartridges (Oasis, HLB, 200 mg, 6 mL) were 

a WWTP A 

Influent 

Screen 

b WWTP B 

Influent 

Screen 

c WWTP C 

Influent 

Screen 

d WWTP D 

Influent 

Screen 

Grit 

Removal 

Grit 

Removal 

Grit 

Removal 

Grit 

Removal 


Table 1 – Information of the WWTPs investigated. 

WWTP Inhabitants 

served 10 3 

Daily flow 

( 10 3 m 3 HRT (h) SRT (d) Secondary treatment Tertiary treatment 

) 

A 814 400 11 12–15 A 2 /O UF/ozone 

B 2400 1000 11 20 A/O SF 

C 480 200 15 12–16 OD – 

D 2415 600 10.67 15 A 2 /O MF/RO 

A/O 

treatment 

Primary 

Clarifier 

Primary 

Clarifier 

Oxidation 

Ditch 

conditioned, wastewater samples, added with internal standards 

and adjusted to pH ¼ 7, were introduced to the cartridge 

via a PTFE tube, at a flow rate of 5–10 mL min 1 . After washing 

by 5 mL of 5.0% methanol solution, the cartridge was dried 

under vacuum for 2 h and eluted with 5 mL of methanol. The 

extract was then concentrated to 0.4 mL under a gentle 

nitrogen stream and stored at 4 C for analysis. Concentrations 

of the target compounds were analyzed using ultraperformance 

liquid chromatography coupled with tandem 

mass spectrometry (UPLC–MS/MS). Analytes were separated 

using Waters Acquity UPLC system (Waters Corporation, USA) 

equipped with Acquity UPLC BEH C18 column (50 2.1 mm, 

particle size of 1.7 mm), and detected by Quattro Premier XE 

tandem quadrupole mass spectrometry (Waters Corp., USA) 

equipped with an electrospray ionization source. The analysis 

was carried out in multiple reaction monitoring (MRM) mode, 

Secondary 

Clarifier 

A 2 /O 

treatment 

A 2 /O 

treatment 

Secondary 

Clarifier 

Secondary 

Clarifier 

Ultrafiltration 

Secondary Effluent 

Secondary 

Clarifier 

Secondary 

Effluent 

Ozonation 

Sand Filtration 


MF 

` 

RO 


Tertiary 

Effluent 

Tertiary 

Effluent 

Tertiary 

Effluent 

Fig. 1 – Schematic diagram of the treatment processes in the four WWTPs and sampling site location (C).

420 

and in general, two precursor ion/product ion transitions were 

monitored for one compound with the purpose of quantification 

and confirmation. 

2.4. Quality control 

For each sampling, 500 mL Milli-Q water in an amber glass 

bottle as a field blank was brought to the WWTPs, exposed to 

the environment where the samples were taken from, and 

then delivered back to the laboratory with samples. For each 

set of samples (normally 10 samples), at least one procedural 

blank was prepared from ultra-pure water in the laboratory. 

Both the field blanks and procedural blanks were run identically 

to the wastewater samples, and the concentrations of 

target compounds were below the limit of quantification 

(LOQ). The absolute recoveries, calculated by comparing the 

concentrations of target compounds in spiked and unspiked 

wastewaters, were proved to be 73–102% and 50–95% in the 

effluent and influent for most compounds, respectively. While 

for several compounds (i.e. sulpiride, gemfibrozil, mefenamic 

acid), the absolute recoveries were not satisfactory. However, 

13 C-phenacetin and 3 D-mecoprop, the surrogate standards 

used for positive and negative ion mode respectively were able 

to compensate for the loss of most analytes, and relative 

recoveries were 67–130% for all the analytes in the effluent 

and 79–140% in the influents except mefenamic acid (251%) 

and nalidixic acid (178%). Therefore, the concentrations of 

these two compounds in the wastewater influents were not 

quantitatively determined and reported. The LOQs were 

0.3–5.5 ng L 1 and 0.7–20 ng L 1 in the effluent and influent, 

respectively. Detailed information about the calibration, 

recoveries, LOQ, matrix effects, etc. were described in Sui et al. 

(in press), and briefly listed in Table 2. As duplicate samples 

were collected at each sampling site, mean concentrations 

were adopted. In most cases, deviations of duplicate samples 

were less than 20%. For some tertiary effluent samples, low 

concentrations of some target compounds (i.e. caffeine, DEET, 

carbamazepine) resulted in slightly higher deviations. 

3. Result and discussion 

3.1. Influents 

As shown in Fig. 2, 12 target compounds were detected in all 

the influent samples from the four WWTPs, while ketoprofen 

was below LOQ in all wastewater samples. The most 

abundant compounds detected were the consumer products, 

caffeine (3.4–6.6 mgL 1 ) and N,N-diethyl-meta-toluamide 

(0.6–1.2 mgL 1 ), probably due to the large consumption of 

drinks containing caffeine (i.e. coffee, tea, etc.) and wide 

application of insect repellent during the summer time when 

we sampled. Diclofenac, trimethoprim, sulpiride, carbamazepine, 

indometacin and metoprolol showed relatively high 

concentrations (Fig. 2). A similar composition distribution 

was observed among all the influents of the four WWTPs. 

The concentrations of target pharmaceuticals except 

diclofenac and trimethoprim, were much lower than those 

reported in the European and North American countries 

(Thomas and Foster, 2005; Lishman et al., 2006; Vanderford 

and Snyder, 2006; Santos et al., 2007; Gomez et al., 2007; Vieno 

et al., 2007b; Huerta-Fontela et al., 2008). For instance, the 

concentrations of ketoprofen in the wastewater influents 

were recorded to be 2.0 0.6 mgL 1 in Finland (Lindqvist et al., 

2005), 200 ng L 1 in Australia (Al-Rifai et al., 2007), and 

300–1360 ng L 1 in Spain (Santos et al., 2007), while in the 

influents of four WWTPs in Beijing, it could not be detected. 

Concerning gemfibrozil, which is used to lower cholesterol 

and triglyceride levels in the blood, the contamination level 

found in the present study was 24–140 ng L 1 , even 1 or 2 order 

of magnitude lower than those in the USA (4770 ng L 1 , 

Vanderford and Snyder, 2006) and Canada (418 ng L 1 , 

Table 2 – Instrumental quantification limit (IQL), limit of quantification (LOQ), absolute recovery (AR), relative recovery (RR) 

and matrix suppression of target compounds. 

Compounds IQL (pg) LOQ (ng L 1 ) AR (n¼6, %) RR (n ¼ 6, %) Matrix effect (%) 

Effluent Influent Effluent a 

Influent a 

Effluent a 

Influent a 

BF 0.5 0.3 0.7 74 (3) 58 (5) 94 (4) 94 (10) 27 

CA 10 5.4 16 74 (4) 50 (4) 94 (6) 82 (8) 30 

CBZ 2.5 1.0 2.8 100 (4) 71 (12) 130 (6) 133 (23) 6 

CF 2.5 1.7 3.3 60 (3) 61 (37) 77 (4) 114 (70) 49 

CP 2.5 1.0 2.3 98 (7) 86 (9) 124 (10) 140 (17) 39 

DEET 2.5 1.3 3.2 76 (3) 63 (6) 99 (5) 118 (14) 24 

DF 10 4.7 9.4 86 (12) 85 (19) 109 (16) 139 (32) 4 

GF 10 4.2 20 95 (12) 40 (11) 120 (16) 65 (18) 28 

IM 2.5 1.3 2.8 80 (12) 73 (13) 101 (16) 119 (22) 1 

KP 10 5.5 18 69 (9) 44 (3) 87 (12) 72 (7) 46 

MA 10 5.5 5.2 73 (11) 154 (24) 92 (15) 251 (41) 13 

MTP 2.5 1.1 3.3 88 (3) 60 (3) 115 (5) 113 (9) 15 

NA 2.5 1.1 2.1 91 (9) 95 (9) 118 (12) 178 (20) 3 

SP 0.5 0.4 1.0 51 (5) 42 (3) 67 (7) 79 (7) 63 

TP 2.5 1.0 2.7 102 (7) 74 (16) 132 (9) 139 (31) 5 

a Value in the brackets refers to the deviation of the recovery. 


a 

Concentration (ng L -1 ) 

b 

Concentration (ng L -1 ) 

10000 

1000 

100 

10 

1 

10000 

1000 

100 

10 

1 

DEET 

DEET 

CF 

CF 

CP 

TP 

BF 

CA 

Lishman et al., 2006). The low levels of target pharmaceuticals 

were probably due to the lower per capita consumption in 

China than in the countries with higher socioeconomic 

statuses, where medical care is more prevalent (Thomas and 

Foster, 2005). The per capita consumption rate of gemfibrozil 

in China is estimated to be 0.036 mg person 1 d 1 (Table 3), 

lower than those in Germany (0.2 mg person 1 d 1 , Ternes, 

1998) and Canada (0.2 mg person 1 d 1 , Lishman et al., 2006). 

Since the levels of target pharmaceuticals were somewhat 

different from those of European and North American 

countries, we theoretically calculate the concentration of 

pharmaceuticals in the wastewater influent by the following 

equation (Lindqvist et al., 2005; Nakada et al., 2006) 

T e% I 1012 

Cpred ¼ 

365 P Q 

where C pred is the predicted concentration of the pharmaceutical 

in wastewater influent (ng L 1 ); T is the total production 

of a pharmaceutical both for human and animal use in 

China per year (ton year 1 ), P is the population of China, e% is 

the amount of the pharmaceutical excreted unchanged, I is 

the number of inhabitants served and Q is the influent flow 

(m 3 d 1 ). The predicted concentrations of gemfibrozil, diclofenac, 

indometacin, ketoprofen, carbamazepine, and sulpiride 

were comparable to those measured in the influents 

(Table 3). Much lower measured concentration than predicted 

concentration of chloramphenicol was probably because it 

had been forbidden for use in food and aquaculture in China 

since 2005, and the available data about the production of 

pharmaceuticals were based on the year of 2004. It should be 

noticed that since there is no available data on the total 

consumption of any pharmaceutical, we used figures for total 

production of individual pharmaceutical instead. Therefore, 

the differences between the amounts actually produced and 

applied as well as the amount used in human and veterinary 

medicine could not be distinguished, which might result in 

overestimation of the theoretical concentration. Nevertheless, 

the comparability between the predicted concentrations 

and measured concentrations illustrates the overall reasonability 

of the approach. 

Table 3 – Outputs, per capita consumption, predicted concentrations (PECs) and measured concentrations (MECs) of some 

pharmaceuticals in the wastewater influents of WWTPs investigated. 

Compound Output a 

(tons year 1 ) 

GF 

DF 

IM 

MTP 

Pharmaceuticals & Consumer Products 

CP 

TP 

BF 

CA 

GF 

DF 

IM 

MA 

MTP 

Influents 

CBZ 

CBZ 

SP 

Secondary effluents 

Pharmaceuticals & Consumer Products 

Fig. 2 – Concentrations of target pharmaceuticals in 

wastewater influents (a) and secondary effluents (b) of four 

WWTPs in Beijing. 


SP 

Per capita 

consumption 

(mg person 1 d 1 ) 

Excreted 

unchanged b (%) 

PEC 

(ng L 1 ) 

(1) 

MEC 

(mean, ng L 1 ) 

CP 1929 4.051 5–10 844 31 

TP 2352 4.939 45 6173 400 

GF 17 0.036 76 76 60 

DF 328 0.689 15 287 318 

IM 277 0.582 10–20 242 129 

KP 92 0.193 2.7 15 n.d. c 

CBZ 395 0.830 2–3 58 113 

SP 98 0.206 15–70 243 157 

a From CMEIN (2005). 

b From Bolton and Null (1981), Ternes (1998), Khan and Ongerth (2004), Niwa et al. (2005), Nakada et al. (2006), Jjemba (2006). 

c n.d. ¼ Not detected.

422 

3.2. Secondary effluent 

Similar to the influent samples, ketoprofen was below the LOQ 

in all the secondary effluent samples. Nalidixic acid and 

chloramphenicol were detected only in one WWTP, with the 

concentration of 8.1 and 19 ng L 1 , respectively. The mean 

concentrations of the other 12 compounds ranged from 5 to 

200 ng L 1 (Fig. 2). Diclofenac, N,N-diethyl-meta-toluamide, 

trimethoprim, sulpiride and indometacin showed high 

concentrations in the secondary effluents. Carbamazepine 

and metoprolol followed, with the concentrations ranging 

from 69 to 120 ng L 1 , and 60 to 108 ng L 1 , respectively. Other 

compounds, such as caffeine, gemfibrozil and mefenamic 

acid, occurred at the lowest levels. Despite of a wide variation 

of trimethoprim from different WWTPs, the composition 

profiles of target pharmaceuticals in secondary effluents from 

the four WWTPs were quite similar (Fig. 3). 

The concentration levels of most pharmaceuticals and 

consumer products detected in the secondary effluent were 

also lower than those reported in the Europe. They were over 

100 ng L 1 , in some cases even up to 500 ng L 1 in the wastewater 

effluents of the European countries (Santos et al., 2007; 

Gomez et al., 2007; Ternes, 1998; Vieno et al., 2007b). While in 

the present study, 10 out of 15 compounds were less than 

100 ng L 1 , and none of them exceeded 400 ng L 1 in any 

effluent samples (Fig. 2). Our results were in agreement with 

those in Japan (Nakada et al., 2006), Korea (Kim et al., 2007) and 

some other cities of China (Xu et al., 2007; Gulkowska et al., 

2008; Chen et al., 2008). For instance, the concentrations of 

chloramphenicol in the effluents of 4 WWTPs in Guangzhou 

were

et al., 2008; Gomez et al., 2007). The removal rate of bezafibrate 

was found to be 87% in six Italian WWTPs in summer time 

(Castiglioni et al., 2006), very similar to that observed in our 

study. The concentrations of DEET were decreased by more 

than 80% during the biological treatment in the WWTP of 

Japan (Okuda et al., 2008), slightly better than our results. A 

second group of pharmaceuticals, including three antiinflammatory 

drugs, clofibric acid, gembrozil, metoprolol and 

sulpiride, had lower removal rates with large variation in 

different WWTPs studied. For instance, 28–53% of diclofenac, 

a representative of the anti-inflammatory drugs, was removed 

by secondary treatment in the WWTPs, which was between 

26% in Finland (Lindqvist et al., 2005) and 69% in Germany 

(Ternes, 1998). The elimination of these compounds may be 

highly dependent on the configurations and operation conditions 

of individual WWTP as well as wastewater characteristics, 

and thus no definitive conclusion could be reached. 

Higher load of carbamazepine was found in the secondary 

effluent than in the primary effluent, indicating negative 

removal efficiency during the secondary treatment. Some 

carbamazepine was found to be excreted as the form of 

conjugates (Vieno et al., 2007b), which was biodegraded to 

carbamazepine by enzymatic processes during the secondary 

treatment, resulting in additional amounts of carbamazepine 

in the secondary effluent. However, as the calculations of all 

the removal efficiencies were based on grab samples that were 

not sampled with a hydraulic lag in the present study, some 

error might be brought in due to diurnal variation of the 

concentration. Therefore, the present study only provided 

a snapshot of the removal of pharmaceuticals and consumer 

products in the WWTPs of Beijing. To better illustrate that, 

24-h composite samples that are lagged by HRT should be 

collected and analyzed in further studies. 

It has been reported that high HRT (>12 h) and SRT (>10 d) 

may contribute to an increased removal rate of pharmaceuticals 

(Jones et al., 2007; Vieno et al., 2007b). In the present 

study, the WWTP C, in which the HRT was higher than the 

others, was the best in removing these compounds, due to 

increased contact time of target compounds and the microorganisms. 

On the other hand, the different SRTs did not have 

significant effects on the removal efficiency, probably because 

the SRTs in all the four WWTPs were relatively high (>10 d), 

and without large differences. In addition, it is noteworthy 

that the WWTP C employed oxidation ditches, which showed 

better removal of natural estrogens and estrogenic activity 

than A/O (Hashimoto et al., 2007). It also could be the reason 

for the higher removal efficiencies in the WWTP C. Further 

investigation for different types of WWTPs is necessary to 

confirm the results mentioned above. 

3.4. Removal efficiency in advanced treatment processes 

The removal efficiencies of the pharmaceuticals during the SF, 

UF/ozonation, as well as MF/RO treatment in three corresponding 

WWTPs are listed in Table 5. 

Generally, sand filtration was not effective for these 

compounds. Only trimethoprim, DEET and gemfibrozil were 

removed slightly during this treatment process. It should be 

noticed that these compounds were efficiently removed in the 

secondary treatment, indicating that the biodegradation on 


Table 5 – Removal efficiencies (%) of target 

pharmaceuticals and consumer products by advanced 

treatment processes in studied WWTPs. 

Compound WWTP A WWTP B WWTP D 

UF Ozone SF MF/RO 

DEET 0–50 50–80 0–50 >90 

CF 90 

BF 0–50 0–50 0–50 >90 

CA 90 90 

IM 0–50 >90 90 

MA 0–50 80–90 90 

SP 0–50 >90 0–50 >90 

the biofilm present on the sand particle, rather than the 

removal with particles, may be the main reason for their 

elimination (Gobel et al., 2007). 

The results showed that ozonation is effective in removing 

most of the target compounds, probably due to the operation 

conditions employed in WWTP A (ozone dosage: 5 mg L 1 , 

contact time: 15 min). Carbamazepine, diclofenac, indomethacin, 

sulpiride and trimethoprim were significantly 

eliminated, with the removal rates of above 95%. The double 

bond in the azepine ring of carbamazepine and pyrrole ring of 

indomethacin, and the non-protonated amine of diclofenac 

and trimethoprim were susceptible to ozone attack (Vieno 

et al., 2007a; Nakada et al., 2007; Westerhoff et al., 2005). The 

removal efficiencies of DEET and metoprolol were modest. 

The amide group, which is not reactive with ozone, could be 

the reason for the modest removal of DEET (Nakada et al., 2007). 

Low removal efficiencies were found for bezafibrate, clofibric 

acid, as well as caffeine. Only 14% of bezafibrate disappeared 

in the ozone process, consistent with its low rate constants 

with ozone (590 50 M 1 S 1 , Huber et al., 2003). The reaction 

site of bezafibrate is the R-oxysubstituent (–O–C(CH 3) 2COOH) 

on one of the aromatic rings. However, as the pK a of bezafibrate 

is 3.6, the R-oxysubstituent cannot be deprotonated and 

consequently the overall rate constant at pH > 4 is much lower 

(Huber et al., 2003). It should be noticed that during the 

ozonation, most of the pharmaceuticals were not mineralized 

but transformed to the oxidation products. For instance, three 

oxidation products containing quinazoline-based functional 

groups were identified during the ozonation of CBZ (Mcdowell 

et al., 2005). 

The good performance of ozonation in the present study 

was consistent with Ternes et al. (2003), Huber et al. (2005) 

and Okuda et al. (2008). When 5 mg L 1 ozone was applied to 

the effluent of a municipal WWTP in Germany (contact time: 

18 min), target compounds, such as trimethoprim, carbamazepine, 

indomethacin, clofibric acid, were removed by 

more than 50% (Ternes et al., 2003). Huber et al. (2005) 

conducted a pilot study on the oxidation of pharmaceuticals 

during ozonation of conventional activated sludge (CAS) and 

membrane bio-reactor (MBR) effluents with various ozone 

dosages, and found that macrolide and sulfonamide antibiotics, 

estrogens, and acidic pharmaceuticals diclofenac,

424 

naproxen and indomethacin were oxidized by more than 

90–99% for ozone doses 2mgL 1 in all effluents. 

The elimination by ultrafiltration in the WWTP A was low 

for all the investigated compounds. The molecular weight cutoff 

(MWCO) of UF membranes was much higher than 1000 Da, 

thus UF membranes showed poor retention of all the investigated 

pharmaceuticals, of which the molecular weight are 

less than 400 Da. The removal of individual target compound 

was less than 50%, and might be due to the adsorption onto 

the membrane. It has been also demonstrated that UF 

membrane typically had less than 40% retention of 27 PPCPs, 

and the mass balances calculated based on the concentration 

of each compound in feed, permeate and retentate showed 

the observed retention was significantly governed by adsorption 

(Yoon et al., 2006). 

In contrast, MF/RO employed in WWTP D was very effective. 

In the effluent of MF/RO, all the target compounds except 

caffeine were not detected. Generally, one or combination of 

three basic mechanisms could be involved during the rejection 

of solute by NF/RO membrane: steric effect, charge 

exclusion and adsorption (Radjenovic et al., 2008). For most 

pharmaceuticals, the rejections were considered to be dominated 

by steric interaction in ‘‘tight’’ NF or RO membrane 

filtration (Nghiem et al., 2005; Radjenovic et al., 2008). As most 

investigated compounds have molecular weights about 

200–400 Da, smaller than MWCO of RO membrane applied, 

excellent rejection of most pharmaceuticals by RO membrane 

was observed in this study as well as in previous studies 

(Kimura et al., 2004; Al-Rifai et al., 2007; Radjenovic et al., 

2008). Besides, membrane fouling and the presence of organic 

matter in the wastewater effluents likely contributed to higher 

rejections of pharmaceuticals, especially for some hydrophobic 

ionogenic compound (Nghiem and Coleman, 2008; 

Comerton et al., 2008). 

Nevertheless, the rejections of two compounds, caffeine 

and mefenamic acid were slightly lower (i.e. 50–80% and 

0–50%, respectively). The concentration of mefenamic acid in 

feed wastewaters of MF/RO membrane process was very low, 

only a bit higher than its LOQ in the wastewater effluent, 

which could be the reason for the low rejection rate. The low 

retention of caffeine in the present study was inaccordance 

with Drewes et al. (2005). They found that in two full-scale RO 

facilities, target EDCs and PPCPs were efficiently rejected to 

below detection limit except for caffeine, still detected in the 

permeates. The physiochemical properties might explain the 

low rejection rate of caffeine. As a representative of hydrophilic 

and non-ionic compounds, the rejection driven by 

charge exclusion and adsorption is negligible, and steric 

exclusion is solely responsible for the retention of caffeine 

(Nghiem et al., 2005). However, the molecular weight of 

caffeine is 195 Da, smaller than other target compounds, and 

might result in the decreased removal efficiency during the RO 

membrane filtration process. 

Compared to the other two, the WWTPs employing ozone 

and RO membrane filtration as advanced treatment were 

more efficient in removing pharmaceuticals. For these 

WWTPs, the advanced treatment made a significant contribution 

to the total elimination of most pharmaceuticals 

(Fig. 5). Therefore, the utility of efficient advanced treatment 

could be considered as a tool to reduce pharmaceuticals in the 


a 

Removal Contribution (%) 

b 

Contribution to removal efficiency (%) 

120 

100 

80 

60 

40 

20 

0 

-20 

180 

160 

140 

120 

100 

80 

60 

40 

20 

0 

-20 

-40 

-60 

DEET 

DEET 

CF 

CP 

municipal wastewater treatment plants. However, the problems 

of membrane fouling and further treatment or disposal 

of retentate challenge the application of RO membrane 

filtration (Van der Bruggen et al., 2008). For ozonation, as most 

of the pharmaceuticals could not be mineralized, and oxidation 

products are formed from parent pharmaceutical 

compounds (Mcdowell et al., 2005), more research is required 

to identify the oxidation products and their potential toxicity 

during the partial oxidation process (Nakada et al., 2007). 

Besides, economic feasibility should be evaluated by estimating 

the energy consumption and investment and 

operation costs for both advanced treatment processes (Joss 

et al., 2008). 

4. Conclusion 

Tertiary treatment 

Conventional treatment 

CF 

TP 

TP 

BF 

BF 

13 out of 15 pharmaceuticals and consumer products from 

eight classes were detected at four WWTPs in Beijing, China. 

The concentrations of most compounds in the influent and 

secondary effluent were lower than those reported in the USA 

and Europe, but consistent with the production profile of the 

CA 

GF 

DF 

Pharmaceutical 

CA 

GF 

DF 


IM 

IM 

MTP 

MTP 

CBZ 

CBZ 

SP 

SP 

Tertiary treatment 

Secondary treatment 

Primary treatment 

Fig. 5 – Contributions of primary, secondary (or 

conventional treatment) and tertiary treatment to the total 

elimination of selected pharmaceuticals in WWTP A (a) 

and WWTP D (b).

pharmaceuticals in China. According to the result of risk 

assessment for the secondary effluent, only diclofenac might 

pose a risk to the aquatic environment. The removal efficiencies 

by the conventional treatment varied for different 

compounds, depending on their chemical structures, physiochemical 

properties, as well as the specific treatment 

processes utilized at each WWTP. Further removal could be 

achieved by adopting some advanced treatment processes, 

such as ozonation and MF/RO. However, others, such as 

sand filtration, showed low efficiency in removing these 

compounds from secondary effluent. 

Acknowledgement 

This study was supported by the National Science Fund for 

Distinguished Young Scholars (No. 50625823). 

Appendix. 


Supplementary information related to this article can be 

found at doi:10.1016/j.watres.2009.07.010. 

references 

Al-Rifai, J.H., Gabelish, C.L., Schafer, A.I., 2007. Occurrence of 

pharmaceutically active and non-steroidal estrogenic 

compounds in three different wastewater recycling schemes 

in Australia. Chemosphere 69 (5), 803–815. 

Bolton, S., Null, G., 1981. Caffeine: psychological effects, use and 

abuse. Orthomolecular Psychiatry 10 (3), 202–211. 



treatment plants in Italy. Environmental Science and 

Technology 40 (1), 357–363. 

Chen, H.C., Wang, P.L., Ding, W.H., 2008. Using liquid 

chromatography–ion trap mass spectrometry to determine 

pharmaceutical residues in Taiwanese rivers and 

wastewaters. Chemosphere 72 (6), 863–869. 

China Medicine Economic Information Net (CMEIN), 2005. 

Chinese Medical Statistical Yearbook. Products section. China 

Medicine Economic Information Net (CMEIN), Beijing, China. 

Comerton, A.M., Andrews, R.C., Bagley, D.M., Hao, C.Y., 2008. The 

rejection of endocrine disrupting and pharmaceutically active 

compounds by NF and RO membranes as a function of 

compound and water matrix properties. Journal of Membrane 

Science 313 (1–2), 323–335. 

Daughton, C.G., Ternes, T.A., 1999. Pharmaceuticals and personal 

care products in the environment: agents of subtle change? 

Environmental Health Perspectives 107 (S6), 907–938. 

Drewes, J.E., Bellona, C., Oekekoven, M., Xu, P., Kim, T.U., Amy, G., 

2005. Rejection of wastewater-derived micropollutants in 

high-pressure membrane applications leading to indirect 

potable reuse. Environmental Progress 24 (4), 400–409. 

European Environment Agency, 1998. Environmental Risk 

Assessment – Approaches, Experiences and Information 

Source. European Environment Agency, London, pp. 68–86. 

Ferrari, B., Paxeus, N., Giudice, R.L., Pollio, A., Grric, J., 2003. 

Ecotoxicological impact of pharmaceuticals found in treated 


wastewaters: study of carbamazepine, clofibric acid, and 

diclofenac. Ecotoxicology and Environmental Safety 55 (3), 

359–370. 

Gobel, A., Mcardell, C.S., Joss, A., Siegrist, H., Giger, W., 2007. Fate 

of sulfonamides, macrolides, and trimethoprim in different 

wastewater treatment technologies. Science of the Total 

Environment 372 (2–3), 361–371. 

Gomez, M.J., Bueno, M.J.M., Lacorte, S., Fernandez-Alba, A.R., 

Aguera, A., 2007. Pilot survey monitoring pharmaceuticals and 

related compounds in a sewage treatment plant located on the 

Mediterranean coast. Chemosphere 66 (6), 993–1002. 

Grung, M., Kallqvist, T., Sakshaug, S., Skurtveit, S., Thomas, K.V., 

2008. Environmental assessment of Norwegian priority 

pharmaceuticals based on the EMEA guideline. Ecotoxicology 

and Environmental Safety 71 (2), 328–340. 

Gulkowska, A., Leung, H.W., So, M.K., Taniyasu, S., Yamashita, N., 

Yeung, L.W.Y., Richardson, B.J., Lei, A.P., Giesy, J.P., Lam, P.K.S., 

2008. Removal of antibiotics from wastewater by sewage 

treatment facilities in Hongkong and Shenzhen, China. Water 

Research 42 (1–2), 395–403. 

Hashimoto, T., Onda, K., Nakamura, Y., Tada, K., Miya, A., 

Murakami, T., 2007. Comparison of natural estrogen removal 

efficiency in the conventional activated sludge process and 

the oxidation ditch process. Water Research 41 (10), 

2117–2126. 

Huber, M.M., Canonica, S., Park, G.Y., Gunten, U.V., 2003. 

Oxidation of pharmaceuticals during ozonation and advanced 

oxidation processes. Environmental Science and Technology 

37 (5), 1016–1024. 

Huber, M.M., Gobel, A., Joss, A., Hermann, N., Loffler, D., 

Mcardell, C.S., Ried, A., Siegrist, H., Ternes, T.A., Gunten, U.V., 

2005. Oxidation of pharmaceuticals during ozonation of 

municipal wastewater effluents: a pilot study. Environmental 

Science and Technology 39 (11), 4290–4299. 

Huerta-Fontela, M., Galceran, M.T., Martin-Alonso, J., Ventura, F., 

2008. Occurrence of psychoactive stimulatory drugs in 

wastewaters in north-eastern Spain. Science of the Total 

Environment 397 (1–3), 31–40. 

Huschek, G., Hansen, P.D., Maurer, H.H., Krengel, D., Kayser, A., 

2004. Environmental risk assessment of medicinal products 

for human use according to European Commission 

recommendations. Environmental Toxicology 19 (3), 226–240. 

Jjemba, P.K., 2006. Excretion and ecotoxicity of pharmaceutical 

and personal care products in the environment. Ecotoxicology 

and Environmental Safety 63 (1), 113–130. 

Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2005. Human 

pharmaceuticals in wastewater treatment processes. Critical 

Reviews in Environmental Science and Technology 35 (4), 

401–427. 

Jones, O.A.H., Voulvoulis, N., Lester, J.N., 2007. The occurrence 

and removal of selected pharmaceutical compounds in 

a sewage treatment works utilising activated sludge 

treatment. Environmental Pollution 145 (3), 738–744. 

Joss, A., Siegrist, H., Ternes, T.A., 2008. Are we about to upgrade 

wastewater treatment for removing organic micropollutants. 

Water Science and Technology 57 (2), 251–255. 

Kasprzyk-Hordern, B., Dinsdale, R.M., Guwy, A.J., 2009. The 

removal of pharmaceuticals, personal care products, 

endocrine disruptors and illicit drugs during wastewater 

treatment and its impact on the quality of receiving waters. 

Water Research 43 (2), 363–380. 

Khan, S.J., Wintgens, T., Sherman, P., Zaricky, J., Schafer, A.I., 

2004. Removal of hormones and pharmaceuticals in the 

advanced water recycling demonstration plant in Queensland, 

Australia. Water Science and Technology 50 (5), 15–22. 

Khan, S.J., Ongerth, J.E., 2004. Modelling of pharmaceutical 

residues in Australian sewage by quantities of use and 

fugacity calculations. Chemosphere 54 (3), 355–367.

426 

Kim, S.D., Cho, J., Kim, I.S., Vanderford, B.J., Snyder, S.A., 2007. 

Occurrence and removal of pharmaceuticals and endocrine 

disruptors in South Korean surface, drinking, and waste 

waters. Water Research 41 (5), 1013–1021. 





Lishman, L., Smyth, S.A., Sarafin, K., Kleywegt, S., Toito, J., 

Peart, T., Lee, B., Servos, M., Beland, M., Seto, P., 2006. 

Occurrence and reductions of pharmaceuticals and personal 

care products and estrogens by municipal wastewater 

treatment plants in Ontario, Canada. Science of the Total 

Environment 367 (2–3), 544–558. 

Lindqvist, N., Tuhkanen, T., Kronberg, L., 2005. Occurrence of 

acidic pharmaceuticals in raw and treated sewages and in 

receiving waters. Water Research 39 (11), 2219–2228. 

Mcdowell, D.C., Huber, M.M., Wagner, M., Gunten, U.V., 

Ternes, T.A., 2005. Ozonation of carbamazepine in drinking 

water: identification and kinetic study of major oxidation 

products. Environmental Science and Technology 39 (20), 

8014–8022. 

Nakada, N., Tanishima, T., Shinohara, H., Kiri, K., Takada, H., 

2006. Pharmaceutical chemicals and endocrine disrupters in 

municipal wastewater in Tokyo and their removal during 

activated sludge treatment. Water Research 40 (17), 3297–3303. 

Nakada, N., Shinohara, H., Murata, A., Kiri, K., Managaki, S., 

Sato, N., Takada, H., 2007. Removal of selected 

pharmaceuticals and personal care products (PPCPs) and 

endocrine-disrupting chemicals (EDCs) during sand filtration 

and ozonation at a municipal sewage treatment plant. Water 

Research 41 (19), 4372–4382. 

Nghiem, L.D., Schafer, A.I., Elimelech, M., 2005. Pharmaceutical 

retention mechanisms by nanofiltration membrane. 


Nghiem, L.D., Coleman, P.J., 2008. NF/RO filtration of the 

hydrophobic ionogenic compound triclosan: transport 

mechanisms and the influence of membrane fouling. 

Separation and Purification Technology 62 (3), 709–716. 

Niwa, T., Inoue, S., Shiraga, T., Takagi, A., 2005. No inhibition of 

cytochrome P450 activities in human liver microsomes by 

sulpiride, an antipsychotic drug. Biological and 

Pharmaceutical Bulletin 28 (1), 188–191. 

Okuda, T., Kobayashi, Y., Nagao, R., Yamashita, N., Tanaka, H., 

Tanaka, S., Fujii, S., Konishi, C., Houwa, I., 2008. Removal 

efficiency of 66 pharmaceuticals during wastewater treatment 

process in Japan. Water Science and Technology 57 (1), 65–71. 

Paxeus, N., 2004. Removal of selected non-steroidal antiinflammatory 

drugs (NSAIDs), gemfibrozil, carbamazepine, 

beta-blockers, trimethoprim and triclosan in conventional 

wastewater treatment plants in five EU countries and their 

discharge to the aquatic environment. Water Science and 

Technology 50 (5), 253–260. 

Quinn, B., Gagne, F., Blaise, C., 2008. An investigation into the acute 

and chronic toxicity of eleven pharmaceuticals (and their 

solvents) found in wastewater effluent on the cnidarian, Hydra 

attenuata. Science of the Total Environment 389 (2–3), 306–314. 

Radjenovic, J., Petrovic, M., Ventura, F., Barcelo, D., 2008. Rejection 

of pharmaceuticals in nanofiltration and reverse osmosis 


membrane drinking water treatment. Water Research 42 (14), 

3601–3610. 

Santos, J.L., Aparicio, I., Alonso, E., 2007. Occurrence and risk 

assessment of pharmaceutically active compounds in 

wastewater treatment plants. A case study: Seville city 

(Spain). Environment International 33 (4), 596–601. 

Sui, Q., Huang, J., Deng, S. B., Yu, G. Rapid determination of 

pharmaceuticals from multiple therapeutic classes in 

wastewater by solid-phase extraction and ultra-performance 

liquid chromatography tandem mass spectrometry. Chinese 

Science Bulletin, in press. 

Ternes, T.A., 1998. Occurrence of drugs in German sewage 

treatment plants and rivers. Water Research 32 (11), 

3245–3260. 

Ternes, T.A., Meisenheimer, M., Mcdowell, D., Sacher, F., 

Brauch, H.J., Haist-gulde, B., Preuss, G., Wilme, U., 

Zulei-seibert, N., 2002. Removal of pharmaceuticals during 

drinking water treatment. Environmental Science and 

Technology 36 (17), 3855–3863. 

Ternes, T.A., Stuber, J., Herrmann, N., Mcdowell, D., Ried, A., 

Kampmann, M., Teiser, B., 2003. Ozonation: a tool for removal 

of pharmaceuticals, contrast media and musk fragrances from 

wastewater? Water Research 37 (8), 1976–1982. 

Thomas, P.M., Foster, G.D., 2005. Tracking acidic 

pharmaceuticals, caffeine, and triclosan through the 

wastewater treatment process. Environmental Toxicology and 

Chemistry 24 (1), 25–30. 

Vanderford, B.J., Snyder, S.A., 2006. Analysis of pharmaceuticals 

in water by isotope dilution liquid chromatography/tandem 

mass spectrometry. Environmental Science and Technology 

40 (23), 7312–7320. 

Van der Bruggen, B., Manttari, M., Nystrom, M., 2008. 

Drawbacks of applying nanofiltration and how to avoid them: 

a review. Separation and Purification Technology 63 (2), 

251–263. 

Vieno, N.M., Harkki, H., Tuhkanen, T., Kronberg, L., 2007a. 

Occurrence of pharmaceuticals in river water and their 

elimination in a pilot-scale drinking water treatment plant. 


Vieno, N., Tuhkanen, T., Kronberg, L., 2007b. Elimination of 

pharmaceuticals in sewage treatment plants in Finland. Water 

Research 41 (5), 1001–1012. 

Watkinson, A.J., Murby, E.J., Costanzo, S.D., 2007. Removal of 

antibiotics in conventional and advanced wastewater 

treatment: implications for environmental discharge and 

wastewater recycling. Water Research 41 (18), 4164–4176. 

Westerhoff, P., Yoon, Y., Snyder, S., Wert, E., 2005. Fate of 

endocrine-disruptor, pharmaceutical, and personal care 

product chemicals during simulated drinking water treatment 

processes. Environmental Science and Technology 39 (17), 

6649–6663. 

Xu, W.H., Zhang, G., Li, X.D., Zou, S.C., Li, P., Hu, Z.H., Li, J., 2007. 

Occurrence and elimination of antibiotics at four sewage 

treatment plants in the Pearl River Delta, South China. Water 

Research 41 (19), 4526–4534. 

Yoon, Y., Westerhoff, P., Snyder, S.A., Wert, E.C., 2006. 

Nanofiltration and ultrafiltration of endocrine disrupting 

compounds, pharmaceuticals and personal care products. 

Journal of Membrane Science 270 (1–2), 88–100.

Application of the combination index (CI)-isobologram 

equation to study the toxicological interactions of lipid 

regulators in two aquatic bioluminescent organisms 

Ismael Rodea-Palomares a,1 , Alice L. Petre b,1 , Karina Boltes b , Francisco Leganés a , 

José Antonio Perdigón-Melón b , Roberto Rosal b , Francisca Fernández-Piñas a, * 

a 

Departamento de Biología, Facultad de Ciencias, Universidad Autónoma de Madrid, 2 Darwin Street, Cantoblanco, 28049 Madrid, Spain 

b 

Departamento de Ingeniería Química, Universidad de Alcalá, Alcalá de Henares, E-28871 Madrid, Spain 

article info 


Received 26 March 2009 


13 July 2009 



Keywords: 

Antagonism 

Combination index-isobologram 

equation 

Cyanobacterium 

Fibrates 

Synergism 

Vibrio fischeri 


abstract 

Fibrates and statins (HMG-CoA reductase inhibitors) are the 

main lipid-lowering drugs prescribed either alone or in 

combination therapy in order to decrease plasma cholesterol 

levels and reduce the incidence of coronary heart disease. 

Although partially displaced by statins, the total number of 

fibrate prescriptions is in constant increase in the United 

States (Holoshitz et al., 2008). Fibric acids are the active forms 





E-mail address: francisca.pina@uam.es (F. Fernández-Piñas). 

1 

Both authors contributed equally to this work. 


doi:10.1016/j.watres.2009.07.026 

Pharmaceuticals in the aquatic environment do not appear singly and usually occur as 

complex mixtures, whose combined effect may exhibit toxicity to the aquatic biota. We 

report an environmental application of the combination index (CI)-isobologram equation, 

a method widely used in pharmacology to study drug interactions, to determine the nature 

of toxicological interactions of three fibrates toward two aquatic bioluminescent organisms, 

Vibrio fischeri and the self-luminescent cyanobacterial recombinant strain Anabaena 

CPB4337. The combination index-isobologram equation method allows computerized 

quantitation of synergism, additive effect and antagonism. In the Vibrio test, the fibrate 

combinations showed antagonism at low effect levels that turned into an additive effect or 

synergism at higher effect levels; by contrast, in the Anabaena test, the fibrate combinations 

showed a strong synergism at the lowest effect levels and a very strong antagonism at high 

effect levels. We also evaluated the nature of the interactions of the three fibrates with a real 

wastewater sample in the cyanobacterial test. We propose that the combination indexisobologram 

equation method can serve as a useful tool in ecotoxicological assessment. 


of fibrates and belong to the nuclear receptor superfamily of 

ligand-activated transcription factors. Gemfibrozil and fenofibrate 

are the fibrates currently marketed in the US, whereas 

bezafibrate is also available in Europe and other developed 

countries (Lambropoulou et al., 2008). Fenofibric acid, 2-[4-(4chlorobenzoyl)phenoxy]-2-methylpropanoic 

acid, is the active 

metabolite of fenofibrate, the inactive prodrug marketed and 

dispensed. Gemfibrozil, 5-(2,5-dimethylphenoxy)-2,2-dimethylpentanoic 

acid and bezafibrate, p-[4-[chlorobenzoylamino-

428 

ethyl]-phenoxy]-b-methylpropionic acid, are also fibric acid 

derivatives with similar pharmacokinetic behaviour (Miller and 

Spence, 1998). 

The occurrence of lipid regulators in the discharge of 

treated urban and municipal wastewater has been relatively 

well documented. Bezafibrate has been detected in effluents 

of two British STP with averages up to 230 ng/L (Kasprzyk- 

Hordern et al., 2009). Metcalfe et al. (2003) found around 1 mg/L 

of gemfibrozil in effluents of Canadian STP, whereas fenofibrate 

has been reported in concentrations up to 0.5 mg/L in the 

influent of several Brazilian STP (Stumpf et al., 1999). 

Andreozzi et al. (2003) found lipid regulators in the effluent of 

several European STP at concentrations up to 4.76 mg/L (gemfibrozil), 

1.07 mg/L (bezafibrate) and 0.16 mg/L (fenofibrate). 

Rosal et al. (2008), reported the occurrence of bezafibrate and 

gemfibrozil at levels of 139 and 608 ng/L respectively in the 

effluent of a Spanish STP. In the same plant Rodríguez et al. 

(2008) found 165 ng/L of fenofibric acid, 61 ng/L of bezafibrate 

and 143 ng/L of gemfibrozil. 

It is also significant that removal efficiencies observed in 

current STP are not always high. Fent et al. (2006) reported 

maximum removal rates of 50–75% for fenofibric acid and 

gemfibrozil and somewhat higher for bezafibrate, although for 

the later, efficiencies below 15% have also been reported. 

Stumpf et al. (1999) reported a 45% removal of fenofibric acid 

by an activated sludge conventional treatment. Kasprzyk- 

Hordern et al. (2009) encountered an average degradation of 

bezafibrate not higher than 67%. On the other hand, Castiglioni 

et al. (2006) reported that the removal efficiency of 

bezafibrate during an activated sludge treatment greatly 

varied from 15% in winter to 87% in summer. 

At measured environmental concentrations as those 

reported above (mostly in the ng/L and mg/L range), many 

studies have shown that the risk of acute toxicity is unlikely 

(Fent et al., 2006; Han et al., 2006; Borgmann et al., 2007); 

however, there is a lack of data on chronic toxicity effects. 

Moreover, pharmaceuticals in the aquatic environments occur 

as complex mixtures from different classes, not as single 

contaminants (Gros et al., 2007); thus, although the concentration 

of individual pharmaceuticals is low, their mixture 

could prove ecotoxicologically significant (Brain et al., 2004). 

Current methods of risk assessment usually focus on the 

assessment of single chemicals, which may underestimate 

the risk associated with toxic action of mixtures; probably for 

this reason, in the last years there is an increasing number of 

studies dealing with complex mixtures of pharmaceuticals 

(Cleuvers, 2003, 2004; Crane et al., 2006; Han et al., 2006; 

Borgmann et al., 2007; Christensen et al., 2007; Pomati et al., 

2008; Quinn et al., 2009). However, assessment of combined 

toxicities is not an easy issue. Basically, two different models 

are in use for the prediction of mixture toxicity, i.e., concentration 

addition, when pharmaceuticals have a similar mode 

of toxic action, and response addition or independent action, 

when pharmaceuticals have different modes of toxic action 

(Cleuvers, 2003; Teuschler, 2007). However, toxicological 

interactions, synergisms or antagonisms, between the pharmaceuticals 

and their effects can occur independently of 

mode of action; moreover, in most cases, the pharmacological 

mechanisms of action is known but the toxic mode of action 

may remain unknown (Cleuvers, 2003; Chou, 2006). In an effort 


to overcome this limitation, we report an environmental 

application of a method widely used in pharmacology to 

interpret drug interactions; this method, termed as the 

median-effect/combination index (CI)-isobologram equation 

(Chou, 2006) allows quantitative determinations of chemical 

interactions where CI 1 indicate synergism, 

additive effect and antagonism, respectively. One important 

property of the method is that previous knowledge of the 

mechanisms of action of each chemical is not required. 

Besides, the method takes into account both the potency and 

the shapes of the dose-effect curve of each chemical. The 

method has been computerized allowing an automated 

simulation of synergism and antagonism at different 

concentrations and at different effect levels of the chemicals 

in a mixture. 

The aim of our study was to assess the nature of the 

toxicological interactions of three fibrates, gemfibrozil, bezafibrate 

and fenofibric acid, by the method of combination 

index (CI)-isobologram equation. The three pharmaceuticals 

were used singly or in two- and three-drug combinations. As 

toxicity endpoint we have chosen the bioluminescent 

response of two prokaryotes, the naturally luminescent Vibrio 

fischeri and the recombinant bioluminescent cyanobacterium 

Anabaena sp. PCC 7120 CPB4337 (hereinafter, Anabaena 

CPB4337), both bioluminescent organisms have proved very 

useful in evaluating the toxicity of individual fibrates in 

a previous study (Rosal et al., 2009). For Anabaena CPB4337, we 

also evaluated the nature of the interactions of the three 

fibrates with a wastewater sample from a local STP, which 

already proved very toxic to the cyanobacterium (Rosal et al., 

2009). 


2.1. Materials 

Gemfibrozil (þ99%) and bezafibrate (þ98%) were purchased 

from Sigma–Aldrich. Fenofibric acid was produced from 

fenofibrate (Sigma–Aldrich, þ99% purity) by hydrolysis. 

A suspension of fenofibrate in isopropanol (30 wt.%, 400 mL) 

was refluxed during 4 h with an aqueous sodium hydroxide 

solution (2.0 M, 200 mL). After cooling to less than 70 C, 

a solution of hydrochloric acid (1.0 M, 325 mL) was slowly 

added while keeping the temperature over 60 C. The product 

crystallized after cooling and keeping at room temperature 

during 4 or more h. The product was filtered and rinsed with 

water and dried overnight at 60 C under nitrogen. The purity 

of the product was over 97% checked by HPLC. Solubility of 

acidic drugs in water is strongly pH dependent with few data 

considering this variable. Comerton et al. (2007) reported 

a solubility of 10.9 mg/L of gemfibrozil in water, but we could 

solve over 125 mg/L in 2 mM MOPS (3-[N-morpholino] propanesulfonic 

acid) at pH 6 and higher quantities for the pH at 

which V. fischeri bioassays were performed. In all cases, we 

avoided the use of solvents and the upper limit for the 

concentrations of the studied compounds was their solubility 

in pure water or wastewater at the pH of the bioassay. 

Wastewater samples were collected from the secondary 

clarifier of a STP located in Alcalá de Henares (Madrid) that

eceives domestic wastewater with a minor contribution of 

industrial effluents from facilities located near the city. This 

STP used a conventional activated sludge treatment and has 

been designed for a total capacity of 375,000 equivalent 

inhabitants with a maximum flow rate of 3000 m 3 /h. In 

a recent previous study (Rosal et al., 2009), we found that this 

wastewater was very toxic to Anabaena cells with a wastewater 

dilution as low as 0.11 causing 50% luminescence 

inhibition (wastewater EC 50). 

2.2. Toxicity tests 

Bioassays with the photo-luminescent bacteria Vibrio fischeri 

were carried out according to ISO 11348-3 standard protocol 

(ISO, 2007). This bioassay measures, during the prescribed 

incubation period, the decrease in bioluminescence induced 

in the cell metabolism due to the presence of a toxic 

substance. The bacterial assay used the commercially available 

Biofix Lumi test (Macherey-Nagel, Germany). The bacterial 

reagent is supplied freeze-dried (Vibrio fischeri NRRL-B 

11177) and was reconstituted and incubated at 3 C for 5 min 

before use. The desired pH was set by using NaOH or HCl. The 

analysis media was 0.34 M NaCl (2% w/v) and tests were 

performed at 15 C and the measurements of light were made 

using a luminometer (Optocomp I). The effect of toxicants or 

toxicant mixtures (i.e., fibrates or fibrate combinations) was 

measured as percent inhibition with respect to the light 

emitted under test conditions in the absence of any toxic 

influence. Toxicity values were routinely obtained after 

30 min exposure. Phenol and ZnSO 4 7H 2O have been used as 

toxicity standards and all tests have been replicated to ensure 

reproducibility. 

The bioassays using the recombinant bioluminescent 

cyanobacterium Anabaena CPB4337 were based on the inhibition 

of constitutive luminescence caused by the presence of 

any toxic substance (Rodea-Palomares et al., 2009; Rosal et al., 

2009). Anabaena CPB4337 was routinely grown at 28 C in the 

light, ca. 65 mmol photons m 2 s 1 on a rotary shaker in 50 mL 

AA/8 (Allen and Arnon, 1955) supplemented with nitrate 

(5 mM) in 125 ml Erlenmeyer flasks and 10 mg/mL of neomycin 

sulphate (Nm). Luminescence inhibition-based toxicity assays 

were performed as follows: 160 mL from five to seven serial 

dilutions of each tested toxicant or toxicant mixture (i.e.; 

fibrates or fibrate combinations) plus a control (ddH 2O buffered 

with MOPS at pH 5.8) were disposed in an opaque white 

96-well microtiter plates. 40 mL cells, grown as described, were 

washed twice and resuspended in ddH2O buffered with MOPS 

at pH 5.8 and were added to the microtiter plate wells to reach 

a final cell density at OD750 nm of 0.5. The luminescence of 

each sample was recorded every 5 min for up to 1 h in the 

Centro LB 960 luminometer. Three independent experiments 

with duplicate samples were carried out for all Anabaena 

toxicity assays. CuSO 4 has been used as toxicity standard and 

all tests have been replicated to ensure reproducibility. 

2.3. Experimental design of fibrate combinations 

Solutions of gemfibrozil (Gm), bezafibrate (Bz) and fenofibric 

acid (Fn) prepared as described above were used singly and in 

two (Bz þ Gm; Fn þ Gm; Fn þ Bz) and three (Fn þ Gm þ Bz) 


combinations. Anabaena and Vibrio fischeri cells were treated 

with serial dilutions of each fibrate individually and with 

a fixed constant ratio (1:1), based on the individual EC50 values, 

in their binary and ternary combinations. Five dilutions (serial 

dilution factor ¼ 2) of each fibrate and combination plus 

a control were tested in three independent experiments with 

replicate samples. 

For evaluating the nature of the interaction of fibrates with 

wastewater, binary combinations of each fibrate plus wastewater 

(Fn þ WW; Gm þ WW; Bz þ WW) and a quaternary 

combination of the three fibrates plus wastewater (Fn þ Gm þ 

Bz þ WW) were also prepared and tested for Anabaena 

CPB4337. Anabaena cells were treated with serial dilutions of 

each fibrate and wastewater individually and with a fixed 

constant ratio (1:1), based on the individual EC50 values, in 

their binary and quaternary combinations. Five dilutions 

(serial dilution factor ¼ 2) of each fibrate and wastewater and 

their combinations plus a control were tested in three independent 

experiments with replicate samples. The experimental 

design is shown in Table 1. 

All individual fibrate, wastewater and their combination 

assays were carried out at the same time as recommended by 

Chou (2006) to maximize computational analysis of data. 

2.4. Median-effect and combination index (CI)isobologram 

equations for determining combined 

fibrate interactions 

The results were analyzed using the median-effect/combination 

index (CI)-isobologram equation by Chou (2006) and Chou 

and Talalay (1984) which is based on the median-effect principle 

(mass-action law) (Chou, 1976) that demonstrates that 

there is an univocal relationship between dose and effect 

independently of the number of substrates or products and of 

the mechanism of action or inhibition. This method involved 

plotting the dose-effect curves for each compound and their 

combinations in multiple diluted concentrations by using the 

median-effect equation: 

fa D 

¼ 

fu Dm 

m 

Where D is the dose, Dm is the dose for 50% effect (e.g., 50% 

inhibition of bioluminescence or EC 50), fa is the fraction 

affected by dose D (e.g., 0.75 if cell bioluminescence is inhibited 

by 75%), fu is the unaffected fraction (therefore, fa ¼ 1 fu), 

and m is the coefficient of the sigmoidicity of the dose-effect 

curve: m ¼ 1, m > 1, and m < 1 indicate hyperbolic, sigmoidal, 

and negative sigmoidal dose-effect curve, respectively. 

Therefore, the method takes into account both the potency 

(Dm) and shape (m) parameters. If Eq. (1) is rearranged, then: 

D ¼ Dm½fa=ð1 faÞŠ 1=m 

The Dm and m values for each fibrate are easily determined by 

the median-effect plot: x ¼ log (D) versus y ¼ log ( fa/fu) which 

is based on the logarithmic form of Eq. (1). In the medianeffect 

plot, m is the slope and log (Dm) is the x-intercept. The 

conformity of the data to the median-effect principle can be 

ready manifested by the linear correlation coefficient (r) of the 

data to the logarithmic form of Eq. (1) (Chou, 2006). 

(1) 

(2)

Table 1 – Experimental design for determining toxicological interactions of fenofibric acid [Fn (D)1], gemfibrozil [Gm (D)2], bezafibrate [Bz (D)3] and their binary and ternary 

combinations for Vibrio fischeri and Anabaena CPB4337 bioluminescence tests. 

Pure fibrate experiments Fibrates plus wastewater experiments 

Vibrio fischeri Anabaena CPB4337 Anabaena CPB4337 

Dilutions Single toxicant Single toxicant Single toxicant 

Fn Gm Bz Fn Gm Bz Fn Gm Bz WW** 

(D) 1 (D) 2 (D) 3 (D) 1 (D) 2 (D) 3 (D) 1 (D) 2 (D) 3 (D) 4 

1 

⁄4 (EC50) 0.4 8.75 37.5 2.5 2.5 12.5 

1 

⁄4 (EC50) 2.5 2.5 12.5 0.025 

1 

⁄2 (EC50) 0.8 17.5 75 5 5 25 

1 

⁄2 (EC50) 5 5 25 0.05 

1 (EC50) 1.6 35 150 10 10 50 1 (EC50) 10 10 50 0.1 

2 (EC50) 3.2 70 300 20 20 100 2 (EC50) 20 20 100 0.2 

4 (EC50) 6.4 140 600 40 40 200 4 (EC50) 40 40 200 0.4 

Two toxicant combo Two toxicant combo Two toxicant combo 

(D) 1 þ (D) 2 (1.6:35) (D) 1 þ (D) 2 (1:1) (D) 1 þ (D) 4 (1:0.01) 

1 

⁄4 (EC50) 0.4 8.75 2.5 2.5 

1 

⁄4 (EC50) 2.5 0.025 

1 

⁄2 (EC50) 0.8 17.5 5 5 

1 

⁄2 (EC50) 5 0.05 

1 (EC50) 1.6 35 10 10 1 (EC50) 10 0.1 

2 (EC50) 3.2 70 20 20 2 (EC50) 20 0.2 

4 (EC50) 6.4 140 25* 25* 4 (EC50) 40 0.4 

(D) 1 þ (D) 3 (1.6:150) (D) 1 þ (D) 3 (1:5) (D) 2 þ (D) 4 (1:0.01) 

1 

⁄4 (EC50) 0.4 37.5 2.5 12.5 

1 

⁄4 (EC50) 2.5 0.025 

1 

⁄2 (EC50) 0.8 75 5 25 

1 

⁄2 (EC50) 5 0.05 

1 (EC50) 1.6 150 10 50 1 (EC50) 10 0.1 

2 (EC50) 3.2 300 20 100 2 (EC50) 20 0.2 

4 (EC50) 6.4 600 30* 150* 4 (EC50) 40 0.4 

(D)2 þ (D)3 (35:150) (D)2 þ (D)3 (1:5) (D)3 þ (D)4 (1:0.002) 

1 

⁄4 (EC50) 8.75 37.5 2.5 12.5 

1 

⁄4 (EC50) 12.5 0.025 

1 

⁄2 (EC50) 17.5 75 5 25 

1 

⁄2 (EC50) 25 0.05 

1 (EC50) 35 150 10 50 1 (EC50) 50 0.1 

2 (EC50) 70 300 20 100 2 (EC50) 100 0.2 

4 (EC50) 140 600 40 200 4 (EC50) 200 0.4 

Three toxicant combo Three toxicant combo Four toxicant combo 

(D)1 þ (D)2 þ (D)3 (1.6:35:150) (D)1 þ (D)2 þ (D)3 (1:1:5) (D)1 þ (D)2 þ (D)3 þ (D)4 (1:1:5:0.01) 

1 

⁄4 (EC50) 0.4 8.75 37.5 2.5 2.5 12.5 1/8 (EC50) 1.25 1.25 6.25 0.0125 

1 

⁄2 (EC50) 0.8 17.5 75 5 5 25 

1 

⁄4 (EC50) 2.5 2.5 12.5 0.025 

1 (EC50) 1.6 35 150 10 10 50 

1 

⁄2 (EC50) 5 5 25 0.05 

2 (EC50) 3.2 70 300 20 20 100 1 (EC50) 10 10 50 0.1 

4 (EC50) 6.4 140 600 40 40 200 2 (EC50) 20 20 100 0.2 

For the Anabaena test, the design for the experiment with the wastewater [WW (D4)] sample is also included. The experimental design is based on EC50 ratios as proposed by Chou and Talalay (1984). 

EC 50 is the effective concentration of a toxicant which caused a 50% bioluminescence inhibition. The combination ratio was approximately equal to the EC 50 ratio of the combination components (i.e., 

close to their equipotency ratio). *Upper maximal possible dose due to the solubility limit of fibrates in pure water. **EC 50 for wastewater is the dilution which caused 50% luminescence inhibition. (D) 1, 

(D)2 and (D)3 in mg/L, (D)4 is the dilution of wastewater in ddH2O. 

430 


These parameters were then used to calculate doses of the 

fibrates and their combinations required to produce various 

effect levels according to Eq. (1); for each effect level, combination 

index (CI) values were then calculated according to the 

general combination index equation for n chemical combination 

at x% inhibition (Chou, 2006): 

n ðCIÞx ¼ Xn 

j¼1 

ðDÞj ðDxÞj ¼ Xn 

j¼1 

ðDmÞ j 

ðDxÞ1 n ½DŠj = Pn 1 ½DŠ 

n h 

fax = 1 j 

fax j 

io 1=mj 

where n (CI) x is the combination index for n chemicals (e.g., 

fibrates) at x% inhibition (e.g., bioluminescence inhibition); 

(Dx) 1 n is the sum of the dose of n chemicals that exerts x% 

inhibition in combination, {[Dj]/ Pn 1 ½DŠ} is the proportionality 

of the dose of each of n chemicals that exerts x% inhibition in 

combination; and (Dm)j {( fax)j/[1 ( fax)j]} 1/mj is the dose of each 

drug alone that exerts x% inhibition. From Eq. (3), CI1 indicates synergism, additive effect and antagonism, 

respectively. 

2.5. Analysis of results 

Computer program CompuSyn (Chou and Martin, 2005, 

Compusyn Inc, USA) was used for calculation of dose-effect 

curve parameters, CI values, fa-CI plot (plot representing CI 

(3) 

versus fa, the fraction affected by a particular dose; see Eq. (1)) 

and polygonograms (a polygonal graphic representation 

depicting synergism, additive effect and antagonism for three 

or more drug combinations). Linear regression analyses were 

computed using MINITAB Release 14 for Windows (Minitab 

Inc; USA). 

3. Results 

3.1. Toxicological interactions of fibrate combinations in 

Vibrio fischeri and Anabaena CPB4337 bioluminescence 

tests 

Applying the combination index-isobologram method, we 

evaluated the nature of gemfibrozil (Gm), fenofibric acid (Fn) 

and bezafibrate (Bz) interactions both in Vibrio fischeri and 

Anabaena CPB4337 bioluminescence tests. Table 2 shows the 

dose-effect curve parameters (Dm, m and r) of the three 

fibrates singly and their binary and ternary combinations, as 

well as mean combination index (CI) values of fibrate combinations. 

Dm was the dose required to produce the medianeffect 

(analogous to the EC 50); Dm values for Fn were the 

lowest both, in Vibrio and Anabaena tests, Dm values for Gm 

were in the same range for both Vibrio and Anabaena while Bz 

Table 2 – Dose-effect relationship parameters and mean combination index (CI) values (as a function of fractional inhibition 

of luminescence) of gemfibrozil (Gm), fenofibric acid (Fn), and bezafibrate (Bz) individually and of their binary and ternary 

combinations on Vibrio fischeri and Anabaena CPB4337 bioluminescence tests. 

Drug combo Vibrio fischeri 

Dose-effect parameters CI values 

Dm m r EC10 EC50 EC90 

mg/L (mM) 

Fn 1.45 (4.01) 0.78 0.989 – – – 

Gm 20.58 (82.11) 1.53 0.966 – – – 

Bz 252.07 (696.46) 1.15 0.975 – – – 

Gm þ Bz 78.20 (234.20) 1.54 0.991 1.13 0.13 Add 0.97 0.04 Add 0.86 0.05 Syn 

Fn þ Bz 153.79 (424.93) 1.09 0.981 2.98 0.15 Ant 1.71 0.03 Ant 1.17 0.06 Ant 

Fn þ Gm 9.84 (38.74) 1.15 0.973 0.99 0.17 Add 0.75 0.05 Syn 0.86 0.08 Syn 

Fn þ Gm þ Bz 55.69 (166.69) 1.23 0.993 1.46 0.06 Ant 1.01 0.02 Add 0.99 0.03 Add 

Anabaena CPB4337 

Dose-effect parameters CI values 

Dm m r EC 10 EC 50 EC 90 

mg/L (mM) 


Fn 8.53 (23.62) 0.96 0.971 – – – 

Gm 10.69 (42.67) 0.81 0.959 – – – 

Bz 12.56 (34.70) 1.08 0.990 – – – 

Gm þ Bz 19.17 (56.88) 0.84 0.972 1.06 0.15 Add 1.57 0.06 Ant 2.5 0.22 Ant 

Fn þ Bz 13.92 (38.49) 0.76 0.965 0.55 0.06 Syn 1.19 0.04 Ant 2.59 0.14 Ant 

Fn þ Gm 12.26 (41.45) 0.46 0.955 0.13 0.02 Syn 1.29 0.05 Ant 12.9 2.33 Ant 

Fn þ Gm þ Bz 6.62 (19.45) 0.53 0.960 0.09 0.01 Syn 0.57 0.02 Syn 3.92 0.19 Ant 

The parameters m, Dm and r are the antilog of x-intercept, the slope and the linear correlation coefficient of the median-effect plot, which 

signifies the shape of the dose-effect curve, the potency (EC50), and conformity of the data to the mass-action law, respectively (Chou, 1976; Chou 

and Talalay, 1984; Chou, 2006). Dm and m values are used for calculating the CI values (Eq. (3)); CI 1 indicate synergism (Syn), 

additive effect (Add), and antagonism (Ant), respectively. EC 10,EC 50 and EC 90, are the doses required to inhibit bioluminescence 10, 50 and 90%, 

respectively. Computer software CompuSyn was used for automated calculation and simulation.

432 

Dm values were an order of magnitude higher in the Vibrio test 

(Rosal et al., 2009); m was the Hill coefficient used to determine 

the shape of the dose-response curve, hyperbolic (m ¼ 1), 

sigmoidal (m > 1) or negative sigmoidal (m < 1); also shown in 

the table, linear regression correlation coefficients (r-values) 

of the median-effect plots were >0.95 in all cases, indicating 

the conformity of the data to the median-effect principle 

which qualifies for further studies using this method. 

The Dm and m values for single fibrates and for their 

combination mixtures were used for calculating synergism or 

antagonism based on the CI Eq. (3) (Chou, 2006). Fig. 1 shows 

the fa-CI plot of fibrate interactions both for Vibrio (Fig. 1a) and 

Anabaena tests (Fig. 1b); the fa-CI plot depicts the CI value 

a 

Combination index, CI 

b 


3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

0.0 

0.0 0.2 0.4 0.6 0.8 1.0 

3.0 

2.5 

2.0 

1.5 

1.0 

0.5 

Fraction affected, fa 

0.0 

0.0 0.2 0.4 0.6 0.8 1.0 

Fraction affected, fa 

Fig. 1 – Combination index plot ( fa-CI plot) for a set of three 

fibrate combinations: Fn D Bz (–6–), Bz D Gm (–B–), 

Fn D Gm (–,–) and Fn D Gm D Bz (–;–) for Vibrio fischeri 

test (a) and Anabaena CPB4337 test (b). CI values are plotted 

as a function of the fractional inhibition of 

bioluminescence ( fa) by computer simulation (CompuSyn) 

from fa [ 0.10 to 0.95. CI 1 indicates 

synergism, additive effect and antagonism, respectively. 

At least three independent experiments with two 

replicates were used. The vertical bars indicate 95% 

confidence intervals for CI values based on sequential 

deletion analysis (SDA) (Chou and Martin, 2005). 

Fn [ fenofibric acid, Bz [ bezafibrate and 

Gm [ gemfibrozil. 


versus fa (effect level or fraction of luminescence inhibited by 

a fibrate singly or in combination with respect to the control) 

for two (Fn þ Bz; Fn þ Gm and Bz þ Gm) and three fibrate 

(Fn þ Gm þ Bz) combinations. The fa-CI plot is an effectoriented 

plot that shows the evolution of the kind of interaction 

(synergism, antagonism, additive effect) as a function of 

the level of the effect ( fa) of a particular toxicant on the 

reference organism ( fa, where ECa ¼ fa 100; i.e., 

EC 10 ¼ f10 100). In the Vibrio test (Fig. 1a), the Bz þ Gm and 

Fn þ Gm binary combination showed a slight antagonism at 

very low fa values and slight synergism (Fn þ Gm) or nearly 

additive effects (Bz þ Gm) at the highest fa values, the Fn þ Bz 

combination showed a strong antagonism at low effect levels 

but the antagonism decreased and approached an additive 

kind of interaction at the highest fa levels; the ternary 

combination (Fn þ Gm þ Bz) showed a moderate antagonism 

at low fa values that also turned into a nearly additive effect at 

fa values above 0.4. Correlation analyses were made between 

CI values of the fibrate ternary combination and CI values of 

each of the fibrate binary combinations to determine which 

binary combination interaction was predominant in the 

ternary mixture (Table 3); the highest correlation coefficient 

was found for the Fn þ Bz combination (r ¼ 0.91), suggesting 

that this combination interaction predominated in the three 

fibrate mixture. The fa-CI plot of the Anabaena test (Fig. 1b) 

showed the opposite pattern of interactions as the three 

binary and the ternary combinations showed from slight to 

strong synergism at the lowest fa values that turned into a very 

strong antagonism at fa values over 0.5; the ternary combination 

(Fn þ Gm þ Bz) closely followed the interaction pattern 

of the binary Fn þ Gm combination, this is confirmed by the 

highest correlation coefficient found between the CI values of 

the ternary combination and the CI values of the Fn þ Gm 

combination (r ¼ 0.996) which suggests that in the Anabaena 

test, this particular combination seemed to be the predominant 

in the ternary toxicological interaction. Selected average 

CI values for both Vibrio fischeri and Anabaena CPB4337 tests at 

three representative dose levels (EC10, EC50 and EC90) and the 

combined effects are summarized in Table 2. 

3.2. Toxicological interactions of wastewater and 

fibrate combinations in the Anabaena CPB4337 

bioluminescence test 

In a recent previous study (Rosal et al., 2009), we found that 

a wastewater sample collected from a local STP was very toxic 

to Anabaena cells with a wastewater dilution of 0.11 causing 

50% luminescence inhibition (wastewater EC50). The observed 

toxicity was attributed to the combined toxicities of over thirty 

micropollutants, which included fibrates as well as other 

pharmaceuticals (Rosal et al., 2008). We sought to investigate 

the nature of the interaction between the wastewater (WW) 

and the three fibrates in binary (Fn þ WW; Bz þ WW and 

Gm þ WW) and quaternary (Fn þ Gm þ Bz þ WW) combinations; 

for these experiments, the wastewater itself was 

regarded as a toxicant; the experimental design was analogous 

to the one for the three fibrate interactions and is also shown 

in Table 1. The r-values of the median-effect plots were >0.95 

in all cases, indicating that the data conformed to the medianeffect 

principle (not shown). Fig. 2 shows the fa-CI plot for each

Table 3 – Correlation analyses between CI values of fibrate ternary and fibrate D wastewater quaternary combinations ( y) 

and their binary combinations (x) for Vibrio fischeri and Anabaena CPB4337 tests. 

Test organism Combinations Regression parameters 

of the binary fibrate-wastewater combination and the 

quaternary combination; as can be observed, in a broad range 

of fa values, the binary combinations showed a strong synergism; 

however, at fa values above 0.8, the binary Fn þ WW and 

Bz þ WW combinations approached an additive effect and at 

fa values above 0.95, these two combinations yielded antagonism; 

by contrast, the Gm þ WW combination became even 

more synergistic. The quaternary combination interaction 

showed a strong synergism through a broad range of fa values 

but also turned into slight antagonism at fa values above 0.95, 


closely resembling the pattern of the Fn þ WW and Bz þ WW 

interactions which is confirmed by the highest r value 

(r ¼ 0.999) in the correlation analyses (Table 3), which suggests 

a predominant effect of Fn and Bz in the quaternary 

interaction. 

The computer software CompuSyn (Chou and Martin, 2005) 

displays a type of graphic termed polygonogram, which is 

a semiquantitative method of representing interactions 

between three or more compounds at a determined fa value. 

This graphic allows a simplified visual presentation of the 

overall results. Fig. 3 shows the polygonogram for the three 

fibrates and the wastewater at four fa values; synergism is 

indicated by solid lines and antagonism by broken ones; the 

thickness of the lines indicates the strength of the interaction. 

The polygonogram clearly shows the synergistic interaction of 

wastewater in combination with each of the three fibrates at 

low fa values and the antagonistic interaction that appeared at 

the highest fa value, 0.99, for the Fn þ WW and the Bz þ WW 

combinations. 

The same wastewater sample collected from a local STP 

was proved as responsible of stimulation of the bioluminescence 

activity of Vibrio fischeri to 110–120% of that of the 

control. Moreover, the EC50 values for the fibrates in the 

wastewater were higher than those for fibrates in pure water 

(Rosal et al., 2009). The same trend was observed comparing 

the dose-effect curve parameters (Dm, m and r) for the ternary 

combination (Fn þ Gm þ Bz) of fibrates in ddH 2O and wastewater. 

The dose required to produce the median-effect (Dm)in 

Vibrio fischeri test when (Fn þ Gm þ Bz) were solved in wastewater 

was 131.936 compared to 55.6951 mg/L required when 

ddH2O was employed. CI values could not be calculated for 

Vibrio fischeri due to the fact that the wastewater itself was not 

toxic to this bacterium; synergism or antagonism could not be 

properly estimated (Chou, 2006). 

4. Discussion 

xo m r 

V. fischeri Fn þ Gm þ Bz versus Gm þ Bz 0.614 1.77 0.83 

Fn þ Bz 0.594 0.281 0.91 

Fn þ Gm 0.067 1.40 0.81 

Anabaena CPB4337 Fn þ Gm þ Bz versus Gm þ Bz 5.876 4.39 0.91 

Fn þ Bz 2.716 3.00 0.941 

Fn þ Gm 0.282 0.247 0.996 

WW þ Fn þ Gm þ Bz versus Gm þ Bz 0.079 0.372 0.999 

Fn þ Bz 0.199 0.246 0.998 

Fn þ Gm 0.464 0.017 0.897 

Fn þ WW 0.253 1.31 0.999 

Gm þ WW 2.131 2.41 0.89 

Bz þ WW 0.003 0.865 0.999 

Fn ¼ fenofibric acid, Bz ¼ bezafibrate, Gm ¼ gemfibrozil, WW ¼ wastewater. The parameters of linear regression equations: x0 (value of y when 

x ¼ 0); m (slope) and r (correlation coefficient) with all p-values of 0.001. Analyses were computed using MINITAB Release 14 for Windows. 

2.0 

1.5 

1.0 

0.5 

0.0 

0.0 0.2 0.4 0.6 0.8 1.0 

Fractions affected, fa 

Fig. 2 – Combination index plot ( fa-CI plot) for a set of three 

fibrates and toxic wastewater sample in their binary and 

quaternary combinations: Gn D WW (–6–), Fn D WW (–B–), 

Bz D WW (–,–) and Fn D Gm D Bz D WW (–;–) for the 

Anabaena CPB4337 test. CI values are plotted as a function of 

the fractional inhibition of bioluminescence ( fa) by 

computer simulation (CompuSyn) from fa [ 0.10 to 0.95. 

CI 1 indicates synergism, additive effect and 

antagonism, respectively. At least three independent 

experiments with two replicates were used. The vertical 

bars indicate 95% confidence intervals for CI values based on 

sequential deletion analysis (SDA) (Chou and Martin, 2005). 

Fn [ fenofibric acid, Bz [ bezafibrate, Gm [ gemfibrozil, 

WW [ wastewater. 


The three fibrates that we have used in our study are lipid 

modifying agents that are effective in lowering elevated

434 

Fig. 3 – Polygonograms showing the toxicological interactions of three fibrates and a toxic wastewater sample in their 

binary combinations (Fn D Bz, Bz D Gm, Fn D Gm, Gm D WW, Fn D WW, Bz D WW) as calculated by CompuSyn (Chou and 

Martin, 2005) for Anabaena CPB4337 test at four effect levels: fa [ 0.1 (a), fa [ 0.5 (b), fa [ 0.9 (c) and fa [ 0.99 (d). Solid lines 

indicate synergism, broken lines indicate antagonism. The thickness of the line represents the strength of synergism or 

antagonism. Figure generated by CompuSyn (Chou and Martin, 2005). 

plasma triglycerides and cholesterol in humans (Staels et al., 

1998). These pharmaceuticals are highly used, ubiquitous and 

persistent (Daughton and Ternes, 1999), they are found at ng/L 

to mg/L levels in many STP effluents, surface waters, estuaries 

of rivers and even in sea water (for a review, see Hernando 

et al., 2007). Although non-target organisms; the continuous 

release of these substances into the environment may cause 

acute or chronic toxicity to the aquatic biota. Regarding 

fibrates, in the recent literature there are many reports dealing 

with individual toxicity of different fibrates in a range of 

aquatic organisms from primary producers to consumers; 

a great variability has been found in the sensitivity of the 

different test organisms toward these pharmaceuticals 

(Hernando et al., 2007). However, pharmaceuticals such as 

fibrates do not occur singly in a polluted environment and are 

usually found as mixtures, therefore, for risk assessment 

strategies it is important to know the combined effects of 

pharmaceuticals in non-target organisms (Teuschler, 2007). 

There are two concepts widely used for the prediction of 

mixture toxicity: concentration addition (CA) and independent 

action (IA) (Backhaus et al., 2003; Vighi et al., 2003; 

Backhaus et al., 2004; Junghans et al., 2006). CA is used for 

mixtures whose components act in a similar mode of action 

while IA is based on the idea of dissimilar action, meaning 

that the compounds have different mechanisms of action; 

however, as discussed by Cleuvers (2003) the terms similar/ 


dissimilar action may be misleading. Pharmaceuticals such as 

fibrates may have the same pharmacological mechanism of 

action [i.e., interaction with the binding peroxisome proliferator-activated 

receptor a (PPARa)] in their target organism, 

humans; however, if fibrates released in the aquatic environments 

prove toxic to different non-target organisms, the 

exact mechanism of toxicity (probably different to the pharmacological 

mode of action) should be investigated in depth 

before choosing which approach, CA or IA, to use. In fact, only 

if toxicity is regarded as non-specific at all, the concept of CA 

may be used although it may also have limitations. Cleuvers 

(2003) found that two totally different pharmaceuticals, 

a fibrate and an anti-epileptic drug, followed the concept of CA 

in the Daphnia toxicity test and the concept of IA in an algal 

test; both pharmaceuticals apparently shared the same nonspecific 

toxic mode of action for both organisms; so it 

appeared that the concept of CA or IA did not depend on 

a similar/dissimilar mode of action but on the tested 

organism. The author also discussed that by definition, when 

using CA, substances applied below their individual noneffect 

concentration (NOEC) will contribute to the total effect 

of the mixture while when using IA, substances applied below 

their NOEC will not contribute to the total effect of the 

mixture, meaning that any combination effect will probably 

be higher if the substances follow the concept of CA and this 

may be misleading when considering the terms synergism or

antagonism because as also discussed by Chou (2006), synergism 

or antagonism may occur independently of a similar or 

dissimilar mode of action. In this context, Fent et al. (2006) 

tested mixtures of different kinds of pharmaceuticals 

(including fibrates) that might have estrogenic activity in 

a yeast reporter system; they applied the CA model and found 

that it had severe limitations when the dose-response curves 

of the individual pharmaceuticals were not identical or at low 

effect concentrations. As pharmaceuticals released in the 

environment may have such diverse dose-effect relationships, 

the lack of appropriate prediction suggests limitation of 

the CA mixtures concept. 

To study the nature of the combined fibrate interactions 

(synergism, additive effect, antagonism) for the Vibrio fischeri 

and Anabaena CPB4337 bioluminescence tests, we have followed 

the combination index (CI)-isobologram equation 

method of Chou (2006) and Chou and Talalay (1984); a method 

widely used to study drug interactions in pharmacology. This 

method may be considered a fractional analysis technique for 

drug interactions (Berenbaum, 1981; Bovill, 1998) that is 

independent of the mode of action and considers both the 

potency (EC 50, Dm) and the shape (m) of the dose-effect curve 

for each drug. The method allows prediction of synergism/ 

antagonism at all effect levels ( fa) for a combination of n 

drugs; in contrast with the classical graphical isobologram 

method (Berenbaum, 1981; Bovill, 1998) that cannot be used 

for more than three compounds and have also graphical 

limitations to show all effect levels. By using this method, we 

have been able to determine the nature of interactions for 

a wide range of effect levels of three fibrates in binary and 

ternary combinations in two different bioluminescent organisms. 

However, the nature of these interactions was not 

uniform along the fa levels range in any of the two organisms. 

In Vibrio fischeri, antagonism predominated at low and intermediate 

fa levels but at the highest effect levels, interactions 

became additive or slightly synergistic. In Anabaena, a dual 

synergistic/antagonistic behaviour was observed with synergism 

predominating at fa levels below 0.4–0.5 and strong 

antagonism above these fa values. It is difficult to give an 

explanation to this phenomenon because the combination 

index method only allows quantitative determination of 

synergism or antagonism and the elucidation of the mechanism 

by which synergism or antagonism occurs is a separate 

issue that needs a different kind of approach. However, 

tentatively, antagonism, which could be considered the 

predominant interaction in Vibrio fischeri and Anabaena, might 

be explained by the structural similarity of fibrates which are 

related pharmaceuticals that share a common structural 

motif, a cyclic head and a hydrophobic tail (Rosal et al., 2009); 

at the fa levels where antagonism is found in both organisms, 

fibrates may compete with one another for the same target/ 

receptor sites. The slight synergism found at very high levels 

in Vibrio fischeri could perhaps be explained by the fact that at 

very high concentrations, fibrates may somehow combine to 

increase toxicity by an unspecific way of action that is probably 

not related to their pharmacological mechanism. 

Perhaps, the most puzzling interaction is the observed high 

synergism at very low fa levels in Anabaena; the mechanism of 

such synergistic interaction is not readily apparent. One could 

speculate that these fibrates at very low concentrations could 


involve what Jia et al. (2009) in their extensive review of 

mechanisms of drug combinations call ‘‘facilitating actions’’ 

that means that secondary actions of one drug enhances the 

activity or level of another drug in the mixture or alternatively 

‘‘complementary actions’’ when drugs act at the same target 

at different sites, at overlapping sites or at different targets of 

the same pathway. However, in the literature there are very 

few reports on possible targets of fibrates on the prokaryotic 

cell; English et al. (1994) reported that peroxisome proliferators 

such as fibrates have been shown to induce cytochrome 

P450BM-3 which catalyzes the hydroxylation of fatty 

acids, in Bacillus megaterium. Garbe (2004) reported that 

fibrates induced methyltransferase Rv0560c with a function in 

the biosynthesis of isoprenoid compounds in Mycobacterium 

tuberculosis; Garbe (2004) suggested that both effects may act 

on the plasma membrane, modulating its properties. In 

mitochondria, which have significant features that resemble 

those of prokaryotes, fibrates have been found to inhibit 

respiratory complex I (NDH-1 complex) and to interfere with 

mitochondrial fatty acid oxidation (Scatena et al., 2007). 

Whether fibrates may exert similar effects in Vibrio fischeri and 

Anabaena to those observed in Bacillus or mitochondria needs 

further research. In this context, we have found that, as the fa- 

CI plots show, fibrate interactions do not follow the same 

pattern in both bacteria, this may be due to the different origin 

and position in the food web of Vibrio fischeri, a heterotrophic 

marine prokaryote and Anabaena CPB4337, a recombinant 

strain of an obligate phototrophic freshwater prokaryote; in 

fact, Anabaena presents intracellular photosynthetic 

membranes called thylakoids where several functionally 

distinct NDH-1 complexes have been found with roles both in 

respiration and photosynthesis (Battchikova and Aro, 2007). If 

fibrates are also affecting NDH-1 complexes in Anabaena, their 

effects might be very different to those in Vibrio fischeri; so, 

although we have measured the same toxicity endpoint in 

both bacteria, i.e., luminescence inhibition, the combined 

effects of fibrates seem to depend on the test organism. 

Ince et al. (1999) assessed toxic interactions of heavy 

metals in binary mixtures on Vibrio fischeri and the freshwater 

aquatic plant Lemna minor and found that most binary metal 

mixtures exhibited only antagonistic interactions in the plant 

opposed to fewer antagonistic and some synergistic interactions 

in the heterotrophic bacterium. These authors also 

found that in the bacterium, the nature of the interaction 

(synergism or antagonism) also changed with the effect level 

of the binary metal combinations, although the authors did 

not provide a mechanistic explanation for this variability. 

Cheng and Lu (2002) made a comparison of joint interactions 

of organic toxicants in binary mixtures in Escherichia coli and 

Vibrio fischeri and found that toxicants with the same mechanisms 

of toxicity displayed mostly additive or antagonistic 

interactions in E. coli and Vibrio fischeri; however a synergistic 

interaction was found between glutardialdehyde and butyraldehyde 

in Vibrio. Synergistic effects in both bacteria were 

mostly associated with toxicants with different mechanisms 

of toxicity, although antagonism clearly predominated. They 

also found that for a total of 44 organic binary mixtures, only 

six mixtures resulted in identical type of interaction in both 

bacteria. From our results and those of other authors’ (Ince 

et al., 1999; Cheng and Lu, 2002; Cleuvers, 2003) one may

436 

conclude that previous knowledge of the mechanism of toxic 

action of a compound is not useful enough to predict which 

kind of interactions it will display when combined with other 

toxicants with the same or different toxic mechanism; also, as 

we have shown, the nature of the interaction may depend on 

the effect level of the mixture. In addition, different types of 

organisms will show completely different responses to 

mixtures of potential toxicants. 

We previously found that a local wastewater was very toxic 

for the Anabaena CPB4337 test but non-toxic at all for the Vibrio 

fischeri or Daphnia magna tests. This wastewater is a mixture of 

over thirty micropollutants, mostly pharmaceuticals of 

different therapeutics groups that, besides the fibrates used in 

this study, included antibiotics, analgesics/anti-inflammatories, 

b-blockers, antidepressants, anti-epileptics/psychiatrics, 

ulcer healing compounds, diuretics and bronchodilators; 

personal care products and some priority organic pollutants 

are also present (Rosal et al., 2008). The method of Chou allows 

to combine one drug mixture with another drug mixture and 

determine their interactions; therefore, we studied the nature 

of the interaction of fibrates and wastewater in the Anabaena 

bioluminescence test; interestingly, we found that in a wide 

range of effect levels, the interaction of wastewater and the 

three fibrate combination was synergistic; particularly, at very 

low fa values which means that fibrates are at low concentrations 

and the wastewater is diluted several-fold, the method 

predicted a strong synergism; this may be due, as discussed 

above, to the observed synergistic interactions of fibrates with 

one another as well as interactions with some of the detected 

micropollutants when present at very low concentrations. This 

observed synergism may be environmentally relevant since 

most pharmaceuticals such as fibrates do not usually show 

acute toxicity on non-target organisms when tested at real 

environmental concentrations (Hernando et al., 2007) but in 

a mixture, if they act synergistically, they could prove toxic for 

a test organism even at low concentrations; these results agree 

with those found by Hernando et al. (2004) who reported 

synergistic toxic effects for Daphnia magna test when wastewater 

was spiked with environmental concentrations of 

several pharmaceuticals including fibrates. By contrast, our 

results show that at high fa values ( fa > 0.8), the combined 

interaction of the quaternary fibrates þ wastewater combination, 

the binary Fn þ WW and Bz þ WW combinations 

approached an additive effect and eventually became antagonistic; 

in our previous study, the wastewater itself decreased 

Anabaena bioluminescence by 84% with a lower confidence 

limit of 76% and an upper confidence limit of 91%; when the 

wastewater was spiked with increasing concentrations of each 

fibrate we found that, with the exception of gemfibrozil, the 

EC50 values for the fibrates in the wastewater were higher than 

those for fibrates in pure water; this was attributed either to 

reduced bioavailability or to antagonistic effects of fibrates 

with other chemicals present in the wastewater; although we 

did not use the method of Chou, we obtained similar results to 

the ones we report in this study; that is, at high effect levels 

(>84% luminescence inhibition) the interaction of fibrates with 

wastewater, except the Gm þ WW combination, showed 

antagonism. 

Based on our results, we propose that the combination 

index (CI)-isobologram equation, a method widely used in 


pharmacology both for in vitro and in vivo bioassays, may also 

be applied in environmental toxicology as a general method to 

define interactions of potential toxicants in mixtures in any 

test organism and/or toxicological endpoint of interest and 

could be especially useful for risk assessment strategies that 

take into account the toxicological interactions of substances 

in a mixture. 


We report an environmental application of the combination 

index (CI)-isobologram equation to study the nature of the 

interactions of fibrate combinations in two bioluminescent 

aquatic organisms. The method allowed calculating synergism 

or antagonism of binary and ternary fibrate combinations at all 

effect levels simultaneously; we could also test the method 

with a real wastewater sample in binary and quaternary 

combination with the fibrates, finding that at very low effect 

levels, the fibrates acted synergistically with the wastewater in 

the Anabaena test. The proposed method may be used with 

other test organisms and/or toxicological endpoints and could 

be particularly useful for risk assessment approaches to 

toxicity of complex mixtures. 


The research was funded by the Spanish Ministry of Education 

through grants CTM2005-03080/TECNO and CSD2006-00044 

and the Comunidad de Madrid, grants 0505/AMB-0395 and 

0505/MB/0321. 

references 

Allen, M.B., Arnon, D.I., 1955. Studies on nitrogen-fixing blue 

green algae. I Growth and nitrogen fixation by Anabaena 

cylindrica Lemm. Plant Physiol 30 (4), 366–372. 

Andreozzi, R., Raffaele, M., Nicklas, P., 2003. Pharmaceuticals in 

STP effluents and their solar photodegradation in aquatic 

environment. Chemosphere 50 (10), 1319–1330. 

Backhaus, T., Altenburger, R., Arrhenius, A˚ ., Blanck, H., Faust, M., 

Finizio, A., Gramatica, P., Grote, M., Junghans, M., Meyer, W., 

Pavan, M., Porsbring, T., Scholze, M., Todeschini, R., Vighi, M., 

Walter, H., Horst Grimme, L., 2003. The BEAM-project: 

prediction and assessment of mixture toxicities in the aquatic 

environment. Continental Shelf Res. 23 (17–19), 1757–1769. 

Backhaus, T., Arrhenius, A., Blanck, H., 2004. Toxicity of a mixture 

of dissimilarly acting substances to natural algal 

communities: predictive power and limitations of 

independent action and concentration addition. Environ. Sci. 

Technol 38 (23), 6363–6370. 

Battchikova, N., Aro, E.-M., 2007. Cyanobacterial NDH-1 

complexes: multiplicity in function and subunit composition. 

Physiol. Plant 131 (1), 22–32. 

Berenbaum, M.C., 1981. Criteria for analyzing interactions 

between biologically active agents. Adv. Cancer Res. 35, 

269–335. 

Borgmann, U., Bennie, D.T., Ball, A.L., Palabrica, V., 2007. Effect of 

a mixture of seven pharmaceuticals on Hyalella azteca over 

multiple generations. Chemosphere 66 (7), 1278–1283.

Bovill, J.G., 1998. Analysis of drug interactions. Baillière’s Clin. 

Anaesthesiol 12 (2), 135–168. 

Brain, R.A., Johnson, D.J., Richards, S.M., Hanson, M.L., 

Sanderson, H., Lam, M.W., Young, C., Mabury, S.A., Sibley, P.K., 

Solomon, K.R., 2004. Microcosm evaluation of the effects of an 

eight pharmaceutical mixture to the aquatic macrophytes Lemna 

gibba and Myriophyllum sibiricum. Aquat. Toxicol 70 (1), 23–40. 



treatment plants in Italy. Environ. Sci. Technol 40 (1), 

357–363. 

Chen, C.-Y., Lu, C.-L., 2002. An analysis of the combined effects of 

organic toxicants. Sci. Total Environ 289 (1–3), 123–132. 

Chou, T.C., 1976. Derivation and properties of Michaelis–Menten 

type and Hill type equations for reference ligands. J. Theor. 

Biol. 59 (2), 253–276. 

Chou, T.C., 2006. Theoretical basis, experimental design, and 

computerized simulation of synergism and antagonism in 

drug combination studies. Pharmacol. Rev. 58 (3), 621–681. 

Chou, T.C., Martin, N., 2005. CompuSyn for Drug Combinations: 

PC Software and User’s Guide: A Computer Program for 

Quantification of Synergism and Antagonism in Drug 

Combinations and the Determination of IC 50 and ED 50 and 

LD 50 Values. ComboSyn, Inc., Paramus, NJ. 

Chou, T.C., Talalay, P., 1984. Quantitative analysis of dose-effect 

relationships: the combined effects of multiple drugs or 

enzyme inhibitors. Adv. Enzyme Regul 22, 27–55. 

Christensen, A.M., Faaborg-Andersen, S., Ingerslev, F., Baun, A., 

2007. Mixture and single-substance toxicity of selective 

serotonin reuptake inhibitors toward algae and crustaceans. 

Environ. Toxicol. Chem. 26 (1), 85–91. 

Cleuvers, M., 2003. Aquatic ecotoxicity of pharmaceuticals 

including the assessment of combination effects. Toxicol. Lett. 

142 (3), 185–194. 

Cleuvers, M., 2004. Mixture toxicity of the anti-inflammatory 

drugs diclofenac, ibuprofen, naproxen, and acetylsalicylic 

acid. Ecotoxicol. Environ. Saf 59 (3), 309–315. 

Comerton, A.M., Andrews, R.C., Bagley, D.M., Yang, P., 2007. 

Membrane adsorption of endocrine disrupting compounds 

and pharmaceutically active compounds. J. Memb. Sci. 303 

(1–2), 267–277. 

Crane, M., Watts, C., Boucard, T., 2006. Chronic aquatic 

environmental risks from exposure to human 

pharmaceuticals. Sci. Total Environ 367 (1), 23–41. 



Environ. Health Perspect 107 (6), 907–938. 

English, N., Hughes, V., Wolf, C.R., 1994. Common pathways of 

cytochrome P450 gene regulation by peroxisome proliferators 

and barbiturates in Bacillus megaterium ATCC14581. J. Biol. 

Chem. 269 (43), 26836–26841. 

Fent, K., Escher, C., Caminada, D., 2006. Estrogenic activity of 

pharmaceuticals and pharmaceutical mixtures in a yeast 

reporter gene system. Reprod. Toxicol 22 (2), 175–185. 

Garbe, T.R., 2004. Co-induction of methyltransferase Rv0560c by 

naphthoquinones and fibric acids suggests attenuation of 

isoprenoid quinone action in Mycobacterium tuberculosis. Can. J. 

Microbiol. 50 (10), 771–778. 

Gros, M., Petrovic, M., Barcelo, D., 2007. Wastewater treatment 

plants as a pathway for aquatic contamination by 

pharmaceuticals in the Ebro river basin (northeast Spain). 


Han, G.H., Hur, H.G., Kim, S.D., 2006. Ecotoxicological risk of 

pharmaceuticals from wastewater treatment plants in Korea: 

occurrence and toxicity to Daphnia magna. Environ. Toxicol. 

Chem. 25 (1), 265–271. 

Hernando, M.D., Petrovic, M., Fernández-Alba, A.R., Barceló, D., 

2004. Analysis by liquid chromatography-electrospray 


ionization tandem mass spectrometry and acute toxicity 

evaluation for beta-blockers and lipid-regulating agents in 

wastewater samples. J. Chromatogr. A 1046 (1–2), 133–140. 

Hernando, M.D., Agüera, A., Fernández-Alba, A.R., 2007. LC-MS 

analysis and environmental risk of lipid regulators. Anal. 

Bioanal. Chem. 387 (4), 1269–1285. 

Holoshitz, N., Alsheikh-Ali, A.A., Karas, R.H., 2008. Relative safety 

of gemfibrozil and fenofibrate in the absence of concomitant 

cerivastatin use. Am. J. Cardiol 101 (1), 95–97. 

Ince, N.H., Dirilgen, N., Apikyan, I.G., Tezcanli, G., Üstün, B., 1999. 

Assessment of toxic interactions of heavy metals in binary 

mixtures: a statistical approach. Arch. Environ. Contam. 

Toxicol 36, 365–372. 

International Organization for Standardization, 2007. Water 

quality - Determination of the inhibitory effect of water 

samples on the light emission of Vibrio fischeri (Luminescent 

bacteria test), ISO 11348-3 Revised version, Geneva, 

Switzerland. 

Jia, J., Zhu, F., Ma, X., Cao, Z.W., Li, Y.X., Chen, Y.Z., 2009. 

Mechanisms of drug combinations: interaction and network 

perspectives. Nat. Rev. Drug Discov 8 (2), 111–128. 

Junghans, M., Backhaus, T., Faust, M., Scholze, M., Grimme, L.H., 

2006. Application and validation of approaches for the 

predictive hazard assessment of realistic pesticide mixtures. 

Aquat. Toxicol 76 (2), 93–110. 





Water Res. 43 (2), 363–380. 

Lambropoulou, D.A., Hernando, M.D., Konstantinou, I.K., 

Thurman, E.M., Ferrer, I., Albanis, T.A., Fernandez-Alba, A.R., 

2008. Identification of photocatalytic degradation products of 

bezafibrate in TiO 2 aqueous suspensions by liquid and gas 

chromatography. J. Chromatogr. A 1183 (1–2), 38–48. 

Metcalfe, C.D., Miao, X.S., Koenig, B.G., Struger, J., 2003. 

Distribution of acidic and neutral drugs in surface waters near 

sewage treatment plants in the lower Great Lakes, Canada. 


Miller, D.B., Spence, J.D., 1998. Clinical pharmacokinetics of fibric 

acid derivatives (fibrates). Clin. Pharmacokinet 34 (2), 155–162. 

Pomati, F., Orlandi, C., Clerici, M., Luciani, F., Zuccato, E., 2008. 

Effects and interactions in an environmentally relevant 

mixture of pharmaceuticals. Toxicol. Sci. 102 (1), 129–137. 

Quinn, B., Gagne, F., Blaise, C., 2009. Evaluation of the acute, 

chronic and teratogenic effects of a mixture of eleven 

pharmaceuticals on the cnidarian, Hydra attenuata. Sci. Total 

Environ 407 (3), 1072–1079. 

Rodea-Palomares, I., González-García, C., Leganés, F., Fernández- 

Piñas, F., 2009. Effect of pH, EDTA, and anions on heavy 

metal toxicity toward a bioluminescent cyanobacterial 

bioreporter. Arch. Environ. Contam. Toxicol. doi:10.1007/ 

s00244-008-9280-9. 

Rodríguez, A., Rosal, R., Perdigón, J.A., Mezcua, M., Agüera, A., 

Hernando, M.D., Letón, P., Fernández-Alba, A.R., García- 

Calvo, E., 2008. In: Barceló, D., Petrovic, M. (Eds.), Emerging 

Contaminants from Industrial and Municipal Waste. Springer- 

Verlag, Berlin, pp. 127–175. 

Rosal, R., Rodríguez, A., Perdigón-Melón, J.A., Mezcua, M., 

Hernando, M.D., Letón, P., García-Calvo, E., Agüera, A., 

Fernández-Alba, A.R., 2008. Removal of pharmaceuticals and 

kinetics of mineralization by O 3/H 2O 2 in a biotreated 

municipal wastewater. Water Res. 42 (14), 3719–3728. 

Rosal, R., Rodea-Palomares, I., Boltes, K., Fernández-Piñas, F., 

Leganés, F., Gonzalo, S., Petre, A., 2009. Ecotoxicity assessment 

of lipid regulators in water and biologically treated 

wastewater using three aquatic organism. Environ. Sci. Pollut. 

Res. doi:10.1007/s11356-009-0137-1.

438 

Scatena, R., Bottoni, P., Botta, G., Martorana, G.E., Giardina, B., 

2007. The role of mitochondria in pharmacotoxicology: 

a reevaluation of an old, newly emerging topic. Am. J. Physiol. 

Cell Physiol 293 (1), C12–C21. 

Staels, B., Dallongeville, J., Auwerx, J., Schoonjans, K., Leitersdorf, E., 

Fruchart, J.C., 1998. Mechanism of action of fibrates on lipid and 

lipoprotein metabolism. Circulation 98 (19), 2088–2093. 

Stumpf, M., Ternes, T.A., Wilken, R.D., Rodrigues, S.V., Baumann, W., 

1999. Polar drug residues in sewage and natural waters in the 

state of Rio de Janeiro. Brazil. Sci. Total Environ 225 (1–2), 135–141. 


Teuschler, L.K., 2007. Deciding which chemical mixtures risk 

assessment methods work best for what mixtures. Toxicol. 

Appl. Pharmacol 223 (2), 139–147. 

Vighi, M., Altenburger, R., Arrhenius, A., Backhaus, T., 

Bodeker,W.,Blanck,H.,Consolaro,F.,Faust,M., 

Finizio, A., Froehner, K., Gramatica, P., Grimme, L.H., 

Gronvall, F., Hamer, V., Scholze, M., Walter, H., 2003. 

Water quality objectives for mixtures of toxic chemicals: 

problems and perspectives. Ecotoxicol. Environ. Saf 54 (2), 

139–150.

Toxication or detoxication? In vivo toxicity assessment of 

ozonation as advanced wastewater treatment with the 

rainbow trout 

Daniel Stalter a, *, Axel Magdeburg a , Mirco Weil b , Thomas Knacker b ,Jörg Oehlmann a 

a 

Goethe University Frankfurt am Main, Biological Sciences Division, Department of Aquatic Ecotoxicology, Siesmayerstrasse 70, 60054 

Frankfurt, Hessen, Germany 

b 

ECT Oekotoxikologie GmbH, Böttgerstrasse 2–4, 65439 Flörsheim, Germany 

article info 




16 July 2009 



Keywords: 

Sewage treatment 


Oxidation byproducts 

Vitellogenin 

Emerging contaminants 

Advanced oxidation process 

Rainbow trout 

Fish early life stage toxicity test 


abstract 

Wastewater (WW) is one of the major sources of micropollutants 

introduced into the aquatic environment 

(Schwarzenbach et al., 2006). The large spectrum of pharmaceuticals 

and personal care products (PPCPs) occurring in 





E-mail address: stalter@bio.uni-frankfurt.de (D. Stalter). 


doi:10.1016/j.watres.2009.07.025 

Ozonation as advanced wastewater treatment method is an effective technique for 

micropollutant removal. However, the application of this method carries the inherent 

danger to produce toxic oxidation byproducts. For an ecotoxicological assessment 

conventionally treated wastewater, wastewater after ozonation and ozonated wastewater 

after sand filtration were evaluated in parallel at an operating treatment plant via the fish 

early life stage toxicity test (FELST) using rainbow trout (Oncorhynchus mykiss). 

The FELST revealed a considerable developmental retardation of test organisms 

exposed to ozonated WW. This was accompanied by a significant decrease in body weight 

and length compared to reference water, to the conventionally treated WW and to the 

ozonated water after sand filtration. Hence sand filtration obviously prevents from adverse 

ecotoxicological effects of ozonation. 

An additional test with yolk-sac larvae resulted in a significant reduction of vitellogenin 

levels in fish exposed to ozonated wastewater compared to fish reared in conventionally 

treated wastewater. This demonstrates the effective removal of estrogenic activity by 

ozonation. 

Adverse ozonation effects may have been a result of the conversion of chemicals into 

more toxic metabolites. However, sand filtration reduced toxication effects indicating that 

these oxidation byproducts are readily degradable or adsorbable. The results indicate that 

in any case ozonation should not be applied without subsequent post treatment appropriate 

for oxidation byproducts removal (e.g. sand filtration). 


surface waters (Daughton and Ternes, 1999), particularly with 

regard to mixture toxicity, may pose a potential threat to 

aquatic wildlife. But also single substances may have the 

capability to endanger the ecosystem as for example diclofenac 

can lead to an impairment of the general health condition 

of rainbow trout at environmentally relevant concentrations

440 

Nomenclature 

AOC assimilable organic carbon 

C control 

COD chemical oxygen demand 

DOC dissolved organic carbon 

EE2 17a-ethinylestradiol 

ELISA enzyme linked immunosorbent assay 

EQS environmental quality standard 

FELST fish early life stage toxicity test 

FS wastewater after final sedimentation 

GJIC gap junctional intercellular communication 

MS222 tricaine methanesulfonate 

NH4–N ammonium nitrogen 

(Schwaiger et al., 2004; Triebskorn et al., 2004). 17a-ethinylestradiol 

(EE2) is considered to be the most potent estrogenic 

active compound for fish with the lowest observed effect 

concentration of 0.1 ng/L for vitellogenesis in rainbow trout 

(Purdom et al., 1994). Moreover it leads to a reduced egg fertilisation 

success and skewed sex ratio toward females of fat 

head minnows at a LOEC of 0.32 ng/L (Parrott and Blunt, 2005). 

Furthermore, industrial surfactants like alkylphenolic ethoxylates 

are likely to be responsible for feminization effects in 

rainbow trout (Jobling et al., 1996) and complex mixtures of 

estrogenic chemicals present in the environment have been 

shown to act additively (Brian et al., 2005). The concentrations 

of such compounds in WW often exceed effect concentrations 

(Ternes et al., 1999; Ying et al., 2002; Tixier et al., 2003) and 

thus even treated WW contains high amounts of estrogenic 

substances that impair the endocrine system of exposed fish 

(Larsson et al., 1999; Rodgers-Gray et al., 2001) resulting in 

feminization effects of males and consequently reduced 

fertility of field populations (Jobling et al., 2002). 

Besides, in the European Union the reduction of the 

contamination of surface waters with hazardous substances 

is defined by the Water Framework Directive requiring a ‘‘good 

status’’ for all coastal and inland waters until 2015 (European 

Commission, 2000). One tool is the implementation of environmental 

quality standards (EQSs) for mono substance 

pollutants exhibiting a significant risk to the aquatic 

ecosystem. The discharge of such compounds, classified as 

priority substances, is envisaged to be progressively reduced 

by 2025 or 5 years after inclusion in the list for priority 

substances, respectively. 

However, till now so called ‘‘emerging contaminants’’ are 

only marginally addressed and PPCPs are not included in this 

list. But the latter has to be reviewed at least every 4 years and 

provisions have been made to add several hazardous 

emerging contaminants. Amongst others the inclusion of 

diclofenac, EE2 and carbamazepine is discussed and additionally 

recommended by the European Parliament since 2007 

(Jensen, 2007). Moreover EQSs meanwhile have been proposed 

for several PPCPs based on reliable ecotoxicity data (e.g. 

diclofenac: 0.1 mg/L, EE2: 0.03 ng/L, carbamazepine: 0.5 mg/L; 

Jahnel et al., 2006; Moltmann et al., 2007). Unfortunately, some 

PPCPs exceed the EQSs in surface waters of some urbanised 


NO 3–N nitrate nitrogen 

O wastewater after ozonation 

OECD Organisation for Economic Cooperation and 

Development 

OS wastewater after ozonation and sand filtration 

P phosphate 

PAH polycyclic aromatic hydrocarbons 

PAC powdered activated carbon 

PPCPs pharmaceuticals and personal care products 

rcf relative centrifugal force 

SPM suspended particulate matter 

VTG vitellogenin 

WW wastewater 

WWTP wastewater treatment plant 

regions. That is, inter alia, a result of the fact that conventional 

WW treatment plants are not designed for appropriate 

removal of PPCPs (Daughton and Ternes, 1999). Nevertheless 

in many countries surface waters serve as raw water resources 

for drinking water supply. Due to the public’s demand for safe 

drinking water and ecosystem health advanced WW treatment 

methods are required to allow for sufficient removal of 

micropollutants from sewage treatment effluents. Therefore, 

end of pipe techniques like activated carbon filtration and 

ozonation of WW might be crucial to achieve regulative 

requirements in a medium-term perspective. 

Huber et al. (2005) demonstrated that ozonation of 

secondary effluent is an effective tool for the removal of 

a wide range of pharmaceuticals among them diclofenac, 

carbamazepine and the estrogens estradiol, estrone and EE2, 

with degradation rates of more than 90% for ozone doses of 

2 mg/L. Merely iodinated X-ray contrast media and a few 

acidic pharmaceuticals were oxidized only partially. Nakada 

et al. (2007) confirmed the effective elimination of 22 pharmaceuticals 

during ozonation in an operating treatment 

plant. A further advantage of the ozonation is the sanitizing 

property of the method (Tyrrell et al., 1995). The effective 

removal of micropollutants and pathogens indicates that 

ozonation is a suitable technique for advanced WW treatment 

to reduce the contamination of the aquatic environment. 

Actually a reduced toxicity of a mixture of six pharmaceuticals 

treated by ozonation towards algae and rotifer was 

demonstrated by Andreozzi et al. (2004). Reduced estrogenic 

activity of an EE2-containing WW after ozonation was shown 

by Huber et al. (2004) and ozone treatment of WW significantly 

decreased the toxic potency of urban pollutants to the freshwater 

mussel Elliptio complanata (Gagne et al., 2007). However, 

ozone treatment typically transforms chemical compounds 

but does not mineralize them entirely. Consequently, not only 

the concentrations of mother compounds but also the inclusion 

of transformation products and their mixtures should be 

assessed for toxicity, too. 

For a risk-benefit analysis an extensive ecotoxicological 

evaluation of ozonated WW is essential. In vitro bioassays 

covering different modes of toxic action (e.g. estrogenicity, 

acetylcholine esterase activity or non-specific baseline 

toxicity) and performed with enriched water samples are

Table 1 – Wastewater quality summary (mg/L). 

SPM COD P NH4–N NO3–N pH 

Final sedimentation 

Median 4.8 17 0.19 0.1 8.3 

10% percentile 3.6 15 0.17 0.04 5.4 8.12 

90% percentile 8.96 23.8 0.22 0.27 10.1 8.4 

n 37 23 37 37 2 23 

Ozonation and sand filtration 

Median 2 15 0.17 0.04 9.8 8.3 

10% percentile 1.6 12.2 0.14 0.04 7.28 8.02 

90% percentile 2.4 19.8 0.19 0.15 12.48 8.4 

n 37 23 37 37 37 23 

suitable tools for toxicity characterization of WW as shown by 

Escher et al. (2008). These methods were shown to be highly 

sensitive even for identification and characterization of low 

toxic water samples. Nevertheless, as typical screening tools 

they are not designed to replace chronic in vivo tests of whole 

effluent samples. To allow for a more comprehensive and 

integrative assessment of the potentially hazardous impact of 

WW ozonation we applied the fish early life stage toxicity test 

(FELST) with rainbow trout (Oncorhynchus mykiss). 


Ozonated and conventionally treated WW were tested in 

parallel on site at the wastewater treatment plant (WWTP) 

Wüeri (Regensdorf, Switzerland) with the fish early life stage 

toxicity test (FELST) in a flow-through system. 

2.1. Characterization of the wastewater treatment plant 

The municipal WWTP Wüeri operates experimentally with 

a full scale ozonation reactor after final sedimentation and 

with a sand filtration step after the ozone reactor. Table 1 

shows the WW quality parameters. The water parameters 

after the final sedimentation and after the ozone reactor are 

on the same level and therefore exemplarily shown for final 

sedimentation water. Low ammonium and phosphate 

concentrations indicate that the treatment plant is already 

working well. The dissolved organic carbon (DOC) ranged from 

5.4 to 5.9 mg/L and the pH was nearly constant in all test 

waters. The WWTP serves for a population equivalent of 

25,000. The median discharge in the experimental period was 

6190 m 3 /day (10th percentile: 4430 m 3 /day, 90th percentile: 

10,500, n ¼ 109) and the applied ozone concentration was in 

a range between 0.4 and 1 mg O 3/mg DOC. 


2.2. Experimental setup 

WW from three different sampling points of serial treatment 

steps was tested (Fig. 1): after the final sedimentation (FS), 

after the ozone reactor (O) and after additional sand filtration 

(OS). 

Test waters were passively transported through high-grade 

steel pipes to aerated high-grade steel reservoir tanks. The 

retention time in the pipes was adjusted to be at least 45 min 

to avoid ozone residuals reaching the exposure vessels. 

Indeed no ozone was detected by indigo blue method (Bader 

and Hoigne, 1981) in the reservoir tank during maximum 

required flow-through and during maximum applied ozone 

concentration (1 mg O3/mg DOC). From reservoir tanks test 

waters were transported through polytetrafluoroethylene 

tubes via a peristaltic pump (IPC24, Ismatec, Wertheim– 

Mondfeld, Germany) to the exposure vessels each equipped 

with a passive discharge device and tempered using 

a temperature-controlled water bath. Constant temperature 

conditions in the water bath were achieved with a flowthrough 

cooling system (Van der Heijden, Dörentrup, 

Germany). The flow-through rates in the exposure vessels 

ranged from 11 mL per minute up to 44 mL per minute (2–8 fold 

water exchange per day in the exposure vessels) depending on 

the fish size to match the loading rate criteria (OECD, 1992b). 

Reconstituted water according to OECD guideline 203 (1992a) 

was used as control water (C). 

The FELST was performed with the rainbow trout (Oncorhynchus 

mykiss) according to OECD guideline 210 (1992b) with 

a constant water temperature of 10 2 C as well as darkness 

for embryo development and 12 2 C with 12/12 h light/dark 

photoperiod post hatch. Sixty newly fertilized eggs per replicate 

were exposed to the test waters for 65 days in high-grade 

steel 10 L tanks. One test series was performed with unfiltered 

WW and a second with membrane filtered WW (pore size: 

0.4 mm, Kubota Corp., Osaka, Japan) to minimize microbial 

impacts. Macromolecules and organic compounds were not 

retained. However, it has to be considered that membrane 

filtration additionally removes suspended particulate matter 

and consequently all particle bound pollutants. The filter 

membranes were placed in the reservoir vessels. 

A third test was performed with yolk-sac fry (5 days post 

hatch, 30 larvae per exposure vessel) and non-filtered test 

waters because we postulated a reduced sensitivity to pathogen 

contamination of the larvae compared to the egg stage. 

The test duration was 64 days. All tests were performed with 

undiluted WW to increase the probability to detect differences 

between the treatment groups. With the beginning of swim up 

(the swim up process marks a developmental transition from 

larval stage to juvenile fish stage and is characterized with the 

Fig. 1 – Sampling points at the wastewater treatment plant. Abbreviations: FS, after final sedimentation; O, after the ozone 

reactor; OS, after sand filtration.

442 

beginning of exogenous ingestion) the fish were fed four times 

per day (trout starter, 4% body weight per day). 

The tests were run with four (C) and three (FS, O, OS) 

replicates in the first FELST with unfiltered WW and two (C, 

OS) and three replicates (FS, O) per test water in the remaining 

two tests. The latter were performed in parallel and therefore 

it was not possible to use consistently three replicates per 

WW-group due to limited space capacities. In each test replicates 

were placed randomized in the water bath. Observations 

on egg coagulation, hatching, mortality, swim up, malformation 

and abnormal behaviour were recorded daily. Fish were 

humanely killed by MS222 (tricaine methanesulfonate, Sigma– 

Aldrich, St. Louis, USA) overdose. Individual fish were blotted 

dry, weight and body length were determined. Afterwards fish 

were frozen in liquid nitrogen and stored at 80 C until 

vitellogenin detection in whole body homogenates. 

2.3. Vitellogenin detection 

Whole body homogenates of 11 fish (from the third test with 

yolk-sac fry) per replicate were prepared as described by 

Holbech et al. (2006) with slight modifications. Aliquots of 0.3 g 

frozen fish, excised between head and pectoral fin, were 

mixed with 10 fold of the body weight of homogenisation 

buffer (50 mM Tris–HCl pH 7.4; 1% protease inhibitor cocktail 

(P 8340, Sigma–Aldrich, St. Louis, USA)) and homogenated 

with a dispersing apparatus (T18 basic Ultra–Turrax Ò , IKA, 

Staufen, Germany). The homogenate was centrifuged for 

30 min at 20,000 rcf and the supernatant was used for vitellogenin 

analysis. Vitellogenin (VTG) was detected with 

a rainbow trout vitellogenin ELISA test kit (Biosense, Bergen, 

Norway) using a 1:20 dilution. 

2.4. Statistical analysis 

Complete statistical analyses (Kruskal–Wallis with Dunn’s 

post test, Fisher’s exact test, non-linear regression with variable 

slope) were performed using GraphPad Prism version 5.0 

for windows (GraphPad software, San Diego, CA, USA). All 

given error values indicate the standard error (SE) and in all 

figures error bars display the SE. Kruskal–Wallis with Dunn’s 

post test was chosen to test on significant differences because 

data were not normally distributed in all cases. For quantal 

data Fisher’s exact test was applied (mortality, swim up, 

coagulation). When toxicity endpoints could be analysed 

individually for each test animal (biomass, body length, 

vitellogenin), the collected data are presented and statistically 

evaluated on a per specimen basis. 

3. Results 

3.1. FELST with unfiltered wastewater 

Unfiltered WW caused an increased coagulation rate of the 

exposed eggs in the FELST (Fig. 2). The eggs exposed to WW 

after final sedimentation (FS) were completely coagulated 

after 18 days with a 50% coagulation time of 12.1 days. The 

coagulation of the eggs exposed to WW after the ozone reactor 

(O) was considerably delayed compared to FS. However the 


Fig. 2 – Cumulative coagulation of Oncorhynchus mykiss 

eggs (mean values ± SE) exposed to differently treated 

wastewaters. Abbreviations: C, control water; FS, final 

sedimentation; O, ozonation; OS, ozonation and sand 

filtration; SE, standard error. Time scale is logarithmized. 

After 40 days all treatment groups differ significantly from 

each other (Fisher’s exact test: p < 0.001; n [ 240 (C), 180 

(FS, O, OS)). 

coagulation after 40 days achieves still an average rate of 

87.2 4.8% with a 50% coagulation time of 17.6 days. The 

lowest coagulation rate in the WW treatments occurred after 

sand filtration (OS; 64.4 4.0% after 40 days; 50% coagulation 

time: 25.8 days). After 10–15 days exposure, fungus mycelia 

(first appearing in the FS vessels) were observed in all WW 

exposure vessels. Mycelia were found on and between eggs as 

well as vorticellas on the eggs whereas the reference water (C) 

remained observably free from mycelia and vorticellas. 

Hatching success in the control group (53.4 5.2%) did not 

meet validity criteria (>66%; OECD guideline 210). 

3.2. FELST with membrane filtered wastewater 

To exclude microbial impairment the second FELST was performed 

with membrane filtered WW and eggs were obtained 

from another fish hatchery. 

All test vessels remained free from microbial contamination 

throughout the test duration. The egg coagulation rates in 

the WW treatment groups of this experimental series were 

reduced compared to the first FELST with a maximum of 25% 

(O, OS) and only 20.1 1.1% after FS (Fig. 3A). Egg coagulation 

was significantly increased and hatching success significantly 

decreased in the WW treatment groups compared to the 

control ( p < 0.05, Fisher’s exact). The coagulation rates were 

slightly but not significantly increased in O and OS compared 

to FS (Fig. 3A). The hatching progress was slightly delayed in O 

compared to FS and OS but hatching success achieved at least 

75% in all treatments (Fig. 3B). The control group met the 

validity criteria according to OECD guideline 210 (egg coagulation: 

10.0 1.7%, hatching success: 90.0 1.7%).

Fig. 3 – Cumulative coagulation of eggs (A) and cumulative hatching of larvae (B) from Oncorhynchus mykiss (mean 

values ± SE) exposed to differently treated and membrane filtered wastewaters. Abbreviations: C, control water; FS, final 

sedimentation; O, ozonation; OS, ozonation and sand filtration; SE, standard error. Time scale is logarithmized. After 40 and 

36 days, respectively all wastewater treatment groups are significantly different from the control (Fisher’s exact test: 

p < 0.01–0.05; n [ 120 (C, OS), 180 (FS, O)). 

Fig. 4 shows the cumulative swim up of hatched fish 

beginning after 45 days of exposure. The swim up is considerably 

delayed in all WW treatment groups compared to the 

control. This effect is most notable in O, even if compared to 

Fig. 4 – Cumulative swim up of Oncorhynchus mykiss larvae 

(mean values ± SE) exposed to differently treated and 

membrane filtered wastewaters. Abbreviations: C, control 

water; FS, final sedimentation; O, ozonation; OS, ozonation 

and sand filtration; SE, standard error. Time scale is 

logarithmized. After 64 days the O group differs 

significantly from the other treatment groups (Fisher’s 

exact test: p < 0.001–0.01; n [ 95 (C), 116 (FS), 103 (O), 75 

(OS)). 


FS and OS. At the end of the experiment only 83.3 1.6% of the 

fish swam up in O while 100% swam up in C and FS and 

97.6 2.4% in OS. Hereby the swim up success after 64 days is 

significantly decreased in O compared to the other treatment 

groups ( p < 0.01, Fisher’s exact). The 50% swim up time in the 

control is 47.4 1.0 days, which is only slightly increased in FS 

and OS (50.2 1.0; 49.3 1.0) but obviously increased in O 

(57.5 1.04). 

The biomass as well as the body length of fish is significantly 

decreased ( p < 0.001, Kruskal–Wallis with Dunn’s post 

test) in all WW treatments compared to the control (Fig. 5). 

Both endpoints in fish exposed to O are furthermore significantly 

decreased compared to the FS group ( p < 0.001) and 

significantly decreased compared to OS ( p < 0.05). 

Generally the mortality is comparatively low in all WW 

treatment groups (Fig. 6), as they are fulfilling the validity 

criteria for controls according to OECD guideline 210 (survival 

after hatch 70%). Nevertheless the mortality in the WW 

treatment groups is slightly increased compared to C. The 

highest and significantly increased mortality rate was detected 

in the ozonated water (24.0 4.0%, p < 0.05, Fisher’s exact). 

3.3. Fish test starting with yolk-sac fry & vitellogenesis 

The fish test starting with the yolk-sac stage revealed no 

statistically significant differences in development between 

WW treatments and the control. The biomass exhibited no 

major deviations between exposure groups (

444 

Fig. 5 – Biomass (A) and body length (B) in percentage relative to the control (0.28 ± 0.01 g and 34.1 ± 0.2 mm, respectively) of 

Oncorhynchus mykiss (mean values ± SE) exposed to differently treated and membrane filtered wastewaters. Abbreviations: 

C, control water; FS, final sedimentation; O, ozonation; OS, ozonation and sand filtration; SE, standard error. Significant 

differences to the control are indicated with white asterisks, between treatments with black asterisks (Kruskal–Wallis with 

Dunn’s post test: +, p < 0.05; +++, p < 0.001; n [ 73–114). 

Kruskal–Wallis with Dunn’s post test) compared to fish maintained 

in reference water (11.0 6.6 ng/mL) whereas it was 

significantly decreased in the O and OS groups (4.6 1.9 and 

4.8 1.9 ng/mL, respectively; p 0.01) compared to FS (Fig. 7). 

4. Discussion 

4.1. FELST with unfiltered wastewater 

Unfiltered WW led to an increased coagulation rate of the 

exposed eggs in the FELST (Fig. 2). High coagulation rates 

Fig. 6 – Mortality of Oncorhynchus mykiss post hatch (mean 

values ± SE) exposed to differently treated and membrane 

filtered wastewaters. Abbreviations: C, control water; FS, 

final sedimentation; O, ozonation; OS, ozonation and sand 

filtration; SE, standard error. Significant differences to 

control are indicated with white asterisks (Fisher’s exact 

test: +, p < 0.05; n [ 108 (C), 143 (FS), 135 (O), 90 (OS)). 


presumably are a result of microbial contamination as the 

reference water (C) remained free from mycelia and vorticellas 

whereas in the WW treatment groups fungi mycelia and 

vorticellas were found after 10–15 days exposure. However, 

the hatching success in the control (53.4 5.2%) did not 

comply with the validity criteria (>66%) according to OECD 

guideline (1992b) indicating that egg quality was not sufficient 

to run a valid test. The disinfectant effect of ozonation (Tyrrell 

et al., 1995) is likely to be responsible for the delayed egg 

Fig. 7 – Whole body vitellogenin concentration of 

Oncorhynchus mykiss (mean values ± SE) after 60 days 

exposure to differently treated wastewaters starting with 

the yolk-sac stage. Abbreviations: FS, final sedimentation; 

O, ozonation; OS, ozonation and sand filtration; C, control 

water; SE, standard error. Significant differences to control 

are indicated with white asterisks, between treatments 

with black asterisks (Kruskal–Wallis with Dunn’s post test: 

+, p < 0.05; ++, p < 0.01; n [ 22 (C, OS) – 33 (FS, O)).

coagulation in the O group compared to FS. Ozonation is also 

known to produce high amounts of assimilable organic 

carbons (AOC) which may have allowed a fast regrowth of 

microorganisms (Huang et al., 2005), resulting in still high 

coagulation rates in the O group. Sand filtration reduces the 

amount of suspended particulate matter (SPM) and of AOC 

(Wang and Summers, 1996). This may have reduced microbial 

development and resulting coagulation rates. 

4.2. FELST with membrane filtered wastewater 

Exposure to membrane filtered WW revealed a distinct delay 

of the swim up process in the O group compared to FS and 

OS (Fig. 4) accompanied by a significant decrease in body 

weight and body length (Fig. 5). The reduced biomass and 

body length after ozonation is most likely a result of the 

delayed development because fish start exogenous ingestion 

at the onset of the swim up process. Consequently trout that 

swim up later may suffer from developmental disadvantages. 

Oxidative byproducts in ozonated WW may have impeded 

embryonic and/or larval development of fish. Possibly 

because sand filtration is able to remove ozonation metabolites 

(e.g. aldehydes, glyoxal, AOC; Wang and Summers, 1996) 

the effect in OS was reduced. However, no single compound 

could be clearly identified for the retardation effect. Possibly 

the sum of aldehydes (e.g. formaldehyde, glyoxal, acetaldehyde), 

carboxylic acids (e.g. formate), ketones and brominated 

organic compounds formed due to ozonation (Huang 

et al., 2005; Wert et al., 2007) caused these effects. Unfortunately, 

no chronic toxicity data for juvenile rainbow trout are 

available in literature for these compounds. Nevertheless, 

Wang and Summers (1996) were able to demonstrate that 

sand filtration efficiently removes aldehydes produced 

during ozone application, supporting this assumption. 

Besides, the formation of more complex organic metabolites 

evoked by ozonation may result in an increased toxicity 

compared to chemical precursors as it has already been 

documented for polycyclic aromatic hydrocarbons (PAHs). In 

the experiments of Luster-Teasley et al. (2002, 2005) chrysene 

and pyrene and byproducts as a result of ozonation were 

examined for their ability to disrupt gap junctional intercellular 

communication (GJIC), an indicator for tumor 

promoting properties. Among the transformation products, 

aldehydic compounds exhibited an increased toxicity 

compared to the precursor substance while their carboxylic 

structure analogue showed no GJIC disrupting activity. For 

the antiepileptic drug carbamazepine McDowell et al. (2005) 

identified three new oxidation products of unknown toxicity 

after ozone treatment. Furthermore, a possible increase in 

mutagenicity after ozonation of WW, as observed by Monarca 

et al. (2000) with the Ames test, verifies the potential of 

ozonation to produce toxic oxidation byproducts. In this 

context conversion of the widely used fungicide tolylfluanide 

into the carcinogen N-nitrosodimethylamine during ozonation 

was discovered by Schmidt and Brauch (2008). Based on 

these results it is likely that ozonation of WW leads to an 

increased number of unknown and potentially toxic metabolites 

depending on the composition of WW at the outset and 

the post treatment (e.g. sand filtration). Petala et al. (2006) 

demonstrated with the Vibrio fischeri bioluminescence test, 


that toxicity originating from ozonated WW decreases with 

increasing storage time, indicating a rapid decomposition of 

toxic metabolites. 

In spite of the developmental retardation and the reduced 

biomass no significant differences in mortality rates were 

observed between the WW treatment groups (Fig. 6). Mortality 

was highest directly after ozonation (O, 24.0 4.0%). Considering 

the developmental retardation and the decreased 

biomass of fish exposed to ozonated water the increased 

mortality of O group specimens compared to FS and OS is 

noticeable and, although not significant, potentially a result of 

a reduced fitness of the fish in the ozonated water. 

It might be assumed that the retarded development is 

a consequence of an unspecific and general impairment of the 

fish’s health condition. This may increase their sensitivity 

towards environmental and anthropogenic stressors, thus 

leading to an increased mortality. Furthermore, under field 

conditions the delayed development possibly increases the 

risk for the fish to fall prey to predators since before swim up 

the larvae rest on the bottom, not capable to abscond effectively. 

Based on the assumption that the compounds causing 

adverse effects after ozonation are readily degradable it is 

possible that the detoxication after the ozone reactor will 

occur in the river as well. However, the discharge of ozonated 

WW without sand filtration would bear the risk to endanger 

fish populations in a considerable range of the receiving river, 

depending on WW load and flow velocity. Therefore, ozonation 

should always be followed by a sand filtration and further 

studies should focus on whether sand filtration is capable to 

remove oxidation byproducts sufficiently. 

4.3. Fish test starting with yolk-sac fry & vitellogenesis 

The fish test with non-filtered WW starting with the yolk-sac 

stage resulted in negligible and insignificant deviations of 

development and biomass between treatment groups. Ozonation 

had no effect on these toxicity endpoints. These results 

indicate that yolk-sac fry 5 days post hatch are less sensitive 

to ozonation metabolites, compared to fish embryos and 

newly hatched fish, respectively. Nevertheless, the amount of 

suspended particulate matter in the exposure vessels and in 

the Teflon tubes was accompanied by an increased biofilm 

formation. This may have contributed to an increased detoxication 

of the ozonated WW due to (bio-) degradation of 

oxidation byproducts. 

The significant increase of vitellogenin content in fish 

exposed to FS compared to the reference water indicates an 

environmentally relevant contamination of the WW with 

estrogenic active compounds (Fig. 7). After ozonation the VTG 

content decreased even below the control level. The formation 

of antiestrogenic compounds is not likely because the VTG 

decrease compared to control level is not significant and 

phenols, as an important functional group interacting with 

the estrogen receptor (Nishihara et al., 2000), are particularly 

susceptible to ozone attack (von Gunten, 2003). These results 

confirm the high efficiency of ozonation to eliminate estrogenic 

contamination in WW as already shown by Huber et al. 

(2004) with an in vitro test system. Based on these results it 

can be assumed that ozonation is well suited to reduce the 

estrogenic burden of WW below environmental relevance.

446 

Regarding ecosystem health it should be considered that 

estrogen and xenoestrogen concentrations detected in the 

environment have been shown to impair the sustainability of 

wild fish populations (Kidd et al., 2007), arguing for the 

establishment of an end of pipe technique for estrogen 

removal to protect fish populations in surface waters receiving 

high amounts of WW. 

4.4. Ozonation vs. alternatives 

Currently there is a controversy regarding the appropriate 

advanced WW treatment techniques to reduce contamination 

of the aquatic ecosystem with micropollutants. Alternative 

processes other than ozonation also deliver promising results 

with regard to the removal of contaminants. Wastewater 

treatment with powdered activated carbon (PAC) achieves 

reduction rates of 75–90% for pharmaceuticals (including 

X-ray contrast media) with PAC dosages of 10–20 mg/L 

(Püttmann et al., 2008). The main advantage of PAC treatment 

is that a reduced chemical concentration is equivalent to 

removal of chemicals whereas ozonation leads basically to 

a transformation of the compounds. According to cost estimations 

PAC treatment is expected to be about 30% more 

expensive than ozonation (Joss et al., 2008). However, the 

main disadvantage of PAC treatment might be the need for the 

disposal of the used and contaminated carbon. Membrane 

filtration is suitable for micropollutants retention but as 

a result of considerably higher requirement for energy and 

technical equipment economically not competitive with 

ozonation or activated carbon treatment (Joss et al., 2008). 

Overall, end of pipe techniques are presumably a suitable 

solution to reduce toxicity of hazardous WW in a mediumterm 

perspective. Nevertheless in the long term source control 

strategies such as wastewater separation (e.g. urine separation), 

ecologically correct disposal of drugs by end users, reuse 

or recycling by the pharmaceutical industry or alternative 

medical treatments to drug therapies (Daughton, 2003; Joss 

et al., 2006) could offer environmentally friendly and affordable 

options. 


The sanitizing effect of ozonation was confirmed. Coagulation 

rates of eggs exposed to ozonated wastewater were on 

a lower level compared to those exposed to conventionally 

treated wastewater. Disinfection is presumably only efficient 

when linked to sand filtration because of rapid 

recovery of microorganisms due to increased assimilable 

organic carbon formation as a result of ozonation. 

Membrane filtered wastewater (for removal of microorganisms) 

reveals developmental retardation directly after 

ozonation. After sand filtration this adverse effect disappears. 

The reduced biomass and body length in fish exposed 

to ozonated wastewater is most probably a result of the 

formation of toxic oxidation byproducts. 

The mortality in the ozonated wastewater was significantly 

increased as a result of retarded development. Impairment of 

the fish’s health condition may increase the sensitivity 

towards environmental and anthropogenic stressors. 


Developmental retardation might increase the risk for the fish 

to fall prey to predators because swim up stage is delayed. 

Ozonation should not be applied without appropriate 

barrier for oxidation byproducts. Effectiveness of sand 

filtration for removal of oxidation byproducts should be 

further evaluated. 

Reduction of vitellogenin content in fish exposed to ozonated 

wastewater on control level confirms the suitability of 

this technique to reduce estrogenic activity, possibly below 

environmental relevance. 


The authors should like to express their gratitude to the staff 

from WWTP Wüeri for their technical cooperation and 

Adriano Joss from the EAWAG for helpful suggestions and 

technical support. Ulrike Schulte-Oehlmann is acknowledged 

for critical comments and helpful suggestions on the manuscript. 

This study was part of the EU project Neptune (contract 

no 036845, SUSTDEV-2005-3.II.3.2) within the Energy, Global 

Change and Ecosystems Programme of the Sixth Framework 

(FP6-2005-Global-4) and co-funded by Bundesamt für Umwelt 

(BAFU), Bern (CH) within the Strategy MicroPoll Programme 

(contract no 05.0013.PJ/F471-0916). 

references 

Andreozzi, R., Campanella, L., Fraysse, B., Garric, J., Gonnella, A., 

Lo Giudice, R., Marotta, R., Pinto, G., Pollio, A., 2004. Effects of 

advanced oxidation processes (AOPs) on the toxicity of 

a mixture of pharmaceuticals. Water Science and Technology 

50 (5), 23–28. 

Bader, H., Hoigne, J., 1981. Determination of ozone in water by the 

indigo method. Water Research 15 (4), 449–456. 

Brian,J.V.,Harris,C.A.,Scholze,M.,Backhaus,T.,Booy,P., 

Lamoree,M.,Pojana,G.,Jonkers,N.,Runnalls,T.,Bonfa,A., 

Marcomini, A., Sumpter, J.P., 2005. Accurate prediction of 

the response of freshwater fish to a mixture of estrogenic 

chemicals. Environmental Health Perspectives 113 (6), 

721–728. 

Daughton, C.G., 2003. Cradle-to-cradle stewardship of drugs for 

minimizing their environmental disposition while 

promoting human health. II. Drug disposal, waste reduction, 

and future directions. Environmental Health Perspectives 

111 (5), 775–785. 



Environmental Health Perspectives 107, 907–938. 

Escher, B.I., Bramaz, N., Quayle, P., Rutishauser, S., 

Vermeirssen, E.L.M., 2008. Monitoring of the ecotoxicological 

hazard potential by polar organic micropollutants in sewage 

treatment plants and surface waters using a mode-of-action 

based test battery. Journal of Environmental Monitoring 10 (5), 

622–631. 

European Commission, 2000. Directive 2000/60/EC of the 

European Parliament and of the Council of 23 October 2000 

establishing a framework for Community action in the field 

of water policy. Official Journal of the European 

Communities L327. Retrieved from: http://eur-lex.europa.eu/

LexUriServ/LexUriServ.do?uri¼OJ: L:2000:327:0001:0072:EN: 

PDF [12.02.2009]. 

Gagne, F., Andre, C., Cejka, P., Gagnon, C., Blaise, C., 2007. 

Toxicological effects of primary-treated urban wastewaters, 

before and after ozone treatment, on freshwater mussels 

(Elliptio complanata). Comparative Biochemistry and Physiology 

C – Toxicology & Pharmacology 145 (4), 542–552. 

Holbech, H., Kinnberg, K., Petersen, G.I., Jackson, P., Hylland, K., 

Norrgren, L., Bjerregaard, P., 2006. Detection of endocrine 

disrupters: evaluation of a fish sexual development 

test (FSDT). Comparative Biochemistry and Physiology 

C – Toxicology & Pharmacology 144 (1), 57–66. 

Huang, W.J., Fang, G.C., Wang, C.C., 2005. The determination 

and fate of disinfection by-products from ozonation of 

polluted raw water. Science of the Total Environment 345 

(1-3), 261–272. 

Huber, M.M., Ternes, T.A., von Gunten, U., 2004. Removal of 

estrogenic activity and formation of oxidation products during 

ozonation of 17a-ethinylestradiol. Environmental Science & 

Technology 38 (19), 5177–5186. 


McArdell, C.S., Ried, A., Siegrist, H., Ternes, T.A., von 

Gunten, U., 2005. Oxidation of pharmaceuticals during 

ozonation of municipal wastewater effluents: a pilot study. 


Jahnel, J., Neamtu, M., Schudoma, D., Frimmel, F.H., 2006. 

Scientific risk assessment of considered water relevant 

substances (Bestimmung von Umweltqualitaetsnormen fuer 

potenziell gewaesserrelevante Stoffe). Acta Hydrochimica et 

Hydrobiologica 34 (4), 389–397. 

Jensen, G.K., 2007. European Parliament increases number of 

chemicals earmarked for priority action by European 

Commission. The Lancet Oncology 8 (5), 382–383. 

Jobling, S., Sheahan, D., Osborne, J.A., Matthiessen, P., 

Sumpter, J.P., 1996. Inhibition of testicular growth in rainbow 

trout (Oncorhynchus mykiss) exposed to estrogenic 

alkylphenolic chemicals. Environmental Toxicology and 

Chemistry 15 (2), 194–202. 

Jobling, S., Coey, S., Whitmore, J.G., Kime, D.E., Van Look, K.J.W., 

McAllister, B.G., Beresford, N., Henshaw, A.C., Brighty, G., 

Tyler, C.R., Sumpter, J.P., 2002. Wild intersex roach (Rutilus 

rutilus) have reduced fertility. Biology of Reproduction 67 (2), 

515–524. 

Joss, A., Klaschka, U., Knacker, T., Liebig, M., Lienert, J., Ternes, T.A., 

Wennmalm, A., 2006. Source control, source separation. In: 

Ternes, T.A., Joss, A. (Eds.), Human Pharmaceuticals, Hormones 

and Fragrances. The Challenge of Micropollutants in Urban 

Water Management. ISBN 1843390930, pp. 353–384. 

Joss, A., Siegrist, H., Ternes, T.A., 2008. Are we about to upgrade 

wastewater treatment for removing organic micropollutants? 


Kidd, K.A., Blanchfield, P.J., Mills, K.H., Palace, V.P., Evans, R.E., 

Lazorchak, J.M., Flick, R.W. (2007). Collapse of a fish population 

after exposure to a synthetic estrogen. Proceedings of the 

National Academy of Sciences of the United States of America 

104(21), 8897–8901. 

Larsson, D.G.J., Adolfsson-Erici, M., Parkkonen, J., 

Pettersson, M., Berg, A.H., Olsson, P.E., Forlin, L., 1999. 

Ethinylestradiol – an undesired fish contraceptive? Aquatic 

Toxicology 45 (2-3), 91–97. 

Luster-Teasley, S.L., Yao, J.J., Herner, H.H., Trosko, J.E., Masten, S.J., 

2002. Ozonation of chrysene: evaluation of byproduct mixtures 

and identification of toxic constituent. Environmental Science 

& Technology 36 (5), 869–876. 

Luster-Teasley, S.L., Ganey, P.E., DiOrio, M., Ward, J.S., 

Maleczka, R.E., Trosko, J.E., Masten, S.J., 2005. Effect of 

byproducts from the ozonation of pyrene: biphenyl-2,2 0 ,6,6 0 - 

tetracarbaldehyde and biphenyl-2,2 0 ,6,6 0 -tetracarboxylic acid 


on gap junction intercellular communication and neutrophil 

function. Environmental Toxicology and Chemistry 24 (3), 

733–740. 

McDowell, D.C., Huber, M.M., Wagner, M., Von Gunten, U., 

Ternes, T.A., 2005. Ozonation of carbamazepine in drinking 

water: identification and kinetic study of major oxidation 

products. Environmental Science & Technology 39 (20), 

8014–8022. 

Moltmann, J.F., Liebig, M., Knacker, T., Keller, M., Scheurer, M., 

Ternes, T., 2007. Relevance of Endocrine Disrupting 

Substances and Pharmaceuticals in Surface Waters 

(Gewässerelevanz endokriner Stoffe und Arzneimittel). Final 

report. Federal Environment Agency. Funding code 20524205. 

Monarca, S., Feretti, D., Collivignarelli, C., Guzzella, L., Zerbini, I., 

Bertanza, G., Pedrazzani, R., 2000. The influence of different 

disinfectants on mutagenicity and toxicity of urban 

wastewater. Water Research 34 (17), 4261–4269. 



pharmaceuticals and personal care products (PPCPs) and 



Research 41 (19), 4373–4382. 

Nishihara, T., Nishikawa, J., Kanayama, T., Dakeyama, F., Saito, K., 

Imagawa, M., Takatori, S., Kitagawa, Y., Hori, S., Utsumi, H., 

2000. Estrogenic activities of 517 chemicals by yeast two-hybrid 

assay. Journal of Health Science 46 (4), 282–298. 

OECD, 1992a. Fish Acute Toxicity Test. OECD Guidelines for the 

Testing of Chemicals. No. 203. Organisation for Economic 

Cooperation and Development, Paris. 

OECD, 1992b. Fish Early Life Stage Toxicity Test. OECD Guidelines 

for the Testing of Chemicals. No. 210. Organisation for 

Economic Cooperation and Development, Paris. 

Parrott, J.L., Blunt, B.R., 2005. Life-cycle exposure of fathead 

minnows (Pimephales promelas) to an ethinylestradiol 

concentration below 1 ng/L reduces egg fertilization success 

and demasculinizes males. Environmental Toxicology 20 (2), 

131–141. 

Petala, M., Samaras, P., Zouboulis, A., Kungolos, A., 

Salkellaropoulos, G., 2006. Ecotoxicological properties of 

wastewater treated using tertiary methods. Environmental 

Toxicology 21 (4), 417–424. 

Püttmann, W., Keil, F., Oehlmann, J., Schulte-Oehlmann, U., 2008. 

Strategy to reduce pharmaceuticals in drinking water – 

technical approach (Wassertechnische Strategien zur 

Reduzierung der Trinkwasserbelastung durch 

Arzneimittelwirkstoffe). Umweltwissenschaften und 

Schadstoff-Forschung 20 (3), 209–226. 

Purdom, C.E., Hardiman, P.A., Bye, V.V.J., Eno, N.C., Tyler, C.R., 

Sumpter, J.P., 1994. Estrogenic effects of effluents from sewage 

treatment works. Chemical Ecology 8 (4), 275–285. 

Rodgers-Gray, T.P., Jobling, S., Kelly, C., Morris, S., Brighty, G., 

Waldock, M.J., Sumpter, J.P., Tyler, C.R., 2001. Exposure of 

juvenile roach (Rutilus rutilus) to treated sewage effluent 

induces dose-dependent and persistent disruption in gonadal 

duct development. Environmental Science & Technology 35 

(3), 462–470. 

Schmidt, C.K., Brauch, H.J., 2008. N, N-dimethosulfamide as 

precursor for N-nitrosodimethylamine (NDMA) formation 

upon ozonation and its fate during drinking water treatment. 


Schwaiger, J., Ferling, H., Mallow, U., Wintermayr, H., 

Negele, R.D., 2004. Toxic effects of the non-steroidal antiinflammatory 

drug diclofenac Part 1: histopathological 

alterations and bioaccumulation in rainbow trout. Aquatic 

Toxicology 68 (2), 141–150. 

Schwarzenbach, R.P., Escher, B.I., Fenner, K., Hofstetter, T.B., 

Johnson, C.A., von Gunten, U., Wehrli, B., 2006. The challenge

448 

of micropollutants in aquatic systems. Science 313 (5790), 

1072–1077. 

Ternes, T.A., Stumpf, M., Müller, J., Haberer, K., Wilken, R.D., 

Servos, M., 1999. Behavior and occurrence of estrogens in 

municipal sewage treatment plants – I. Investigations in 

Germany, Canada and Brazil. Science of the Total 

Environment 225 (1-2), 81–90. 

Tixier, C., Singer, H.P., Oellers, S., Müller, S.R., 2003. Occurrence 

and fate of carbamazepine, clofibric acid, diclofenac, 

ibuprofen, ketoprofen, and naproxen in surface waters. 


Triebskorn, R., Casper, H., Heyd, A., Eikemper, R., Köhler, H.R., 

Schwaiger, J., 2004. Toxic effects of the non-steroidal antiinflammatory 

drug diclofenac Part II. Cytological effects in 

liver, kidney, gills and intestine of rainbow trout (Oncorhynchus 

mykiss). Aquatic Toxicology 68 (2), 151–166. 


Tyrrell, S.A., Rippey, S.R., Watkins, W.D., 1995. Inactivation of 

bacterial and viral indicators in secondary sewage 

effluents, using chlorine and ozone. Water Research 29 

(11), 2483–2490. 

von Gunten, U., 2003. Ozonation of drinking water: Part I. 

Oxidation kinetics and product formation. Water Research 37 

(7), 1443–1467. 

Wang, J.Z., Summers, R.S., 1996. Biodegradation behavior of 

ozonated natural organic matter in sand filters. Revue des 

Sciences de l’Eau 9 (1), 3–16. 

Wert, E.C., Rosario-Ortiz, F.L., Drury, D.D., Snyder, S.A., 2007. 

Formation of oxidation byproducts from ozonation of 


Ying, G.G., Williams, B., Kookana, R., 2002. Environmental fate of 

alkylphenols and alkylphenol ethoxylates – a review. 

Environment International 28 (3), 215–226.

The role of organic matter in the removal of emerging 

trace organic chemicals during managed aquifer recharge 

T. Rauch-Williams, C. Hoppe-Jones, J.E. Drewes* 

Advanced Water Technology Center (AQWATEC), Colorado School of Mines, Environmental Science and Engineering Division, 

Golden, CO 80401-1887, United States 

article info 





Accepted 20 August 2009 

Available online 27 August 2009 

Keywords: 

Trace organic chemicals (TOrC) 

Groundwater recharge 

Effluent organic matter 

Managed aquifer recharge 

Riverbank filtration 

Biotransformation 

Co-metabolism 

Primary substrate 


abstract 

Managed aquifer recharge (MAR) systems, such as riverbank 

filtration (RBF) and soil aquifer treatment (SAT), are widely 

used natural processes for drinking water augmentation 

projects using source water that might be impaired by 

wastewater discharge. Previous studies have demonstrated 

that MAR systems are effective in dampening and reducing 




* Corresponding author. 

E-mail address: jdrewes@mines.edu (J.E. Drewes). 


doi:10.1016/j.watres.2009.08.027 

This study explored the effect of different bulk organic carbon matrices on the fate of trace 

organic chemicals (TOrC) during managed aquifer recharge (MAR). Infiltration through 

porous media was simulated in biologically active column experiments under aerobic and 

anoxic recharge conditions. Wastewater effluent derived organic carbon types, differing in 

hydrophobicity and biodegradability (i. e., hydrophobic acids, hydrophilic carbon, organic 

colloids), were used as feed substrates in the column experiments. These carbon substrates 

while fed at the same concentration differed in their ability to support soil biomass growth 

during porous media infiltration. Removal of degradable TOrC (with the exception of 

diclofenac and propyphenazone) was equal or better under aerobic versus anoxic porous 

media infiltration conditions. During the initial phase of infiltration, the presence of 

biodegradable organic carbon (BDOC) enhanced the decay of degradable TOrC by 

promoting soil biomass growth, suggesting that BDOC served as a co-substrate in a cometabolic 

transformation of these contaminants. However, unexpected high removal 

efficiencies were observed for all degradable TOrC in the presence of low BDOC concentrations 

under well adopted oligotrophic conditions. It is hypothesized that removal under 

these conditions is caused by a specialized microbial community growing on refractory 

carbon substrates such as hydrophobic acids. Findings of this study reveal that the 

concentration and character of bulk organic carbon present in effluents affect the degradation 

efficiency for TOrC during recharge operation. Specifically aerobic, oligotrophic 

microbiological soil environments present favorable conditions for the transformation of 

TOrC, including rather recalcitrant compounds such as chlorinated flame retardants. 


the concentrations of dissolved organic carbon (DOC) as well 

as various trace organic contaminants (TOrC) that might be 

present in impaired source waters (Drewes and Fox, 1999; 

Brauch et al., 2000; Grünheid et al., 2005). The presence of 

TOrC has become a key concern for drinking water augmentation 

projects during the past decade (Kolpin et al., 2002; 

Heberer, 2002; Focazio et al., 2008). Although adverse human 

health effects caused by these compounds at concentrations

450 

commonly observed in impaired water resources are very 

unlikely (Schwab et al., 2005), minimizing exposure of 

wastewater derived contaminants in these projects is desired. 

Previous research on the fate of TOrC during MAR has 

primarily focused on collecting anecdotal and site specific 

information on their occurrence and removal (Drewes et al., 

2002; Montgomery-Brown et al., 2003; Grünheid et al., 2005; 

Dillon et al., 2008). Studies delineating the mechanisms and 

boundary conditions for the transformation of wastewater 

derived TOrC during MAR are lacking. 

Previous studies demonstrated that the type and bioavailability 

of effluent organic matter (EfOM) controls the extent of 

soil biomass growth in MAR systems (Rauch and Drewes, 2004; 

Rauch and Drewes, 2005). EfOM may consequently effect the 

degradation of TOrC by serving as a co-substrate in microbiologically 

facilitated transformations (Stratton et al., 1983). 

The diversity and expression of the soil microbial community 

also depends on composition and concentration of the organic 

carbon substrate controlling trophic cycles in the subsurface 

(Preuß and Nehrkorn, 1996; Szewzyk et al., 1998). The 

composition of EfOM (i.e. in terms of its bioavailability) is 

primarily determined by the degree of wastewater treatment 

employed (Drewes and Fox, 1999), which can vary widely from 

primary to conventional to advanced wastewater treatment. 

As a result, effluent qualities fed to MAR systems can vary in 

biodegradable dissolved organic carbon (BDOC) concentrations 

from less than 1 up to 15 mg/L or more. As a consequence, 

soil microbial communities growing on different 

levels of BDOC can differ widely in total biomass and diversity. 

The objectives of this research were to investigate the role 

of 1) abiotic vs. biotic conditions, 2) BDOC and 3) the type of 

organic carbon matrices on the removal of select TOrC, such 

as pharmaceutical residues, personal care products, and 

household chemicals, during MAR. 

2. Methodology 

2.1. Target organic contaminants 

Compounds selected for this study represent small molecular 

weight organic chemicals (180 to 360 Dalton) that are hydrophilic 

at neutral pH regimes as indicated by an octanol/water 

partition coefficient at pH 7 (log DpH¼7 of less than 2.6). TOrC 

with these properties have a high potential to migrate into 

groundwater and are not expected to be adsorbed onto porous 

media. The molecular structures and physicochemical properties 

of the target compounds chosen for this study are presented 

in Table 1. These compounds cover a wide range of 

biodegradability as previously reported for soil/water systems. 

The anticonvulsants carbamazepine and primidone have 

been classified as recalcitrant during wastewater treatment 

and MAR in earlier studies (Heberer, 2002; Drewes et al., 2003; 

Clara et al., 2004). The chlorinated phosphate esters tris 

(1-chloroisopropyl)-phosphate (TCPP) and tris(2-chloroethyl)phosphate 

(TCEP) are two widely used flame retardants and 

are persistent in the aquatic environment (Heberer et al., 2001; 

Fries and Püttmann, 2003). During bank filtration in Germany, 

however, TCPP and TCEP exhibited a significant reduction, 

which was attributed to biotransformation in the aquifer 


(Heberer et al., 2003). Amy and Drewes (2006) also reported 

removal of TCEP to concentrations below the detection limit 

after two years of subsurface travel in an MAR facility supporting 

that chlorinated flame retardants can be biotransformed 

under anoxic conditions. Propyphenazone is 

a poorly biodegradable analgesic that persists during RBF 

(Heberer et al., 2003) and SAT (Drewes et al., 2003). For diclofenac, 

a popular analgesic drug, low removal due to biodegradation 

or adsorption was reported (Buser et al., 1998; Möhle 

et al., 1999) unless soils contain a high organic carbon content 

(Drillia et al., 2003). Several studies report a faster degradation 

of diclofenac under anoxic conditions as compared to aerobic 

conditions (Zwiener and Frimmel, 2003; Hua et al., 2003). 

Opposing results were reported by Schmidt et al. (2004) in that 

diclofenac was almost completely removed during aerobic 

bank filtration but recalcitrant during anaerobic recharge. 

Ibuprofen, ketoprofen, and naproxen are common analgesics 

that are well degradable during wastewater treatment (Buser 

et al., 1999; Zwiener and Frimmel, 2003; Carballa et al., 2004) 

and during soil infiltration (Sedlak and Pinkston, 2001; Drewes 

et al., 2003). Gemfibrozil is a commonly prescribed blood lipid 

regulator found in unconfined shallow aquifers impacted by 

wastewater infiltration in the low to moderate ng/L-concentration 

range (Heberer and Stan, 1997; Drewes and Shore, 

2001; Heberer, 2002). Gemfibrozil was removed below the limit 

of detection within a few weeks during groundwater recharge 

using SAT (Drewes et al., 2003). 

2.2. Analytical methods 

2.2.1. GC/MS analysis 

The following TOrC were analyzed by gas chromatography 

coupled with mass spectrometry (GC/MS) using a HP 6890 

gas chromatograph and a HP 5973 quadrupole mass spectrometer 

from Agilent Technologies (Waldbronn, Germany) 

adopting a method published by Reddersen and Heberer 

(2003): gemfibrozil, primidone, diclofenac sodium salt, 

carbamazepine, ketoprofen, naproxen, phenacetine, tris 

(2-chloroethyl)phosphate (TCEP) (Sigma Aldrich Chemicals), 

tris(chloropropyl)phosphate (TCPP), and propyphenazone 

(Pfaltz & Bauer, Inc.). Stock solutions were prepared by dissolving 

the compounds in milli-Q water adjusted to a pH of 

10, during sonification and in the dark. A volume of 500 mL 

of sample was collected and filtered (0.45 mm, Whatman) 

prior to solid phase extraction. 10,11-dihydrocarbamazepine 

(Sigma Aldrich Chemicals) and 2-(m-chlorophenoxy) propionic 

acid (Sigma Aldrich Chemicals) were used as surrogate 

standards. The limits of detection (LofD) and limit of 

quantification (LoQ) for the target compounds ranged from 5 

to 10 ng/L and from 10 to 50 ng/L, respectively. Only parent 

target compounds were investigated in this study, metabolites 

and conjugates were not analyzed. 

2.2.2. HPLC analysis 

A high performance liquid chromatography (HPLC) HP 1100 

(Agilent Technologies) combined with a UV diode array detector 

(DAD) was used to quantify concentrations of primidone, phenacetine, 

carbamazepine, naproxen, diclofenac, and ibuprofen 

in the lower mg/L range during adsorption breakthrough tests. 

The samples were directly injected without filtration or other

Table 1 – Physical–chemical properties of target organic compounds. 

Compound Structure pKa log P 

(log DpH7) 

Carbamazepine 

(anticonvulsant) 

Diclofenac sodium 

(analgesic/antiinflammatory) 

Gemfibrozil (blood 

lipid regulator) 

Ibuprofen 


Ketoprofen 


Naproxen 


Phenacetine 

(antipyretic) 

Primidone 

(anticonvulsant) 

Propyphenazone 

(analgesic) 


Expected 

degradability 

13.9 2.67 Persistent 

4.0 

4.8 

4.4 

4.2 

4.0 

n/a 

12.3 

n/a 

4.06 

(1.06) 

4.39 

(2.19) 

3.72 

(1.12) 

2.81 

(0.01) 

3.00 

(0.00) 

1.63 

(1.63) 

0.4 

(0.4) 

1.74 

(1.74) 

Degradable 

(anoxic) 

Relatively 

persistent 

Degradable 

(better 

aerobic) 

Degradable 

Degradable 

Relatively 

persistent 

Persistent 

Poor/ 

persistent 

References 

Mersmann et al., 

2002; Schmidt 

et al., 2004 

Buser et al., 1998; 

Möhle et al., 1999; 

Drillia et al., 2003; 

Zwiener and Frimmel, 

2003, Hua et al., 

2003; Schmidt et al., 2004 

Heberer and Stan, 1997; 

Drewes and Shore, 2001; 

Heberer, 2002; Drewes et al., 2003 



2003; Carballa et al., 2004; 

Sedlak and Pinkston, 2001; 

Drewes et al., 2003 





















Heberer et al., 2003; 



452 


Compound Structure pKa log P 

(log DpH7) 

Tris(2chloroethyl)phosphate 

(TCEP) (flame 

retardant) 

sample preparation. The mobile phase consisted of two 

solvents: 1) type 1 water (adjusted to pH 2.5 with o-phosphoric 

acid (HPLC grade, Fisher Scientific) and buffered with 25 mmol 

potassium phosphate monobasic (Fisher Scientific)), and 2) 

acetonitrile (HPLC grade, Mallinckrodt ChromAR HPLC) with the 

same concentration of o-phosphoric acid as solvent 1). Sample 

injection (100 mL) occurred in triplicates for each sample at a flow 

rate of 2 mL/min. A mobile phase gradient was applied during 

the run for solvent 2 (acetonitrile) of 30% at t ¼ 0 min., 80% at 

t ¼ 12 min., and 30% at t ¼ 15 min. with a post-run time of 7 min. 

for column cleaning (30% of solvent 2). The detection wavelengths 

(band width 10 nm) were set to 205 nm at t ¼ 0min.for 

quantification of primidone, phenacetine, and carbamazepine, 

230 nm at t ¼ 6.8 min. (naproxen), and 205 nm at 7.8 min. for 

detection of diclofenac and ibuprofen with a reference wavelength 

of 330 nm (band width 60 nm) for all compounds. Standards 

were run in the range of 5–500 mg/L. The detection limits 

for each compound were 5 mg/L. 

2.2.3. DOC/UV absorbance/nitrate/ammonium 

A Sievers 800 Total Organic Carbon Analyzer (GE, Boulder) 

was used for DOC quantification after microfiltration 

(0.45 mm, Whatman) (Standard Method 5310C). UV absorbance 

(UVA) measurements were conducted at 254 nm using 

a Beckman Coulter DU 800 Spectrophotometer after 0.45 mm 

filtration (Standard method 5910B). The specific UV absorbance 

(SUVA) was calculated as the ratio of UVA and DOC. 

Ammonium and nitrate concentrations were measured 

using the Nessler and the chromotropic acid method (HACH) 

with a detection range of 0.02–2.0 mg/l and 0.2–30 mg/L, 

respectively. 

2.2.4. Biomass 

Soil biomass was determined as total viable biomass (i.e. 

viable, not necessarily active bacteria) using phospholipid 

extraction (PLE) as described in Rauch and Drewes (2005). 

Analyses were conducted in triplicates from the top-soil 

(0–2 cm, infiltration zone) of the columns. 

2.2.5. Wastewater effluent 

Secondary treated wastewater effluent served as the feed to 

column systems and as source for subsequent organic carbon 

fractionation. The secondary treated effluent was collected 

from a local wastewater treatment plant employing nitrification 

and partial denitrification. The average DOC concentration 

of the effluent was 8.74 1.44 mg/L. 


n/a 

0.48 

(0.48) 

n/a not applicable. 

pKa and log P values calculated by software ACD. 

log D pH¼7 values calculated using equations proposed by Scherrer and Howard, (1977). 

Expected 

degradability 

Relatively 

persistent 

2.2.6. Organic carbon fractionation 

Organic matter (less than 1 mm in size as defined for this study) 

of the secondary effluent sample was isolated into three 

organic fractions: colloidal organic carbon, hydrophobic acids 

(HPO-A), and hydrophilic carbon (HPI) following the procedure 

described in Rauch and Drewes (2005) with some modifications. 

The sample was concentrated by vacuum rotary evaporation 

at 45 C (concentration factor: 40). The concentrate 

was separated by dialysis (Spectra/Por, Spectrum, 6000– 

8000 Dalton) at pH 4–5 into organic colloids and DOC. The 

permeate of the dialysis (containing DOC) was collected and 

further separated into HPI and HPO-A by XAD-8 fractionation 

according to Leenheer et al. (2000) using a capacity factor of 

k 0 ¼ 4. A carbon mass balance was performed for each carbon 

fractionation based on UVA and DOC measurements for 

quality control. In average, the secondary treated effluent 

contained 16 percent colloidal organic carbon, 38 percent 

HPO-A, 37 percent HPI, and 8 percent hydrophobic neutrals 

(HPO-N) (Rauch and Drewes, 2004). Organic carbon isolates 

were diluted to 3 mg/L DOC using milli-Q water, adjusted for 

ion strength, macro nutrient (nitrogen, phosphate) and micro 

nutrient concentrations (Rauch and Drewes, 2004) and 

promptly utilized as column feed waters. 

2.3. Experimental set-up 

References 

Heberer et al., 2001; 

Heberer et al., 2003, 

Amy and Drewes, 2006; 

Fries and Püttmann, 2003 

2.3.1. Anoxic column system 

The columns (PC system) were operated to study TOrC removal 

during simulated MAR under anoxic conditions and up to 3–4 

weeks of retention time in the subsurface. The anoxic columns 

consisted of four 1-m plexiglass columns (15 cm i.d.) in series 

filled with aquifer material (d50 ¼ 0.8 mm, foc ¼ 0.003%) (Fig. SI-1, 

Supplemental Information). The column system was operated 

in flow-through mode at a loading rate of 0.065 m/d under 

saturated, anoxic (denitrifying) flow conditions (Table 2). The 

hydraulic retention time of the system using four columns in 

series was previously determined as 25 days. The columns had 

been continuously fed with secondary or tertiary treated effluents 

for over 6 years and for 5 months with the secondary 

treated effluent quality employed in this study prior to spiking of 

TOrC. The column influent was regularly purged with nitrogen 

gas to keep dissolved oxygen concentrations below 0.5 mg/L. 

Samples were collected once to twice a week from column 

influent and the four column effluents, respectively, and 

analyzed for DOC, UV absorbance, pH, conductivity and TOrC 

concentrations.

Table 2 – Operational parameters and water quality changes of columns used in this study. a 

Label Feed water 

matrix 

PC Anoxic columns 

(EfOM) 

C1 Hydrophobic acids 

(HPO-A) 

C2 Hydrophilic carbon 

(HPI) 

C3 Effluent organic 

matter (EfOM) 

C4 Organic colloids 

(Org. Colloids) 

Abiotic Abiotic column, 

Type I water 

Length 

(m) 

4 1m¼ 

4m 

2.3.2. Organic carbon fraction columns 

Four columns (C1–C4) were operated to examine the effect of 

different organic carbon matrices on soil biomass growth and 

TOrC removal during aerobic conditions and short retention 

times (less than 1 day) in the subsurface. This system consisted 

of four parallel plexiglass columns (L ¼ 30 cm, i.d. ¼ 5 cm) filled 

with silica sand (d50 ¼ 0.65 mm, foc ¼ 0.004%). The columns 

were operated under saturated, aerobic flow conditions at 

a loading rate of 0.9 m/d representing a retention time of about 

19 h in the porous media. The system had been acclimated for 

over two years to the infiltration of the four organic carbon 

fractions derived from secondary effluent (HPO-A, HPI, organic 

colloids, EfOM), respectively. Approximately six weeks prior to 

spiking TOrC to the column systems, the feed water concentrations 

of all four fractions was adjusted to 3 mg/L as DOC. 

Samples for DOC, UVA and pH were taken once to twice a week 

from the four column influents and effluents. Three consecutive 

TOrC spiking events were conducted. For each event, 

hydraulically corresponding samples of column influents and 

effluents were collected, respectively, after a complete breakthrough 

of the spike solution was observed using conductivity 

as a conservative tracer (typically after 24 h). After the 1st 

spiking event (day 0), the 2nd and 3rd spikes occurred on day 

11 and 22, respectively. Set-up and operation conditions of the 

different columns employed in this study are summarized in 

Table 2. All small column systems received feed waters that 

were adjusted to a pH of 8.0 0.2 to minimize adsorption 

effects of acidic TOrC onto porous media. 

2.3.3. Adsorption test 

An additional column was operated under abiotic conditions 

to assess the role of physical adsorption of selected TOrC onto 

porous media. The column set-up and operation was the same 

as described for the biologically active columns C1–C4. The 

adsorption column was filled with aquifer material from an 

RBF site (d50 ¼ 0.55 mm, bulk density ¼ 1.81 g/cm 3 , effective 

porosity ¼ 0.36, fOC ¼ 0.066%). Six TOrC were spiked into 

secondary effluent in a concentration range of 175–226 mg/L 

adjusted to a pH of 7.0 (feed water). To minimize biotransformation 

during this experiment, the column was kept 

Organic Carbon Predominant 

Feed water Effluent BOC 

Redox Condition 

(mg/L) (mg/L) (mg/L) 

6.9 See Fig. 1 for 

DOC column profile 

abiotic by adding sodium azide at a concentration of 2 mM. 

Sodium bromide was added as a conservative tracer and 

bromide was measured by ion chromatography (Dionex, 

Sunnyvale, CA) in the column effluent. The column was 

operated in single flow-through mode under saturated flow 

conditions at a hydraulic loading rate of 0.33 m/d. Effluent 

samples were collected using a fraction collector and analyzed 

by HPLC analysis. Retardation factors (Rd) were fitted to the 

step input breakthrough curves for each compound using the 

software program CXTFIT (Toride et al., 1995). Adsorption 

coefficients (K d) were calculated according to the following 

relationship 

Rd ¼ rKd 

þ 1 (1) 

3 

where Rd is the retardation factor, r is the sand bulk density, 

and 3 is the sand porosity. 


3.1. Adsorption behavior of target TOrC 

In order to assess whether physical retardation would be 

a contributing factor in the removal of select TOrC in the 

column experiments, an abiotic column experiment was 

conducted to study sorption of the compounds onto the virgin 

porous media used in the columns. None of the six TOrC 

included in this experiment exhibited significant retardation 

during flow through porous media as compared to the 

conservative tracer bromide (indicated by R d values very close 

to 1, Table 3). These findings reveal that sorption did not play 

an important role in the attenuation of these compounds onto 

the porous media used in the columns. 

3.2. Organic carbon removal 

Total Viable 

Soil Biomass b 

(nmol PO4 3 /g d.w soil) 

Anoxic 17.3 1.3 

0.3 3.1 0.9 2.7 0.02 0.4 0.9 Aerobic 3.3 0.4 

0.3 2.8 0.4 2.6 0.1 0.3 0.4 Aerobic 10.5 5.3 

0.3 3.0 0.4 2.7 0.4 0.4 0.4 Aerobic 20.6 2.5 

0.3 3.1 0.5 2.2 0.3 0.8 0.5 Aerobic 27.2 1.8 

0.3

454 

Table 3 – Retardation and sorption coefficients of select 

TOrC derived during abiotic column study. 

TOrC Rd Kd (mL/g) Mass recovery (%) 

Naproxen 1.30 0.08 94.2 

Ibuprofen 1.10 0.03 92.7 

Phenacetine 1.02 n.a. n.a. 

Diclofenac 1.41 0.01 93.6 

Primidone 1.00

Table 4 – Summary of TOrC concentrations observed in column studies. 

TOrC (ng/L) Feed 

water 

1-m 

Column 

effluent 

Anoxic Column HPO-A (C1) HPI (C2) EfOM (C3) Org. colloids (C4) 

2-m 

Column 

effluent 

3-m 

Column 

effluent 

4-m 

column 

effluent 

Column 

influent 

Column 

effluent 

Column 

influent 

Column 

effluent 

Column 

influent 

Column 

effluent 

Column 

influent 

Carbamazepine 279–296 (n ¼ 3) n.a. n.a. n.a. 318 n.a. n.a. n.a. n.a. 

Diclofenac sodium 307–416 (n ¼ 2) 75 134 110 80–120 (n ¼ 2) 604 538 n.a. n.a. 710 682 

538 190 70

456 

Fig. 2 – Removal of TOrC in anoxic column system. 

first-order degradation kinetics are indicative of organic 

carbon degradation at concentrations below the half-velocity 

coefficient Ks, or under mass transport limitations that keep 

biodegradation rates below the biological capacity. Zero-order 

reaction kinetics are indicative of degradation below the biological 

degradation potential, where either substrate concentrations 

of trace contaminants are significantly higher than 

the affinity for the specific substrate, or the limiting substrate 

is delivered at a constant rate (Simoni et al., 2001; Alexander, 

1999). It should be noted that, while the column results 

ng/L removed in column 

900 

800 

700 

600 

500 

400 

300 

200 

100 

0 

Gemfibrozil Ketoprofen Naproxen 

3.3 10.4 27.2 

HPO-A HPI Org. Colloids 

3- -1 

nmol org-PO4 g 

Fig. 3 – Relationship between TOrC removal, soil biomass, 

and organic carbon substrate in columns C1, C2, and C4 

(1st spike event). 


indicate a linear removal profile for ketoprofen, naproxen, 

gemfibrozil, and ibuprofen, the removal profile might also 

have been affected by a change in soil microbial community 

composition with column depth caused by changing organic 

carbon substrate composition and concentrations along the 

flow path. Thus, definite conclusions about the kinetic order 

are not possible without additional experiments. 

Due to the low ambient concentrations of trace organic 

contaminants (occurring in the ng/L range) it is unlikely that 

energy from TOrC metabolism is sufficient for biomass 

maintenance and growth. It was hypothesized that at trace 

levels, TOrC were transformed by co-metabolism (co-utilization). 

In this context the term co-metabolism describes 

a transformation of a trace compound, which is by itself 

unable to support cell replication, and requires the presence of 

another transformable organic compound (primary carbon 

source) that allows microorganisms to obtain energy for their 

metabolism and growth (Brandt et al., 2003; Stratton et al., 

1983; Clara et al., 2005). Because the primary substrate 

(biodegradable bulk organic carbon) was present in the 

column experiment in concentrations two orders of magnitude 

above common concentrations of TOrC in the environment, 

co-metabolism offers a possible explanation for the 

observed removal of the target TOrC, that were fed at trace 

concentrations. This hypothesis was further tested by examining 

the effect of organic carbon subgroups present in EfOM 

and differing in biodegradability on soil biomass growth and 

TOrC removal. 

3.6. Effect of aqueous organic carbon matrices 

on TOrC removal 

In order to further study the effect of soil biomass and aqueous 

organic carbon matrices on the removal of TOrC, aerobic 

column experiments were conducted using different organic 

carbon bulk fractions (EfOM, HPO-A, HPI, and organic colloids) 

isolated from secondary treated effluent. The different organic 

bulk substrates were fed at similar DOC concentrations to 

different column systems and were able to support different 

amounts of viable soil biomass (Table 2). During three consecutive 

spiking tests, hydraulically corresponding influent and 

effluent samples were collected from each column (Fig. 3)with 

the exception of the column fed with organic colloids for which 

only two spiking tests were conducted. 

The results of the first spiking test for the HPI and organic 

colloid columns are presented in Fig. 3. Findings suggest that 

the more viable soil biomass was present on the column 

media, the more complete was the removal for gemfibrozil, 

ketoprofen, and naproxen. In other terms the concentration of 

biodegradable effluent derived organic carbon substrate 

limited the transformation of TOrC. (Phenacetine was 

removed below the quantification limit in all columns and 

removal could, thus, not be related to biomass activity in the 

columns.) Organic colloids (BDOC content 0.86 0.48 mg/L) 

appeared to have promoted the degradation of TOrC by 

serving as the primary substrate and establishing a relatively 

higher microbial population that was able to co-metabolize 

TOrC better as compared to HPI organic carbon. TOrC removal 

in the HPI column was possibly lower because HPI substrate 

(BDOC content 0.25 0.41 mg/L) supported less biomass

emoval % 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

growth to metabolize TOrCs. During longer exposure times to 

the contaminants the performance of the HPI column for 

TOrC removal improved. In the 2nd and 3rd spiking event, 

ketoprofen, phenacetine, gemfibrozil, and naproxen exhibited 

increasingly better removal exceeding 95 percent 22 days after 

the first spike (Fig. 4) suggesting that the microbial community 

had adapted to metabolizing TOrC. 

In accordance with the performance of the anoxic column 

system (Fig. 2), ibuprofen was equally efficient removed (>85%) 

in the presence of all organic carbon fractions and in all three 

spiking events (Table 4 and Fig. 4) under both aerobic and anoxic 

conditions. This confirms previously reported findings that 

ibuprofen is very well degradable under various field conditions 

(Sedlak and Pinkston, 2001; Drewes et al., 2003). In contrast to the 

anoxic column system, propyphenazone was not eliminated in 

any of the four short columns(Table 4). Consistent with previous 

studies, propyphenazone was found to require longer retention 

times (>1 week) to exhibit partial removal under MAR conditions 

(Drewes et al., 2003; Heberer et al., 2003). The anticonvulsant 

drugs were not eliminated in any of the columns 

investigated (Table 4). Based on these study results, effluentderived 

BDOC does not seem to be a suitable substrate to stimulate 

co-metabolic decay of the selected anticonvulsant drugs 

(i.e. primidone, carbamazepine) under environmentally relevant 

bulk and TOrC concentrations. Whereas TCPP was not 

removed in the anoxic columns, this flame retardant seems to 

be removed in the aerobic HPO-A, HPI, and EfOM column after 

a lag time of two to three weeks after starting to feed this 

compound to the column, possibly indicating biological acclimatization 

leading to some degree of attenuation of this 

compound (Table 4). However, results are inconsistent when 

comparing influent and effluent concentrations in the different 

columns, therefore further studies are recommended to solidify 

this observation. 

Given that the first column of the anoxic system was 

operated at a longer residence time (about 6 days) compared to 

the aerobic columns (less than 1 day), it is noticeable that most 

degradable TOrC (i.e. gemfibrozil, ibuprofen, ketoprofen, 

naproxen) were significantly better transformed under 

aerobic conditions as compared to anoxic recharge 

87 

1st spike 2nd spike (day 11 after 1st spike) 

3rd spike (day 22 after 1st spike) 

561 

925 


206 

544 

700 

550 

506 

Ketoprofen Naproxen Phenacetine Gemfibrozil Ibuprofen 

508 

113 

200 

243 

490 495 499 

Fig. 4 – Microbial adaptation to TOrC in HPI column (values on top of bars represent removal in ng/L). 

conditions. Diclofenac exhibited consistent removal during 

anoxic conditions but not during aerobic recharge, which is 

consistent with findings reported by Zwiener and Frimmel 

(2003) and Hua et al. (2003). 

3.7. Removal of TOrC in oligotrophic environments 

Although the HPO-fraction provided little soil biomass growth 

(

458 

Because oligotrophic organisms are more likely to be prevalent 

in native environments characterized by low-carbon 

fluxes than in wastewater treatment plants, the degradation 

of TOrC may be more efficient during soil percolation than in 

wastewater treatment plants (Daughton and Ternes, 1999). 

It is anticipated that oligotrophic organisms reside in MAR 

systems in soil depths below the immediate infiltration zone 

(which in contrast is likely dominated by microbial organisms 

growing primarily on easily degradable EfOM). Further investigations 

directed at the interrelation between effluent derived 

organic carbon and the succession of biocommunities in soil 

based on trophic carbon sequences seem key to the understanding 

of mechanisms and limitations to TOrC removal in 

MAR systems. 


Hydrophilic, small-molecular size TOrC can survive conventional 

and advanced wastewater treatment processes and are 

one of the key concerns in drinking water augmentation 

projects using impaired source waters. More research is 

needed to better understand the trophodynamic role of organic 

carbon on the removal of TOrC in MAR applications. In this 

study we conducted column experiments to illuminate key 

factors for the metabolic removal of twelve TOrC during soil 

infiltration. Experiments were designed to resemble natural 

environments (flow through systems, long biological acclimatization 

times, complex aqueous carbon compositions, environmentally 

relevant bulk and TOrC concentrations) in order 

to reveal interactions between bulk organic carbon and TOrC 

removal that are relevant for full-scale MAR conditions. 

Removal efficiencies varied widely among the target 

compounds but were generally higher during aerobic soil 

infiltration with the exception of diclofenac that was faster 

degraded during anoxic recharge. Propyphenazone required 

a longer residence time for removal than was provided in the 

aerobic columns employed in this study. The biological reactions 

for all other degradable TOrC were fast at ambient 

concentrations (ng/L level) and completed within the first 2 m 

of porous media infiltration. Threshold concentrations below 

which further degradation was negligible were observed for 

acidic TOrC under anoxic conditions in the lower ng/L-range. 

Intensive research has been conducted on stimulating the 

degradation of certain pollutants by artificially adding carbon 

substrates (Horvath, 1973, Semprini, 1997). Findings of this 

study demonstrated that naturally available organic matter in 

aqueous environments can promote the degradation of 

certain TOrC by serving as a secondary carbon substrate for 

their removal. The composition of effluent-derived bulk 

organic carbon had a strong effect on the biological removal of 

the target TOrC and is based on several mechanisms. BDOC, 

prevalent in form of colloidal and hydrophilic carbon, stimulates 

soil biomass growth and induces secondary substrate 

utilization of TOrC. Soil biomass in aerobic systems was able 

to adapt to most acidic drugs showing an improved removal 

over time. Soil biomass growth occurred even in response to 

the infiltration of recalcitrant hydrophobic acids. This organic 

carbon substrate induced an oligotrophic microbial community 

that was more effective in removing TOrC than during 


copiotrophic metabolism in the presence of higher concentrations 

of BDOC. Consistent with previous observations, the 

anticonvulsant drugs carbamazepine and primidone exhibited 

no removal under any conditions examined in this study. 

Acknowledgments 

Partial funding for this study was provided by the Gwangju 

Institute of Technology in Korea. We are thankful for analytical 

support during this study provided by Dr. Thomas Heberer 

at the Technical University in Berlin, Germany, Matt 

Oedekoven, and Stephan Wagner. 

Appendix. Supplementary data 

Supplementary data associated with this article can be found 

in the online version at doi:10.1016/j.watres.2009.08.027. 

references 

Amy, G., Drewes, J.E., 2006. Soil-aquifer treatment (SAT) as 

a natural and sustainable wastewater reclamation/reuse 

technology: fate of wastewater effluent organic matter (EfOM) 

and trace organic compounds. Environmental Monitoring and 

Assessment 129 (1–3), 19–26. 

Alexander, M., 1999. Biodegradation and Bioremediation. 

Academic Press, San Diego, California. 

Brandt, B.W., Van Leeuwen, I.M.M., Kooijman, S.A.L.M., 2003. 

A general model for multiple substrate biodegradation. 

Application to co-metabolism of structurally non-analogous 

compounds. Water Research 37, 4843–4854. 

Brauch, H.-J., Sacher, F., Denecke, E., Tacke, T., 2000. Effectiveness 

of bank filtration for the removal of polar organic trace 

contaminants. GWF Wasser Abwasser 141 (4) 2260–234 (in 

German). 

Buser, H.R., Poiger, T., Müller, M.D., 1998. Occurrence and fate of 

the pharmaceutical drug diclofenac in surface waters: rapid 

photodegradation in a lake. Environmental Science and 

Technology 32 (22), 3449–3456. 

Buser, H.R., Poiger, T., Müller, M.D., 1999. Occurrence and 

environmental behavior of the chiral pharmaceutical drug 

ibuprofen in surface waters and in wastewater. 


Carballa, M., Omil, F., Lema, J.M., Lllompart, M., Jares, C.G., 

Rodriguez, I., Gomez, M., Ternes, T., 2004. Behaviour of 

pharmaceuticals, cosmetics and hormones in a sewage 

treatment plant. Water Research 38, 2918–2926. 

Clara, M., Strenn, B., Kreuzinger, N., 2004. Carbamazepine as 

a possible anthropogenic marker in the aquatic environment: 

investigations on the behaviour of carbamazepine in 

wastewater treatment and during groundwater infiltration. 

Water Research 38, 947–954. 

Clara, M., Kreuzinger, N., Strenn, B., Gans, O., Kroiss, H., 2005. The 

solids retention time – a suitable design parameter to evaluate 

the capacity of wastewater treatment plants to remove 

micropollutants. Water Research 39, 97–106. 



Environmental Health Perspectives 107 (Suppl. 6) December 

1999.

Dillon, P., Page, D., Vanderzalm, J., Pavelic, P., Toze, S., Bekele, E., 

Sidhu, J., Prommer, H., Higginson, S., Regel, R., Rinck- 

Pfeiffer, S., Purdie, M., Pitman, C., Wintgens, T., 2008. A critical 

evaluation of combined engineered and aquifer treatment 

systems in water recycling. Water Science and Technology 57 

(5), 753–762. 

Drewes, J.E., Fox, P., 1999. Fate of natural organic matter (NOM) 

during groundwater recharge using reclaimed water. Water 


Drewes, J.E., Shore, L.S., 2001. Concerns about pharmaceuticals in 

water reuse, groundwater recharge, and animal waste. In: 

American Chemical Society (Ed.), Pharmaceuticals and 

personal care products in the environment. Symp. Ser, vol. 

791, pp. 206–228. 

Drewes, J.E., Heberer, T., Reddersen, K., 2002. Fate of 

pharmaceuticals during indirect potable reuse. Water Science 

& Technology 46 (3), 73–80. 

Drewes, E.J., Heberer, T., Rauch, T., Reddersen, K., 2003. Fate of 

pharmaceuticals during groundwater recharge. Ground Water 

Monitoring and Remediation 23 (3), 64–72. 

Drillia, P., Stamatelatou, K., Lyberatos, G., 2003. Mobility of 

Pharmaceuticals in Soil, 8th Conference on Environmental 

Science and Technology, Lemnos, Greece. 

Focazio, M.J., Kolpin, D.W., Barnes, K.K., Furlong, E.T., Meyer, M.T., 

Zaugg, S.D., Barber, L.B., Thurman, M.E., 2008. A national 

reconnaissance for pharmaceuticals and other organic 

wastewater contaminants in the United States–II untreated 

drinking water sources. Science of the Total Environment 402 

(2–3), 201–216. 

Fries, E., Püttmann, W., 2003. Monitoring of the three 

organophosphate esters TBP, TCEP and TBEP in river water 

and ground water (Oder, Germany). Journal of Environmental 

Monitoring 5 (2), 346–352. 

Grünheid, S., Amy, G., Jekel, M., 2005. Removal of bulk dissolved 

organic carbon (DOC) and trace organic compounds by bank 

filtration and artificial recharge. Water Research 39, 3219–3228. 

Heberer, T., Stan, H.-J., 1997. Determination of clofibric acid and 

N-(phenylsulphonyl)-sarcosine in sewage, river and drinking 

water. International Journal of Environmental Analytical 

Chemistry 67, 113–124. 

Heberer, T., Verstraeten, I.M., Meyer, M.T., Mechlinski, A., 

Reddersen, K., 2001. Occurrence and fate of pharmaceuticals 

during bank filtration–preliminary results from investigations in 

Germany and the United States. Water Resources Update 20, 4–17. 

Heberer, T., 2002. Occurrence, fate and removal of 

pharmaceutical residues in the aquatic environment: a review 

of recent research data. Toxicology Letters 131, 5–17. 

Heberer, T., Mechlinski, A., Fanck, B., 2003. NASRI – Occurrence 

and Fate of Pharmaceuticals During Bank Filtration, 

Conference Wasser Berlin, KompetenzZentrum Wasser Berlin, 

pp. 41–49. 

Horvath, R.S., 1973. Enhancement of co-metabolism of 

chlorobenzoates by the co-substrate enrichment technique. 

Applied Microbiology 25 (6), 961–963. 

Hua, J., An, P., Winter, J., Gallert, C., 2003. Elimination of COD, 

microorganisms and pharmaceuticals from sewage by 

trickling through sandy soil below leaking sewers. Water 

Research 37, 4395–4404. 

Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S. 

D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, 

hormones, and other organic wastewater contaminants in U. 

S. streams, 1999–2000: a national reconnaissance. 

Environmental Science & Technology 36, 1202–1211. 

Leenheer, J.A., Croue, J.P., Benjamin, M., Korshin, G.V., Hwang, C.., 

Bruchet, A., Aiken, G.R., 2000. Comprehensive isolation of 

natural organic matter from water for spectral characterizations 

and reactivity testing. In: Barett, S.E., Krasner, S.W., Amy, G.L. 


(Eds.), Natural Organic Matter and Disinfection By-Products, vol. 

761. ACS Symposium Series, pp. 68–83. 

Margesin, R., Schinner, F., 2001. Bioremediation (natural 

attenuation and biostimulation) of diesel-oil-contaminated 

soil in an alpine glacier skiing area. Applied and 

Environmental Microbiology 67 (7), 2127–3133. 

McCarty, P.L., Reinhard, M., Rittmann, B.E., 1981. Trace organics in 

groundwater. Environmental Science and Technology 15 (1), 40–51. 

Mersmann, P., Scheytt, T., Heberer, T., 2002. Column experiments 

on the transport of pharmaceutically active compounds in the 

saturated zone. Acta Hydrochimica et Hydrobiologica 30 (5–6), 

275–284 (in German). 

Möhle, E., Kempter, C., Kern, A., Metzger, J.W., 1999. 

Examination of the degradation of drugs in municipal sewage 

plants using liquid chromatography-electrospray mass 

spectrometry. Acta Hydrochimica et Hydrobiologica 27 (6), 

430–436 (in German). 

Montgomery-Brown, J., Drewes, J.E., Fox, P., Reinhard, M., 2003. 

Behavior of alkylphenol polyethoxlate metabolites during soil 

aquifer treatment. Water Research 37, 3672–3681. 

Preuß, G., Nehrkorn, A., 1996. Succession of microbial 

communities during bank filtration and artificial groundwater 

recharge Artificial Recharge of Groundwater. Proceedings of 

an International Symposium, A.-L. Kivimäki and T. Suokko, 

eds., Helsinki, Finland, June 3–5, 1996, Nordic Hydrological 

Programme, pp. 215–221. 

Rauch, T., Drewes, J.E., 2004. Assessing the removal potential of 

soil-aquifer treatment systems for bulk organic matter. Water 


Rauch, T., Drewes, J.E., 2005. A novel approach for quantifying 

bulk organic carbon removal in groundwater recharge 

systems. Journal of Environmental Engineering 131 (6), 

909–923. 

Reddersen, K., Heberer, T., 2003. Multi-compound methods for 

the detection of pharmaceutical residues in various waters 

applying solid phase extraction (SPE) and gas chromatography 

with mass spectrometric (GC-MS) detection. Journal of 

Separation Sciences 26, 1443–1450. 

Rittmann, B.E., McCarty, P.L., 2001. Environmental Biotechnology. 

Principles and Applications. McGraw-Hill, Boston. 

Scherrer, R.A., Howard, S.M., 1977. Use of distribution coefficients 

in quantitative structure-activity relationships. Journal of 

Medical Chemistry 20 (1), 53–58. 

Schmidt, C.K., Lange, F.T., Brauch, H.-J., 2004. Assessing the 

Impact of Different Redox Conditions and Residence Times on 

the Fate of Organic Micropollutants during Riverbank 

Filtration, Fourth International Conference on 

Pharmaceuticals and Endocrine Disrupting Chemicals in 

Water, October 13–15, Minneapolis, Minnesota. 

Schwab, B.W., Hayes, E.P., Fiori, J.M., Mastrocco, F.J., Roden, N.M., 

Cragin, D., Meyerhoff, R.D., D’aco, V.J., Anderson, P.D., 2005. 

Human pharmaceuticals in US surface waters: a human 

health risk assessment. Regulatory Toxicology and 

Pharmacology 42 (3), 296–312. 

Sedlak, D.L., Pinkston, K.E., 2001. Factors affecting the 

concentrations of pharmaceuticals released to the aquatic 

environment. Journal of Contemporary Water Research and 

Education 120, 56–64. September. 

Semprini, L., 1997. Strategies for the aerobic co-metabolism of 

chlorinated solvents. Current Opinion in Biotechnology 8 (3), 

296–308. 

Simoni, S.F., Schäfer, A., Harms, H., Zehnder, A.J.B., 2001. Factors 

affecting mass transfer limited biodegradation in saturated 

porous media. Journal of Contaminant Hydrology 50, 99–120. 

Stratton, R.G., Namkung, E., Rittmann, B.E., 1983. Secondary 

utilization of trace organic by biofilms on porous media. 

Journal American Water Works Association 75, 463–469.

460 

Szewzyk, U., Kalmbach, S., Manz, W., La˚ ngmark, J., and 

Stenström, T.-A., 1998. Artificial groundwater as the future 

water supply of greater Stockholm IV: Phylogenetic 

diversity of the unsaturated zones of experimental sand 

filter columns. Artificial recharge of groundwater. 

Proceedings of the third international symposium on 

artificial recharge of groundwater–Tisar 98, J.H. Peters, ed., 

Amsterdam, Netherlands, 21–25 September 1998, 

Rotterdam, pp. 391–393. 

Toride, N., Leij, F.J., van Genuchten, M.T., 1995. The CXTFIT Code 

for Estimating Transport Parameters from Laboratory or Field 


Tracer Experiments. Version 2.0. U.S. Salinity Laboratory Res. 

Rep. 137, U.S. Salinity, Riverside, CA. 

Van der Meer, J.R., Roelofsen, W., Schraa, G., Zehnder, A.J.B., 1987. 

Degradation of low concentrations of dichlorobezenes and 1,2,4thrichlorobenzene 

by Pseudomonas sp. strain P51 in nonsterile 

soil columns. FEMS Microbiology Letters 45 (6), 333–341. 

Zwiener, C., Frimmel, F.H., 2003. Short-term tests with a pilot 

sewage plant and biofilm reactors for the biological 

degradation of the pharmaceutical compounds clofibric acid, 

ibuprofen, and diclofenac. Science of the Total Environment 

309, 201–211.

Toxicological relevance of emerging contaminants for 

drinking water quality 

Merijn Schriks a, *, Minne B. Heringa a , Margaretha M.E. van der Kooi a , Pim de Voogt a,b , 

Annemarie P. van Wezel a 

a KWR Watercycle Research Institute, P.O. Box 1072, 3430 BB Nieuwegein, The Netherlands 

b Institute for Biodiversity and Ecosystem Dynamics, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV, 

Amsterdam, The Netherlands 

article info 







Keywords: 


Drinking water 

(Provisional) drinking water 

guideline value 

Threshold of Toxicological 

Concern (TTC) 


abstract 

Due to anthropogenic activities, freshwater systems worldwide 

are confronted with thousands of compounds. In the 

European Union, for example, there are more than 100 000 

registered chemicals (EINECS), of which 30 000–70 000 are in 

daily use. About 300 million tons of synthetic compounds 

annually used in industrial and consumer products, partially 

find their way to natural waters (Schwarzenbach et al., 2006). 

A major contribution to chemical contamination originates 

from wastewater discharges that impact surface water quality 




* Corresponding author. Tel.: þ31 30 6069564. 

E-mail address: merijn.schriks@kwrwater.nl (M. Schriks). 


doi:10.1016/j.watres.2009.08.023 

The detection of many new compounds in surface water, groundwater and drinking water 

raises considerable public concern, especially when human health based guideline values 

are not available it is questioned if detected concentrations affect human health. In an 

attempt to address this question, we derived provisional drinking water guideline values 

for a selection of 50 emerging contaminants relevant for drinking water and the water 

cycle. For only 10 contaminants, statutory guideline values were available. Provisional 

drinking water guideline values were based upon toxicological literature data. The 

maximum concentration levels reported in surface waters, groundwater and/or drinking 

water were compared to the (provisional) guideline values of the contaminants thus 

obtained, and expressed as Benchmark Quotient (BQ) values. We focused on occurrence 

data in the downstream parts of the Rhine and Meuse river basins. The results show that 

for the majority of compounds a substantial margin of safety exists between the maximum 

concentration in surface water, groundwater and/or drinking water and the (provisional) 

guideline value. The present assessment therefore supports the conclusion that the 

majority of the compounds evaluated pose individually no appreciable concern to human 

health. 


with incompletely removed organic contaminants (Kolpin 

et al., 2004; Snyder et al., 2001). Additional contamination 

comes from diffuse agricultural activities, in which over 140 

million tons of fertilizers and several million tons of pesticides 

are applied each year, and from atmospheric deposition. Such 

contamination can become an increasing problem for 

drinking water supplies, especially since the European REACH 

legislation may drive producers to develop newly designed 

less lipophilic/bioaccumulative chemicals that will be inherently 

more difficult to remove by traditional drinking water 

treatment techniques.

462 

Recently, Loos et al. (2009) presented an EU-wide monitoring 

study on 35 organic compounds in European river 

waters in concentrations up to 40 mg/L. In addition, we have 

shown the occurrence of emerging polar contaminants, 

such as benzotriazoles and metabolites of illicit drugs (e.g. 

benzoylecgonine, desalkylflurazepam and 9-carbonic acid-d- 

9-THC) in groundwaters and surface waters in the 

Netherlands (Hogenboom et al., 2009; Van Leerdam 

et al., 2009). 

Many of these emerging contaminants raise considerable 

toxicological and public concern, especially when human 

health based guideline values are unavailable. At present, 

the World Health Organization (WHO) and the US Environmental 

Protection Agency (US EPA) have derived approximately 

125 statutory guideline values for contaminants in 

drinking water (Cotruvo, 1988; US EPA, 2006; WHO, 2006). 

However, the potential health effects of many emerging 

contaminants present in the water cycle and the potential 

human health concern associated with direct water ingestion 

have not been evaluated and statutory standards are 

not available. Therefore, we proposed earlier to assess the 

potential human health concern of unknown non-genotoxic 

compounds (lacking structural alerts that raise concern for 

potential genotoxicity) by comparing the environmental or 

drinking water concentration to a TTC (Threshold of Toxicological 

Concern) derived target value (Mons et al., 2008). 

The TTC concept was developed in the context of food 

safety to obtain a first clue on risks of unregulated chemicals 

present at low levels. The TTC thus derived is based on the 

molecular structure of the chemical involved and its related 

mode of toxic action (Munro et al., 1996). Assuming a daily 

intake of 2 l/day of drinking water, and a maximum contribution 

of 10% from drinking water to the total exposure – 

both of which are standard assumptions for deriving 

drinking water-quality guidelines (WHO, 2006) –TTCbased 

target values proposed for drinking water are 0.1 mg/L for 

non-genotoxic compounds and 0.01 mg/L for genotoxicants 

(Mons et al., 2008). This value is based on a TTC level of 

1.5 mg/person/d (0.15 mg/person/d for substances containing 

structural alerts that raise concern for potential genotoxicity 

with an acceptable lifetime cancer risk of 10 6 ) for 

compounds in food (Kroes et al., 2004). Since the TTC based 

value is rather conservative, an over-estimation of the 

actual risk may be the result. For emerging contaminants, 

a more profound human health based assessment may 

therefore be very valuable. The objectives of the present 

study were twofold. The first objective was to collect existing 

drinking water guideline values for a selection of 50 

emerging contaminants relevant for the water cycle. If 

existing guideline values were not available, provisional 

guideline values were derived with the aid of relevant 

toxicological literature data. The second aim was to 

compare the maximum concentration levels reported in 

surface water, groundwater and/or drinking water to the 

(provisional) guideline values of the contaminants thus 

obtained, and express this as a Benchmark Quotient (BQ) 

value (further abbreviated as ‘‘BQ value’’). The present study 

does not attempt to quantify mixture interactions, since for 

compounds with an unknown mode of action there is no 

accepted methodology for such an assessment. 



The toxicological assessment of the compounds presented in 

this paper comprises of a tiered approach in five consecutive 

steps (Fig. 1). First, the compounds to be assessed were 

selected. Second, n-octanol–water partition coefficients 

(log K ow) were obtained and compounds with a log K ow > 3 

were excluded from further assessment. This log K ow cut off 

value is applied as a default threshold; for compounds with 

a log K ow above 3 it is less likely that they pass drinking water 

treatment plants (Westerhoff et al., 2005). Third, if available, 

statutory drinking water guideline values were obtained from 

websites of competent authorities; else provisional guideline 

values were derived with the aid of toxicological data relevant 

for humans as reported in literature. Fourth, measured 

maximum surface water, groundwater and/or drinking water 

concentrations were obtained from various sources and 

compared to (provisional) guideline values. Finally, a BQ value 

was calculated from the maximum concentrations reported 

and the (provisional) drinking water guideline values 

obtained. These steps are described in more detail below. 

2.1. Selection of compounds for assessment 

A priority list representing a broad range of chemical classes 

was formulated with more than 100 compounds of interest. 

The arguments for inclusion were (i) questions related to 

toxicity posed by Dutch drinking water companies, (ii) potential 

low removal efficiency during drinking water production, 

(iii) appearance in recent literature and (iv) occurrence in 

surface waters, groundwaters and drinking water as determined 

by ourselves and others in various screening studies. 

2.2. Collection of compound-specific data 

2.2.1. n-Octanol–water partition coefficients (log K ow) 

All log Kow values were obtained with the aid of the estimation 

program KOWWIN (US EPA, v1.67). An exception was 

made for perfluoroctane sulfonate (PFOS) and perfluoroctanoic 

acid (PFOA), for which accurate log K ow values 

cannot be calculated with estimation software. For these 

compounds the log Kow values were obtained from a database 

(Krop and de Voogt, 2008). 

2.2.2. Toxicological data 

As illustrated in Fig. 1 (step 3), the first step was to obtain existing 

statutory drinking water guideline values from e.g. the US EPA 

(URL1) and the WHO (URL2). If not available, the second step was 

to obtain an established (by an (inter)national organization) 

Tolerable Daily Intake (TDI), Acceptable Daily Intake (ADI) or 

Reference Dose (RfD) and subsequently a provisional drinking 

water guideline value was derived as further described in 

Section 2.3. If not available, in a third step toxicity data collection 

focused primarily on established (by an (inter)national organization) 

lowest/no observed (adverse) effect levels (LO/NO(A)ELs) 

and subsequently a TDI was calculated as further described in 

Section 2.3. Finally, in a fourth step, miscellaneous toxicological 

information was collected and a TDI was calculated accordingly. 

In the case of insufficient human relevant toxicological data the

compound of interest was not further evaluated and removed 

from the list. To facilitate the interpretation for which 

compounds the toxicity database is strong and less strong, all 

compounds were categorized (Table 2); (A) representing 

compounds with a statutory drinking water guideline value, (B) 

representing compounds with an established TDI, ADI or RfD, 

(C) representing compounds for which the TDI was calculated 

with an established LO(A)EL or NO(A)EL and (D) representing 

compounds for which the TDI was calculated with miscellaneous 

toxicological information. 

TDIs, ADIs, RfDs and/or other chronic toxicity data were 

sourced from peer-reviewed scientific papers and from other 

sources such as ‘‘grey literature’’. In addition, a literature 

search was performed on the internet and/or toxicological 

relevant data were obtained from the US EPA IRIS database, 

the European Chemicals Bureau (ECB), the Organization for 

Economic Cooperation and Development (OECD), the Dutch 

National Institute for Health and the Environment (RIVM), the 

US Food and Drug Administration (US FDA), the Joint Meeting 

FAO/WHO Meetings on Pesticide Residues (JMPR), the Dutch 

Expert Committee for Occupational Standards (DECOS), the 

Dutch board for authorization of plant protection products 

and biocides (CTGB), the Scientific Committee on Occupational 

Exposure Limits (SCOEL), the US National Toxicology 

Program (NTP), the Joint FAO/WHO Committee on Food 

Additives (JECFA), the Food and Agriculture Organization 

(FAO), the Danish veterinary and food administration, the 

European Union (EU), the US National Research Council (NRC), 

the EFSA scientific panel on contaminants in the food chain 

(CONTAM) and the Hazardous Substance Data Bank (HSDB). 


Fig. 1 – Flow diagram of the assessment conducted in the present study. Abbreviations: log K ow, n-octanol–water partition 

coefficient; GLV, drinking water guideline value; ADI, acceptable daily intake; RfD, reference dose; TDI, tolerable daily intake; 

LO/NO(A)EL, lowest/no observed (adverse) effect level; numbers correspond to consecutive steps as described Section 2. 

Compound categories: see Section 2.2.2. 

2.2.3. Occurrence data 

Collection of occurrence data focused primarily on maximum 

concentrations of compounds measured in the downstream 

parts of the Rhine and Meuse river basins during the past 

decade. If not available, maximum concentrations in other 

surface waters and/or groundwaters were sought. The 

primary source of occurrence data of compounds in surface 

waters were the annual reports of the Dutch Association of 

River Water Companies (RIWA) and the German Association 

of River Water Companies (ARW). Occurrence data of 

compounds in drinking water were obtained from the REWAB 

data set (restricted water-quality data from the Dutch water 

companies). If not available, alternative resources were 

searched such as ‘‘grey literature’’, unpublished data, or 

publicly available sources such as RIVM, the Dutch ministry of 

transport, public works and water management (Rijkswaterstaat) 

and the WHO. Additional data on the occurrence of 

compounds was obtained from peer-reviewed scientific 

papers and an internet based literature search. 

2.3. Derivation of provisional drinking water guideline 

values 

A drinking water guideline value represents the concentration 

of a constituent that does not exceed tolerable risk to the 

health of the consumer over a lifetime (WHO, 2006). In some 

cases, an odour-threshold value may be much lower than the 

health based guideline value. To calculate a provisional health 

based guideline value, the general methodology was applied

464 

Table 1 – List of compounds assessed in the present study. 

No. Compound CAS log K ow No. Compound CAS Log K ow 

1 1,4-Dioxane 2 

2 2,6-Dichlorobenzamide (BAM) 3 

3 4-Methylbenzenesulfonamide 

(p-toluenesulfonamide, 

4-tolylsulfonamide) 2 

4 Acetylsalicylate 

(aspirin, acetyl salicylic acid) 6 

5 Alpha-amino-3-hydroxy- 

5-methyl-4-isoxazole propionic 

acid (AMPA) 3 

6 Amidotrizoic acid (diatrizoic acid) 1 

7 Bentazone 3 

8 Benzene 2 

9 Benzothiazole 2 

10 Benzotriazole (1H-benzotriazole) 2 

11 Bis(2-chloroisopropyl)ether (BCIPE) 2 

12 Carbamazepine 6 

13 Carbendazim 3 

14 Chloridazon (pyrazon) 3 

15 Clofibric acid 6 

16 Dichlorophenoxyacetic acid (2,4-D) 3 

17 Diethyl phthalate 2 

18 Diethyl toluamide (DEET) 3 

19 Diethylamine (DEA) 2 

20 Diethylene glycol dimethyl ether 

(diglyme, bis(2-methoxy ethyl)ester)) 2 

21 Diethylene triamine penta acetic 

acid (DTPA) 2 

22 Dimethenamid 3 

23 Dimethylamine (DMA) 2 

24 Diuron 3 

25 Ethyl tert-butyl ether (ETBE) 4 

123-91-1 0.3 a 

2008-58-4 0.8 a 

70-55-3 0.9 b 

50-78-2 1.2 a 

77521-29-0 2.2 b 

117-96-4 1.4 b 

25057-89-0 2.3 a 

71-43-2 2.1 a 

95-16-9 2.0 a 

95-14-7 1.4 a 

108-60-1 2.5 a 

298-46-4 2.5 a 

10605-21-7 1.5 a 

1698-60-8 1.1 a 

882-09-7 2.6 a 

94-75-7 2.8 a 

84-66-2 2.4 a 

134-62-3 2.2 a 

109-89-7 0.6 a 

111-96-6 0.4 a 

as described by Van Leeuwen (2000) and the WHO (2006). For 

compounds without a statutory drinking water guideline 

value, first the Tolerable Daily Intake (TDI) was determined. 

The point of departure (POD) for calculating the TDI was 

mostly a chronic LO(A)EL, NO(A)EL, benchmark dose level, 

maximum tolerated dose (MTD) or lowest effective safe dose. 

In case only inhalatory toxicity data could be found, a routeto-route 

extrapolation was carried out according to toxicological 

methods as described by Stokinger and Woodward 

(1958). An appropriate safety factor to extrapolate between 

species (inter-species differences), inter-individual differences 

(intraspecies differences), exposure route/duration and 

quality of the data was utilized as part of the TDI calculation 

(Van Leeuwen and Vermeire, 2007). Secondly, a drinking water 

equivalent level (DWEL) was calculated by multiplying the TDI 


67-43-6 4.2 b 

87674-68-8 2.2 a 

124-40-3 0.4 a 

330-54-1 2.7 a 

637-92-3 1.9 b 

26 Ethylenediamine tetra 

acetic acid (EDTA) 2 

27 Glyphosate 3 

28 Imidacloprid 3 

29 Iohexol 1 

30 Iomeprol (iomeron) 1 

31 Iopamidol 1 

32 Iopromide 1 

33 Isoproturon 3 

34 Methyl tert-butyl ether (MTBE) 4 

35 Metoprolol 6 

36 n-Butylbenzenesulphonamide 2 

37 Nicosulfuron 3 

38 n-Nitrosodimethylamine 

(NDMA) 2 

39 p,p 0 -Sulfonyldiphenol 2 

40 Perfluoroctanesulfonate 

(PFOS) (potassium-salt) 5 

41 Perfluorooctanoic acid 

(PFOA) 5 

42 Phenazone 6 

43 Simazine 3 

44 Sulfamethoxazole 6 

45 Tolyltriazole 2 

46 Trichloroethene 2 

47 Triethylphosphate 

(ethylphosphate) (TEP) 2 

48 Triphenylphosphine 

oxide (TPPO) 2 

49 Tris(2-chloroethyl) 

phosphate (TCEP) 2 

50 Urotropine (methenamine, 

hexamine) 2 

60-00-4 3.9 b 

1071-83-6 4.0 a 

138261-41-3 0.6 a 

66108-95-0 3.1 a 

78649-41-9 1.4 b 

62883-00-5 2.4 a 

73334-07-3 2.1 a 

34123-59-6 2.9 a 

1634-04-4 0.9 a 

37350-58-6 1.9 a 

3622-84-2 2.3 b 

111991-09-4 1.2 b 

62-75-9 0.6 a 

80-09-1 1.7 b 

2795-39-3 1.08 c 

335-67-1 2.8 c 

60-80-0 0.4 a 

122-34-9 2.2 a 

723-46-6 0.9 a 

29385-43-1 NA 

79-01-6 2.4 a 

78-40-0 0.8 a 

791-28-6 2.8 a 

115-96-8 1.4 a 

100-97-0 4.2 b 

NA not available. Chemical categories: 1 Iodinated contrast media; 2 Miscellaneous organic compounds; 3 Miscellaneous pesticides; 4 Oxygenated 

gasoline additives; 5 Perfluorinated organic compounds; 6 Pharmaceuticals. 

a log K ow values were derived from US EPA’s KOWWIN experimental database. 

b log Kow values were calculated with the aid of US EPA’s KOWWIN. 

c log Kow values were obtained from Krop and de Voogt (2008). 

by a typical average body weight of 70 kg and division by 

a daily water consumption of 2 l. Finally, to account for the 

fraction of the TDI allocated to drinking water, the DWEL was 

multiplied by an allocation factor to give the provisional 

guideline value. In most cases, when there was insufficient 

exposure information to derive chemical-specific allocation 

factors, a default allocation factor of 10% was used. 

2.4. Evaluation of water-quality data in the context of 

human health 

Of each compound the maximum concentration level reported 

in surface waters, groundwater and/or drinking water was 

compared to its (provisional) guideline value and was 

expressed as a BQ value (concentration in water divided by

Table 2 – Parameters used for derivation of (provisional) drinking water guideline values. 

Compound Point of departure (POD) Category a SUF b TDI, ADI or RfD 

(mg/kg bw/d) 

1,4-Dioxane POD is an oral slope factor as derived by US EPA (1988a) of 0.011 per mg/kg bw/d from a 110-week study 

in rats (NCI, 1978). 

2,6-Dichlorobenzamide (BAM) POD is a NOAEL of 4.5 mg/kg bw/d for decreased body weight in both sexes and increased liver weight 

in males as derived by the Danish EPA (2004) from a study in which dogs were fed BAM in the diet for 2 

years (Wilson and Thorpe, 1971), with an uncertainty factor of 300 (100 for inter- and intraspecies 

variation and 3 for uncertainties in the dataset). 

4-Methylbenzenesulfonamide 


4-tolylsulfonamide) 

Acetylsalicylate (aspirin, acetyl 

salicylic acid) 

Alpha-amino-3-hydroxy-5-methyl- 

4-isoxazole propionic acid (AMPA) 

POD is a NOAEL of 300 mg/kg bw/d for reproductive effects (decrease lactation index and litter weight at 

birth) as derived from a GLP compliant study in which rats were administered 4methylbenzenesulfonamide 

via oral gavage for 42 days (OECD/SIDS, 1994), with an uncertainty factor 

of 400 (100 for inter- and intraspecies variation and 4 for extrapolation to chronic exposure). 

POD is a LOEL of 10 mg/person from a human study as derived by the Dutch National Institute for 

Health and the Environment (RIVM) (Versteegh et al., 2007). 

POD is a NOAEL of 32 mg/kg bw/d as derived by the WHO (2005a) from a 26-month toxicity study in rats 

(study reference unknown). 

Amidotrizoic acid (diatrizoic acid) POD is the highest therapeutic dose of 50 mg/person/d (0.71 mg/kg bw/d) as derived by the Dutch 

National Institute for Health and the Environment (RIVM) (Versteegh et al., 2007). 

Bentazone POD is a NOAEL of 9 mg/kg bw/d as derived by the WHO (1998) from a 2-year dietary toxicity study in 

rats (study reference unknown). 

Benzene POD is a risk estimate as derived by the WHO (2003a) from a 2-year gavage study in rats and mice (NTP, 

1986). 

Benzothiazole POD is a NOEL of 5.1 mg/kg bw/d as derived by the WHO/JECFA (2003) from a study in which rats were 

administered benzothiazole in the diet for 90 days (Morgareidge, 1971). Daily observations revealed no 

treatment related effects in histopathological/haematological parameters, body weight, food 

consumption and liver/kidney weights. An uncertainty factor of 200 was applied (100 for inter- and 

intraspecies variation and 2 for extrapolation to chronic exposure). 

Benzotriazole (1H-benzotriazole) POD is a LOAEL of 295 mg/kg bw/d for histological changes in the liver, decreased body weight gain and 

inflammation of the prostate/uterus as derived by the Dutch Expert Committee for Occupational 

Standards (DECOS, 2000) from a study in which rats were administered benzotriazole in the diet for 78 

weeks (BIBRA Toxicology International, 1995), with an uncertainty factor of 1000 (100 for inter- and 

intraspecies variation and 10 for extrapolation of a LOAEL to a NOAEL). 

Bis(chloroisopropyl)ether (BCIPE) POD is a NOAEL of 35.8 mg/kg bw/d as derived by the US EPA (1989) from a 24-month chronic toxicity 

study in mice (Mitsumori et al., 1979). 

Carbamazepine POD is a maximum tolerated dose (MTD) of 250 mg/kg bw/d as derived by Snyder et al. (2008) from a 2year 

study in rats showing evidence of carcinogenicity (Singh et al., 2005). 

Carbendazim POD is a NOAEL of 2.5 mg/kg bw/d as derived by the WHO/JECFA (1995) from a 2-year study in dogs 

(Sherman, 1972). 

Chloridazon (pyrazon) POD is a NOAEL of 5.4 mg/kg bw/d for adverse histopathological/haematological changes, decreased 

food intake, lower body weight gain and higher organ weights (liver, kidney, thyroid gland) as derived 

from a study in which rats were orally administered chloridazon (method of administration 

unspecified) for 7 weeks (ECB, 2000b), with an uncertainty factor of 100 (inter- and intraspecies 

variation). 

Clofibric acid POD is a LOEL of 1 mg/kg bw/d as derived by the Dutch National Institute for Health and the 

Environment (RIVM) (Versteegh et al., 2007) from an 8-week oral study in humans (Larsen et al., 1994). 

(Provisional) 

guideline 

value (mg/L) 

B NA NA 30 d 

C 300 0.015 52.5 

D 400 0.75 2600 

B 20 0.007 25 

A 100 0.3 900 c 

B 10 NA 250 000 

A 100 0.1 300 c 

A NA NA 10 c,d 

C 200 0.026 90 

C 1000 0.295 1000 

B 1000 0.04 140 

B NA 0.00034 1 

B 100 0.03 105 

D 100 0.054 189 

B 100 0.01 30 


water research 44 (2010) 461–476 465



(mg/kg bw/d) 

Dichlorophenoxyacetic acid (2,4-D) POD is a NOAEL of 1 mg/kg bw/d as derived by the WHO (2003b) from a 1-year study of toxicity in dogs 

and a 2-year study of toxicity and carcinogenicity in rats (study reference unknown). 

Diethyl phthalate POD is a NOAEL of 750 mg/kg bw/d as derived by US EPA (1988b) from a 16-week toxicity study in rats 

(Brown et al., 1978). 

Diethyl toluamide (DEET) POD is a NOEL of 100 mg/kg/bw/d based on clinical signs, reduced haemoglobin/haematocrit levels and 

histological changes in liver, lymph nodes and uterus as derived by the California Environmental 

Protection Agency (California EPA, 2000) from a study in which beagle dogs were orally administered 

DEET (gelatin capsules) for 1 year (Goldenthal, 1994), with an uncertainty factor of 56 (14 for inter- and 

intraspecies variation and 4 for extrapolation to chronic exposure). 

Diethylamine (DEA) POD is a LOAEL of 75 mg/m 3 for reduced mean body weights and adverse histopathological effects 

(lesions of the nasal mucosa) as derived by the scientific committee on occupational exposure limits 

(SCOEL, 2002) from a study in which rats were exposed to DEA via the inhalatory route for 24 weeks 

(6.5 h/d, 5d/wk) (Lynch et al., 1986), with an uncertainty factor of 50 (5 for the absence of human data 

and a NOAEL and 10 for route-to-route extrapolation uncertainties). 

Diethylene glycol dimethyl ether 

(diglyme, bis(2-methoxy ethyl)ester)) 

Diethylene triamine penta acetic acid 

(DTPA) 

POD is a NOAEL 25 mg/kg bw/d for developmental effects (adversely affected implants per liter and 

decreased weight gain) as derived by the WHO (2002) from a study in which rabbits were administered 

diglyme via oral gavage for 13 days (NTP, 1987), with an uncertainty factor of 500 (100 for inter- and 

intraspecies variation and 5 for uncertainties in the dataset). 

POD is a NOAEL of 100 mg/kg bw/d for developmental effects (increased fetal deformations) as derived 

from a study according to OECD guideline 414 in which rats were administered DTPA (in its sodium 

form) via oral gavage during day 6–15 of pregnancy (ECB, 2000c), with an uncertainty factor of 1000 (100 

for inter- and intraspecies variation and 10 for extrapolation to chronic exposure). 

Dimethenamid POD is a NOAEL of 7 mg/kg bw/d as derived by the WHO/JMPR (2005) from a 24-month study in rats 

given diets containing racemic dimethenamid (study reference unknown). 

Dimethylamine (DMA) POD is a LOAEL of 19 mg/m 3 for concentration-related lesions in the respiratory/olfactory mucosa as 

derived by the scientific committee on occupational exposure limits (SCOEL, 1991) from a study in 

which rats and mice were exposed to DMA via the inhalatory route for 2 years (6 h/d, 5d/wk) (CIIT, 

1990), with an uncertainty factor of 50 (5 for the absence of human data and a NOAEL and 10 for routeto-route 

extrapolation uncertainties). 

Diuron POD is a NOEL of 0.625 mg/kg bw/d as derived by US EPA (1988c) from a 2-year feeding study in dogs 

(DuPont, 1964). 

Ethyl tert-butyl ether (ETBE) POD is a NOAEL of 500 ppm (29.1 mg/kg bw/d h ) for testes degeneration as derived from a study in which 

rats were exposed to ETBE via the inhalatory route for 13 weeks (Medinsky et al., 1999), with an 

uncertainty factor of 200 (100 for inter- and intraspecies variation and 2 for extrapolation to chronic 

exposure). 

Ethylenediamine tetra acetic Acid 

(EDTA) 

POD is a NOAEL of 250 mg/kg bw/d (190 mg/kg bw/d as the free acid) as derived by the WHO/JECFA 

(1973) from a 2-year toxicity study in rats (study reference unknown). 

Glyphosate POD is a NOAEL of 32 mg/kg bw/d as derived by the WHO (2005a) from a 26-month study of toxicity in 

rats fed technical-grade glyphosate (study reference unknown). 

Imidacloprid POD is a NOAEL of 5.7 mg/kg bw/d as derived by the WHO/JMPR (2001) from a 2-year study of toxicity 

and carcinogenicity in rats (study reference unknown). 

Iohexol POD is a safe dose of 75 g/person/d (1.07 g/kg bw/d) as derived by the Dutch National Institute for 


(Provisional) 

guideline 

value (mg/L) 

A 100 0.01 30 c 

B 1000 0.8 2800 

C 56 1.8 6250 

C 50 2.14 f 

750 

C 500 0.05 175 

D 1000 0.1 350 

B 100 0.07 245 

C 50 0.54 g 

190 

B 300 0.002 7 

D 200 0.15 525 e 

A 100 1.9 600 b,c 

A 100 0.3 900 c 

B 100 0.06 210 

B 10 NA 375 000 

466 


Iomeprol (iomeron) POD is a NOEL of 2 g Iodine/kg bw/d (equal to 4 g iomeprol/kg bw/d) for adverse effects on liver and 

kidney (non-lipid cytoplasmic vacuolization of hepatocytes and renal tubular epithelium cells) as 

derived from a study in which dogs were intravenously exposed to iomeprol for 28 days (Morisetti et al., 

1994), with an uncertainty factor of 2100 (35 for inter- and intraspecies, 10 for route-to-route 

extrapolation and 6 for extrapolation to chronic exposure). 

D 2100 1.9 6700 

Iopamidol POD is a safe dose of 83 g/person/d (1.19 g/kg bw/d) as derived by the Dutch National Institute for 


B 10 NA 415 000 

Iopromide POD is a safe dose of 50 g/person/d (0.71 g/kg bw/d) as derived by the Dutch National Institute for 


B 10 NA 250 000 

Isoproturon POD is a NOAEL of 3 mg/kg bw/d as derived by the WHO (2003c) from a 90-day study in dogs and a 2-year 

feeding study in rats (study reference unknown). 

A 1000 0.003 9 c 

Methyl tert-butyl ether (MTBE) POD is a NOAEL of 300 mg/kg bw/d as derived by the Dutch National Institute for Health and the 

Environment (RIVM) (Swartjes et al., 2004) from a 90-day oral toxicity study in rats (Robinson et al., 

1990). 

B 1000 0.3 9400 e 

Metoprolol POD is a LOEL of 100 mg/person/d (1.42 mg/kg bw/d) as derived by the Dutch National Institute for 


B 100 0.014 50 

n-Butylbenzenesulphonamide POD is a NOAEL of 50 mg/kg bw/d for adverse treatment related macro- or microscopic effects (liver 

enlargement, hepatocyte hypertrophy, thymic atrophy and lymphocytolysis) as derived from a study 

according to OECD guideline 407 (GLP compliant) in which rats were administered nbutylbenzenesulphonamide 

via oral gavage for 28 days (Proviron Fine Chemicals, 2003), with an 

uncertainty factor of 600 (100 for inter- and intraspecies differences, 6 for extrapolation to chronic 

exposure). 

D 600 0.083 292 

Nicosulfuron POD is a NOAEL of 199 mg/kg bw/d as derived from a 2-year study in rats (URL3). B 1000 0.2 700 

n-Nitrosodimethylamine (NDMA) POD is a tumor dose (TD05, dose level that causes 5% in increase in tumour incidence over the 

background) of 18 mg/kg bw/d as derived by the WHO (2008) from a detailed 2-year cancer dose– 

response study in rats (Peto et al., 1991a,b). 

A NA NA 0.1 c,d 

p,p 0 -Sulfonyldiphenol POD is a NOEL of 10 mg/kg bw/d for histopathological effects (hyperplasia of the mucosal epithelium, 

hypertrophy of hepatocytes), decreased food consumption/body weight gain and increased liver 

weights as derived from a study according to OECD guideline 421 in which male and female rats were 

administered p,p 0 -sulfonyldiphenol via oral gavage for respectively 45 days and from 14 days before 

mating to day 3 of lactation (Mitsubishi Chemical Safety Institute Ltd, date of study unknown), with an 

uncertainty factor of 600 (100 for inter- and intraspecies differences, 6 for extrapolation to chronic 

exposure). 

Perfluoroctane sulfonate (PFOS) POD is a NOAEL of 0.03 mg/kg bw/d as derived by the Scientific Panel on Contaminants in the Food 

Chain (CONTAM) (EFSA, 2008) from a 182-day study in Cynomolgus monkeys (Seacat et al., 2002). 

Perfluorooctanoic acid (PFOA) POD is a benchmark dose level (BMDL10) of 0.3 mg/kg bw/d as derived by the Scientific Panel on 

Contaminants in the Food Chain (CONTAM) (EFSA, 2008) from a number of studies in mice and male 

rats (study references unknown). 

Phenazone POD is a LOEL of 250 mg/person/day (3.57 mg/kg bw/d) as derived by the Dutch National Institute for 


Simazine POD is a NOAEL of 0.52 mg/kg bw/d as derived by the WHO (1996) from a 2-year combined chronic 

toxicity/oncogenicity study in rats (Ciba-Geigy, 1988; unpublished study submitted to WHO). 

Sulfamethoxazole POD is a NOAEL of 25 mg/kg/d as derived by Schwab et al. (2005) from a 60-week study in rats (Swarm 

et al., 1973). 

Tolyltriazole POD is a NOAEL of 150 mg/kg bw/d for observed mild apathy as derived from a study in which rats were 

administered tolyltriazole via oral gavage for 29 days (Benzotriazoles Coalition, 2001; ECB, 2000a), with 

an uncertainty factor of 600 (100 for inter- and intraspecies variation and 6 for extrapolation to chronic 

exposure). 

D 600 0.017 60 

B 200 0.00015 0.5 

B 200 0.0015 5.3 

B 100 0.036 125 

A 1000 0.52 2 c 

B 200 0.13 440 

D 600 0.25 875 


water research 44 (2010) 461–476 467



(mg/kg bw/d) 

Trichloroethene POD is a benchmark dose level (BMDL10) of 0.146 mg/kg bw/d as derived by the WHO (2005b)/Health 

Canada (2003) from a developmental toxicity study in rats (Dawson et al., 1993). 

Triethylphosphate (ethylphosphate) 

(TEP) 

POD is a NOEL of 335 mg/kg bw/d for fertility effects (effects on litter size) as derived from a study in 

which rats were administered TEP via the food for a unknown period (OECD/SIDS, 1998), with an 


exposure). 

Triphenylphosphine oxide (TPPO) POD is a NOAEL of 8 mg/kg bw/d for salivation, vomiting, diarrhea, histopathological (liver damage and 

skeletal muscle atrophy)/haematological (elevated GPT, GOT and alkaline phosphatase activities, 

reduced haemoglobin/haematocrit levels) parameters as derived from a study in which dogs were 

administered TPPO via the food for 3 months (ECB, 2000d), with an uncertainty factor of 1000 (100 for 

inter- and intraspecies variation and 10 for extrapolation to chronic exposure). 

Tris(2-chloroethyl)phosphate (TCEP) POD is a NOAEL of 22 mg/kg bw/d for increased relative liver and kidney weights as derived from 

a study in which rats were administered TCEP via oral gavage for 16 weeks (NTP, 1991), with an 


exposure and uncertainty in genotoxic potential). 

Urotropine (methenamine, hexamine) POD is a NOAEL of 15 mg/kg bw/d as derived by the WHO/JECFA (1974) from a teratogenicity study 

(exposure from the fourth to fifty-sixth day after mating) in dogs (Hurni and Ohder, 1973). 

(Provisional) 

guideline 

value (mg/L) 

A 100 0.0015 20 c 

D 600 0.56 1950 

D 1000 0.008 28 

D 1000 0.022 77 

B 100 0.15 500 

NA not available. 

a Categories: A) statutory drinking water guideline value available; B) established TDI, ADI or RfD available; C) TDI calculated with a established LO(A)EL or NO(A)EL; D) TDI calculated with 

miscellaneous toxicological information. 

b Uncertainty factors. 

c WHO drinking water guidelines (WHO, 2006). 

d Based on a specific cancer risk level of 10 5 . 

e The odour-threshold value for drinking water preparation is 15 mg/L for MTBE and w1 mg/L for ETBE (Swartjes et al., 2004; Van Wezel et al., 2009). 

f Using the Stokinger–Woodward approach (Stokinger and Woodward, 1958), a TDI of 150 mg/person/d (2.14 mg/kg bw/d) can be calculated from the 8-h threshold limit value (15 mg/m 3 ) assuming 

100% oral/inhalatory absorption and a 8-h total workshift ventilation of 10 m 3 . 

g Using the Stokinger–Woodward approach (Stokinger and Woodward, 1958), a TDI of 40 mg/kg/person/d (0.54 mg/kg bw/d) can be calculated from the 8-hour threshold limit value (3.8 mg/m 3 ) 

assuming 100% oral/inhalatory absorption and a 8-h total workshift ventilation of 10 m 3 . 

h A TDI of 0.15 mg/kg bw/day can be calculated from the NOAEL (500 ppm ¼ 29.1 mg/kg/bw/d) assuming 100% oral/inhalatory absorption, a specific lung retention of 25% and a minute ventilation 

volume for rat of 45 mL. 

468 


guideline value) in order (i) to provide perspective on what the 

occurrence of emerging contaminants might signify to human 

health and (ii) to help prioritize further investigations. A BQ 

value of 1 represents a (drinking) water concentration equal to 

the (provisional) guideline value. Compounds with a BQ value 

of 1 in drinking water may be of potential human health 

concern if the water were to be consumed over a lifetime 

period. Compounds with a BQ value 0.1 in drinking water 

were identified as those that may warrant further investigation; 

this is consistent with various US State and Federal 

practices (Toccalino, 2007). For compounds found in surface 

waters and groundwater the BQ value threshold to carry out 

an additional assessment was set at an arbitrary value of 0.2, 

since these source waters are purified in drinking water 

treatment plants which provides extra safety. Compounds in 

surface waters/groundwater or drinking water with a BQ value 

of 0.2 or 0.1 respectively, are presumed to present no 

appreciable concern to human health. 

3. Results 

3.1. Selection of compounds 

For only 50 compounds out of the original list, statutory 

guideline values or useful toxicity and occurrence data could 

be found. These compounds constitute the final list, which 

includes compounds from various groups such as iodinated 

contrast media, pharmaceuticals, oxygenated gasoline additives, 

perfluorinated organic compounds, miscellaneous 

organic compounds and pesticides (Table 1). Natural and 

synthetic steroid hormones such as 17b-estradiol, 17a-ethynylestradiol 

and estrone were not included in this assessment, 

as they are removed relatively easily in drinking water 

purification processes (Nghiem et al., 2004). 

3.2. (Provisional) drinking water guideline values 

For 10 compounds WHO statutory drinking water guideline 

values were available and these compounds were classified 

as category A. For the remaining 40 compounds a provisional 

guideline value was established with the aid of toxicological 

data. An established TDI, ADI or RfD was available for 22 

compounds (category B). In 7 cases when there was no TDI, 

ADI or RfD available, an established NO(A)EL or LO(A)EL was 

used to calculate a TDI and subsequently a provisional 

drinking water guideline value (category C). For the remaining 

11 compounds, miscellaneous toxicological data was 

used to calculate a TDI and subsequently a provisional 

drinking water guideline value (category D). As tabulated in 

Table 2, (provisional) guideline values ranged from 

0.0001 mg/L for NDMA to 415 mg/L for the iodinated contrast 

medium iopamidol. All iodinated contrast media had relatively 

high provisional guideline values, ranging from 6.7 mg/ 

L (iomeprol) to 415 mg/L (iopamidol). In the cases of MTBE 

and ETBE, the human health based guideline values were at 

least one order of magnitude higher than the corresponding 

odour-threshold based guideline values of respectively 15 mg/ 

L and w1 mg/L (Swartjes et al., 2004; Van Wezel et al., 2009). 

For two compounds (DMA and DEA) the provisional guideline 


values were established by route-to-route extrapolation of 

inhalatory LOAELs to oral LOAELs. Since benzene was evaluated 

to be genotoxic/carcinogenic (IARC group I) and NDMA 

(IARC group 2A) and 1,4-dioxane (IARC group 2B) are respectively 

suspected non-genotoxic and genotoxic carcinogens, 

their corresponding (provisional) guideline values are 

provided as an upper bound lifetime cancer risk to an individual 

of 10 5 , or the odds that one case of cancer would 

result for every 100 000 persons subjected to continuous 

exposure over a 70-year lifetime. 

3.3. Concentration of compounds in surface waters, 

groundwaters and drinking water 

The maximum concentrations of compounds reported in 

surface waters and/or groundwaters are summarized in 

Table 3. Measured maximum surface water concentrations 

were available for 37 of the 50 compounds in the annual 

reports of RIWA and ARW. For two compounds (MTBE and 

clofibric acid) maximum concentrations in Dutch groundwater 

are reported. For the remaining compounds, the 

maximum concentration reported in surface waters was 

taken from other sources (see Section 2.2.3). 

The six compounds with the highest reported maximum 

concentrations in surface waters were EDTA (29 mg/L), DTPA 

(12.2 mg/L), p,p 0 -sulfonyldiphenol (10 mg/L), urotropine (10 mg/ 

L), 1,4-dioxane (10 mg/L) and AMPA (5 mg/L), whereas in 

groundwater a relatively high concentration was found for 

MTBE (27.3 mg/L) showing the environmental relevance of this 

compound. The highest maximum concentration of iodinated 

contrast media in surface waters was reported for 

iomeprol (0.97 mg/L). 

Table 3 also summarizes the maximum concentrations of 

compounds reported in drinking water. Data on the occurrence 

of compounds in drinking water were relatively scarce, 

and limited to 35 compounds. For 18 compounds, drinking 

water concentrations were obtained from the Dutch REWAB 

database and for 17 compounds drinking water concentrations 

were taken from reports by others. Drinking water 

concentrations for the remaining compounds could not be 

found. The highest maximum concentration reported was for 

EDTA (13.6 mg/L), followed by DTPA (9 mg/L), metoprolol 

(2.1 mg/L) and BCIPE (1.9 mg/L). 

3.4. Comparison of compound concentrations to 

(provisional) guideline values (BQ value) 

For all compounds found in surface waters, groundwaters 

and drinking water the calculated BQ value was

Table 3 – Reported concentrations in surface waters, groundwaters and drinking water and comparison to (provisional) drinking water guideline values expressed as 

Benchmark Quotient (BQ) values. 

Compound Surface waters and groundwaters Drinking water 

Max conc (mg/L) (number 

of measurements, year) 

Source Ref BQ 

value a 

Max conc (mg/L) (number 

of measurements, year) 

Source Ref BQ 

value a 

1,4-Dioxane 10 (NA, 1997) SW, NL (4) 0.3 0.5 (NA) UDW, NL (4) 0.02 

2,6-Dichlorobenzamide (BAM) 0.05 (40, 2002-2006) SW, NL (11) 0.001 0.23 (14, 2002) FDW, USA (10) 0.004 

4-Methylbenzenesulfonamide 


4-tolylsulfonamide) 

0.06 (20, 2005) SW, NL (13) 0.00002 NA 

Acetylsalicylate (aspirin, acetyl 

salicylic acid) 

0.065 (NA, 2007) SW, NL (15) 0.003 0.12 (12, 2007) FDW, NL (15) 0.005 

Alpha-amino-3-hydroxy-5-methyl- 

4-isoxazole propionic acid (AMPA) 

5 (499, 2005) SW, NL (11) 0.006 1.1 (6, 2001) FDW, NL (10) 0.001 

Amidotrizoic acid (diatrizoic acid) 0.63 (189, 2006) SW, GER (2) 0.000003 0.25 (6, 2006) FDW, NL (10) 0.000001 

Bentazone 0.1 (126, 2007) SW, NL (11) 0.0003 0.28 (11, 2006) FDW, NL (10) 0.0009 

Benzene 0.74 (116, 2001) SW, NL (11) 0.07 0.96 (12, 2005) TW, NL (10) 0.1 

Benzothiazole 0.03 (3, 2008) SW, NL (6) 0.0003 0.01 (10, 2007) FDW, NL (14) 0.0001 

Benzotriazole (1H-benzotriazole) 0.54 (11, 2007) SW, NL (11) 0.0005 0.2 (10, 2007) FDW, NL (14) 0.0002 

Bis(chloroisopropyl)ether (BCIPE) 2.9 (15, 1984-1985) GW, NL (6) 0.02 1.9 (9, 1982-1984) FDW, NL (6) 0.01 

Carbamazepine 0.227 (263, 2003) SW, NL (11) 0.2 0.03 (2, 2007) FDW, NL (15) 0.03 

Carbendazim 1.5 (111, 2006) SW, BE (11) 0.01 NA 

Chloridazon (pyrazon) 0.3 (68, 2002) SW, BE (11) 0.002 NA 

Clofibric acid 0.091 (NA, 2007) BFGW, NL (15) 0.003 0.14 (2, 2007) FDW, NL (15) 0.005 

Dichlorophenoxyacetic acid (2,4-D) 0.2 (34, 2006) SW, NL (11) 0.007 0.11 (5, 2002) TW, NL (10) 0.004 

Diethyl phthalate 0.9 (8, 2005) SW, NL (11) 0.0003 NA 

Diethyl toluamide (DEET) 0.06 (36, 2005) SW, NL (11) 0.00001 0.03 (1, 2005) TW, NL (10) 0.000005 

Diethylamine (DEA) 0.29 (38, 2007) SW, NL (11) 0.0004 NA 

Diethylene glycol dimethyl 

ether (diglyme, bis(2-methoxy ethyl)ester)) 

3.64 (11, 2007) SW, NL (11) 0.02 0.15 (13, 2007) UDW, NL (1) 0.0009 

Diethylene triamine penta acetic acid (DTPA) 12.2 (53, 2005) SW, NL (11) 0.03 9 (2, 2001) FDW, NL (10) 0.03 

Dimethenamid 0.12 (4, 2005) SW, NL (9) 0.0005 NA 

Dimethylamine (DMA) 0.34 (42, 2005) SW, NL (11) 0.002 NA 

Diuron 0.68 (386, 2002) SW, BE (11) 0.1 0.08 (2, 2005) TW, NL (10) 0.01 

Ethyl tert-butyl ether (ETBE) 1.2 (97, 2006) SW, GER (2) 0.002 (1.2 b ) NA 

Ethylenediamine tetra acetic Acid (EDTA) 29 (192, 2005) SW, NL (11) 0.05 13.6 (7, 2001) FDW, NL (10) 0.02 

Glyphosate 1.2 (291, 2006) SW, NL (11) 0.001 0.46 (3, 2006) TW, NL (10) 0.0005 

Imidacloprid 0.06 (1, 2007) SW, NL (11) 0.0003 NA 

Iohexol 0.5 (180, 2005) SW, NL (11) 0.000001 0.06 (1, 2007) FDW, NL (15) 0.0000002 

Iomeprol (iomeron) 0.97 (172, 2007) SW, NL (11) 0.0001 0.01 (1, 2006) FDW, NL (10) 0.000001 

Iopamidol 0.714 (188, 2006) SW, NL (11) 0.000002 0.1 (6, 2006) FDW, NL (10) 0.0000002 

Iopromide 0.56 (186, 2004) SW, NL (11) 0.000002 0.04 (2, 2007) FDW, NL (15) 0.0000002 

Isoproturon 0.31 (256, 2002) SW, NL (11) 0.03 0.02 (1, 2004) TW, NL (10) 0.002 

Methyl tert-butyl ether (MTBE) 27.3 (14, 2003-2005) GW, NL (7) 0.003 (1.8 b ) 1.25 (27, 2006) FDW, NL (10) 0.0001 (0.08 b ) 

Metoprolol 0.2 (114, 2006) SW, NL (11) 0.004 2.1 (2, 2005) FDW, NL (10) 0.04 

n-Butylbenzenesulphonamide 0.78 (300-400, 2004-2006) SW, NL (6) 0.003 0.05 (2, 2004) TW, NL (13) 0.0002 

Nicosulfuron 0.17 (5, 2007) SW, NL (11) 0.0002 NA 

n-Nitrosodimethylamine (NDMA) 0.0071 (38, 2006) SW, NL (11) 0.07 0.002 (21, 2007) UDW, NL (5) 0.02 

470 


p,p0-Sulfonyldiphenol 10 (300, 2005-2006) SW, BE (6) 0.1 NA 

Perfluoroctane sulfonate (PFOS) 0.11 (NA, 2008) SW, NL (8) 0.2 0.02 (NA, 2005/2006) TW, GER (12) 0.04 

Perfluorooctanoic acid (PFOA) 0.647 (NA, 2005/2006) SW, GER (13) 0.1 0.52 (NA, 2005/2006) TW, GER (12) 0.1 

Phenazone 0.11 (26, 2005) SW, NL (12) 0.0009 0.03 (8, 2005) FDW, NL (15) 0.0002 

Simazine 0.13 (124, 2006) SW, BE (11) 0.07 0.06 (4, 2004) FDW, NL (10) 0.03 

Sulfamethoxazole 0.11 (170, 2005) SW, NL (11) 0.0003 0.03 (4, 2007) FDW, NL (15) 0.00007 

Tolyltriazole 0.29 (11, 2007) SW, NL (11) 0.0003 NA 

Trichloroethene 1.35 (206, 2006) SW, NL (11) 0.07 1.75 (8, 2005) FDW, NL (10) 0.09 

Triethylphosphate (ethylphosphate) (TEP) 0.189 (25, 2007) SW, NL (11) 0.0001 NA 

Triphenylphosphine oxide (TPPO) 0.344 (88, 2003) SW, NL (11) 0.01 0.13 (6, 2007) FDW, NL (6) 0.005 

Tris(2-chloroethyl)phosphate (TCEP) 0.29 (8, 2006) SW, NL (11) 0.004 NA 

Urotropine (methenamine, hexamine) 10 (NA, 1997-1998) SW, GER (3) 0.02 NA 

Sources: bank-filtrated groundwater (BFGW); groundwater (GW); surface water (SW); finished drinking water (FDW); tap water (TW); unspecified drinking water (UDW). Locations: Belgium (BE); 

Germany (GER); the Netherlands (NL); United States of America (USA). References: (1) Anonymous data Dutch drinking water companies; (2) ARW database; URL4; (3) Brauch et al., 2000; (4) ECB, 2002; (5) 

Kleinneijenhuis and Puijker, 2008; (6) KWR internal data (KWR, 2009); (7) De Voogt et al., 2008; (8) Loos et al., 2009; (9) Mout et al., 2007; (10) REWAB database, 2009; (11) RIWA database; URL5 (12) 

Skutlarek et al., 2006; (13) van Beelen, E.S.E. (HWL, the Netherlands), personal communication; (14) Van Leerdam et al., 2009; (15) Versteegh et al., 2007. NA: Not available; an absence of measured 

concentrations above the detection limit. 

a BQ value: ratio of reported maximum concentration to the (provisional) guideline value (cf. Table 2). 

b Based on an odour-threshold value (cf. Table 2). 


there is a concern for drinking water production, because 

consumers will not accept odourous drinking water. The BQ 

values of the remaining compounds calculated for surface 

waters and groundwater ranged between 0.1 (p,p 0 -sulphonylphenol, 

diuron) and 0.000001 (iohexol). For all iodinated 

contrast media the concentration in surface waters was at 

least three orders of magnitude less than the (provisional) 

guideline value. For drinking water, the BQ values of these 

compounds were even lower (Fig. 2B). For two compounds 

(benzene and PFOA) found in drinking water, the BQ value 

was equal to 0.1 indicating that additional assessments such 

as establishing trends may be warranted. However, for 15 

compounds occurrence data in drinking water were not 

available and therefore the human health concern associated 

with drinking water consumption due to presence of any of 

these compounds remains unknown. 

4. Discussion 

Rapid new developments in analytical chemistry lead to the 

detection and quantification of many emerging contaminants 

in drinking water and its environmental sources (surface 

water and groundwater). Since toxicological information is 

often absent, such compounds are a growing concern for 

drinking water companies and their customers. 

The present study attempts to address potential human 

health concern associated with water containing emerging 

contaminants. The 50 compounds included in this study 

represent a broad range of chemical classes for which 

maximum concentrations in surface waters, groundwater 

and/or drinking water were obtained in the downstream parts 

of the Rhine and Meuse basins. The results as presented in 

Fig. 2 indicate that a substantial margin exists between the 

(provisional) guideline value and the maximum concentrations 

of most compounds reported in surface waters, 

groundwaters and/or drinking water. 

The compounds evaluated with a relatively high BQ value 

(i.e. a high potential human health concern) and a known 

carcinogenic action are 1,4-dioxane, benzene and NDMA. The 

(provisional) guideline values for 1.4-dioxane (30 mg/L), benzene 

(10 mg/L) and NDMA (0.1 mg/L) used in the present study are 

based on a specific cancer risk level of 10 5 . However, when 

applying a specific risk level of 10 6 ,asiscommonpracticeinthe 

Netherlands, the provisional guidelines value would be 3 mg/L, 

1 mg/L and 0.01 mg/L, respectively. This would result in BQ values 

much higher than the arbitrary thresholds for surface waters 

and drinking water employed in the present study. This indicates 

that very low concentrations of these compounds in 

drinking water could lead to a potential carcinogenic effect, and 

we conclude that for these compounds it is important to 

monitor trends in their (environmental) occurrence. 

Furthermore, Fig. 2 illustrates that the (provisional) guideline 

values of the majority of non-genotoxic compounds are at least 

two orders of magnitude above the Threshold of Toxicological 

Concern (TTC)-based drinking water target value for non-genotoxic 

compounds (0.1 mg/L). In addition, the (provisional) 

guideline values (expressed as a specific risk level of 10 6 )ofthe 

three compounds with known carcinogenic action (1,4-dioxane, 

benzene and NDMA) are equal (NDMA) or much higher (1,4

472 

Fig. 2 – Comparison of compound concentrations in (A) surface/groundwaters and (B) drinking water to (provisional) 

guideline values. Benchmark Quotient (BQ) thresholds are indicated with dashed lines. Threshold of Toxicological Concern 

(TTC) based target value for non-genotoxic compounds (0.1 mg/L) is indicated with a dotted line. Numbers correspond to 

compounds as tabulated in Table 1. 

dioxane and benzene) than the TTC derived target value of 

0.01 mg/L for genotoxic compounds. This illustrates that the TTC 

based drinking water target value may be a conservative value 

ideally suited for exposure based waiving of compounds for 

which there is no sufficient toxicological information, which 

can be followed up by a more data-intensive evaluation. 

Two perfluorinated organic compounds were evaluated in 

the present study. For PFOA in drinking water a BQ value equal 

to the arbitrary threshold of 0.1 was calculated, whereas for 

PFOS in surface waters a BQ value of 0.2 was calculated. These 

persistent compounds are becoming a global problem, and 

PFOA and PFOS have already been detected in the ng/L range 

in, e.g. European and Japanese tapwaters (Ericson et al., 2007; 

Loos et al., 2007; Norimitsu et al., 2004). Recently, Skutlarek 

et al. (2006) observed at sampling site Neheim (river Ruhr 

catchment, a tributary of the river Rhine, Germany) 


concentrations of PFOA of 0.65 mg/L in Lake Moehne, and 

0.53 mg/L in corresponding drinking water, respectively. The 

authors concluded that water treatment steps may not 

effectively eliminate perfluorinated compounds to a sufficient 

extent, although approximately 50% of the waterworks at the 

Ruhr river are equipped with activated carbon filters. Hence, 

more research should be devoted to the behavior of perfluorinated 

organic compounds in drinking water treatment 

processes. 

Several structurally related iodinated contrast media 

(iopamidol, iohexol, iomeprol and iopromide) were evaluated 

in the present research. However, their calculated BQ values 

are much lower than the BQ threshold above which further 

investigations would be warranted. Iopromide, for example, is 

a relatively non-toxic compound with a reported safe dose 

(intravenous) of 50 g/person/d (Versteegh et al., 2007). Despite

the absence of human health effects, these compounds may 

deserve further attention from an environmental impact 

point of view. Since environmental sublethal effects of 

iodinated contrast media to organisms are largely unknown, 

taken together with high persistence and environmental 

presence at relatively high concentrations, additional environmental 

assessments may be necessary. 

For compounds with a low (provisional) guideline value as 

identified in the present study (e.g. carbamazapine), additional 

environmental monitoring may be warranted to characterize 

concentrations and to establish trends in their occurrence. As 

shown by Walraven and Laane (2009), river flow rates may 

influence contaminant concentrations seasonally, thus 

resulting in substantially varying BQ values. For example, it can 

be observed that the riverine concentration of the fuel 

oxygenate MTBE is highly dependent on the flow of the river 

Meuse. Similar patterns may occur for other compounds, 

resulting in (temporarily) exceedance of the BQ threshold. 

The evaluation as presented here supports the conclusion 

that the majority of the selected compounds as found in 

surface waters, groundwater and drinking water do not pose 

an appreciable concern to human health. This finding of no 

adverse effect to human health from exposure to trace 

quantities of compounds (e.g. pharmaceuticals) in surface 

waters and/or drinking water is supported by other results 

reported in the literature. Kingsbury et al. (2008) recently 

evaluated the potential health effects of 148 organic 

compounds in source water and finished water. The authors 

showed that the annual mean concentration of all compounds 

detected in finished water was less than the established 

human health benchmarks. Furthermore, Snyder et al. (2008) 

arrived at the same findings after evaluating human health 

effects associated with potential drinking water exposure of 

a suite of 62 indicator pharmaceuticals and potential endocrine 

disrupting compounds. 

Despite the absence of any concern to human health, 

drinking water remains a major point of consumer concern 

and some residual uncertainties need further exploration. For 

example, drinking water guideline values are developed using 

toxicity information for single compounds. Hence, the longterm 

cumulative dose-additive or synergistic effects of low 

concentrations of contaminants co-occurring as mixtures on 

human health and potentially sensitive sub-populations 

remain currently unknown. Understanding and implementing 

of such information is important for the development of 

future (enforceable) guideline values. Finally, the relatively 

large data gap on occurrence of compounds in drinking water 

should compel further research and assessment, especially 

for those compounds with a low (provisional) guideline value. 

5. Major conclusions 

For most compounds evaluated in the present assessment, 

a substantial margin exists between the (provisional) 

guideline value and the maximum concentrations in 

surface waters, groundwaters and/or drinking water. 

The TTC based drinking water target values (0.1 mg/L and 

0.01 mg/L for non-carcinogenic compounds, respectively) as 

proposed earlier are the conservative values they are meant 


to be. They are optimally suited to provide exposure based 

waiving. 

The concentrations in drinking water of compounds such as 

MTBE, ETBE, 1,4-dioxane, NDMA and benzene should be 

monitored closely, since their guideline values are easily 

exceeded. 

Alkylated perfluorinated compounds such as PFOA and 

PFOS are environmentally persistent compounds and their 

increasing occurrence in (the sources of) drinking water 

should be monitored closely. 

For compounds with a very low (provisional) guideline value 

(e.g. mutagenic and carcinogenic compounds) it is important 

to better establish trends in their environmental occurrence. 

From a toxicological point of view iodinated contrast media 

as present in drinking water, such as amidotrizoic acid 

iopamidol, iohexol and iopromide, are not a direct concern 

for human health. However, further environmental assessment 

may be necessary, especially since the sublethal 

(ecological) effects of these compounds are largely unknown. 

Better understanding of the potential mixture effects of 

emerging compounds present in drinking water is important 

for the development of future guideline values. 


The authors wish to acknowledge the significant contribution 

to this work by Leo Puijker (KWR) and Ruud Jansen. Thanks 

also to Dr. Corine Houtman (HWL, the Netherlands) for her 

valuable comments on the manuscript. The research 

described was funded by the Joint Research Programme of the 

Dutch Water utilities (BTO). 

references 

Benzotriazoles Coalition, 2001. Benzotriazoles category 

justification and testing rationale. Synthetic Organic Chemical 

Manufacturers Organization. http://www.epa.gov/oppt/ 

chemrtk/pubs/summaries/benzo/c13456.pdf (accessed June 

2009). 

BIBRA Toxicology International, 1995. Toxicity Profile of 

Benzotriazole. BIBRA Toxicology International, Carshalton 

(Surrey), UK. 

Brauch, H.J., Sacher, F., Denecke, E., Tacke, T., 2000. Efficiency of 

river-bank filtration for the removal of polar organic 

compounds (Werksamkeit der Uferfiltration fuer die 

Enfernung von polaren organischen Spurenstoffen). Wasser- 

Abwasser 141 (4), 226–234. 

Brown, D., Butterworth, K.R., Gaunt, I.F., Grasso, P., Gangolli, S.D., 

1978. Short-term oral toxicity study of diethyl phthalate in the 

rat. Food Cosmet. Toxicol. 16, 415–422. 

California EPA (California Environmental Protection Agency), 

2000. N,N-diethyl-meta-toluamide (DEET) risk 

characterization document. Department of pesticide 

regulation. Document number RCD 00-01, Sacramento, 

California, USA. 

CIIT (Chemical Industry Institute of Toxicology), 1990. Twenty 

four month final report inhalation toxicity of dimethylamine 

in F344 rats and B6C3F1 mice. CIIT Archives, Docket #11957. 

Ciba-Geigy, 1988. Combined chronic toxicity/oncogenicity study 

in rats. Basle, Ciba-Geigy, 1988 (unpublished study submitted 

to WHO).

474 

Cotruvo, J.A., 1988. Drinking water standards and risk 

assessment. Regul. Toxicol. Pharmacol. 8 (3), 288–299. 

Danish EPA (Danish Environmental Protection Agency), 2004. 

Evaluation of Health Hazards by Exposure to BAM (2,6- 

Dichlorobenzamide) and Risk Characterisation of Drinking 

Water Exposure. Environmental Project Nr. 943 2004. 

Dawson, B.V., Johnson, P.D., Goldberg, S.J., Ulreich, J.B., 1993. 

Cardiac teratogenesis of halogenated 

hydrocarbon-contaminated drinking water. J. Am. Coll. 

Cardiol. 21 (6), 1466–1472. 

De Voogt, P., Vink, C., Puijker, L.M., 2008. MTBE and ETBE in Dutch 

groundwater (MTBE en ETBE in Nederlands grondwater), KWR 

08.043, KWR, Nieuwegein, The Netherlands. 

DuPont de Nemours and Company, Inc., 1964. Report MRID No. 

00017763, 00080899. 

DECOS (Dutch Expert Committee for Occupational Standards), 

2000. 1,2,3-benzotriazole. Health-based recommended 

occupational exposure limit. No. 2000/14OSH, The Hague, The 

Netherlands. 

ECB (European Chemicals Bureau), 2000a. IUCLID dataset on 

benzotriazoles. 

ECB (European Chemicals Bureau), 2000b. IUCLID dataset on 

Chloridazon. 

ECB (European Chemicals Bureau), 2000c. IUCLID dataset on 

Diethylene triamine penta acetic acid. 

ECB (European Chemicals Bureau), 2000d. IUCLID dataset on 

triphenylphosphine oxide. 

ECB (European Chemicals Bureau), 2002. European Risk 

Assessment Report 1,4-dioxane. 

EFSA (European Food Safety Authority), 2008. Opinion of the 

scientific panel on contaminants in the food chain on 

perfluorooctane sulfonate (PFOS), perfluorooctanoic acid 

(PFOA) and their salts. EFSA J. 53, 1–131. 

Ericson, I., Nadal, M., van Bavel, B., Lindström, G., Domingo, J.L., 

2007. Levels of perfluorochemicals in water samples from 

Catalonia, Spain: is drinking water a significant contribution 

to human exposure? Environ. Sci. Pollut. Res. 15, 614–619. 

Goldenthal, E.I. (International Research and Development Corp.), 

1994. Evaluation of DEET in a one-year chronic oral toxicity 

study in dogs. DEET Joint Venture/Chemical Specialties 

Manufacturers Assoc. DPR vol. 50191-164, Rec. No. 131265. 

Health Canada, 2003. Benchmark Dose for TCE in Drinking Water. 

Health Canada, Healthy Environments and Consumer Safety 

Branch, Biostatistics Unit, April, Ottawa, Ontario. 

Hogenboom, A.C., Van Leerdam, J.A., De Voogt, P., 2009. Accurate 

mass screening and identification of emerging contaminants 

in environmental samples by liquid chromatography–hybrid 

linear ion trap Orbitrap mass spectrometry. J. Chromatogr. A 

1216 (3), 510–519. 

Hurni, H., Ohder, H., 1973. Reproduction study with formaldehyde 

and hexamethylenetetramine in beagle dogs. Food Cosmet. 

Toxicol. 11 (3), 459–462. 

Kingsbury, J.A., Delzer, C.D., Hopple, J.A., 2008. Anthropogenic 

organic compounds in source water of nine community water 

systems that withdraw from streams, 2002–2005, USGS, 

Reston, Virginia, USA, Report 2008–5208. 

Kleinneijenhuis, A.J., Puijker, L.M., 2008. Monitoring program for 

DMSA, DMST, N,N-dimethylsulfamide and its ozone reaction 

product NDMA in untreated and ozone-treated water. TNO 

report 7686, Delft, The Netherlands. 

Kolpin, D.W., Skopec, M., Meyer, M.T., Furlong, E.T., Zaugg, S.D., 

2004. Urban contribution of pharmaceuticals and other 

organic wastewater contaminants to streams during differing 

flow conditions. Sci. Total Environ. 328 (1–3), 119–130. 

Kroes, R., Renwick, A.G., Cheeseman, M., Kleiner, J., 

Mangelsdorf, I., Piersma, A., Schilter, B., Schlatter, J., 

Van Schothorst, F., Vos, J.G., Wurtzen, G., 2004. Structurebased 

thresholds of toxicological concern (TTC): guidance for 


application to substances present at low levels in the diet. 

Food Chem. Toxicol. 42, 65–83. 

Krop, H. de Voogt, P., 2008. Physicochemical parameters and 

source markers of PFAS. Report IVAM ISO doc O/0818, 

Amsterdam, The Netherlands, pp. 1–147. 

KWR Watercycle Research Institute, 2009. KWR internal data, 

Nieuwegein, the Netherlands. 

Larsen, M.L., Illingworth, D.R., O’Malley, J.P., 1994. Comparative 

effects of gemfibrozil and clofibrate in type III 

hyperlipoproteinemia. Atherosclerosis 106, 235–240. 

Loos, R., Wollgast, J., Huber, T., Hanke, G., 2007. Polar herbicides, 

pharmaceutical products, PFOS, PFOA, and nonylpheniol and 

its carboxylates and ethoxylates in surface and tap waters 

around Lake Maggiore in Northern Italy. Anal. Bioanal. Chem. 

387, 1469–1478. 

Loos, R., Gawlik, B.M., Locoro, G., Rimaviciute, E., Contini, S., 

Bidoglio, G., 2009. EU-wide survey of polar organic persistent 

pollutants in European river waters. Environ. Pollut. 157 (2), 

561–568. 

Lynch, D.W., Moorman, W.J., Stober, P., Lewis, T.R., Iverson, W.O., 

1986. Subchronic inhalation of diethylamine vapour in Fisher- 

344 rats: organ toxicity. Fund. Appl. Toxicol. 6 (3), 559–565. 

Medinsky, M.A., Wolf, D.C., Cattley, R.C., Wong, B., Janszen, D.B., 

Farris, G.M., Wright, G.A., Bond, J.A., 1999. Effects of 

a thirteen-week inhalation exposure to ethyl tertiary 

butyl ether on Fischer-344 rats and CD-1 mice. Toxicol. Sci. 

51, 108–118. 

Mitsubishi Chemical Safety Institute Ltd. (date of study unknown). 

4,4’-sulfonyldiphenol. Summary of repeated dose and 

developmental studies. Kashima Laboratory, Ibaraki, Japan. 

Mitsumori, K., Usui, T., Takahashi, K., Shirasu, Y., 1979. Twentyfour 

month chronic toxicity studies of dichlorodiisopropyl 

ether in mice. J. Pestic. Sci. 4 (3), 323–335. 

Mons, M., Van der Hoek, J.P., Stoks, P., Van der Kooij, D., 2008. 

Vision on and guideline values for xenobiotic substances in 

drinking water (Visie op en streefwaarden voor 

milieuvreemde stoffen in drinkwater). H 2O 4, 1–4 (in Dutch). 

Morgareidge, K., 1971. 90-Day feeding study with benzothiazole in 

rats. Unpublished report from Food and Drug Research 

Laboratories, Inc., Maspeth, New York, USA. Submitted to 

WHO by Flavor and Extract Manufacturers’ Association of the 

United States. 

Morisetti, A., Tirone, P., Luzzani, F., de Haen, C., 1994. 

Toxicological safety assessment of iomeprol, a new X-ray 

contrast agent. Eur. J. Radiol. 1 (18 Suppl), S21–31. 

Mout, L., Jansen, L., van der Pijll, S., Nierop, S., 2007. Implementation 

program dimethenamid-p. Clean sources, now and in the future 

(uitvoeringsprogramma dimethenamid-p. Schone bronnen, nu 

en in de toekomst: tweede reeks knelpunten (http://www. 

schonebronnen.nl/). The Hague, The Netherlands. 

Munro, I.C., Ford, R.A., Kennepohl, E., Sprenger, J.G., 1996. 

Correlation of structural class with no-observed effect levels: 

a proposal for establishing a threshold of concern. Food Chem. 

Toxicol 34, 829–867. 

NCI (National Cancer Institute), 1978. Bioassay of 1,4-dioxane for 

possible carcinogenicity, CAS No. 123-91-1. NCI 

Carcinogenesis Technical Report Series No. 80. DHEW 

Publication No. (NIH) PB-285-711, Bethesda, Maryland, USA. 

Nghiem, L.D., Schafer, A.I., Elimelech, M., 2004. Removal of 


measurement, modeling, and mechanisms. Environ. Sci. 

Technol. 38 (6), 1888–1896. 

Norimitsu, S., Kouji, H., Kayoko, I., Kazuaki, S., Takeo, Y., Akio, K., 

2004. Perfluorooctanoate and perfluorooctane sulfonate 

concentrations in surface water in Japan. J. Occup. Health 46, 

49–59. 

NTP (National Toxicology Program), 1986. Toxicology and 

carcinogenesis studies of benzene in F344/N rats and B6C3F1

mice (gavage studies). Research Triangle Park, NC, USA, US 

Department of Health and Human Services. (Technical report 

series No. 289). 

NTP (National Toxicology Program), 1987. Teratologic evaluation 

of diethylene glycol dimethyl ether (CAS No. 111-96-6) 

administered to New Zealand White rabbits on gestation days 

6 through 19. Research Triangle Park, NC, USA, National 

Institute of Environmental Health Sciences, National 

Toxicology Program (NTP-87-108; PB 87–209532). 

NTP (National Toxicology Progam), 1991. Toxicology and 

carcinogenesis studies of tris(2-chloroethyl)phosphate in 

F344/N rats and B6C3F1 mice. Technical report series No. 391. 

NTP TR 391, Research Triangle Park, NC, USA. 

OECD/SIDS (Organization for Economic Development and 

Cooperation/Screening Information DataSet), 1994. p- 

Toluenesulfonamide. UNEP publications, Nairobi, Kenya. 

OECD/SIDS (Organization for Economic Development and 

Cooperation/Screening Information DataSet), 1998. 

Triethylphosphate. UNEP publications, Nairobi, Kenya. 

Peto, R., Gray, R., Brantom, P., Grasso, P., 1991a. Effects on 4080 

rats of chronic ingestion of N-nitrosodiethylamine or Nnitrosodimethylamine: 

a detailed dose–response study. 

Cancer Res. 51, 6415–6451. 

Peto, R., Gray, R., Brantom, P., Grasso, P., 1991b. Dose and time 

relationships for tumor induction in the liver and esophagus 

of 4080 inbred rats by chronic ingestion of Nnitrosodiethylamine 

or N-nitrosodimethylamine. Cancer Res. 

51, 6452–6469. 

Proviron Fine Chemicals, 2003. IUCLID dataset on Nbutylbenzenesulphonamide. 

REWAB Database, 2009. KWR Watercycle Research Institute, 

Nieuwegein, The Netherlands. 

Robinson, M., Bruner, R.H., Olson, G.R., 1990. Fourteen- and 

ninety-day oral toxicity studies of methyl-t-butyl ether in 

Sprague-Dawley rats. J. Am. Coll. Toxicol. 9, 525–540. 


Johnson, C.A., Von Gunten, U., Wehrli, B., 2006. The challenge 


1072–1077. 

SCOEL (Scientific Committee on Exposure Limits), 1991. No 11b. 

Recommendations of the SCOEL for DMA. 

SCOEL (Scientific Committee on Exposure Limits), 2002. No 91. 

Recommendations of the SCOEL for DEA. 

Schwab, B.W., Hayes, E.P., Fiori, J.M., Mastrocco, F.J., Roden, N.M., 

Cragin, D., Meyerhoff, R.D., D’Aco, V.J., Anderson, P.D., 2005. 

Human pharmaceuticals in US surface waters: a human 

health risk assessment. Regul. Toxicol. Pharmacol. 42 (3), 

296–312. 

Seacat, A.M., Thomford, P.J., Hansen, K.J., Olsen, G.W., Case, M.T., 

Butenhoff, J.L., 2002. Subchronic toxicity studies on 

perfluorooctanesulfonate potassium salt in cynomolgus 

monkeys. Toxicol. Sci. 68 (1), 249–264. 

Sherman, H., 1972. Long-term feeding studies in rats and dogs 

with 2-benzimadazole carbamic acid, methyl ester (INE-965) 

(50% and 70% MBC wettable powder formulations). 

Unpublished report from E.I. DuPont de Nemours and Co., Inc., 

Haskell Laboratory, Newark, Delaware, USA. 

Singh, G., Driever, P.H., Sander, J.W., 2005. Cancer risk in people 

with epilepsy: the role of antiepileptic drugs. Brain 128 (Pt 1), 

7–17. 

Skutlarek, D., Exner, M., Farber, H., 2006. Perfluorinated 

surfactants in surface and drinking waters. Environ. Sci. 

Pollut. Res. 13 (5), 299–307. 

Snyder, S.A., Villeneuve, D.L., Snyder, E.M., Giesy, J.P., 2001. 

Identification and quantification of estrogen receptor agonists 

in wastewater effluents. Environ. Sci. Technol. 35 (18), 

3620–3625. 

Snyder, S.A., Trenholm, R.A., Snyder, E.M., Bruce, G.M.P., 

Hemming, J.D.C., 2008. Toxicological Relevance of EDCs and 


Pharmaceuticals in Drinking Water. Awwa Research 

Foundation, Denver, USA, p. 484. 

Stokinger, H.E., Woodward, R.L., 1958. Toxicologic methods for 

establishing drinking water standards. J. Am. Water Works 

Ass. 50, 515–529. 

Swarm, R.L., Roberts, G.K., Levy, A.C., Hines, L.R., 1973. 

Observations on the thyroid gland in rats following the 

administration of sulfamethoxazole and trimethoprim. 

Toxicol. Appl. Pharm. 24 (3), 351–363. 

Swartjes, F.A., Baars, A.J., Fleuren, R.H.L.J., Otte, P.F., 2004. Risk 

limits for MTBE (methyl tertiair-butyl ether) in soil, sediment, 

groundwater, surface water, drinking water and for drinking 

water production (risicogrenzen voor MTBE (Methyl tertiair- 

Butyl Ether) in bodem, sediment, grondwater, 

oppervlaktewater, drinkwater en voor drinkwaterbereiding) 

RIVM report 711701039/2004, Bilthoven, The Netherlands. 

Toccalino, P.L., 2007. Development and application of healthbased 

screening levels for use in water-quality assessments. 

US Geological Survey Scientific Investigations Report 2007- 

5106, 12 p, Reston, Virginia, USA. 

URL1: http://www.epa.gov/waterscience/criteria/drinking/ 

dwstandards.pdf, accessed June 2009. 

URL2: http://www.who.int/water_sanitation_health/dwq/ 

guidelines/en/, accessed June 2009. 

URL3: http://www.ctb.agro.nl/ctb_files/11995_01.html, accessed 

May 2009 (In Dutch). 

URL4: http://www.iawr.org/index.php, publications, accessed 

January 2009. 

URL5: http://www.riwa-rijn.org/riwa_en.html, publications, 

accessed January 2009. 

US EPA (United States Environmental Protection Agency) IRIS 

database (1988a). 1,4-dioxane. 


database, 1988b. Diethyl phthalate. 


database, 1988c. Diuron. 


database, 1989. Bis(chloroisopropyl)ether. 

US EPA (United States Environmental Protection Agency), 2006. 

National Primary Drinking Water Standards. http://www.epa. 

gov/safewater/consumer/pdf/mcl.pdf (accessed August 2009). 

Van Leerdam, J.A., Hogenboom, A.C., Kooi van der, M.E., de 

Voogt, P., 2009. Determination of polar 1-H-benzotriazoles and 

benzothialzoles in water by solid-phase extraction and liquid 

chromatography LTQ Orbitrap mass spectrometry. Int. J. Mass 

Spectrom. 282 (3), 99–107. 

Van Leeuwen, F.X., 2000. Safe drinking water: the toxicologist’s 

approach. Food Chem. Toxicol. 38 (1 Suppl), S51–S58. 

Van Leeuwen, C.J., Vermeire, T., 2007. Risk Assessment of 

Chemicals, second ed., Springer, ISBN 978-4020-6101-1. 

Van Wezel, A.P., Puijker, L., Vink, C., Versteegh, A., de Voogt, P., 

2009. Odour and flavour thresholds of gasoline additives 

(MTBE, ETBE and TAME) and their occurrence in Dutch 

drinking water collection areas. Chemosphere 76, 672–676. 

Versteegh, J., Van Der Aa, N., Dijkman, E., 2007. Pharmaceuticals 

in drinking water and sources of drinking water - results of 

monitoring campaign 2005/2006 (geneesmiddelen in 

drinkwater en drinkwaterbronnen - resultaten van het 

meetprogramma 2005/2006), 53, RIVM, Bilthoven, The 

Netherlands. Number: R(T)-1-67. 

Walraven, N., Laane, R.W., 2009. Assessing the discharge of 

pharmaceuticals along the Dutch coast of the North sea. Rev. 

Environ. Contam. Toxicol. 199, 1–18. 




processes. Environ. Sci. Technol. 37 (17), 6649–6663. 

WHO (World Health Organization), 1996. Simazine in drinkingwater. 

Background document for development of WHO

476 

Guidelines for drinking-water quality. WHO/SDE/WSH/03.04/ 

42, Geneva, Switzerland. 

WHO (World Health Organization), 1998. Guidelines for 

drinking water quality. Second edition, volume 1 

recommendations: addendum, ISBN 92 4 154514 3, Geneva, 

Switzerland. 

WHO (World Health Organization), 2002. Diethylene glycol 

dimethyl ether. Concise International Chemical Assessment 

Document 41, Geneva, Switzerland. 

WHO (World Health Organization), 2003a. Benzene in drinking 

water. WHO report WHO/SDE/WSH/03.04/24, Geneva, 

Switzerland. 

WHO (World Health Organization), 2003b. 2,4-D in drinking water. 

WHO report WHO/SDE/WSH/03.04/70, Geneva, Switzerland. 

WHO (World Health Organization), 2003c. Isoproturon in 

Drinking-water. Background document for development of 

WHO Guidelines for Drinking-water Quality WHO/SDE/WSH/ 

03.04/37, Geneva, Switzerland. 

WHO (World Health Organization), 2005a. Glyphosate and AMPA 

in drinking-water. Background document for development of 

WHO Guidelines for Drinking-water Quality WHO/SDE/WSH/ 


WHO (World Health Organization), 2005b. Trichloroethene in 

Drinking-water, Background document for development of 

WHO Guidelines for Drinking-water Quality. WHO/SDE/WSH/ 


WHO (World Health Organization), 2006. Guidelines for drinking 

water quality, First Adendum to third ed., Recommendations, 

vol. 1, 2006, Geneva, Switzerland. 

WHO (World Health Organization), 2008. N- 

Nitrosodimethylamine in drinking-water. Background 


document for preparation of WHO Guidelines for drinkingwater 

quality. Geneva, World Health Organization (WHO/HSE/ 

AMR/08.03/8), Geneva, Switzerland. 

WHO/JECFA (World Health Organization/joint FAO/WHO expert 

committee on food additives), 1973. Food Additives Series 32; 

Toxicological Evaluation of Certain Food additives and 

Contaminants. Prepared by the forty-first meeting of the joint 

FAO/WHO expert committee. 


committee on food additives), 1974. Food Additives Series No. 

5; Seventeenth Report of the Joint FAO/WHO Expert 

Committee on Food Additives, World Health Organization 

technical report series, 1974, No. 539; FAO Nutrition Meetings 

Report Series, 1974, No. 53. 


committee on food additives), 1995. 892. Carbendazim 

(Pesticide residues in food: 1995 evaluations Part II 

Toxicological & environmental). 


committee on food additives), 2003. WHO Food Additives 

Series No. 50; sulfur-containing heterocyclic compounds. 

WHO/JMPR (World Health Organization/Joint Meeting on 

Pesticide Residues), 2001. Toxicological evaluations 

imidacloprid. Pesticide residues in food. 

WHO/JMPR (World Health Organization/Joint Meeting on 

Pesticide Residues), 2005. Dimethenamid-p. pp. 189–239. 

Wilson, A.B., Thorpe, E., 1971. Toxicity studies on the ‘‘Prefix’’ 

residue 2,6-dichlorobenzamide: two year oral experiment with 

dogs. Tunstall Laboratory. Report number: Lab Project 

Number: T507531; Group research report: TLGR.0028.71, pp. 

1–28.

Monitoring the biological activity of micropollutants during 

advanced wastewater treatment with ozonation and activated 

carbon filtration 

M. Macova a , B.I. Escher a, *, J. Reungoat b , S. Carswell a,c , K. Lee Chue a , 

J. Keller b , J.F. Mueller a 

a 

University of Queensland, National Research Centre for Environmental Toxicology (EnTox), 39 Kessels Road, Brisbane, QLD 4108, Australia 

b 

University of Queensland, Advanced Water Management Centre (AWMC), Brisbane, QLD 4072, Australia 

c 

Queensland Health Forensic and Scientific Services, Organics Laboratory, 39 Kessels Road, Brisbane, QLD 4108, Australia 

article info 







Keywords: 

Bioassays 

Baseline toxicity 

Phytotoxicity 

Estrogenicity 

Genotoxicity 

Toxic equivalency concept 




abstract 

* Corresponding author. Tel.: þ61 (0) 7 3274 9180; fax: þ61 (0) 7 3274 9003. 

E-mail address: b.escher@uq.edu.au (B.I. Escher). 


doi:10.1016/j.watres.2009.09.025 

A bioanalytical test battery was used to monitor the removal efficiency of organic micropollutants 

during advanced wastewater treatment in the South Caboolture Water Reclamation 

Plant, Queensland, Australia. This plant treats effluent from a conventional sewage treatment 

plant for industrial water reuse. The aqueous samples were enriched using solid-phase 

extraction to separate some organic micropollutants of interest from metals, nutrients and 

matrix components. The bioassays were chosen to provide information on groups of chemicals 

with a common mode of toxic action. Therefore they can be considered as sum indicators 

to detect certain relevant groups of chemicals, not as the most ecologically or human health 

relevant endpoints. The baseline toxicity was quantified with the bioluminescence inhibition 

test using the marine bacterium Vibrio fischeri. The specific modes of toxic action that were 

targeted with five additional bioassays included aspects of estrogenicity, dioxin-like activity, 

genotoxicity, neurotoxicity, and phytotoxicity. While the accompanying publication discusses 

the treatment steps in more detail by drawing from the results of chemical analysis as well as 

the bioanalytical results, here we focus on the applicability and limitations of using bioassays 

for the purpose of determining the treatment efficacy of advanced water treatment and for 

water quality assessment in general. Results are reported in toxic equivalent concentrations 

(TEQ), that is, the concentration of a reference compound required to elicit the same response 

as the unknown and unidentified mixture of micropollutants actually present. TEQ proved to 

be useful and easily communicable despite some limitations and uncertainties in their derivation 

based on the mixture toxicity theory. The results obtained were reproducible, robust 

and sensitive. The TEQ in the influent ranged in the same order of magnitude as typically seen 

in effluents of conventional sewage treatment plants. In the initial steps of the treatment 

chain, no significant degradation of micropollutants was observed, and the high levels of 

dissolved organic carbon probably affected the outcome of the bioassays. The steps of coagulation/flocculation/dissolved 

air flotation/sand filtration and ozonation decreased the effectbased 

micropollutant burden significantly. 


478 


Water quantity and water quality issues have led many 

countries to explore water reuse and water reclamation as 

alternatives to provide water for direct use in industrial 

applications as well as agricultural and gardening use, and 

for indirect use to supplement water reservoirs. Australia is 

amongst the countries most affected by uneven distribution 

of water resources and droughts. Therefore, various water 

treatment and desalination plants have been put into 

operation, among them the Water Reclamation Plant in 

South Caboolture, Queensland, Australia (van Leeuwen 

et al., 2003). 

Besides hygiene issues, organic micropollutants are of 

major concern because classical wastewater treatment with 

biological activated sludge only partially removes hazardous 

chemicals. Furthermore there is the potential that persistent 

and toxic micropollutants and transformation products could 

build up during the recycling process. 

Urban and agricultural wastewaters contain a high 

number of organic micropollutants, including pharmaceuticals 

and personal care products, hormones and industrial 

chemicals, biocides and pesticides (Schwarzenbach et al., 

2006). A large number of studies have investigated the fate 

and removal of various individual and specific groups of 

micropollutants by treatment processes like biological activated 

sludge treatment (Ternes et al., 2004; Joss et al., 2005), 

ozonation and advanced oxidation processes (von Gunten, 

2003; Huber et al., 2004; Huber et al., 2005; Lee et al., 2008) or 

activated carbon treatment (Sanchez-Polo et al., 2006; Snyder 

et al., 2007) but little is known how these treatment processes 

change the composition and biological activity of the 

micropollutants. 

Toxicity testing may provide complementary information 

to chemical analysis on the sum of micropollutants present in 

treated water. However, whole-effluent toxicity testing, which 

is a valuable tool for hazard assessment of industrial and 

municipal effluents (Chapman, 2000), is inappropriate to 

evaluate the fate of micropollutants during advanced water 

treatment because acute toxicity tests are not sensitive 

enough for purified water and long-term chronic tests are too 

expensive and time-consuming for routine monitoring. In 

addition, the physical treatment of wastewater alters the 

matrix and electrolyte composition and it becomes virtually 

impossible to differentiate the effect caused by micropollutants 

from that induced by the matrix, such as organic 

matter content, pH, ionic strength or nutrient content. 

Therefore, over the last years, we have developed 

a framework for bioanalytical quantification of organic 

micropollutants, which has three main cornerstones: Firstly, 

the aqueous samples are enriched using solid-phase extraction 

(SPE) to separate the organic micropollutants of interest 

from metals and matrix components. This SPE method was 

previously validated for recovery and extraction yield for 

selected chemicals including pesticides and pharmaceuticals 

(Bengtson Nash et al., 2005a; Escher et al., 2005; Escher et al., 

2008a). Secondly, a battery of bioanalytical methods were 

selected that cover a non-specific bioassay, the bioluminescence 

inhibition of the marine bacterium Vibrio fischeri and 


five bioassays for specific modes of toxic action that are 

indicative for groups of chemicals of particular relevance for 

human and environmental health, including aspects of 

estrogenicity, dioxin-like activity, genotoxicity, neurotoxicity, 

and phytotoxicity (Muller et al., 2007; Muller et al., 2008; 

Escher et al., 2008a; Escher et al., in press). Table 1 summarizes 

the key features of the bioassays and more details are 

given in the Supporting Information. 

Finally,we developed a coherent data evaluation method that 

is based on the toxic equivalency concept (Villeneuve et al., 2000; 

Escher et al., 2008b). Thus, the effects at various steps of the 

treatment chain can be easily compared and treatment efficiency 

can be compared between the different effect endpoints. 

Bioassays in combination with SPE have been successfully 

applied by several groups to evaluate various water treatment 

options (Escher et al., 2008a; Cao et al., 2009). In particular, the 

bioluminescence inhibition test with Vibrio fischeri and the 

Salmonella based test for genotoxicity have been widely used 

despite their recognized limitations (Johnson, 2005; Petala 

et al., 2006; Petala et al., 2008). The present work is more 

extensive covering several additional toxicity endpoints, 

allowing high-throughput monitoring applications while 

remaining cost-efficient. 

Here we present this framework for the first time to evaluate 

all steps of enhanced water treatment in the South 

Caboolture Water Reclamation Plant (van Leeuwen et al., 

2003). This paper focuses on the advantages and limitations in 

the application of the bioanalytical test battery, and the 

interpretation and significance of the results. In the accompanying 

paper, the results obtained from the bioanalytical 

tools are compared with chemical analytical data for a variety 

of pharmaceuticals and pesticides and conclusions for 

advanced wastewater treatment are drawn (Reungoat et al., 

2010). 


2.1. Samples and sites 

Four consecutive sets of 24-hour composite samples were 

collected on 11-07-08, 22-07-08, 27-07-08 and 06-08-08 using 

refrigerated autosamplers at 10 sites of the South Caboolture 

Water Reclamation Plant (Table 2). 

2.2. Solid phase extraction 

Immediately after sampling, concentrated HCl (36%) was 

added to a final concentration of 5 mM for preservation. It was 

demonstrated in earlier work that a pharmaceutical cocktail 

in a wastewater matrix had highest recoveries for HLB at pH 3 

(Escher et al., 2005). 

Samples were extracted using 1 g OASIS Ò HLB solid phase 

material in 20 mL cartridges (Waters, Australia) following 

centrifugation (only S1 sample) and filtration with a glass fibre 

filter. After conditioning the cartridges with 10 mL methanol 

and 20 mL of 5 mM HCl in MilliQ water, a known volume of 

sample was percolated under vacuum. The cartridges were 

dried overnight under vacuum and were eluted with 10 mL

Table 1 – Bioanalytical test battery (adapted from Muller (2008)). 

Assay Measured by Target mode of 

toxic action 

Bioluminescence 

inhibition test 

Acetylcholinesterase 

(AChE) 

Inhibition Assay 

Imaging-PAM 

Assay 

methanol and 10 mL hexane:acetone (1:1). All eluates were 

evaporated to dryness and reconstituted with 0.5 mL of ethanol. 

2.3. Bioanalytical tools 

The bioassays were performed as previously reported (Escher 

et al., 2005; Muller et al., 2007; Muller et al., 2008; Escher et al., 

2008a; Escher et al., in press) using the methods cited in Table 

1 and described in detail in the Supporting Information. 

2.4. Relative enrichment factor of the samples REF 

For each sample, the enrichment factor of the solid phase 

extraction was calculated using (Eq. 1) which represents the 

enrichment of the extract compared to the source water. 

enrichment factorSPE ¼ Vwater 

Reduction in luminescence of 

the naturally bioluminescent 

marine bacteria Vibrio fischeri 

Product of the enzymatic 

reaction (colorimetric) 

Increased Chl a fluorescence, 

inversely proportional to PS II 

photosynthetic yield 

E-SCREEN Cellular proliferation in 

estrogenic dependent cell line 

AhR-CAFLUX Induction of green 

fluorescent protein under the 

control of Ah receptor; 

(fluorescence) 

umuC assay Induction of ß-galactosidase 

enzyme as an indicator of 

DNA damage; 

Product of the enzymatic 

reaction (colorimetric) 

Vextract 

Sample volume varied from 2.0 to 2.5 L. The final volume of 

each extract was 0.5 mL therefore the enrichment factor SPE 

was between 3500 and 5500 depending on the sample. 

An aliquot of the enriched sample extracts were then 

added to the microtiter plate of the respective bioassay and 

serially diluted by a test medium to obtain a concentrationeffect 

curve. A dilution factor of each bioassay was calculated 

using (Eq. 2). The dilution factors ranged from 0.05 to 0.007 for 

the highest dose depending on the bioassay. 


Effect on energy 

status through 

cytotoxicity and/or 

baseline toxicity 

Inhibition of the 

enzyme, which 

hydrolyses the 

neurotransmitter 

acetylcholine 

PS II derived 

photosynthesis 

inhibition 

(1) 

Chemicals that 

give response 

Literature reference 

All chemicals International Standard 

Organisation, 1998; 

Johnson, 2005; 

Farré et al., 2006; Escher 

et al., 2008a 

Organophosphates 

and carbamate 

insecticides 

Triazine and phenylurea 

herbicides 

Estrogenicity Estrogens, 

estrogenic industrial 

chemicals 

Dioxin-like activity; 

Aryl hydrocarbon 

(Ah) receptor 

activation 

Genotoxicity; 

DNA damage 

(SOS-response) 

Polychlorinated 

dibenzodioxins/ 

furans (PCDD/F) 

and biphenyls (PCB) 

or polycyclic 

aromatic 

hydrocarbons (PAH) 

Chlorinated 

byproducts, 

aromatic amines, 

polycyclic aromatic 

hydrocarbons (PAH) 

Ellman et al., 1961; 

Deutsche Norm, 1995; 

Hamers et al., 2000 

Schreiber et al., 2002; 

Schreiber et al., 2007 

Soto et al., 1995; Körner 

et al., 1999 

Nagy et al., 2002; Zhao 

and Denison, 2004 

Oda et al., 1985; 

Reifferscheid et al., 1991; 

International Standard 

Organisation, 2000 

volume of extract added to bioassay 

dilution factorbioassay ¼ 

total volume of bioassay 

(2) 

The final relative enrichment factor REF is the combination 

of the enrichment of the extraction and the dilution in the 

bioassay (Eq. 3) (Escher et al., 2006; Muller et al., 2007) and 

represents the enrichment or dilution of the original sample 

in each bioassay (Table 3). The REF is equivalent to concentration 

and is expressed in the units [Lwater sample/Lbioassay]. 

REF ¼ dilution factorbioassay enrichment factorSPE (3) 

2.5. Evaluation of the concentration-effect curves 

Due to the nature of the endpoints, the six bioanalytical tests 

of the battery could not be evaluated with a single data evaluation 

model (Table 3). In general, all concentration effect 

curves of the reference compounds and sample extracts followed 

a log-logistic function (Eq. 4), which was fitted using 

Prism 5.0 software (GraphPad, San Diego, CA, USA). Adjustable 

parameters are the minimum and maximum of the effect, 

min and max, the slope s, and the effect concentration 

inducing 50% of the maximum effect, EC50. If the concentration-effect 

curve dropped sharply at the range of higher doses 

indicating that the sample has a cytotoxic effect, the cytotoxic 

concentrations were excluded from the concentration-effect 

assessment.

480 

Table 2 – Sample details. Samples were each taken after the described process. More details are given in Reungoat et al. 

(2010). TOC/DOC data represent the average of two samplings (22-07-08 and 06-08-08). 

Sample ID Site description TOC (mg/L) DOC (mg/L) 

effect¼min 

þ 

max min 

1þ10 s ðlog EC 50 log ðconcentrationof reference compoundor REF of sampleÞÞ 

Three bioassays, the inhibition of bioluminescence of Vibrio 

fischeri, inhibition of AChE and inhibition of PSII photosynthetic 

yield, all produced effects between 0% and 100%, therefore the 

maximum effect in (Eq. 4) could be set to 100% (Table 3). 

Two bioassays, the AhR-CAFLUX and the E-SCREEN, yield 

information on the induction of a receptor or proliferation of 

cells and there is no absolute maximum effect. Dependent on 

the mode of action, the maximum effect can be higher or 

lower than the maximum effect of the reference compounds. 

The extent of difference of maximally attainable effect yields 

some additional information on the mode of action and the 

type of compounds that induce the effect. Details of each 

assay are discussed below. The curve fit is then performed in 

such a way that the maximum effect in equation 4 is an 

additional fitting parameter. 

Finally the umuC assay exhibited a linear concentrationeffect 

curve. At low effect levels the log-logistic concentration 

effect curve (Eq. 4) is congruent with a linear concentrationeffect 

curve. Unlike the other assays, where effects were 

compared to a reference compound, genotoxicity is expressed 

as the concentration that induces the threshold of induction 

defined by the EN ISO guideline (International Standard 

Organisation, 2000). This way of data evaluation needs further 

refinement in the future because it is not directly comparable 

to the results in the other bioassays and the effect concentration 

cannot be translated into TEQ. 

The detection limit of the bioassays was defined as three 

times the standard deviation of the response using the lowest 

concentration of the standard that induced an effect significantly 

different from the control. 

2.6. Estimating equivalent concentrations 

in the samples 

For an easy-to-follow reporting of the effect, in all but one 

bioassay, the effects were reported in terms of toxic equivalent 

(4) 

concentrations (Villeneuve et al., 2000; Escher et al., 2008b). The 

equivalent concentration represents the concentration of 

the reference compound required to produce the same effect as 

the mixture of different compounds in the sample. 

Equivalent concentrations were calculated from concentration-effect 

curves of the reference compound and the 

samples. Both reference compounds and samples generally 

followed a sigmoidal log–concentration-effect curve (Eq. 4), 

which is equivalent to a linear function with respect to nonlogarithmic 

concentration at small effect levels. Sample 

extracts are normally composed of a mixture of unknown 

substances at unknown concentrations, so the concentration 

unit of the sample in the concentration effect curve is based 

on the Relative Enrichment Factor (REF) in units of 

[Lwater sample/Lbioassay]. 

Toxic equivalent concentrations (TEQ) were calculated as 

the ratio of EC50 values of the reference compound to the EC50 

of the sample (Eq. 5). 

TEQ ¼ EC50ðreference compoundÞ 

EC50ðsampleÞ 

2 g 

3 

Lbioassay 4 ¼ 

g 

5 ð5Þ 

L water sample 

L bioassay 

Lwater sample 

Since the EC50(reference compound) is expressed in concentration 

units such as [g/Lbioassay] and the EC50(sample) in 

concentration units [Lwater sample/Lbioassay], representing the 

enrichment or dilution (REF) of the sample required to elicit the 

50% effect, the toxic equivalent concentration is expressed in 

[g/Lwater sample]. 

In principle, any effect level can be used to derive TEQ, 

provided that the concentration-effect curves are parallel i.e. 

have the same slope, which is in fact one of the prerequisites for 

the validity of the toxic equivalency concept (Villeneuve et al., 

2000). Therefore, if the effect of the sample in the bioassay does 

not reach 50%, toxic equivalency can be estimated using EC20 of 

the sample and EC20 of the corresponding reference compound. 

2.7. QA/QC 

avg sd avg sd 

Blank MilliQ water n.a. n.a. n.a. n.a. 

S1 Influent (Effluent from WWTP) 20 4 17 4 

S2 Denitrification 24 n.a. 21 n.a. 

S3 Pre-Ozonation 21 7 21 4 

S4 Coagulation/flocculation/DAFF 11 1 11 0.5 

S5 Main ozonation 9.9 0.9 10 0.9 

S6 Activated carbon filtration 6.6 0.6 6.6 0.9 

S7 Effluent 7.1 0.5 6.8 0.7 

C3 Biological sand filter fed with S4 water 8.5 1 8.8 0.8 

C4 Biological activated carbon filter fed with S4 water 4.0 0.4 4.2 0.04 

C5 Biological activated carbon filter fed with S5 water 3.6 0.07 4.1 0.08 

avg - average; sd - standard deviation; n.a. - not applicable. 


To determine background levels of contamination associated 

with the extraction process as well as assessing any effect of

Table 3 – Evaluation of the concentration effect curves and expression of the bioanalytical results. 

Assay Concentration-effect 

assessment 

Baseline Toxicity – 

Bioluminescence 

inhibition in 

Vibrio 

fischeri 

Neurotoxicity – 

AChE 

Phytotoxicity – 

Max-I-PAM 

Estrogenicity – 

E-SCREEN 

Ah-Receptor – 

AhR-CAFLUX 

Genotoxicity – 

Umu 

Top: 100 

Bottom: 0 

Slope: 1 

Top: w 

Bottom: w 

Slope: 1 

Top: at least 

100 and then w 

Bottom: w 

Slope: 1 

the solvent, 2.0 L of MilliQ water were extracted identically to 

the samples and analysed in all bioassays as a procedural 

blank. 

For quality control and assurance purposes two replicates 

of randomly selected samples (2 sites per each sampling; for 

details see Supporting information Table S1–S6) were 

collected and analysed in all bioassays on a different day than 

the first replicate to evaluate repeatability of the SPE and the 

bioassays. Each sample extract was tested in the bioassays in 

duplicates or triplicates per run, depending on the assay. 

All individual data sets are summarised in the Supporting 

information Table S1–S6. Since the variability among the four 

sampling events was not higher than the variability between 

the sample replicates collected at the same time, we report in 

the main manuscript the average sd of four sample extracts 

at four different consecutive sampling events (n ¼ 4, whith 

exception of S2, where n ¼ 3). Analysis of variance (ANOVA) 

was used to test the differences among the average TEQ of 

samples S1-C5, including the blank, in comparison to S1. 


REF 

range 

3.1. Baseline toxicity – bioluminescence inhibition 

in vibrio fischeri 

Reference compound; 

Positive control 

0.4–38 Virtual baseline 

toxicant; 

Phenol a 

The EC50 of phenol was 130 37 mg/L over 8 plates in 3 different 

repetitions of the assay, confirming the reproducibility and 

repeatability of the assay. While in all other assays, the reference 

compound served two purposes, as quality control and to 

define the reference for the TEQ, in the bioluminescence inhibition 

test, phenol was only used for quality control. 

EC 50 of the 

reference 

compound 

12 mg/L; 

130 mg/L 

Result 

expression 

Baseline toxicity 

equivalent 

concentrations 

(baseline-TEQ) 

0.04–115 Parathion 120 mg/l Parathion equivalent 

concentrations (PTEQ) 

1.5–258 Diuron 16 mg/L Diuron equivalent 


(DEQ) 

0.0006–55 17ß-Estradiol (E2) 2 ng/L Estradiol equivalent 

concentrations (EEQ); 

Relative proliferative 

effect (RPE) 

0.004–55 2,3,7,8- 

Tetrachlorodibenzodioxin 

(TCDD) 

Linear regression 2.2–110 (-S9) 2-amino 

anthracene a 

(þS9) 4-nitroquinoline- 

N-oxide a 

a Only used as positive control, not as reference compound. 

b Quantification limit; w variable; n.a. - not applicable. 


8 ng/L TCDD equivalent 


(TCDDEQ) 

n.a. REF that elicits genotoxic 

effect with threshold 

Induction Ratio of 

1.5 

Detection limit 

20% inhibition 

of bioluminescence 

0.3 mg/L 

0.01 mg/L 

0.02 ng/L 

0.04 ng/L 

Maximum REF 

achieved in the 

assay b 

Toxicity of the sample was expressed in baseline toxicity 

equivalents (baseline-TEQ) derived from a baseline toxicity 

QSAR (quantitative structure-activity relationship) (Fig. 1A) 

using a virtual compound with log K ow of 3 and molecular 

weight of 300 g/mol as a reference, which equates to an EC50 of 

12 mg/L (Escher et al., 2008b). The rationale behind this choice 

is the fact that there is not a singular positive control, as every 

chemical exhibits baseline toxicity. For specifically acting 

compounds the baseline effect is marginal and cannot be 

resolved if the single chemical is tested. However, in a mixture 

with a large number of chemicals with a variety of specific 

modes of action, as is the case in a wastewater sample, the 

baseline toxicity might actually dominate the overall mixture 

effect (Warne and Hawker, 1995). Thus, baseline toxicity will 

provide an integrative measure of the combination of chemicals 

that act together in the mixture and each chemical’s 

contribution is essentially weighted by its hydrophobicity only 

(Escher and Schwarzenbach, 2002). Fig. 1A depicts the relationship 

between hydrophobicity and effect concentration on 

a log-log plot and how the EC50 of a virtual baseline toxicant is 

derived (note that the hydrophobicity scale is based on liposome-water 

partitioning not octanol-water partitioning), 

therefore the virtual baseline toxicant with log K ow of 3 is 

situated slightly higher than 3 on the log-hydrophobicity 

scale. 

Samples were tested for baseline toxicity at 8 different 

concentrations after serial dilution 1:2, with the REF ranging 

from 0.4 to 38. All previously tested baseline toxicants 

exhibited a common slope of 1 (Escher et al., 2008b), and 

therefore the slope of the samples was fixed to 1, too. Effect 

concentrations (EC50) of the samples used to calculate the 

toxic equivalent concentrations were expressed in REFs,

482 

Fig. 1 – Concentration-effect curves of the reference compounds and selected samples tested in bioluminescence inhibition 

assay (A, B), AChE inhibition assay (C, D) and I-PAM (E, F) assay. Bottom and top of each curve was fixed to 0 and 100, 

respectively. Quality of the fit expressed as R 2 was >0.90, except for S7-effluent sample tested in AChE assay, where the R 2 

was 0.74. (A) Baseline toxicity QSAR for the bioluminescence inhibition test with Vibrio fischeri and derivation of the EC50 of 

the virtual baseline toxicant that is used as reference compound in this assay (data and QSAR from Escher et al. (2008b)). 

Error bars indicate SD. 

representing the enrichment or dilution of the sample 

required to elicit the 50% effect, and are summarised in Table 4. 

The EC 50 of the blank was 59, which means that if MilliQ 

water were enriched 59 times, the blank extract would elicit 

50% bioluminescence inhibition. This EC 50 was extrapolated 

from experimental data, where the highest measured REF was 

38, which caused a 25% effect. It is not reasonable to enrich the 

samples to a higher REF because the enrichment of impurities 

of solvents and materials or contamination during the 

enrichment procedure would mask the measurable effect. 

The influent, denitrification and ozonation samples all 

exhibited EC50 in the range of 4–5.6, i.e. the samples had to be 


enriched by approximately five times to see an effect. The 

concentration window for a visible effect is relatively narrow. 

A sample that is not enriched (REF ¼ 1) or even diluted would 

not give any visible response. The necessary enrichment is not 

the only reason for using SPE for sample preparation: 

a majority of matrix compounds are removed by SPE, such as 

salts and particulates. All organisms are very sensitive to 

variable electrolyte concentration. In this assay too little salt 

would decrease the bioluminescence of the marine bacterium 

(Escher et al., 2008a), and salting up is difficult because the salt 

content of water samples is unknown. In contrast to salts and 

particulate matter, we assume that dissolved organic carbon

Table 4 – Sample enrichment required to produce a 50% effect in each assay expressed as the average ± sd of four 24 h 

sampling events. 

ID Site description Bioluminescence inhibition AChE I-PAM 

(DOC), in particular the low molecular weight fraction, is 

partially retained by SPE. 

Baseline-TEQ increased slightly from influent (2.3 mg/L) to 

denitrification (2.8 mg/L) and after pre–ozonation (3.2 mg/L). 

However the stepwise increase was not statistically significant 

(ANOVA, p ¼ 0.05). Significant decrease by more than 

a factor of two in TEQ was observed after coagulation/flocculation/DAFF 

(1.4 mg/L). Baseline-TEQ after the main ozonation, 

activated carbon filtration and in the effluent ranged 


EC50 (REF) EC50 (REF) EC50 (REF) 

avg sd avg sd avg sd 

Blank 59 0.81 BDL (

484 

Fig. 2 – Relationship between DOC concentration and 

baseline-TEQ. The black diamonds are the data from this 

study (from Tables 1 and 5). The empty squares refer to 

a previous study performed in Switzerland (all data refer to 

secondary effluent) (Escher et al., in press). Error bars 

indicate SD. 

matter that is co-extracted with SPE. Assimilable organic carbon 

may be taken up by microorganisms and contribute to bioluminescence 

inhibition, while higher molecular-weight fractions 

of organic matter may bind micropollutants, potentially 

causing a decrease in their bioavailability and thus decrease the 

response. Note also that organic micropollutants contribute to 

DOC in DOC measurements. Therefore it is virtually impossible 

to differentiate between assimilable organic carbon and organic 

micropollutants. 

The TEQ of the influent of the reclamation plant, which is the 

effluent of a conventional wastewater treatment plant (WWTP), 

was higher than of the secondary effluent of a WWTP in 

Switzerland (Escher et al., 2008a; Escher et al., in press). It must 

also be noted that the DOC levels were significantly lower in the 

Swiss WWTP effluent than in the present study. In fact, 

the baseline-TEQ of the Swiss study (Escher et al., in press)fallon 

the lower left corner of Fig. 2, but they fit the correlation. This 

observation supports the conclusion that DOC contributed to 

the effects measured in this bioassay. 

3.2. Neurotoxicity – acetylcholinesterase 

inhibition assay 

The concentration-effect curve of the reference compound, 

parathion, followed a log-logistic function with the slope 1 and 

the EC50 of 120 32 mg/L (Fig. 1C), using a commercially available 

acetylcholinesterase enzyme from Electrophorus electricus. 

The EC50 of parathion in our work was higher than data 

reported in literature: 27 mg/L (Escher et al., in press) and 26 mg/L 

(Escher et al., 2008a) with the use of bovine acetylcholinesterase. 

The difference in the sensitivity of the assay might be 

due to a different origin of acetylcholinesterase enzyme. 

The samples were tested at 12 different concentrations 

after serial dilution 1:2 with the REF range from 0.04 to 115. 

The best-fit slope of the samples’ concentration-effect curve 

was 1 (Fig. 1D) as it was for the reference compound. Effective 

concentrations required to elicit 50% inhibition are summarised 

in Table 4. Influent samples had to be enriched 52 times to 

elicit 50% inhibition, in contrast to effluent, which would 


require an enrichment of 1100 times to elicit the same effect. 

As discussed above, we did not apply such high enrichment 

factors but had to extrapolate the EC50. The effect of the blank 

and samples treated with activated carbon was less than 20%, 

defined as the detection limit of the assay. Therefore the 

effective concentrations were not extrapolated from the 

concentration-effect curves. 

Results of the acetylcholinesterase inhibition assay were 

expressed as parathion equivalent concentrations (PTEQ) and 

are summarised in Table 5. PTEQ of the samples showed firstly 

a slight increase after pre-ozonation to 4.2 mg/L in comparison 

with the influent PTEQ of 3.1 mg/L, and then a significant 

decrease to 1.9 mg/L after the main ozonation. Activated 

carbon filtration further reduced PTEQ below the detection 

limit (

measurement is somewhat compromised, when several nonspecifically 

acting chemicals in the mixture mask the phytotoxic 

effect (Escher et al., 2008a). In fact, when single baseline 

toxicants (Escher et al., 2008b) were tested for the inhibition of 

photosynthesis yield they showed a response when 

measuring Y after 2 hours incubation, albeit at higher 

concentrations than for the corresponding growth inhibition 

endpoint, and the curve was steeper with a slope of 1.9. Since 

it is impossible to quantitatively differentiate between the 

contributions from phytotoxicity and cytotoxicity (unless 

cytotoxicity is so overwhelming that it makes the photosynthesis 

yield measurements impossible), we used the slope of 

the reference compound diuron to fit the samples even 

though the quality of fit was sub-optimal. 

The phytotoxic response of the samples after 2 hours of 

incubation expressed as diuron equivalent concentration DEQ 

is summarised in Table 5. Denitrification significantly 

increased the DEQ (0.29 mg/L) in comparison with the influent 

DEQ of 0.11 mg/L. The pre-ozonation step itself had no significant 

effect (0.34 mg/L). After coagulation/flocculation/DAFF, 

the DEQ decreased by 85% to 0.05 mg/L. DEQ decreased in small 

steps after the main ozonation and activated carbon filtration 

stages and in the effluent samples. No significant additional 

decrease in DEQ was observed after sand filter or biological 

activated carbon filters. All DEQ after coagulation/flocculation/DAFF 

are close to the detection limit and are based on 


Fig. 4 – Concentration-effect curves of selected samples and corresponding standards tested in E–SCREEN (A, B) and AhR- 

CAFLUX (C, D) assay. The slope of the curves was fixed to 1. Bottom of the curves was an adjustable parameter in both 

assays. The top of the curve was adjustable in the E–SCREEN assay, while in AhR-CAFLUX, the top was set to the level of at 

least 100% of the TCDD maximal effect. Quality of the fit expressed as R 2 ranged for selected samples and reference 

compounds from 0.96 to 0.99, except for Blank and S7-effluent samples tested in E-SCREEN assay, where the R 2 was 0.3 and 

0.4, respectively due to a very low response of both samples. Error bars indicate SD. 

a single data point, therefore should be interpreted with some 

caution. 

The initial slight but not statistically significant increase in 

TEQ from sample S1 to S3 is unexpected at first sight. It is 

interesting to note, however, that this small increase was 

common for baseline toxicity and slightly less clear also for 

neurotoxicity (Figs. 2 and 3). The increase in baseline toxicity 

was explained by an increase of DOC, which is presumably 

partially co-extracted at low pH. The small increase in DEQ 

could also be caused by baseline toxicants interfering with the 

measurement of the photosynthesis yield. We have not 

observed such an effect in our previous work (Escher et al., 

2008a; Escher et al., in press). However, we have never 

encountered DOC concentrations exceeding 10 mg/L in any of 

our previous work. It is unclear how much of the DOC is coextracted 

with the micropollutants by SPE. The effect of DOC 

needs to be investigated in future work but is beyond the 

scope of the present study. Possibly, extraction at a higher pH 

might reduce the fraction of co-extracted DOC due to the 

higher negative charge of DOC under these conditions. 

Previous work on comparing herbicides concentrations 

with DEQ often showed a relatively good correlation and 

a substantial fraction (> 50% of the response) could be 

explained by the measured herbicides (Bengtson Nash et al., 

2005b; Bengtson Nash et al., 2006; Escher et al., 2006; Muller 

et al., 2008). There were some cases, such as samples collected

486 

Table 6 – Sample enrichment required to produce a defined effect in each assay: 50% effect in E–SCREEN assay and 20% of 

maximal TCDD effect in AhR-CAFLUX assay. Results represent the average ± sd of 4 samplings. 

ID Site description E-SCREEN AhR-CAFLUX 

from the Thames River, where the biological DEQ were much 

higher than the chemically analysed DEQ (Bengtson Nash 

et al., 2006). The authors suggested this difference was due to 

the presence of additional herbicides that were not analysed 

with the analytical chemical methodology, but given the new 

evidence it is also possible that additional cytotoxicity by nonspecifically 

acting micropollutants and DOC could have 

contributed to the discrepancy. 

3.4. Estrogenic activity: E-SCREEN assay 

In the E-SCREEN, bottom and top of the concentration-effect 

curve (Eq. 4), both adjustable parameters, refer to the 

minimum or maximum stimulatory response, i.e., proliferation 

of the estrogen-dependent cells in E–SCREEN. The slopes 

of the concentration-effect curves of the reference compound, 

17ß-estradiol, and the samples were fixed to 1 (Figs. 4A and B). 

The maximum of the curve yields information on the mode of 

action, thus being able to distinguish among full agonist, 


Before Clean-up After Clean-up 

RPE Mode of Action EC50 (REF) EC20 TCDD (REF) EC20 TCDD (REF) 

avg sd avg sd avg sd avg sd 

Blank 0.05 – Not estrogenic BDL (

Table 7 – Sample enrichment required to produce 

a genotoxic effect [ Induction Ratio of 1.5 (IR1.5) in umuC 

assay without and with metabolic activation. Results 

represent the average ± sd of 4 samplings. 

ID Site description umuC -S9 umuC þ S9 

EC IR1.5 (REF) EC IR1.5 (REF) 

avg sd avg sd 

Blank >78 >78 

S1 Influent (Effluent from WWTP) 5.2 2 17 3 

S2 Denitrification 6.9 2 22 3 

S3 Pre-ozonation 6.3 2 18 3 

S4 Coagulation/flocculation/DAFF 15 7 28 8 

S5 Main ozonation 78 20 >87 

S6 Activated carbon filtration >84 >84 

S7 Effluent >69 >73 

C3 Biological sand 

filter fed with S4 water 

24 13 34 16 

C4 Biological activated 

carbon filter 

fed with S4 water 

>80 >80 

C5 Biological activated 

carbon filter 

fed with S5 water 

>79 >79 

(1667 times dilution) to 55 (55 times enrichment of the original 

water sample). Sample REF required to produce a 50% effect 

(EC50) in the E–SCREEN assay ranged from 0.27 (equivalent to 4 

times dilution of the original water sample) to 6.7 (equivalent to 

6.7 times enrichment of the original water sample) (Table 3). The 

wide range of REF windows allows one to monitor a wide range 

of samples collected throughout the treatment chain. 

The estrogenicity of the samples was quantified with two 

parameters: the relative proliferative effect (RPE) and the 

estradiol equivalent concentration (EEQ). RPE of the samples, 

representing the ratio of the maximum proliferation induced 

by test samples compared to 17b-estradiol (also referred to as 

relative efficacy), are summarised in Table 6. The samples from 

the influent up to ozonation were fully or partially agonistic, 

while the remainder of the samples were not classified estrogenic 

and no EEQ could be derived for these samples. 

When a sample was classified as either a partial or full 

agonist, an estradiol equivalent concentration (EEQ) was 

calculated from the EC 50. Influent to the reclamation plant 

(representing the effluent of the WWTP) showed full agonistic 

activity (RPE > 0.8) with the estradiol equivalent concentration 

EEQ of 6 ng/L (Table 5). This was comparable with the EEQs 

reported for WWTP effluent in the literature: 2 ng/L) and the resulting effluent of the ozonation contained 

0.6 ng/L EEQ (Escher et al., in press). 

Estrogenicity was further markedly altered after activated 

carbon filtration. Estrogenic effect of the sample treated with 

activated carbon (S6) was less than 20% of the estradiol effect 

(RPE < 0.2), which corresponds to the detection limit of the 

assay, therefore the EEQ was less than 0.02 ng/L. 

Estrogenicity (RPE) of the effluent sample was slightly 

increased, but did not reach 50% effect of estradiol, therefore 

the sample was classified as not estrogenic and the EEQ was 

not quantified. The estrogenic effect was below quantification 

limit. Taking into account the REF of the sample in this assay, 

the EEQ of the effluent was less than 0.06 ng/L. 

Besides the main treatment chain, additional treatment 

processes were assessed in this study: sand filtration of the 

coagulation/flocculation/DAFF effluent (C3) and biological 

activated carbon filtration of the coagulation/flocculation/ 

DAFF effluent (C4) and main ozonation effluent (C5). Sand 

filtration significantly decreased the EEQ of the coagulation/ 

flocculation/DAFF effluent from 9.8 ng/L to 0.63 ng/L. Biological 

activated carbon filtration decreased the estrogenic effect 

of both effluents to below detection limit (

488 

reproducibility and repeatability of the assay and is in good 

agreement with the previous levels from the literature of 

5.9 ng/L (Nagy et al., 2002) and 3.9 ng/L (Zhao and Denison, 

2004). 

TCDD was chosen as the reference compound as it is one of 

the most potent agonists for this assay. However, high affinity 

ligands for the AhR also include dibenzofurans and biphenyls as 

well as a variety of polycyclic aromatic hydrocarbons, benzoflavones 

and other chemicals (Denison and Nagy, 2003). To 

assess the contribution of chemicals other than dioxins, furans 

and dioxin-like PCBs to the potency estimates, sulfuric acid silica 

gel clean-up was performed in this study, which removed most 

organic chemicals (e.g. PAHs) except persistent chemicals such 

as dioxins, furans and PCBs. Original sample extracts, as well as 

the extracts cleaned with sulfuric acid silica gel were tested with 

the AhR-CAFLUX assay at 5 different concentrations after serial 

dilution 1:10, with the REF ranging from 55 to 0.004. The response 

of many samples, especially after clean-up, did not reach 50% 

TCDD response. Therefore effective concentrations of the 

samples required to elicit 20% effect of TCDD (EC 20 TCDD) were 

interpolated from concentration-effect curves (Table 6). The EC 20 

TCDD of the samples ranged from 1.4 to 8.2. The effective 

concentration of the same samples was significantly altered 

after clean up and ranged from 14 to 22, i.e. samples after acid 

silica gel clean up had to be enriched approximately 3–10 times 

more to elicit the same effect. 

The response in the AhR-CAFLUX assay was expressed as the 

2,3,7,8-TCDD-equivalent concentration (TCDDEQ). TCDDEQ was 

significantly decreased by the main ozonation to 0.44 ng/L in 

comparison with the TCDDEQ of 0.83 ng/L in the influent. Sand 

filtration did not change the TCDDEQ of the coagulation/flocculation/DAFF 

effluent, while treatment with the biological activated 

carbon filters reduced the TCDD equivalent concentration 

to 0.33 ng/L a level not significantly different from the blank. 

The TCDDEQ of the samples after the sulfuric acid silica gel 

clean-up were reduced to levels not significantly different 

from the blank (Table 5). These results indicate that the 

observed effect in the AhR-CAFLUX assay prior to the sulfuric 

acid silica gel clean-up step may be attributed to chemicals 

other than dioxins, furans and PCBs. 

3.6. Genotoxicity-UmuC assay 

The umuC genotoxicity assay can detect both cytotoxic and 

genotoxic effects. The extracts were tested both with (þS9) and 

without ( S9) exogenous metabolic activation. Response on 

this assay is determined as an induction ratio (IR) (Fig. 5). An 

IR 1.5 is considered genotoxic, providing the sample is not 

cytotoxic (growth < 0.5). For samples that demonstrate significant 

genotoxic response, the effect concentrations EC IR 1.5 

were derived from the linear concentration-effect curve as 

depicted in Fig. 5. Effect concentrations are in dimensionless 

units of REF and represent how many times the samples must 

be concentrated or diluted to elicit a threshold IR of 1.5 in the 

assay (Table 7). Results are expressed as 1/ECIR 1.5 therefore 

a higher number represents a higher genotoxic effect (Table 5). 

Unfortunately no TEQ could be derived for this endpoint but in 

theory this should be possible and in the future we will work on 

implementing the TEQ concept for this assay. This is somehow 


problematic because of the frequently observed cytotoxicity of 

the samples, which can only be partially accounted for. 

Overall, the umuC test without metabolic activation gave 

higher responses than the test with metabolic activation 

indicating that some of the micropollutants that are responsible 

for genotoxicity are detoxified by the liver enzyme fraction 

S9. As in all other bioassays, the denitrification and preozonation 

did not affect the activity in the umuC test but 

coagulation/flocculation/DAFF decreased the genotoxicity 

without S9 by 58% and that with S9 by 35%. Subsequent main 

ozonation completely removed the effect for the test with prior 

S9 treatment and drastically reduced it in the absence of S9. 

4. Conclusion 

A similar test battery has been previously used for various 

applications in wastewater treatment efficiency and water 

quality assessment. To apply the test battery to a treatment 

train with very subtle steps and for relatively pure water was 

a challenge. This challenge was met by the following measures: 

A comprehensive framework was developed that centres 

around the bioassays but links to sample preparation with 

SPE prior to testing and to the data evaluation scheme. 

Quality control was implemented and all effects were 

related to reference compounds, thereby correcting for 

slight day-to-day variability of the test results. 

The use of SPE allows a relatively non-specific extraction of 

water samples to concentrate a wide range of relevant 

pollutants to levels where they can exhibit a measurable 

response, while matrix components such as salts and 

metals are removed from the sample during SPE. 

The detection window of the bioassays is adjustable. For the 

application in the present study the concentration range of 

the sample in the bioassay ranged between 1700 times dilution 

and 250 times enrichment of the sample. The detection 

window depends on the constitution of the sample and the 

type of bioassay and is only limited by the effect of the blank 

at very high enrichment and nonspecific cytotoxicity of the 

sample for each of the assays. By dynamically changing 

the relative enrichment window of the sample throughout 

the study one can adapt the various bioassays to a wide 

variety of samples and results are robust and reproducible. 

Results are reported as toxic equivalent concentrations, 

a format that allows for comparison to results from chemical 

analysis. This is also a key feature of the bioanalytical toolsthey 

are no direct indicators for environmental health but 

they are indicators of the presence of mixtures of chemicals, 

accounting for the mixture toxicity of the unidentified 

chemicals in an environmental sample. To report bioanalytical 

results as toxic equivalent concentrations is better 

suited for hazard assessment and extrapolation than the 

commonly reported % effect of the water sample in a given 

bioassay because it is too difficult to extrapolate effects while 

it is possible to compare effect concentrations of chemicals 

across different level of biological organisation. So for 

example, if a sample elicits 0.5 mg/L DEQ one can compare 

this value with the environmental quality standard for

diuron to estimate what potential long-term effects could be 

expected from such an exposure. 

The study also highlighted some difficulties that remain and 

some lessons are to be learned. One of the weak points of the 

approach is the sample preparation with SPE. However, this is 

a common problem for chemical analysis and can only be circumvented 

if there are recovery standards for every single 

compound to be analysed. Such an approach is not possible for 

bioassays because spiked chemicals would add to the mixture 

toxicity. A shortcoming of the SPE method that was noticed here 

for the first time is the co-extraction of an unknown fraction of 

dissolved organic matter, an issue that requires further attention 

in the future. 

In addition, volatile compounds will evaporate during SPE 

but they would also cross contaminate in a 96 well plate, so it 

is a procedural advantage to have removed the volatile 

chemicals prior to toxicity testing. It must also be noted that 

oxidation by-products formed during ozonation are often 

more hydrophilic than their parent compound and might 

therefore not be extracted efficiently by SPE. 

In addition, one must be aware that mixture effects are 

assessed when applying bioanalytical tools to complex samples 

and there is always the problem that the subtle effect of newly 

formed by-products might be shielded by other micropollutants 

present in the mixture. Some of the oxidation byproducts 

stemming from ozonation are very reactive, for example aldehydes 

and other electrophiles formed from ß-blockers during 

ozonation (Benner and Ternes, 2009), others are very hydrophilic 

organics such as NDMA (N-nitrosodimethylamine) or 

inorganics such as bromate (Hollender et al., in press) noneof 

these would be caught with the bioanalytical tools used here but 

require targeted chemical analysis for quantification. However, 

ozonation often only mildly oxidizes organic micropollutants 

and many resulting transformation products are likely to retain 

some of their toxic potential (Escher et al., 2008c) or lose their 

specific receptor-binding effect but retain non-specific baseline 

toxicity (Lee et al., 2008) and will thus contribute to the mixture 

toxicity in the bioanalytical tools applied in the present study. 

While the results are discussed in more detail and 

compared to chemical analysis in the accompanying paper 

(Reungoat et al., 2010), some general conclusions can be 

drawn: 

The influent of the water reclamation plant had similar 

levels of effects for the different endpoints as we have 

encountered in previous studies for wastewater after 

secondary treatment (Escher et al., 2008a; Escher et al., in 

press). 

The effluent of the water reclamation plant had strongly 

reduced effect levels compared to the influent, reducing the 

baseline-TEQ by 79%, the DEQ by 76%, the estrogenicity to 

below the detection limit of 0.02 ng/L EEQ, the PTEQ by 88% 

and the genotoxicity to below the detection limit. Only the 

AhR-CAFLUX assay did not yield any information on the 

removal efficiency of the treatment because the level of 

persistent inducers of the arylhydrocarbon receptor 

remained close to the detection limit. This was not too 

unexpected given that these are typically very hydrophobic 

compounds, which are usually adsorbed and not present in 


the water column. In contrast, the non-persistent inducers 

of the AhR were clearly reduced by a factor of two during 

ozonation. 

The most efficient steps with respect to removal of the toxic 

responses in all selected assays proved to be the coagulation/flocculation/DAFF, 

the main ozonation and (biological) 

activated carbon filtration steps. 

This paper demonstrates the first comprehensive application 

of the bioanalytical tool framework on an advanced water 

treatment plant. One future goal would be to automate the 

assays to a degree that they can be used routinely by 

commercial laboratories or water companies themselves. At 

this stage, there remain some scientific questions to be 

resolved, including (1) the sample treatment and dosing, (2) 

extension of the battery and (3) data interpretation. (1) SPE and 

dosing by solvent spiking will always change the composition 

of an undefined mixture. A combination of passive sampling to 

a biomimetic polymer and passive dosing directly from the 

polymer into the bioassay might be a way to overcome this 

limitation. (2) A battery of bioassays will never be comprehensive 

because of the myriads of receptors and regulatory 

pathways in an organism. Nevertheless the battery would 

benefit from inclusion of additional endpoints, such as chemically 

induced immunosuppression or even other examples of 

mode-of-action categories already included, e.g. other endocrine 

endpoints. (3) At this stage we can say how much of 

a toxic equivalent concentrations is removed by a certain 

treatment step and we can give comparative results but we will 

never be able to deduce ecological or health consequences 

from that. This discussion must be lead independently. One 

solution would be to compare the TEQ of the mixture with the 

defined Environmental Water Quality Criteria (EQS) for single 

substances of the Water Framework Directive (European 

Commission, 2006) or guideline values for drinking water (e.g. 

NHMRC–NRMMC, 2004) but not for all reference compounds 

are such EQS and guideline values available. 


This work was co-funded by the Urban Water Security Research 

Alliance under the Enhanced Treatment Project, the CRC Water 

Quality and Treatment Project No. 2.0.2.4.1.1. Dissolved Organic 

Carbon Removal by Biological Treatment and by EnTox. The 

National Research Centre for Environmental Toxicology (EnTox) 

is a joint venture of the University of Queensland and Queensland 

Health Forensic and Scientific Services (QHFSS). 

The authors acknowledge the Moreton Bay Water for access to 

the South Caboolture Water Reclamation Plant; Ray McSweeny 

and Paul McDonnell (Moreton Bay Water) for their help during 

sampling and Chris Pipe-Martin (Ecowise) for providing information 

on the South Caboolture Water Reclamation Plant. 

The authors are grateful to Renee Muller (Gold Coast 

Water), Fred Leusch (Griffith University/EnTox), Karen 

Kennedy (EnTox) and Pam Quayle (EnTox) for constructive 

discussions and for help with the data analysis method, Marita 

Goodwin (EnTox) for running various assays and for helpful 

comments Christina Carswell (QHFSS) for extracting the

490 

samples and Chris Paxman (EnTox) for technical assistance. 

We would like to thank also Michael Denison (University of 

California Davis, USA) for providing the H4G1.1c2 cells, Georg 

Reifferscheid (German Federal Institute of Hydrology, 

Germany) for providing the bacteria Salmonella typhimurium 

TA1535/pSK1002, and Ana Soto (Tufts University, USA) for 

providing the MCF7-BOS cells. 

Appendix. Supporting information 

Supporting information related to this article can be found at 

doi:10.1016/j.watres.2009.09.025. 

references 

Bengtson Nash, S.M., Schreiber, U., Ralph, P.J., Mueller, J.F., 2005a. 

The combined SPE: ToxY-PAM phytotoxicity assay; 

application and appraisal of a novel biomonitoring tool for the 

aquatic environment. Biosensors and Bioelectronics 20 (7), 

1443–1451. 

Bengtson Nash, S.M., McMahon, K., Eaglesham, G., Mueller, J.F., 

2005b. Application of a novel phytotoxicity assay for the 

detection of herbicides in Hervey Bay and the Great Sandy 

Straits. Marine Pollution Bulletin 51, 351–360. 

Bengtson Nash, S.M., Goddard, J., Muller, J.F., 2006. Phytotoxicity 

of surface waters of the Thames and Brisbane river estuaries: 

a combined chemical analysis and bioassay approach for the 

comparison of two systems. Biosensors and Bioelectronics 21, 

2086–2093. 

Benner, J., Ternes, T.A., 2009. Ozonation of propranolol: formation 

of oxidation products. Environmental Science & Technology 

43, 5086–5093. 

Cao, N., Yang, M., Zhang, Y., Hu, J., Ike, M., Hirotsuji, J., Matsui, H., 

Inoue, D., Sei, K., 2009. Evaluation of wastewater reclamation 

technologies based on in vitro and in vivo bioassays. 

Ecotoxicology and Environmental Safety 407 (5), 1588–1597. 

Chapman, P.M., 2000. Whole effluent toxicity testing – usefulness, 

level of protection, and risk assessment. Environmental 

Toxicology and Chemistry 19 (1), 3–13. 

Denison, M.S., Nagy, S.R., 2003. Activation of the aryl hydrocarbon 

receptor by structurally diverse exogenous and endogenous 

chemicals. Annual Reviews in Pharmacology and Toxicology 

43, 309–334. 

Deutsche Norm, 1995. In: DIN, N.W.N.i. (Ed.), Suborganismische 

Verfahren: Teil 1: Bestimmung von Cholinesterase- 

Hemmenden Organophosphat- und Carbamat-Pestiziden 

(Cholinesterase-Hemmtest) DIN 38415-1: 1995-02, Berlin, 

Germany. 

Ellman, G.L., Courtney, K.D., Andres, V., Featherstone, R.M., 1961. A 

new and rapid colorimetric determination of acetylcholinesterase 

activity. Biochemical Pharmacology 7, 88–95. 

Escher, B., Schwarzenbach, R.P., 2002. Mechanistic studies on 

baseline toxicity and uncoupling as a basis for modeling 

internal lethal concentrations in aquatic organisms. Aquatic 

Sciences 64, 20–35. 

Escher, B.I., Bramaz, N., Maurer, M., Richter, M., Sutter, D., von 

Känel, C., Zschokke, M., 2005. Screening test battery for 

pharmaceuticals in urine and wastewater. Environmental 

Toxicology and Chemistry 24 (3), 750–758. 

Escher, B.I., Quayle, P., Muller, R., Schreiber, U., Mueller, J., 2006. 

Passive sampling of herbicides combined with effect analysis 

in algae using a novel high-throughput phytotoxicity assay 


(Maxi-Imaging- PAM). Journal Environmental Monitoring 8, 

456–464. 

Escher, B.I., Bramaz, N., Quayle, P., Rutishauser, S., Vermeirssen, E., 

2008a. Monitoring of the ecotoxicological hazard potential by 

polar organic micropollutants in sewage treatment plants and 

surface waters using a mode-of-action based test battery. 

Journal Environmental Monitoring 10, 622–631. 

Escher, B.I., Bramaz, N., Mueller, J.F., Quayle, P., Rutishauser, S., 

Vermeirssen, E., 2008b. Toxic equivalent concentrations 

(TEQs) for baseline toxicity and specific modes of action as 

a tool to improve evaluation of ecotoxicity tests on 

environmental samples. Journal Environmental Monitoring 

10, 612–621. 

Escher, B.I., Baumgartner, R., Lienert, J., Fenner, K., 2008c. 

Predicting the ecotoxicological effects of transformation 

products. SpringerLink online. In: Boxall, A. (Ed.), Handbook of 

Environmental Chemistry – Degradation of Synthetic 

Chemicals in the Environment. Springer, Heidelberg, 

Germany. doi:10.1007/6998_2_015. http://www.springerlink. 

com/content/48r2t364q183q654/?p ¼ ec460152bfd 

14906adc89378463e4997&pi ¼ 5 published Wednesday, 

April 30, 2008. 

Escher, B.I., Bramaz, N. and Ort, C., 2009. Monitoring the 

treatment efficiency of a full scale ozonation on a 

sewage treatment plant with a mode-of-action based test 

battery. Journal Environmental Monitoring, doi: 10.1039/ 

b907093a. 

European Commission, 2006. Proposal for a Directive of the 

European Parliament and the Council on Environmental 

Quality Standards in the Field of Water Policy and Amending 

Directive 2000/60/EC COM(2006) 398 Final. COM(2006) 398 

Final. European Parliament, Committee on the Environment, 

Public Health and Food Safety, Brussels. 

Farré, M., Martinez, E., Hernando, M., Fernandez-Alba, A., Fritz, J.I., 

Unruh, E., Mihail, O., Sakkas, V., Morbey, A., Albanis, T.A., 

Brito, F., Hansen, P.D., Barcelo, D., 2006. European ring 

exercise on water toxicity using different bioluminescence 

inhibition tests based on Vibrio fischeri, in support to the 

implementation of the water framework directive. Talanta, 

323–333. 

GWRC, 2008. In: Leusch, F.D.L. (Ed.), Tools to Detect Estrogenic 

Activity in Environmental Waters. Global Water Research 

Coalition, London, UK, p. 84. 

Hamers, T., Molin, K.R.J., Koeman, J.H., Murk, A.J., 2000. A smallvolume 

bioassay for quantification of the esterase inhibiting 

potency of mixtures of organophosphates and carbamate 

insceticides in rainwater: development and optimization. 

Toxicological Sciences 58, 60–67. 

Hollender, J., Zimmermann, S.G., Koepke, S., Krauss, M., 

McArdell, C.S., Ort, C., Singer, H., von Gunten, U., Siegrist, H., 

2009. Elimination of organic micropollutants in a municipal 

wastewater treatment plant upgraded with a full-scale postozonation 

followed by sand filtration. Environmental Science 

& Technology, Articles ASAP (Article), in press. doi: 10.1021/ 

es9014629. 

Huber, M.M., Ternes, T.A., von Gunten, U., 2004. Removal of 

estrogenic activity and formation of oxidation products during 

ozonation of 17 alpha-ethinylestradiol. Environmental 

Science & Technology 38, 5177–5186. 

Huber, M.M., Göbel, A., Joss, A., Hermann, N., Löffler, D., 





International Standard Organisation 1998 Water Quality– 

determination of the inhibitory effect of water samples on the 

light emission of Vibrio fischeri (luminescent bacteria test), EN 

ISO 11348-3, Geneva, Switzerland.

International Standard Organisation 2000 Water quality – 

Determination of the genotoxicity of water and waste water 

using the umu-test, EN ISO 13829(2000) and 38415-3 (1996), 

Geneva, Switzerland. 

Johnson, B.T., 2005. Microtox acute toxicity test. In: Blaise, C., 

Ferard, J.-F. (Eds.), Small-Scale Freshwater Toxicity 

Investigations. Toxicity Test Methods, vol. 1. Springer. 

Joss, A., Keller, E., Alder, A.C., Gobel, A., McArdell, C.S., Ternes, T., 

Siegrist, H., 2005. Removal of pharmaceuticals and fragrances 

in biological wastewater treatment. Water Research 39 (14), 

3139–3152. 

Körner, W., Hanf, V., Schuller, W., Kempter, C., Metzger, J., 

Hagenmaier, H., 1999. Development of a sensitive E-screen 

assay for quantitative analysis of estrogenic activity in 

municipal sewage plant effluents. Ecotoxicology and 

Environmental Safety 225 (1–2), 33–48. 

Körner, W., Bolz, U., Sussmuth, W., Hiller, G., Schuller, W., 

Hanf, V., Hagenmaier, H., 2000. Input/output balance of 

estrogenic active compounds in a major municipal sewage 

plant in Germany. Chemosphere 40 (9–11), 1131–1142. 

Körner, W., Spengler, P., Bolz, U., Schuller, W., Hanf, V., 

Metzger, J.W., 2001. Substances with estrogenic activity in 

effluents of sewage treatment plants in southwestern 

Germany. 2. Biological analysis. Environmental Toxicology 

and Chemistry 20 (10), 2142–2151. 

Lee, Y., Escher, B.I., von Gunten, U., 2008. Efficient removal of 

estrogenic activity during oxidative treatment of waters 

containing steroid estrogens. Environmental Science & 

Technology 42 (17), 6333–6339. 

Legler, J., Zeinstra, L.M., Schuitemaker, F., Lanser, P.H., Bogerd, J., 

Brouwer, A., Vethaak, A.D., De Voogt, P., Murk, A.J., Van der 

Burg, B., 2002. Comparison of in vivo and in vitro reporter gene 

assays for short-term screening of estrogenic activity. 


Leusch, F.D., Chapman, H.F., van den Heuvel, M.R., Tan, B.L., 

Gooneratne, S.R., Tremblay, L.A., 2006. Bioassay-derived 

androgenic and estrogenic activity in municipal sewage in 

Australia and New Zealand. Ecotoxicology and Environmental 

Safety 65 (3), 403–411. 

Muller, R., Tang, J., Thier, R., Mueller, J., 2007. Combining passive 

sampling and toxicity testing for evaluation of mixtures of 

polar organic chemicals in sewage treatment plant effluents. 

Journal Environmental Monitoring 9, 104–109. 

Muller, R. 2008 Passive sampling and bioanalysis of organic 

chemicals in surface water. Ph.D. thesis. University of 

Queensland, St. Lucia, Qld, Australia, 201pp. Accessible at 

http://library.uq.edu.au/search~S7?/XRenee þ 

Muller&searchscope ¼ 7&SORT ¼ AZ/XRenee þ 

Muller&searchscope ¼ 7&SORT ¼ AZ&extended ¼ 0&SUBKEY 

¼ Renee%20Muller/21%22C23%22C23%22CB/frameset&FF ¼ 

XRenee þ Muller&searchscope ¼ 27&SORT ¼ 

AZ&21%22C21%22C 

Muller, R., Schreiber, U., Escher, B.I., Quayle, P., Bengtson Nash, S., 

Mueller, J., 2008. Rapid exposure assessment of PSII herbicides 

in surface water using a novel chlorophyll a fluorescence 

imaging assay. Science Total Environment 401, 51–59. 

Murk, A.J., Legler, J., van Lipzig, M.M.H., Meerman, J.H.N., 

Belfroid, A.C., Spenkelink, A., van der Burg, B., Rijs, G.B.J., 

Vethaak, D., 2002. Detection of estrogenic potency in 

wastewater and surface water with three in vitro bioassays. 

Environmental Toxicology and Chemistry 21 (1), 16–23. 

Nagy, S., Sanborn, J., Hammock, B., Denison, M., 2002. 

Development of a green fluorescent protein-based cell 

bioassay for the rapid and inexpensive detection and 

characterization of Ah receptor agonists. Toxicological 

Sciences 65, 200–210. 

NHMRC–NRMMC, 2004. Australian Drinking Water Guidelines 

(ADWG). National Health and Medical Research Council 


(NHMRC) in Collaboration With the Natural Resource 

Management Ministerial Council (NRMMC), Commonwealth 

of Australia. www.nhmrc.gov.au/publications/synopses/_ 

files/adwg_11_06.pdf. 

Oda, J., Nakamura, S.I., Oki, I., Kato, T., Shinagawa, H., 1985. 

Evaluation of the new system (umu-test) for the detection of 

environmental mutagens and carcinogens. Mutatation 

Research 147, 219–229. 


Salkellaropoulos, G., 2006. Ecotoxicological properties of 

wastewater treated using tertiary methods. Environmental 

Toxicology 21 (4), 417–424. 


Sakellaropoulos, G.P., 2008. Influence of ozonation on the in 

vitro mutagenic and toxic potential of secondary effluents. 

Water Research 42 (20), 4929–4940. 

Reifferscheid, G., Heil, J., Oda, Y., Zahn, R.K., 1991. A microplate 

version of the SOS/umu-test for rapid detection of genotoxins 

and genotoxic potentials of environmental samples. Mutation 

Research 253 (3), 215–222. 

Reungoat, J., Macova, M., Escher, B.I., Carswell, S., Mueller, J.F., 

Keller, J., 2010. Removal of micropollutants and reduction of 

biological adverse effects in a full scale reclamation plant 

using ozonation and activated carbon filtration. Water 

Research 44 (2), 625–637. 

Rutishauser, B.V., Pesonen, M., Escher, B.I., Ackermann, G.E., 

Aerni, H.-R., Suter, M.J.-F., Eggen, R.I.L., 2004. Comparative 

analysis of estrogenic activity in sewage treatment plant 

effluents involving three in vitro assays and chemical analysis of 

steroids. Environmental Toxicology and Chemistry 23, 857–864. 

Sanchez-Polo, M., Salhi, E., Rivera-Utrilla, J., von Gunten, U., 2006. 

Combination of ozone with activated carbon as an alternative 

to conventional advanced oxidation processes. Ozone-Science 

& Engineering 28 (4), 237–245. 

Schreiber, U., Mueller, J.F., Haugg, A., Gademann, R., 2002. New 

type of dual-channel PAM chlorophyll fluorometer for highly 

sensitive water toxicity biotests. Photosynthesis Research 74, 

317–330. 

Schreiber, U., Quayle, P., Schmidt, S., Escher, B.I., Mueller, J., 2007. 

Methodology and evaluation of a highly sensitive algae 

toxicity test based on multiwell chlorophyll fluorescence 

imaging. Biosensors and Bioelectronics 22, 2554–2563. 


Johnson, C.A., von Gunten, U., Wehrli, B., 2006. The challenge 


1072–1077. 

Snyder, S.A., Adham, S., Redding, A.M., Cannon, F.S., DeCarolis, J., 

Oppenheimer, J., Wert, E.C., Yoon, Y., 2007. Role of 

membranes and activated carbon in the removal of endocrine 

disruptors and pharmaceuticals. Desalination 202, 156–181. 

Soto, A.M., Sonnenschein, C., Chung, K.L., Fernandez, M.F., Olea, N. 

, Serrano, F.O., 1995. The E-SCREEN assay as a tool to identify 

estrogens: an update on estrogenic environmental pollutants. 

Environmental Health Perspectives 103 (Suppl. 7), 113–122. 

Tan, B.L.L., Hawker, D.W., Muller, J.F., Leusch, F.D.L., Tremblay, L.A., 

Chapman, H.F., 2007. Comprehensive study of endocrine 

disrupting compounds using grab and passive sampling at 

selected wastewater treatment plants in South East Queensland, 

Australia. Environment International 33 (5), 654–669. 

Ternes, T.A., Joss, A., Siegrist, H., 2004. Scrutinizing pharmaceuticals 

and personal care products in wastewater treatment. 

Environmental Science & Technology 38 (17), 393A–399A. 

van Leeuwen, J., Pipe-Martin, C., Lehmann, R.M., 2003. Water 

reclamation at South Caboolture, Queensland, Australia. 

Ozone: Science & Engineering 25, 107–120. 

Villeneuve, D.L., Blankenship, A.L., Giesy, J.P., 2000. Derivation and 

application of relative potency estimates based on in-vitro 

bioassays. Environmental Toxicology Chemistry 19, 2835–2843.

492 

von Gunten, U., 2003. Ozonation of drinking water: part I. 

Oxidation kinetics and product formation. Water Research 37 

(7), 1443–1467. 

Warne, M.S.J., Hawker, D.W., 1995. The number of components 

in a mixture determines whether synergistic and 

antagonistic or additive toxicity predominate – the funnel 


hypothesis. Ecotoxicology and Environmental Safety 31 (1), 

23–28. 

Zhao, B., Denison, M., 2004. Dvelopment and characterization of 

a green fluorescence protein-based rat cell bioassay system 

for detection of Ah receptor ligands. Organohalogen 

Compounds 66, 3332–3337.

An evaluation of a pilot-scale nonthermal plasma advanced 

oxidation process for trace organic compound degradation 

Daniel Gerrity*, Benjamin D. Stanford, Rebecca A. Trenholm, Shane A. Snyder 

Applied Research and Development Center, Southern Nevada Water Authority, River Mountain Water Treatment Facility, P.O. Box 99954, 

Las Vegas, NV 89193-9954, USA 

article info 




3 September 2009 



Keywords: 

Advanced oxidation process (AOP) 

Nonthermal plasma (NTP) 

Trace organic compound 


Endocrine disrupting compound 

(EDC) 


abstract 

Wastewater-derived contaminants, including pharmaceuticals 

and personal care products (PPCPs), endocrine disrupting 

compounds (EDCs), and other trace organic compounds, have 

been a significant aspect of environmental research efforts for 

decades (Snyder et al., 2003). However, adverse impacts on 

aquatic ecosystems and increased public awareness of trace 

organic compounds in water supplies have stimulated recent 

interest in this issue (Snyder et al., 2003). Due to rapid population 

growth and increasingly stressed water supplies, 

particularly in semi-arid regions, many cities are turning to 

alternative sources of drinking water such as indirect potable 

reuse. In some systems, wastewater effluent is having 

a greater impact on reservoir water quality due to drought and 

overdraft (Benotti et al., 2009a). Over time, these trends may 




* Corresponding author. Tel.: þ1 (702) 856 3666; fax: þ1 (702) 856 3647. 

E-mail address: Daniel.Gerrity@snwa.com (D. Gerrity). 


doi:10.1016/j.watres.2009.09.029 

This study evaluated a pilot-scale nonthermal plasma (NTP) advanced oxidation process 

(AOP) for the degradation of trace organic compounds such as pharmaceuticals and 

potential endocrine disrupting compounds (EDCs). The degradation of seven indicator 

compounds was monitored in tertiary-treated wastewater and spiked surface water to 

evaluate the effects of differing water qualities on process efficiency. The tests were also 

conducted in batch and single-pass modes to examine contaminant degradation rates and 

the remediation capabilities of the technology, respectively. Values for electrical energy per 

order (EEO) of magnitude degradation ranged from

494 

ozone, granular activated carbon (GAC), nanofiltration/reverse 

osmosis, and advanced oxidation) can achieve significant 

removal or transformation (Ternes et al., 2002; Westerhoff 

et al., 2005; Snyder et al., 2007). Advanced oxidation processes 

(AOPs) are often used to achieve further reductions in trace 

organics, target recalcitrant compounds, and comply with 

certain water reuse regulations (e.g., California Department of 

Public Health Title 22 requirements for NDMA and 1,4dioxane). 

The most common AOPs consist of UV/peroxide and 

ozone/peroxide, but emerging technologies such as 

nonthermal plasma (NTP) provide viable alternatives. 

This study focuses on a novel pilot-scale NTP reactor 

developed by Aquapure Technologies, Ltd. (Upper Galilee, 

Israel). The NTP uses high-voltage electrical pulses across 

fiber-like electrodes to form a corona discharge, which in turn 

generates UV light, ozone, and hydroxyl radicals (Locke et al., 

2006). The light generated by the corona discharge spans 

wavelengths of approximately 250 nm to 1,000 nm with 

multiple peaks at wavelengths associated with radical 

formation (Locke et al., 2006). Although the light is lowintensity 

(Locke et al., 2006), it has been credited with 

a capacity for NDMA destruction (Even-Ezra et al., 2009), 

which is generally inefficient in ozone-based oxidation 

processes (Mitch et al., 2003). 

NTP is considered to be highly efficient because little energy 

is lost in heating the surrounding fluid, which allows the 

energy to be focused on the excitation of electrons (Pekarek, 

2003). In contrast to the electron beam technology, which 

introduces electrons from an external source, the NTP used in 

this study employs a corona discharge that excites electrons in 

the ambient air directly above the target water matrix 

(Pekarek, 2003). Other NTP configurations are also available, 

including those that create a corona discharge in aerosols or 

within the target water matrix (Pekarek, 2003; Locke et al., 

2006). The level of treatment can be adjusted by changing the 

frequency (e.g., from 500 to 1,000 Hz) or voltage (up to 40 kV) of 

the electrical pulses. The ionizing capability of the electrical 

pulses generates singlet oxygen atoms, which in turn form 

ozone and hydroxyl radicals. One of the most significant 

benefits of NTP is the fact that oxidants can be generated 

without the addition of costly chemicals or UV lamps, which 

require cleaning and are hindered by high turbidity and matrix 

absorbance. However, the NTP technology has limited 

commercial availability and has not been tested in full-scale 

drinking water and wastewater treatment applications so its 

long-term reliability, efficiency, and effectiveness are still 

unknown. In addition, the NTP technology is currently limited 

to low-flow applications due to its thin-film configuration. 

Although plasma-based technologies have been studied 

extensively for the degradation of volatile organic compounds 

(VOCs) (Masuda et al., 1995; Hwang and Jo, 2005; Ighigeanu 

et al., 2005; Thevenet et al., 2007; Kim et al., 2008), many of 

these studies have focused on air quality and/or bench-scale 

configurations. Due to limitations associated with conventional 

AOPsdprimarily the need for chemical additiond 

studies of novel AOPs for water and wastewater treatment 

applications are gaining popularity. A recent pilot-scale study 

utilizing the Aquapure technology evaluated the efficacy of 

NTP in degrading trichloroethylene (TCE), N-nitrosodimethylamine 

(NDMA), 1,4-dioxane, and methyl tert-butyl 


ether (MTBE) in batch experiments and single-pass remediation 

(Even-Ezra et al., 2009). The current study uses the same 

pilot-scale system to evaluate the degradation of PPCPs and 

EDCs in water and wastewater. Despite the lack of PPCP/EDC 

data for plasma-based technologies, other AOPs, particularly 

UV/peroxide (Huber et al., 2003; Pereira et al., 2007; Snyder 

et al., 2007; Benotti et al., 2009b), ozone/peroxide (Acero and 

von Gunten, 2001; Snyder et al., 2007), and photocatalysis 

(Benotti et al., 2009b; Westerhoff et al., 2009), have been tested 

extensively with respect to trace organic degradation. In their 

review of NTP applications, Locke et al. (2006) discussed the 

need for comparisons of NTP with more common AOPs to 

determine whether there is a net benefit associated with 

employing NTP technologies. Among other research needs, 

Locke et al. (2006) specifically identified energy consumption 

and applicability to different waste matrices as useful points 

of comparison. To this end, the objective of this study is to 

expand the NTP knowledge base by evaluating its efficacy and 

energy efficiency in degrading seven PPCPs and EDCs in 

tertiary-treated wastewater and spiked surface water. 

2. Experimental section 

2.1. Selection of wastewater-derived contaminants 

In order to narrow the scope of this research, a subset of the 

numerous compounds detected in previous occurrence 

studies was selected for evaluation. The indicator compounds 

were selected based on their magnitude and frequency of 

occurrence in water and wastewater (Snyder et al., 2007), 

varying physical/chemical characteristics and resulting 

susceptibility to treatment (Ternes et al., 2002; Westerhoff 

et al., 2005; Snyder et al., 2007), and ease of analytical methods 

(Trenholm et al., 2009). The seven compounds targeted in this 

study include meprobamate (anti-anxiety), dilantin (also 

known as phenytoin; anticonvulsant), primidone (anticonvulsant), 

carbamazepine (anticonvulsant), atenolol (betablocker), 

trimethoprim (antibiotic), and atrazine (herbicide 

and suspected EDC). Their structures and characteristics are 

described in Snyder et al. (2007). Although these compounds 

have generated considerable interest in the research, treatment, 

and regulatory arenas, only atrazine is currently regulated 

by the United States (U.S.) Environmental Protection 

Agency (EPA) at a maximum contaminant level (MCL) of 3 mg/L. 

2.2. Nonthermal plasma pilot unit 

The pilot-scale unit depicted in Fig. 1 is an ‘‘electrode-to-plate’’ 

NTP prototype (Locke et al., 2006) developed by Aquapure 

Technologies, Ltd.; the pilot skid contains two reactors connected 

in series. Water flows in a thin film (z5 mm) along the 

stainless steel ground electrode (anode) where it is exposed to 

high-voltage electrical pulses. The electrical pulses from the 

generator have frequencies ranging from 500 to 1,000 Hz, 

maximum voltage of 8.0 kV, maximum current of 100 A, 

maximum energy of 1 J, and rise time of approximately 18 ns 

(Locke et al., 2006; Even-Ezra et al., 2009). Depending on the 

applied voltage and frequency, each reactor draws between 0.4 

and 1.0 kW, and the external pumping and controls draw

approximately 2.8 kW. The electrical pulses generate a corona 

discharge that stretches approximately 1 cm from the carbon 

fiber-like cathode to the water surface. The distance between 

the carbon electrode and water surface can be adjusted at 

multiple points along the reactor to level the system and 

optimize the process for a particular water matrix. The 1-cm 

distance was selected to achieve a consistent corona discharge 

and minimize sparking within the reactor, which would be 

destructive to the carbon fibers. Despite the harsh conditions 

experienced by the gas-phase electrode, no visible wear was 

detected on the carbon fibers over the duration of the experiments. 

With continuous, long-term operation of the NTP 

reactor, it is likely that the carbon fibers would eventually 

require replacement, but this life span is currently unknown. 

After traveling through the first reactor, the water is 

collected in a storage tank before being pumped into the 

ozone injector system or the second reactor. The ozone 

injector system combines a portion of the treated water 

from the first reactor with ozone-rich air (z2 g/m 3 ) from 

the reactor headspace using a Venturi inductor. For the 

single-pass configuration, this mixture is collected in 

another storage tank for additional ozone contact time 

before being mixed and discharged with effluent from the 

first reactor. Excess air from the ozone contactor is then 

drawn out of the system by vacuum and quenched in an 

ozone destruct unit. The hydraulic residence times in the 

reactor, storage tank, and ozone contactor are approximately 

0.5 min, 4–7.5 min, and 1.5 min, respectively. The 

hydraulic residence times in the reactor and storage tank 

depend on the process flow rate, whereas the residence 

time in the ozone contactor remains relatively constant 

because the injector flow rate (z40 L/min) is controlled by 


Fig. 1 – Schematic of NTP pilot unit. The bottom figure illustrates a profile view of the electrodes within the reactor. 

a separate pump. The process is then repeated in the 

second reactor before being pumped out of the pilot skid. 

For the batch configuration, the effluent pump is replaced 

by a recirculation pump that allows the water to cycle 

continuously through a single reactor and ozone contactor. 

Additional details related to the NTP reactor used in this 

study are provided in Even-Ezra et al. (2009). 

2.3. Experimental design 

The study was divided into three experiments: (1) a batch 

experiment with 150 L of tertiary-treated wastewater (i.e., 

grit removal, primary settling, activated sludge, secondary 

settling, and dual media filtration) at a recirculation rate of 

8.0 L/min, (2) a single-pass experiment with tertiary-treated 

wastewater at a flow rate of 15.5 L/min, and (3) a single-pass 

experiment with spiked surface water at a flow rate of 

11.4 L/min. For experiment (3), spiked surface water (Colorado 

River water from Lake Mead, NV) was pumped into the 

reactor from a 3,000-gallon polyethylene water tank 

(American Tank Company, Windsor, CA). The spiking solution 

was prepared by dissolving neat standards in water in 

order to avoid the introduction of radical scavenging 

solvents such as methanol. The spiking solution was added 

to the 3,000-gallon tank followed by recirculation for 24 h to 

facilitate homogenization. The two water types (wastewater 

and surface water) were selected to provide an assessment 

of NTP treatment efficiency for matrices with different 

organic loadings. The initial pH, total alkalinity, total organic 

carbon (TOC), UV254 absorbance, and trace organic 

compound concentrations for the three experiments are 

provided in Table 1. In all tables, ‘‘Batch’’ refers to the batch

496 

Table 1 – Initial water quality conditions. 

Parameter Batch (1) WW (2) a,c 

SSW (3) b,c 

pH 7.0 7.0 8.0 

Total Alkalinity 126 mg/L as 126 mg/L as 137 mg/L as 

CaCO3 CaCO3 CaCO3 TOC 5.9 mg-C/L 5.6 mg-C/L 2.6 mg-C/L 

UV254 Abs. 0.119 cm 1 

N/A 0.036 cm 1 

Meprobamate 275 ng/L 256 25 e ng/L 933 15 ng/L 

Dilantin 119 ng/L 165 10 ng/L 1,005 74 ng/L 

Primidone 124 ng/L 146 6 ng/L 1,375 96 ng/L 

Carbamazepine 176 ng/L 219 9 ng/L 1,600 115 ng/L 

Atenolol 378 ng/L 413 8 ng/L 1,018 57 ng/L 

Trimethoprim 36 ng/L 39 2 ng/L 1,200 82 ng/L 

Atrazine d 

N/A N/A 1,118 159 ng/L 

a WW (2) ¼ Single-pass wastewater experiment. 

b SSW (3) ¼ Single-pass spiked surface water experiment. 

c Four influent samples were taken throughout the experiment to 

calculate a representative ambient concentration and evaluate 

temporal variability in target compound concentrations. 

d Atrazine was spiked in the SSW experiment but was not detected 

in the wastewater. 

e Errors represent 1 standard deviation. 

wastewater experiment, ‘‘WW’’ refers to the single-pass 

wastewater experiment, and ‘‘SSW’’ refers to the single-pass 

spiked surface water experiment. 

The generator settings for the NTP pilot differed for each of 

the experiments, as described in Table 2. For the batch 

experiment (1), a single reactor/generator was operated at 

a frequency of 500 Hz and a voltage of 8.0 kV. To assess the 

degradation rate, ten samples were collected at generator 

(AOP-specific) energy consumption values ranging from 0 to 

7.3 kWh/m 3 . For the single-pass wastewater experiment (2), 

four different generator scenarios provided treatment levels 

ranging from 0.7 to 1.8 kWh/m 3 for the generators and 3.8 to 

4.8 kWh/m 3 for the overall pilot skid. Samples were collected 

from the effluent line exiting the pilot skid. Residual oxidants 

in each sample were quenched with calcium thiosulfate. 

During the wastewater experiments, the dissolved ozone 

concentration in the NTP reactor effluent ranged from 

approximately 0.05 mg/L to 0.10 mg/L depending on the 

generator settings. The single-pass surface water experiment 

(3) incorporated nine different generator scenarios ranging 

from 0.6 to 2.6 kWh/m 3 for the generators and 4.8 to 6.8 kWh/ 

m 3 for the pilot skid. For the surface water, samples were 

taken from Reactor 1, Ozone Tank 1, Reactor 2, and the pilot 

skid effluent to evaluate the extent of contaminant degradation 

in each system component. During the surface water 

experiment, the dissolved ozone concentrations in the reactor 

effluent ranged from 0.25 to 0.50 mg/L depending on the 

generator settings. 


Ozone gas concentrations in the reactor headspace were 

determined by online measurements with an OMC20 ozone 

monitor and IN-2000-L2-LC ozone analyzer (IN USA, Norwood, 

MA). Dissolved ozone residuals were measured with the 

indigo trisulfonate method (Yates and Stenstrom, 2000). TOC 

and UV 254 absorbance were analyzed according to standard 

methods 5310B and 5910, respectively. Excitation-emission 

matrices (EEM) were also developed for the single-pass 

wastewater samples using a QuantaMaster UV-Vis QM4 

Steady State Spectrofluorometer (Photon Technology International, 

Inc., Birmingham, NJ). 

Trace organic compounds were extracted and analyzed 

using on-line solid phase extraction and liquid chromatography 

with tandem mass spectrometry (SPE-LC-MS/MS) 

according to Trenholm et al. (2009). All standards and reagents 

were of the highest purity commercially available. All 

Table 2 – Experimental test conditions. 

Experiment Generator 1 Generator 2 Total generator 

energy 

(1) Batch a 

Frequency 

(Hz) 

Voltage 

(kV) 


Energy 

(kWh/m 3 ) 

Frequency 

(Hz) 

Voltage 

(kV) 

Energy 

(kWh/m 3 ) 

Total skid 

energy 

(kWh/m 3 ) (kWh/m 3 ) 

Off Off 0.00 500 8.0 N/A N/A N/A 

(2) WW a – 1 Off Off 0.00 500 8.0 0.71 0.71 3.75 

WW – 2 800 8.0 1.04 Off Off 0.00 1.04 4.08 

WW – 3 600 8.0 0.87 500 8.0 0.71 1.58 4.62 

WW – 4 800 8.0 1.04 500 8.0 0.71 1.75 4.79 

(3) SSW b – 1 Off Off 0.00 600 5.9 0.63 0.63 4.78 

SSW – 2 Off Off 0.00 800 5.9 0.85 0.85 5.01 

SSW – 3 500 8.0 0.97 Off Off 0.00 0.97 5.12 

SSW – 4 Off Off 0.00 1000 5.9 1.07 1.07 5.23 

SSW – 5 1000 8.0 1.53 Off Off 0.00 1.53 5.68 

SSW – 6 500 8.0 0.97 600 5.9 0.63 1.60 5.75 

SSW – 7 500 8.0 0.97 1000 5.9 1.07 2.04 6.19 

SSW – 8 1000 8.0 1.53 600 5.9 0.63 2.16 6.31 

SSW – 9 1000 8.0 1.53 1000 5.9 1.07 2.60 6.75 

a The batch and WWTP experiments were conducted with ambient concentrations of trace organics in tertiary-treated wastewater. 

b The SSW experiments were conducted with raw surface water (Lake Mead) spiked with a suite of trace organics.

pharmaceuticals were obtained from Sigma-Aldrich (St. Louis, 

MO, USA). Meprobamate-d3 and trimethoprim-d9 were 

obtained from Toronto Research Chemicals (Ontario, Canada). 

Phenytoin-d10 and atenolol-d7 were obtained from C/D/N 

Isotopes (Pointe-Claire, Canada). Carbamazepine-d10 and primidone-d5 

were obtained from Cambridge Isotope Laboratories 

(Andover, MA, USA). All solvents were trace analysis grade 

from Burdick and Jackson (Muskegon, MI). Reagent water was 

obtained using a Milli-Q Ultrapure Water Purification System 

(Millipore, Bedford, MA, USA). All concentrated stocks were 

prepared in methanol and stored at 20 C, while mixed 

spiking solutions were prepared in reagent water and stored 

at 4 C. 

On-line SPE was performed using a SymbiosisÔ Pharma 

(Spark Holland, Emmen, the Netherlands) automated SPE in 

XLC mode. Briefly, a 10-mL volume of sample was measured 

in a volumetric flask at which time isotopically-labeled 

standards were spiked at 100 ng/L. This provided sufficient 

sample volume for duplicates, matrix spikes, and dilutions, if 

necessary. A 1.5-mL aliquot of each sample was then 

transferred to a 2-mL autosampler vial, although only 1.0 mL 

was used for extractions. Extractions were performed using 

Waters Oasis HLB Prospekt cartridges (30 mm, 2.5 mg, 

10 1 mm, 96 tray) (Milford, MA). Prior to sample loading, 

each cartridge was sequentially conditioned with 1 mL of 

dichloromethane (DCM), MTBE, methanol, and reagent water 

(Milli-Q). Samples were loaded onto the SPE cartridge at 

1 mL/min after which the cartridge was washed with 1 mL of 

reagent water. After sample loading, the analytes were 

eluted from the SPE cartridge to the LC column with 200 mL 

methanol, using the LC peak focusing mode. A 5-mM 

ammonium acetate solution and methanol gradient was 

used for LC mobile phases with a flow rate of 800 mL/min. 

Analytes were separated using a 150 4.6 mm Luna C18(2) 

with a 5-mm particle size (Phenomenex, Torrance, CA). 

Method reporting limits (MRLs) were chosen at 3 to 5 times 

the method detection limit (MDL). MRLs were 10 ng/L for all 

compounds, except for atenolol, which was 25 ng/L. All 

analytes were quantified using isotope dilution. Stringent 

QA/QC protocols (i.e., matrix spikes, duplicate samples, field 

blanks, and laboratory blanks) were followed throughout the 

duration of the experiment. Using the on-line SPE method, 

the concentrations of duplicate samples rarely varied by 

greater than 5%. 


3.1. Experiment 1 – batch experiment with tertiarytreated 

wastewater 

Contaminant degradation profiles during the batch wastewater 

experiment are provided in Fig. 2. The initial ambient 

concentrations for the target contaminants ranged from 

36 ng/L for trimethoprim to as high as 378 ng/L for atenolol 

(Table 1). Although included in the analytical method, atrazine 

was below the method reporting limit (MRL) for both 

wastewater experiments. As indicated in Table 2, the generator 

was operated at a frequency of 500 Hz, a voltage of 8.0 kV, 

and a recirculation rate of 8.0 L/min. 


C/C 0 

1.0 

0.8 

0.6 

0.4 

0.2 

0.0 

0 1 2 3 4 5 6 7 8 

Generator Energy Consumption (kWh/m 3 ) 

Meprobamate Dilantin Primidone Carbamazepine Atenolol Trimethoprim 

Fig. 2 – Contaminant degradation profiles from the batch 

experiment (1) with ambient concentrations of trace 

organic compounds in tertiary-treated wastewater. This 

graph illustrates the relationship between contaminant 

degradation and generator (AOP-specific) energy 

consumption. The dashed lines represent samples that 

were below the MRL for those compounds. 

The degradation of each detected contaminant was 

assumed to be pseudo first order based on a plot of ln(C/C 0) 

versus generator energy consumption. Degradation rates, 95% 

confidence intervals, and R 2 values are provided in Table 3. 

Due to the low initial trimethoprim concentration, the extent 

of degradation was limited, as indicated by the dashed line 

representing the MRL in Fig. 2. Despite the limited degradation, 

it is apparent that trimethoprim and carbamazepine are 

highly susceptible to NTP treatment. This result is consistent 

with previous studies assessing oxidation strategies for trace 

organic compounds. Westerhoff et al. (2005) reported nearly 

100% degradation of trimethoprim and carbamazepine with 

standard chlorination and ozonation practices. Dilantin and 

atenolol degraded more slowly during the NTP treatment, 

which is also consistent with oxidation by chlorine or ozone 

(Westerhoff et al., 2005). Finally, meprobamate has been 

shown to be highly resistant to chlorination and ozonation in 

comparison to other trace organic compounds (Westerhoff 

et al., 2005). This conclusion is supported by the NTP results in 

that meprobamate required nearly double the energy to 

Table 3 – Pseudo first order rate constants for batch 

experiment (1) in tertiary-treated wastewater. 

Compound k (m 3 /kWh) 95% CI b 

R 2 

Meprobamate 0.36 0.03 0.98 9 

Dilantin 0.52 0.15 0.92 5 

Primidone 0.40 0.15 0.88 5 

Carbamazepine 1.03 0.35 0.93 4 

Atenolol 0.61 0.22 0.88 5 

Trimethoprim 0.71 N/A 0.92 2 

Atrazine a 

N/A N/A N/A N/A 

a Atrazine concentrations were below MRL in all samples. 

b 95% CI ¼ 95% confidence interval. 

N

498 

achieve a specified level of degradation. Primidone was not 

quite as resistant as meprobamate, but its degradation was 

slower and more variable than the other compounds. 

With respect to bulk organic parameters, there was no 

significant change in TOC over the duration of the experiment. 

The consistent TOC level was expected considering that 

mineralization is highly energy intensive in hydroxyl radicaldominated 

processes (Gerrity et al., 2009), and high CT values 

are required in ozone-dominated processes (Rosal et al., 2009). 

The limited energy consumption, low dissolved ozone 

concentrations, and limited ozone contact time were likely 

insufficient to induce measurable organic mineralization in 

the NTP reactor. 

Although there was no reduction in TOC, UV254 absorbance 

was reduced by more than 65%; this change was correlated to 

the degradation of target compounds in Fig. 3 as a potential 

surrogate measure of process efficacy. Quantifying trace 

organic compounds requires specialized equipment and 

training, and the analytical methods are very expensive and 

time-consuming. For this reason, it is generally impractical for 

water and wastewater utilities to employ trace contaminant 

monitoring programs for their systems. Implementation of 

this type of surrogate framework (i.e., reduction in UV 254 

absorbance in lieu of trace contaminant degradation) would 

allow utilities to assess the performance of certain treatment 

processes for emerging contaminants. For oxidation 

processes, UV254 absorbance is an easily measured parameter 

that appears to show consistent correlations with the degradation 

of trace organic compounds (Wert et al., 2009). Fig. 3 

provides support for this surrogate framework because it 

demonstrates strong correlations between the degradation of 

the target compounds and reduction in UV 254 absorbance. The 

slopes of the regression lines indicate that relative contaminant 

degradation (i.e., percent degradation) is actually more 

rapiddmore than twice as fast for nearly all compoundsdthan 

the reduction in UV254 absorbance. The slopes 

also show a general agreement with the degradation profiles 

% Degradation of Trace Organic Compounds 

100% 

80% 

60% 

40% 

20% 


in Fig. 2. Although trimethoprim was expected to degrade 

rapidly, its slope of 6.25 is likely an outlier and artifact of its 

small sample size. With the exception of trimethoprim, the 

magnitude of the slopes and the negative intercepts are 

similar to those reported in the Wert et al. (2009) ozonation 

study. The negative intercepts indicate that there are initial 

reductions in UV254 absorbance that precede degradation of 

the target compounds. This may be attributable to initial 

oxidant/radical scavenging by the bulk organic matter. 

Values for electrical energy per order (EEO) of magnitude 

degradation (Bolton et al., 1996), which is often measured in 

kWh/m 3 -log, are also provided in Table 4. Since the batch 

experiment involved a continuous treatment, the EEO values 

were calculated based on first order degradation rates and 

linear regression. In other words, a linear regression was 

generated for ln(C/C0) against generator energy consumption, 

and the EEO was calculated as ln(0.1) divided by the slope of 

the linear regression. The EEO values ranged from 2.2 kWh/m 3 

for carbamazepine to 6.4 kWh/m 3 for meprobamate. It is 

important to note that the batch EEO values may be overestimates 

because the batch system requires approximately 

10–15 min to achieve its maximum ozone concentration after 

starting the generator. 

3.2. Experiment 2 – single-pass experiment with 

tertiary-treated wastewater 

Contaminant degradation profiles for the single-pass wastewater 

experiment are illustrated in Fig. 4. The initial ambient 

concentrations for the target contaminants were similar to 

the batch experiment and ranged from 39 ng/L for trimethoprim 

to as high as 413 ng/L for atenolol (Table 1). As indicated 

in Table 2, this experiment involved a single-reactor 

configuration and both reactors operating in series at a flow 

rate of 15.5 L/min. All scenarios for this experiment involved 

voltages of 8.0 kV. The reactor consumed 0.7 kWh/m 3 with 

Generator 2 operating at a frequency of 500 Hz during the 

Trimethoprim 6.25 -90% 1.00 2 

0% 

0% 20% 40% 60% 80% 100% 

% Reduction in UV 254 Absorbance 

Linear Regression Models: 

Contaminant Slope Intercept R 2 N 

Meprobamate 1.56 -14% 0.97 8 

Dilantin 2.09 -22% 0.99 4 

Primidone 2.08 -28% 0.98 4 

Carbamazepine 2.75 -30% 1.00 3 

Atenolol 2.68 -44% 0.99 4 

Fig. 3 – Correlation between the degradation of target compounds and UV 254 absorbance during the batch experiment (1). 

This analysis excludes samples that were below the MRL and the final five carbamazepine samples (>3 kWh/m 3 ; see Fig. 2) 

due to high variability near the MRL. The slopes were all significant at the 0.03 level or better, and intercepts were all 

significant at the 0.07 level or better.

Table 4 – AOP-specific EEO values (kWh/m 3 -log). 

Contaminant (1) (2) (3) UV c 

Batch 

SSW 

a 

Min. 

WW b 

Med. 

WW b 

Max. 

WW b 

Min. 

SSW b 

Med. 

SSW b 

Max. 

SSW b 

WW-1 scenario, and the reactor consumed 1.0 kWh/m 3 with 

Generator 1 operating at a frequency of 800 Hz during the 

WW-2 scenario. WW-3 and WW-4 involved both reactors 

operating in series at 600/500 Hz (1.6 kWh/m 3 ) and 800/500 Hz 

(1.8 kWh/m 3 ), respectively. The energy consumption values 

for the entire skid, which include energy related to the 

generators, pumps, and controls, are also provided for each 

scenario. 

The relative degradation rates were identical to those of 

the batch experiment. For example, carbamazepine and 

trimethoprim experienced the most rapid degradation, and 

primidone and meprobamate were the most recalcitrant 

compounds. At the highest level of energy consumption 

(1.8 kWh/m 3 ), every compound except meprobamate experienced 

at least 70% degradation. The evaluation of 

C/C 0 

UV/H 2O 2 c,d 

SSW 

UV/TiO 2 c,e 

SSW 

Meprobamate 6.4 4.6 10 14 2.1 3.5 5.3 6.6 1.0 6.8 

Dilantin 4.4 2.7 3.1 3.5 1.1 2.0 3.1 2.1 1.0 2.2 

Primidone 5.8 2.8 4.3 4.8 1.1 2.2 3.3 3.7 0.6 3.9 

Carbamazepine 2.2 1.3 1.8 2.1

500 

Fig. 5 – Excitation-emission matrices as a function of generator energy consumption for the single-pass experiment (2) with 

tertiary-treated wastewater. The regions are described as follows: (I) Aromatic protein, (II) aromatic protein II, (III) fulvic 

acid-like, (IV) soluble microbial byproduct-like, and (V) humic acid-like (Chen et al., 2003). 

scope of this study, but future studies should address this 

issue in greater detail. 

3.3. Experiment 3 – single-pass experiment with spiked 

surface water 

Contaminant degradation profiles, including that of atrazine, 

for the single-pass spiked surface water experiment are 

illustrated in Fig. 6. The initial spiked concentrations were 

much higher than the ambient wastewater concentrations as 

all of the compounds were spiked at greater than 900 ng/L 

(Table 1). Similar to the single-pass wastewater experiment, 

this experiment involved a single-reactor configuration and 

both reactors operating in series at a flow rate of 11.4 L/min 

(Table 2). For this experiment, Generator 1 was operated at 

a voltage of 8.0 kV, and Generator 2 was operated at a voltage 

of 5.9 kV. The generator and overall skid energy consumption 

values ranged from 0.6 to 2.6 kWh/m 3 and 4.8 to 6.8 kWh/m 3 , 

respectively. 

The relative degradation rates were similar to those of the 

wastewater experiments. For example, meprobamate was the 

most difficult to degrade, while trimethoprim and carbamazepine 

rapidly dropped below the MRL in all treated samples. 

The lower organic loading of the surface water (i.e., lower 

initial TOC concentration and UV254 absorbance), which 


resulted in a higher residual ozone concentration (0.25 to 

0.50 mg/L), likely contributed to the improved degradation. In 

contrast to the wastewater experiments, atenolol degradation 

was more similar to trimethoprim and carbamazepine, while 

primidone degradation was more similar to that of dilantin. 

As expected, the degradation of atrazine, which is considered 

a recalcitrant compound (Westerhoff et al., 2005), was almost 

identical to that of meprobamate. For all of the compounds 

with concentrations above the MRL, the degradation peaked at 

about 70% for the most recalcitrant compounds and 85% for 

the moderately recalcitrant compounds after reaching an 

energy consumption of 1.6 kWh/m 3 . At the highest energy 

level, the effluent concentrations of the residual compounds 

ranged from 140 ng/L for dilantin to 390 ng/L for atrazine. 

During Experiment 3, there was also greater fluctuation for 

several time points. Since the samples represent discrete, 

rather than continuous (i.e., batch), treatment levels, one 

would anticipate greater variability, but there was still a relatively 

consistent decrease in contaminant concentration over 

the range of energy values tested. 

The EEO values for all of the compounds were dramatically 

lower in the surface water experiment than the wastewater 

experiments, presumably due to the lower scavenging 

capacity of the surface water. Meprobamate and atrazine had 

the highest EEOs with maximum values of 5.3 and 6.3 kWh/

C/C 0 

1 

0.8 

0.6 

0.4 

0.2 

0 

0.0 0.5 1.0 1.5 2.0 2.5 3.0 

m 3 -log, respectively. The maximum values generally occurred 

during the experiments when one or both of the generators 

were operating at 1,000 Hz. Although operating both generators 

at 1,000 Hz achieved the highest level of degradation, 

increasing the pulse frequency generally yielded diminishing 

rates of return with respect to efficiency. Operating Generator 

2 at a frequency of 600 Hz provided the most efficient treatment 

conditions for all of the target compounds. However, 

Generator 1 had been operated immediately prior to that 

scenario, so the residual ozone from Reactor 1 likely contributed 

to this increased efficiency. Although this overestimates 

the anticipated degree of degradation, it indicates that it may 

be beneficial to incorporate ozone contact tanks before and 

after each reactor to maximize degradation. Based on this 

Generator Energy Consumption (kWh/m 3 ) 

Meprobamate Dilantin Primidone Carbamazepine Atenolol Trimethoprim Atrazine 

Fig. 6 – Contaminant degradation profiles from the single-pass experiment (3) with spiked surface water. This graph 

illustrates the relationship between contaminant degradation and generator (AOP-specific) energy consumption. 

Concentration (ng/L) 

1600 

1400 

1200 

1000 

800 

600 

400 

200 

0 


observation, the median values are likely the most 

appropriate EEO estimates since they always corresponded 

to both reactors operating simultaneously. For the 

compounds with reportable concentrations, the median 

EEO values ranged from 1.0 to 3.7 kWh/m 3 for atenolol and 

atrazine, respectively. 

A subset of samples was also analyzed to assess the 

contribution of the system components (i.e., reactor vs. ozone 

contactor) to contaminant degradation. The results from one 

set of samples (SSW-8) are provided in Fig. 7; the other sample 

sets yielded similar trends and therefore are not shown. Based 

on these results, it appears that the ozone contactors provided 

the majority of the degradation for the more recalcitrant 

compounds, including meprobamate, dilantin, primidone, 

Influent Reactor 1 Ozone 1 Reactor 2 Effluent 

Sample Location 

Meprobamate Dilantin Primidone Carbamazepine Atenolol Trimethoprim Atrazine 

Fig. 7 – Comparison of the pilot system components in degrading the target compounds during the single-pass experiment 

(3) with spiked surface water. This data refers to samples from SSW-8 with generator and skid energy consumption values 

of 2.2 and 6.3 kWh/m 3 , respectively. The small increase in concentrations from Ozone 1 to Reactor 2 is explained by mixing 

of Reactor 1 and Ozone 1 effluent prior to passage into Reactor 2 (see Fig. 1).

502 

% Degradation of Trace Organic Compounds 

100% 

80% 

60% 

40% 

20% 

Linear Regression Models: 

Contaminant Slope Intercept R 2 

N 

Meprobamate 2.40 0 0.72 9 

Dilantin 1.35 45% 0.84 9 

Primidone 1.52 40% 0.81 9 

Atrazine 2.27 0 0.74 9 

0% 

0% 5% 10% 15% 20% 25% 30% 35% 

atenolol, and atrazine. However, carbamazepine and 

trimethoprim experienced rapid degradation in the reactor 

itself with the additional oxidation pathways (i.e., UV light, 

hydroxyl radicals, and ozone). 

The spiked surface water experiment was also evaluated 

based on general water quality parameters. For example, the 

initial bromide concentration in the surface water was 

approximately 62 mg/L, and additional analyses indicated that 

bromate increased only slightly (maximum of 2.4 mg/L) above 

the ambient surface water concentration of 1.7 mg/L. Similar 

to the wastewater experiments, there was no significant 

change in TOC over the duration of the experiment, but the 

reduction in UV254 absorbance was correlated to the degradation 

of the target contaminants in Fig. 8. In a manner 

similar to Fig. 3, Fig. 8 illustrates strong positive correlations 

between contaminant degradation and reduction in UV 254 

absorbance. Furthermore, the degradation of target contaminants 

was still more rapid than the reduction in UV 254 

absorbance. In contrast to the batch wastewater experiment 

and data provided in Wert et al. (2009), the regression lines 

for dilantin and primidone have positive intercepts, while the 

intercepts for meprobamate and atrazine were found to be 

insignificant ( p > 0.09). Wert et al. (2009) proposed that the 

correlations between contaminant degradation and UV254 

absorbance were independent of wastewater quality, but it 

appears that the correlations differ when the matrices 

themselves are drastically different (i.e., wastewater vs. 

surface water). Specifically, the positive intercepts for 

dilantin and primidone in the surface water experiment 

indicate that these target compounds started to degrade 

(z40–45%) prior to any reduction in UV 254 absorbance. In 

addition to contributing to their smaller slopes, this may 

indicate that there was less oxidant/radical scavenging and 

that the bulk organic matter was more recalcitrant in the 

surface water compared to the wastewater. It is also interesting 

to note that meprobamate has the smallest absolute 

intercept for the wastewater experiment and an insignificant 


% Reduction in UV254 Absorbance 

Meprobamate Dilantin Primidone Atrazine 

Fig. 8 – Correlation between the degradation of detected target compounds and UV254 absorbance during the single-pass 

experiment (3) with spiked surface water. The slopes were all significant at the 0.001 level, the intercepts for primidone and 

dilantin were significant at the 0.001 level, and the intercepts for meprobamate and atrazine were considered insignificant 

( p > 0.09; regressions for meprobamate and atrazine were forced through the origin). 

intercept for the surface water experiment, and atrazine also 

has an insignificant intercept for the surface water experiment. 

Therefore, the initial degradation of these highly 

resistant compounds is more similar to the initial degradation 

of the bulk organic matter. For the surface water 

experiments, this supports the theory that the aromatic 

organic matter in the surface water was more recalcitrant 

than that of the wastewater. Based on these observations, it 

appears that the UV 254 surrogate framework is still valid, but 

there may be differences in the correlations for dissimilar 

water matrices. 

Table 4 provides EEO values for the three experiments in 

this study, and it also provides EEO values from Benotti et al. 

(2009b), which evaluated the efficacy of photolysis, UV/H2O2, 

and photocatalysis in degrading PPCPs and EDCs. Benotti et al. 

(2009b) provides an interesting point of comparison since the 

water matrix used for that study (spiked surface water from 

Lake Mead), target compounds, and analytical methods were 

identical. As indicated by the data, NTP was generally more 

efficient than photolysis and photocatalysis for each of the 

compounds, but was either comparable or slightly less efficient 

than UV/H2O2 at a peroxide dose of 10 mg/L. It is 

important to note that the NTP technology does not require 

UV lamps or costly peroxide feeds to generate the oxidative 

species. Therefore, NTP may be a viable alternative to the 

more common UV/H2O2 process despite the slightly higher 

energy requirements for some compounds. Although the NTP 

achieved comparable efficiencies to those of UV/H 2O 2 in these 

experiments, identification of the best alternative for 

a particular application also depends on capital costs and 

reliability. Since the NTP technology has not been fully 

commercialized, it is difficult to compare its capital costs and 

reliability with more established technologies. Future evaluations 

of large-scale NTP reactors, including long-term 

studies, will provide valuable information in determining how 

NTP compares to more established technologies on a life-cycle 

basis.


With respect to treatment efficacy, the pilot-scale NTP system 

demonstrated rapid degradation of the target compounds. 

The relative degradation rates were consistent between the 

three phases of the study with carbamazepine and trimethoprim 

as the most susceptible compounds and meprobamate, 

primidone, and atrazine as the most recalcitrant compounds. 

Atenolol and dilantin were generally more resistant to 

oxidation than carbamazepine and trimethoprim. These 

similar relative degradation rates justify the use of indicator 

compounds in evaluations of PPCP and EDC treatment. The 

EEO values for the optimal scenarios were generally less than 

5 kWh/m 3 -log for all of the target compounds, particularly in 

the spiked surface water with a lower organic loading. With 

process optimization (i.e., optimization of electrode height for 

each water matrix) and limited modifications to the system, 

including more effective use of ambient ozone, the EEO values 

could likely be reduced further. 

More common AOPs, including UV/H 2O 2 and ozone/H 2O 2, 

require significant chemical addition and residual H 2O 2 

quenching, which represent a significant portion of their 

operational costs. The primary benefit of the NTP AOP is the 

ability to generate UV light, ozone, and hydroxyl radicals 

without chemical addition or the use of UV lamps. However, 

the true capital costs and reliability of large-scale NTP reactors 

for water and wastewater treatment are still relatively unclear. 

Also, since oxidation by molecular ozone may be the dominant 

mechanism for many compounds, future studies should 

provide direct comparisons of NTP with conventional ozone 

generators to compare system performance. These expanded 

studies are necessary before one can fully determine the 

applicability of the novel NTP technology for water and 

wastewater treatment. However, these preliminary results 

indicate that NTP may be a viable alternative to more common 

AOPs due to its chemical-free operation and comparable 

energy requirements for trace contaminant degradation. 


We would like to thank the SNWA Applied Research and 

Development Center staff, particularly Janie Zeigler, Shannon 

Ferguson, Christy Meza, and Elaine Go, for assisting with the 

various analyses. We would also like to thank Dvir Solnik, Itay 

Even-Ezra, Gal Bitan, and Shahar Nuriel of Aquapure Technologies 

for use of the nonthermal plasma pilot unit and for 

providing crucial training and guidance during the study. 

references 

Acero, J.L., von Gunten, U., 2001. Characterization of oxidation 

processes: ozonation and the AOP O 3/H 2O 2. J. AWWA 93 (10), 

90–100. 

Benotti, M.J., Stanford, B.D., Snyder, S.A., 2009a. Increased loading 

of pharmaceuticals and endocrine disrupting compounds to 

source drinking water due to climate change. J. Environ. Qual. 

Accepted. 


Benotti, M.J., Stanford, B.D., Wert, E.C., Snyder, S.A., 2009b. 

Evaluation of a photocatalytic reactor membrane pilot system 

for the removal of pharmaceuticals and endocrine disrupting 

compounds from water. Water Res. 43 (6), 1513–1522. 

Benotti, M.J., Trenholm, R.A., Vanderford, B.J., Holaday, J.C., 

Stanford, B.D., Snyder, S.A., 2009c. Pharmaceuticals and 

endocrine disrupting compounds in U.S. drinking water. 

Environ. Sci. Technol. 43 (3), 597–603. 

Bolton, J.R., Bircher, K.G., Tumas, W., Tolman, C.A., 1996. Figuresof-merit 

for the technical development and application of 

advanced oxidation processes. J. Adv. Oxid. Technol. 1 (1), 

13–17. 

Chen, W., Westerhoff, P., Leenheer, J.A., Booksh, K., 2003. 

Fluorescence excitation-emission matrix regional integration 

to quantify spectra for dissolved organic matter. Environ. Sci. 

Technol. 37 (24), 5701–5710. 

Even-Ezra, I., Mizrahi, A., Gerrity, D., Snyder, S., Salveson, A., 

Lahav, O., 2009. Application of a novel plasma-based advanced 

oxidation process for efficient and cost effective destruction of 

refractory organics in tertiary effluents and contaminated 

groundwater. Desalin. Water Treat. Accepted. 

Gerrity, D., Mayer, B., Ryu, H., Crittenden, J., Abbaszadegan, M., 

2009. A comparison of pilot-scale photocatalysis and 

enhanced coagulation for disinfection byproduct mitigation. 

Water Res. 43, 1597–1610. 

Huber, M.M., Canonica, S., Park, G., von Gunten, U., 2003. 


oxidation processes. Environ. Sci. Technol. 37, 1016–1024. 

Hwang, Y.H., Jo, Y.M., 2005. Decomposition of odorous gases in 

a pilot-scale nonthermal plasma reactor. J. KOSAE 21 (2), 

57–65. 

Ighigeanu, D., Martin, D., Zissulescu, E., Macarie, R., Oproiu, C., 

Cirstea, E., Iovu, H., Calinescu, I., Iacob, N., 2005. SO 2 and NO x 

removal by electron beam and electrical discharge induced 

non-thermal plasmas. Vacuum 77, 493–500. 




waters. Water Res. 41, 1013–1021. 

Kim, H., Han, J., Sakaguchi, Y., Minami, W., 2008. Simultaneous 

oxidation of NO x and SO 2 by a new non-thermal plasma 

reactor enhanced by catalyst and additive. Plasma Sci. 

Technol. 10 (1), 53–56. 

Kolpin, D.W., Furlong, E.T., Meyer, M.T., Thurman, E.M., Zaugg, S. 

D., Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, 

hormones, and other organic wastewater contaminants in U. 

S. streams, 1999–2000: A national reconnaissance. Environ. 

Sci. Technol. 36, 1202–1211. 

Locke, B.R., Sato, M., Sunka, P., Hoffmann, M.R., Chang, J.S., 2006. 

Electrohydraulic discharge and nonthermal plasma for water 

treatment. Ind. Eng. Chem. Res. 45, 882–905. 

Masuda, S., Hosokawa, S., Tu, X., Wang, Z., 1995. Novel plasma 

chemical technologiesdPPCP and SPCP for control of gaseous 

pollutants and air toxics. J. Electrostatics 34, 415–438. 

Mitch, W.A., Sharp, J.O., Trussell, R.R., Valentine, R.L., Alvarez- 

Cohen, L., Sedlak, D.L., 2003. N-nitrosodimethylamine (NDMA) 

as a drinking water contaminant: a review. Environ. Eng. Sci. 

20 (5), 389–404. 

Pekarek, S., 2003. Non-thermal plasma ozone generation. Acta 

Polytech. 43 (6), 47–51. 

Pereira, V.J., Linden, K.G., Weinberg, H.S., 2007. Evaluation of UV 

irradiation for photolytic and oxidative degradation of 

pharmaceutical compounds in water. Water Res. 41 (19), 

4413–4423. 

Rosal, R., Rodriguez, A., Perdigon-Melon, J.A., Petre, A., Garcia- 

Calvo, E., 2009. Oxidation of dissolved organic matter in the 

effluent of a sewage treatment plant using ozone combined 

with hydrogen peroxide (O 3/H 2O 2). Chem. Eng. J. 149, 311–318.

504 

Snyder, S.A., Wert, E.C., Lei, H., Westerhoff, P., Yoon, Y., 2007. 

Removal of EDCs and Pharmaceuticals in Drinking and 

Reuse Treatment Processes. Water Research Foundation, 

Denver, CO. 

Snyder, S.A., Westerhoff, P., Yoon, Y., Sedlak, D.L., 2003. 

Pharmaceuticals, personal care products, and endocrine 

disruptors in water: implications for the water industry. 

Environ. Eng. Sci. 20 (5), 449–469. 


treatment plants and rivers. Water Res. 32 (11), 3245–3260. 

Ternes, T.A., Meisenheimer, M., McDowell, D., Sacher, F., 

Brauch, H., Haist-Gulde, B., Preuss, G., Wilme, U., Zulei- 

Seibert, N., 2002. Removal of pharmaceuticals during drinking 

water treatment. Environ. Sci. Technol. 36, 3855–3863. 

Thevenet, F., Guaitella, O., Puzenat, E., Herrmann, J.M., 

Rousseau, A., Guillard, C., 2007. Oxidation of acetylene by 

photocatalysis coupled with dielectric barrier discharge. Catal. 

Today 122, 186–194. 


Trenholm, R.A., Vanderford, B.J., Snyder, S.A., 2009. On-line SPE 

LC-MS/MS analysis of pharmaceutical indicators in water: 

a green alternative to conventional methods. Talanta 79, 

1425–1432. 

Wert, E.C., Rosario-Ortiz, F.L., Snyder, S.A., 2009. Using ultraviolet 

absorbance and color to assess pharmaceutical oxidation 

during ozonation of wastewater. Environ. Sci. Technol. 43, 

4858–4863. 




processes. Environ. Sci. Technol. 39, 6649–6663. 

Westerhoff, P., Moon, H., Minakata, D., Crittenden, J., 2009. 

Oxidation of organics in retentates from reverse osmosis 

wastewater reuse facilities. Water Res. 43 (16), 3992–3998. 

Yates, R.S., Stenstrom, M.K., 2000. Gravimetric sampling 

procedure for aqueous ozone concentrations. Water Res. 34 

(4), 1413–1416.

Single-walled carbon nanotubes dispersed in aqueous media 

via non-covalent functionalization: Effect of dispersant 

on the stability, cytotoxicity, and epigenetic toxicity 

of nanotube suspensions 

Alla L. Alpatova a , Wenqian Shan a , Pavel Babica b , Brad L. Upham b,c , 

Adam R. Rogensues a , Susan J. Masten a , Edward Drown d,1 , Amar K. Mohanty d,2 , 

Evangelyn C. Alocilja e , Volodymyr V. Tarabara a, * 

a 

Department of Civil and Environmental Engineering, Michigan State University, East Lansing, MI 48824, USA 

b 

Center for Integrative Toxicology, Michigan State University, East Lansing, MI 48824, USA 

c 

Department of Pediatrics and Human Development, Michigan State University, East Lansing, MI 48824, USA 

d 

School of Packaging, Michigan State University, East Lansing, MI 48824, USA 

e 

Department of Biosystems and Agricultural Engineering, Michigan State University, East Lansing, MI 48824, USA 

article info 


Received 2 June 2009 





Keywords: 

Single-walled carbon nanotubes 

Dispersion 

Non-covalent functionalization 

Cytotoxicity 

Epigenetic toxicity 




abstract 

As the range of applications for carbon nanotubes (CNTs) rapidly expands, understanding 

the effect of CNTs on prokaryotic and eukaryotic cell systems has become an important 

research priority, especially in light of recent reports of the facile dispersion of CNTs in 

a variety of aqueous systems including natural water. In this study, single-walled carbon 

nanotubes (SWCNTs) were dispersed in water using a range of natural (gum arabic, 

amylose, Suwannee River natural organic matter) and synthetic (polyvinyl pyrrolidone, 

Triton X-100) dispersing agents (dispersants) that attach to the CNT surface non-covalently 

via different physiosorption mechanisms. The charge and the average effective hydrodynamic 

diameter of suspended SWCNTs as well as the concentration of exfoliated SWCNTs 

in the dispersion were found to remain relatively stable over a period of 4 weeks. The 

cytotoxicity of suspended SWCNTs was assessed as a function of dispersant type and 

exposure time (up to 48 h) using general viability bioassay with Escherichia coli and using 

neutral red dye uptake (NDU) bioassay with WB-F344 rat liver epithelia cells. In the E. coli 

viability bioassays, three types of growth media with different organic loadings and salt 

contents were evaluated. When the dispersant itself was non-toxic, no losses of E. coli and 

WB-F344 viability were observed. The cell viability was affected only by SWCNTs dispersed 

using Triton X-100, which was cytotoxic in SWCNT-free (control) solution. The epigenetic 

toxicity of dispersed CNTs was evaluated using gap junction intercellular communication 

(GJIC) bioassay applied to WB-F344 rat liver epithelial cells. With all SWCNT suspensions 

except those where SWCNTs were dispersed using Triton X-100 (wherein GJIC could not be 

measured because the sample was cytotoxic), no inhibition of GJIC in the presence of 

SWCNTs was observed. These results suggest a strong dependence of the toxicity of 

* Corresponding author. Tel.: þ517 432 1755; fax: þ517 355 0250. 

E-mail address: tarabara@msu.edu (V.V. Tarabara). 

1 

Present address: Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, MI 48824, USA. 

2 

Present address: Department of Plant Agriculture, University of Guelph, 50 Stone Rd. E., Guelph, Ontario, N1 G 2W1 Canada. 


doi:10.1016/j.watres.2009.09.042

506 


The discovery (Iijima, 1991) and subsequent extensive characterization 

of carbon nanotubes (CNTs) have revealed a class 

of materials with extraordinary electrical, mechanical, and 

thermal properties (Tasis et al., 2006). The wider application of 

CNTs in electronic, optical, sensing, and biomedical fields has 

been impeded by the low solubility of as-produced CNTs in 

polar liquids and by the strong tendency of CNTs to aggregate 

due to hydrophobic–hydrophobic interactions (Lin et al., 2004). 

Dispersing CNTs and ensuring long-term stability of CNTs 

suspended in a liquid medium have proved especially challenging 

for aqueous systems. 

1.1. Dispersing CNTs in aqueous media 

Recent efforts on the development of efficient and facile 

methods of dispersing CNTs in aqueous media have been 

focused on the hydrophilization of CNT with molecules that 

bind to the CNT surface non-covalently (O’Connell et al., 2001; 

Bandyopadhyaya et al., 2002; Star et al., 2002; Islam et al., 2003; 

Moore et al., 2003; Didenko et al., 2005; Wang et al., 2005; 

Bonnet et al., 2007; Grossiord et al., 2007; Hyung et al., 2007; Liu 

et al., 2007). Such non-covalent functionalization has great 

promise as the modification-induced changes in the electronic 

and mechanical properties of CNTs are minimized 

(Yang et al., 2007). Various surfactants (Islam et al., 2003; 

Moore et al., 2003; Grossiord et al., 2007), synthetic polymers 

(e.g., polyvinyl pyrrolidone (O’Connell et al., 2001; Didenko 

et al., 2005), poly(ethylene glycol) (Vaisman et al., 2006), polyphosphazene 

(Park et al., 2006)), natural organic matter 

(NOM) (Hyung et al., 2007; Liu et al., 2007; Saleh et al., 2009), 

biomolecules (e.g., proteins (Karajanagi et al., 2006; Zong et al., 

2007), aminoacids (Georgakilas et al., 2002), DNA (Enyashin 

et al., 2007)), and carbohydrates (e.g., cyclodextrines (Dodziuk 

et al., 2003), amylose (Bonnet et al., 2007), starch (Star et al., 

2002), and GA (Bandyopadhyaya et al., 2002)) have been evaluated 

as dispersants for CNTs. Dispersion via non-covalent 

functionalization is based on the direct contact between 

a CNT and a dispersant molecule (Liu et al., 1998; Grossiord 

et al., 2007). Such modification of the CNT surface facilitates 

the disaggregation (i.e. debundling) of CNT bundles into 

smaller diameter bundles (Liu et al., 1998) or even individual 

CNTs (O’Connell et al., 2002; Hyung et al., 2007) and leads to 

the stabilization of suspended CNTs via steric or electrostatic 

repulsion mechanisms or both. (See Supporting Documentation 

(SD), Section S.1.1, for a brief review of mechanisms of 

non-covalent functionalization of CNTs.) 

The dispersion of CNTs in water has been enhanced by 

mixing (Didenko et al., 2005; Hyung et al., 2007), sonication 

(Bandyopadhyaya et al., 2002; Liu et al., 2007; Salzmann et al., 

2007), or mixing followed by sonication (O’Connell et al., 2001, 


SWCNT suspensions on the toxicity of the dispersant and point to the potential of noncovalent 

functionalization with non-toxic dispersants as a method for the preparation of 

stable aqueous suspensions of biocompatible CNTs. 


2002; Star et al., 2002; Islam et al., 2003; Moore et al., 2003; 

McDonald et al., 2006; Grossiord et al., 2007). These treatments 

were applied both in the absence (Salzmann et al., 2007) and in 

the presence of solubilizing agents: NOM (Hyung et al., 2007), 

Triton X-100 (Islam et al., 2003), Triton X-405 (Chappell et al., 

2009), PVP-1300 (Didenko et al., 2005), GA (Bandyopadhyaya 

et al., 2002), and starch (Star et al., 2002). A three-step 

approach to solubilizing single-walled carbon nanotubes 

(SWCNTs) with amylose was suggested by Kim et al. (Kim 

et al., 2004): dispersion of SWCNT in water by sonication followed 

by treatment with amylose in dimethylsulfoxide 

(DMSO)–H2O mixture, followed by sonication allowing for 

molecularly controlled encapsulation of CNTs. 

1.2. Stability of CNT suspensions in water 

Previous studies of the long-term changes in suspensions of 

dispersed CNTs have focused on monitoring changes in the 

concentration of suspended CNTs (Jiang and Gao, 2003; Tseng 

et al., 2006; Lee et al., 2007; Marsh et al., 2007). By measuring 

UV–vis absorption at certain wavelength: 253 nm (Jiang and 

Gao, 2003), 300 nm (Sinani et al., 2005), 500 nm (Bahr et al., 

2001; Lee et al., 2007), 530 nm (Marsh et al., 2007), and 800 nm 

(Hyung et al., 2007) the change in the concentration of suspended 

CNTs with time was determined. 

Aqueous suspensions of non-functionalized CNTs are 

known to be unstable. There are considerable quantitative 

differences, however, in the reported stability data for nonfunctionalized 

CNTs. The concentration of unmodified multiwalled 

carbon nanotubes (MWCNTs) suspended in deionized 

water was reported to decline 86 % over 2 h in one study 

(Marsh et al., 2007) and only 50% over 500 h in another study 

(Jiang and Gao, 2003). The suspensions of unmodified SWCNTs 

in deionized water were found to completely precipitate after 

only 4 h (Tseng et al., 2006). 

It was found that non-covalent modification of CNT 

surface drastically improved the stability of CNT suspensions 

(Jiang and Gao, 2003; Tseng et al., 2006; Hyung et al., 2007; Lee 

et al., 2007; Marsh et al., 2007). The stability of the suspension 

of non-covalently functionalized CNTs was found to be 

a function of the type (Hyung et al., 2007; Lee et al., 2007) and 

concentration (Chappell et al., 2009) of the dispersant, CNT 

length (Marsh et al., 2007) and the presence of inorganic salts 

in the solution (Saleh et al., 2009). SWCNTs stabilized with 

oligothiophene-terminated poly(ethylene glycol) produced 

suspensions that were significantly more stable than CNTs 

dispersed in water with the aid of sodium dodecyl sulfate 

(SDS) and Pluronic F127 (Lee et al., 2007). CNT suspensions in 

model Suwannee River NOM (SRNOM) solutions and in 

Suwannee River water were found to be considerably more 

stable than suspensions of CNTs dispersed in 1% aqueous 

solution of SDS (Hyung et al., 2007). Chappell et al. showed

that the stability of MWCNTs dispersed using two types of 

humic acids, Triton X-405, Brij 35, and SDS was enhanced in 

dose-dependent manner (Chappell et al., 2009). In a study of 

the aggregation kinetics of MWCNTs in aquatic systems, 

Suwannee River humic acid was shown to significantly 

enhance stability of MWCNTs suspensions in the presence of 

monovalent and divalent salts (Saleh et al., 2009). While 

studying the stability of MWCNTs in water, Marsh and coworkers 

(Marsh et al., 2007) observed that wrapping of the 

annealed CNTs with 1% SDS or charge doping increased their 

stability in the suspension; the authors demonstrated that 

shorter CNTs form more stable dispersions than longer CNTs 

of the same diameter. In summary, there is growing evidence 

that by the appropriate choice of a dispersant that modifies 

CNT surface non-covalently, highly stable CNT suspensions 

can be produced. However, to date there have been no systematic 

investigations that correlated long-term changes in size and charge 

of CNTs dispersed via non-covalent functionalization with the 

dispersion stability. 

1.3. Toxicity of CNTs 

1.3.1. Toxicity of CNTs towards eukaryotic cells: cytotoxicity 

Most nanomaterial toxicity studies have been performed with 

mammalian cells, in particular with lung and skin cell 

cultures reflecting the understanding that the most likely 

routes of an organism’s exposure to nanomaterials are 

respiratory and dermal contact. With the development of 

methods of CNT dispersion in aqueous media, the assessment 

of the toxicity of such dispersed CNTs becomes very important 

in view of their increased mobility and potential to enter 

water supplies. While the toxicity of CNTs was studied with 

respect to the type of the CNT (MWCNT versus SWCNT) (Ding 

et al., 2005; Jia et al., 2005; Kang et al., 2008a), surface functionalization 

(Sayes et al., 2006; Yu et al., 2007; Meng et al., 

2009; Yun et al., 2009), CNT length (Muller et al., 2005; Kang 

et al., 2008b; Simon-Deckers et al., 2008), exposure dose 

(Cherukuri et al., 2006; Flahaut et al., 2006; Sayes et al., 2006; 

Kang et al., 2007; Pulskamp et al., 2007; Yang et al., 2008, 2009; 

Ye et al., 2009), degree of purification (Fiorito et al., 2006; 

Flahaut et al., 2006; Elias et al., 2007; Simon-Deckers et al., 

2008), and degree of dispersion (Wick et al., 2007), only very 

limited information exists on the biocompatibility of CNTs as 

a function of surface coating. 

Several hypotheses have been put forth to explain the likely 

pathways of CNT toxicity: (i) oxidative stress induced by the 

formation of reactive oxygen species (ROS) generated at the 

surface of CNTs (Manna et al., 2005; Yang et al., 2008; Ye et al., 

2009); (ii) the presence of residual catalyst, which is used during 

the CNT manufacturing process (Kang et al., 2008a); 

(iii) physical contact between a CNT and a cell (Kang et al., 2007); 

or (iv) a combination of these factors (Kang et al., 2008a). The 

dispersant used for the stabilization of CNTs in suspension was 

suggested as another possible cause of the observed nanotube 

toxicity (Sayes et al., 2006) (SeeSD, Section S.1.2 for a brief 

review of possible mechanisms of CNT toxicity). 

In most studies of the toxicity of non-covalently functionalized 

CNTs the dispersants were surfactants. In vivo 

assessment of the toxicity of SWCNTs modified with Pluronic 

F108 administered intravenously to rabbits showed the 


absence of the acute toxicity (Cherukuri et al., 2006). Pluronic 

F-127-coated MWCNTs did not affect cell viability, apoptosis, 

and ROS formation in mouse and human neuroblastoma cells 

(Vittorio et al., 2009), and mouse cerebral cortex (Bardi et al., 

2009). In contrast, MWCNTs dispersed using Pluronic F68 

caused cell death, changes in cell size and complexity, ROS 

production, interleukin-8 (IL-8) gene expression and nuclear 

factor (NF)-jB activation (Ye et al., 2009). CNTs dispersed using 

Tween 80 were toxic to mesothelioma cells (Wick et al., 2007), 

led to an inflammation of murine allergic airway with 

augmented humoral immunity (Inoue et al., 2009), and 

induced inflammatory and fibrotic responses when intracheally 

administrated to rats (Muller et al., 2005). More ROS 

were observed upon exposure of human lung epithelial or 

primary bronchial epithelial cells to SWCNTs dispersed using 

dipalmitoylphosphatidylcholine, a major component of lung 

surfactant (Herzog et al., 2009), as compared to dipalmitoylphosphatidylcholine-free 

samples. Low toxicity was observed 

in in vivo experiments when SWCNTs were dispersed using 

Tween 80 and intravenously injected into mice (Yang et al., 

2008). In another study, SWCNTs dispersed in 1% SDS aqueous 

solution showed no cytotoxicity with respect to the human 

alveolar epithelial cells (Wörle-Knirsch et al., 2006). CNTs 

dispersed using Pluronic PF-127 solution did not affect 

viability, apoptosis and ROS generation in the human neuroblastoma 

cells after 3 days of incubation; however, cell 

proliferation decreased as incubation time increased to 2 

weeks (Vittorio et al., 2009). In the only paper that mentioned 

the potential toxicity effect of the dispersant, the authors 

suggested that excess surfactant was responsible for the 

observed increase in toxicity; in this work, the controlled 

exposure of cells to 1% Pluronic F108 produced a 10% decrease 

in the cells viability (Sayes et al., 2006). 

Very little is known about the effect of non-covalent wrapping 

with dispersants other than surfactants on the biocompatibility of 

CNTs. In vitro cytotoxicity of CNTs wrapped with poly(methyl 

vinylketone) decorated with a-N-acetylgalactosamine (Chen 

et al., 2006), nano-1 peptide (Chin et al., 2007) and cholesterolend-capped 

poly(2-methacryloyloxyethyl phosphorylcholine) 

(Xu et al., 2008) were examined after their contact with human 

lung epithelial-like cells, normal skin fibroblasts and human 

umbilical vein endothelial cell line. No impact on cell growth 

and proliferation was demonstrated in all three cases. Cytotoxicity 

of GA-stabilized MWCNTs upon exposure to A549 cells 

was observed by LDH, XTT and MTT bioassays in (Simon- 

Deckers et al., 2008). Two possible reasons were hypothesized 

to be responsible for this effect: (i) increased availability of GAstabilized 

MWCNTs and (ii) different intercellular accumulation 

pathway of GA-stabilized MWCNTs as compared to when 

carbon nanotubes were exposed in bundles. 

1.3.2. Toxicity of CNTs towards eukaryotic cells: genotoxicity 

and epigenetic toxicity 

Most toxicological studies of CNTs focused on the evaluation of 

cytotoxicity (Sayes et al., 2006; Elias et al., 2007; Gutierrez et al., 

2007; Kang et al., 2007; Wick et al., 2007; Kang et al., 2008a,b; 

Yang et al., 2008; Bardi et al., 2009; Herzog et al., 2009; Inoue et al., 

2009; Kang et al., 2009; Vittorio et al., 2009; Yang et al., 2009). 

However, genotoxicity (Kang et al., 2008a; Di Sotto et al., 2009; 

Lindberg et al., 2009; Wirnitzer et al., 2009; Yang et al., 2009)and

508 

epigenetic toxicity (see SD, Section S.1.3) (Upham et al., 1994); 

(Trosko et al., 1998) are other possible causes of cell damage or 

other adverse effects. The genotoxicity of SWCNTs dispersed 

using fetal bovine serum at (5–10) mg/mL concentration, with 

respect to primary mouse embryo fibroblasts was demonstrated 

using Comet assay (Yang et al., 2009). In another recent 

study, CNTs (dispersed in BEGM cell culture medium and subjected 

to ultrasonication) induced a dose-dependent increase in 

DNA damage as indicated by Comet assay and caused a significant 

increase in micronucleated cells (micronucleus assay) in 

human bronchial epithelial cells (Lindberg et al., 2009). Kang 

et al. observed high levels of stress-related gene products in 

Escherichia coli upon its exposure to CNTs, with the quantity and 

magnitude of expression being much higher in the presence of 

SWCNTs (Kang et al., 2008a); in this study CNTs were either 

deposited on a PVDF membrane surface or dispersed in saline 

solution. No mutagenic effect was observed in Salmonella 

microcosme test with baytubes Ò (high purity MWCNTs 

agglomerates, sonicated 10 min in deionized water) (Wirnitzer 

et al., 2009) and in bacterial reverse mutation assay (Ames test) 

with Salmonella typhimurium TA 98 and TA 100 strains and 

with E. coli WP2uvrA strain exposed to MWCNTs dispersed in 

DMSO (Di Sotto et al., 2009). There have been no reports to date on the 

epigenetic toxicity of carbon nanomaterials. 

1.3.3. Toxicity of CNTs in prokaryotic cells 

Most toxicity studies have focused on the effect of CNTs on 

mammalian cell lines. Only limited information exists on the 

cytotoxic effects of CNTs towards bacterial cells. Recently, 

antimicrobial activity of SWCNTs (Kang et al., 2007, 2008a, 

2009) and MWCNTs (Kang et al., 2008a,b, 2009) either suspended 

in aqueous solution (Kang et al., 2007, 2008a) or 

deposited on the surface of a PVDF microfilter (Kang et al., 

2007, 2008a,b, 2009) towards gram-negative bacteria (E. coli 

and Pseudomonas aeruginosa (Kang et al., 2007, 2008a,b, 2009) 

and gram-positive (Staphylococcus epidermidis and Bacillus subtilis 

(Kang et al., 2009)) bacteria was reported. It was suggested 

that membrane damage to the cells was caused by the direct 

physical contact between the CNT and cell (Kang et al., 2007) 

or by a combination of direct physical contact and oxidative 

stress (Kang et al., 2008a). No effect on the percentage of E. coli 

inactivation was observed upon exposure of SRNOM-stabilized 

SWCNTs as compared of SWCNTs dispersed in the 

absence of SRNOM (Kang et al., 2009). Significant antimicrobial 

activity of CNTs composite films against Staphylococcus aureus 

and Staphylococcus warneri was reported in a separated study 

(Narayan et al., 2005). In contrast to the findings of above 

studies, no inhibition of E. coli growth and proliferation was 

reported in a study where microchannel-structured MWCNTs 

scaffolds were immersed into the culture medium with the 

cells (Gutierrez et al., 2007). 

1.4. Effect of growth media on the toxicity of 

nanoparticles 

The existing literature highlights the importance of the 

growth media in cytotoxicity testing and links the apparent 

cytotoxic effect of the nanomaterial to the salt and organic 

content of the culture and related physicochemical characteristics 

of the nanomaterial (Lyon et al., 2005; Tong et al., 


2007). Lyon et al. showed that in media with high salt 

content, the size of nC60 aggregates tends to increase as 

compared to that observed in low salt media (Lyon et al., 

2005). In another study, MWCNTs were found to stimulate 

growth of unicellular protozoan Tetrahymena pyriformis in 

growth medium, which contained proteose peptone, yeast 

extract, and glucose; the increase in growth was attributed to 

the formation of peptone-MWCNTs conjugates, which were 

taken up by the microorganism (Zhu et al., 2006). While the 

aforementioned studies indicate that salt and organic 

composition of the medium, in which exposure studies are 

performed, may influence the interaction of CNTs with 

bacteria, there have been no studies that comparatively evaluated 

CNT toxicity in growth media with different organic loadings and 

salt contents. 

1.5. Objectives of this study 

This study addressed some of the knowledge gaps identified 

above. Aqueous suspensions of SWCNTs functionalized by 

a range of non-covalently bound dispersants of natural (NOM, 

GA, amylose) and synthetic (PVP, Triton X-100) origin, were 

prepared and evaluated in terms of their physicochemical and 

toxicity properties. The study pursued the following 

objectives: 

(i) To evaluate the long-term stability and its physicochemical 

determinants for non-covalently functionalized SWCNTs as 

a function of dispersant type. The evolution of the concentration, 

size, and charge of suspended SWCNTs was 

studied over the period of 28 days. 

(ii) To assess time-dependent cytotoxicity of non-covalently functionalized 

SWCNTs with respect to bacteria and mammalian 

cells as a function of dispersant type and growth media. The 

effect of SWCNTs on E. coli was studied by measuring cell 

viability after 3 h, 24 h, and 48 h of exposure in three 

types of growth media with different organic loadings 

and salt contents. The cytotoxicity of dispersed SWCNTs 

towards mammalian cells was assessed in NDU bioassay 

with rat liver epithelial cells after 30 min and 24 h of 

incubation. 

(iii) To assess time-dependent epigenetic toxicity of non-covalently 

functionalized SWCNTs as a function of dispersant type. In this 

study, we evaluated epigenetic toxicity of the SWCNTs to 

rat liver epithelial cells as a function of dispersant type in 

GJIC bioassay after 30 min and 24 h of incubation. 

2. Approach 

The set of chemically diverse dispersants was chosen based 

on their demonstrated effectiveness in solubilizing CNTs 

(Bandyopadhyaya et al., 2002; O’Connell et al., 2002; Star et al., 

2002; Islam et al., 2003; Moore et al., 2003; Hyung et al., 2007; 

Liu et al., 2007) and potential adverse effects (Burnette, 1960; 

Chourasia and Jain, 2004; Dayeh et al., 2004; Schmitt et al., 

2008). NOM, GA and PVP have LD50 doses of (54.8–58.5) mg 

(intravenous administration in mice) (EMEA, 1999), 2,000 mg 

(oral administration in rats) (Schmitt et al., 2008), and 

100,000 mg (oral administration in rats) (Burnette, 1960) per kg

of body weight, respectively. Amylose has been reported to be 

used in a colon-specific drug delivery due to its low toxicity 

and high biodegradability (Chourasia and Jain, 2004). Triton 

X-100 was reported to be toxic to protozoa, fish, and 

mammalian cells (Dayeh et al., 2004). 

Three batches of dispersed SWCNTs were prepared. The 

first batch was used to comprehensively evaluate the longterm 

stability of SWCNT suspensions over a period of 4 weeks 

in terms of concentration, effective hydrodynamic diameter, 

and z-potential of stabilized SWCNTs. 

The second batch was used to evaluate the viability of E. 

coli cells after their contact with dispersed SWCNTs by the 

quantification of colony forming units. To elucidate the 

effects of ionic and organic composition of the growth 

medium, the cytotoxicity of the SWCNTs suspended in three 

growth media of different organic and salt compositions 

was evaluated. First, in order to assess the acute cytotoxicity 

of dispersed SWCNTs, we conducted experiments in 

0.1 M NaCl. Then, in order to investigate the effect of 

SWCNTs on the ability of E. coli form colonies over time, we 

employed three types of growth media: (i) LB medium with 

higher organic chemical and salt content, (ii) MD medium 

with low salt content and organic load, and (iii) 0.1 M NaCl 

solution. 

The third batch was used to evaluate cyto- and epigenetic 

toxicity of the dispersed SWCNTs against rat liver epithelial 

cells by the (i) NDU and (ii) GJIC bioassays. The NDU assesses 

cell viability by measuring the accumulation of neutral red 

dye in lysosomes, which depends on the cell’s capacity to 

maintain pH gradients through the maintaining membrane 

integrity and production of adenosine triphosphate. The 

amount of dye incorporated by the cell is quantified spectrophotometrically 

(Borenfreund and Puerner, 1985). The NDU 

bioassay has been used to assess cytotoxicity of CNTs (Flahaut 

et al., 2006). The GJIC bioassay (Borenfreund and Puerner, 

1985; Weis et al., 1998; Herner et al., 2001; Satoh et al., 2003) 

utilizes the ability of epigenetic tumor promoters to alter level 

of GJIC (Yamasaki, 1990; Trosko et al., 1991). The degree of GJIC 

was quantified by measuring the distance (area) the fluorescent 

dye ( 1200 Da) travels (occupies) between the cells after 

a given time. Both NDU and GJIC bioassays were carried out 

with WB-F344 rat liver epithelial cells exposed to dispersed 

SWCNTs for different periods of time. The WB-F344 cell line 

was chosen because these normal diploid rat liver epithelial 

cells have already used in numerous studies of cytotoxic or 

epigenetic effects (e.g., Herner et al., 2001) thus allowing us to 

have a comparative basis. 


3.1. CNTs 

SWCNTs (purity > 90%), produced by catalytic chemical vapor 

deposition, were obtained from Cheap Tubes, Inc (Brattleboro, 

VT) and used as received. The SWCNTs were used as obtained, 

allowing us to mimic what might occur in the environment. 

As indicated by the manufacturer, the SWCNTs had an inside 

diameter in the range of 0.8 nm to 1.6 nm, an outer diameter in 

the range of 1 nm to 2 nm and were 5 mm to30mm in length. 


3.2. Non-covalent functionalization: dispersants and 

dispersion procedures 

3.2.1. Dispersants 

Gum arabic (approx. 250 kDa; a complex mixture of saccharides 

and glycoproteins obtained from the acacia tree), PVP 

(approx. 29 kDa), Triton X-100 (approx. 625 Da; polyethylene 

glycol p-(1,1,3,3-tetramethylbutyl)-phenyl ether) and amylose 

(molecular weight not specified; a polymeric form of glucose 

and a constituent of potato starch) were purchased from 

Sigma–Aldrich (Milwaukee, WI). SRNOM reverse osmosis 

isolation was obtained from International Humic Substances 

Society (St Paul, MN). The molecular structure of dispersants 

used in this study is presented in Fig. 1. 

3.2.2. Dispersion procedure 

Aqueous solutions of PVP, Triton X-100 were prepared by 

dissolving PVP (40 mg) and Triton X-100 (0.4 mL) in 40 mL of 

water and adjusting the pH of both solutions to 7 with 0.1 M 

HCl. The aqueous solution of GA was prepared by mixing 1 g of 

GA in 50 mL of water and adjusting the pH of the mixture to 7 

with 0.1 M HCl; this mixture was left to settle for 24 h and then 

40 mL of supernatant was collected for use in the stability or 

toxicity studies. To prepare NOM solutions, two flasks of 

40 mL of water, each containing 8 mg of NOM, were stirred for 

24 h. Each solution was then filtered through a 0.22 mm filter 

under vacuum. The pH of one solution was kept at its original 

value (approx. 3.5) while the pH in another solution was 

adjusted to 7 with 0.1 M NaOH. SWCNTs were added to the 

above solutions to result in a 1 mg (SWCNT)/mL loading and 

were sonicated in Aquasonic 50 T water bath (VWR Scientific 

Products Corp, West Chester, PA) for 20 min. 

SWCNTs were dispersed with amylose based on modified 

version of the three-step approach reported by Kim et al. (Kim 

et al., 2004). SWCNTs (40 mg) were sonicated in 25 mL of water 

(pH 7) at approximately 75 W for 5 min using Sonicator 3000 

(Misonix, Inc., Farmingdale, NY) equipped with a microprobe. 

100 mg of amylose in 6.28 mL of DMSO were prepared and 

added to the sonicated suspension of SWCNT in water so that 

the DMSO/water ratio was 20% by volume. The resulting 

mixture was sonicated for another 5 min in order to remove 

the excess amylose and DMSO. The suspension was sonicated, 

centrifuged and refilled with water four times (See SD, 

Section S.2.3). 

Following the sonication, SWCNTs suspensions were 

divided into 12 mL aliquots and each aliquot was transferred 

into 15 mL centrifuge tube. After 24 h standing time, the top 

4 mL of each suspension was withdrawn in order to be used in 

the stability or toxicity studies. Overall, three batches of 

dispersed SWCNT suspensions were prepared and were used, 

correspondingly, for 1) the study of the long-term stability of 

SWCNTs suspensions, 2) the general viability bioassay with 

E. coli, and 3) NDU and GJIC bioassays. 

3.3. Physicochemical characterization of SWCNT 

suspensions 

In order to quantitatively assess the long-term stability 

SWCNT suspensions, a combination of characterization 

methods was employed.

510 

a 

GALP 

1 

3 

ARAF 

1 

3 

3 

ARAF 

ARAF 

1 

1 

3 6 

1 

GALP 

3 1 

GALP 

3 6 1 

GALP 

3 1 

GALP 

6 

6 

3 1 

3 1 

RHAP GALP 

RHAP GALP 

6 6 

1 

1 

GA 

GA 

4 4 

1 

ARAF 

GALP 

1 

3 

ARAF 

1 

1 

ARAF 

GALP = D-GALACTOPYRANOSE ARAF = L-ARABOFURANOSE 

GA = D-GLUCURONIC ACID RHAP = L-RHAMNOPYRANOSE 

b 

d 

C 8 H 17 

3.3.1. UV–vis spectrophotometry. Quantification of SWCNTs 

concentration in suspension 

The absorbance of SWCNT suspensions over the (200–800) nm 

wavelength range was measured over a period of 4 weeks 

(Multi-spec 1501, Shimadzu, Kyoto, Japan). For each type of 

SWCNT suspension prepared, the absorbance of SWCNT-free 

solution of the corresponding dispersant was used as 

a baseline. 

SWCNT suspensions used in our toxicity studies likely 

contained both individual and bundled SWCNTs. In fact, DLS 

measurements indirectly confirmed (see Section 4.2) the 

presence of dispersed SWCNT bundles and direct TEM 

imaging (see SD, Section S.3.3, Fig. S5) also showed the presence 

of SWCNTs bundles (note that TEM results need to be 

interpreted with caution as evaporation-induced aggregation 

could have contributed to bundling during TEM sample 

preparation). Incomplete exfoliation is the most likely and 

most environmentally relevant scenario; unfortunately, it also 

N 

O 

n 

O 


OH 

n 

OH 

OH 

H 2 COH 

HO 

O 

N 

H 2 

(a-1) 

CH 2 OH 

OH 

O 

(a-3) 

O 

OH 

O OH 

HOOC 

HO 

OH 

HO 

COOH 

COH 

OH 

(a-2) 

CH 3 

OH OH 

HOOC 

(a-4) 

entails difficulties with quantifying the total concentration of 

suspended nanotubes as larger CNT bundles tend to separate 

from the suspension. 

In this study, we used UV–vis spectrophotometry to estimate 

the concentration of exfoliated SWCNTs in suspensions. 

It is known that only fully exfoliated SWCNTs absorb in the 

(200–1200) nm wavelength range. Bundled NTs do not absorb 

significantly in this range due to the tunnel coupling between 

metallic and semiconductive SWNTs (Lauret et al., 2004; 

Grossiord et al., 2007). Therefore, the UV absorbance by 

a SWCNT suspension can be used to selectively measure the 

concentration of individually dispersed (i.e. fully exfoliated or 

debundled) SWCNTs. 

To determine the concentration of suspended exfoliated 

SWCNTs, we first prepared a separate set of suspensions with 

ten-fold reduced SWCNT loading with respect to the CNT 

loading used in long-term stability and toxicity studies. Four 

suspensions (with PVP, NOM pH 3.5, NOM pH 7, and GA used 

O 

O 

OH 

OH 

OH OH 

O O 

OH 

O 

HO 

OH 

O O 

Fig. 1 – Molecular structure of solubilizers: (a) Gum arabic (a-1) galactose, (a-2) rhamnose, (a-3) arabinose, (a-4) glucuonic 

acid; (b) Poly(vinyl pyrrolidone); (c) A building block of humic acids, which has a compositional similarity to SRNOM; Dots 

represent chiral centers; (d) Triton X-100; (e) Amylose. 

c 

e 

n 

OH 

COOH 

OH 

OH

as dispersants) with 10-fold reduced SWCNT content (0.1 g/L) 

were prepared using the same procedures as described in 

Section 3.2 except that they were subjected to rigorous, prolonged 

sonication for 1 h at power of (70–80) W using horn 

sonicator (Sonicator 3000, Misonix, Inc., Farmingdale, NY). No 

settling of SWCNTs was observed over the short term 

following the application of this treatment indicating 

complete dispersion of SWCNTs. UV–vis absorbance spectra 

were recorded at different dilutions and a calibration curve for 

absorption at 500 nm (Bahr et al., 2001; Huang et al., 2002; 

Sinani et al., 2005; Lee et al., 2007; Salzmann et al., 2007) as 

a function of concentration was constructed, and coefficient 

of molecular extinction, 3, was determined. This coefficient 

was used to estimate the concentration of exfoliated SWCNTs 

in suspensions used in long-term stability and toxicity 

experiments. Note that by subjecting suspended SWCNTs to 

a very intense sonication treatment we aimed at maximizing 

the extent of exfoliation; however, it was not possible to 

ascertain that the exfoliation was complete. Thus coefficients 

of molecular extinction and exfoliated SWCNT concentrations 

determined using the recorded calibration curves were estimated 

values. SWCNT characteristics measured upon the 

preparation of CNT suspensions are given in Table S1 in SD. 

3.3.2. Size and charge of dispersed SWCNTs 

The effective hydrodynamic diameter and z-potential of suspended 

SWCNTs were determined after 1, 4, 7, 14, 21, and 28 

days of settling. The size and charge were measured by 

dynamic light scattering (DLS) and phase analysis light scattering 

techniques, respectively (ZetaPALS, BI_MAS Option, 

Brookhaven Instrument Corp., Holtsville, NY). The Smoluchowski 

equation was applied to convert the measured 

electrophoretic mobility of dispersed SWCNT to z-potential. 

Transmission electron microscopy (TEM) imaging was used as 

an auxiliary method to aid in the interpretation of dynamic 

light scattering data (see SD, Sections S.2.5 and S.3.4). 

3.4. Toxicity assessment: E. coli viability assay 

3.4.1. Media preparation 

Luria–Bertani (LB) growth medium was prepared according to 

the standard procedure (Atlas, 1993). Minimal Davis medium 

with 90% reduced potassium phosphate concentration 

(MD medium) was prepared (Fortner et al., 2005). 0.1 M NaCl 

solution was prepared by dissolving 9 g of NaCl in 1 L of water 

(pH 7) and autoclaving it for 15 min at 1 bar and 121 C. For the 

toxicity assessments, each component of LB, MD and NaCl 

media was prepared in 4-fold higher concentration as 

compared to original protocol and the aliquot part of the corresponding 

medium was added to SWCNTs suspension 

(as further described in ‘‘Quantification of cell viability’’ 

Section). Luria–Bertani Petri plates were prepared according to 

the published method (Atlas, 1993). 

3.4.2. Preparation of E. coli culture 

E. coli K12 stock was prepared in glycerol and stored at 80 C. 

Prior to use, the stock was defrosted, and 30 mL of LB medium 

were inoculated with 5 mL of the stock. After overnight growth 

at 37 C, 5 mL of this suspension was spread onto LB agar plate 

and cultured at 37 C. Once distinct colonies were formed, the 


agar plate was transferred to the refrigerator and kept at 4 C 

for up to one month. E. coli suspensions to be used in SWCNT 

cytotoxicity studies were prepared by scraping one colony 

from the surface of a Petri plate by aseptic loop and immersing 

the loop into 10 mL of LB or MD media in a 50 mL centrifuge 

tube. Tubes were placed on a shaker in an incubator 37 C for 

12 h. When 0.1 M NaCl was used as an exposure medium in 

colony forming units bioassay, E. coli were first grown in the LB 

medium, centrifuged for 5 min at 2250 rpm and washed with 

0.1 M NaCl as follows: the supernatant was decanted and 

replaced with an equal volume of 0.1 M NaCl, vortexed and 

resuspended by centrifugation. The washing procedure was 

repeated twice, presuming that after this treatment most of 

the remaining organic constituents of LB medium were 

removed from the E. coli suspension. 

3.4.3. Quantification of cell viability 

The SWCNTs suspension (1.425 mL), growth medium (475 mL) 

and 100 mL of the stock E. coli suspension were transferred into 

a 15 mL centrifuge tube and incubated under gentle shaking at 

37 C. Samples were taken after 3, 24, and 48 h and a series of 

dilutions (10 4 –10 6 ) was prepared for each sample. Five samples 

of 10 mL and 20 mL from each dilution were placed onto an agar 

plate and incubated at 37 C until distinct colonies developed. 

Colony forming units (CFU)/1 mL were calculated for each 

sample. Each experiment was run in triplicates with negative 

(bacterial suspension with the corresponding amount of 

ultrapure water) and vehicle (solution of the corresponding 

dispersant) controls, herein called vehicle control I and 

vehicle control II, respectively. The results are reported as 

a fraction of control (FOC) standard deviation (STD), calculated 

as the ratio of the average number of colonies grown 

after E. coli exposure to SWCNTs suspensions or vehicle 

control II to the average number of colonies grown in vehicle 

control I plates. 

3.5. Toxicity assessment: neutral red dye uptake 

bioassay 

For the NDU bioassay we adapted a published procedure 

(Borenfreund and Puerner, 1985; Weis et al., 1998; Satoh et al., 

2003). A solution (0.015 w/v) of neutral red dye (3-amino-7- 

(dimethylamino)-2-methylphenazine hydrochloride) in 

D-medium was incubated at 37 C for 2 h and filtered through 

a 0.22 mm syringe filter (Millipore Corp., New Bedford, MA) to 

remove undissolved dye and ensure sterile conditions. 

Confluent WB-F344 cells (see SD, Section S.2.6 for details on the 

preparation of the cells) were exposed to 500 mL of each of 

SWCNTs suspension and incubated at 37 C for 30 min and 24 h 

under gentle shaking in a humidified atmosphere containing 

5% CO 2. After cells were exposed to SWCNTs, the exposure 

medium was removed by aspiration and the cells were washed 

with 1 mL of phosphate saline buffer (PBS). Following washing, 

2 mL of the neutral red dye solution per plate was added and 

the cells were incubated for 1 h at 37 C in the humidified 

atmosphere containing 5% CO2. Upon incubation, the cells 

were rinsed three times with PBS, and 2 mL of aqueous solution 

containing 1% acetic acid and 50% ethanol was added to each 

plate to lyse the cells. 1.5 mL of the lysate was transported into 

2 mL microcentrifuge tube and the optical density was

512 

recorded at 540 nm using a Beckman DU 7400 diode array 

detector (Beckman Instruments, Inc., Schaumburg, IL). The 

background absorbance was measured at 690 nm and then 

subtracted from the original absorbance. Each experiment was 

conducted in triplicates. The neutral red dye uptake was 

reported as the FOC (absorption of neutral red in the chemically 

treated sample divided by the absorption of neutral red in the 

nontreated control I sample). FOC values of 1.0 indicates noncytotoxic 

response while FOC values

a b 

Concentration (mg/L) 

c 

-potential (mV) 

200 

160 

120 

80 

40 

0 

0 

0 7 14 21 28 0 7 14 21 28 

Time (d) 

Time (d) 

0 

-15 

-30 

-45 

Triton X-100 stabilizes nanotubes by the formation of hemimicelles 

that cover nanotube surface with benzene rings 

providing p–p stacking between the surfactant molecule and 

nanotube core (Islam et al., 2003). The exfoliation (debundling) 

of SWCNTs in the presence of SRNOM was suggested to occur 

through the interaction of the aromatic moieties of natural 

organic matter and nanotube surface (Hyung et al., 2007; Liu 

et al., 2007; Hyung and Kim, 2008). Despite the differences in 

the physiosorption mechanisms responsible for the stabilization 

of CNTs in aqueous suspensions, all dispersants were 

effective, albeit to somewhat different extents. 

For all SWCNT suspensions, the calculated values of the 

concentration of exfoliated SWCNTs correlated well to the 

visual observations described above. As seen from Fig. 2a, the 

dispersion efficiency in terms of the concentration of 

dispersed SWCNTs was a function of the type of the dispersant, 

with GA producing suspensions having highest 

concentration of SWCNTs. The data were consistent among 

the three prepared batches of SWCNTs (see SD, Table S1). 

Didenko et al. (Didenko et al., 2005) suggested that after 

covering one single carbon nanotube with a long polymeric 

molecule, the remaining strands would react with other 

uncovered or partially covered SWCNTs thus bundling several 

nanotubes together. We speculate that this mechanism where 


Effective diameter (nm) 

-60 

-45 

0 7 14 21 28 0 7 14 21 28 

Time (d) 

Time (d) 

Fig. 2 – The time-dependent characterization of (a) estimated concentration of the exfoliated SWCNTs in the dispersion, 

(b) effective diameter and (c) z-potential of dispersed SWCNTs, and (d) z-potential of dispersants alone (-B- for GA/SWCNTs, 

-,- for PVP/SWCNTs, -6- for NOM3.5/SWCNTs, -:- for NOM7/SWCNTs, --- for Triton X-100/SWCNTs, -C- for Amylose/ 

SWCNTs). Notes: (1) The z-potential measurement was not conducted for amylose solution, because amylose is not watersoluble 

under room temperature. (2) Reported are results of only those measurements that were statistically significant, i.e. 

when sufficiently high photon count rates were recorded in dynamic light scattering measurements. (3) The absorption by 

suspensions of amylose-stabilized SWCNTs was not measured because the nanotube content in these suspensions could 

not be precisely determined; this was due to the incomplete separation at the centrifugation step of the suspension 

preparation process. 

d 

-potential (mV) 

1000 

800 

600 

400 

200 

0 

-15 

-30 

one polymer molecule links two or more nanotubes may be 

responsible for the higher dispersing efficiency of GA (by far 

the largest molecule among the target dispersants) towards 

SWCNTs compared to other dispersants. This hypothesis is 

supported by the measurements of the effective hydrodynamic 

diameter (Fig. 2b), and the estimation of the length of 

suspended particles with the size and the length of GA/ 

SWCNTs being higher than those of PVP, Triton X-100 and 

SRNOM (both pHs). As the measured diameter of GA-stabilized 

SWCNTs aggregates (which could also be expected to be 

rather porous) was still relatively small, gravity settling did 

not lead to a significant removal of GA-stabilized SWCNTs 

from the dispersion. 

When comparing the solubilizing ability of SRNOM at 

different pH, one could see that SRNOM is a more efficient 

dispersant at pH 3.5 than at pH 7. This is consistent with the 

findings by Huyng et al. (Hyung and Kim, 2008) who reported 

that adsorption of SRNOM to MWCNTs increased when pH 

decreased due to more compact and coiled conformation of 

NOM at acidic pHs. When pH increases, carboxylic and 

phenolic groups of SRNOM deprotonate resulting in higher 

electrostatic repulsion between a CNT and a SRNOM molecule 

and, as a consequence, in a lower amount of organic matter 

adsorbed on the CNT surface. The better dispersion of

514 

SWCNTs at pH 3.5 could also be attributed, in part, to the steric 

hindrance imposed by SRNOM when a higher surface density 

of NOM on the surface of CNT bundles results in stronger 

repulsion between CNTs. This observation is supported by the 

long-term z-potential measurements (Fig. 2c), where the 

surface charge of stabilized SRNOM/SWCNTs at pH 3.5 

became less negative up to day 7, and then stabilized with 

increasing settling time. At the same time the surface charge 

of SWCNTs dispersed in SRNOM solution at pH 7 gradually 

increased after day 5. Indeed, had the electrostatic repulsion 

been the sole mechanism of SWCNTs stabilization in SRNOM 

solutions, the surface charge on the SWCNTs would have 

more rapidly become less negative at pH 3.5 than at pH 7. In 

summary, in terms of the effectiveness of SWCNT stabilization, 

the dispersants were ranked as follows: GA > Triton 

X-100 > PVP > NOM (pH 3.5) > NOM (pH 7). 

4.2. Hydrodynamic size of dispersed SWCNTs 

The hydrodynamic size of CNTs in a suspension is an 

important characteristic that affects the stability and, 

possibly, toxicity of dispersed CNTs. It should be noted that 

the effective hydrodynamic diameter of suspended SWCNTs 

measured using DLS can only be used as a very rough estimate. 

While the DLS data are interpreted with the assumption 

that the primary scatterers are spherical with an aspect ratio 

of 1, dispersed SWCNTs are long and tubular, with a very high 

aspect ratio (see SD, Fig. S3). To obtain a better approximation 

of the size distribution of dispersed SWCNT, the multimodal 

size distribution model was used. TEM imaging was employed 

to obtain auxiliary information on the size and morphology of 

stabilized SWCNT. Even though the measurement of effective 

hydrodynamic diameter cannot be relied on to compare the 

sizes of suspended SWCNTs in different dispersion media, 

these measurements can be used to compare how the average 

particle size for a given dispersant changes with settling time. 

The values of effective diameter as a function of time for 

different SWCNT suspensions are given in Fig. 2b. 

Generally, the effective size of SWCNTs dispersed using 

Triton X-100, GA and NOM did not change significantly with 

increasing settling time (Fig. 2b). In the case of amylosedispersed 

SWCNTs, a slight decrease in the effective size was 

observed due to settling of larger and unstable SWCNT 

aggregates, which left behind more uniformly sized smaller 

amylose-stabilized CNT clusters. The concentration of 

SWCNTs in the suspension was negatively correlated with the 

effective size of SWCNT. The notable apparent exception was 

the suspension of SWCNTs dispersed using GA. However, it 

should be noted that the aqueous suspension of GA 

contained GA colloids of the size that was 1) comparable to the 

size of dispersed SWCNTs and 2) decreased during the 4 weeks 

of settling (see SD, Fig S5). Thus, the effective size measurements 

for SWCNTs dispersed using GA should be interpreted 

with caution. 

4.3. Charge of dispersed SWCNTs 

For all suspensions, the z-potential of dispersed SWCNTs was 

negative and nearly constant over the entire duration of the 

experiment (Fig. 2c and SD, Table S1). Each z-potential 


a 

Fraction of control (FOC) 

b 


c 


1.2 

1 

0.8 

0.6 

0.4 

0.2 

0 

1.2 

1 

0.8 

0.6 

0.4 

0.2 

0 

1.2 

1 

0.8 

0.6 

0.4 

0.2 

0 

NOM 3.5 NOM 7 GA Amylose PVP Triton 

NOM 3.5 NOM 7 GA Amylose PVP Triton 

NOM 3.5 

NOM 7 GA Amylose PVP Triton 

Fig. 3 – Viability of E. coli in (a) 0.1 M NaCl, (b) LB medium, 

and (c) MD medium. Exposure time: (a) 3 h, (b) 48 h, (c) 48 h. 

Open, hatched and cross-hatched bars correspond to 

control I, control II and dispersed SWCNTs. In the case of 

amylose ultrapure water was used as control. 

measurement for SWCNT suspensions was accompanied by 

a measurement of the z-potential of the dispersants in 

a SWCNT-free aqueous solution (Fig. 2d). After 1 d, only the GA 

solution scattered light sufficiently to enable a statistically 

meaningful z-potential measurement. After 4 d, solutions of 

NOM (pH 7), PVP, and Triton X-100 also scattered light

a 1.2 

b 


1 

0.8 

0.6 

0.4 

0.2 

0 

NOM3.5 NOM7 GA Amylose PVP Triton 

sufficiently. After 7 d, all the dispersants except for NOM (pH 

3.5) achieved acceptable count rates during z-potential 

measurements. The gradual change in the scattering ability of 

the dispersant solutions may correspond to the aggregation of 

the dispersant molecules. The dispersants were ranked in 

order from the most to the least negative z-potential of 

dispersed SWCNTs: NOM (pH 7) > NOM (pH 3.5) > GA z Triton 

X-100 > PVP z amylose. The highly negative charge of NOMstabilized 

SWCNTs was likely due to the charge of NOM. The 

low stability of amylose-dispersed SWCNTs could be attributed 

to the combination of low surface charge and steric 

repulsion between stabilized SWCNTs in the dispersion. 

4.4. Assessment of cell toxicity in prokaryotic and 

eukaryotic systems 

4.4.1. General viability bioassay with E. coli (prokaryotic) 

Immediately after E. coli cells were exposed to SWCNTs suspended 

in growth medium as well as after 3 h of exposure, no 

SWCNT aggregation was visually observed (see SD, Table S2) 

in experiments with all three types of the media – 0.1 M NaCl, 

MD (medium with lower salt organics content) and LB 

(medium with higher salt and organics content). After 24 h of 

incubation, limited precipitation was observed for all SWCNT 



1.2 

1 

0.8 

0.6 

0.4 

0.2 

0 

NOM3.5 NOM7 GA Amylose PVP Triton 

Fig. 4 – Neutral red dye uptake by WB-F344 cells as a function of dispersant type (a – 30 min, b – 24 h of exposure). Open, 

hatched and cross-hatched bars correspond to control I, control II and dispersed SWCNTs. In the case of amylose control I 

and control II were ultrapure water. 

suspensions in LB and MD media. After 48 h of incubation 

more precipitation occurred in each type of media. 

No inhibition of E. coli colony forming ability was observed 

after 3 h of incubation with amylose-, NOM-, GA- and PVPstabilized 

SWCNTs in 0.1 M NaCl (Fig. 3a). The FOC values did 

not decline in any of these samples and no significant difference 

in the influence on CFU counts was found between 

dispersed SWCNTs and solutions of corresponding dispersants. 

When the bacteria suspension was brought in contact 

with Triton X-100-stabilized SWCNTs, approx. 25% loss of the 

cell viability was measured as compared to vehicle control I 

samples. However, there was no statistical difference between 

FOC of Triton-stabilized SWCNTs and the control (solution of 

Triton X-100 only). It remains unclear if the observed cytotoxicity 

of the these suspensions was due to i) SWCNTs 

stabilized by Triton X-100 or ii) the residual ‘‘free’’ (i.e. not 

associated with suspended SWCNTs) Triton X-100 potentially 

present in the solution or iii) the combined effect of both 

Triton X-100/SWCNT and dissolved Triton X-100. The 

complete separation of the Triton X-100/SWCNT and 

dissolved Triton X-100 could not be accomplished using 

centrifugation. Even for very long centrifugation times, the 

supernatant had grayish color indicating that some fraction of 

SWCNTs was not removed. 

Fig. 5 – Representative phase contrast images of WB-F344 cells incubated with 500 mLofH 2O (control I), 500 mL of Triton X-100 

(control II) and 500 mL of Triton X-100-stabilized SWCNTs. Black dots observed in Control II and Triton X-100-solubilized 

SWCNTs samples correspond to dead WB-F344 cells. All images were taken at 2003 magnification. Scale bar is 50 mm.

516 

Fig. 6 – Representative phase contrast (upper row) and UV epifluorescent (bottom row) images of WB-F344 cells incubated 

with 500 mLofH2O (control I), 500 mL of NOM pH 3.5 (control II) and 500 mL of NOM-stabilized SWCNTs (pH 4). All images were 

taken at 2003 magnification. Bright dots correspond to cells that absorb the dye; the absorption is indicative of cellular 

health. All images were taken at 2003 magnification. Scale bar is 50 mm. 

The ability of E. coli to grow and to form colonies in the 

presence of amylose-, GA-, PVP-, and NOM-stabilized SWCNTs 

in LB medium mimicked both control samples regardless of 

the contact time (see SD, Fig. 3b and Fig. S6). On the contrary, 

21 and 18 % mortality rates after 3 h of incubation were 

observed for E. coli in Triton X-100/SWCNTs suspension and 

Triton X-100 solution, respectively. After 24 h of contact, the 

number of colonies grown on the Petri plates decreased by 

30 % when bacteria were in contact with Triton X-100-stabilized 

SWCNTs and by 27 % when bacteria were in Triton X-100 

only solution (see SD, Fig. S6) as compared to vehicle control I 

a 1.2 

b 


1 

0.8 

0.6 

0.4 

0.2 

0 

NOM 3.5 NOM 7 GA Amylose PVP 


plates. As the exposure time increased to 48 h, E. coli resumed 

its growth; with FOC of both Triton X-100 suspensions 

approaching values for vehicle control I sample (Fig. 3b). When 

dispersed SWCNTs were tested in MD medium, no losses in 

cell viability for amylose-, GA-, PVP- and NOM-stabilized 

SWCNTs were measured (Fig. 3c; also see SD, Fig. S7). In the 

case of Triton X-100 suspensions, the reduction in E. coli 

survival was observed after 3 h of exposure; comparable 

losses of the E. coli viability were also observed for both Triton 

X-100-stabilized SWCNTs and the Triton X-100 only solution. 

However, after 24 h of exposure, fewer E. coli colonies were 


1.2 

1 

0.8 

0.6 

0.4 

0.2 

0 

NOM 3.5 NOM 7 GA Amylose PVP 

Fig. 7 – GJIC in WB-F344 cells as a function of dispersant type (a – 30 min, b – 24 h of exposure). Open, hatched and 

cross-hatched bars correspond to control I, control II and dispersed SWCNTs. In the case of amylose control I and control II 

were ultrapure water.

observed on Petri plates from suspensions of Triton X-100stabilized 

SWCNTs and Triton X-100 solution only (vehicle 

control II) (33 and 44 %, respectively) (see SD, Fig. S7). As the 

incubation time increased to 48 h, bacterial CFU counts in the 

presence of both Triton X-100 samples increased and 

approached that of vehicle control I samples (Fig. 3c). 

As was previously mentioned, the exposure of Triton 

X-100-stabilized SWCNTs to MD medium for 24 h resulted in 

formation of large flake-like aggregates, which settled out of 

the suspension. It is possible that this low salt medium 

induced specific interactions between bacterial cells and 

Triton X-100-coated SWCNTs, which were suppressed in the 

medium having a higher ionic strength. This led to a larger 

decrease in E. coli viability in these samples in comparison 

with the losses in Triton X-100 solution only. Damaged or dead 

cell attached to the SWCNTs aggregates and formed debris at 

the bottom of the testing tubes. The simultaneous effect of 

bacterial and nanoparticles settling out of the solution has 

been previously described (Tang et al., 2007) where the 

response of Shewanella oneidensis to C 60–NH 2 nanoparticles 

was studied. 

The fact that we observed inhibition of E. coli viability after 

24 h of exposure in MD medium and did not observed this 

effect in LB medium highlights the importance of growth 

medium in cytotoxicity testing. Similarly, E. coli exposure to 

SWCNTs, stabilized with amylose, GA, PVP and NOM (both pH) 

did not affect CFU counts in any of the three media pointing to 

the importance of the dispersant in the assessments of 

biocompatibility of CNTs. 

4.4.2. NDU cytotoxicity bioassay on eukaryotic cells 

The assessment of cytotoxicity was determined by the 

viable uptake of neutral red into F344-WB rat liver epithelial 

cells. Nonviable cells lack the ability to absorb this dye. 

Results (Fig. 4) indicated that SWCNTs dispersed using NOM 

(pH 3.5), NOM (pH 7), GA, amylose and PVP were not 

cytotoxic. However, Triton X-100 detergent induced 

a significant cytotoxic effect independent of SWCNTs. This 

latter observation can also be seen under phase contrast 

microscopy of cells treated with Triton X-100 for 30 min 

(Fig. 5). The morphology of the cells drastically changed 

from that of the normal control cells, in which the cells 

became clearer due to loss of most cytosolic contents, 

except for the cell nuclei. After a 24 h treatment with Triton 

X-100, the cells completely detached or solubilized from the 

culture plates (data not shown). 

4.5. Assessment of epigenetic toxicity on eukaryotic cells: 

GJIC bioassay 

A representative fluorescent micrograph of the GJIC assay 

after treatment with (NOM þ SWCNTs) is shown in Fig. 6. 

The migration of the fluorescent dye through several cell 

layers is an indicator that the gap junction channels are 

open. After WB-F344 cells were incubated with dispersed 

SWCNTs suspensions for (1) 30 min and (2) 24 h, there was 

no significant affect on GJIC for all suspensions tested 

(Fig. 7). Due to Triton X-100-stabilized SWCNTs being 

cytotoxic, GJIC was not measured in cells exposed to this 

mixture. Also, when cells were examined under phase 


contrast, no changes in size and shape of the individual 

cells were detected (Fig. 6). Hence, it can be concluded that 

under these experimental conditions WB-F344 cells exposed 

to dispersed SWCNTs retain normal intercellular communication 

irrespective of the applied treatment. 


The stability of aqueous suspensions of SWCNTs non-covalently 

functionalized by a range of natural (gum arabic, 

amylose, Suwannee River natural organic matter) and 

synthetic (polyvinyl pyrrolidone, Triton X-100) dispersants 

that bind to SWCNT surface via different physiosorption 

mechanisms was evaluated. Despite the differences, all 

SWCNT suspensions remained relatively stable over a period 

of 4 weeks as indicated by the evolution of the concentration 

of suspended exfoliated SWCNT and by the fact that both the 

average effective hydrodynamic diameter and z-potential of 

suspended SWCNTs were relatively constant. The epigenetic 

toxicity of dispersed SWCNTs was evaluated for the first time 

using a gap junctional intercellular communication assay 

with WB-F344 rat liver epithelial cells resulting in no inhibition 

by SWCNTs non-covalently functionalized using GA, PVP, 

amylose, and SRNOM. There were no cytotoxic effects of these 

SWCNTs suspensions on either the prokaryotic, bacterial (E. 

coli), or eukaryotic (WB-F344 rat liver epithelial) cell types. 

Only when the dispersant itself was toxic, were losses of cell 

viability observed. Bacterial CFU counts or neutral red 

uptake was affected only by SWCNTs dispersed using Triton 

X-100, which showed cytotoxicity in the SWCNT-free 

solution (vehicle control II). The results suggest a strong 

dependence of the toxicity of SWCNT suspensions on the 

toxicity of the solubilizing agent and point to the potential of 

non-covalent functionalization with non-toxic dispersants as 

a method for the preparation of aqueous suspensions of 

biocompatible CNTs. 


This work was supported by the National Science Foundation 

research grants CBET-0608320, CBET-056828 and OISE- 

0530174, by the National Institute of Environmental Health 

Sciences grants R01 ES013268-01A2 and Grant-in-Aid for 

Science Research. We thank Michael Housinger for his assistance 

with the preparation of SWCNT suspensions. ARR 

acknowledges the support by Michigan AWWA Fellowship for 

Water Quality and Treatment Study. The contents of this work 

are solely the responsibility of the authors and do not necessarily 

represent the official views of the National Institute of 

Environmental Health Sciences. 

Appendix. 



in the online version, at doi: doi:10.1016/j.watres.2009.09.042.

518 

references 

Atlas, R.M., 1993. Handbook of Microbiological Media. CRC Press, 

Boka Raton. 

Bahr, J.L., Mickelson, E.D., Bronikovski, M.J., Smalley, R.E., Tour, J.M., 

2001. Dissolution of small diameter single-wall carbon 

nanotubes in organic solvents. Chemical Communications, 

193–194. 

Bandyopadhyaya, R., Nativ-Roth, E., Regev, O., Yerushalmi- 

Rozen, R., 2002. Stabilization of individual carbon nanotubes 

in aqueous solutions. Nano Letters 2 (1), 25–28. 

Bardi, G., Tognini, P., Ciofani, G., Raffa, V., Costa, M., 

Pizzorusso, T., 2009. Pluronic-coated carbon nanotubes do 

not induce degeneration of cortical neurons in vivo and in 

vitro. Nanomedicine: Nanotechnology, Biology, and Medicine 

5, 96–104. 

Bonnet, P., Albertini, D., Bizot, H., Bernard, A., Chauvet, O., 2007. 

Amylose/SWNT composites: from solution to film – synthesis, 

characterization and properties. Composite Science and 

Technology 67, 817–821. 

Borenfreund, E., Puerner, J.A., 1985. Toxicology determined in 

vitro by morphological alterations and neutral red absorption. 

Toxicology Letters 24, 119–124. 

Burnette, L.W., 1960. A review of the physiological properties of 

PVP. Proceedings of the Scientific Section of the Toilet Goods 

Association 38, 1–4. 

Chappell, M.A., George, A.J., Dontsova, K.M., Porter, B.E., Price, C.L., 

Zhou, P., Morikawa, E., Kennedy, A.J., Steevens, J.A., 2009. 

Surfactive stabilization of multi-walled carbon nanotube 

dispersions with dissolved humic substances. Environmental 

Pollution 157, 1081–1087. 

Chen, X., Tam, U.C., Czlapinski, J.L., Lee, G.S., Rabuka, D., Zettl, A., 

Bertozzi, C.R., 2006. Interfacing carbon nanotubes with living 

cells. Journal of American Chemical Society 128, 6292–6293. 

Cherukuri, P., Gannon, C.J., Leeuw, T.K., Schmidt, H.K., 

Smalley, R.E., Curley, S.A., Weisman, R.B., 2006. Mammalian 

pharmacokinetics of carbon nanotubes using intrinsic nearinfrared 

fluorescence. Proceedings of the National Academy of 

Sciences 103, 18882–18886. 

Chin, S.F., Baughman, R.H., Dalton, A.B., Diekmann, G.R., 

Draper, R.K., Mikoryak, C., Musselman, I.H., Poenitzsch, V.Z., 

Xie, H., Pantano, P., 2007. Amphiphilic helical 

peptide enhances the uptake of single-walled carbon 

nanotubes by living cells. Experimental Biological Medicine 

232 (9), 1236–1244. 

Chourasia, M.K., Jain, S.K., 2004. Polysaccharides for colon 

targeted drug delivery. Drug Delivery 11 (2), 129–148. 

Dayeh, V.R., Chow, S.L., Schirmer, K., Lynn, D.H., Bols, N.C., 

2004. Evaluating the toxicity of Triton X-100 to protozoan, 

fish, and mammalian cells using fluorescent dyes as 

indicators of cell viability. Ecotoxicology and Environmental 

Safety 57 (3), 375–382. 

Di Sotto, A., Chiaretti, M., Carru, G.A., Bellucci, S., Mazzantia, G., 

2009. Multi-walled carbon nanotubes: lack of mutagenic 

activity in the bacterial reverse mutation assay. Toxicology 

Letters 184, 192–197. 

Didenko, V.V., Moore, V.C., Baskin, D.S., Smalley, R.E., 2005. 

Visualization of individual single-walled carbon nanotubes by 

fluorescent polymer wrapping. Nano Letters 5 (8), 1563–1567. 

Ding, L.H., Stilwell, J., Zhang, T.T., Elboudware, J.O., Jiang, H.J., 

Selegue, J.P., Cooke, P.A., Gray, J.W., Chen, F.F., 2005. Molecular 

characterization of the cytotoxic mechanism of multiwall 

carbon nanotubes and nanoonions on human skin fibroblast. 

Nano Letters 5, 2448–2464. 

Dodziuk, H., Ejchart, A., Anczewski, W., Ueda, H., Krinichnaya, E., 

Dolgonosa, G., Kutner, W., 2003. Water solubilization, 


determination of the number of different types of single-wall 

carbon nanotubes and their partial separation with respect to 

diameters by complexation with h-cyclodextrin. Chemical 

Communications, 986–987. 

El-Fouly, M.H., Trosko, J.E., Chang, C.C., 1987. Scrape-loading and 

dye transfer: a rapid and simple technique to study gap 

junctional intercellular communications. Experimental Cell 

Research 168, 422–430. 

Elias, A.L., Carrero-Sanchez, J.C., Terrones, H., Endo, M., 

Laclette, H.P., Mauricio, M., 2007. Viability studies of pure 

carbon- and nitrogen-doped nanotubes with Entamoeba 

Histolytica: from amoebicidal to biocompatible structures. 

Small 3 (10), 1723–1729. 

Enyashin, A.N., Gemming, S., Seifert, G., 2007. DNA-wrapped 

carbon nanotubes. Nanotechnology 18, 245702. 

Fiorito, S., Serafino, A., Andreola, F., Bernier, P., 2006. Effects of 

fullerenes and single-wall carbone nanotubes on murine and 

human macrophages. Carbon 44, 1100–1105. 

Flahaut, E., Durrieu, M.C., Remy-Zolghadri, M., Bareille, R., 

Baquey, C., 2006. Investigation of the cytotoxicity of CCVD 

carbon nanotubes towards human umbilical vein endothelial 

cells. Carbon 44 (6), 1093–1099. 

Fortner, J.D., Lyon, D.Y., Sayes, C.M., Boyd, A.M., Falkner, J.C., 

Hotze, E.M., Alemany, L.B., Tao, Y.J., Guo, W., Ausman, K.D., 

Colvin, V.L., Hughes, J.B., 2005. C 60 in water: nanocrystal 

formation and microbial response. Environmental Science 

and Technology 39 (11), 4307–4316. 

Georgakilas, V., Tagmatarchis, N., Pantarotto, D., Bianco, A., 

Briand, J.-P., Prato, M., 2002. Amino acid functionalisation of 

water soluble carbon nanotubes. Chemical Communications, 

3050–3051. 

Grossiord, N., Schhoo, P., Meuldijk, J., Koning, C.E., 2007. 

Determination of the surface coverage of exfoliated carbon 

nanotubes by surfactant molecules in aqueous solution. 

Langmuir 23, 3646–3653. 

Gutierrez, M.C., Garcia-Carvajal, Z.Y., Hortiguela, M.J., Yuste, L., 

Rojo, F., Ferrer, M.L., Monte, F., 2007. Biocompatible MWCNT 

scaffolds for immobilization and proliferation of E. coli. Journal 

of Materials Chemistry 17, 2992–2995. 

Herner, H.A., Trosko, J.E., Masten, S.J., 2001. The epigenetic 

toxicity of pyrene and related ozonation byproducts 

containing an aldehyde functional group. Environmental 


Herzog, E., Byrne, H.J., Davoren, M., Casey, A., Duschl, A., 

Oostingh, G.J., 2009. Dispersion medium modulates oxidative 

stress response of human lung epithelial cells upon exposure 

to carbon nanomaterial samples. Toxicology and Applied 

Pharmacology 236, 276–281. 

Huang, W.J., Taylor, S., Fu, K.F., Lin, Y., Zhang, D.H., Hanks, T.W., 

Rao, A.M., Sun, Y.P., 2002. Attaching proteins to carbon 

nanotubes via diimide-activated amidation. Nano Letters 2 (4), 

311–314. 

Humic acids and their sodium salts. EMEA-The European agency 

for the evaluation of medicinal products. Available via. http:// 

www.emea.europa.eu/pdfs/vet/mrls/055499en.pdf, 1999. 

Hyung, H., Fortner, J.D., Hughes, J.B., Kim, J.H., 2007. Natural organic 

matter stabilizes carbon nanotubes in the aqueous phase. 


Hyung, H., Kim, J.-H., 2008. Natural organic matter (NOM) 

adsorption to multi-walled carbon nanotubes: effect of NOM 

characteristics and water quality parameters. Environmental 


Iijima, S., 1991. Helical microtubules of graphitic carbon. Nature 

354 (6348), 56–58. 

Inoue, K.-i., Koike, E., Yanagisawa, R., Hirano, S., Nishikawa, M., 

Takano, H., 2009. Effects of multi-walled carbon nanotubes on 

a murine allergic airway inflammation model. Toxicology and 

Applied Pharmacology 237 (3), 306–316.

Islam, M.F., Rojas, E., Bergey, D.M., Johnson, A.T., Yodh, A.G., 

2003. High weight fraction surfactant solubilization of 

single-wall carbon nanotubes in water. Nano Letters 3 (2), 

269–273. 

Jia, G., Wang, H., Yan, L., Wang, X., Pei, R., Yan, T., Zhao, Y., 

Guo, X., 2005. Cytotoxicity of carbon nanomaterials: singlewall 

nanotube, multi-wall nanotube, and fullerene. 

Environmental Science and Technology 39, 1378–1383. 

Jiang, J., Gao, L., 2003. Production of aqueous colloidal dispersions 

of carbon nanotubes. Journal of Colloid and Interface Science 

260, 89–94. 

Kang, S., Herzberg, M., Rodrigues, D.F., Elimelech, M., 2008a. 

Antibacterial effects of carbon nanotubes: size does matter! 

Langmuir 24 (13), 6409–6413. 

Kang, S., Mauter, M.S., Elimelech, M., 2008b. Physicochemical 

determinants of multiwalled carbon nanotube bacterial 

cytotoxicity. Environmental Science and Technology 42 (19), 

7528–7534. 

Kang, S., Mauter, M.S., Elimelech, M., 2009. Microbial cytotoxicity 

of carbon-based nanomaterials: implications for river water 

and wastewater effluent. Environmental Science and 

Technology 43 (7), 2648–2653. 

Kang, S., Pinault, M., Pfefferle, L.D., Elimelech, M., 2007. Singlewalled 

carbon nanotubes exhibit strong antimicrobial activity. 

Langmuir 23 (17), 8670–8673. 

Karajanagi, S.S., Yang, H.C., Asuri, P., Sellitto, E., Dordick, J.S., 

Kane, R.S., 2006. Protein-assisted solubilization of singlewalled 

carbon nanotubes. Langmuir 22 (4), 1392–1395. 

Kim, O.-K., Je, J., Baldwin, J.W., Kooi, S., Pehrsson, P.E., Buckley, L.J., 

2004. Solubilization of single-wall carbon nanotubes by 

supramolecular encapsulation of helical amylose. Journal of 

American Chemical Society 125, 4426–4427. 

Lauret, J.-S., Voisin, C., Cassabois, G., Roussignol, P., Delalande, C., 

Capes, L.E., Valentin, E., Jost., O., 2004. Photocreated carrier 

dynamics in isolated carbon nanotubes. Semiconductor 

Science and Technology 19, S486–S488. 

Lee, J.U., Huh, J., Kim, K.H., Park, C., Jo, W.H., 2007. Aqueous 

suspension of carbon nanotubes via non-covalent 

functionalization with oligothiophene-terminated 

poly(ethylene glycol). Carbon 45, 1051–1057. 

Lin, Y., Taylor, S., Li, H.P., Fernando, K.A.S., Qu, L.W., Wang, W., 

Gu, L.R., Zhou, B., Sun, Y.P., 2004. Advances toward 

bioapplications of carbon nanotubes. Journal of Materials 

Chemistry 14 (4), 527–541. 

Lindberg, H.K., Falck, G.C.-M., Suhonen, S., Vippola, M., 

Vanhalab, E., Catalána, J., Savolainena, K., Norppa, H., 2009. 

Genotoxicity of nanomaterials: DNA damage and micronuclei 

induced by carbon nanotubes and graphite nanofibres in 

human bronchial epithelial cells in vitro. Toxicology Letters 

186, 166–173. 

Liu, J., Rinzler, A.G., Dai, H.J., Hafner, J.H., Bradley, R.K., Boul, P.J., 

Lu, A., Iverson, T., Shelimov, K., Huffman, C.B., Rodriguez- 

Macias, F., Shon, Y.S., Lee, T.R., Colbert, D.T., Smalley, R.E., 

1998. Fullerene pipes. Science 280 (5367), 1253–1256. 

Liu, Y.Q., Gao, L., Zheng, S., 2007. Debundling of single-walled 

carbon nanotubes by using natural polyelectrolytes. 

Nanotechnology 18 (36), 365702. 

Lyon, D.Y., Fortner, J.D., Sayes, C.M., Colvin, V.L., Hughes, J.B., 

2005. Bacterial cell association and antimicrobial activity of 

a C60 water suspension. Environmental Toxicology and 

Chemistry 24, 2757–2762. 

Manna, S.K., Sarkar, S., Barr, J., Wise, K., Barrera, E.V., Jejelowo, O., 

Rice-Ficht, A., Ramesh, G.T., 2005. Single-walled carbon 

nanotube induces oxidative stress and activates nuclear 

transcription factor-kB in human keratinocytes. Nano Letters 

5, 1676–1684. 

Marsh, D.H., Rance, G.A., Zaka, M.H., Whitby, R.J., Khlobystov, A.N., 

2007. Comparison of the stability of multiwalled carbon 


nanotube dispersions in water. Physical Chemistry Chemical 

Physics 9 (40), 5490–5496. 

McDonald, T.J., Engtrakul, C., Jones, M., Rumbles, G., Heben, M.J., 

2006. Kinetics of PL quenching during single-walled 

carbon nanotube rebundling and diameter-dependent 

surfactant interactions. Journal of Physical Chemistry B 110, 

25339–25346. 

Meng, J., Yanga, M., Song, L., Kong, H., Wang, C.Y., Wang, R., 

Wang, C., Xie, S.S., Xua, H.Y., 2009. Concentration control of 

carbon nanotubes in aqueous solution and its influence on the 

growth behavior of fibroblasts. Colloids and Surfaces B: 

Biointerfaces 71, 148–153. 

Moore, V.C., Strano, M.S., Haroz, E.H., Hauge, R.H., Smalley, R.E., 

Schmidt, J., Talmon, Y., 2003. Individually suspended singlewalled 

carbon nanotubes in various surfactants. Nano Letters 

3 (10), 1379–1382. 

Muller, J., Huaux, F., Moreau, N., Misson, P., Heilier, J.-F., Delos, M., 

Arras, M., Fonseca, A., Nagy, J.B., Lison, D., 2005. Respiratory 

toxicity of multi-wall carbon nanotubes. Toxicology and 

Applied Pharmacology 207 (31), 221–231. 

Narayan, R.J., Berry, C.J., Brigmon, R.L., 2005. Structural and 

biological properties of carbon nanotube composite films. 

Materials Science and Engineering B 123, 123–129. 

O’Connell, M.J., Bachilo, S.M., Huffman, C.B., Moore, V.C., 

Strano, M.S., Haroz, E.H., Rialon, K.L., Boul, P.J., Noon, W.H., 

Kittrell, C., Ma, J., Hauge, R.H., Weisman, R.B., Smalley, R.E., 

2002. Band gap fluorescence from individual single-walled 

carbon nanotubes. Science 297, 593–596. 

O’Connell, M.J., Boul, P.J., Ericson, L.M., Huffman, C., Wang, Y., 

Haroz, E., 2001. Reversible water-solubilization of singlewalled 

carbon nanotubes by polymer wrapping. Chemical 

Physics Letters 342, 265–271. 

Park, H.J., Heo, H.Y., Lee, S.C., Park, M., Lee, S.-S., Kim, J., Chang, J.Y., 

2006. Dispersion of single-walled carbon nanotubes in water 

with polyphosphazene polyelectrolyte. Journal of Inorganic and 

Organometallic Polymers and Materials 16 (4) 359–364. 

Pulskamp, K., Worle-Knirsch, J.M., Hennrich, F., Kern, K., Krug, H.F., 

2007. Human lung epithelial cells show biphasic oxidative burst 

after single-walled carbon nanotube contact. Carbon 45, 2241– 

2249. 

Saleh, N.D., Pfefferle, L.D., Elimelech, M., 2009. Aggregation 

kinetics of multiwalled carbon nanotubes in aquatic systems: 

measurements and environmental implications. 


Salzmann, C.G., Chu, B.T.T., Tobias, G., Llewellyn, S.A., 

Green, M.L.H., 2007. Quantitative assessment of carbon 

nanotube dispersions by Raman spectroscopy. Carbon 45 (5), 

907–912. 

Satoh, A.Y., Trosco, J., Masten, S.J., 2003. Epigenetic toxicity of 

hydroxylated biphenyls and hydroxylated polychlorinated 

biphenyls on normal rat liver epithelial cells. Environmental 

Science and Technology 37, 2727–2733. 

Sayes, C.M., Liang, F., Hudsona, J.L., Mendeza, J., Guob, W., 

Beach, J.M., Moore, V.C., Doyle, C.D., West, J.L., Billups, W.E., 

Ausmanb, K.D., Colvin, V.L., 2006. Functionalization density 

dependence of single-walled carbon nanotubes cytotoxicity in 

vitro. Toxicology Letters 161, 135–142. 

Schmitt, D., Tran, N., Riefler, S., Jacoby, J., Merkel, D., Marone, P., 

Naouli, N., 2008. Toxicologic evaluation of modified gum 

acacia: mutagenicity, acute and subchronic toxicity. Food and 

Chemical Toxicology 46, 1048–1054. 

Simon-Deckers, A., Gouget, B., Mayne-L’Hermiteb, M., Herlin- 

Boimeb, N., Reynaudb, C., Carrièrea, M., 2008. In vitro 

investigation of oxide nanoparticle and carbon nanotube 

toxicity and intracellular accumulation in A549 human 

pneumocytes. Toxicology 253, 137–146. 

Sinani, V.A., Gheith, M.K., Yaroslavov, A.A., Rakhnyanskaya, A.A., 

Sun, K., Mamedov, A.A., Wicksted, J.P., Kotov, N.A., 2005.

520 

Aqueous dispersions of single-wall and multiwall carbon 

nanotubes with designed amphiphilic polycations. Journal of 

American Chemical Society 127 (10), 3463–3472. 

Star, A., Steuerman, D.W., Heath, J.R., Stoddart, J.F., 2002. 

Starched carbon nanotubes. Angewandte Chemie 

International Edition 41 (14), 2508–2512. 

Tang, Y.J., Ashcroft, J.M., Chen, D., Min, G., Kim, C.-H., 

Murkhejee, B., Larabell, C., Keasling, J.D., Chen, F.F., 2007. 

Charge-associated effects of fullerene derivatives on microbial 

structural integrity and central metabolism. Nano Letters 7 (3), 

754–760. 

Tasis, D., Tagmatarchis, N., Bianco, A. and Prato, M., 2006. 

Chemistry of carbon nanotubes. 106 (3), 1105–1136. 

Tong, Z., Bischoff, M., Nies, L., Applegate, B., Turco, R., 2007. 

Impact of fullerene (C60) on a soil microbial community. 


Trosko, J.E., Chang, B.V., Madhucar, J.E., Klauning, J.E., 1991. 

Chemical, oncogene, and growth factor inhibition of gap 

junctional intercellular communication: an alternative 

hypothesis to carcinogenesis. Pathobiology 58, 265–278. 

Trosko, J.E., Chang, C.-C., Upham, B., Wilson, M., 1998. Epigenetic 

toxicology as toxicant-induced changes in intracellular 

signalling leading to altered gap junctional intercellular 

communication. Toxicology Letters 102–103, 71–78. 

Tseng, C.-H., Wang, C.-C., Chen, C.-Y., 2006. Modification of 

multi-walled carbon nanotubes by plasma treatment and 

further use as templates for growth of CdS nanocrystals. 

Nanotechnology 17, 5602–5612. 

Upham, B.L., Masten, S.J., Lochwood, B.R., Trosko, J.E., 1994. 

Nongenotoxic effects of polycyclic aromatic hydrocarbons and 

their ozonation by-products on the intercellular 

communication of rat liver epithelial cells. Fundamental and 

Applied Toxicology 23, 470–475. 

Vaisman, L., Marom, G., Wagner, H.D., 2006. Dispersions of 

surface-modified carbon nanotubes in water-soluble and 

water-insoluble polymers. Advanced Functional Materials 16 

(3), 357–363. 

Vittorio, O., Raffa, V. and Cuschieri, A., Influence of purity and surface 

oxidation on cytotoxicity of multi-wall carbon nanotubes with 

human neuroblastoma cells. Nanomedicine: Nanotechnology, 

Biology, and Medicine, in press, doi:10.1016/j.nano.2009.02.006. 

Wang, Y., Iqbal, Z., Malhotra, S.V., 2005. Functionalization of 

carbon nanotubes with amines and enzymes. Chemical 

Physics Letters 402, 96–101. 

Weis, L.M., Rummel, A.M., Masten, S.J., Trosko, J.E., Upham, B.L., 

1998. Bay or baylike regions of polycyclic aromatic 

hydrocarbons were potent inhibitors of gap junctional 


intercellular communication. Environmental Health 

Perspectives 106 (1), 17–22. 

Wick, P., Manser, P., Limbach, L.K., Dettlaff-Weglikowskab, U., 

Krumeich, F., Roth, S., Stark, W.J., Bruinink, A., 2007. The 

degree and kind of agglomeration affect carbon nanotube 

cytotoxicity. Toxicology Letters 168, 121–131. 

Wirnitzer, U., Herbold, B., Voetz, M., Ragot, J., 2009. Studies on the 

in vitro genotoxicity of baytubes Ò , agglomerates of engineered 

multi-walled carbon-nanotubes (MWCNT). Toxicology Letters 

186, 160–165. 

Wörle-Knirsch, J.M., Pulskamp, K., Krug, H.F., 2006. Oops they did 

it again! Carbon nanotubes hoax scientists in viability assays. 

Nano Letters 6 (6), 1261–1268. 

Xu, F.-M., Xu, J.-P., Ji, J., Shen, J.-C., 2008. A novel biomimetic 

polymer as amphiphilic surfactant for soluble and 

biocompatible carbon nanotubes (CNTs). Colloids and 

Surfaces B: Biointerfaces 67, 67–72. 

Yamasaki, Y., 1990. Gap junctional intercellular communication 

and carcinogenesis. Carcinogenesis 11, 1051–1058. 

Yang, H., Liu, C., Yang, D., Zhanga, H., Xia, Z., 2009. Comparative 

study of cytotoxicity, oxidative stress and genotoxicity induced 

by four typical nanomaterials: the role of particle size, shape 

and composition. Journal of Applied Toxicology 29, 69–78. 

Yang, S.-T., Wang, X., Jia, G., Guc, Y., Wang, T., Nie, H., Ge, C., 

Wang, H., Liu, Y., 2008. Long-term accumulation and low 

toxicity of single-walled carbon nanotubes in intravenously 

exposed mice. Toxicology Letters 181, 182–189. 

Yang, Z., Chen, X.H., Chen, C.S., 2007. Noncovalent-wrapped 

sidewall multi-walled carbon nanotubes functionalization 

with polyimide. Polymer Composites 28 (1), 36–41. 

Ye, S.-F., Wu, Y.-H., Hou, Z.-Q., Zhang, Q.-Q., 2009. ROS and NF-jB 

are involved in upregulation of IL-8 in A549 cells exposed to 

multi-walled carbon nanotubes. Biochemical and Biophysical 

Research Communications 379, 643–648. 

Yu, B.Z., Yang, J.S., Li, W.X., 2007. In vitro capability of multi-walled 

carbon nanotubes modified with gonadotrophin releasing 

hormone on killing cancer cells. Carbon 45 (10), 1921–1927. 

Yun, Y., Dong, Z., Tan, Z., Schulz, M.J., Shanov, V., 2009. Fibroblast 

cell behavior on chemically functionalized carbon 

nanomaterials. Materials Science and Engineering C 29, 719–725. 

Zhu, Y., Ran, T., Li, Y., Guo, J., Li, W., 2006. Dependence of the 

cytotoxicity of multi-walled carbon nanotubes on the culture 

medium. Nanotechnology 16, 4668–4674. 

Zong, S., Cao, Y., Jua, H., 2007. Direct electron transfer of hemoglobin 

immobilized in multiwalled carbon nanotubes enhanced 

grafted collagen matrix for electrocatalytic detection of 

hydrogen peroxide. Electroanalysis 19 (7–8), 841–846.

Oxidation of atenolol, propranolol, carbamazepine and 

clofibric acid by a biological Fenton-like system mediated by 

the white-rot fungus Trametes versicolor 

Ernest Marco-Urrea a , Jelena Radjenović b , Gloria Caminal c , Mira Petrović b,d , 

Teresa Vicent a, *, Damià Barceló b,e 

a 

Departament d’Enginyeria Química and Institut de Ciència i Tecnologia Ambiental, Universitat Autònoma de Barcelona (UAB), 

08193 Bellaterra, Spain 

b 

Department of Environmental Chemistry, IDAEA-CSIC, c/Jordi Girona 18–26, 08034 Barcelona, Spain 

c 

Unitat de Biocatàlisis Aplicada associada al IQAC (CSIC-UAB), Escola Tècnica Superior d’Enginyeria, UAB, 08193 Bellaterra, Spain 

d 

Institució Catalana de Recerca i Estudis Avançats (ICREA), Barcelona, Spain 

e 

Institut Català de Recerca de l’Aigua (ICRA), Parc Científic i Tecnológic de la Universitat de Girona, Pic de Peguera, 15, 17003 Girona, Spain 

article info 







Keywords: 

Trametes versicolor 


Hydroxyl radical 

Carbamazepine 

Beta-blockers 

Clofibric acid 


abstract 

The presence of pharmaceuticals and personal care products 

(PPCPs) and their metabolites in waters has become an 

emerging environmental issue. To date, most attention has 

been focused on identification, fate and distribution of PPCPs 




Biological advanced oxidation of the pharmaceuticals clofibric acid (CA), carbamazepine 

(CBZP), atenolol (ATL) and propranolol (PPL) is reported for the first time. Extracellular 

oxidizing species were produced through a quinone redox cycling mechanism catalyzed by 

an intracellular quinone reductase and any of the ligninolytic enzymes of Trametes versicolor 

after addition of the lignin-derived quinone 2,6-dimethoxy-1,4-benzoquinone (DBQ) 

and Fe 3þ -oxalate in the medium. Time-course experiments with approximately 10 mg L 1 

of initial pharmaceutical concentration resulted in percent degradations above 80% after 

6 h of incubation. Oxidation of pharmaceuticals was only observed under DBQ redox 

cycling conditions. A similar degradation pattern was observed when CBZP was added at 

the environmentally relevant concentration of 50 mgL 1 . Depletion of DBQ due to the attack 

of oxidizing agents was assumed to be the main limiting factor of pharmaceutical degradation. 

The main degradation products, that resulted to be pharmaceutical hydroxylated 

derivatives, were structurally elucidated. The detected 4- and 7-hydroxycarbamazepine 

intermediates of CBZP degradation were not reported to date. Total disappearance of 

intermediates was observed in all the experiments at the end of the incubation period. 


in municipal wastewater treatment plants (WWTPs), which 

are commonly found at very low concentrations (low ppb 

levels) (Radjenovic et al., 2007, 2009, Reemtsma et al., 2006). To 

avoid their potential adverse health effects on both humans 

and environment, research efforts are underway to develop 

efficient techniques for their removal. Among the alternatives 

* Corresponding author. Departament d’Enginyeria Química and Institut de Ciència i Tecnologia Ambiental. Universitat Autònoma de 

Barcelona (UAB). 08193 Bellaterra, Spain. Tel.: þ34 935812142; fax: þ34 935812013. 

E-mail address: Teresa.Vicent@uab.cat (T. Vicent). 


doi:10.1016/j.watres.2009.09.049

522 

to traditional water treatment processes, advanced oxidation 

processes (AOPs) have received great attention to degrade 

PPCPs (Klavarioti et al., 2009; Song et al., 2008). AOPs can be 

broadly described as aqueous phase oxidation methods based 

on the intemediacy of highly reactive species such as hydroxyl 

radicals ( , OH). This radical acts as a non-selective, very strong 

oxidant agent with the ability to react with chemicals giving 

dehydrogenated or hydroxylated derivates, until achieving 

their total mineralization. To date, the production of , OH 

radicals was based on chemical, photochemical, photocatalytical 

or electrochemical techniques, but little is known 

about the possibility to use biological systems to produce , OH 

radicals for pollutant degradation (Klavarioti et al., 2009). This 

biological approach was proposed recently for several whiterot 

fungi (WRF) including Trametes versicolor and Pleurotus 

eringyi through a simple quinone redox cycling mechanism 

that is referred to as advanced biooxidation (Gómez-Toribio 

et al., 2009a). 

WRF are basidiomycetes that are capable of extensive 

aerobic lignin depolymeration and mineralization due to the 

low substrate specificity and high reactivity of the enzymes 

they produce (principally laccase and peroxidases). Such 

combinations of enzymes with auxiliary oxidoreductive 

enzymes such as cytochrome P450 system are commonly 

involved in the degradation of several PPCPs as for example 

carbamazepine, clofibric acid, ibuprofen and several endocrine 

disrupting chemicals (Cabana et al., 2007; Marco-Urrea 

et al., 2009a). 

In this work, the anti-epileptic drug carbamazepine (CBZP), 

lipid regulator clofibric acid (CA), and b-blockers propranolol 

(PPL) and atenolol (ATL) were the selected pharmaceuticals due 

to their long-term use in Europe and North America and their 

subsequent occurrence in the aquatic environment (Alder 

et al., 2006). The metabolism of these pharmaceuticals in 

mammals has been well established. It is known that CBZP 

undergoes extensive hepatic metabolism by the cytochrome 

P450 system, with oxidation to 10,11-dihydro-10,11-epoxy 

carbamazepine, its further hydration to 11-dihydro-10,11epoxy 

carbamazepine and conjugation of the latter one with 

glucuronide being the main metabolic pathway of CBZP (Kerr 

et al., 1994). CA is the pharmacologically active derivative of 

clofibrate and several other fibrates, and it is metabolized to 

reactive acylating derivatives that have been shown to transacylate 

glutathione forming the corresponding glutathione 

conjugate (Grillo and Benet, 2001). The primary metabolic 

pathways of PPL are glucuronidation, side-chain oxidation and 

ring oxidation, with the main metabolites arising from Ndealkylation 

of the isopropanolamine side-chain, naphthalene 

ring hydroxylation, and side-chain O-glucuronidation (Luan 

et al., 2005). ATL is excreted eagerly via urine as an unchanged 

compound (90%) (Dollery, 1991), with a small percentage of 

atenolol-glucuronide (0.8–4.4%) and hydroxyatenolol (1.1– 

4.4%, hydroxylation of the benzilic position). 

Limited removal of ATL and CA and no removal of CBZP 

have been frequently observed in municipal WWTPs and 

subsequently they deserve more attention due to the high risk 

of passing through later barriers in partly closed water cycles 

(Radjenovic et al., 2007; Reemtsma et al., 2006; Ternes, 1998). 

Joss et al. (2003) found that CBZP was not expected to produce 

acute toxic effects in the aquatic biota, but chronic and 


synergistic effects with other chemicals could not be 

excluded. According to their results and regarding the present 

European legislation on the classification and labelling of 

chemicals (92/32/EEC), they classified CBZP as ‘‘R52/53 Harmful 

to aquatic organisms and may cause long-term adverse 

effects in the aquatic environment’’. On the other side, CA has 

been reported to be a non-hazardous, yet environmentally 

persistent compound (estimated persistence of 21 years in the 

environment) (Buser et al., 1998; Ferrari et al., 2003). Nevertheless, 

the possibility of endocrine disruption activity of CA 

through interference with cholesterol synthesis was reported 

(Pfluger and Dietrich, 2001), which is particularly important 

regarding its long-term effects. Although specific environmental 

effects of b-blockers are low, in a mixture with other bblockers 

the effect of concentration addition can occur, as 

shown in tests with Daphnia magna and phototoxicity assays 

with green algae (Escher et al., 2006). Considering these facts, 

removal techniques need to be developed to effectively 

degrade the selected pharmaceuticals. 

The objective of this study was to demonstrate for the first 

time the degradation of the aforementioned pharmaceuticals 

by biologically induced oxidizing species, produced by the 

WRF T. versicolor. The advanced biooxidation strategy consists 

of the incubation of fungi with a lignin-derived quinone (2,6,dimethoxy-1,4-benzoquione, 

DBQ) and chelated ferric ion 

(Fe 3þ -oxalate). Under these conditions, fungi catalyze the 

conversion of the quinone into hydroquinone (DBQH2) byan 

intracellular quinone reductase, and subsequent oxidation of 

DBQH2 to semiquinone radicals (DBQ, ) is performed in the 

extracellular medium by any of the lignin modifying enzymes 

of the WRF (laccase and peroxidases). Then, Fenton’s reagent 

is formed by DBQ , autoxidation catalyzed by Fe 3þ , in which 

Fe 2þ , 

and superoxide anion radical (O2 ) are generated 

(DBQ , þ Fe 3þ -oxalate / DBQ þ Fe 2þ -oxalate; and Fe 2þ - 

oxalate þ O2 # Fe 3þ -oxalate þ O2 , ), followed by O2 , 

dismutation 

(O2 , þ HO2 , þ H þ / O2 þ H2O2). Production of , OH 

via quinone redox cycling described above was previously 

proposed by estimating production of 2-thiobarbituric acid 

reactive substances (TBARS) from 2-deoxyribose and hydroxylation 

of 4-hydroxybenzoic acid producing 3,4-dimethoxybenzoic 

acid, showing a high correlation between formation 

of Fenton’s reagent and TBARS production (Gómez-Toribio 

et al., 2009a). The inhibitory effect of , OH scavengers and 

catalase on TBARS production rate further strengthened , OH 

generation by this mechanism (Gómez-Toribio et al., 2009a; 

Guillén et al., 2000). However, besides production of , OH 

radicals, other oxidizing species such as ferryl ion might also 

be produced in typical Fenton reactions under certain conditions 

of pH and concentration of organic and inorganic ligands 

(Hug and Leupin, 2003).Therefore, more research is needed to 

ascertain the relative contribution of , OH produced under 

quinone redox cycling conditions. 

This quinone redox cycle mechanism applied to WRF 

increases the range of environmental pollutants susceptible to 

be degraded by these microorganisms due to the high oxidation 

power and low substrate specificity of the oxidizing species 

generated in comparison with their ligninolytic enzymes 

(Marco-Urrea et al., 2009b). These oxidizing species are 

produced rapidly in the extracellular medium and almost total 

destruction of pollutants can be achieved during the first hours

of incubations without the need of adaptation of fungi to 

pollutants (Gómez-Toribio et al., 2009a, b). Furthermore, 

degradation metabolites were elucidated, which were further 

oxidized either by the induced oxidizing agents or by the ligninolytic 

enzyme system of fungus. Studies concerning degradation 

of pharmaceuticals by fungi are very scarce, with 1- and 

2-hydroxy ibuprofen, and 1,2-dihydroxy ibuprofen being the 

only fungal metabolites of pharmaceutical reported up to date 

(Marco-Urrea et al., 2009a). Nevertheless, the knowledge of 

these metabolites is necessary for safe application of fungi 

biocatalyst as bioremediation technology, as well as for the 

elucidation of the key steps of the degradation process. 



CA (CAS No. 882-09-7), CBZP (CAS No. 298-46-4), ATL (CAS No. 

29122-68-7), PPL (CAS No. 3506-09-0) and DBQ (CAS No. 530-55-2) 

were obtained from Sigma–Aldrich Co. All other chemicals used 

were of analytical grade. 

2.2. Fungus and culture conditions 

T. versicolor (ATCC#42530) was maintained by subculturing on 

2% malt extract agar slants (pH 4.5) at room temperature. 

Subcultures were routinely made every 30 days. 

Pellets of T. versicolor were produced by inoculating 1 mL of 

a mycelial suspension, prepared as described previously 

(Marco-Urrea et al., 2008), in 1 L Erlenmeyer flask containing 

250 mL of malt extract medium. This was shaken (135 rpm, 

r ¼ 25 mm) at 25 C for 5 days. Subsequent pellets formed by 

this process were transferred to another 1 L Erlenmeyer flask 

containing 250 mL of a defined medium described elsewhere 

(Marco-Urrea et al., 2008) and were also incubated for 2 days in 

shaking conditions. 

2.3. Degradation experiments 

In time-course experiments, induction of oxidizing agents in 

T. versicolor via quinone redox cycling was routinely performed 

as follows. Mycelial pellets from each sample flask 

were collected by filtration and washed three times with 

MilliQ water. Appropriate amounts of 2-day old washed 

mycelium pellets were incubated in the reaction mixture (see 

figure legends). Reaction mixture contained 500 mM DBQ, 100– 

300 mM Fe 3þ -oxalate, and 100 mM Mn 2þ in 25 mL 20 mM phosphate 

buffer, pH 5, based on previous optimization of , OH 

production in quinone redox cycling (Gómez-Toribio et al., 

2009b). Twenty mL of a solution containing the corresponding 

pharmaceutical in acetonitrile was added into the flasks to 

give the desired final pharmaceutical concentration (approximately 

10 mg L 1 ). The flasks were incubated at 25 oC and 

130 rpm and samples were taken at each point for analysis. 

The samples were filtered through a Millex-GV (Millipore) 

0.22 mm filter and subsequently analyzed by HPLC. 

In order to investigate the degradation of pharmaceuticals 

present at environmentally relevant concentrations, CBZP 

was selected as a model compound and tests flasks were 


amended with CBZP to a 50 mg L 1 concentration. Erlenmeyer 

flasks containing 50 mL of the corresponding reaction mixture 

were sacrificed at each experiment time and were filtered 

through 0.45 mm glass fiber filter from Whatman. The target 

compound was extracted in one step by solid phase extraction 

with Oasis HLB cartridges (60 mg adsorbent, Waters, Barcelona, 

Spain) as is described elsewhere (Radjenovic et al., 2007). 

Briefly, the cartridges were preconditioned sequentially with 

5 mL of methanol and 5 mL of deionized water at neutral pH. 

The cartridge was dried under vacuum and was eluted with 

two 2-mL portions of methanol and subsequently concentrated 

to dryness under a gentle nitrogen stream. The extracts 

were reconstituted with 0.5 mL 25:75 (v/v) acetonitrile-water. 

To obviate the possible influence of light on pharmaceuticals 

stability, all the experiments were carried out in the dark. 

Each experiment included control flasks (without Fe 3þ - 

oxalate). All the results included in the text and shown in 

figures are the mean and standard deviation of duplicate 

experiments. 

2.4. Analytical procedures 

2.4.1. Analysis of pharmaceuticals and DBQ 

Analysis of the pharmaceuticals and DBQ were performed 

using a Dionex 3000 Ultimate HPLC equipped with a UV 

detector at 230 nm. The column temperature was 30 C and 

a sample volume of 20 mL was injected from a Dionex autosampler. 

Chromatographic separation of CBZP, PPL, CA and 

DBQ was achieved on a GraceSmart RP 18 column 

(250 mm 4 mm, particle size 5 mm). The mobile phase consisted 

of 6.9 mmol L 1 acetic acid adjusted to pH 4 (by NaOH) 

with 35% v/v acetonitrile. It was delivered isocratically at 1 mL 

min 1 as was described elsewhere (Stafiej et al., 2007). ATL 

analysis was performed with a Ascentis C18 column 

(150 mm 4.6 mm, particle size 5 mm). Mobile phase of 0.01 M 

ammonium acetate (pH 7) and mobile phase B (acetonitrile) 

were delivered at flow rate of 1.2 mL min 1 and used for 

gradient elution of ATL (t ¼ 0 min A ¼ 95%, t ¼ 20 min A ¼ 80%). 

The detection limit was calculated to be

524 

Fig. 1 – Time-course of CBZP (a) and CA (b) degradation in T. versicolor cultures under quinone redox cycle conditions. 

Symbols: (B) CBZP and CA in the reaction mixture (25 mL 20 mM phosphate buffer containing 500 mM DBQ, 100–300 mM 

Fe 3D -oxalate, 100 mM Mn 2D , and 10 mg L L1 of the corresponding pharmaceutical); (C) CBZP and CA in the control treatment 

(non-containing DBQ and Fe 3D -oxalate); (-) DBQ(H2) in the reaction mixture; (:) metabolite formation expressed as relative 

area (A/A0) where A is the corresponding metabolite and A0 is the parent drug in the control treatment at time zero. In the 

case of CBZP where two hydroxylated isomers were found, black and white triangles refer to P254A and B, respectively. 

Incubations were carried out with 1.5 ± 0.1 and 1.7 ± 0.2 mg dry weight mL L1 for CBZ and CA, respectively. 

increased to 60% at 11 min, further increased to 90% of A in the 

next 3 min and held isocratically for 2 min. Together with 

returning to initial conditions, the total run time was 20 min. 

CA was analyzed in the negative ion (NI) mode with mobile 

phases consisting of (A) methanol; and (B) water. The elution 

started at 5% of A for the first min, which was raised to 70% in 

the next 7 min, to 90% in the next 2 min, and then returned to 

initial conditions. The total run time was 13 min. The injection 

volume of the sample was 10 mL. 

The mass spectrometry analysis on the QqToF instrument 

was performed in wide pass quadrupole mode, for MS 

experiments, with the ToF data being collected between m/z 

50–800. The capillary and cone voltages were set to 3000 and 

25–30 V, respectively. Data were collected in the centroid 

mode, with a scan accumulation time of 1 s. The instrument 

was operated at a resolution of 5000 (FWHM). The nebulisation 

gas was set to 500 L h 1 at a temperature of 350 C, 


the cone gas was set to 50 L h 1 , and the source temperature 

to 120 C. All analyses were acquired using an independent 

reference spray via the LockSpray interference to ensure 

accuracy and reproducibility. Sulfaguanidine (C7H10N4O2S) 

was used as the internal lock mass in both PI mode 

([M þ H] þ ¼ m/z 215.0602) and NI mode ([M H] ¼ m/z 

213.0446). The LockSpray frequency was set at 11 s. Fragmentation 

of precursor ions was done by applying collision 

energies in the range of 10–35 eV, using argon as a collision 

gas at a pressure of ~20 psi. 

Elemental compositions of the molecular ions and their 

fragments were determined and exact masses were calculated 

with the help of MassLynx V4.1 software incorporated in the 

instrument. Since the software calculation of the accurate 

mass of cation is performed by adding a hydrogen atom 

instead of proton, mass of one electron (i.e., 0.0005) was subtracted 

from the calculated mass.

2.4.3. Mycelial dry weight 

Mycelial dry weights were determined by vacuum filtering the 

cultures through reweighed glass filters (Whatman GF/C, 

Maidstone, England). The filters containing the mycelial mass 

were placed in glass dishes and dried at 100 C to constant 

weight. 

3. Results and discussions 

3.1. Oxidation of CA, CBZP, PPL and ATL by T. 

versicolor under quinone redox cycling conditions 

The strategy proposed here to degrade the selected pharmaceuticals 

couples both AOP and biological strategies by 

inducing the production of oxidizing agents such as , OH in the 

white-rot fungus T. versicolor. In the experiments conducted, 


Fig. 2 – Time-course of PPL (a) and ATL (b) degradation in T. versicolor cultures under quinone redox cycle conditions. Culture 

conditions were as described in the legend for Fig. 1. Symbols: (B) PPL and ATL in the reaction mixture; (C) PPL and ATL in 

the control treatment; (-) DBQ(H2) in the reaction mixture; (:) metabolite formation expressed as relative area (A/A0) where 

A is the corresponding metabolite and A0 is the parent drug in the control treatment at time zero. Incubations were carried 

out with 1.9 ± 0.1 mg dry weight mL L1 for both PPL and ATL. 

oxalate was used as chelant agent of Fe 3þ instead of EDTA and 

Mn 2þ was added into the incubation mixture, since the 

oxidizing power under these conditions was substantially 

improved as discussed previously (Gómez-Toribio et al., 

2009b). Figs. 1 and 2 show the time-courses degradation of the 

selected pharmaceuticals in test flasks amended with 

approximately 10 mg L 1 of the target pharmaceuticals, when 

incubated under DBQ redox cycling conditions. Control 

treatments in the absence of DBQ and Fe 3þ but containing 

fungus were performed to demonstrate both the positive 

involvement of oxidizing agents formed from the quinone 

redox cycling mechanism and to discard the role of fungal 

sorption in the removal of pharmaceuticals. Furthermore, 

control treatments permit us to rule out the intracellular and 

extracellular enzymatic system of T. versicolor in the degradation 

of the pharmaceuticals. Previously CA and CBZP were 

reported to be degraded by the cyt P450 system of T. versicolor 

grown in a chemically define medium (Marco-Urrea et al.,

526 

Fig. 3 – Time-course of CBZP degradation added at low 

concentration of 50 mg L L1 in T. versicolor cultures under 

quinone redox cycle conditions. Symbols: (;) CBZP in the 

reaction mixture; (B) CBZP in the control treatment; (C) 

CBZP in uninoculated bottles. Incubations were carried out 

with 3.8 ± 0.1 mg dry weight mL L1 . Values plotted are 

means ± standard deviations for triplicate cultures. 

2009a) but as can be observed in Figs. 1 and 2 degradation did 

not occur in control treatments for the incubation period 

studied. 

Results in Figs. 1 and 2 show a steep decrease of the 

concentration of pharmaceuticals in the first hours of incubation 

when the fungus was incubated under quinone redox 

cycling conditions. Approximately 50% of CBZP, CA and ATL 

were degraded the first hour and a plateau was reached above 

80% after the sixth hour of incubation. However, the incubation 

was continued to 24 h to investigate the further degradation 

of possible metabolites formed during the degradation 

process (see Section 3.3). 

The consumption of DBQ was also investigated in the same 

samples used for the determination of pharmaceutical 

degradation. This analysis included the determination of DBQ 

and its counterpart (DBQH 2) levels and is shown in Figs. 1 and 

2 as the sum of both: DBQ(H 2). A high correlation between 

DBQ(H2) and pharmaceutical degradation was found, suggesting 

that DBQ(H2) consumption was due to the attack of the 

induced oxidizing agents. Furthermore, since quinone redox 

cycling and therefore the production of oxidizing species 

cannot be sustained without the presence of DBQ(H2), the 

plateau observed in the pharmaceutical degradation rate can 

be ascribed to DBQ(H 2) depletion. 

In an effort to discuss the relevance of this strategy in 

regards to the normal concentrations of pharmaceuticals 

found in aqueous environments (that are frequently detected 

at levels from ng L 1 to low mg L 1 ), one experiment was 

carried out at 50 mg L 1 . For this purpose CBZP was chosen as 

model compound since it is found to be recalcitrant in sewer 

treatment facilities and in the environment (Zhang et al., 

2008). As is shown in Fig. 3 the concentration of CBZP was 

decreased to approximately 10 mg L 1 , due to the mechanisms 

of fungal sorption and oxidation-induced degradation. In this 


experiment, adsorption and transformation of pharmaceutical 

by fungus cannot be clearly distinguished, since live 

cultures can also degrade adsorbed pollutants by intracellular 

mechanisms (Blánquez et al., 2004). In this case, adsorption 

was higher than that observed at 10 mg L 1 probably due to 

the highest concentration of fungal mycelium in the experimental 

flasks (see figure legend). Similarly to that described 

for higher concentrations, the plateau in CBZP degradation 

under quinone redox cycling conditions was observed at the 

fifth hour, when DBQ(H 2) was expected to be depleted under 

the tested conditions. 

3.2. QqToF-MS 2 fragmentation pathways of 

pharmaceuticals and their metabolites 

Table 1 summarizes the exact masses of molecular and fragment 

ions, together with recalculated mass errors and double 

bond equivalents (DBEs) given by the software (mass 

measurements accuracy threshold of 5 mg L 1 ). Since all 

fragment ions had an even number of electrons (i.e., DBE was 

half integer number in all cases), all neutral losses had likewise 

even electron configurations. These data were obtained under 

optimized conditions of collision energy and cone voltage in 

ESI(þ) and ESI( )MS 2 experiments on the QqToF instrument. 

In ESI(þ) full-scan experiments performed at UPLC-QqToF 

instrument, ATL (molecular ion [M þ H] þ 267) eluted at 

3.4 min. Another peak was observed at 3.2 min, with the 

molecular ion [M þ H] þ 283, denominated as metabolite P282. 

Considering that the mass of P282 was shifted 16 Da upwards 

relative to the parent compound, hydroxylation by , OH radical 

attack was assumed. In Table 1 are presented calculated and 

measured masses of fragment ions of ATL and its metabolite 

P282, determined in ESI(þ)-MS 2 experiments at the QqToF-MS 

instrument. The most abundant fragment ions in the MS 2 

spectrum of ATL were detected m/z 190 (loss of 77 Da, isopropylamine 

and water) and m/z 145 (further cleavage of CO 

and intermolecular cyclization and rearrangement) (see 

Fig. 4c). In the fragmentation pattern of the molecular ion 

[M þ H] þ 283, characteristic losses of isopropyl group (42 Da) 

and ammonia (17 Da) afforded the fragment ions at m/z 241 

and 224 (see Fig. 4d). The subsequent losses of water and CO 

from the fragment ion m/z 224 led to the formation of m/z 206 

and 178 fragment ions. The fragment ion m/z 161 was formed 

through cleavage of isopropylamine, water, CO and ammonia 

and intramolecular cyclization, as suggested in the insert in 

Fig. 4d. The fragment ion m/z 116 was also present in the MS 2 

spectrum of P282, indicating that the hydroxylation had to 

occur at the p-hydroxyphenylacetamide fraction of ATL. From 

the comparison of the two fragmentation patterns, it was 

deduced that the –OH group was attached to the alkyl sidechain, 

at the C-atom next to the ether oxygen. 

In the case of PPL, MS 2 fragmentation of the molecular ion 

[M þ H] þ 260 rendered intense signals at m/z 183 and 157, 

corresponding to the subsequent cleavages of aminoisopropyl 

moiety and water, and C2H2, respectively (see Fig. 4a). The less 

intense fragment ions at m/z 242 and 218 were formed through 

cleavages of water and isopropylamine, respectively. The 

fungal metabolite of PPL eluted 0.6 min earlier, with the 

molecular ion [M þ H] þ 276, and was denominated as P275. 

Besides the molecular ion with mass 16 Da greater than

Table 1 – Accurate mass measurement of product ions of carbamazepine (CBZP), clofibric acid (CA), propranolol (PPL), and atenolol (ATL) and their degradation products 

P254, P230, P275 and P282, respectively, as determined by UPLC-QqToF in ESI(D) MS 2 mode for CBZP, PPL and ATL, and ESI(L) MS 2 mode for CA. 

Comp. Precursor ion/product ion Elemental formula Mass (m/z) Error DBE a 

Exp. Theor. mDa ppm 

CBZP [M þ H] þ 

C15H13N2O 237.1019 237.1022 0.3 1.3 10.5 8.93 

[M þ Na] þ 

C15H12N2ONa 259.0844 259.0842 0.2 0.8 10.5 

[M þ H NH3] þ 

C15H10NO 220.0761 220.0757 0.4 1.8 11.5 

[M þ H CONH] þ 

C14H12N 194.0960 194.0964 0.4 2.1 9.5 

P254 [M þ H] þ 

C15H13N2O2 253.0976 253.0972 0.4 1.6 10.5 7.24; 7.81 

[M þ Na] þ 

C15H12N2O2Na 275.0799 275.0791 0.8 2.9 10.5 

[M þ H NH3] þ 

C15H10NO2 236.0710 236.0706 0.4 1.7 11.5 

[M þ H CONH] þ 

C14H12NO 210.0906 210.0913 0.7 3.3 9.5 

[M þ H CONH3] þ 

C14H10NO 208.0762 208.0762 0 0 10.5 

[M þ H CONH CO] þ 

C13H12N 182.0965 182.0964 0.1 0.5 8.5 

CA [M H] C10H10ClO3 213.0327 213.0324 0.3 1.4 5.5 5.90 

[M H C4H6O2] C6H4ClO 126.9944 126.9956 1.2 9.4 4.5 

[M H C6H4ClO] C4H6O2 85.0294 85.0290 0.4 4.7 2.5 

P230 [M H] C10H10ClO4 229.0254 229.0268 1.4 6.1 5.5 6.40 

[M H C4H6O2] C6H4ClO2 142.9909 142.9905 0.4 2.8 4.5 

PPL [M þ H] þ 

C16H22NO2 260.1653 260.1645 0.8 3.1 6.5 6.7 

[M þ H H2O] þ 

C16H20NO 242.1547 242.1539 0.8 3.3 7.5 

[M þ H C(CH3)2] þ 

C13H16NO2 218.1170 218.1176 0.6 2.7 6.5 

[M þ H H2O NH2CH(CH3) 2] þ 

C13H11O 183.0814 183.0804 1.0 5.5 8.5 

[M þ H H2O NH2CH(CH3) 2 C2H2] þ 

C11H9O 157.0648 157.0648 0 0 7.5 

[M þ H C10H8O] þ 

C6H14NO 116.1066 116.1070 0.4 3.4 0.5 

P275 [M þ H] þ 

C16H22NO3 276.1583 276.1594 1.1 4.0 6.5 6.1 

[M þ H H2O] þ 

C16H20NO2 258.1492 258.1489 0.3 1.2 7.5 

[M þ H H2O O NH2CH(CH3) 2] þ 

C13H11O 183.0795 183.0804 0.9 4.9 8.5 

[M þ H H2O O NH2CH(CH3)2 C2H2] þ 

C11H9O 157.0664 157.0648 1.6 10.1 7.5 

[M þ H C10H8O2] þ 

C6H14NO 116.1079 116.1070 0.9 7.7 0.5 

ATL [M þ H] þ 

C14H23N2O3 267.1696 267.1704 0.8 3.0 4.5 3.4 

[M þ H C(CH3)2] þ 

C11H17N2O3 225.1222 225.1234 5.3 1.2 4.5 

[M þ H C(CH3)2 NH3] þ 

C11H14NO3 208.0975 208.0969 0.6 2.9 5.5 

[M þ H C(CH3) 2 H2O] þ 

C11H12NO2 190.0874 190.0863 1.1 5.8 6.5 

[M þ H C(CH3) 2 NH3 H2O CO NH3] þ 

C10H9O 145.0650 145.0648 0.2 1.4 6.5 

[M þ H CHCONH2 H2O C(CH3)2 NH3] þ 

C9H9O 133.0658 133.0648 1.0 7.5 5.5 

[M þ H C8H9NO2] þ 

C6H14NO 116.1070 116.1071 0.1 0.9 0.5 

P282 [M þ H] þ 

C14H23N2O4 283.1663 283.1652 1.1 3.9 4.5 3.2 

[M þ H C(CH3) 2] þ 

C11H17N2O4 241.1186 241.1183 0.3 1.2 4.5 

[M þ H C(CH3)2 NH3] þ 

C11H14NO4 224.0901 224.0917 1.6 7.1 5.5 

[M þ H C(CH3)2 NH3 H2O] þ 

C11H12NO3 206.0807 206.0812 0.5 2.4 6.5 

[M þ H C(CH3) 2 NH3 H2O CO] þ 

C10H12NO2 178.0862 178.0863 0.1 0.6 5.5 

[M þ H C(CH3)2 NH3 CO H2O NH3] þ 

C10H9O2 161.0617 161.0597 2.0 12.4 6.5 

[M þ H C8H9NO3] þ 

C6H14NO 116.1055 116.1070 1.5 12.9 0.5 

Only product ions with abundances higher than 10% are taken into account. 

a DBE, double bond equivalent. 

b tr-UPLC retention time. 

tr b ,min 

water research 44 (2010) 521–532 527

528 

Fig. 4 – Spectra obtained in ESI(D)-MS/MS experiments at QqToF instrument (cone voltages 20–25 V, collision energies 15– 

25 eV) for: (a) standard mixture of PPL in methanol/water (v/v, 25/75) at 10 mg L L1 , (b) degradation product of PPL, P275, in 

the sample taken after 2 h, (c) standard mixture of ATL in methanol/water (v/v, 25/75) at 10 mg L L1 , and (d) degradation 

product of ATL, P282, in the sample taken after 4 h. 

[M þ H] þ 260, fragment ion at m/z 258 was observed, equivalent 

to the m/z 242 in the MS 2 spectrum of PPL (Fig. 4b). The rest of 

the observed fragment ions of the molecular ion [M þ H] þ 276 

were the same (i.e., m/z 116, 157 and 183), indicating that the 

naphthalene moiety was the site of , OH radical attack. 

In full-scan experiments CBZP eluted at 8.94 min, whereas 

two new peaks emerged at 7.24 and 7.81 min. The identical 

fragmentation patterns of these two peaks and their molecular 

weight (MW) 16 Da higher than the parent compound 

(see Figs. 5c,d) suggested that the formed products were two 

hydroxylated isomers of CBZP, denominated as P254A and B. 

In Table 1 exact and measured masses were presented for 

a more intense M254B. The collision-induced-dissociation 

experiments with CBZP (molecular ion [M þ H] þ 237) revealed 

the formation of only two fragment ions at m/z 220 and 194, as 

a consequence of the loss of ammonia and CONH group, 

respectively. Equivalent fragments were observed for P254A,B, 

at m/z 236 and 210. The strong signal observed at m/z 208 

suggested that the –OH group was probably attached at the C4 

and C7-atom for the two observed isomers, since the loss of 2 

H-atoms could be favoured by intramolecular cyclization as 

presented in the insert of Fig. 5d. Further cleavage of CO in the 


fragment ion m/z 208 afforded the signal at m/z 182. Nevertheless, 

based on the retention times only it could not be 

deduced which peak belongs to which product. Probably the 

formation of intramolecular hydrogen bond between the 

attached oxygen and hydrogen atoms of the amide group 

caused one of the isomers to elute ~0.6 min later than the 

other. 

Finally, ESI( ) full-scan screenings of samples from the 

experiments with CA (molecular ion [M þ H] þ 213) revealed 

a new signal appearing 0.5 min after the CA peak, at 6.40 min. 

Similar to the other metabolites identified, its MW was 

deferring for 16 Da relative to the original drug, with the 

molecular ion [M þ H] þ 229 (i.e., P230). Very poor MS 2 fragmentation 

was observed for both parent compound and its 

fungal metabolite. The collision of [M þ H] þ 213 rendered two 

fragment ions at m/z 126 and 85, representing the cleavage of 

the ether bond (see insert in Fig. 5a), whereas in the spectrum 

of [M þ H] þ 229 only one fragment ion was detected at m/z 142, 

equivalent to the m/z 126 but with mass shifted upwards for 

16 Da. Nevertheless, this was enough evidence to assume 

hydroxylation of the benzene ring, although the exact position 

of the attached –OH group could not be deduced.

The degradation intermediates of each pharmaceutical 

detected in this section are found in Fig. 6. 

3.3. Discussion of the results on the fungal metabolites 

of pharmaceuticals 

The formation of metabolites was depicted in Figs. 1 and 2 and 

was expressed as relative area (A/A0) measured by integration 

of LC-MS peaks of the corresponding metabolite (A) and the 

parent drug in the control treatments at time zero (A0), since 

due to the lack of authentic analytical standards for the newly 

identified products their quantitative determination was not 

possible. Examination of the profiles show qualitatively 

similar formation pattern of metabolites across the incubation 

time, obtaining a maximum value of A/A 0 between the 

first and fourth hour and disappearing completely in all the 

cases after 24 h. The metabolites appeared to be formed at low 

concentration levels in comparison with the parent drug, with 

a maximum A/A0 values for each case in the range of 0.012 

(PPL, Fig. 2a) to 0.27 (CA, Fig. 1b). 

The same compound that was identified in our experiments 

as the fungal metabolite of ATL, P282, was previously 

reported as one of the intermediate products of solar photo- 


Fig. 5 – Spectra obtained in MS/MS experiments at QqToF instrument (cone voltages 15–30 V, collision energies 15–25 eV) 

for: (a) standard mixture of CA in methanol/water (v/v, 25/75) at 10 mg L L1 , (b) degradation product of CA, P230, in the 

sample taken after 2 h, (c) standard mixture of CBZP in methanol/water (v/v, 25/75) at 10 mg L L1 , and (d) degradation product 

of CBZP, P254, in the sample taken after 2 h. 

Fenton and TiO2 photocatalytic treatment of ATL in pilotscale 

compound parabolic collectors (Radjenovic et al., 2009). 

In the case of these photocatalytic treatments, P282 was 

observed together with its keto tautomer (P280), and was 

formed through , OH-mediated reactions. Furthermore, Song 

et al. (2008) reported a product of ATL with MW 282, formed in 

the reaction of ATL with , OH radicals. This product was 

identified by 60 Co g-irradiation and LC-MS, although the , OH 

attacks was assumed to occur at the benzene ring. In the 

same study hydroxylated derivative of PPL was also detected 

with MW 275 and –OH group attached at the naphthalene 

ring, which coincides with the structure of P275 identified in 

the present study. Hydroxylation of naphthalene moiety is 

phase I reaction in the human metabolism of PPL, in which 4hydroxy 

PPL is formed (Luan et al., 2005). Cytochrome P450 

isozymes were found to be responsible for oxidative metabolic 

pathways of PPL in humans proceeding through naphthalene 

ring-hydroxylations at the 4-, 5-, and 7-positions and 

side-chain N-desisopropylation (Masubuchi et al., 1994). 

However, in our study no metabolites were found in fungal 

control treatments indicating that P275 was generated via 

induction of oxidizing agents instead of fungal cyt P450 

mechanism.

530 

Fig. 6 – Selected pharmaceuticals and proposed structures of their degradation products. 

Sirés et al. (2007) proposed the formation of a hydroxylated 

intermediate as the observed in our study after the 

hydroxylation of CA on its C 2-position by electro-Fenton and 

photoelectro-Fenton. This proposed metabolite was not 

identified by pure standards in the mentioned study but it 

was assigned to a dehydrated species of 2-(4-chloro-2hydroxyphenoxy)-2-methylpropionic 

acid on the basis of its 

mass fragmentation spectra. On the other hand, by using an 

HPLC-DAD and FLD, Doll and Frimmel (2004) proposed TiO2 

photocatalytic degradation pathway of CA, which did not 

include the hydroxylated product as the one observed in our 

study. During photocatalysis CA underwent substitution of 

–Cl by –OH group in the para-position, whereas products such 

as 2-(4-hydroxyphenoxy)-isobutyric acid and hydroquinone 

were formed. On the other side, cleavage of isobutyric acid 

from the side-chain afforded 4-chlorophenol and 4-chlorocatechol. 

Interestingly, in the abovementioned reports, 

substrates of laccase such as phenol and 4-chlorophenol 

have been described as typical metabolites of CA degradation 

by AOP and once produced they could therefore be mineralized 

by the ligninolytic enzymatic system of WRF. But 

perhaps even more interesting is the production of different 


quinone metabolites (also described for CA degradation by 

, OH attack) that could theoretically be incorporated into the 

quinone redox cycling enhancing the degradation rates of 

the target pollutant. 

As far as CBZP is concerned, there are few studies 

regarding identification of products in , OH-induced degradation 

pathway. Vogna et al. (2004) reported the formation of 

toxic acridine intermediates in UV/H2O2 treatment of CBZP, 

formed in , OH and , HO2-mediated reactions. Chiron et al. 

(2006) also found acridine as major photodegradation intermediate 

of CBZP after direct photolysis, but in the presence of 

Fe þ3 

they identified a hydroxycarbamazepine structure 

derived from a , OH attack on the C 6 aromatic ring. The 

formation of 2- and 3-hydroxylated derivatives of CBZP has 

been reported in the metabolism of CBZP by the cytochrome 

P450 in mammals (Mandrioli et al., 2001; Pearce et al., 2002) as 

well as in the two model fungi Cunninghamella elegans and 

Umbelopsis ramanniana used to study mammalian metabolism 

for many drugs (Kang et al., 2008). As far as we know, the two 

hydroxylated metabolites of CBZP in the 4- and 7-position 

(P254A and B) detected in this study are reported for the first 

time.


Biological advanced oxidation of ATL, CA, CBZP and PPL was 

demonstrated for the first time by inducing extracellular 

oxidizing species in T. versicolor. Under quinone redox cycling 

conditions, pharmaceuticals degradation (added at 10 mg L 1 ) 

reached a plateau after 6 h of incubation achieving degradation 

levels up to 80%. The plateau observed was due to DBQ 

depletion, which was shown to be concomitant to pharmaceutical 

degradation. The feasibility of this novel technology 

to degrade pharmaceuticals at low levels commonly found in 

the environment (50 mg L 1 ) was also demonstrated. Hydroxylated 

intermediates of all pharmaceuticals were identified 

and they disappeared after the incubation period, suggesting 

the total mineralization of the target compound. 


This work was supported by the Spanish MICINN (project 

CTM2007-60971/TECNO) and MMAMRM (project 010/PC08/3-04). 

The Department of Chemical Engineering of the UAB is the 

Unit of Biochemical Engineering of the XRB de la Generalitat de 

Catalunya. E. Marco-Urrea, T. Vicent and G. Caminal are 

members of a Consolidated Research Group of Catalonia 

(2005SGR 00220). Jelena Radjenović gratefully acknowledges 

the JAE Program (CSIC-European Social Funds). 

references 

Alder, A.C., Bruchet, A., Carballa, M., Clara, M., Joss, A., Löffler, D., 

McArdell, C.S., Miksch, K., Omil, F., Tuhkanen, T., Ternes, T.A., 

2006. Consumption and occurrence. In: Ternes, T.A., Joss, A. 

(Eds.), Human Pharmaceuticals, Hormones and Fragrances. 

IWA Publishing, London, pp. 15–54. 

Blánquez, P., Casas, N., Font, X., Gabarrell, X., Sarrà, M., 

Caminal, G., Vicent, T., 2004. Mechanism of textile metal dye 

biotransformation by Trametes versicolor. Water Research 38 

(8), 2166–2172. 

Buser, H.-R., Müller, M.D., Theobald, N., 1998. Occurrence of the 

pharmaceutical drug clofibric acid and the herbicide 

mecoprop in various Swiss Lakes and in the North Sea. 


Cabana, H., Jones, J.P., Agathos, S.N., 2007. Elimination of 

endocrine disrupting chemicals using white-rot fungi and 

their lignin modifying enzymes: a review. Engineering in Life 

Sciences 7 (5), 429–456. 

Chiron, S., Minero, C., Vione, D., 2006. Photodegradation 

processes of the antiepileptic drug carbamazepine, relevant to 

estuarine waters. Environmental Science and Technology 40 

(19), 5977–5983. 

Doll, T.E., Frimmel, F.H., 2004. Kinetic study of photocatalytic 

degradation of carbamazepine, clofibric acid, iomeprol and 

iopromide assisted by different TiO 2 materials – determination 

of intermediates and reaction pathways. Water Research 38 (4), 

955–964. 

Dollery, C., 1991. Therapeutic Drugs. Churchill Livingstone, 

Edinburgh. 

Escher, B.I., Bramaz, N., Richter, M., Lienert, J., 2006. Comparative 

ecotoxicological hazard assessment of beta-blockers and their 


human metabolites using a mode-of-action-based test battery 

and a QSAR approach. Environmental Science and Technology 

40 (23), 7402–7408. 

Ferrari, B., Paxéus, N., Giudice, R.L., Pollio, A., Garric, J., 2003. 

Ecotoxicological impact of pharmaceuticals found in treated 

wastewaters: study of carbamazepine, clofibric acid, and 

diclofenac. Ecotoxicology and Environmental Safety 55 (3), 

359–370. 

Gómez-Toribio, V., García-Martín, A.B., Martínez, M.J., 

Martínez, A.T., Guillén, F., 2009a. Induction of extracellular 

hydroxyl radical production by white-rot fungi through 

quinone redox cycling. Applied and Environmental 

Microbiology 75 (12), 3944–3953. 

Gómez-Toribio, V., García-Martín, A.B., Martínez, M.J., 

Martínez, A.T., Guillén, F., 2009b. Enhancing the production of 

hydroxyl radicals by Pleurotus eryngii via quinone redox cycling 

for pollutant degradation. Applied and Environmental 

Microbiology 75 (12), 3954–3962. 

Grillo, M.P., Benet, L.Z., 2001. Interaction of gglutamyltranspeptidase 

with clofibryl- S-acyl-glutathione in 

vitro and in vivo in rat. Chemical Research in Toxicology 14 (8), 

1033–1040. 

Guillén, F., Gómez-Toribio, V., Martínez, M.J., Martínez, A.T., 2000. 

Production of hydroxyl radical by the synergistic action of 

fungal laccase and aryl alcohol oxidase. Archives of 

Biochemistry and Biophysics 383 (1), 142–147. 

Hug, S.J., Leupin, O., 2003. Iron-catalyzed oxidation of arsenic(III) 

by oxygen and by hydrogen peroxide: pH-dependent 

formation of oxidants in the Fenton reaction. Environmental 


Joss, A., Repetto, G., Rios, J.C., Hazen, M.J., Molero, M.L., del 

Peso, A., Salguero, M., Fernandez-Freire, P., Perez-Martin, J.M., 

Camean, A., 2003. Ecotoxicological evaluation of 

carbamazepine using six different model systems with 

eighteen endpoints. Toxicology in Vitro 17 (5-6), 525–532. 

Kang, S.-I., Kang, S.-Y., Hur, H.-G., 2008. Identification of fungal 

metabolites of anticonvulsant drug carbamazepine. Applied 

Microbiology and Biotechnology 79 (4), 663–669. 

Kerr, B.M., Thummel, K.E., Wurden, C.J., Klein, S.M., Kroetz, D.L., 

Gonzalez, F.J., Levy, R.H., 1994. Human liver carbamazepine 

metabolism: role of CYP3A4 and CYP2C8 in 10,11-epoxide 

formation. Biochemical Pharmacology 47 (11), 1969–1979. 

Klavarioti, M., Mantzavinos, D., Casinos, D., 2009. Removal of 

residual pharmaceuticals from aqueous systems by advanced 

oxidation processes. Environment International 35 (2), 

402–417. 

Luan, L.-J., Shao, Q., Ma, J.-Y., Zeng, S., 2005. Stereoselective 

urinary excretion of S-( )- and R-(þ)-propranolol glucuronide 

following oral administration of RS-propranolol in Chinese 

Han subjects. World Journal of Gastroenterology 11 (12), 

1822–1824. 

Mandrioli, R., Albani, F., Casamenti, G., Sabbioni, C., Raggi, M.A., 

2001. Simultaneous high-performance liquid chromatography 

determination of carbamazepine and five of its metabolites in 

plasma of epileptic patients. Journal of Chromatography B 762 

(2), 109–116. 

Marco-Urrea, E., Parella, T., Gabarrell, X., Caminal, G., Vicent, T., 

Reddy, C.A., 2008. Mechanistics of trichloroethylene 

mineralization by the white-rot fungus Trametes versicolor. 

Chemosphere 70 (3), 404–410. 

Marco-Urrea, E., Pérez-Trujillo, M., Vicent, T., Caminal, G., 2009a. 

Ability of white-rot fungi to remove selected pharmaceuticals 

and identification of degradation products of ibuprofen by 

Trametes versicolor. Chemosphere 74 (6), 765–772. 

Marco-Urrea, E., Aranda, E., Caminal, G., Guillén, F., 2009b. 

Induction of hydroxyl radical production in Trametes versicolor 

to degrade recalcitrant chlorinated hydrocarbons. Bioresource 

Technology 100 (23), 5757–5762.

532 

Masubuchi, Y., Hosokawa, S., Horie, T., Suzuki, T., Ohmori, S., 

Kitada, M., Narimatsu, S., 1994. Cytochrome P450 isozymes 

involved in propranolol metabolism in human liver 

microsomes. The role of CYP2D6 as ring-hydroxylase and 

CYP1A2 as N-desisopropylase. Drug Metabolism and 

Disposition 22 (6), 909–915. 

Pearce, R.E., Vakkalagadda, G.R., Steven Leeder, J., 2002. Pathways 

of carbamazepine bioactivation in vitro I. Characterization of 

human cytochromes P450 responsible for the formation of 

2- and 3-hydroxylated metabolites. Drug Metabolism and 

Disposition 30 (11), 1170–1179. 

Pfluger, P., Dietrich, D.R., 2001. Effects on pharmaceuticals in the 

environmentdan overview and principle considerations. In: 

Kümmerer, K. (Ed.), Pharmaceuticals in the 

EnvironmentdSources, Fate, Effects and Risks. Springer, Berlin, 

pp. 11–17. 

Radjenovic, J., Petrovic, M., Barceló, D., 2007. Analysis of 

pharmaceuticals in wastewater and removal using 

a membrane bioreactor. Analytical and Bioanalytical 

Chemistry 387 (4), 1365–1377. 

Radjenovic, J., Sirtori, C., Petrovic, M., Barceló, D., Sixto, M., 2009. 

Solar photocatalytic degradation of persistent 

pharmaceuticals at pilot-scale: Kinetics and characterization 

of major intermediate products, Applied catalysis B: 

Environmental 89 (1–2), 255–264. 


Reemtsma, T., Weiss, S., Mueller, J., Petrovic, M., González, S., 

Barceló, D., Ventura, F., Knepper, T.P., 2006. Polar pollutants 

entry into the water cycle by municipal wastewater: 

a European perspective. Environmental Science and 

Technology 40 (17), 5451–5458. 

Sirés, I., Arias, C., Cabot, P.Ll., Centellas, F., Garrido, J.A., 

Rodríguez, R.M., Brillas, E., 2007. Degradation of clofibric acid 

in acidic aqueous medium by electro-Fenton and 

photoelectro-Fenton. Chemosphere 66 (9), 1660–1669. 

Song, W., Cooper, W.J., Mezyk, S.P., Greaves, J., Peake, B.M., 2008. 

Free radical destruction of b-blockers in aqueous solution. 


Stafiej, A., Pyrzynska, K., Regan, F., 2007. Determination of antiinflammatory 

drugs and estrogens in water by HPLC with UV 

detection. Journal of Separation Science 30 (7), 985–991. 


treatment plants and rivers. Water Research 32 (11), 3245– 

3260. 

Vogna, D., Marotta, R., Andreozzi, R., Napolitano, A., D’Ischia, M., 

2004. Kinetic and chemical assessment of the UV/H 2O 2 

treatment of antiepileptic drug carbamazepine. Chemosphere 

54 (4), 497–505. 

Zhang, Y., Geissen, S.U., Gal, C., 2008. Carbamazepine and 

diclofenac: removal in wastewater treatment plants and 

occurrence in water bodies. Chemosphere 73 (8), 1151–1161.

Loadings, trends, comparisons, and fate of achiral and 

chiral pharmaceuticals in wastewaters from urban tertiary 

and rural aerated lagoon treatments 

Sherri L. MacLeod a,1 , Charles S. Wong a,b, * 

a 

Department of Chemistry, University of Alberta, Edmonton, Alberta, T6G 2G2, Canada 

b 

Environmental Studies Program and Department of Chemistry, Richardson College for the Environment, University of Winnipeg, 

Winnipeg, Manitoba, R3B 2E9, Canada 

article info 







Keywords: 


Passive samplers 

Wastewater 

Loadings 

Trends 

Chiral drugs 


abstract 

Pharmaceutical occurrence, fate, and toxicity in aquatic 

ecosystems has been studied for more than 10 years (Daughton 

and Ternes, 1999; Halling-Sorensen et al., 1998; Kolpin 




A comparison of time-weighted average pharmaceutical concentrations, loadings and 

enantiomer fractions (EFs) was made among treated wastewater from one rural aerated 

lagoon and from two urban tertiary wastewater treatment plants (WWTPs) in Alberta, 

Canada. Passive samplers were deployed directly in treated effluent for nearly continuous 

monitoring of temporal trends between July 2007 and April 2008. In aerated lagoon effluent, 

concentrations of some drugs changed over time, with some higher concentrations in 

winter likely due to reduced attenuation from lower temperatures (e.g., less microbially 

mediated biotransformation) and reduced photolysis from ice cover over lagoons; however, 

concentrations of some drugs (e.g. antibiotics) may also be influenced by changing use 

patterns over the year. Winter loadings to receiving waters for the sum of all drugs were 

700 and 400 g/day from the two urban plants, compared with 4 g/day from the rural plant. 

Per capita loadings were similar amongst all plants. This result indicates that measured 

loadings, weighted by population served by WWTPs, are a good predictor of other effluent 

concentrations, even among different treatment types. Temporal changes in chiral drug 

EFs were observed in the effluent of aerated lagoons, and some differences in EF were 

found among WWTPs. This result suggests that there may be some variation of microbial 

biotransformation of drugs in WWTPs among plants and treatment types, and that the 

latter may be a good predictor of EF for some, but not all drugs. 


et al., 2002; Stamm et al., 2008). While drugs have been 

consistently detected in surface water, the effects of these 

complex mixtures of bioactive pollutants on non-target 

aquatic organisms are not fully understood (Fent et al., 2006). 

Accurate exposure scenarios are essential to assess toxicity 

Abbreviations: CR, Capital Region; DL, method detection limit; EF, enantiomer fraction; GB, Gold Bar; LLB, Lac La Biche; PEC, predicted 

environmental concentration; POCIS, polar organic chemical integrative sampler; SD, standard deviation; TWA, time-weighted average; 

WWTP, wastewater treatment plant. 

* Corresponding author. Richardson College for the Environment, University of Winnipeg, Winnipeg, Manitoba, R3B 2E9, Canada. 

E-mail addresses: smacleod@ualberta.ca (S.L. MacLeod), wong.charles.shiu@alum.mit.edu (C.S. Wong). 

1 Department of Chemistry, Université de Montréal, Montréal, Quebec, H3C 3J7, Canada. 


doi:10.1016/j.watres.2009.09.056

534 

and assign risk. Exposure, in turn, can be determined by either 

prediction or by measurement of wastewater or surface water 

concentrations. Prescription and sales information are not 

readily available, and may not be ideal for predicting environmental 

concentrations of some drugs (Coetsier et al., 2009) 

because of the uncertainty associated with use and metabolism 

(Letzel et al., 2009). In addition, removal of drugs from 

wastewater may range from 0 to 100% depending on the drug 

and the treatment process (Batt et al., 2007; Letzel et al., 2009, 

Lishman et al., 2006), even for the same compound such as 

diclofenac (Letzel et al., 2009), further complicating prediction 

of environmental concentrations. In Europe, where a tiered 

approach is taken for requirements of fate and effects studies, 

predicted environmental concentrations (PECs) are calculated 

as a worst-case scenario, without accounting for any metabolism, 

biodegradation, or retention by treatment (EMEA, 2006). 

For PEC values above 10 ng/L, higher tiered assessment is 

required (EMEA, 2006), and while in a French study, PECs were 

accurate for some drugs (carbamazepine, diclofenac, 

propranolol), local measured environmental concentrations 

of others were found to be much lower than those predicted, 

even very close (10 m) to effluent outfalls (Coetsier et al., 2009). 

Wastewater and surface water concentrations may be 

heavily affected by abiotic transformation processes such as 

photolysis, which depends on light intensity over the seasons 

(Vieno et al., 2005); as well as biotransformation, which would 

depend on microbial activity and on temperature. The latter is 

particularly significant for the numerous chiral drugs. As with 

other chiral pollutants (Wong, 2006), the enantiomers of 

pharmaceuticals are affected identically by abiotic processes, 

but may be affected differently by biologically mediated 

processes in wastewater treatment (Fono and Sedlak, 2005; 

MacLeod, 2009; MacLeod et al., 2007b; Nikolai et al., 2006). Drug 

enantiomers may also exhibit differential toxicity to aquatic 

life (Stanley et al., 2006). Thus, a robust means by which to 

predict environmental loadings and fate of drugs from wastewater 

effluents is valuable for exposure and risk assessment. 

Direct measurement of drug loadings from wastewater 

treatment plants (WWTPs) has proven accurate for prediction 

of drug concentrations in receiving waters (Letzel et al., 2009; 

MacLeod, 2009). Such measurements can be accomplished via 

repeated grab sampling (Sacher et al., 2008), which is tedious 

and time-consuming, or continuous flow-proportional effluent 

collection (Letzel et al., 2009). However, passive sampling 

devices such as the Polar Organic Chemical Integrative Sampler 

(POCIS) (Alvarez et al., 2004) can provide a simple means of 

obtaining time-weighted average (TWA) loadings of polar 

contaminants, such as drugs, over longer time periods with 

lower detection limits (Jones-Lepp et al., 2004; MacLeod et al., 

2007a; Zhang et al., 2008). While a high-frequency, discrete 

sampling approach may capture the real-time heterogeneity of 

contaminant loadings, passive samplers have the advantage of 

providing integrated, continuous sample collection, and are 

thus useful as an alternative or complementary sampling 

approach for dissolved phase chemicals. 

Most research on drugs in the environment has, thus far, 

focused on measurement of drugs in the waste or receiving 

waters from larger (10,000 to millions) population centers 

(Carballa et al., 2007; Chen et al., 2006; Hua et al., 2006; Kimura 

et al., 2007; Lajeunesse et al., 2008; Letzel et al., 2009; Lissemore 


et al., 2006; Loraine and Pettigrove, 2006; Managaki et al., 2007; 

Sacher et al., 2008; Vieno et al., 2005), with little attention paid 

to the drug output from the wastewater of smaller (

communities. Treated effluent from LLB is discharged 

continuously to Field Lake which discharges into Lac La Biche 

(the lake) via Red Deer Brook, while GB and CR both discharge 

treated effluent to the North Saskatchewan River and we 

have previously measured drugs in both of these receiving 

waters (MacLeod, 2009; MacLeod et al., 2007a). To our 

knowledge, this is the first study to use POCIS to undertake 

long-term, nearly continuous, monitoring of drugs, particularly 

chiral drugs, in municipal effluents from a small population 

center. 


2.1. Materials 

We used POCIS (Environmental Sampling Technologies, 

St. Joseph, MO) in the ‘pharmaceutical’ configuration: 200 mg 

Oasis HLB sorbent (Waters, Milford, MA) between two polyethersulfone 

membranes (45.8 cm 2 total surface area) held 

together by stainless steel rings and attached to a spindle 

holder. This configuration was used as POCIS sampling rates 

are available for the target analytes (Tables 1 and S2), all 

commonly used drugs. Chemical analytes (>98% purity) and 

used as received (sources in Table S2). Solvents and reagents 

for chromatography (Fisher Scientific, Ottawa, Canada) were 

of HPLC grade; nanopure water (18 MU cm) was supplied by 

a Nanopure Ultrapure system (Barnsteam/Thermolyne, 

Dubuque, IA). Details are published elsewhere (MacLeod, 2009; 

MacLeod et al., 2007a, b). 

Table 1 – Pooled winter time-weighted average 

concentrations for drugs detected at all three WWTPs in 

each of the three deployment periods. 

Analyte Gold 

Bar 

(ng/L) SD 

Capital 

Region 

(ng/L) SD 

Lac la 

Biche 

(ng/L) SD 

Atenolol 72 270 120 440 36 160 

Carbamazepine 200 130 290 190 63 39 

Celecoxib 17 8 28 13 22 10 

Citalopram 23 5 67 8 5.7 1.7 

Clarithromycin 220 140 420 270 120 90 

Codeine 450 380 920 710 990 710 

Diclofenac 1200 680 2,000 1300 460 280 

Erythromycin 13 10 19 15 3.1 2.8 

Gemfibrozil 150 59 190 72 72 34 

Metoprolol 58 47 65 52 29 23 

Naproxen 66 58 530 510 280 250 

Paroxetine 0.7 0.9 1.4 1.8 0.2 0.2 

Propranolol 4.7 3.0 14 8.5 1.4 1.1 

Sotalol 27 17 66 22 16 11 

Temazepam 130 71 160 76 120 60 

Triclosan 1.5 2.3 53 69 6.9 8.7 

Trimethoprim 38 39 65 70 10 12 

Sum of drugs 

(mg/L) 

2.2 5.0 2.2 

Pooled values calculated from n ¼ 6 for Gold Bar and Capital Region 

WWTPs, n ¼ 9 for Lac la Biche WWTP. 


2.2. Passive sampling in WWTP effluent 

Sampling of WWTP effluent took place between July 2007 and 

April 2008 (Table S1). 

During each sampling period at the CR and LLB WWTPs, 

two and three POCIS, respectively, were protected in perforated 

stainless steel cages secured with nylon rope, and 

deployed in final effluent. During each sampling period at GB 

WWTP, two caged samplers were deployed in a tank into 

which treated effluent was continuously pumped. Field blanks 

were brought to all sites at deployment and retrieval, at which 

time POCIS were individually put on ice until return to the 

laboratory, where they were stored at 20 C until chemical 

extraction. 

2.3. Chemical extraction and instrumental analysis 

Details on chemical extraction and instrumental analysis are 

provided elsewhere (MacLeod, 2009; MacLeod et al., 2007a, b) 

and in Supplemental Information. In brief, POCIS sorbent was 

washed into a glass column, and methanol used to elute target 

analytes. The extract was filtered, reduced to 5–10 mL via 

rotary evaporation, filtered through 0.22 mm Acrodisc polytetrafluoroethylene 

syringe filters (Pall Life Sciences, Ann 

Arbor, MI, USA), and nitrogen-evaporated to dryness. Internal 

standards (Table S2) were added, and extracts were reconstituted 

to 1 mL with methanol. Achiral chromatography for 

determining concentrations (MacLeod et al., 2007a) was 

performed with an Ultra C18 reversed-phase column 

(150 mm 4.6 mm internal diameter (i.d.) 5 mm particle size 

(dp), Restek, Bellefonte, PA, USA), while enantioselective 

chromatography for determining enantiomer compositions 

was performed with a Chirobiotic V column (250 mm 4.6 mm 

i.d. 5 mm d p, Advanced Separation Technologies, Whippany, 

NJ, USA) for all chiral analytes except temazepam, for which 

a Chiralpak AD-RH column (MacLeod, 2009) was used 

(150 mm 4.6 mm i.d. 5 mm d p, Daicel Chemical Industries, 

West Chester, PA). Analytes were detected by electrospray 

ionization-tandem mass spectrometry using an Applied Biosystems 

QTrap triple-quadruple instrument (Foster City, CA) in 

multiple reaction monitoring mode. 

2.4. Quality assurance/quality control and 

data handling 

Detection limits (DLs, Table S2) were determined as the 

analyte concentration in 1 mL POCIS extract providing 

a signal-to-noise ratio of 3, and ranged from 0.001 to 1.6 ng/ 

POCIS. Drugs were not detected in any blanks. Calculated 

time-weighted average concentrations and loadings (hereafter 

called ‘‘concentrations’’ and ‘‘loadings’’, respectively) are 

presented as mean standard deviation (SD). Data analysis 

was carried out as previously described elsewhere (MacLeod, 

2009; MacLeod et al., 2007a, b). Statistical comparisons 

between two groups were carried out by t-test, while three or 

more groups were compared by ANOVA, with either Tukey’s 

test for multiple comparisons or with Dunnett’s test for 

comparison with a control, as indicated. Differences noted are 

statistically significant (all p < 0.05). As expected, analytes for 

which the POCIS sampling rate was known with higher

536 

precision (Table S2) were generally more conducive to statistical 

significance testing. 

For chiral drugs atenolol, citalopram, fluoxetine, metoprolol, 

nadolol, pindolol, propranolol, salbutamol, sotalol, and 

temazepam, enantiomer fractions (EFs) described enantiomer 

composition (Harner et al., 2000): 

EF ¼ 

E1 

E1 þ E2 

ðþÞ 

¼ 

ðþÞþð Þ 

where E 1 and E 2 are the first and second eluted enantiomers, 

respectively, when the enantiomer optical rotation was 

unknown (i.e., all analytes except atenolol, propranolol, and 

fluoxetine). Further information on data quality and handling 

is available in Supplementary Data. 


3.1. Time-weighted average pharmaceutical 

concentrations in treated effluents 

3.1.1. Aerated lagoons: Lac la Biche 

Most studies to date have relied on grab or short-term 

composite sampling to assess temporal trends in pharmaceutical 

concentrations in effluents (Batt et al., 2007; Chen et al, 

2006; Hua et al., 2006; Loraine and Pettigrove 2006, Stülten et al., 

2008; Vieno et al., 2005). However, temporal trends assumed 

from grab sampling may not be representative of longer time 

periods. Because POCIS can provide longer and continuous 

temporal integration, it may provide more accurate assessment 

of temporal trends. However, passive samplers cannot 

provide reliable information on concentration fluctuations 

over time scales shorter than sample integrative periods, 

which can result in very different conclusions about temporal 

changes in concentration over time (Shaw and Mueller, 2009). 

Drug concentrations at LLB (Fig. 2) measured over six 

sampling periods of 35–51 days (Table S1) ranged from 0.07 ng/L 

(paroxetine, propranolol) to 1100 ng/L (codeine). Fenoprofen, 

fluoxetine, ketoprofen, omeprazole, roxithromycin, sildenafil, 

sulfadimethoxine, sulfamethazine, sulfapyridine, sulfisoxazole, 

tadalafil and vardenafil were not detected in LLB treated 

effluent above their respective DLs in any sampling period 

(Table S2). Concentrations of atenolol and propranolol were 

similar to effluent grab sampled from LLB in September 2005 

(Nikolai et al., 2006); however, metoprolol was higher (200 ng/L). 

Diclofenac and naproxen concentrations at LLB were also 

similar to that in effluents of the activated sludge WWTP of 

Aura, Finland, which has a similar population (4000) to LLB 

(Vieno et al., 2005). Concentrations were also generally similar 

to those in an aerated lagoon WWTP in Louisiana, USA (Conkle 

et al., 2008) for drugs common to both studies: atenolol, carbamazepine, 

gemfibrozil, metoprolol, naproxen, and sotalol. 

Thus, our TWA concentrations are likely representative of 

those expected in small community WWTP effluents. 

There were statistically significant temporal concentration 

changes (Fig. 2) for celecoxib, citalopram, clarithromycin, 

codeine, diclofenac, erythromycin, gemfibrozil, naproxen, 

propranolol, sotalol, and temazepam. Temporal concentration 

trends were similar for clarithromycin, codeine, diclofenac, 


(1) 

erythromycin, naproxen and propranolol, with higher 

concentrations during winter sampling periods (December to 

April) than at other times (July to December, Fig. 2). For atenolol, 

carbamazepine, metoprolol, triclosan and trimethoprim, 

the effluent concentrations did not change significantly 

over the sampling periods (Fig. 2). For drugs with temporal 

differences in concentration, the changes in concentration 

ranged from 2 (propranolol) to 960 ng/L (codeine), whereas 

drugs without statistically significant temporal differences 

had smaller concentration changes over time, from 8 (triclosan) 

to 85 ng/L (atenolol). 

While the average daily volume of treated effluent 

discharged decreased over the six sampling periods (Table S1), 

this decrease was not statistically significant. Thus, TWA 

loadings are reflective of concentrations. Of the drugs studied, 

photodegradation likely plays a major role in the dissipation 

of diclofenac (Andreozzi et al., 2003), naproxen (Lin et al., 2006; 

Lin and Reinhard, 2005), and propranolol (Andreozzi et al., 

2003) in aquatic systems. The higher effluent concentrations 

of those drugs in winter (Fig. 2) may be at least partially the 

result of decreased photolysis, due to ice cover on lagoons 

from November to March (Siebold, 2008) blocking sunlight 

from the water column (Vieno et al., 2005), and fewer hours of 

sunlight during those months (MacLeod, 2009; Vieno et al., 

2005). It is impossible to ascertain the amount of photodegradation 

in the lagoon during ice-free periods with our 

data. While the LLB lagoon wastewater was not clear in color, 

it is likely that photodegradation of drugs took place at least at 

the lagoon surface during ice-free periods, and continuous 

aeration of lagoon waters for oxygenation purposes would 

bring water at depth to the surface to provide some photodegradation 

treatment. Lower winter temperatures also likely 

influenced elimination of pharmaceuticals, through lower 

reaction rate constants for abiotic transformation, and 

reduced microbial activity for biotransformation (Kim and 

Carlson, 2007). Effluent pH was 8 0.5 over the six sampling 

periods and effluent temperatures ( C) were 21 3, 16 2, 

4 3, 1 0, 1.4 0.5, 4 2, chronologically for the six 

sampling periods (Siebold, 2008). The concentrations of those 

three drugs decreased again in March/April, as temperatures 

and number of daylight hours increased, and ice cover was 

reduced or absent on the lagoons (Fig. 2). Since naproxen can 

be readily biodegraded in sewage (Yu et al., 2006), decreased 

biodegradation may also contribute to the higher effluent 

concentrations for naproxen in winter. 

It is possible that changes in temperature may affect 

calculated TWA concentrations, as sampling rates used in 

this study were determined at room temperature (22–28 C, 

MacLeod et al., 2007a), and WWTP temperatures during 

colder months were considerably less. However, we cannot 

presently quantify the compound-dependent effect of such 

changes on our observed trends. Lower temperatures would 

result in lower analyte aqueous diffusion coefficients and 

hence lower sampling rates, estimated to be roughly 75% 

over a 20 C ambient temperature change (MacLeod et al., 

2007a). When applied to water colder than under the calibration 

condition, sampling rates would then provide an 

underestimate of the calculated TWA concentrations, the 

effect of which would be compound-dependent. Such an 

effect may be pronounced in the case of drugs with higher

concentrations in winter than in other times (e.g., antibiotics 

such as clarithromycin and erythromycin, Fig. 2). Temporal 

effects on concentrations were less for other analytes, and 

may either be insignificant (e.g., triclosan and gemfibrozil) or 

of little practical significance even if statistically significant, 

given the relatively narrow range (e.g., from a factor of two, 

to an order of magnitude) in concentrations over the seasons 

studied (Fig. 2). 


Fig. 2 – Time-weighted average concentrations (ng/L ± SD; ) and enantiomer fractions (EF) of chiral drugs (,) in LLB treated 

wastewater during six sampling periods from July 2007 to August 2008 (Table S1). Hash marks on x-axis indicate break in 

continuous sampling between September and October 2008. Significant concentration differences are indicated by lowercase 

letters while EF differences are indicated by upper-case letters. Plots without letters do not have concentrations or EFs, 

as appropriate, statistically different from one another over sampling periods. Non-racemic EFs are indicated by an asterisk, 

and EFs that were corrected for enantiomer-specific matrix effects are indicated by ^. RS, racemic standards. Error bars (SD) 

represent propagated error from replicate POCIS, analytical processing and instrumental uncertainty, and error associated 

with POCIS sampling rates. 

Without measurements of influent concentration, we 

cannot discern whether temporal concentrations in effluent 

are more influenced by changes in removal efficiency, as 

suggested as likely by many studies (Hua et al., 2006; Kim and 

Carlson, 2007; Loraine and Pettigrove, 2006; Sacher et al., 2008; 

Vieno et al., 2005) or in use patterns over time, which may be 

the case for antibiotics. In a Swiss study (McArdell et al., 2003), 

higher macrolide concentrations were found in winter than in

538 

other periods, consistent with our observations for the macrolides 

clarithromycin and erythromycin (Fig. 2). In the Swiss 

study, as with this study, raw wastewater concentrations 

were not measured so removal efficiency could not be 

assessed. However, the Swiss authors cited a personal 

communication which indicated that macrolide sales were 

higher in winter, indicating that increased use may contribute 

to higher environmental concentrations. 

The winter concentrations in LLB treated effluent, which 

were sampled mostly at the same time as those at GB and CR 

(see Section 3.1.3), changed over time for some but not all of 

our target analytes (Fig. 2). Concentration changes were less 

prevalent among the three winter sampling months, 

compared with other periods. Among the sampling periods 

between December and April (Tukey’s test), there were 

statistically significant changes in effluent concentration for 

only three drugs. The celecoxib concentration in December/ 

January was three times that in January/March and twice that 

in March/April with magnitudes of difference of 20 and 15 ng/L, 

respectively. Sotalol had December/January concentrations 

were eight times that in both January/March and March/April 

and magnitudes of 33 ng/L for both comparisons. While the 

concentration of citalopram in January/March was 50% higher 

than that in March/April, the magnitude of this difference was 

only 2 ng/L. The general lack of effluent concentration change 

over the winter indicates that there is no change in drug use 

patterns and/or in the ability of aerated lagoons to remove 

drugs effectively over this season. 

It is important to note that concentration changes in 

WWTPs serving small communities may well be sensitive to 

changes in use patterns of the population served. If a drug was 

completely excreted from the body in unchanged form and 

did not undergo any removal during wastewater treatment, 

then the addition of a single additional individual’s twice-aday, 

100 mg active ingredient dose at LLB would result in an 

effluent concentration change of 100 ng/L, which would rival 

existing concentrations at LLB (Fig. 2)! In actuality, only 

a fraction of most drugs are excreted as parent compounds, 

and WWTP treatment provides some elimination. For 

example, 3% of carbamazepine is excreted unmetabolized 

from humans, with 30% removal based on a survey of Canadian 

WWTPs (Miao et al., 2005). Assuming these conditions 

hold for LLB, then a 2 ng/L change in effluent concentration, or 

a 3–10% increase (Fig. 2), would result from the same dosage. 

However, without specific usage information, excretion data, 

and WWTP-specific removal efficiencies, we cannot make 

more specific predictions of the sensitivity of drug concentrations 

in wastewater on usage in small communities. 

3.1.2. Tertiary treatment: Gold Bar and Capital Region 

The concentration of detectable drugs (Table 1, Table S3) in GB 

effluent ranged from 0.03 (omeprazole) to 1350 ng/L (diclofenac) 

while that in CR ranged from 0.09 (roxithromycin) to 

2520 ng/L (diclofenac). These concentrations are similar to 

grab sample measurements for our analytes in other Canadian 

WWTP effluents (Hua et al., 2006; Lishman et al., 2006; 

Miao et al., 2005), indicating that our passive sampler derived 

TWA concentrations were comparable with previous work. 

However, concentrations of chiral drugs in this study were 

generally lower by a factor of at least 10 compared with a grab 


sample from GB in April 2007 (Macleod et al., 2007b), suggesting 

possible annual or other temporal changes in 

concentrations as discussed below. 

Since only two samplers were deployed in each of GB and 

CR during each sampling period (Table S1, Table S3), temporal 

trends could not be statistically assessed. However, few 

temporal trends were evident. In GB effluent, the temazepam 

concentration in March was double (range >100 ng/L) that in 

December/January, and in CR effluent, the naproxen concentration 

increased seven-fold over the winter (range >700 ng/L). 

Also in CR effluent, the roxithromycin concentration in 

December/January was nine times higher than that in both 

February/March and March (range 0.8 ng/L), whereas the 

opposite trend was seen for sulfamethazine (increased three 

times over time, range 3 ng/L). Fenoprofen, ketoprofen, sulfadimethoxine, 

sulfisoxazole, and tadalafil were not detected 

in either urban WWTP above their respective DLs. 

When the TWA concentrations for each drug were pooled 

for each WWTP (n ¼ 6 for each WWTP), statistical comparisons 

revealed no significant differences between urban 

effluent concentrations for any drug (Table 1). As with the 

aerated lagoon effluent, there was little evidence of changes in 

drug concentration over the monthly integrated sampling 

periods. Seasonality in removal efficiency and/or or drug use 

patterns could not be assessed for the tertiary plants as only 

winter data was collected. The concentrations of each drug 

were similar between the two tertiary plants (Table 1), likely 

a reflection of similar per capita drug and water use between 

the populations since treatment processes are nearly identical 

in the two plants (Letzel et al., 2009). We do not expect 

concentrations in these large-community WWTPs to be 

influenced significantly by use changes in a small number of 

individuals; using the same assumptions as in the previous 

section, the same individual’s carbamazepine dose change 

would result in a 0.02 ng/L increase in GB concentrations, an 

insignificant amount. 

3.1.3. Comparison of concentrations in aerated lagoon and 

tertiary treated effluents in winter 

Drug concentrations in effluents depend on a number of 

factors. Removal efficiency during treatment is known to vary 

by drug and by WWTP (Batt et al., 2007; Letzel et al., 2009). 

Differences in sewage retention time (Batt et al., 2007; Göbel 

et al., 2007; Kimura et al., 2007), and/or the co-metabolic 

activity of microbes (Joss et al., 2006; Kim and Carlson, 2007; 

Kimura et al., 2007; Yu et al., 2006) can produce differences in 

the efficacy of wastewater treatment for removal of some 

drugs. For other chemicals, sorption capacity, rather than 

retention time or biodegradation, can be most important for 

removal during treatment (Carballa et al., 2007; Joss et al., 

2006). Differences in sunlight attenuation while wastewater is 

in outdoor tanks (tertiary) and lagoons may also play a role for 

some drugs, as the turbidity of the wastewater may vary, thus 

varying the amount of sunlight attenuation and thus the 

amount available for photolysis. Drugs in GB and CR were 

exposed to sunlight for less than 24 h, and received at least 8 s 

of low power UV exposure (City of Edmonton, 2009) for 

disinfection of effluent before release. The drugs in LLB are 

exposed to sunlight for as long as the daylight hours of 90 days 

(Duigou, 2006), providing more time for photolysis.

Temperature can also play a role, as previously noted. In 

addition, both GB and CR use secondary activated sludge 

treatment. These tanks are warmed in the winter to maintain 

efficiency, especially in cold climates such as that in Alberta. 

Thus, the water temperatures in those tanks may be higher 

than that in the LLB lagoon pond under ice (w4 C). As a result, 

biological activity capable of biodegrading drugs may be more 

active in GB and CR compared with LLB. 

Notwithstanding these potential differences in incidental 

removal during treatment, the concentrations of several drugs 

detected at all three plants were generally comparable, with 

CR having the highest total concentration (Table 1). Most 

drugs (17 out of 24 detectable) were found in all three WWTPs 

in winter (Table 1). No differences in winter pooled TWA 

concentrations (three winter sampling periods with n ¼ 9 for 

LLB, and n ¼ 6 for GB and for CR) among the three effluents for 

atenolol, celecoxib, codeine, metoprolol, naproxen, paroxetine, 

temazepam, triclosan and trimethoprim. For eight other 

drugs, the concentrations in urban effluents were from 2 to 12 

times higher than that in rural effluent, with concentration 

differences that ranged from 9 to 1600 ng/L (Table 1). For citalopram 

and gemfibrozil, the concentrations at GB were 

greater than those at LLB. For carbamazepine, citalopram, 

clarithromycin, diclofenac, erythromycin, gemfibrozil, 

propranolol, and sotalol, the concentrations in CR effluent 

were greater than those in LLB effluent. Between the two 

urban WWTPs, the concentrations of citalopram, propranolol 

and sotalol at CR were greater than those at GB. Several drugs 

(fluoxetine, paroxetine, roxithromycin, sildenafil, sulfamethazine, 

sulfapyridine, vardenafil), infrequently or not detected 

in the rural WWTP effluent, were consistently detected in 

the urban WWTP effluents (Tables S2–S3). These drugs were 

generally present at low concentrations (100 ng/L). The discrepancy with sulfapyridine may be 

explained by differences in drug use among the communities, 

but is more likely caused by treatment differences. Sulfapyridine 

is generally administered via sulfasalazine, which is 

cleaved in the colon to release sulfapyridine (10–35%), while 

another 30–50% is released as sulfasalazine and a metabolite 

(Neumann, 1989), both of which may be cleaved in biological 

wastewater treatment to release sulfapyridine (Göbel et al., 

2005). It is possible that such cleavage does not occur in the 

aerated lagoons, leading to a lower concentration of sulfapyridine 

in LLB effluent. For the other drugs, differences 

reflect differences between the communities in terms of water 

and/or drug use, and/or a difference in the fate of drugs in the 

WWTPs (Hua et al., 2006; Kim and Carlson, 2007; Loraine and 

Pettigrove, 2006; Sacher et al., 2008; Vieno et al., 2005). 

3.2. Loadings of pharmaceuticals to surface waters from 

aerated lagoon and tertiary treated effluents 

Average daily loadings (Table S4) in winter were calculated for 

each plant by multiplying the average daily water discharge by 

the TWA concentration of each drug during each sampling 

period. Total winter loadings for GB, CR, and LLB were 700, 400 

and 4 g/day, respectively. For the winter sampling periods, 

average daily loadings from GB were generally twice that 

from CR. Average daily loadings from LLB were on the order 


of mg/day. This was at least an order of magnitude lower than 

loadings from GB and CR, as expected based on the populations 

served (Letzel et al., 2009). The above result suggests 

that population-weighted per capita TWA daily loadings 

would be similar among the three WWTPs. Indeed, this was 

the case for most drugs, for which there were no statistically 

significant differences among population-weighted average 

daily loadings (Fig. 3), e.g., atenolol, celecoxib, clarithromycin, 

codeine, gemfibrozil, metoprolol, naproxen, temazepam, 

triclosan, and trimethoprim. Urban per capita average daily 

loadings were greater than those from the rural WWTP in 

a few cases: CR was greater than LLB (carbamazepine, 

citalopram, diclofenac, propranolol, erythromycin, sotalol), 

GB was greater than LLB (citalopram), while the per capita 

loading from CR was greater than GB for three drugs 

(citalopram, propranolol, sotalol). These per capita loading 

differences, although statistically significant, may not be 

practically significant, as all were all less than an order of 

magnitude (two to eight times) and the absolute differences 

ranged from 2 (propranolol, CR > GB) to 390 mg/person per day 

(diclofenac, CR > LLB). All of these differences in loadings 

between WWTPs were also noted as differences in concentration. 

Loading differences were not found for clarithromycin 

and gemfibrozil, for which differences in concentration were 

found among WWTPs. Thus, we conclude that population is 

a good predictor of human-use drug release into receiving 

waters, at least for over-the-counter or long-term prescription 

drugs such as those in this study. 

Such predictions are likely to be more accurate when based 

on direct measurements, particularly from demographically 

similar regions (Stamm et al., 2008), and most accurate when 

also based on comparable wastewater treatment. In a ten-year 

study of drugs in the Rhine, concentrations of diclofenac and 

carbamazepine were found to be relatively constant and given 

consistency in use patterns, and although raw wastewater 

concentrations were not measured, the authors attributed any 

temporal variations to differences in WWTP removal 

efficiency (Sacher et al., 2008). Their conclusion was that any 

measures that had been implemented to reduce loadings were 

thus far ineffective. Because our estimates are based on 

continuous measurements of pharmaceutical concentrations 

(Letzel et al., 2009; Miao et al., 2005; Stamm et al., 2008), rather 

than instantaneous grab sampling, they are likely to be more 

robust as TWA concentrations from passive samplers are less 

susceptible to short-term fluctuations in concentration over 

periods shorter than the sampler deployment periods. Efforts 

have been undertaken to find a suitable mechanism for 

removal of drugs from the waste stream (Ikehata et al., 2008), 

and such efforts should continue, because as our results 

further highlight, even sophisticated WWTPs like GB and CR 

produce wastewater with drug concentrations that are 

comparable with water treated by aerated lagoons. 

As previously noted, loadings at WWTPs serving small 

communities are potentially more sensitive to small changes 

in use patterns than larger communities. The twice-a-day, 

100 mg/dose from one individual at LLB would result in 

a loading increase of 50 mg/person per day at 100% excretion 

and no WWTP removal, or 1 mg/person per day (4% increase, 

Fig. 3) using the carbamazepine assumptions previously 

given. At GB, the equivalent increase in loadings would

540 


Fig. 3 – Time-weighted average per capita daily loadings (:, mg/person per day) and enantiomer fractions (,, EF), both ± SD, 

of drugs at three WWTPs: GB, Gold Bar; CR, Capital Region (n [ 6), LLB, Lac La Biche. Differences in daily loadings are 

indicated by lower-case letters while EF Statistically significant differences are indicated by upper-case letters. Plots without 

letters do not have loadings or EFs, statistically different from one another over sampling periods. Asterisks indicate EFs 

that are significantly non-racemic compared with racemic standards. Error bars (SD) represent propagated error from 

replicate POCIS, analytical processing and instrumental uncertainty, and error associated with POCIS sampling rates.

generally be insignificant (Fig. 3), at 0.27 and 0.006 mg/person 

per day, respectively. 

3.3. Enantiomer fractions of chiral drugs in 

treated effluents 

3.3.1. Temporal trends in enantiomer fractions of chiral drugs 

in aerated lagoon effluent 

We measured EFs in LLB to discern whether temporal changes 

existed in enantiomer composition of chiral drugs released by 

WWTPs. Changes in EF may arise from differential interaction 

of chiral drug enantiomers with other chiral chemicals (e.g., 

enzymes). These may result from temporal differences in 

enantiomer-specific metabolism of patients (Mehvar and 

Brocks, 2001; Schmekel et al., 1999), and/or from enantiomerspecific, 

microbial enzyme mediated biotransformations 

during wastewater treatment, as previously noted for atenolol 

(MacLeod et al., 2007b; Nikolai et al., 2006), citalopram 

(MacLeod et al., 2007b) and propranolol (Fono and Sedlak, 

2005; MacLeod et al., 2007b; Nikolai et al., 2006), but not 

metoprolol (Fono and Sedlak, 2005; MacLeod et al., 2007b; 

Nikolai et al., 2006), salbutamol (MacLeod et al., 2007b), or 

sotalol (MacLeod et al., 2007b). Because of enantiomer-specific 

metabolism by the target organism, chiral pharmaceuticals 

may not necessarily enter the environment with the same 

enantiomeric composition as the dose. This is in contrast to 

chiral persistent organic pollutants, which generally enter the 

environment as a racemate, or sometimes as a single enantiomer. 

Thus, changes in enantiomeric composition from the 

excreted EF, not just from a racemic value of 0.5, would be 

environmentally significant as such a change would indicate 

post-release biochemical weathering. However, given that 

few environmental measurements of drug EFs exist to date, 

our discussion here focuses on changes in EF from the racemate, 

the most likely enantiomeric composition of formulated 

chiral drugs not sold as single enantiomers (e.g., naproxen). 

With the exception of tempazepam, chiral drugs were 

generally non-racemic in LLB effluent (Fig. 2) compared with 

racemic standards (Dunnett’s test). Atenolol was enriched in 

( )-atenolol during some sampling periods, with changes over 

time, consistent with previous observations in 2005 at LLB 

(Nikolai et al., 2006). Citalopram was mainly enriched in the 

first-eluted enantiomer except during August/September. 

Metoprolol was enriched in the E2-enantiomer when nonracemic, 

which was not previously observed at LLB in individual 

grab samples (Nikolai et al., 2006). Salbutamol was 

always enriched in the second eluted enantiomer, except 

during January/March wherein it was E1-salbutamol that was 

enriched. Sotalol was always non-racemic, with no change in 

EF over time. Temazepam was racemic with one exception, 

wherein E 2-temazepam was enriched. Propranolol EFs could 

not be measured (supplemental data). 

Temporal changes in EF were observed (Tukey’s test) for all 

drugs except sotalol in LLB effluent (Fig. 2, Table S5). Without 

direct measurement of the EF in influent and during the 

treatment process, we cannot conclusively determine 

whether EFs in effluent arose from metabolism prior to bodily 

excretion or from processes during treatment. The latter may 

be plausible, given changes in activity and composition of 

microbial consortia during biological sewage treatment from 


temperature changes over the seasons. For drugs such as 

citalopram and salbutamol, changes in EF may also reflect 

differences in availability or use of single-enantiomer 

formulations (Maier et al., 2001). Because of the temporal 

changes in EF observed, our results show that an EF measured 

at one time point may not necessarily be characteristic of the 

enantiomer composition of chiral drugs discharged from 

a WWTP, and hence insufficient to characterize exposures in 

receiving waters. 

3.3.2. Comparison of chiral drug EFs from aerated lagoons 

and tertiary treated wastewater 

Winter EFs from LLB and from the urban tertiary WWTPs were 

compared to gain insight on whether EFs could be predicted by 

treatment type, and may possibly be caused by similar human 

metabolism and release and/or microbial degradation during 

WWTP treatment. Due to low sample sizes, statistical tests and 

trend analyses were not carried out on the temporal EF data 

from the urban tertiary WWTPs. Instead, winter EFs were pooled 

for each drug from GB and CR and compared with EF standard 

(Dunnett’s test), to each other (t-test, with Welch’s correction for 

unequal variances as needed), and to the pooled EFs from the 

three winter sampling periods at LLB (Tukey’s test). Standard 

deviations in pooled data reflect propagated variability from 

both temporal changes in EFsample and from replicates. 

The two tertiary WWTP effluents had similar EFs during 

winter for all chiral drugs except sotalol (Fig. 3), suggesting 

that both treatment plants were processing those drugs in the 

same stereoselective manner during microbial treatment. 

At all three WWTPs (Fig. 3), metoprolol EFs were racemic in 

effluents, while sotalol was enriched in the E 2-enantiomer and 

differed in EF between the tertiary effluents. Racemic metoprolol 

was previously observed in both GB and LLB effluents 

(MacLeod, 2009; MacLeod et al., 2007b; Nikolai et al., 2006) and 

in the effluent of a WWTP in Dallas TX (Fono et al., 2006), with 

no change in EF from raw to treated wastewater (MacLeod 

et al., 2007b). After correction for enantiomer-specific matrix 

effects (MacLeod et al., 2007b), sotalol was previously found to 

be racemic in GB effluent (MacLeod et al., 2007b) via grab 

samples. This observation suggests that changes in sotalol EFs 

might have occurred over time in GB effluent; however, we 

cannot unequivocally conclude that this must be the case 

given our earlier caveats that a single EF cannot necessarily 

characterize the enantiomer composition. No statistically 

significant differences were observed for salbutamol EFs in 

the three effluents, despite enrichment of E 2-salbutamol in 

both tertiary effluents. This observation is largely a result of 

the high variability in salbutamol EFs in LLB and CR. In turn, 

the variability may be a function of the drug’s input to 

WWTPs, given that salbutamol is available as an R-salbutamol-only 

formulation, while S-salbutamol is metabolized 

slower than its antipode (Schmekel et al., 1999). Citalopram 

(Fig. 3) was enriched in the E 1-enantiomer in all effluents, with 

differences in EF between tertiary and aerated lagoon output. 

As with salbutamol, this difference could be influenced 

by differences in prescription patterns, given the availability 

of escitalopram, the S-enantiomer formulation. Citalopram 

was previously found with similar EFs (0.55–0.65) in GB 

effluent (MacLeod et al., 2007b). Temazepam was enriched in 

the E2-enantiomer at both GB and LLB, but racemic in CR,

542 

wherein its EF was different from that in LLB, but not from that 

in GB. No information regarding temazepam EFs in treated 

effluent is available, except our previous results (MacLeod, 

2009). For atenolol, all effluents had EFs that were different 

from one another, with GB containing racemic atenolol, CR 

containing more (þ)-atenolol and LLB containing more 

( )-atenolol. The EF of atenolol from healthy renal excretion is 

0.476 (Mehvar and Brocks, 2001), and 90% of atenolol is 

excreted unchanged in urine (Bourne, 1981). Thus, a process 

other than healthy renal excretion, and potentially specific to 

microbial communities of WWTPs, was influencing atenolol 

EFs in effluent. Likewise, the differences in EFs observed 

among WWTP effluents is likely due, at least in part, to variations 

in enantioselective microbial biotransformation of 

drugs during wastewater treatment. 

As we have noted, it is not clear what the enantiomeric 

composition of drugs may be upon excretion from humans, 

but non-racemic release is likely. As such, we must delineate 

between environmental biotransformation and in changes in 

enantiomeric composition of the source material in assessing 

the environmental fate of drug enantiomers, as has been done 

for chiral pesticides (e.g., metolachlor) also released in nonracemic 

amounts (Kurt-Karakus et al., 2008). 


Although other studies have used POCIS to monitor drugs in 

treated wastewater, none have done so for more than 30 days 

(Jones-Lepp et al., 2004; Zhang et al., 2008). Herein, we monitored 

LLB effluent for 271 days and found that the concentration 

of drugs in effluent varied over seasonal time scales. Other 

studies have mainly focused on larger urban centers. By 

comparing drugs in both rural and urban wastewater effluents, 

we find that pharmaceutical loadings to aquatic ecosystems 

scale with population. However, we also found that the 

concentrations of drugs in effluent from WWTPs with biological 

activated sludge treatment and UV disinfection were 

comparable with concentrations in effluent from simpler 

treatment methods like aerated lagoons. Although absolute 

loadings were generally lower, drugs were still detectable in 

effluent from small communities. Thus, if pharmaceutical 

contamination of surface water presents a risk to aquatic life, 

that risk is also present near smaller centers, especially those 

which are landlocked and not discharging to a large surface 

water body to provide effluent dilution. 

Overall, temporal changes in EF were found for all chiral 

drugs except sotalol, with temporal changes in concentration 

for all chiral drugs except atenolol and metoprolol. The 

concentration of atenolol and metoprolol in LLB effluent did 

not change over time, but the EFs of these drugs did change. 

Conversely, although the concentration of sotalol in LLB 

effluent changed over time, the EF remained constant. 

Although temporal changes in concentration may not be 

occurring, the relative proportions of enantiomers being 

released to the environment may be changing over time. Also, 

for some drugs the EF may remain constant, but the concentration 

of the drug may change. 

Similar conclusions can be drawn from the comparison 

between EFs and per capita daily loadings among the three 


WWTPs. Loading and EF differences were found among 

WWTPs for citalopram and sotalol, while for atenolol and 

temazepam, there were no differences in loading but there 

were changes in EF, and for metoprolol there were no differences 

in loading or EF among the three WWTPs. Simply put, 

although concentrations and loadings may be accurately 

predicted based on population-weighted measured effluent 

concentrations in other treatment plants, the EFs of chiral 

drugs released by one WWTP may not necessarily reflect that 

released by another, probably from site-specific biochemical 

activity. Thus, enantiomer-specific monitoring of chiral drugs 

may be useful, particularly for drugs with enantiomer-specific 

toxicity to aquatic life (Stanley et al., 2006). 


We thank D. Bleackley, V. Cooper, M. Ross, B. Asher and 

E. McClure (University of Alberta), R. Litwinow and D. Seehagel 

(GB WWTP), D. Cikaluk (CR WWTP) and G. Siebold (LLB WWTP) 

for sampling assistance, and H. Li (Trent University) for 

providing unpublished POCIS sampling rates for citalopram and 

sotalol. Funding was provided to C.S.W. from the Canada 

Research Chairs Program, Canada’s Natural Sciences and Engineering 

Research Council (NSERC), and the Society of Environmental 

Toxicology and Chemistry (SETAC) Early Career Award 

for Applied Ecological Research, cosponsored by the American 

Chemistry Council; and to S.L.M. as fellowships from NSERC, 

Alberta Ingenuity Fund, and the American Chemical Society 

Division of Analytical Chemistry, sponsored by DuPont. 

Appendix. 


Information about sampling locations and times, analytical 

and data treatment procedures, concentrations, loadings, and 

EFs are available in the online supplementary data to 


in the online version at doi:10.1016/j.watres.2009.09.056. 

references 

Alberta Capital Region Wastewater Commission, 2009. http:// 

www.acrwc.ab.ca accessed February 25. 

Alvarez, D.A., Petty, J.D., Huckins, J.N., Jones-Lepp, T.L., 

Getting, D.T., Goddard, J.P., Manahan, S.E., 2004. Development 

of a passive, in situ, integrative sampler for hydrophilic organic 

contaminants in aquatic environments. Environ. Toxicol. 

Chem. 23, 1640–1648. 

Andreozzi, R., Raffaele, M., Nicklas, P., 2003. Pharmaceuticals in 

STP effluents and their solar photodegradation in the aquatic 

environment. Chemosphere 50, 1319–1330. 

Batt, A.L., Kim, S., Aga, D.S., 2007. Comparison of the occurrence of 

antibiotics in four full-scale wastewater treatment plants with 

varying designs and operations. Chemosphere 68, 428–435. 

Bourne, G.R., 1981. The metabolism of b-adrenoreceptor blocking 

drugs. In: Bridges, J.W., Chasseaud, L.F. (Eds.), Progress in Drug 

Metabolism 6. John Wiley & Sons Ltd, pp. 77–110.

Carballa, M., Omil, F., Ternes, T., Lema, J.M., 2007. Fate of 

pharmaceutical and personal care products (PPCPs) during 

anaerobic digestion of sewage sludge. Water Res. 41, 2139–2150. 

Chen, M., Ohman, K., Metcalfe, C., Ikonomou, M.G., Amatya, P.L., 

Wilson, J., 2006. Pharmaceuticals and endocrine disruptors in 

wastewater treatment effluents and in the water supply 

system of Calgary, Alberta, Canada. Water Qual. Res. J. Can. 

41, 351–364. 

City of Edmonton, 2009. http://www.edmonton.ca. accessed 

February 25. 

Coetsier, C.M., Spinelli, S., Lin, L., Roig, B., Touraud, E., 2009. 

Discharge of pharmaceutical products (PPs) through 

a conventional biological sewage treatment plant: MECs vs 

PECs? Environ. Int 35, 787–792. 

Conkle, J.L., White, J.R., Metcalfe, C.D., 2008. Reduction of 

pharmaceutically active compounds by a lagoon wetland 

wastewater treatment system in Southeast Louisiana. 




Environ. Health. Perspect 107, 907–938. 

Duigou, E., 2006. Town of Lac La Biche, Personal communication. 

EMEA, 2006. Note for guidance on environmental risk assessment of 

medicinal products for human use. The European Agency for the 

Evaluation of Medicinal Products (EMEA) CMPC/SWP/4447/00. 

Environment Canada, 2009. http://www.ec.gc.ca/soer-ree/ 

English/Indicator_series/new_issues.cfm?issue_id¼7&tech_ 

id¼28#bio_pic. accessed February 27. 

Fent, K., Weston, A.A., Caminada, D., 2006. Ecotoxicology of 

human pharmaceuticals. Aquat. Toxicol 76, 122–159. 

Fono, L.J., Sedlak, D.L., 2005. Use of the chiral pharmaceutical 

propranolol to identify sewage discharges into surface waters. 

Environ. Sci. Technol 39, 9244–9252. 

Fono, L.J., Kolodziej, E.P., Sedlak, D.L., 2006. Attenuation of 

wastewater-derived contaminants in a effluent-dominated 

river. Environ. Sci. Technol 40, 7257–7262. 

Göbel, A., McArdell, C.S., Joss, A., Siegrist, H., Giger, W., 2007. Fate 

of sulfonamides, macrolides, and trimethoprim in different 

wastewater treatment methodologies. Sci. Total Environ 372, 

361–371. 

Göbel, A., Thomsen, A., McArdell, C.S., Joss, A., Giger, W., 2005. 

Occurrence and sorption behavior of sulfonamides, 

macrolides, and trimethoprim in activated sludge treatment. 


Halling-SBrensen, B., Nielsen, S.N., Lanzky, P.F., Ingerslev, F., 

Holten-LützhBft, H.C., JBrgensen, S.E., 1998. Occurrence, fate 

and effects of pharmaceutical substances in the environment 

– a review. Chemosphere 36, 357–393. 

Harner, T., Wiberg, K., Norstrom, R., 2000. Enantiomer fractions 

are preferred to enantiomer ratios for describing chiral 

signatures in environmental analysis. Environ. Sci. Technol 

34, 218–220. 

Hua, W.Y., Bennett, E.R., Miao, X.-S., Metcalfe, C.D., Letcher, R.J., 

2006. Seasonality effects on pharmaceuticals and s-triazine 

herbicides in wastewater effluent and surface water from the 

Canadian side of the Upper Detroit River. Environ. Toxicol. 

Chem. 25, 2356–2365. 

Ikehata, K., El-Din, M.G., Snyder, S.A., 2008. Ozonation and 

advanced oxidation treatment of emerging organic pollutants 

in water and wastewater. Ozone Sci. Eng 30, 21–26. 

Jones-Lepp, T.L., Alvarez, D.A., Petty, J.D., Huckins, J.N., 2004. 

Polar organic chemical integrative sampling and liquid 

chromatography-electrospray/ion-trap mass spectrometry for 

assessing selected prescription and illicit drugs in treated 

sewage effluents. Arch. Environ. Contam. Toxicol 47, 427–439. 

Joss, A., Zabcynski, S., Göbel, A., Hoffmann, B., Löffler, D., 

McArdell, C.S., Ternes, T.A., Thomsen, A., Siegrist, H., 2006. 

Biological degradation of pharmaceuticals in municipal 


wastewater treatment: proposing a classification scheme. 

Water Res. 40, 1686–1696. 

Kim, S.-C., Carlson, K., 2007. Temporal and spatial trends in 

the occurrence of human and veterinary antibiotics in 

aqueous and river sediment matrices. Environ. Sci. Technol 

41, 50–57. 

Kimura, K., Hara, H., Watanabe, Y., 2007. Elimination of selected 

acidic pharmaceuticals from municipal wastewater by an 

activated sludge system and membrane bioreactors. Environ. 

Sci. Technol 41, 3708–3714. 


Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, hormones, 

and other organic wastewater contaminants in U.S. streams, 

1999–2000: a national reconnaissance. Environ. Sci. Technol 36, 

1202–1211. 

Kurt-Karakus, P.B., Bidleman, T.F., Muir, D.C.G., Cagampan, S.J., 

Struger, J., Sverko, E., Small, J.M., Jantunen, L.M., 2008. Chiral 

current-use herbicides in Ontario streams. Environ. Sci. 

Technol 42, 8452–8458. 

Lajeunesse, A., Gagnon, C., Sauvé, S., 2008. Determination of basic 

antidepressants and their N-desmethyl metabolites in raw 

sewage and wastewater using solid-phase extraction and 

liquid chromatography-tandem mass spectrometry. Anal. 

Chem. 80, 5325–5333. 

Letzel, M., Metzner, G., Letzel, T., 2009. Exposure assessment of 

the pharmaceutical diclofenac based on long-term 

measurements of the aquatic input. Environ. Int 35, 363–368. 

Lin, A.Y.-C., Reinhard, M., 2005. Photodegradation of common 

environmental pharmaceuticals and estrogens in river water. 


Lin, A.Y.-C., Plumlee, M.H., Reinhard, M., 2006. Natural 

attenuation of pharmaceuticals and alkylphenol 

polyethoxylate metabolites during river transport: 

photochemical and biological transformation. Environ. 

Toxicol. Chem. 25, 1458–1464. 

Lishman, L., Smyth, S.A., Sarafin, K., Kleywegt, S., Toito, J., 

Peart, T., Lee, B., Servos, M., Beland, M., Seto, P., 2006. 

Occurrence and reductions of pharmaceuticals and personal 

care products and estrogens by municipal wastewater 

treatment. Sci. Total Environ 367, 544–588. 

Lissemore, L., Hao, C., Yang, P., Sibley, P.K., Mabury, S., 

Solomon, K.R., 2006. An exposure assessment for selected 

pharmaceuticals within a watershed in Southern Ontario. 


Loraine, G.A., Pettigrove, M.E., 2006. Seasonal variations in 

concentrations of pharmaceuticals and personal care 

products in drinking water and reclaimed wastewater in 

Southern California. Environ. Sci. Technol 40, 687–695. 

MacLeod, S.L., 2009. Pharmaceutical occurrence and fate in 

wastewater and receiving surface waters in two Alberta 

watersheds. PhD dissertation. University of Alberta, 

Edmonton, Alberta. 

MacLeod, S.L., McClure, E.L., Wong, C.S., 2007a. Laboratory 

calibration and field deployment of the Polar Organic 

Chemical Integrative Sampler for pharmaceuticals and 

personal care products in wastewater and surface water. 


MacLeod, S.L., Sudhir, P., Wong, C.S., 2007b. Stereoisomer 

analysis of wastewater-derived beta-blockers, selective 

serotonin re-uptake inhibitors, and salbutamol by highperformance 

liquid chromatography-tandem mass 

spectrometry. J. Chromatogr. A 1170, 23–33. 

Maier, N.M., Franco, P., Lindner, W., 2001. Separation of 

enantiomers: needs, challenges, perspectives. J. Chromatogr. 

A 906, 3–33. 

Managaki, S., Murata, A., Takada, H., Tuyen, B.C., Chiem, N.H., 

2007. Distribution of macrolides, sulfonamides, and 

trimethoprim in tropical waters: ubiquitous occurrence of

544 

veterinary antibiotics in the Mekong Delta. Environ. Sci. 

Technol 41, 8004–8010. 

McArdell, C.S., Molnar, E., Suter, M.J.-F., Giger, W., 2003. 

Occurrence and fate of macrolide antibiotics in wastewater 

treatment plants and in the Glatt Valley Watershed. 

Switzerland. Environ. Sci. Technol 37, 5479–5486. 

Mehvar, R., Brocks, D.R., 2001. Stereospecific pharmacokinetics 

and pharmacodynamics of beta-adrenergic blockers in 

humans. J. Pharm. Pharm. Sci. 4, 185–200. 

Miao, X.-S., Yang, J.-J., Metcalfe, C.D., 2005. Carbamazepine and its 

metabolites in wastewater and in biosolids in a municipal 

wastewater treatment plant. Environ. Sci. Technol 39, 7469–7475. 

Neumann, J., 1989. Untersuchungen zur Bioverfügbarkeit und 

Pharmakokinetik von Sulfasalazin und seinen Metaboliten. 

PhD thesis. Free University of Berlin. 

Nikolai, L.N., McClure, E.L., MacLeod, S.L., Wong, C.S., 2006. 

Stereoisomer quantification of the b-blocker drugs atenolol, 

metoprolol, and propranolol in wastewaters by chiral highperformance 

liquid chromatography-tandem mass 


Sacher, F., Ehmann, M., Gabriel, S., Graf, C., Brauch, H.-J., 2008. 

Pharmaceutical residues in the river Rhine-results of a onedecade 

monitoring programme. J. Environ. Monit. 10, 664–670. 

Schmekel, B., Rydberg, I., Norlander, B., Sjöswärd, K.N., Ahlner, J., 

Anderson, R.G.G., 1999. Stereoselective pharmacokinetics of 

S-salbutamol after administration of the racemate in healthy 

volunteers. Eur. Respir. J 13, 1230–1235. 

Shaw, M., Mueller, J.F., 2009. Time integrative passive sampling: 

how well do Chemcatchers integrate fluctuating pollutant 

concentrations? Environ. Sci. Technol 43, 1443–1448. 

Siebold, G., 2008. Lac La Biche County, Personal communication. 


Stamm, C., Alder, A.C., Fenner, K., Hollender, J., Kraus, M., 

McArdell, C.S., Ort, C., Schneider, M.K., 2008. Spatial and 

temporal patterns of pharmaceuticals in the aquatic 

environment. Geog. Compass 2/3, 920–955. 

Stanley, J.K., Ramirez, A.J., Mottaleb, M., Chambliss, C.K., 

Brooks, B.W., 2006. Enantiospecific toxicity of the b-blocker 

propranolol to Daphnia magna and Pimephales promelas. 


Statistics Canada, 2009. http://www12.statcan.ca. accessed 

February 25. 

Stülten, D., Zühlke, S., Lamshöft, M., Spiteller, M., 2008. 

Occurrence of diclofenac and selected metabolites in sewage 

effluents. Sci. Total Environ 405, 310–316. 

Vieno, N.M., Tuhkanen, T., Kronberg, L., 2005. Seasonal variation 

in the occurrence of pharmaceuticals in effluents from 

a sewage treatment plant and in the recipient water. Environ. 

Sci. Technol 39, 8220–8226. 

Vasskog, T., Anderssen, T., Pedersen-Bjergaard, S., Kallenborn, R., 

Jensen, E., 2008. Occurrence of selective serotonin reuptake 

inhibitors in sewage and receiving waters at Spitsbergen and 

in Norway. J. Chromatogr. A 1185, 194–205. 

Wong, C.S., 2006. Environmental fate processes and biochemical 

transformations of chiral emerging organic pollutants. Anal. 

Bioanal. Chem. 386, 544–558. 

Yu, J.T., Bouwer, E.J., Coelhan, M., 2006. Occurrence and 

biodegradability studies of selected pharmaceuticals and personal 

care products in sewage effluent. Agric. Water Mgmt 86, 72–80. 

Zhang, Z., Hibberd, A., Zhou, J.L., 2008. Analysis of emerging 

contaminants in sewage effluent and river water: 

comparison between spot and passive sampling. Anal. Chim. 

Acta 607, 37–44.

Degradation of fifteen emerging contaminants at mgL L1 initial 

concentrations by mild solar photo-Fenton in MWTP effluents 

N. Klamerth a,b , L. Rizzo c , S. Malato a, *, Manuel I. Maldonado a , 

A. Agüera b , A.R. Fernández-Alba b 

a 

Plataforma Solar de Almería-CIEMAT. Carretera Senés km 4, 04200 Tabernas (Almería), Spain 

b 

Pesticide Residue Research Group, University of Almería, 04120 Almería, Spain 

c 

Department of Civil Engineering, University of Salerno, via Ponte don Melillo, 84084 Fisciano (SA), Italy 

article info 






Available online 4 October 2009 

Keywords: 

Emerging contaminants 

Pharmaceuticals treatment 

Photo-Fenton 

Solar photocatalysis 


abstract 

Due to their growing use, pharmaceuticals, like the antiinflammatory 

ibuprofen, the antibiotic flumequine and the 

antiepileptic carbamazepine, endocrine disruptors like 

bisphenol A and atrazine, personal care products like oxybenzone 

and parabens (PHBA), synthetic musks and 

fragrances like musk xylene and galaxolide, pesticides like 

isoproturon and endosulphan and illicit drugs like THC aznd 

cocaine, to name just a few, and other xenobiotic substances, 

are found in increasing quantities in wastewater, surface 

water, and even in drinking water (Kasprzyk-Hordern et al., 

2009; Kim et al., 2007; Mitch et al., 2003; Esplugas et al., 2007; 

Ternes, 1998). Since the use of these substances cannot be 

controlled or eliminated as they are ever present in our daily 




* Corresponding author. Tel.: þ34 950387940; fax: þ34 950365015. 

E-mail address: sixto.malato@psa.es (S. Malato). 


doi:10.1016/j.watres.2009.09.059 

The degradation of 15 emerging contaminants (ECs) at low concentrations in simulated and 

real effluent of municipal wastewater treatment plant with photo-Fenton at unchanged pH 

and Fe ¼ 5mgL 1 in a pilot-scale solar CPC reactor was studied. The degradation of those 

15 compounds (Acetaminophen, Antipyrine, Atrazine, Caffeine, Carbamazepine, Diclofenac, 

Flumequine, Hydroxybiphenyl, Ibuprofen, Isoproturon, Ketorolac, Ofloxacin, Progesterone, 

Sulfamethoxazole and Triclosan), each with an initial concentration of 100 mgL 1 , 

was found to depend on the presence of CO3 2 and HCO3 (hydroxyl radicals scavengers) and 

on the type of water (simulated water, simulated effluent wastewater and real effluent 

wastewater), but is relatively independent of pH, the type of acid used for release of 

hydroxyl radicals scavengers and the initial H2O2 concentration used. Toxicity tests with 

Vibrio fisheri showed that degradation of the compounds in real effluent wastewater led to 

toxicity increase. 


lives, their release into the environment has to be optimized 

and restricted, as they pose risks to the environment, public 

health and aquatic systems and they are responsible for 

building up microbiological resistance, feminisation of higher 

organisms and ecotoxicological issues (Laville et al., 2004). 

Particularly relevant examples of such emerging contaminants 

(ECs), such as those mentioned above, which are ubiquitously 

present in influents and effluents of MWTPs in the 

high ng L 1 to low mgL 1 range, do not need to be persistent to 

be hazardous, because they are introduced continuously into 

the environment (Fono et al., 2006; Jackson and Sutton, 2008; 

Nakada et al., 2008; Petrovic et al., 2003). 

Conventional MWTPs, typically based on biological 

processes, are capable of removing some substances, but nonbiodegradable 

compounds may escape the treatment and be

546 

released into the environment (Göbel et al., 2007; Carballa 

et al., 2004; Ternes et al., 2007). ECs have been found in the 

MWTP effluents at mean concentrations ranging from 0.1 to 

20 mgL 1 (Castiglioni et al., 2006; Martínez Bueno et al., 2007; 

Radjenović et al., 2007; Richardson, 2007; Zhao et al., 2009). 

Concern about the growing problem of the continuously rising 

concentrations of these compounds must be emphasized, and 

therefore, the application of more thorough wastewater 

treatment protocols, including the use of new and improved 

technologies, is a necessary task. 

Reusable water should be free of these persistent, toxic, 

endocrine-disrupting or non-biodegradable substances, 

(Radjenović et al., 2007; Teske and Arnold, 2008), and therefore, 

an effective tertiary treatment is required to remove 

these substances completely. 

Among the advanced technologies that may be used to 

remove these pollutants, (Gogate and Pandit, 2004; Saritha 

et al., 2007; Huber et al., 2005) advanced oxidation processes 

(AOPs), through the generation of hydroxyl radicals which are 

able to mineralise most organic molecules yielding CO 2 and 

inorganic ions as final products, are a particularly attractive 

option. (Farré et al., 2005; Gebhardt and Schröder, 2007; Gültekin 

and Ince, 2007; Ning et al., 2007; Rosenfeldt and Linden, 

2004; Rosenfeldt et al., 2007; Ternes et al., 2003) The generation 

of the OH radicals can be achieved: electrochemically (Cañizares 

et al., 1999; Pelegrini et al., 2001; Zhou et al., 2005), 

sonochemically (Mantzavinos et al., 2004; Papadaki et al., 

2004; Lesko et al., 2006), photochemically (Esplugas et al., 2005; 

Bali et al., 2003; Bremner et al., 2006), and by homogeneous or 

heterogenous catalysis (Zepp et al., 1992; Martinez et al., 2005) 

in acid or basic media (Glaze et al., 1987; Hislop and Bolton, 

1999; Neyens and Baeyens, 2003). Most of the AOPs make use 

of a combination of either oxidants and irradiation (O 3/H 2O 2/ 

UV), or a catalyst and irradiation (Fe 2þ /H 2O 2; UV/TiO 2). The 

drawbacks which make them economically disadvantageous 

depend on the specific AOP: (i) High electricity demand (e.g. 

ozone and UV-based AOPs), (ii) the relatively large amounts of 

oxidants and/or catalysts (e.g. ozone, hydrogen peroxide and 

iron-based AOPs), and (iii) the pH operating conditions (e.g. 

Fenton and photo-Fenton). This is why, although AOPs are 

well known for their capacity for oxidising and mineralising 

almost any organic contaminant, commercial applications are 

still scarce. Processes like photo-Fenton may be applied to 

commercial applications by using solar energy as a light 

source, optimizing the pH range and the amounts of oxidant/ 

catalyst required. 

AOP efficiency in the removal of ECs has typically been 

studied in demineralised water and bench scale, at initial 

concentrations in the milligram-to-gram range, which is not 

realistic compared to the concentrations detected in real 

water and wastewater (Farré et al., 2005; Lapertot et al., 2007; 

Kassinos et al., 2009; Malato et al., 2007). This work focused on 

solar photo-Fenton degradation of the ECs typically found in 

the effluents of MWTPs, leaving the treated wastewater suitable 

for reuse. Moreover, to make the process of interest for 

practical applications high iron concentrations (mM range), 

excessive amounts of H 2O 2 and a pH under 3 must be avoided 

(Pignatello et al., 2006). A new approach aimed at finding 

a very mild photo-Fenton treatment (low iron concentration 

and H2O2 dose at neutral pH), has been proposed (Moncayo- 


Lasso et al., 2008; Klamerth et al., 2009). In this paper, a pilotscale 

solar photo-Fenton treatment was run at a pH between 6 

and 7, with starting concentrations of 5 mg L 1 Fe, and 

50 mg L 1 H2O2. Synthetic water (SW), simulated effluent 

wastewater (SE) and real effluent wastewater (RE) were tested 

in this study to which a mixture of 15 ECs, consisting of 

pharmaceuticals, pesticides and personal care products, 

selected from a list of 80 compounds found in MWTP effluents 

in previous studies (Martínez Bueno et al., 2007), were added at 

low concentrations (100 mgL 1 each). 

2. Material and methods 

2.1. Reagents 

All reagents used for chromatographic analyses, acetonitrile, 

methanol, and ultrapure (MilliQ) water, were HPLC grade. 

Analytical standards for chromatography analyses were 

purchased from Sigma-Aldrich. Table 1 and Scheme 1 list the 

15 compounds (pharmaceuticals, pesticides and personal care 

products) used. Photo-Fenton experiments were performed 

using iron sulphate (FeSO4$7H2O), reagent-grade hydrogen 

peroxide (30% w/v), sulphuric acid and hydrochloric acid for 

carbonate stripping, all provided by Panreac. The filters used 

were syringe driven 0.2 mm Millex nylon membrane filters 

from Millipore. The reagents used in the preparation of the 

synthetic water (NaHCO 3, CaSO 4$2H 2O, MgSO 4, and KCl), and 

simulated effluent wastewater (Peptone, Meat extract, Urea, 

K 2HPO 4, CaCl 2$2H 2O, MgSO 4$7H 2O and NaCl) were provided by 

Panreac. 

2.2. Polluted waters 

Taking into consideration that typical EC concentrations in the 

effluent are in the 0.1–20.0 mgL 1 range, it was decided to work 

at 100 mgL 1 , which is a compromise between (i) a high enough 

concentration to characterise kinetics using conventional 

Table 1 – LOD, LOQ and absorption l of the selected 

compounds. 

Name LOD 

(mgL 1 ) 

LOQ 

(mgL 1 ) 

Absorption 

l [nm] 

Acetaminophen 1.4 2.8 245 

Antipyrine 2.1 4.2 205 

Atrazine 0.6 1.2 223 

Caffeine 0.7 1.5 205 

Carbamazepine 0.8 1.6 211 

Diclofenac 2.7 5.5 277 

Flumequine 2.1 4.3 248 

Hydroxybiphenyl 1.6 3.2 243 

Ibuprofen 2.5 5.0 222 

Isoproturon 1.5 3.0 205 

Ketorolac 2.1 4.2 321 

Ofloxacin 1.3 2.6 295 

Progesterone 2.0 4.0 248 

Sulfamethoxazole 1.4 2.7 267 

Triclosan 5.0 10.0 280

Name Structure Name Structure 

Acetaminophen 

Ibuprofen 

analgesic / 

antipyretic 

Antipyrine 

analgesic 

Atrazine 

herbicide 

Caffeine 

stimulant 

Carbamazepine 

anticonvulsant 

Diclofenac 

antiinflammatory 

Flumequine 

broadspectrum 

antibiotic 

Hydroxybiphenyl 

biocide 

HO NH 

O 

N 

Cl 

N 

O 

N 

N N 

O 

N 

NH 2 

O 

nonsteroidal 


Isoproturon 

phenylurea 

herbicide 

H 

H 

N N N Ketorolac 

O 

N 

H 

N 

Cl 

Cl 

N 

N 

O 

OH 


Ofloxacin 

gramnegative 

antibiotic 

Progesterone 

steroid 

hormone 

Sulfamethoxaz 

ole 

bacteriostatic 

antibiotic 

F COOH Triclosan 

OH 


O 

N 

anti-bacterial/ 

fungal agent 

N 

N 

O 

O 

F 

N 

O 

H 2 N S 

Scheme 1 – Name and structure of the 15 selected compounds. 

Cl 

O 

Cl 

O 

O 

N 

H 

N 

H 

N 

O 

O 

O 

OH 

OH 

N 

N 

O 

O 

O 

O 

OH 

Cl 

OH

548 

C/C 0 

1.0 

0.5 

0.0 

1.0 

0.5 

0.0 

1.0 

0.5 

0.0 

Fenton 

-40 -20 0 20 40 60 80 100 

t EXP , min 

illumination 

t 30W , min 

Acetaminophen 

Caffeine 

Ofloxacine 

Antipyrine 

Sulfamethoxazole 

Carbamazepine 

Flumequine 

Ketorolac 

Atrazine 

Isoproturon 

Hydroxybenzone 

Diclofenac 

Ibuprofen 

Progesterone 

Triclosan 

Fig. 1 – Degradation of the 15 ECs (0.1 mg L L1 each) by 

photo-Fenton with 5 mg L L1 Fe without pH adjustment in 

SW acidified for release of carbonates. 

analytical techniques, and (ii) low enough to simulate real 

conditions. 

Synthetic water (SW), simulated effluent wastewater (SE) and 

real effluent wastewater (RE) to which a mixture of the 15 ECs 

previously listed had been added at low concentrations 

(100 mgL 1 ) were tested. The experimental protocol was designed 

to study solar photo-Fenton with a relatively uncomplicated 

aqueous matrix (SW) first, before gradually increasing 

complexity by studying SE and finally RE, in order to acquire 

information on process operation. 

Demineralised water used for preparing SW and SE in the 

pilot plant was supplied by the Plataforma Solar de Almería 

C/C 0 

1.0 

0.5 

0.0 

1.0 

0.5 

0.0 

1.0 

0.5 

0.0 

Fenton 

t EXP , min 

illumination 


(PSA) distillation plant (conductivity < 10 mScm 1 ,Cl ¼ 0.7– 

0.8 mg L 1 , NO3 ¼ 0.5 mg L 1 , organic carbon < 0.5 mg L 1 ). 

The preparation of standard moderately-hard freshwater 

(SW) was carried out taking into account the characteristics of 

ground water in the province of Almería (southern Spain). SW 

was prepared by mixing 96 mg L 1 of NaHCO3, 60mgL 1 of 

CaSO4$2H2O, 60 mg L 1 of MgSO4, and 4 mg L 1 of KCl (Standard 

Methods, 1998). The simulated effluent wastewater (SE) 

consists of SW and Peptone (32 mg L 1 ), Meat extract 

(22 mg L 1 ), Urea (6 mg L 1 ), K 2HPO 4 (28 mg L 1 ), CaCl 2$2H 2O 

(4 mg L 1 ), MgSO 4$7H 2O(2mgL 1 ) and NaCl (7 mg L 1 ) which 

derives to an initial DOC (dissolved organic carbon) of 

25 mg L 1 (OECD, 1999). RE was taken downstream of the 

Almería MWTP secondary biological treatment and used as 

received within the next 2 days. Initial COD and DOC were at 

60 mg L 1 and 25 mg L 1 , respectively. 

2.3. Solar photo-Fenton pilot plant 

Photo-Fenton experiments were performed at the Plataforma 

Solar de Almería in a pilot compound parabolic collector (CPC) 

solar plant designed for solar photocatalytic applications. This 

reactor is composed of two 11dL modules with twelve Pyrex 

glass tubes (30 mm O.D.) mounted on a fixed platform tilted 37 

(local latitude). The water flows (20 L min 1 ) directly from one 

module to the other and finally to a 10dL reservoir. The piping 

and valves (3 L) between the reactor and the tank are black 

HDPE, which is highly resistant to chemicals, weatherproof 

and opaque, preventing any photochemical effect from outside 

the collectors. The total illuminated area is 3 m 2 , the total 

volume (two modules þ reservoir tank þ piping and valves) is 

35 L (VT) and the irradiated volume is 22 L (Vi). Solar ultraviolet 

radiation (UV) was measured by a global UV radiometer 

(KIPP&ZONEN, model CUV 3) mounted on a platform tilted 37 

consumed H 2O 2 

0 60 120 180 240 300 

t 30W , min 

Acetaminophen 

Caffeine 

Ofloxacine 

Antipyrine 

Sulfamethoxazole 

Carbamazepine 

Flumequine 

Ketorolac 

Atrazine 

Isoproturon 

Hydroxybenzone 

Diclofenac 

Ibuprofen 

Progesterone 

Triclosan 

Fig. 2 – Degradation of the 15 ECs (0.1 mg L L1 each) by photo-Fenton (5 mg L L1 Fe, no pH adjustment) and hydrogen peroxide 

consumption in SE acidified for carbonate release. 

120 

90 

60 

30 

0 

consumed H 2 O 2 , mg L -1

Table 2 – Final concentration of ECs in two different experiments with SE after similar irradiation times, with 

[Fe]0 [ 5mgL L1 : 300 min with [H2O2]0 [ 50 mg L L1 and addition of 50 mg L L1 when the initial dose was almost consumed 

and so on; 336 min with [H2O2]0 [ 5mgL L1 and addition of H2O2 every 30–60 min until the end of the experiment. Fenton 

degradation (%) in the dark and first order kinetic constants (k) for photo-Fenton are also included. 

[H2O2]0 ¼ 50 mg L 1 

[H2O2]0 ¼ 5mgL 1 

(the same as the CPCs). The temperature inside the reactor was 

continuously recorded by a temperature probe (Crioterm 

PT-100 3H) inserted in the piping. A plant diagram has been 

published elsewhere (Kositzi et al., 2004). With Eq. (1), combination 

of the data from several days’ experiments and their 

comparison with other photocatalytic experiments is possible, 

UV Vi 

t30W;n ¼ t30W;n 1 þ Dtn ; Dtn ¼ tn tn 1; t0 ¼ 0ðn ¼ 1Þ (1) 

30 VT 

where tn is the experimental time for each sample, UV is the 

average solar ultraviolet radiation (l < 400 nm) measured 

between tn 1 and tn, and t30W is a ‘‘normalized illumination 

time’’. In this case, time refers to a constant solar UV power of 

30 Wm 2 (typical solar UV power on a perfectly sunny day 

around noon). 

2.4. Experimental setup 

t 30W C/C 0 Fenton 

deg. X [%] 

The mixture of the 15 compounds (100 mgL 1 each) dissolved in 

Methanol at 2.5 g L 1 (mother solution) was added directly 

(1.4 mL) into the pilot plant and well homogenized by turbulent 

recirculation for 15 min. Methanol was used to dissolve the 

contaminants (TOC from methanol was 12 mg L 1 ). The pH in 

SW was 7.6, in SE 7.8 and in RE 8.0, which favoured bicarbonate 

ions, so at the beginning of the process, when the collectors 

were still covered, enough acid to remove CO 3 2 and HCO3 , 

known OH radical scavengers, was added. Recirculation time 

for this process was between 15 and 60 min. When the total 

inorganic carbon (TIC) concentration was below the desired 

level (10 mg L 1 ), 50 mg L 1 of peroxide was added and homogenised 

by recirculation for 15 min. Finally, Fe 2þ as FeSO4$7H20 

was added (5 mg L 1 ). After 15 min of recirculation, in which 

the Fenton reaction started, the collectors were uncovered and 

photo-Fenton began. Hydrogen peroxide and iron were 

measured in every sample taken. The experiments normally 


k [min 1 ] t 30W C/C 0 Fenton 

deg. X [%] 

lasted as long as there was H 2O 2 to be consumed (usually 3–4 h), 

in the case of experiments which lasted longer than that, the 

experiment was covered at the end of the day and the next day, 

after recirculation for 15 min, a sample was taken, peroxide 

was added (50 mg L 1 ) and the covers again removed. 

2.5. Analytical measurements 

k [min 1 ] 

Acetaminophen 261

550 

Table 3 – Final EC concentration in RE with 

[Fe]0 [ 5mgL L1 ,[H2O2]0 [ 50 mg L L1 after an irradiation 

time of t30W [ 276 min. Fenton degradation (%) in the 

dark and first order kinetic constant (k) for photo-Fenton 

are also included. 

t30W C/C0 Fenton k [min 

deg. X [%] 

1 ] 

Acetaminophen 109

eactor (t ExP ¼ 15 min). Removal of all compounds but atrazine 

(18%) was in the range between 30% (caffeine) and 60% 

(ibuprofen). Fenton degradation was very quick after adding just 

5mgL 1 of Fe 2þ but it did not proceed further until the reactor 

was uncovered, as Fe 3þ reduction to Fe 2þ was inefficient without 

illumination. All the pollutants but triclosan (87%) and atrazine 

(77%) were efficiently degraded (removal over 94%) under photo- 

Fenton conditions, at t30W ¼ 90 min. Moreover, some pollutants 

(ofloxacin and diclofenac) were degraded to below the LOD as 

early as t 30W ¼ 32 min. The total amount of H 2O 2 consumed was 

37 mg L 1 and the iron concentration decreased due to precipitation 

from 5 mg L 1 to 3.1 mg L 1 , probably due to colloidal 

aggregation, while the pH varied from an original 7.6–5.3 (after 

adding the FeSO4$7H2O) to a final pH ¼ 3.8 at t30W ¼ 90 min. 

Similar results were found by adding H2SO4 (56 mg L 1 ) 

instead of HCl. H2SO4 has certain advantages (handling, 

volatility, etc.) over HCl, and Cl ions may act as radical 

scavengers. Experiments did not show any significant difference 

between using HCl or H2SO4. Indeed, the total concentration 

of chloride and sulphate measured in SW was 

43 mg L 1 Cl when HCl was added and 145 mg L 1 SO 4 2 when 

H 2SO 4 was added respectively. It is well known that inorganic 

species (chloride, sulphate, phosphates, etc) are usually 

detrimental to the photo-Fenton reaction rate (Pignatello 

et al., 2006) due to complexation of these ions with Fe 2þ or 

Fe 3þ , or to the scavenging of hydroxyl radicals and formation 

of less reactive inorganic ions (De Laat et al., 2004), but it has 

also been observed that they significantly reduce the photo- 

Fenton reaction rate only if they occur at concentrations over 

500 mg L 1 , usually very far from the concentration found in 

natural water (Bacardit et al., 2007; Zapata et al., 2009). 

3.2. Photo-Fenton tests using SE 

The following experiment dealt with the photo-Fenton treatment 

of SE, which was characterised by original TOC and COD 

of 37 mg L 1 and 80 mg L 1 , respectively. Two different 

approaches to the H2O2 dose were studied: (i) 50 mg L 1 were 

added at the beginning and again when the initial dose was 

almost consumed and so on; (ii) 5 mg L 1 were added every 

30–60 min (also depending on consumption) until the end of 

the experiment. In this case, 56 mg L 1 H 2SO 4 acid were added. 

A recirculation time of 15 min was necessary for carbonates to 

be released (TIC < 5mgL 1 ). As can be seen in Fig. 2, degradation 

is much slower with both Fenton (t < 0) and photo- 

Fenton (t30W > 0) than in the experiment with SW shown in 

Fig. 1. The residual concentration of the contaminants at the 

end of the experiments (t30W ¼ 300 min and 336 min, respectively) 

can be seen in Table 2. The pH varied during the 

experiment with high H 2O 2 concentrations from the original 

7.6–4.9 at the end, while the iron concentration varied from 

5mgL 1 to 3.2 mg L 1 . During the experiment with low H 2O 2 

concentrations the pH varied from 8.1 to 4.9, while iron 

concentration went from 4.8 to 3.1 mg L 1 . The EC degradation 

behaviour did not change significantly, no matter which 

hydrogen peroxide dose approach was used, as can be seen in 

Table 2, although the overall amount of peroxide consumed 

was much lower (110 mg in 50-mg-L 1 additions compared to 

52 mg with the 5-mg-L 1 additions). It should be mentioned, 

although not relevant to the purpose of the experiments, 


because the main purpose was to degrade ECs, that TOC 

mineralisation was rather low in both cases (around 25%). 

3.3. Photo – Fenton treatment of RE 

The initial DOC, TIC and COD were 36 mg L 1 , 106 mg L 1 and 

60 mg L 1 , respectively. In this case, 406 mg L 1 H2SO4 were 

added to reach < 20 mg L 1 TIC and recirculation time in the pilot 

plant necessary to remove carbonate species was 30 min. 

According to previous tests, photo-Fenton treatment at a TIC over 

20 mg L 1 leads to very slow degradation. Interestingly, although 

the degradation rate of all ECs was faster than in the previous 

experiments in SE, as observed in Table 3, DOC degradation was 

similar to SE (around 20%). This can be explained by the presence 

of humic acids in RE which produce solvated electrons and 

hydroxyl radicals upon irradiation (Fukushima and Tatsumi, 

2001; Prosen and Zupančič-Kralj, 2005; Auger et al., 2002). 

3.4. The effect of photo-Fenton treatment on 

V. fisheri toxicity 

In Fig. 3,theresultsofV. fisheri toxicity tests with SE and RE after 

photo-Fenton experiments are compared. Three different steps 

in toxicity behaviour (marked 1, 2 and 3 in Fig. 3) can be 

described. Before photo-Fenton starts, SE toxicity (20–30% inhibition) 

may be due to the mixture of pollutants as well as to the 

formation of intermediates during Fenton in the dark, particularly 

from the fast oxidation of some of the ECs. On the contrary, 

simultaneous activation (negative inhibition) is also observable 

in the RE experiment. This completely different behaviour may 

be explained by the RE – nutrients and salt contents, which in 

someway favoured the growth of V. fisheri until the toxic effect of 

pollutants and their oxidation intermediates was neutralized. 

In both SE and RE, when photo-Fenton starts up, from an of 

H2O2 dose ¼ 0mgL 1 up to H2O2 dose ¼ 50 mg L 1 , the degradation 

of parent pollutants begins forming more toxic intermediates, 

which drastically increase the inhibition rate in SE 

(Fig. 3, Step 2). The previously favourable conditions of RE 

probably can no longer efficiently balance the increasingly 

toxic effect, so inhibition start to gradually increase in RE too. 

In the last step (3) inhibition with RE is quite constant, probably 

because toxic organic intermediates take longer to 

mineralise; on the other hand, toxicity in SE almost totally 

disappeared. If we compare these results with residual 

concentration of the four parent ECs selected for comparison 

(selected for their representativeness of the behaviour of the 

15 different compounds) it may be observed that photo-Fenton 

treatment was more efficient in the RE experiment than in 

the SE. So in terms of toxicity, the higher removal rate in RE 

experiment may result in a correspondingly larger amount of 

more toxic intermediates than in the SE experiment or the 

generation of other intermediates from RE organics. Therefore, 

toxicity assessment be should taken into account as well 

as EC degradation during AOP wastewater treatment. 


- The experiments showed that ECs at low concentrations 

(mgL 1 range) can be successfully degraded to negligible

552 

concentrations with solar photo-Fenton at low iron 

concentrations (5 mg L 1 ) and low initial H2O2 (50 mg L 1 ) 

concentrations without adjusting the pH. 

- One limiting factor for the degradation rate is the presence of 

CO3 2 and HCO3 - which are very efficient OH radical scavengers 

and which have to be removed prior to the photo- 

Fenton reaction. This can be done either with HCl or H2SO4, 

as the range of the final concentration of Cl or SO4 2 does not 

influence the reaction. Therefore, either one may be used as 

economic/operating reasons recommend, but not for any 

kinetic effect. It should be taken into account that around 

400 mg L 1 of H 2SO 4 (0.05 V kg 1 ) were necessary for CO 2 

stripping in RE, which means around 0.025 V m 3 . However, 

these costs should be evaluated on a case by case basis as it 

strongly depends on the particular RE. 

- Changing the operating conditions by lowering H2O2 

concentration (5 mg L 1 ) and adjusted by properly 

controlled dosing could lead to comparable degradation of 

the compounds with lower peroxide consumption. 

- Degradation is faster in RE due to humic acids producing 

solvated electrons and hydroxyl radicals upon irradiation, 

and these contribute to the radicals produced by photo- 

Fenton. 

Toxicity from the formation of oxidation intermediates 

should be taken into account in the treatment of effluents 

from wastewater treatment plants by AOPs, because they 

could increase final toxicity. 


Funding for this work was provided by the Spanish Ministry of 

Science and Innovation under the Consolider-Ingenio 2010 

programme (Project CSD2006–00044 TRAGUA; http://www. 

consolider-tragua.com) and by the Andalusia Regional Government 

(Project no. P06-TEP-02329). Nick Klamerth would to thank 

the University of Almería and CIEMAT for his Ph. D. research 

grant. Luigi Rizzo wishes to thank the University of Salerno 

(Italy) and the Province of Salerno for the research grant which 

allowed him to work in the above mentioned project at the 

Plataforma Solar de Almería. The authors also wish to thank 

Mrs. Deborah Fuldauer for English language correction. 

references 

Auger, J.P., Richard, C., Trubetskaya, O., Trubetskoj, O., Lévèque, J., 

Andreux, F., 2002. Photoinductive efficiency of soil extracted 

humic and fulvic acids. Chemosphere 49, 259–262. 

Bacardit, J., Stötzner, J., Chamarro, E., Esplugas, S., 2007. Effects of 

salinity on the photo-Fenton process. Industrial and 

Engineering Chemistry Research 46, 7615–7619. 

Bali, U., Catalkaya, E.C., Sengul, F., 2003. Photochemical 

degradation and mineralisation of phenol: a comparative 

study. Journal of Environmental Science and Health – Part A 

Toxic/Hazardous Substances and Environmental Engineering 

38, 2259–2275. 

Bremner, D.H., Burgess, A.E., Houllemare, D., Namkung, K., 2006. 

Phenol degradation using hydroxyl radicals generated from 

zero-valent iron and hydrogen peroxide. Applied Catalysis B: 

Environmental 63, 15–19. 


Cañizares, P., Domínguez, J.A., Rodrigo, M.A., Villaseñor, J., 

Rodríguez, F., 1999. Effect of the current intensity in the 

electrochemical oxidation of aqueous phenol wastes at an 

activated carbon and steel anode. Industrial and Engineering 

Chemistry Research 38, 3779–3785. 

Carballa, M., Omil, F., Lema, J.M., Llompart, M., Garcia-Jares, C., 



treatment plant. Water Research 38, 2918–2926. 



treatment plants in Italy. Environmental Science and 


De Laat, J., Truong Le, G., Legube, B., 2004. A comparative study of 

the effects of chloride, sulfate and nitrate ions on the rates of 

decomposition of H 2O 2 and organic compounds by Fe(II)/H 2O 2 

and Fe(III)/H 2O 2. Chemosphere 55, 715–723. 

Esplugas, S., Rodríguez, M., Malato, S., Pulgarin, C., Contreras, S., 

Curcó, D., Giménez, J., 2005. Optimizing the solar photo Fenton 

process in the treatment of contaminated water. 

Determination of intrinsic kinetic constants for scale-up. Solar 

Energy 79, 360–368. 

Esplugas, S., Bila, D., Krause, L., Dezotti, M., 2007. Ozonation and 

advanced oxidation technologies to remove endocrine 

disrupting chemicals (EDCs) and pharmaceuticals and 

personal care products (PPCPs) in water effluents. A review. 

Journal of Hazardous Materials 149, 631–642. 

Farré, M.J., Franch, M.I., Malato, S., Ayllon, J., Peral, J., 

Doménech, X., 2005. Degradation of some biorecalcitrant 

pesticides by homogeneous and heterogeneous photocatalytic 

ozonation. Chemosphere 58, 1127–1133. 

Fono, L.J., Kolodziej, E.P., Sedlak, D.L., 2006. Attenuation of 

wastewater-derived contaminants in an effluent 

dominated river. Environmental Science and Technology 40, 

7257–7262. 

Fukushima, M., Tatsumi, K., 2001. Degradation pathways of 

pentachlorophenol by photo-Fenton systems in the presence 

of iron(III), humic acids and hydrogen peroxide. 


Gebhardt, W., Schröder, H., 2007. Liquid chromatography- 

(tandem) mass spectrometry for the follow-up of the 

elimination of persistent pharmaceuticals during wastewater 

treatment applying biological wastewater treatment and 

advanced oxidation. Journal of Chromatography A 1160, 34–43. 

Glaze, W., Kang, W., Chapin, H., 1987. Chemistry of water 

treatment processes involving ozone, hydrogen peroxide 

and ultraviolet radiation. Ozone Science & Engineering 9, 

335–352. 

Göbel, A., McArdell, C., Joss, A., Siegrist, H., Giger, W., 2007. Fate of 

sulfonamides, macrolides and trimethoprim in different 

wastewater treatment technologies. Science of the Total 

Environment 372, 361–371. 

Gogate, P.R., Pandit, A.B., 2004. A review of imperative 

technologies for wastewater treatment. I: oxidation 

technologies at ambient conditions. Advances in 

Environmental Research 8, 501–551. 

Gültekin, I., Ince, N.H., 2007. Synthetic endocrine disruptors in 

the environment and water remediation by advanced 

oxidation processes. Journal of Environmental Management 

85, 816–832. 

Hislop, K.A., Bolton, J.R., 1999. The photochemical generation of 

OH radicals in the UV–Vis/ferrioxalate/H2O2 system. 


Huber, M., Coronen, S., Ternes, T., von Gunten, U., 2005. 

Oxidation of pharmaceuticals during water treatment with 

chlorine dioxide. Water Research 39, 3607–3617. 

Jackson, J., Sutton, R., 2008. Sources of endocrine disrupting 

chemicals in urban wastewater, Oakland, CA. Science of the 

Total Environment 405, 153–160.

Kassinos, D., Varanva, N., Michael, C., Piera, P., 2009. 

Homogenous oxidation of aqueous solutions of atrazine and 

fenitrothion through dark and photo-Fenton reactions. 










waters. Water Research 41, 1013–1021. 

Klamerth, N., Miranda, N., Malato, S., Agüera, A., Fernández- 

Alba, A.R., 2009. Degradation of emerging contaminants at low 

concentrations in MWTPs effluents with mild solar photo- 

Fenton and TiO 2. Catalysis Today 144, 124–130. 

Kositzi, M., Poulios, I., Malato, S., Cáceres, J., Campos, A., 2004. 

Solar photocatalytic treatment of synthetic municipal 

wastewater. Water Research 38, 1147–1154. 

Lapertot, M., Ebrahimi, S., Dazio, S., Rubinelli, A., Pulgarin, C., 

2007. Photo-Fenton and biological integrated process for 

degradation of a mixture of pesticides. Journal of 

Photochemistry and Photobiology A: Chemistry 186, 34–40. 

Laville, N., Ait-Aissa, S., Gomez, E., Casellas, C., Porcher, J.M., 

2004. Effects of human pharmaceuticals on cytotoxicity, EROD 

activity and ROS production in fish hepatocytes. Toxicology 

196, 41–55. 

Lesko, T., Colussi, A., Hoffmann, M., 2006. Sonochemical 

decomposition of phenol: evidence for a synergistic effect of 

ozone and ultrasound for the elimination of total organic 

carbon from water. Environmental Science and Technology 

40, 6818–6823. 

Malato, S., Blanco, J., Alarcón, D.C., Maldonado, M.I., Fernández- 

Ibáñez, P., Gernjak, W., 2007. Photocatalytic decontamination 

and disinfection of water with solar collectors. Catalysis 

Today 122, 137–149. 

Mantzavinos, D., Vassilakis, C., Pantidou, A., Psillakis, E., 

Kalogerakis, N., 2004. Sonolysis of natural phenolic 

compounds in aqueous solutions: degradation pathways and 

biodegradability. Water Research 38, 3110–3118. 

Martinez, F., Calleja, G., Melero, J.A., Molina, R., 2005. 

Heterogeneous photo-Fenton degradation of phenolic 

aqueous solutions over iron containing SBA-15 catalyst. 

Applied Catalysis B: Environmental 60, 181–190. 

Martínez Bueno, M.J., Agüera, A., Gómez, M.J., Hernando, M.D., 

García-Reyes, J.F., Fernández-Alba, A.R., 2007. Application of 

liquid chromatography/quadrupole-linear ion trap mass 

spectrometry and time-of-flight mass spectrometry to the 

determination of pharmaceuticals and related 

contaminants in wastewater. Analytical Chemistry 79, 

9372–9384. 

Mitch, W.A., Sharp, J.O., Trussel, R.R., Valentine, R.L., Alvarez- 

Cohen, L., Seldak, D.L., 2003. N-Nitrosodimethylamine (NDMA) 

as a drinking water contaminant: A review. Environmental 

Engineering and Science 20, 389–404. 

Moncayo-Lasso, A., Pulgarin, C., Benítez, N., 2008. Degradation of 

DBPs’ precursors in river water before and after slow sand 

filtration by photo-Fenton process at pH 5 in a solar CPC 

reactor. Water Research 42, 4125–4132. 

Nakada, N., Kiri, K., Shinohara, H., Harada, A., Kuroda, K., 

Takizawa, S., Takada, H., 2008. Evaluation of pharmaceuticals 

and personal care products as water-soluble molecular 

markers of sewage. Environmental Science and Technology 

42, 6347–6353. 

Ning, B., Graham, N., Zhang, Y., Nakonechny, M., El-Din, M.G., 

2007. Degradation of endocrine disrupting chemicals by 

ozone/AOPs. Ozone Science & Engineering 29, 153–176. 


Neyens, E., Baeyens, J., 2003. A review of classic fenton’s 

peroxidation as an advanced oxidation technique. Journal of 

Hazardous Materials 98, 33–50. 

Nogueira, R.F.P., Oliveira, M.C., Paterlini, W.C., 2005. Simple and 

fast spectrophotometric determination of H 2O 2 in photo- 

Fenton reactions using metavanadate. Talanta 66, 86–91. 

OECD Guidelines for testing of Chemicals, Simulation Test- 

Aerobic Sewage Treatment 303A, 1999. 

Papadaki, M., Emery, R.J., Abu-Hassan, M.A., Diaz-Bustos, A., 

Metcalfe, I.S., Mantzavinos, D., 2004. Sonocatalytic oxidation 

processes for the removal of contaminants containing 

aromatic rings from aqueous effluents. Separation and 

Purification Technology 34, 35–42. 

Pelegrini, R.T., Freire, R.S., Duran, N., Bertazzoli, R., 2001. 

Photoassisted electrochemical degradation of organic 

pollutants on a DSA type oxide electrode: process test for 

a phenol synthetic solution and its application for the E1 

bleach kraft mill effluent. Environmental Science and 

Technology 35, 2849–2853. 

Petrovic, M., Gonzalez, S., Barceló, D., 2003. Analysis and removal 

of emerging contaminants in wastewater and drinking water. 

Trends in Analytical Chemistry 22, 685. 

Pignatello, E., Oliveros, A., McKay, 2006. Advanced oxidation 

processes for organic contaminant destruction based on the 

Fenton reaction and related chemistry. Critical Reviews in 


Prosen, H., Zupančič-Kralj, L., 2005. Evaluation of photolysis and 

hydrolysis of atrazine and its first degradation product in the 

presence of humic acids. Environmental Pollution 133, 517–529. 

Radjenović, J., Petrović, M., Barceloć, D., Petrović, M., 2007. 

Advanced mass spectrometric methods applied to the study of 

fate and removal of pharmaceuticals in wastewater 

treatment. Trends in Analytical Chemistry 26, 1132–1144. 

Richardson, S., 2007. Water analysis: emerging contaminants and 

current issues. Reviews in Analytical Chemistry 79, 4295–4324. 

Rosenfeldt, E.J., Linden, K.G., 2004. Degradation of endocrine 

disrupting chemicals bisphenol A, ethinyl estradiol and estradiol 

during UV photolysis and advanced oxidation processes. 


Rosenfeldt, E.J., Chen, P.J., Kullman, S., Linden, K.G., 2007. 

Destruction of estrogenic activity in water using UV advanced 

oxidation. Science of the Total Environment 377, 105–113. 

Saritha, P., Aparna, C., Himabindu, V., Anjaneyulu, Y., 2007. 

Comparison of various advanced oxidation processes for the 

degradation of 4-chloro-2-nitrophenol. Journal of Hazardous 

Materials 149, 609–614. 

United Book Press Inc.. In: Clesceri, L.S., Greenberg, A.E., Eaton, A. 

D. (Eds.), Standard Methods for the Examination of Water and 

Wastewater, 20th ed. American Public Health Association, 

American Water Works Association, Water Environment 

Federation, Baltimore, Maryland. 

Ternes, T., 1998. Occurrence of drugs in German sewage 

treatment plants and rivers. Water Research 32, 3245–3260. 

Ternes, T., Struber, J., Herrmann, N., McDowell, D., Ried, A., 


of pharmaceuticals contrast media and musk fragrances from 

wastewater. Water Research 37, 1976–1982. 

Ternes, T., Bonerz, M., Herrmann, N., Teiser, B., Andersen, R., 

2007. Irrigation of treated wastewaters in Braunschweig, 

Germany: an option to remove pharmaceuticals and musk 

fragrances. Chemosphere 66, 894–904. 

Teske, S.S., Arnold, R.G., 2008. Removal of natural and xenoestrogens 

during conventional wastewater treatment. Reviews 

in Environmental Science and Biotechnology 7, 107–124. 

Zapata, A., Oller, I., Bizani, E., Sánchez-Pérez, J.A., Maldonado, I., 

Malato, S., 2009. Evaluation of operational parameters 

involved in solar photo-Fenton degradation of a commercial 

pesticide mixture. Catalysis Today 144, 94–99.

554 

Zepp, R.G., Faust, B.C., Hoigné, J., 1992. Hydroxyl radical 

formation in aqueous reactions pH(3–8) of iron(II) with 

hydrogen peroxide: the photo-fenton reaction. Environmental 


Zhao, J.-L., Ying, G.-G., Wang, L., Yang, J.-F., Yang, X.-B., Yang, L.-H., 

Li, X., 2009. Determination of phenolic endocrine disrupting 

chemicals and acidic pharmaceuticals in surface water of the 


Pearl River in South China by GC-MS. Science of the Total 


Zhou, M., Dai, Q., Lei, L., Ma, C., Wang, D., 2005. Long life 

modified lead dioxide anode for organic wastewater 

treatment: electrochemical characteristics and degradation 

mechanism. Environmental Science and Technology 39, 

363–370.

Oxidative transformation of micropollutants during municipal 

wastewater treatment: Comparison of kinetic aspects of 

selective (chlorine, chlorine dioxide, ferrate VI , and ozone) 

and non-selective oxidants (hydroxyl radical) 

Yunho Lee a , Urs von Gunten a,b, * 

a 

Eawag, Swiss Federal Institute of Aquatic Science and Technology, Ueberlandstrasse 133, P.O. Box 611, 8600 Duebendorf, Switzerland 

b 

Institute of Biogeochemistry and Pollutant Dynamics, ETH Zurich, 8092 Zurich, Switzerland 

article info 


Received 1 October 2009 


13 November 2009 

Accepted 23 November 2009 

Available online 27 November 2009 

Keywords: 

Oxidation processes 

Ozone 

Chlorine 

Chlorine dioxide 

Ferrate VI 

Hydroxyl radical 




abstract 

Chemical oxidation processes have been widely applied to water treatment and may serve 

as a tool to minimize the release of micropollutants (e.g. pharmaceuticals and endocrine 

disruptors) from municipal wastewater effluents into the aquatic environment. The 

potential of several oxidants for the transformation of selected micropollutants such as 

atenolol, carbamazepine, 17a-ethinylestradiol (EE2), ibuprofen, and sulfamethoxazole was 

assessed and compared. The oxidants include chlorine, chlorine dioxide, ferrate VI , and 

ozone as selective oxidants versus hydroxyl radicals as non-selective oxidant. Secondorder 

rate constants (k) for the reaction of each oxidant show that the selective oxidants 

react only with some electron-rich organic moieties (ERMs), such as phenols, anilines, 

olefins, and deprotonated-amines. In contrast, hydroxyl radicals show a nearly diffusioncontrolled 

reactivity with almost all organic moieties (k 10 9 M 1 s 1 ). Due to a competition 

for oxidants between a target micropollutant and wastewater matrix (i.e. effluent 

organic matter, EfOM), a higher reaction rate with a target micropollutant does not 

necessarily translate into more efficient transformation. For example, transformation 

efficiencies of EE2, a phenolic micropollutant, in a selected wastewater effluent at pH 8 

varied only within a factor of 7 among the selective oxidants, even though the corresponding 

k for the reaction of each selective oxidant with EE2 varied over four orders of 

magnitude. In addition, for the selective oxidants, the competition disappears rapidly after 

the ERMs present in EfOM are consumed. In contrast, for hydroxyl radicals, the competition 

remains practically the same during the entire oxidation. Therefore, for a given oxidant 

dose, the selective oxidants were more efficient than hydroxyl radicals for transforming 

ERMs-containing micropollutants, while hydroxyl radicals are capable of transforming 

micropollutants even without ERMs. Besides EfOM, ammonia, nitrite, and bromide were 

found to affect the micropollutant transformation efficiency during chlorine or ozone 

treatment. 


* Corresponding author at: Eawag, Swiss Federal Institute of Aquatic Science and Technology, Ueberlandstrasse 133, P.O. Box 611, 

8600 Duebendorf, Switzerland. Tel.: þ41 44 823 5270; fax: þ41 44 823 5028. 

E-mail address: vongunten@eawag.ch (U. von Gunten). 


doi:10.1016/j.watres.2009.11.045

556 


Effluents of municipal wastewater treatment plants (WWTPs) 

have been identified as a major source of micropollutants, 

such as hormones, pharmaceuticals, and personal care 

products (Ternes and Joss, 2006). To minimize a discharge of 

micropollutants from WWTPs to the receiving waters and 

thus to prevent adverse ecological effects, there have been 

discussions on upgrading WWTPs with additional treatment 

steps such as activated carbon and ozonation (Nowotny et al., 

2007; Hollender et al., 2009). This is also desirable for the 

quality of drinking water if the wastewater-impacted 

receiving waters are used as a drinking water source via 

indirect potable reuse. 

The focus of the current study is on the chemical oxidation. 

Chemical oxidants, such as chlorine, chlorine dioxide, ozone, 

and hydroxyl radicals, have been widely applied for disinfection 

of drinking waters and wastewaters and often for the 

transformation/elimination of undesired micropollutants 

from drinking water (Letterman, 1999). Chlorine is worldwide 

one of the most commonly used disinfectant for drinking 

water and wastewater treatment (Letterman, 1999). Chlorine 

dioxide has been used as an alternative disinfectant for 

chlorine, often to minimize the formation of chlorine-based 

disinfection by-products (e.g. trihalomethane) (Gates, 1998; 

Letterman, 1999). The application of ozone in drinking water 

treatment is also widespread and the main reasons for its use 

are disinfection and oxidation (e.g. micropollutants transformation) 

or a combination of both (Hoigné, 1998; von 

Gunten, 2003a). Even though chlorine, chlorine dioxide, and 

ozone have been widely used for disinfection (Letterman, 

1999), their potentials for transforming micropollutants in 

wastewaters have received attention only in recent years 

(Ternes and Joss, 2006). In addition, ferrate VI is an emerging 

water treatment chemical due to its dual functions as an 

oxidant and a subsequent coagulant/precipitant as ferric 

hydroxide (Lee et al., 2009). Finally, the hydroxyl radical is 

a primary oxidant in advanced oxidation processes (AOPs) 

such as UV/H2O2 and O3/H2O2 and their application are 

considered for transforming various micropollutants in 

drinking waters as well as wastewaters for water reuse 

(Letterman, 1999; Asano et al., 2007). The hydroxyl radical is 

quite different from other oxidants with respect to its high 

reactivity and thus low selectivity (Buxton et al., 1988). In this 

paper, chlorine, chlorine dioxide, ferrate VI and ozone will be 

referred as selective oxidants as opposed to non-selective 

hydroxyl radicals. 

The oxidants mentioned above can achieve a transformation 

of various micropollutants during wastewater 

treatment. The transformation efficiency mainly depends on: 

(i) the reactivity of the oxidant to target micropollutants; and 

(ii) towards matrix components present in water (e.g. dissolved 

organic matter, DOM) which determines the stability of 

the oxidant. In general, the reactions of the oxidants follow 

second-order kinetics, i.e. the reaction rate is proportional to 

the concentrations of both reactants. The corresponding 

second-order reaction rate constants (k) range from 1to 

10 10 M 1 s 1 in aqueous solution (Buxton et al., 1988; Deborde 

and von Gunten, 2008; Lee et al., 2009; Neta et al., 1988; von 


Gunten, 2003a). Currently, k values are available for a wide 

range of organic compounds including many micropollutants 

(Fig. 1 and Table 1 as examples). For ozone reactions more 

than 500 k values have been measured (Neta et al., 1988 von 

Gunten, 2003a), and a few thousand k values are available for 

hydroxyl radical reactions (Buxton et al., 1988). There are also 

a few hundred k values available for chlorine, chlorine 

dioxide, and ferrate VI (Deborde and von Gunten, 2008; Lee 

et al., 2009; Neta et al., 1988). These second order rate 

constants show that the selective oxidants react preferentially 

with electron-rich organic moieties (ERMs), such as activated 

aromatic compounds (i.e. phenol, aniline, and polycyclic 

aromatics), organosulfur compounds, and deprotonated 

amines. Additionally, ozone and ferrate VI react with olefins. 

Contrary to the selective oxidants, hydroxyl radicals react 

with almost all organic moieties (e.g. aliphatic CeH bond) with 

nearly diffusion-controlled rates (i.e. k 10 9 M 1 s 1 ). 

Therefore, micropollutants containing the ERMs 

mentioned above can be potentially transformed by the 

selective oxidants. On the contrary, hydroxyl radicals can be 

effective for transforming any type of micropollutant. 

However, a high reactivity of the oxidants to the ERMs is also 

responsible for a rapid oxidant consumption by the ERMs 

included in the matrix components, such as DOM. In addition 

to DOM, some inorganic species such as ammonia, nitrite, 

sulfide, iron(II), and manganese(II) can be potential 

consumers of the selective oxidants. In the case of hydroxyl 

radicals, their rapid consumption by almost all types of 

matrix components can lead to a significantly lower transformation 

efficiency of micropollutants than the selective 

oxidants. Therefore, the kinetic information (i.e. k values) of 

the oxidants with the ERMs and some other moieties is 

crucial to assess the transformation efficiency of structurallydiverse 

micropollutants by a given oxidant in a real wastewater 

matrix. 

The aim of this study was to compare each of the selective 

oxidants and non-selective hydroxyl radicals with respect to 

their efficiency for transforming micropollutants during the 

treatment of the wastewater effluents. This comparison is 

important to provide criteria for choosing an appropriate 

oxidation process for a specific class of micropollutants based 

on the chemical structure (i.e. presence of ERM). For this 

purpose, the following issues were addressed: 

The k of each oxidant with selected organic model 

compounds and micropollutants (containing various ERMs) 

available in literature were summarized and compared. 

The oxidant consumption kinetics in a wastewater effluent 

were determined. 

The oxidative transformations of selected micropollutants 

containing various ERMs were conducted in a wastewater 

effluent for varying oxidant doses (0–150 mM). The transformation 

efficiency was then compared for the oxidants 

based on the required oxidant dose to achieve a certain 

degree of a micropollutant transformation. 

The effect of ammonia, nitrite, and bromide was investigated 

in respect to the transformation efficiency of selected 

micropollutants during the treatment of a wastewater 

effluent by the selected oxidants and hydroxyl radicals.

-1 -1 

s 

a b c 

k, 

M 

10 

10 

10 8 

10 6 

10 4 

10 2 

10 0 

HO. 

d e f 

-1 -1 

s 

k, 

M 

10 10 

10 8 

10 6 

10 4 

10 2 

10 0 

HO. 

HFeO 4 – 


2.1. Reagents and wastewater 

Atenolol (ATL), carbamazepine (CBZ), 17a-ethinylestradiol 

(EE2), ibuprofen (IBP), and sulfamethoxazole (SMX) were 

obtained from Sigma–Aldrich with a purity higher than 98%. 

Stock solutions of these pharmaceuticals were prepared with 

Milli-Q purified water (Barenstad). Other chemicals and 

solvents were purchased from various commercial suppliers 

and used as received. 

A 10-L grab sample of secondary wastewater effluent 

(RDWW) was obtained from the Regensdorf wastewater 

treatment plant near Zurich, Switzerland. The sample was 

transported to the laboratory within several hours of sampling 

and vacuum-filtered with a 0.45 mm cellulose nitrate 

membrane prior to storage at 4 C. RDWW has a DOC of 5.0 mg 

CL 1 , 1.2 mM ammonia (17 mg NL 1 ), 6 mM nitrite (84 mg NL 1 ), 

5.3 mM carbonate alkalinity, 0.35 mM (28 mg L 1 )ofBr and 

apHofw8. The concentrations of selected micropollutants 

(ATL, CBZ, EE2, IBP, and SMX) originally present in RDWW 

O 3 

phenol aniline butenol (olefin) 

ClO 2 

HOCl 

glycine 

o (1 amine) 

HOCl 

O 3 

HFeO 4 – 


HO. HO. 

HO. 

HOCl 

HFeO 4 – 

HFeO 4 – 

O 3 

ClO 2 

HOCl 

dimethylamine 

o (2 amine) 

5 6 7 8 9 10 5 6 7 8 9 10 5 6 7 8 9 10 

pH pH pH 

O 3 

ClO 2 

were negligible relative to the concentrations spiked for 

oxidation experiments, which were 0.2–1 mM for each 

substrate. 

2.2. Oxidants 

HO. 

HFeO 4 – 

Stock solutions of chlorine (5–20 mM) were prepared by 

diluting a commercial solution of sodium hypochlorite (10% 

active chlorine, Riedel-deHaen, Germany). Chlorine dioxide 

(w5 mM) was produced by mixing potassium peroxodisulfate 

(K 2S 2O 8) with sodium chlorite (NaClO 2) according to a method 

described by Gates (1998). Ozone was produced with a Fischer 

502 ozone generator and its stock solutions (w1.5 mM) were 

produced by sparging ozone-containing oxygen through Milli- 

Q water that was cooled in an ice bath (Bader and Hoigné, 

1981). Potassium ferrate VI of high purity (w90%) was prepared 

by the method of Thompson et al. (1951). Stock solutions of 

ferrate VI (2 mM) were freshly prepared by dissolving solid 

samples of potassium ferrate VI (K2FeO4) in Milli-Q water and 

subsequent filtration (pH z 9.2). Hydroxyl radicals were in 

situ generated by photolysis of H2O2 (0.2–5 mM). Irradiations 

O 3 

trimethylamine 

o (3 amine) 

Fig. 1 – pH dependent second-order rate constants (k) for the reaction of the oxidants, chlorine (HOCl), chlorine dioxide (ClO2), 

ferrate VI (HFeO4 L ), hydroxyl radicals (HO ), and ozone (O3) with (a) phenol, (b) aniline, (c) butenol, (d) glycine, (e) 

dimethylamine, and (f) trimethylamine. k values for chlorine: Deborde and von Gunten, (2008); for chlorine dioxide and 

ozone: Neta et al., (1988); for ferrate VI : Lee et al. (2005a, 2008, 2009); and for hydroxyl radicals: Buxton et al., (1988). Speciesspecific 

second-order rate constants and the pKa values of the compounds used to generate the data are provided in 

Supplementary Information, SI-excel. 

O 3 

ClO 2 

HOCl

558 

Table 1 – Second-order rate constants for the reaction of oxidants with selected pharmaceuticals and inorganic species 

(ammonia, nitrite, and bromide). a 

17a-Ethinylestradiol (EE2), 

synthetic steroid estrogen 

Sulfamethoxazole (SMX), 

sulfonamide antibiotic 

Carbamazepine (CBZ), 

antiepileptic drug 

Atenolol (ATL), b-blocker 

Ibuprofen (IBP), antiphlogistic 


Compounds pK a Oxidant k (species 

of compound), 

M 1 s 1 

k at pH 8, 

M 1 s 1 

10.4 HO 9.8 10 9 

9.8 10 9 

O3 1.8 10 5 (neutral), 

3.7 10 9 1.5 10 

(anion) 

7 

ClO2 4.6 10 8 (anion) 1.8 10 6 

HFeO4 9.4 10 2 (neutral), 

5.4 10 5 4.2 10 

(anion) 

2 

HOCl 3.5 10 5 (anion) 3.4 10 2 

Ref. 

Huber et al. (2003) 

Deborde et al. (2005) 


Lee et al. (2005a) 

Deborde et al. (2004) 

5.7 HO 5.5 10 9 

5.5 10 9 


O3 4.7 10 4 (neutral), 

5.7 10 5 5.7 10 

(anion) 

5 

Dodd et al. (2006) 

ClO2 7.9 10 3 (anion) 7.9 10 3 


HFeO4 9.2 10 3b 

95 Lee et al. (2009) 

HOCl 1.1 10 3 (neutral), 

2.4 10 3 5.7 10 

(anion) 

2 

Dodd and Huang 

(2004) 

– HO 8.8 10 9 

O3 3.0 10 5 

ClO 2 


Compounds pKa Oxidant k (species 

of compound), 

M 1 s 1 

were performed in quartz tubes using a merry-go-round 

photoreactor (Hans Mangels GmbH, Bornheim-Roisdor, 

Germany), equipped with a low pressure mercury lamp (Heraues 

Noblelight model TNN 15/32, nominal power 15 W) 

emitting essentially monochromatic light at l ¼ 254 nm. In the 

case of SMX, a medium pressure mercury lamp (Heraues 

Noblelight model TQ 718, nominal power 500 W) with a 0.25 M 

sodium nitrate filter solution producing light at l > 320 nm 

was used to minimize the direct photo-transformation of SMX 

(less than 10%). Fluence and fluence rate were determined by 

chemical actinometry at low optical density using 5 mM 

aqueous atrazine as an actinometer described by Meunier 

et al. (2006). Stock solutions of each oxidant were standardized 

spectrophotometrically based on their molar absorption 

coefficient: 3 ¼ 350 M 1 cm 1 at 292 nm for OCl (Johnson and 

Margerum, 1991), 3 ¼ 1230 M 1 cm 1 at 359 nm for ClO2 (Gates, 

1998), 3 ¼ 3000 M 1 cm 1 at 260 nm for O3 (Liu et al., 2001), 

3 ¼ 1150 M 1 cm 1 at 510 nm for FeO4 2 

(Lee et al., 2005b), and 

3 ¼ 40 M 1 cm 1 at 240 nm for H2O2 (Bader et al., 1988). 

2.3. Oxidant consumption kinetics and transformation 

of selected micropollutants in wastewater 

Oxidant consumption experiments for the selective oxidants 

were performed in a secondary wastewater effluent (RDWW) 

at pH 8 (20 mM borate buffer) in 50 mL reaction volume. Forty 

to 45 mM of each oxidant was applied to RDWW and 0.2–1 mL 

of the reaction solution was sampled for the measurements of 

residual oxidants over a 1 h reaction time. 

Transformation of the selected micropollutants was tested 

in RDWW spiked with each micropollutant and treated by 

each oxidant at various oxidant doses (0–150 mM) at 24 1 C 

and pH 8. The micropollutants, ATL (amine), CBZ (olefin), EE2 

(phenol), and SMX (aniline), were selected as representative 

substances with different ERMs that belong to different usage 

classes and have an environmental relevance (Ternes and 

Joss, 2006). IBP was chosen as a model micropollutant without 

the ERMs. Each micropollutant was spiked separately at 

concentrations of 0.2–1 mM into 20 mL of the wastewater 

buffered at pH 8. For the selective oxidants, stock solutions of 


k at pH 8, 

M 1 s 1 

Br bromide – HO w1.1 10 9 

w1.1 10 9 

von Gunten (2003b) 

O3 2.6 10 2 

2.6 10 2 

Liu et al. (2001) 

ClO2

560 

irradiation time. For details of the kinetic method, see 

Supplementary Information, SI-word. 


Oxidant concentrations in wastewater effluents were 

measured by several colorimetric methods. Chlorine and 

chlorine dioxide were determined by the ABTS method 

developed by Pinkernell et al. (2000); ozone by the indigo 

method (Bader and Hoigné, 1981); ferrate VI by the ABTS 

method described by Lee et al. (2005b); H2O2 by the DPD/ 

peroxidase method (Bader et al., 1988). Micropollutant (or 

pCBA) concentrations at sub to low mM range were determined 

with an Agilent 1100 series HPLC system equipped with 

a Nucleosil 100-5 C18 column (125 4 or 250 4 mm) and a UV 

(diode array) and/or fluorescence detector. Separation was 

achieved by different gradient programs with the aqueous 

phase consisting either of 10 mM phosphoric acid or 10 mM 

citric acid (pH 5) and the solvent phase either of acetonitrile or 

methanol. The injection volume was 100 mL and the flow rate 

was set to 1 mL min 1 . The repeatability was determined to be 

4% while limit of quantifications (LOQ) were inferred from the 

lowest standard concentrations used. The limits of quantification 

were in all cases sufficient to determine a transformation 

of micropollutants of more than 96% ( 1.4 logtransformation). 


3.1. Rate constants of oxidants with selected organic 

compounds 

Fig. 1 summarizes the pH-dependent k for the reaction of the five 

oxidants (chlorine, chlorine dioxide, ferrate VI , ozone, and 

hydroxyl radical) with the selected ERM-containing model 

compounds, such as phenol and aniline (activated aromatic 

compounds), butenol (olefin), glycine (primary (1 ) amine), 

dimethylamine (secondary (2 ) amine), and trimethylamine 

(tertiary (3 ) amine). If the k value is lower than 1 M 1 s 1 ,itisnot 

shown in Fig. 1. The species-specific k values which were used to 

generate Fig. 1 are provided in Supplementary Information, SIexcel. 

For more details on the kinetics of each oxidant, refer the 

corresponding references provided in SI-excel. 

The k values for phenol and aniline, as two aromatic 

compounds, decrease in the order of hydroxyl radical - 

> ozone > chlorine dioxide > chlorine z ferrate VI (Figs. 1a,b). 

The k values for the other compounds show a slightly different 

reactivity order. Chlorine and chlorine dioxide are not reactive 

to butenol, therefore the reactivity trend for butenol is 

hydroxyl radical > ozone > ferrate VI (Fig. 1c). For three amine 

compounds, chlorine shows a particular high reactivity 

whereas chlorine dioxide has a low reactivity. For glycine as 1 

amine, the k values decrease in the order of hydroxyl radical 

> chlorine > ozone > ferrate VI > chlorine dioxide (Fig. 1d). 

For dimethylamine as 2 amine, the reactivity order is 

hydroxyl radical > ozone z chlorine > ferrate VI > chlorine 

dioxide (Fig. 1e). Finally, for trimethylamine as 3 amine, the 

reactivity order is hydroxyl radical > ozone > chlorine dioxide 

chlorine > ferrate VI (Fig. 1f). 


In the pH range 5–10, hydroxyl radicals react with phenol, 

aniline, and butenol with nearly diffusion-controlled rates. 

However, the k values for the reaction with three amine 

compounds are one or two orders of magnitude lower (10 7 – 

10 8 M 1 s 1 at pH 6–8). This is due to the relative low reactivity of 

hydroxyl radicals with the protonated-amine moieties. Nevertheless, 

the k of hydroxyl radicals with the amine compounds 

with longer hydrocarbon-chains can be higher than 10 9 M 1 s 1 . 

This is due to the high reactivity of hydroxyl radical with CeH 

bonds (k [ 10 8 M 1 s 1 ) (Buxton et al., 1988). In addition, 

hydroxyl radicals react with benzenes (i.e. non-activated 

aromatics) with k > 10 9 M 1 s 1 (Buxton et al., 1988). Ozone 

reacts fast with all ERMs shown in Fig. 1:thek values decrease in 

the order of aniline phenol > butenol dimethylamine z 

trimethylamine > glycine in the pH range 6–8. Chlorine dioxide 

shows a high reactivity only with phenol, aniline, and trimethylamine. 

Chlorine is reactive to all ERMs except butenol. The k 

values decrease in the order of glycine > dimethylamine 

aniline > phenol z trimethylamine in the pH 

range 6–8. Finally, ferrate VI reacts with all ERMs except trimethylamine. 

The k values decrease in the order of aniline > phenol 

z glycine > butenol dimethylamine in the pH range 6–8. 

For the selective oxidants, ozone shows the widest range of k 

values (i.e. w7 orders of magnitude) while ferrate VI shows the 

narrowest range of k values (i.e. w4 orders of magnitude) toward 

the selected ERMs in the pH range 6–8. 

3.2. Consumption kinetics of the selective oxidants in 

a wastewater effluent 

Fig. 2 shows the consumption kinetics of the selective 

oxidants (ozone, ferrate VI , chlorine, and chlorine dioxide) in 

a secondary wastewater effluent from Regensdorf, 

Switzerland (RDWW) at pH 8 for an oxidant dose of 40–45 mM. 

The selected RDWW has a low concentration of inorganic 

nitrogen species ([ammonia] ¼ 1.2 mM and [nitrite] ¼ 6 mM), 

therefore the consumption of each oxidant is mainly caused 

by EfOM ([DOC] ¼ 5mg C L 1 ). Except ferrate VI , all oxidants 

showed fast consumptions within 2 min after the oxidant 

addition. The stability of the oxidants follows the order of 

ferrate VI z chlorine dioxide chlorine [ ozone. Ozone was 

depleted in less than 2 min (Fig. 2a, inset), whereas the other 

oxidants were slowly consumed within 60 min. The 

observed low stability of ozone is consistent with its high 

reactivity to all ERMs shown in Fig. 1. Chlorine and chlorine 

dioxide showed a similar decay pattern: a rapid decrease 

within 2 min after the oxidant addition (termed ‘initial 

phase’), followed by a slow decrease over 60 min of reaction 

time (termed ‘second phase’). Based on the k values shown in 

Fig. 1, the chlorine consumption in the initial phase is 

expected to be mainly caused by aniline- and 1 &2 aminemoieties 

while in the second phase phenolic- and 3 aminemoieties 

become important. In the case of chlorine dioxide, 

the consumption in the initial phase might be mainly caused 

by phenolic- and aniline-moieties while in the second phase 

by 2 &3 amine-moieties. Contrary to chlorine and chlorine 

dioxide, ferrate VI decreased smoothly along the reaction time, 

which is consistent with its narrow distribution of k values 

with the ERMs shown in Fig. 1. The achieved oxidant exposures 

for the conditions shown in Fig. 2 were 4.5 10 2 M 1

a 

C oncentration, 

µM 

c 


µM 

50 

40 

30 

20 

10 

0 

50 

40 

30 

20 

10 

0 

0 

0.0 0.5 1.0 1.5 2.0 2.5 3.0 

Time, min 

0 10 20 30 40 50 60 

0 10 20 30 40 50 60 

Time, min 

s 1 for ferrate VI and chlorine dioxide, 3.6 10 2 M 1 s 1 for 

chlorine, and 6.0 10 4 M 1 s 1 for ozone. 

3.3. Transformation of selected micropollutants in 

a wastewater effluent: comparison between selective 

oxidants and non-selective hydroxyl radicals 


µM 

A comparison was made for the transformation efficiency of 

selected micropollutants in a secondary wastewater effluent 

(RDWW) by the selective oxidants (chlorine, chlorine dioxide, 

ferrate VI , and ozone) and the non-selective hydroxyl radicals. 

The selected micropollutants include compounds with ERMs, 

such as 17a-ethinylestradiol (EE2, phenol), sulfamethoxazole 

(SMX, aniline), carbamazepine (CBZ, olefin), and atenolol (ATL, 

2 amine), and ibuprofen (IBP, no ERM). For the selective 

oxidants, the reaction time of 1 h was given to simulate realistic 

treatment conditions. Ozone was completely consumed 

after 1 h whereas for chlorine, chlorine dioxide, and ferrate VI , 

more than 10% of the oxidant concentration remained after 

1 h if the oxidant doses were higher than 40 mM (Fig. 2). For 

hydroxyl radicals, the oxidant doses were quantified by the 

kinetic method described in the Supplementary Information, 

SI-word. Fig. 3 shows the logarithm of the relative residual 

concentration of each selected micropollutant during treatment 

of RDWW at pH 8 as a function of the oxidant doses. The 

transformation efficiency of micropollutants varied significantly 

depending on the types of the selective oxidants and 

50 

40 

30 

20 

10 


b 

ozone ferrateVI 0 10 20 30 40 50 60 

d 

chlorine chlorine dioxide 

0 10 20 30 40 50 60 

Time, min 

Fig. 2 – Consumption kinetics of the selective oxidants, (a) ozone, (b) ferrate VI , (c) chlorine, and (d) chlorine dioxide, in 

a secondary wastewater effluent (RDWW) at pH 8. Symbols represent measured data and lines connect each data point to 

show the trend. 

micropollutants, and will be discussed in the following 

sections. 

3.3.1. EE2 and SMX 

For these two micropollutants containing activated aromatic 

moieties, all four selective oxidants were more efficient than 

hydroxyl radicals in terms of the oxidant doses required to 

achieve more than one-log (90%) transformation (Figs. 3a,b). 

For the selective oxidants, especially in the case of chlorine 

and ferrate VI , there was a lag-phase at lower oxidant doses 

where the micropollutant concentration decreased little with 

increasing oxidant doses. The low transformation efficiency 

in the lag-phase is attributed to the high competition for the 

selective oxidants between EE2/SMX and the ERMs of EfOM. 

With an increase of the selective oxidant doses higher than 

the initial consumption by ERMs, the residual concentration of 

EE2 and SMX started to decrease significantly. This can be 

understood by the depletion of the most reactive ERMs of 

EfOM and the corresponding smaller competition for EE2 and 

SMX toward the selective oxidants. For hydroxyl radicals, on 

the contrary, the log-relative residual concentration of EE2 

and SMX was linearly proportional to the applied hydroxyl 

radical dose. This indicates that the magnitude of the 

competition for hydroxyl radicals between micropollutants 

and wastewater effluent matrix remained constant during the 

entire oxidation. Therefore, it can be concluded that the initial 

oxidation products of EfOM by hydroxyl radicals have similar 

k values for their reaction with hydroxyl radicals compared to

562 

l 0 

og 

C/ 

C 

log C/ 

C0 

a 

0 

-1 

17a-ethinylestradiol (EE2) 

0 

Sulfamethoxazole (SMX) 

0 

Carbamazepine (CBZ) 

ClO2 HOCl 

O3 - 

HFeO4 ClO2 -2 

HOCl 

ClO2 -2 

O3 HOCl 

-2 

O3 - 

HFeO4 0 20 40 60 80 100 0 30 60 90 120 150 0 30 60 90 120 150 

Oxidant dose, µM Oxidant dose, µM Oxidant dose, µM 

0 

-1 

O 3 

HO. 

Atenolol (ATL) Ibuprofen (IBP) 

0 

ClO2 HOCl 

ClO2HOCl O3 + t-BuOH 

HO. 

- 

HFeO4 -2 

0 30 60 90 120 150 

-2 

0 30 60 90 120 150 

-2 

0 30 60 90 120 150 

Oxidant dose, µM Oxidant dose, µM Oxidant dose, µM 

the parent EfOM. This leads to a constant consumption rate of 

hydroxyl radicals. Hence, selective oxidants can be much 

more efficient than hydroxyl radicals especially when a higher 

degree of transformation is required (e.g. >90%) for EE2 or 

SMX. The characteristics of transformations of EE2 and SMX 

by hydroxyl radicals (i.e. a linear log-transformations vs. 

hydroxyl radical doses) were also observed for CBZ, ATL, and 

IBP (Figs. 3c–e, respectively). 

For EE2 (Fig. 3a), the transformation efficiency to achieve an 

one-log transformation for the oxidants was in the order of 

chlorine dioxide (3 mM) > ozone z ferrate VI (10 mM) > chlorine 

(20 mM) > hydroxyl radical (50 mM) where the value in the 

parenthesis shows the required oxidant dose. The very high 

efficiency of chlorine dioxide for EE2 transformation can be 

understood by its high reactivity to phenolic moieties (Fig. 1a), 

but relative low reactivity to amine moieties (Figs. 1d–f). Since 

phenolic and amine moieties are expected to be the two main 

ERMs present in EfOM, chlorine dioxide can be selective for 

the transformation of the micropollutants containing 

phenolic moieties in presence of EfOM. In contrast, the relatively 

low efficiency of chlorine for EE2 transformation can be 

understood by its high reactivity to amine moieties (Figs. 1d–f), 

but low reactivity to phenolic moieties (Fig. 1a). 

The one-log transformation efficiency in the case of SMX 

(Fig. 3b) was in the order of chlorine dioxide (15 mM) z ozone 

(20 mM) > ferrate VI 

(45 mM) z chlorine (48 mM) > hydroxyl 


b c 

-1 

-1 

HO. 

HO. HO. 

-1 

- 

HFeO4 d e f 

O 3 

- 

HFeO4 0 

-1 

p-chlorobenzoate (pCBA) 

Fig. 3 – Logarithm of the residual concentrations (log(c/c 0)) of selected micropollutants as a function of oxidant doses in 

a secondary wastewater effluent (RDWW) at pH 8: (a) EE2, (b) SMX, (c) CBZ, (d) ATL, and (e) IBP. Symbols represent measured 

data and lines connect each data point to show the trend. The lines for hydroxyl radicals represent the linear regression of 

data. For the selective oxidants, the reaction time of 1 h was given to simulate realistic treatment conditions. Hydroxyl 

radicals were in situ produced by UV254 nm/H2O2 photolysis and their doses were quantified by the kinetic method described 

in the Supplementary Information, SI-word. For (e) IBP, ‘O3 D tBuOH’ indicates that the ozonation was conducted in 

presence of 60 mM tBuOH to exclude the contribution of hydroxyl radicals to the IBP transformation. 

HO. 

radical (80 mM). The relative lower transformation of SMX 

compared to EE2 at the same oxidant dose (Figs. 3a vs. b) are 

consistent with the lower k values of SMX than EE2 for the 

reaction with the oxidants at pH 8 except the case of chlorine 

(Table 1). The sulfonyl-moiety deactivates the aniline-moiety 

of SMX toward oxidation as an electron-withdrawing group. 

Therefore, the k values for the reaction of SMX with the 

selective oxidants are lower than those for aniline (SI-excel). 

In contrast, the cyclic ring-moiety activates the phenolicmoiety 

of EE2 toward oxidation as an electron-donating group. 

Therefore, the k values for the reaction of EE2 with the 

selective oxidants are higher than those for phenol (SI-excel). 

The smaller transformation of SMX than EE2 at the same 

chlorine doses (Figs. 3a vs. b) is not consistent with the 1.7-fold 

higher k for the reaction of chlorine with SMX at pH 8 (Table 1). 

Dodd et al. found that the reaction of chlorine with SMX 

produced the N-chlorinated SMX (i.e. chlorine-adduct to the 

aniline-moiety of SMX) as an initial product (Dodd and Huang, 

2004). The k values reported in Table 1 correspond to the rate 

constant of this initial reaction step. The N-chlorinated SMX 

can be reduced back to the parent, SMX, upon its reaction with 

reductants such as thiosulfate, which was employed as 

a quenching reagent in this study. Therefore, the actual extent 

of SMX transformation during chlorination could be smaller 

than the predictions based on the k values corresponding to 

the initial chlorine addition to SMX. Finally, even though the k

values of ozone or chlorine dioxide for its reaction with EE2 

and SMX are several orders of magnitude higher than those of 

chlorine or ferrate VI (Table 1), the transformation efficiency in 

RDWW among the selective oxidants differed only within 

a factor of 20. This is caused by their competitive consumption 

by micropollutants and EfOM. 

3.3.2. CBZ 

The one-log transformation efficiency in the case of CBZ 

(Fig. 3c) was in the order of ozone (30 mM) > ferrate VI 

(43 mM) > hydroxyl radical (100 mM) [ chlorine z chlorine 

dioxide. Chlorine and chlorine dioxide achieved little transformations 

of CBZ (

564 

3.5.2. Ammonia and nitrite 

In addition to EfOMs, inorganic nitrogen compounds, such as 

ammonia and nitrite, commonly present in poorly-nitrified/ 

denitrified wastewater effluents can rapidly consume some 

oxidants and affect the transformation efficiency of micropollutants. 

The effect of ammonia and nitrite was investigated 

using EE2 as an example. From preliminary experiments, 

oxidant doses to achieve w80% transformation of EE2 in 

RDWW were determined. As a next step, the RDWW spiked 

with additional ammonia or nitrite (up to 100 mM) was treated 

with the pre-determined oxidant dose. In the case of the 

selective oxidants, the residual EE2 concentration was determined 

after 1 h of reaction time. 

Fig. 4a shows that in the case of chlorine, ammonia 

significantly decreased the degree of transformation of EE2. At 

an ammonia concentration higher than 20 mM ([NH4 þ ]0/ 

[HOCl]0 1), the EE2 transformation was less than 10%. This is 

because chlorine reacts rapidly with ammonia, forming 

chloramines (Deborde and von Gunten, 2008). When the 

þ 

[NH4 ]0/[HOCl] 0 1, monochloramine (NH2Cl) is the predominant 

chloramine species which has a low reactivity to EE2 

(k ¼ 0.24 M 1 s 1 ) (Lee et al., 2008). For the other selective 

oxidants and hydroxyl radicals, the EE2 transformation was 

not affected by the presence of ammonia within the tested 

concentration. This is consistent with the low reactivity of 

these oxidants with ammonia (Table 1). 

Fig. 4b shows that nitrite significantly reduced the transformation 

of EE2 in the case of chlorine and ozone, which is 

consistent with the high k values of these two oxidants with 

nitrite (Table 1). For chlorine, the EE2 transformation decreased 

from 80 to 40% with an increase of the nitrite concentration 

from 6 to 30 mM ([NO2 ] 0/[HOCl] 0 ¼ from 0.3 to 1.5) and remained 

rather constant at w40% with a further increase of the nitrite 

concentration even up to 100 mM ([NO2 ]0/[HOCl]0 ¼ 5). The 

oxidation of nitrite by chlorine to nitrate proceeds through 

some reactive intermediates such as NO2Cl and NO2 þ (Johnson 

% transformation 

+ 

[NH4 ]0 /[HOCl] 0 

0 1 2 3 4 5 

100 

80 

60 

40 

20 

HOCl 

O 3 ClO 2 Fe(VI) HO. 

0 

0 20 40 60 80 100 

+ 

[NH4 ]0 , µM 


a b 

100 

80 

60 

40 

20 

and Margerum, 1991). Reactions of these intermediates with 

EE2 might be responsible for the incomplete inhibition of EE2 

transformation by chlorine at the high nitrite concentration. 

For ozone, the EE2 transformation decreased gradually from 80 

to 20% with an increase of the nitrite concentration from 6 to 

100 mM ([NO2 ]0/[O3]0 ¼ from 0.8 to 12.5). Nitrite had little effect 

on the EE2 transformation in the case of chlorine dioxide and 

ferrate VI , which is consistent with the relatively low k values of 

these two oxidants with nitrite (Table 1). For hydroxyl radical, 

nitrite has a slightly positive effect on EE2 transformation even 

though it reacts rapidly with nitrite ( OH þ NO 2 / OH þ NO 2 , 

k ¼ 10 10 M 1 s 1 , Table 1). In RDWW spiked with nitrite at the 

concentration of 100 mM, w80% of the produced hydroxyl 

radicals is scavenged by nitrite during the initial phase of the 

UV/H2O2 treatment. This is based on the relative reaction rate 

of hydroxyl radicals with nitrite vs. the remaining matrix 

components (i.e. EfOM, H2O2, and carbonate). Therefore, the 

slightly enhanced transformation of EE2 in presence of 100 mM 

nitrite might be explained by the reaction of EE2 with some 

intermediates from the reaction of hydroxyl radical with 

nitrite such as NO 2 (i.e. EE2 þ NO 2 / EE2 oxid þ NO 2 ). The k 

value for the reaction of NO 2 with EE2 is estimated to be 

w10 5 M 1 s 1 at pH 8 from the k value for the reaction of NO 2 

with 4-methylphenol (Neta et al., 1988). 

3.5.3. Bromide 

Bromide is present in waters and wastewaters at concentrations 

of 10 to several hundred mgL 1 (von Gunten, 2003b). Since 

the oxidation of bromide typically produces bromine, which is 

highly reactive to phenol- and amine-moieties, the presence of 

bromide during oxidative wastewater treatment can affect the 

transformation efficiency of micropollutants. Among 

the selective oxidants, only chlorine and ozone react with 

bromide (Table 1), generating bromine. Since bromine is about 

three orders of magnitude more reactive toward phenols than 

chlorine (Lee and von Gunten, 2009), transformation of 

− 

[NO2 ]0 /[O3 ] 0 

0 2 4 6 8 10 12 

− 

[NO2 ]0 /[HOCl] 0 

0 1 2 3 4 5 

0 

0 20 40 60 80 100 

− 

[NO2 ]0 , µM 

Fig. 4 – Effect of (a) ammonia (NH 4 D ) and (b) nitrite (NO2 L ) on the transformations of EE2 during treatment of a secondary 

wastewater effluent (RDWW) by different oxidants at pH 8. Preliminary experiments were conducted to determine the 

oxidant dose for each oxidant to achieve a 80% transformation of EE2 in RDWW without additionally spiked ammonia and 

nitrite. They were 20 mM for chlorine, 3 mM for chlorine dioxide, 8 mM for ozone, 8 mM for ferrate VI , and 37 mM for hydroxyl 

radicals. Symbols represent measured data and lines connect each data point to show the trend. 

HO. 

ClO 2 

Fe(VI) 

HOCl 

O 3

phenolic micropollutants can be significantly enhanced during 

chlorination of bromide-containing waters. A recent study 

showed that bromine produced from bromide was mainly 

responsible for 17a-ethinylestradiol transformation during 

chlorination of a wastewater effluent (Lee and von Gunten, 

2009). In the case of ozone, most phenolic- and amine-moieties 

are already oxidized by ozone before the significant formation 

of bromine from bromide. Hence, the presence of bromide 

affects little the transformation efficiency of micropollutants 

during ozonation. However, an enhanced transformation of 1 

amine-containing micropollutants is expected due to the 

relatively low reactivity of ozone versus high reactivity of 

bromine to 1 amines. For example, the k values at pH 7 for the 

reaction with glycine are 1.6 10 2 and 4.7 10 5 M 1 s 1 for 

ozone and bromine, respectively (SI-excel). In addition, like to 

the enhanced ammonia oxidation by ozone in the presence of 

bromide as a catalyst (Haag et al., 1984), N-brominated 1 

amines can be oxidized by ozone with a faster rate than the 

parent 1 amines and the corresponding reaction regenerates 

bromide. Finally, bromide has only a small effect on the 

transformation efficiency of micropollutants during advanced 

oxidation processes. This is due to (i) the low reactivity of 

hydroxyl radicals with bromide ( OH þ Br / OH þ Br , 

k ¼ 1.1 10 9 M 1 s 1 , Table 1) which results in the formation of 

HOBr and in the case of H2O2-based AOPs (ii) the fast reaction of 

HOBr with H2O2 leading to bromide (von Gunten, 2003b). 

Formation of potentially carcinogenic bromate during 

ozonation of bromide-containing waters is one of drawbacks 

of ozonation. In recent two studies, only low concentration of 

bromate formation (i.e.

566 

Buxton, G.V., Greenstock, C.L., Helman, W.P., Ross, W.P., 1988. 

Critical review of rate constants for reactions of hydrated 

electrons, hydrogen atoms and hydroxyl radicals ( OH/ O )in 

aqueous solution. J. Phys. Chem. Ref. Data 17, 513–886. 

Deborde, M., von Gunten, U., 2008. Reactions of chlorine with 

inorganic and organic compounds during water 

treatmentdkinetics and mechanisms: a critical review. Water 

Res. 42, 13–51. 

Deborde, M., Rabouan, S., Gallard, H., Legube, B., 2004. Aqueous 

chlorination kinetics of some endocrine disruptors. Environ. 

Sci. Technol 38, 5577–5583. 

Deborde, M., Rabouan, S., Duguet, J.P., Legube, B., 2005. Kinetics of 

aqueous ozone-induced oxidation of some endocrine 

disruptors. Environ. Sci. Technol 39, 6086–6092. 

Dodd, M.C., Huang, C.H., 2004. Transformation of the 

antibacterial agent sulfamethoxazole in reactions with 

chlorine: kinetics, mechanisms, and pathways. Environ. Sci. 

Technol 38, 5607–5615. 

Dodd, M.C., Buffle, M.O., von Gunten, U., 2006. Oxidation of 

antibacterial molecules by aqueous ozone: moiety-specific 

reaction kinetics and application to ozone-based wastewater 

treatment. Environ. Sci. Technol 40, 1969–1977. 

Gates, D., 1998. The Chlorine Dioxide Handbook. American Water 

Works Association, Denver. 

Haag, W.R., Hoigné, J., Bader, H., 1984. Improved ammonia 

oxidation by ozone in the presence of bromide ion during 

water treatment. Water Res. 18, 1125–1128. 

Hoigné, J., 1998. Chemistry of aqueous ozone and transformation 

of pollutants by ozonation and advanced oxidation processes. 

In: Hrubec, J. (Ed.), The Handbook of Environmental 

Chemistry. Springer-Verlag, Berlin Heidelberg, pp. 83–141. 




wastewater treatment plant upgraded with a full-scale postozonation 

followed by sand filtration. Environ. Sci. Technol 43, 

7862–7869. 

Huber, M.M., Canonica, S., Park, G.Y., von Gunten, U., 2003. 


oxidation processes. Environ. Sci. Technol 37, 1016–1024. 

Huber, M.M., Korhonen, S., Ternes, T.A., von Gunten, U., 2005. 

Oxidation of pharmaceuticals during water treatment with 

chlorine dioxide. Water Res. 39, 3607–3617. 

Johnson, D.W., Margerum, D.W., 1991. Non-metal redox kinetics: 

a reexamination of the mechanism of the reaction between 

hypochlorite and nitrite ions. Inorg. Chem. 30, 4845–4851. 

Kumar, K., Margerum, D.W., 1987. Kinetics and mechanisms of 

general-acid-assisted oxidation of bromide by hypochlorite 

and hypochlorous acid. Inorg. Chem. 26, 2706–2711. 

Lee, Y., Yoon, J., von Gunten, U., 2005a. Kinetics of the oxidation 

of phenols and phenolic endocrine disruptors during water 

treatment with ferrate (Fe(VI)). Environ. Sci. Technol 39, 

8978–8984. 

Lee, Y., Yoon, J., von Gunten, U., 2005b. Spectrophotometric 

determination of ferrate (Fe(VI)) in water by ABTS. Water Res. 

39, 1946–1953. 

Lee, Y., Escher, B.I., von Gunten, U., 2008. Efficient removal of 


containing steroid estrogens. Environ. Sci. Technol 42, 6333–6339. 


Lee, Y., von Gunten, U., 2009. Transformation of 17aethinylestradiol 

during water chlorination: effects of bromide 

on kinetics, products, and transformation pathways. Environ. 

Sci. Technol 43, 480–487. 

Lee, Y., von Gunten, U. Kinetics of ferrate VI reactions with 

inorganic compounds in water. in preparation. 

Lee, Y., Zimmermann, S.G., Kieu, A.T., von Gunten, U., 2009. Ferrate 

(Fe(VI)) application for municipal wastewater treatment: 

a novel process. Environ. Sci. Technol 43, 3831–3838. 

Letterman, R.D., 1999. Water Quality and Treatment, fifth ed. 

McGraw-Hill, New York. 

Liu, Q., Schurter, L.M., Muller, C.E., Aloisio, S., Francisco, J.S., 

Margerum, D.W., 2001. Kinetics and mechanisms of aqueous 

ozone reactions with bromide, sulfite, hydrogen sulfite, iodide, 

and nitrite ions. Inorg. Chem. 40, 4436–4442. 

Meunier, L., Canonica, S., von Gunten, U., 2006. Implications of 

sequential use of UV and ozone for drinking water quality. 

Water Res. 40, 1864–1876. 

Neta, P., Huie, E., Ross, A.B., 1988. Rate constants for reactions of 

inorganic radicals in aqueous solution. J. Phys. Chem. Ref. 

Data 17, 1027–1264. 

Nowotny, N., Epp, B., von Sonntag, C., Fahlenkamp, H., 2007. 

Quantification and modelling of the elimination behavior of 

ecologically problematic wastewater micropollutants by 

adsorption on powdered and granulated activated carbon. 


Pinkernell, U., Nowack, B., Gallard, H., von Gunten, U., 2000. Methods 

for the photometric determination of reactive bromine and 

chlorine species with ABTS. Water Res. 34, 4343–4350. 

Pinkston, K.E., Sedlak, D.L., 2004. Transformation of aromatic 

ether- and amine-containing pharmaceuticals during chlorine 

disinfection. Environ. Sci. Technol 38, 4019–4025. 

Rosario-Ortiz, F.L., Mezyk, S.P., Doud, D.F.R., Snyder, S.A., 2008. 

Quantitative correlation of absolute hydroxyl radical rate 

constants with non-isolated effluent organic matter bulk 

properties in water. Environ. Sci. Technol 42, 5924–5930. 

Sharma, V.K., Bloom, J.T., Joshi, V.N., 1998. Oxidation of ammonia 

by ferrate(VI). J. Environ. Sci. Health A33, 635–650. 

Ternes, T.A., Joss, A., 2006. Human Pharmaceuticals, Hormones 

and Fragrances. The Challenge of Micropollutants in Urban 

Water Management. IWA Publishing, London, New York. 

Thompson, G.W., Ockerman, G.W., Schreyer, J.M., 1951. 

Preparation and purification of potassium ferrate. VI. J. Am. 

Chem. Soc. 73, 1379–1381. 

von Gunten, U., 2003a. Ozonation of drinking water: Part I. Oxidation 

kinetics and product formation. Water Res. 37, 1443–1467. 

von Gunten, U., 2003b. Ozonation of drinking water: Part II. 

Disinfection and by-products formation in presence of 

bromide, iodide or chlorine. Water Res. 37, 1469–1487. 

Wert, E.C., Rosario-Ortiz, F.L., Drury, D.D., Snyder, S.A., 2007. 

Formation of oxidation byproducts from ozonation of 

wastewater. Water Res. 41, 1481–1490. 

Westerhoff, P., Mezyk, S.P., Cooper, W.J., Minakata, D., 2007. 

Electron pulse radiolysis determination of hydroxyl radical rate 

constants with Suwannee river fulvic acid and other dissolved 

organic matter isolates. Environ. Sci. Technol 41, 4640–4646. 

Zepp, R.G., Hoigné, J., Bader, H., 1987. Nitrate-induced 

photooxidation of trace organic chemicals in water. Environ. 

Sci. Technol 21, 443–450.

A review of phytoestrogens: Their occurrence 

and fate in the environment 

Ze-hua Liu a,b, *, Yoshinori Kanjo a , Satoshi Mizutani a 

a 

Department of Urban Engineering, Graduate School of Engineering, Osaka City University, 

3-3-138 Sugimoto, Sumiyoshi-ku, Osaka 558-8585, Japan 

b 

Department of Chemistry and Chemical Engineering, Shanghai University of Engineering Science, 

333 Longteng Road, Songjiang, Shanghai 201620, PR China 

article info 


Received 11 December 2008 


12 March 2009 

Accepted 16 March 2009 

Available online 24 March 2009 

Keywords: 

Phytoestrogens 

Endocrine disrupting compounds 

Wastewater 

Urinary excretion rate 

Monitoring 


abstract 

Endocrine disrupting compounds (EDCs) are chemicals with 

the potential to elicit negative effects on the endocrine 

systems of humans and wildlife. They include a broad class of 

chemicals such as natural estrogens/androgens, synthetic 

estrogens/androgens, industrial chemicals as well as phytoestrogens 

(Liu et al., 2009). As EDCs exist in extremely low 

concentrations (mg/L or ng/L) in the environment, one of the 

key points for the research is sensitive and accurate 

measurement of EDCs. Nowadays, two categories of methods 

are available; chemical analyses such as LC–MS/MS, GC–MS as 

well as HPLC and bioassays (in vitro and in vivo). Chemical 




Phytoestrogens are plant compounds with estrogenic activities. Many edible plants, some 

of which are common in the human diet, are rich in phytoestrogens. Almost all phytoestrogens 

eaten daily by people were reported partly recovered in urine or feces, which can 

be regarded as one of the main sources of their occurrence in municipal wastewaters. As 

they may act as one part of the endocrine disrupting compounds (EDCs) in water systems, 

some phytoestrogens have been monitored and detected in wastewater and other various 

environments. It is very difficult to monitor numerous unknown EDCs in complex wastewater 

samples, and it is helpful if some estimation of target EDCs can be done before 

monitoring. With this in mind, this review will: (1) summarize estrogenic activities or 

estrogenic potencies of phytoestrogens by different bioassays; (2) summarize daily urinary 

excretion rates of phytoestrogens by humans, and compare their urinary excretion rates to 

that of estrone, which suggests that most phytoestrogens may occur in municipal wastewaters; 

(3) collect and summarize published data on the occurrence and fate of phytoestrogens 

in various environments. 


analyses can measure precise concentrations of known target 

EDCs in environmental samples, while bioassays provide 

direct information of the estrogenic activity of complex 

mixtures of EDCs that are likely to occur in water irrespective 

of the causative compounds. For large-scale monitoring 

research, bioassays (in vitro) seem most appropriate due to 

their low cost and high throughput capability. However, for 

the complexity of environmental samples (usually wastewater 

samples), values measured by different bioassays on 

the same samples differed greatly, which have lead to questions 

over the validity and effectiveness of bioassays for 

environmental samples. To find widely acceptable standard 

bioassays for environmental samples, the chemically derived 

* Corresponding author: Department of Urban Engineering, Graduate School of Engineering, Osaka City University, 3-3-138 Sugimoto, 

Sumiyoshi-ku, Osaka 558-8585, Japan. Tel./fax: þ81 6 6605 3048. 

E-mail address: liuzehua78@yahoo.co.jp (Z.-h. Liu). 


doi:10.1016/j.watres.2009.03.025

568 

values are often compared to those of bioassays, which 

accelerates monitoring a multitude of endocrine-like chemicals 

in environmental samples (Furuichi et al., 2004; Nakada 

et al., 2004; Tan, 2006; Nelson et al., 2007; Salste et al., 2007). In 

Fernandez et al. (2007), as many as 30 EDCs have been monitored 

in 9 wastewater treatment plants in Canada and the 

results of chemically derived estrogen equivalents (EEQ) were 

also compared with those measured by recombinant yeast 

assay (RYA). The monitoring of numerous EDCs in wastewater 

is difficult, and would be preferable if some guidelines were 

applicable to wastewater in which EDCs may exist in high 

values, while other EDCs may not. 

Phytoestrogens are defined as any plant compounds 

structurally and/or functionally similar to ovarian and 

placental estrogens and their active metabolites (Whitten and 

Patisaul, 2001). Phytoestrogens with rather high concentrations 

widely exist in numerous plants such as soybeans, 

fruits, cabbages and so on (Liggin et al., 2000; US Department 

of Agriculture, USDA, 2003). Some of these are widelyconsumed 

foods, especially in Asia, where high amounts of 

soy are consumed in direct or indirect ways (Adlercreutz et al., 

1991; Nagata et al., 2007). Although most research on phytoestrogens 

has suggested that they yield positive effects on 

human health (Setchell and Cassidy, 1999; Lissin and Cook, 

2000; Kris-Etherton et al., 2002), some research denoted that 

most phytoestrogens also acted as endocrine disruptors, 

which could cause intersex in aquatic organisms (Kiparissis 

et al., 2003; Tzchori et al., 2004). In addition, genotoxicity of 

phytoestrogens was reviewed by Stopper et al. (2005), who 

pointed out some possible adverse effects of phytoestrogens 

on humans. 

Due to the possible negative effects mentioned above, 

much attention has been paid to the occurrence and fate of 

phytoestrogens by environmental researchers in municipal/ 

industrial wastewater or surface waters, in which they were 

detected with high frequency (Kiparissis et al., 2001; Lagana 

et al., 2004; Bacaloni et al., 2005; Erbs et al., 2007). Liu et al. (in 

press) detected 15 EDCs at two municipal wastewater treatment 

plants, in which the concentrations of daidzein in 

influents were from 3900 to 12,000 ng/L and its chemically 

derived estrogenic equivalents (EEQ) were 0.9–5.3% of those 

measured by the ER-binding assay (bisphenol A (BPA) and 

nonylphenol (NP) totally contributed 0.7–4.1% in the same 

samples), indicating a relatively high proportion in wastewater 

which was not negligible. Like natural estrogens or 

androgens, the main source of phytoestrogens in municipal 

wastewaters can also be attributed to daily excretions of urine 

or feces by humans. Therefore, a thorough understanding of 

phytoestrogens from daily excretions by humans is helpful for 

estimating their possible concentrations in municipal wastewater. 

This estimation has been applied to natural estrogens 

(Johnson et al., 2000; Johnson and Williams, 2004), and it is 

also helpful for suggesting possible priority phytoestrogens 

when monitoring for numerous EDCs in wastewater samples 

is carried out. Although there is much data on the excretion 

rates of phytoestrogens from human urine, few conclusions 

have been drawn for the purpose above. In addition, in terms 

of convenience to the comparison of chemical analyses with 

bioassays, relative binding affinities (RBA) or estrogenic/ 

androgenic potencies are also important. 


The objectives of this review are: (1) to summarize reported 

estrogenic/androgenic activities of phytoestrogens by 

different bioassays; (2) to summarize the daily urinary excretion 

rates of phytoestrogens by humans; and (3) to summarize 

reported data on the occurrence and fate of phytoestrogens in 

the various environments. 

2. Estrogenic/androgenic activities of 

phytoestrogens 

Phytoestrogens are a large family with two major classes, 

flavonoids and lignans. The number of flavonoids is greater 

than the number of lignans. Flavonoids also possess higher 

relative estrogenic activities than those in lignans (Whitten 

and Patisaul, 2001). One interest for researchers of environmental 

engineering relating to EDCs is the estrogenic/androgenic 

activities of phytoestrogens. There is much available 

data on their estrogenic/androgenic activities by different 

bioassays. To give an overview of their estrogenic/androgenic 

activities, some published data are outlined in Table 1, in 

which the estrogen standard substance is 17b-estradiol (E2) 

and the androgen standard substance is testosterone (Te) or 

methyltrienolone (R1881). From Table 1, it is easily seen that 

almost all phytoestrogens are far weaker estrogenic 

substances, except in the transient gene expression assay 

(TGEA), in which the estrogenic potencies of some phytoestrogens 

are at the same level, or even stronger than that of E2. 

Compared to the estrogenic activities of industrial chemicals 

BPA or NP (Blair et al., 2000; Gutendorf and Westendorf, 2001; 

Liu et al., 2009), most phytoestrogens are in the same order of 

magnitude, which denote that the estrogenic activities 

derived from phytoestrogens can not be neglected when they 

remain in high concentrations in water samples. In addition, 

some phytoestrogens also possess androgenic activities 

(Table 1). With the RBA or estrogenic potency in Table 1, we 

can calculate the contribution ratio of phytoestrogens (Ratio) 

to values measured by bioassays using Eq. (1) and the contribution 

ratio of phytoestrogens (Ratioc) to the chemically 

derived EEQ using Eq. (2): 

Ratio ¼ EEQ P 

c RBAiðEPiÞ Ci 

¼ 

(1) 

EEQb EEQb where EEQc and EEQb are the chemically calculated EEQ and 

the measured EEQ by bioassay, respectively; Ci represents the 

concentration of an individual EDC in wastewater sample or 

the urinary excretion rate of an individual phytoestrogen from 

human and RBAi(EP)i is its corresponding estrogenic activity. 

Ratioc ¼ EEQi ¼ 

EEQc RBAiðEPiÞ Ci 

P 

RBAiðEPiÞ Ci 

where EEQ i is the chemically derived EEQ of an individual 

phytoestrogen in urine or an individual EDC in wastewater. 

To monitor phytoestrogens in complex wastewater or 

other water samples, it is very important to estimate their 

concentrations possibly existing in environmental samples. 

For phytoestrogens in municipal wastewater, one of the main 

sources can be regarded as human urine or feces. Therefore, 

their daily urinary excretion rates were concluded as shown 

below. 

(2)

Table 1 – Estrogenic/androgenic activities of phytoestrogens by different bioassays * . 

Class Subclass Phytoestrogen CAS RBA (ER) RBA (AR) Estrogenic potency (EP) 

hER(a) a;b 

hER(b) b 

hAR(a) a 

rAR(a) c 

RGA a (ER(a)) YES d 

(ER(a)) 

TGEA b 

ER(a) ER(b) 

Flavonoids Isoflavone Daidzein 486-66-8 1.8e-3; 1e-3 5e-3 N.B – 6.67e-5 3e-6 0.97 0.8 

Dihydrodaidzein 17238-05-0 –; – – – – – – – – 

Formononetin 485-72-3 2.65e-4;

570 

Table 2 – Daily excretion rates of phytoestrogens from human urine. 

Phytoestrogens Gender condition 

(number) 


Diet Daily excretion rates 

(average) 

mg/d mgE2/d c 

Mean EEQ Reference 

Genistein Men (267) Normal a 

n.d.–7579.4 (132.4) 0.159 Low et al., 2005 

Women (250) Normal – b (2172.7) 2.607 Dai et al., 2002 d 

Women (19) Soy protein diets 1810.6–2837.5 (2357.8) 2.829 Lampe et al., 2001 

Women (268) Normal 31.5–1251.3 (219.8) 0.264 Tonkelaar et al., 2001 d 

Women (18) Normal 183.5–280.8 (227) 0.272 Xu et al., 2000 

7.1 mg total isoflavones/d 362.4–492.9 (422.7) 0.507 

65 mg total isoflavones/d 1521.1–2096.2 (1785.7) 2.143 

132 mg isoflavones/d 3045.8–4144.6 (3553.1) 4.264 

Men (49) Normal – (78.4) 0.094 Lampe et al., 1999 

Women (49) Normal – (45.9) 0.055 

Men (2) Normal 368.6–587.2 (477.9) 0.573 Adlercreutz et al., 1995 e 

Women (3) Normal (2 vegetarians) 482.6–3653.3 (1749.9) 2.100 Adlercreutz et al., 1995 

Women (18) Normal 10.8–281.0 (48.6) 0.058 Lampe et al., 1994 

Daidzein Men (267) Normal n.d.–4309.8 (227.0) 0.409 Low et al., 2005 

Women (250) Normal – (4390.7) 7.903 Dai et al., 2002 


Women (18) Normal 200.8–323.4 (254.7) 0.458 Xu et al., 2000 





Women (49) Normal – (94.1) 0.169 

Men (2) Normal 1074.2–1113.6 (1093.9) 1.969 Adlercreutz et al., 1995 



Women (41) Normal 56.1–322.9 (124.2) 0.224 Adlercreutz et al., 1986 

Dihydrodaidzein Women (250) Normal – (1340.2) – Dai et al., 2002 

Women (18) Normal 151.2–222.7 (183.5) – Xu et al., 2000 

7.1 mg total isoflavones/d 221.7–339.5 (274.4) – 

65 mg total isoflavones/d 853.1–1334.1 (1066.8) – 

132 mg isoflavones/d 1528.8–2341.4 (1891.9) – 

ODMA Men (267) Normal n.d.–381.9 (48.1) – Low et al., 2005 

Women (250) Normal – (1299.1) – Dai et al., 2002 

Women (19) Soy protein diets 206.6–1859.5 (707.0) – Lampe et al., 2001 

women (18) Normal 142.6–214.4 (174.8) – Xu et al., 2000 

7.1 mg total isoflavones/d 206.4–297.8 (247.9) – 

65 mg total isoflavones/d 1045.2–1344.5 (1265.5) – 

132 mg isoflavones/d 2236.9–3227.8 (2687.3) – 

Men (2) Normal 22.8–335.2 (179.0) – Adlercreutz et al., 1995 

Women (3) Normal (2 vegetarians) 171.0–627.1 (379.6) – Adlercreutz et al., 1995 

Women (41) Normal 6.5–27.4 (12.7) – Adlercreutz et al., 1986 

Equol Men (267) Normal n.d.–1280.5 (18.2) 0.157 Low et al., 2005 

Women (19) Soy protein diet n.d.–581.4 (271.0) 2.331 Lampe et al., 2001 






Women (49) Normal – (17.0) 0.146 




Women (41) Normal 15.5–24.7 (20.4) 0.175 Adlercreutz et al., 1986


Phytoestrogens Gender condition 

(number) 

3. Occurrence and fate of phytoestrogens in 

the environment 

3.1. Phytoestrogens in human urine 

Phytoestrogens in human urine are from the diet or dietary 

metabolites, and their concentrations in urine have been 

found dose dependent on their consumption (Karr et al., 

1997; Zhang et al., 1999; Xu et al., 2000; Peeters et al., 2003). 

For disease related items, there has been much research on 

the daily excretion rates of urinary phytoestrogens, in which 

the excretion rates varied greatly with race and dietary 

customs (Horn-Ross et al., 1997; Seow et al., 1998). To give an 

overview, published data on urinary excretion rates are 

outlined in Table 2. In order to compare the estrogenic 

potencies of phytoestrogens from human urine, the authors 


Diet Daily excretion rates 

(average) 

mg/d mgE2/d c 

Mean EEQ Reference 

Glycitein Men (267) Normal n.d.–556.2 (51.3) 0.014 Low et al., 2005 






Coumestrol Women (18) Normal 8.3–9.5 (8.9) 0.023 Xu et al., 2000 




Enterolactone Men (267) Normal n.d.–13218.5 (1020.4) 0.033 Low et al., 2005 



Women (268) Normal 19.0–6157.9 (1276.2) 0.041 Tonkelaar et al., 2001 


7.1 mg total isoflavones/d 347–489.6 (412.3) 0.013 




Women (49) Normal – (584.7) 0.019 



Women (41) Normal 611.6–1244.0 (914.3) 0.029 Adlercreutz et al., 1986 

Naringenin Women (250) Normal – (1217.0) 0.065 Dai et al., 2002 

Men and women (77) Normal – (653.8) 0.035 Nielsen et al., 2002 

Quercetin Men and women (77) Normal – (23.4) – Nielsen et al., 2002 

Phloretin Men and women (77) Normal – (84.8) 0.013 Nielsen et al., 2002 

Kaempferol Men and women (77) Normal – (51.3) 0.015 Nielsen et al., 2002 

a Normal diet. 

b Not available. 

c Calculated as EEQi in Eq. (2), values of RBA based on ER-binding assay (NITE, 2008). 

d Transferred from nmol/mg creatinine or mmol/mol creatinine to mg/d, and creatinine excretion rate of 1 g/d for women was chosen (Jackson, 

1966). 

e Calculated from their free phytoestrogens and their conjugates. 

converted the urinary excretion rate of the individual phytoestrogens 

into a corresponding value, calculated as EEQi in 

the numerators of Eq. (2); we label it ‘‘excretion rate of EEQ’’. 

As is shown in Table 2, out of 12 phytoestrogens, genistein 

and daidzein are the two most important phytoestrogens 

evaluated by both urinary excretion rate and excretion rate 

of EEQ; followed by equol, for which its urinary excretion 

rates by normal diets are low but with a relatively high 

corresponding excretion rate of EEQ. Enterolactone and 

naringenin are the phytoestrogens with high urinary excretion 

rates, but very low estrogenic potencies (low RBA, Table 

1). Therefore, the corresponding excretion rates of EEQ by 

enterolactone and naringenin seem relatively low. Data of 

urinary excretion rates of some phytoestrogens, especially 

for quercetin, phloretin as well as kaempferol, is very 

limited. For better comparison, it would be helpful if more 

data were available.

572 

In addition, biochanin A and formononetin have been 

found existing in high concentrations in red clovers of 

Canada, and have also been detected in many fruits, beans 

and vegetables (Mazur et al., 1996; Sivesind and Seguin, 2005; 

Kuhnle et al., 2007, 2008). Few data on urinary excretion rates 

are available. As a reference, one published paper showed the 

average fecal excretion rate was about 1.5 mg/d for biochannin 

A and 11.9 mg/d for formononetin (Jenner et al., 

2005). A phytosterol b-sitosterol with estrogenic activity 

(Table 1) was found rich in a wide range of fruits, vegetables, 

nuts, and many kinds of seed oils (Grosso et al., 2000; Beveridge 

et al., 2002, 2005; Iwatsuki et al., 2003; Holser et al., 

2004; Phillips et al., 2005; Jimenez-Escrig et al., 2006; Zhang 

et al., 2006). Although there is no available data on the 

urinary excretion rate of b-sitosterol, very high human fecal 

excretion rates of 52–322 mg/d and a high human fecal 

concentration of 121 mg/g in dry feces (average) were reported 

(Eneroth et al., 1964; Leeming et al., 1996), which suggested 

that b-sitosterol might also exist in wastewater with a very 

high concentration. 

From the view of statistics, a rough estimation on the 

arithmetic average of urinary excretion rates was done on 

the data of normal diets in Table 2, which was also applied 

on human urinary excretion rates of natural estrogens and 

androgens (Liu et al., submitted for publication). If the loss of 

phytoestrogens during the course of flow to wastewater 

treatment plants is low, their concentrations in wastewater 

can be compared to those of natural estrogens. Although it is 

well known that 17b-estradiol (E2) is the strongest natural 

estrogen, estrone (E1) is the one which was detected 

frequently with high concentrations in effluent of WWTPs 

(Baronti et al., 2000; Joss et al., 2004; Johnson et al., 2007; Liu 

et al., 2009). Therefore, the average urinary excretion rates of 

phytoestrogens were compared to that of E1 as shown in 

Fig. 1, and the urinary excretion rate of E1 was modified from 

Liu et al. (submitted for publication), for which the urinary 

excretion rates of two pregnant women out of 177 persons 

(11.3&) were also included. From Fig. 1 it is known that the 

urinary excretion rates of phytoestrogens are 1–86 times that 

of E1, except for coumestrol (0.45 times); Among phytoestrogens 

in Table 2, genistein and daidzein, frequently detected 

in wastewaters, are 35 and 86 times that of E1, 

urinary excretion rates(µg/d) 

10000 

1000 

100 

10 

1 

E1 

Genistein 

Daidzein 

Dihydrodaidzein 

ODMA 

Glycitein 

Equol 

Coumestrol 

Enterlactone 

Naringerin 

Quercetin 

Phloretin 

Fig. 1 – Comparison of urinary excretion rates between 

phytoestrogens and E1. 


Kaemperol 

respectively. According to published results, the average 

concentrations of genistein and daidzein in influent of one 

WWTP in Italy were about 3 and 9 times that of the detected 

E1 (Lagana et al., 2004), while the concentrations of daidzein 

in 12 influent samples of two municipal WWTPs in Japan 

showed about 8–46 times (22 times on average) that of the 

detected E1 (Liu et al., in press). Compared to detected 

concentrations for genistein and daidzein, their estimated 

concentrations in Fig. 1 are about 10 times and 2–10 times 

higher. The large differences may result from the different 

loss rates before they reach WWTPs, different excretion 

ratios of urine and feces for phytoestrogens and E1, as well as 

regional dietary customs. Although coumestrol is the leastpresent 

urinary excretion phytoestrogen in Fig 1, it was also 

reported existing in municipal WWTPs (Bacaloni et al., 2005). 

Based on this fact, the other 11 phytoestrogens in Fig. 1 are 

very likely to be present in municipal wastewater. Based on 

EEQ, phytoestrogens such as daidzein, genistein, equol and 

glycitein should not be neglected, for which the excretion 

rates of EEQ are about 35%, 10%, 2% and 1% that of human 

urinary E1. As enterlactone and naringerin are much weaker 

estrogens than daidzein (Table 1), although their urinary 

excretion rates are on the same level as daidzein, the corresponding 

excretion rates of EEQ are just about 0.5% and 

0.7% that of human urinary E1. However, the urinary excretion 

rates of enterlactone in some regions may be much 

higher than that in Fig. 1 (Lampe et al., 1994; Wang, 2002; 

Milder et al., 2005; Valentin-Blasini et al., 2005; Penalvo et al., 

2008), which may yield a higher contribution ratio in wastewater. 

In addition, the urinary excretion rates of dihydrodaidzein 

and ODMA are almost the same as that of 

daidzein, but the data on their estrogenic activities is not 

available; therefore, their corresponding excretion rates of 

EEQ can not be applied here. 

3.2. Phytoestrogens in wastewaters and the other 

water bodies 

As predicted above that most phytoestrogens may exist in 

wastewater, some phytoestrogens have been proven to not 

only exist in wastewater but also in surface waters (Kawanishi 

et al., 2004; Lagana et al., 2004; Bacaloni et al., 2005; Erbs et al., 

2007). Mycoestrogens (Table 1) are zearalenone and its derivatives. 

They are macrocyclic toxins produced by several 

Fusarium species, colonizing maize, sorghum, wheat, barley 

oats and other cereal grains, which were also reported 

occurring in wastewaters and surface waters (Lagana et al., 

2004). Published data on the occurrence and fate of phytoestrogens 

in the course of WWTPs are included in Table 3. 

As stated in Section 1, to give a whole profile of EDCs in the 

wastewater of WWTPs, a combination of chemical analyses 

and bioassays has been used, and estrogenic contribution 

ratios of detected EDCs to those of chemical derived EEQ have 

also been compared. As shown in Table 3, it is known that 

some phytoestrogens have been detected in both influent and 

effluent, and their removal efficiencies varied greatly as 

shown for natural estrogens and androgens (Liu et al., 2009). 

Data in Table 3 also suggested that the estrogenic contribution 

ratios of phytoestrogens could not be ignored in wastewaters 

when present in relatively high concentrations. For example,

Table 3 – Occurrence and fate of phytoestrogens in wastewater. 

Phytoestrogens Country WWTP 

type (n) 

the authors calculated that six phytoestrogens in influents of 

a municipal WWTP contributed over 2.4% of total chemically 

calculated EEQ including E1, E2, E3 and EE2 (RBA values of 

EDCs are chosen from NITE, 2008), and increased to over 4.5% 

for the corresponding effluents (Lagana et al., 2004); In 12 

influent samples of two municipal WWTPs, the contribution 

ratios of daidzein accounted for 3.0–14.8% of total chemically 

calculated EEQ (Liu et al., in press). In the influents of a swine 

wastewater treatment plant, the contribution ratios of phytoestrogen 

equol accounted for as high as about 64% of total 

chemically derived EEQ (Furuichi et al., 2006). In addition, for 

the extremely high concentrations of b-sitosterol in wastewaters 

(Table 3), its estrogenic contribution ratio in wastewater 

seems relatively high and it also can not be neglected. 

Similar to natural estrogens and androgens, about 97– 

100% of phytoestrogens excreted from human urine are 

phytoestrogen conjugates, in which 62–97% of phytoestrogens 

are glucuronide conjugates (Adlercreutz et al., 1995). 

Although there is no information on the occurrence and fate 

of phytoestrogen conjugates in WWTPs, the glycoside 

conjugates of daidzein and genistein (daidzin and genistin), 

which have structures similar to those of their glucuronide 

conjugates, have been proven hydrolyzable to daidzein and 


Methods LOD Influent Effluent Average 

removal (%) 

Reference 

Daidzein Germany AS (2) – 10 – n.d.–10 – Pawlowski et al., 2003 

Italy AS (1) LC–MS/MS 2.5–3.5 75–120 7–16 88 Lagana et al., 2004 

Italy AS (2) LC–MS/MS 3.5 97–1685 5–81 88 or 98 Bacaloni et al., 2005 

Japan AS (2) LC–MS/MS 5–60 b 

3900–12000 5.7–23 99.8 Liu et al., in press 

Daidzin Italy AS (2) LC–MS/MS 9–15 18–74 n.d.–22 69 or 100 Bacaloni et al., 2005 

Genistein Canada – (1) LC–MS/MS – 13100 10100 23 Kiparissis et al., 2001 

Italy As (1) LC–MS/MS 1.3–4.3 195–384 15–83 81 Lagana et al., 2004 

Italy AS (2) LC–MS/MS 5–6 160–954 7–22 97 Bacaloni et al., 2005 

Genistin Italy AS (2) LC–MS/MS 9–13 11–62 n.d.–19 66 or 100 Bacaloni et al., 2005 

Glycitein Italy AS (2) LC–MS/MS 6–8 6–25 n.d.–16 62 or 88 Bacaloni et al., 2005 

Formononetin Italy AS (2) LC–MS/MS 5–8 n.d.–10 n.d. – Bacaloni et al., 2005 

Coumestrol Italy AS (2) LC–MS/MS 2–5 9–23 n.d.–19 66 or 100 Bacaloni et al., 2005 

Biochanin-A Italy As (1) LC–MS/MS 0.9–5.4 8–18 3–5 73 Lagana et al., 2004 

Italy AS (2) LC–MS/MS 5–9 14–76 n.d.–21 67 or 90 Bacaloni et al., 2005 

Equol Japan – (1) LC–MS/MS – 9.4e5–1.1e6 170–410 99.99 Furuichi et al., 2006 

Zearlenone Italy AS (1) LC–MS/MS 0.6–1.7 3–18 3–10 53 Lagana et al., 2004 

a-Zearalanol Italy AS (1) LC–MS/MS 5.4–8.3 n.d.–10 n.d.–7 0 Lagana et al., 2004 

b-Zearalanol Italy AS (1) LC–MS/MS 3.7–8.4 n.d.–8 n.d.–5 33 Lagana et al., 2004 

a-Zearalenol Germany AS (2) – 10 – 20 or 30 – Pawlowski et al., 2003 

b-Sitosterol Switzerland – (1) GC–MS 125 – 400–1300 – Baig et al., 2008 

Canada AS (3); 

TF/SC (1); 

L (1) 

GC–HRMS 12–39 7303–35033 350–42110 – Fernandez et al., 2007 

USA – (2) GC–MS – 32000 a 

– – Cain et al., 2008 

Germany – (2) GC–MS – 14000or 16000 – – Radke, 2005 

Germany AS (2) – 50 – 60 or 90 – Pawlowski et al., 2003 

Germany – (7) GC–MS/MS – 15900 1900 88 Heberer, 2002 

France AS (2) GC–MS – 4194–49920 1100–50165 – Quemeneur and Marty, 

1994 

Enterolactone Australian – (1) LC–MS 1 600 – – Kang et al., 2006 

Finland AS (1) LC–MS/MS – 1223 659 46 Smeds et al., 2007 

AS, activated sludge; TF/SC, trickling filter/solid contact; L, lagoon; –, not available; n.d., not detected; LOD, limit of detection. 

a Maximum concentration. 

b Limit of quantification. 

genistein by the glucosidase of human intestinal bacteria 

(Hur et al., 2000). The detected daidzin and genistin in some 

effluents of WWTPs suggested their incomplete hydrolysis 

during the whole activated sludge process (Table 3). In addition, 

based on the fact that natural estrogen conjugates 

occurred both the influent and effluent of WWTPs (D’Ascenzo 

et al., 2003; Isobe et al., 2003; Komori et al., 2004; Reddy et al., 

2005; Nakada et al., 2006), phytoestrogen conjugates may also 

be present in influent and effluent of WWTPs with high 

frequency. 

Published data on the occurrence of phytoestrogens in 

other various environments are collected in Table 4. From 

Table 4, it is known that most phytoestrogens including 

mycoestrogens were found in rivers, and in some occasions 

the concentrations of phytoestrogens in drainage water were 

much higher than those in river water, which indicated that 

the drainage water was also a source of phytoestrogens in 

river waters. In Table 4 b-sitosterol showed very high 

concentrations in rivers and it was also detected in the 

atmosphere. However, it has not been detected in drinking 

water (Quemeneur and Marty, 1994; Heberer, 2002; Martins 

et al., 2007; Stackelberg et al., 2007). In addition, enterlactone 

was seldom monitored in wastewater as a micro-pollutant,

574 

Table 4 – Phytoestrogens in the other various environments. 

Phytoestrogens River water 

(ng/L) 

whereas the recent investigation denoted that it not only 

existed in wastewater, but also in river water, lake water, sea 

water and drinking water (Table 4). 


Drainage water 

(ng/L) 

Urinary excretion rates of phytoestrogens from humans and 

their occurrence and fate in the environment were summarized 

here. It was concluded that most phytoestrogens may be 

present in municipal wastewaters as can be estimated from 

their urinary excretion rates. Although phytoestrogens are 

very weak estrogens when compared to natural estrogens, 

their possible high concentrations in wastewater may lead 

them to contribute a non-negligible part of EEQ. Compared to 

natural estrogens or industrial chemicals such as BPA, NP and 

Octylphenol (OP), the research of phytoestrogens in wastewaters 

is just beginning. However, with the progress of 

analytic methods, knowledge of phytoestrogens in wastewater, 

such as removal mechanisms of phytoestrogens at 


River sediments 

(ng/g) 

WWTPs, will grow. On some occasions, chemical advanced 

oxidation should also be adopted to degrade phytoestrogens 

when purer water is required. 

references 

Sea water 

(ng/L) 

Drinking 

water 

Air 

(pg/m 3 ) 

Daidzein 2–3 [1]; 5–30 [3] – a 

2–4 [2]; 

42900 [4]; 

Bacaloni, A., Cavaliere, C., Faberi, A., Foglia, P., Samperi, R., 

Lagana, A., 2005. Determination of isoflavones and coumestrol 

in river water and domestic wastewater sewage treatment 

plants. Analytica. Chimica. Acta 531, 229–237. 

Baig, S., Hansmann, G., Paolini, B., 2008. Ozone oxidation of 

estrogenic active substances in wastewater and drinking 

water. Wat. Sci. Technol 58, 451–458. 

Baronti, C., Curini, R., D’Ascenzo, G., Corcia, A.D., Gentili, A., 

Samperi, R., 2000. Monitoring natural and synthetic estrogens 

at activated sludge sewage treatment plants and in a receiving 

river water. Environ. Sci. Technol 34, 5059–5066. 

Beck, M., Radke, M., 2006. Determination of sterols, estrogens and 

inorganic ions in waste water and size-segregated aerosol 

particles emitted from waste water treatment. Chemosphere 

64, 1134–1140. 

Beck, I.C., Bruhn, R., Gandrass, J., Ruck, W., 2005. Liquid 

chromatography-tandem mass spectrometry analysis of 

estrogenic compounds in coastal surface water of the Baltic 

Sea. J. Chromatogr. A 1090, 98–106. 

Beveridge, T.H.J., Li, T.S.C., Drover, J.C.G., 2002. Phytosterol content 

in American ginseng seed oil. J. Agric. Food Chem. 50, 744–750. 

Beveridge, T.H.J., Girard, B., Kopp, T., Drover, J.C.G., 2005. Yield 

and composition of grape seed oils extracted by supercritical 

carbon dioxide and petroleum ether: varietal effects. J. Agric. 

Food Chem. 53, 1799–1804. 

Blair, R.M., Fang, H., Branham, W.S., Hass, B.S., Dial, S.L., 

Moland, C.L., Tong, W., 2000. The estrogen receptor relative 

binding affinities of 188 natural and xenochemicals: structural 

diversity of ligands. Toxicol. Sci. 54, 138–153. 

Bovee, T.F.H., Helsdingen, R.J.R., Rietjens, I.M.C.M., Keijer, J., 

Hoogenboom, R.L.A.P., 2004. Rapid yeast estrogen bioassays 

stably expressing human estrogen receptor a and b, and green 

fluorescent protein: a comparison of different compounds with 

both receptor types. J. Steroid Biochem. Mol. Biol. 91, 99–109. 

Cain, T.G., Kolpin, D.W., Vargo, J.D., Wichman, M.D., 2008. 

Occurrence of antibiotics, pharmaceuticals and sterols at 

select surface and wastewater sites in Iowa. http://info.ngwa. 

org/gwol/pdf/042379985.pdf. 2008.11.24. 

Chou, C.C., Liu, Y.P., 2004. Determination of fecal sterols in the 

sediments of different wastewater outputs by GC-MS. Intern. J. 

Environ. Anal. Chem. 84, 379–388. 

D’Ascenzo, G., Corcia, A.D., Mancini, A.G.R., Mastropasqua, R., 

Nazzari, M., Samperi, R., 2003. Fate of natural estrogen 

conjugates in municipal sewage transport and treatment 

facilities. Sci. Total Environ 302, 199–209. 

Dai, Q., Franke, A.A., Jin, F., Shu, X.O., Hebert, J.R., Custer, L.J., 

Cheng, J.R., Gao, Y.T., Zheng, W., 2002. Urinary excretion of 

phytoestrogens and risk of breast cancer among Chinese 

women in Shanghai. Cancer Epidem. Biomarkers 11, 815–821. 

Devane, M., Saunders, D., Gilpin, B., 2006. Faecal sterols and 

fluorescene whiteners as indicators of the source of faecal 

contamination. Chem. New Zealand 70, 74–77. 

Eneroth, P., Hellstrom, K., Ryhage, R., 1964. Identification and 

quantification of neutral fecal steroids by gas-liquid 

chromatography and mass spectrometry: studies of human 

excretion during two dietary regimens. J. Lipid Res. 5, 245–262. 

Erbs, M., Hoerger, C.C., Hartmann, N., Bucheli, T.D., 2007. 

Quantification of six phytoestrogens at the nanogram per liter 

level in aqueous environmental samples using 13 C 3-labeled 

internal standards. J. Agric. Food Chem. 55, 8339–8345. 

Fang, H., Tong, W., Branham, W.S., Moland, C.L., Dial, S.L.D., 

Hong, H., Xie, Q., Perkins, R., Owens, W., Sheehan, D.M., 2003. 

Study of 202 natural, synthetic, and environmental chemicals 

for binding to the androgen receptor. Chem. Res. Toxicol 16, 

1338–1358. 

Fattore, E., Benfenati, E., Marelli, R., Cools, E., Fanelli, R., 1996. 

Sterols in sediment samples from Venice lagoon. Italy. 



Fernandez, M.P., Ikonomou, M.G., Buchanan, I., 2007. An 

assessment of estrogenic organic contaminants in Canadian 

wastewaters. Sci. Total Environ 373, 250–269. 

Furuichi, T., Kannan, K., Giesy, J.P., Masunaga, S., 2004. 

Contribution of known endocrine disrupting substances to the 

estrogenic activity in Tama River water samples from Japan 

using instrumental analysis and in vitro reporter gene assay. 

Wat. Res. 38, 4491–4501. 

Furuichi, T., Kannan, K., Suzuki, K., Tanaka, S., Giesy, J.P., 

Masunaga, S., 2006. Occurrence of estrogenic compounds in 

and removal by a swine farm waste treatment plant. Environ. 

Sci. Technol 40, 7896–7902. 

Grosso, N.R., Nepote, V., Guzman, C.A., 2000. Chemical 

composition of some wild peanut species (Arachis L.) seeds. J. 

Agric. Food Chem. 48, 806–809. 

Gutendorf, B., Westendorf, J., 2001. Comparison of an array of in 

vitro assays for the assessment of the estrogenic potential of 

natural and synthetic estrogens, phytoestrogens and 

xenoestrogens. Toxicology 166, 79–89. 

Hartmann, N., Erbs, M., Wettstein, F.E., Schwarzenbach, R.P., 

Bucheli, T.D., 2007. Quantification of estrogenic mycotoxins at 

the ng/L level in aqueous environmental samples using 

deuterated internal standards. J. Chromatogr. A 1138, 132–140. 

Heberer, T., 2002. Tracking persistent pharmaceutical residues 

from municipal sewage to drinking water. J. Hydrol 266, 

175–189. 

Holser, R.A., Bost, G., Boven, M.V., 2004. Phytosterol composition 

of hybrid Hibiscus seed oils. J. Agric. Food Chem. 52, 2546–2548. 

Horn-Ross, P.L., Barnes, S., Kirk, M., Coward, L., Parsonnet, J., 

Hiatt, R.A., 1997. Urinary phytoestrogen levels in young 

women from a multiethnic population. Cancer Epidem. 

Biomar 6, 339–345. 

Hur, H.G., Lay Jr., J.O., Beger, R.D., Freeman, J.P., Rafii, F., 2000. 

Isolation of human intestinal bacteria metabolizing the 

natural isoflavone glycosides daidzin and genistin. Arch. 

Microbiol. 174, 422–428. 

Isobe, T., Shiraishi, H., Yasuda, M., Shinoda, A., Suzuki, H., 

Morita, M., 2003. Determination of estrogens and their 

conjugates in water using solid-phase extraction followed by 

liquid chromatography-tandem mass spectrometry. J. 

Chromatogr. A 984, 195–202. 

Iwatsuki, K., Akihisa, T., Tokuda, H., Ukiya, M., Higashihara, H., 

Mukainaka, T., Izuka, M., Hayashi, Y., Kimura, Y., Nishino, H., 

2003. Sterol ferulates, sterols, and 5-alk(en)yl resorcinols from 

wheat, rye and corn bran oils and their inhibitory effects on 

Epstein-Barr virus activation. J. Agric. Food Chem. 51, 6683– 

6688. 

Jackson, S., 1966. Creatinine in urine as an index of urinary 

excretion rate. Health Phys. 12, 843–850. 

Jenner, A.M., Rafter, J., Halliwell, B., 2005. Human fecal water 

content of phenolics: the extent of colonic exposure to 

aromatic compounds. Free Radic. Biol. Med 38, 763–772. 

Jimenez-Escrig, A., Santos-Hidalgo, A.B., Saura-Calixto, F., 2006. 

Common sources and estimated intake of plant sterols in the 

Spanish diet. J. Agric. Food Chem 54, 3462–3471. 

Johnson, A.C., Williams, R.J., 2004. A model to estimate influent 

and effluent concentrations of estradiol. estrone, and 

ethinylestradiol at sewage treatment works. Environ. Sci. 

Technol 38, 3649–3658. 

Johnson, A.C., Belfroid, A., Di Corcia, A., 2000. Estimating steroid 

oestrogen inputs into activated sludge treatment works and 

observations on their removal from the effluent. Sci. Total 

Environ 256, 163–173. 

Johnson, A.C., Williams, R.J., Simpson, P., Kanda, R., 2007. What 

difference might sewage treatment performance make to 

endocrine disruption in rivers. Environ. Pollut 147, 194–202. 

Joss, A., Andersen, H., Ternes, T., 2004. Removal of estrogens in 

municipal wastewater treatment under aerobic and anaerobic

576 

conditions: consequences for plant optimization. Environ. Sci. 

Technol 38, 3047–3055. 

Kang, J.G., Price, W.E., Hick, L.A., 2006. Simultaneous 

determination of isoflavones and lignans at trace levels in 

natural waters and wastewater samples using liquid 

chromatography/electrospray ionization ion trap mass 

spectrometry. Rapid Commun. Mass Spectrom 20, 2411–2418. 

Karr, S.C., Lampe, J.W., Hutchins, A.M., Slavin, J.L., 1997. Urinary 

isoflavonoid excretion in humans is dose dependent at low to 

moderate levels of soy-protein comsumption. Am. J. Clin. Nutr 

66, 46–51. 

Kawanishi, M., Takamura-Enya, T., Ermawati, R., Shimohara, C., 

Sakamoto, M., Matsukawa, K., Matsuda, T., Murahashi, T., 

Matsushi, T., Wakabayashi, K., Watanabe, T., Tashiro, Y., 

Yagi, T., 2004. Detection of genistein as an estrogenic 

contaminant of river water in Osaka. Environ. Sci. Technol 38, 

6424–6429. 

Kiparissis, Y., Hughes, R., Metcalfe, C., Ternes, T., 2001. 

Identification of the isoflavonoid genistein in bleached kraft 

mill effluent. Environ. Sci. Technol 35, 2423–2427. 

Kiparissis, Y., Balch, G.C., Metcalfe, T.L., Metcalfe, C.D., 2003. 

Effects of the isoflavones genistein and equol on the gonadal 

development of Japanese medaka (Oryzias latipes). Environ. 

Health Persp 111, 1158–1163. 

Komori, K., Tanaka, H., Okayasu, Y., Yasojima, M., Sato, C., 2004. 

Analysis and occurrence of estrogen in wastewater in Japan. 

Wat. Sci. Technol 50, 93–100. 

Kostelac, D., Rechkemmer, G., Briviba, K., 2003. Phytoestrogens 

modulate binding response of estrogen receptor a and b to the 

estrogen response element. J. Agric. Food Chem. 51, 7632– 

7635. 

Kris-Etherton, P.M., Hecker, K.D., Bonanome, A., Coval, S.M., 

Binkoski, A.E., Hilpert, K.F., Griel, A.E., Etherton, T.D., 2002. 

Bioactive compounds in foods: their role in the prevention of 

cardiovascular disease and cancer. Am.. J. Med 113 (Suppl), 71–88. 

Kuhnle, G.G.C., Dell’Aquila, C., Kussmaul, M., Bingham, S.A., 2007. 

Extraction and quantification of phytoestrogens in foods using 

automated solid-phase extraction and LC/MS/MS. Anal. Chem. 

79, 9234–9239. 

Kuhnle, G.G.C., Dell’aquila, C., Aspinall, S.M., Runswick, S.A., 

Mulligan, A.A., Bingham, S.A., 2008. Phytoestrogen content of 

beverages, nuts, seeds, and oils. J. Agric. Food Chem. 56, 7311– 

7315. 

Kuiper, G., Lemmen, J.G., Carlsson, B., Corton, J.C., Safe, S.H., van 

der Saag, P.T., van der Burg, B., Gustafsson, J.A., 1998. 

Interaction of estrogenic chemicals and phytoestrogens with 

estrogen receptor b. Endocrinology 139, 4252–4263. 

Lagana, A., Bacaloni, A., De Leva, I., Faberi, A., Fago, G., Marino, A., 

2004. Analytical methodologies for determining the 

occurrence of endocrine disrupting chemicals in sewage 

treatment plants and natural waters. Analytica. Chimica. Acta 

501, 79–88. 

Lampe, J.W., Wartini, M.C., Kurzer, M.S., Adlercreutz, H., Slavin, J. 

L., 1994. Urinary lignan and isoflavonoid excretion in 

premenopausal women consuming flaxseed powder. J. Clin. 

Nutr 60, 122–128. 

Lampe, J.W., Gustafson, D.R., Hutchins, A.M., Martini, M.L., Li, S., 

Wahala, K., Grandits, G.A., Potter, J.D., Slavin, J.L., 1999. 

Urinary isoflavonoid and lignan excretion on a western diet: 

relation to soy, vegetable, and fruit intake. Cancer Epidem. 

Biomarkers 8, 699–707. 

Lampe, J.W., Skor, H.E., Li, S., Wahala, K., Howald, W.N., Chen, C., 

2001. Wheat bran and soy protein feeding do not alter urinary 

excretion of the isoflavan equol in premenopausal women. J. 

Nutr 131, 740–744. 

Leeming, R., Ball, A., Ashbolt, N., Nichols, P., 1996. Using faecal 

sterols from humans and animals to distinguish faecal 

pollution in receiving waters. Wat. Res. 30, 2893–2900. 


Liggin, J., Bluck, L.J.C., Runswick, S., Atkinson, C., Coward, W.A., 

Bingham, S.A., 2000. Daidzein and genistein content of fruits 

and nuts. J. Nutr. Biochem 11, 326–331. 

Lissin, L.W., Cook, J.P., 2000. Phytoestrogens and cardiovascular 

health. J. Am. Coll. Cardiol 35, 1403–1410. 

Liu, Z.H., Kanjo, Y., Mizutani, S., 2009. Removal mechanisms for 

endocrine disrupting compounds (EDcs) in wastewater 

treatment-physical means, biodegradation, and chemical 

advanced oxidation: A review. Sci. Total Environ 407, 731–748. 

Liu, Z.H., Kanjo, Y., Mizutani, S. Excretions of natural estrogens 

and androgens from humans, and their occurrence and fate in 

the environments. Sci. Total Environ, submitted for 

publication. 

Liu, Z.H., Mamoru, I., Kanjo, Y., Yamamoto, A. Profile and removal 

of endocrine disrupting chemicals (EDCs) in two municipal 

wastewater treatment plants by using an ER (AR) competitive 

ligand binding assay and chemical analyses. J. Environ. Sci. in 

press. doi:10.1016/S1001-0742(08)62356-6. 

Low, Y.L., Taylor, J.I., Grace, P.B., Dowsett, M., Folkerd, E., 

Doody, D., Dunning, A.M., Scollen, S., Mulligan, A.A., Welch, A. 

A., Luben, R.N., Khaw, K.T., Day, N.E., Wareham, N.J., 

Bingham, S.A., 2005. Polymorphisms in the CYP19 gene may 

affect the positive correlations between serum and urine 

phytoestrogen metabolites and plasma androgen 

concentrations in men. J. Nutr 135, 2680–2686. 

Markiewicz, L., Garey, J., Adlercreutz, H., Gurpide, E., 1993. In vitro 

bioassays of non-steroidal phytoestrogens. J. Steroid Biochem. 

Mol. Biol. 45, 399–405. 

Martins, C., Fillmann, G., Montone, R.C., 2007. Natural and 

anthropogenic sterols inputs in surface sediments of Patos 

Lagoon. Brazil. J. Braz. Chem. Soc. 18, 106–115. 

Matthews, J., Celius, T., Halgren, R., Zacharewski, T., 2000. 

Differential estrogen receptor binding of estrogenic 

substances: a species comparison. J. Steroid Biochem. Mol. 

Biol. 74, 223–234. 

Mazur, W., Fotsis, T., Wahala, K., Ojalas, S., Salakka, A., 

Adlercreutz, H., 1996. Isotope dilution gas chromatographicmass 

spectrometric method for the determination of 

isoflavonoids, coumestrol, and lignans in food samples. Anal. 

Biochem 233, 169–180. 

Milder, I.E.J., Feskens, E.J.M., Arts, I.C.W., de Mesquita, H.B.B., 

Hollman, P.C.H., Kromhout, D., 2005. Intake of the plant 

lignans secoisolariciresinol, matairesinol, lericiresinol, and 

pinoresinol in Dutch men and women. J. Nutr 135, 1202–1207. 

Mueller, S., Simon, S., Chae, K., Metzler, M., Korach, K., 2004. 

Phytoestrogens and their human metabolites show distinct 

agonistic and antagonistic properties on estrogen 

receptora(ERa) and ER(b) in human cells. Toxicol. Sci. 80, 

14–25. 

Nagata, Y., Miyanaga, N., Okumura, K., Goto, K., Naito, S., 

Fujimoto, K., Hirao, Y., Takahashi, A., Tsukamoto, T., 

Akaza, H., 2007. Dietary isoflavones may protect against 

prostate cancer in Japanese men. J. Nutr 137, 1974–1979. 

Nakada, N., Nyunoya, H., Nakamura, M., Hara, A., Iguchi, T., 

Takada, H., 2004. Identification of estrogenic compounds in 

wastewater effluent. Environ. Toxicol. Chem. 23, 2807–2815. 

Nakada, N., Higashitani, T., Miyajima, K., Komori, K., Suzuki, Y., 

2006. Development of method for identification of major 

substances inducing estrogenic activity contained in sewage 

and river waters. J. Environ. Chem. 16, 389–401 (in Japanese 

with English abstract). 

Nelson, J., Bishay, F., van Roodselaar, A., Ikonomou, M., Law, F.C. 

P., 2007. The use of in vitro bioassays to quantify endocrine 

disrupting chemicals in municipal wastewater treatment 

plants effluents. Sci. Total Environ 374, 80–90. 

Nielsen, S.E., Freese, R., Kleemola, P., Mutanen, M., 2002. 

Flavonoids in human urine as biomarkers for intake of fruits 

and vegetables. Cancer Epidem. Biomarkers 11, 459–466.

Nishihara, T., Nishikawa, J.I., Kanayama, T., Dakeyama, F., 

Saito, K., Imagawa, M., Takatori, S., Kitagawa, Y., Hori, S., 

Utsumi, S., 2000. Estrogenic activities of 517 chemicals by 

yeast two-hybrid assay. J. Health Sci. 46, 282–298. 

NITE, 2008. http://www.safe.nite.go.jp/english/index.html. 

2008.11. 

Pawlowski, S., Ternes, T., Bonerz, M., Kluczka, T., van der Burg, B., 

Nau, H., Erdinger, L., Braunbeck, T., 2003. Combined in situ and 

in vitro assessment of the estrogenic activity of sewage and 

surface water samples. Toxicol. Sci. 75, 57–65. 

Peeters, P.H.M., Keinan-Boker, L., van der Schouw, Y.T., 

Grobbee, D.E., 2003. Phytoestrogens and breast cancer risk. 

Review of the epidemiological evidence. Breast Cancer Res. 

Treat 77, 177–183. 

Penalvo, J.L., Adlercreutz, H., Uehara, M., Ristimaki, A., 

Watanabe, S., 2008. Lignan content of selected foods from 

Japan. J. Agric. Food Chem. 56, 401–407. 

Phillips, K.M., Ruggio, D.M., Ashraf-Khorassan, M., 2005. 

Phytosterol composition of nuts and seeds commonly 

consumed in the United States. J. Agric. Food Chem. 53, 9436– 

9445. 

Quemeneur, M., Marty, Y., 1994. Fatty acids and sterols in 

domestic wastewaters. Wat. Res. 28, 1217–1226. 

Radke, M., 2005. Sterols and anionic surfactants in urban aerosol: 

emissions from wastewater treatment plants in relation to 

background concentrations. Environ. Sci. Technol 39, 4391– 

4397. 

Reddy, S., Iden, C.R., Brownawell, B., 2005. Analysis of steroid 

conjugates in sewage influent and effluent by liquid 

chromatography-tandem mass spectrometry. Anal. Chem. 77, 

7032–7038. 

Salste, L., Leskinen, P., Virta, M., Kronberg, L., 2007. 

Determination of estrogens and estrogenic activity in 

wastewater effluent by chemical analysis and bioluminescent 

yeast assay. Sci. Total Environ 378, 343–351. 

Seow, A., Shi, C.Y., Franke, A.A., Hankin, J.H., Lee, H.P., Yu, M.C., 

1998. Isoflavonoid levels in spot urine are associated with 

frequency dietary soy intake in a population-based sample of 

middle-aged and older Chinese in Singapore. Cancer Epidem. 

Biomar 7, 135–140. 

Setchell, K.D.R., Cassidy, A., 1999. Dietary isoflavones: biological 

effects and relevance to human health. J. Nutr 129 (Suppl), 

758–767. 

Sivesind, E., Seguin, P., 2005. Effects of the environment, cultivar, 

maturity, and preservation method on red clover isoflavone 

concentration. J. Agric. Food Chem. 53, 6397–6402. 

Smeds, A.I., Willfor, S.M., Pietarinen, S.P., Peltonen-Sainio, P., 

Reunanen, M.H.T., 2007. Occurrence of ‘‘mammalian’’ lignans 

in plant and water sources. Planta 226, 639–646. 


Stackelberg, P.E., Furlong, E.T., Meyer, M.T., Zaugg, S.D., 

Hendersen, A.K., Reissman, D.B., 2004. Persistence of 

pharmaceutical compounds and other organic wastewater 

contaminants in a conventional drinking-water-treatment 

plant. Sci. Total Environ 329, 99–113. 

Stackelberg, P.E., Gibs, J., Furlong, E.T., Meyer, M.T., Zaugg, S.D., 

Lippincott, R.L., 2007. Efficiency of conventional drinkingwater-treatment 

processes in removal of pharmaceuticals and 

other organic compounds. Sci. Total Environ 377, 255–272. 

Song, T.T., Hendrich, S., Murphy, P.A., 1999. Estrogenic activity of 

glycitein, a soy isoflavone. J. Agric. Food Chem. 47, 1607–1610. 

Stopper, H., Schmitt, E., Kobras, K., 2005. Genotoxicity of 

phytoestrogens. Mutat. Res. 574, 139–155. 

Tan, B.L.L., 2006. Chemical and biological analyses of selected 

endocrine disruptors in wastewater treatment plants in South 

East Queensland, Australia. Doctorate dissertation. Griffith 

University, Australia. 

Tonkelaar, I.D., Keinan-Boker, L., Veer, P.V., Arts, C.J.M., 

Adlercreutz, H., Thijssen, J.H.H., Peeters, H.M., 2001. Urinary 

phytoestrogens and postmenopausal breast cancer risk. 

Cancer Epidem. Biomarkers 10, 223–228. 

Tzchori, I., Degani, G., Elisha, R., Eliyahu, R., Hurvitz, A., Vaya, J., 

Moav, B., 2004. The influence of phytoestrogens and 

oestradiol-17b on growth and sex determination in the 

European eel (Anguilla anguilla). Aquat. Res. 35, 1213–1219. 

USDA (US Department of Agriculture), 2003. USDA database for 

the flavonoid content of selected foods. http: www.nal.usda. 

gov/fnic/foodcomp. 

Valentin-Blasini, L., Sadowski, M.A., Walden, D., Caltabiano, L., 

Needham, L.L., Barr, D.B., 2005. Urinary phytoestrogen 

concentrations in the U.S. population (1999–2000). J. Expo. 

Anal. Environ. Epidemiol 15, 509–523. 

Wang, L.Q., 2002. Mammalian phytoestrogens: enterodiol and 

enterolactone. J. Chromatogr. B 777, 289–309. 

Whitten, P.L., Patisaul, H.B., 2001. Cross-species and interassay 

comparisons of phytoestrogen action. Environ. Health Persp 

109 (Suppl), 5–20. 

Xu, X., Duncan, A.M., Wangen, K.E., Kurzer, M.S., 2000. Soy 

consumption alters endogenous estrogen metabolism in 

postmenopausal women. Cancer Epidem. Biomarkers 9, 781–786. 

Zhang, Y., Wang, G.J., Song, T.T., Murphy, P.A., Hendrich, S., 1999. 

Urinary disposition of the soybean isoflavones daidzein, 

genistein and glycitein differs among humans with moderate 

fecal isoflavone degradation activity. J. Nutr 129, 957–962. 

Zhang, X., Cambrai, A., Miesch, M., Roussi, S., Raul, F., Aoude- 

Werner, D., Marchioni, E., 2006. Separation of D5-and D7phytoesterols 

by adsorption chromatography liquid 

chromatography for quantitative analysis of phytosterols in 

foods. J. Agric. Food Chem. 54, 1196–1202.

Occurrence of emerging pollutants in urban wastewater 

and their removal through biological treatment followed 

by ozonation 

Roberto Rosal a, *, Antonio Rodríguez a , José Antonio Perdigón-Melón a , Alice Petre a , 

Eloy García-Calvo a , María José Gómez b , Ana Agüera b , Amadeo R. Fernández-Alba b 

a 

Department of Chemical Engineering, University of Alcalá, 28771 Alcalá de Henares, Spain 

b 

Department of Analytical Chemistry, University of Almería, 04010 Almería, Spain 

article info 


Received 24 February 2009 


26 June 2009 



Keywords: 

Emerging pollutants 


Personal care products 

Wastewater treatment 

Ozonation 


abstract 

The presence of a wide variety of pharmaceutical and 

personal care products (PPCP) in water and wastewater has 

been frequently reported after the findings of Ternes (1998) 

and Daughton and Ternes (1999). These compounds are 

a source of concern because they are used and released in 

large quantities and their physical and chemical properties 




This work reports a systematic survey of over seventy individual pollutants in a Sewage 

Treatment Plant (STP) receiving urban wastewater. The compounds include mainly pharmaceuticals 

and personal care products, as well as some metabolites. The quantification in 

the ng/L range was performed by Liquid Chromatography–QTRAP–Mass Spectrometry and 

Gas Chromatography coupled to Mass Spectrometry. The results showed that paraxanthine, 

caffeine and acetaminophen were the main individual pollutants usually found 

in concentrations over 20 ppb. N-formyl-4-amino-antipiryne and galaxolide were also 

detected in the ppb level. A group of compounds including the beta-blockers atenolol, 

metoprolol and propanolol; the lipid regulators bezafibrate and fenofibric acid; the antibiotics 

erythromycin, sulfamethoxazole and trimethoprim, the antiinflammatories diclofenac, 

indomethacin, ketoprofen and mefenamic acid, the antiepileptic carbamazepine 

and the antiacid omeprazole exhibited removal efficiencies below 20% in the STP treatment. 

Ozonation with doses lower than 90 mM allowed the removal of many individual 

pollutants including some of those more refractory to biological treatment. A kinetic model 

allowed the determination of second order kinetic constants for the ozonation of bezafibrate, 

cotinine, diuron and metronidazole. The results show that the hydroxyl radical 

reaction was the major pathway for the oxidative transformation of these compounds. 


contribute to their widespread distribution into the environment. 

The presence of small concentration of PPCP has been 

associated to chronic toxicity, endocrine disruption and the 

development of pathogen resistance. The consequences are 

particularly worrying in aquatic organisms as they are subjected 

to multigenerational exposure (Halling-Sørensen et al., 

1998). The presence of micropollutants also endangers the 

reuse of treated wastewater, a generally proposed solution to 

* Corresponding author. Department of Chemical Engineering, Facultad de Quamica, University of Alcalá, 28771 Alcalá de Henares, 

Madrid, Spain. Tel.: þ34 91 885 5099; fax: þ34 91 885 5088. 

E-mail address: roberto.rosal@uah.es (R. Rosal). 


doi:10.1016/j.watres.2009.07.004

Notation 

Ci concentration of a given organic compound (M) 

CHO$ concentration of hydroxyl radical (M) 

CO3 concentration of dissolved ozone (M) 

CO3 equilibrium concentration of dissolved 

ozone (M) 

DO3 diffusivity of ozone in water, m 2 s 1 

Dow apparent octanol–water partition coefficient 

(dimensionless) 

fOH fraction of a given compound degraded by 

hydroxyl radicals 

Ha Hatta number, defined in Eq. (5) (dimensionless) 

achieve a sustainable water cycle management (Muñoz et al., 

2009). PPCP represent a rising part of the trace organic 

micropollutants found in urban and domestic wastewaters 

that reach sewage treatment plants (STP), either metabolised 

or not (Castiglioni et al., 2006). Many of these substances 

escape to conventional activated sludge wastewater treatments 

allowing them to reach surface water streams and 

distribute in the environment (Tauxe-Wuersch et al., 2005). 

The need for treatment technologies that can provide safe 

treated effluents led to the proposals for upgrading STP and to 

implement new competing technologies for biological degradation 

of organic matter like membrane bioreactor (Radjenovic 

et al., 2009). In addition to these strategies, effective 

tertiary treatment technologies are also needed in order to 

ensure a safe use for reclaimed wastewater. The available 

technologies include oxidation processes alone or combined 

with nanofiltration or reverse osmosis (Ernst and Jekel, 1999). 

Many oxidation processes have been described for the 

removal of organic compounds in wastewater. Ozone-based 

and Advanced Oxidation Processes (AOP) using hydrogen 

peroxide or radiation have been repeatedly proposed for this 

task (Gogate and Pandit, 2004a; Ikehata et al., 2006). Proper 

combinations of AOP can also be considered in order to treat 

the more refractory pollutants. Fenton and Fenton-based 

systems, heterogeneous photocatalysis, and ultraviolet or 

ozone-based oxidation processes have been described (Gogate 

and Pandit, 2004b; Comninellis et al., 2008). The choice of the 

most suitable technology or combination lies on the quality 

required for the reclaimed water. 

Ozone has been largely used as oxidant in drinking water 

treatment and repeatedly proposed to remove organics in 

wastewater treatment (Raknes, 2005; Beltrán, 2004). The ozone 

molecule can react with many organic compounds, particularly 

those unsaturated or containing aromatic rings or 

heteroatoms also being able to decompose in water to form 

hydroxyl radicals. In a previous work (Rosal et al., 2008a) we 

studied the ozonation of wastewater from the secondary 

clarifier of urban and domestic STP by using ozone (pH w 8) and 

ozone–hydrogen peroxide (O3/H2O2). The presence of hydrogen 

peroxide improved the mineralization of dissolved carbon 

from 15% to over 90% after one hour, from which most part took 

place during the first five minutes. The disappearance of 

a selected group of over thirty pharmaceuticals revealed 

removal efficiencies were over >99% for most compounds after 


kL liquid-phase individual mass transfer 

coefficient, m s –1 

kO3 second order ozone-based rate constant for 

a direct ozonation reaction (M 1 s 1 ) 

kHO$ second order rate constant for a reaction with 

hydroxyl radical (M 1 s 1 ) 

kR second order rate constant for the ozonation of 

a given compound (M 1 s 1 ) 

kLa volumetric mass transfer coefficient (s 1 ) 

Kow octanol–water partition coefficient for neutral 

species (dimensionless) 

TOD dose of ozone transferred to the liquid (M) 

five minutes on stream, with slightly better results in the 

absence of hydrogen peroxide. This result is consistent with 

the fact that many PPCP directly react with ozone with large 

second order kinetic constants (Huber et al., 2003). 

The objective of this research was to verify the occurrence 

and fate of 84 pollutants of different classes, mainly PPCP 

traced during wastewater treatment in a conventional urban 

STP. These include: pharmaceuticals (analgesics, antidepressants, 

antiinflammatories, antibiotics, antiepileptics, betablockers 

and lipid regulators among others), personal care 

products (sunscreen agents, synthetic musks), stimulants 

(caffeine, nicotine) and some metabolites (clofibric acid, cotinine, 

several metabolites of dipyrone). The effectiveness of 

the STP process for the removal of these compounds has been 

assessed in a monitoring program undertaken in Alcalá de 

Henares (Madrid) during a one-year period with samples 

taken before and after a biological activated sludge with 

nutrient removal process. The research was also trying to 

identify the impact of ozone exposure on individual pollutants 

encountered in the secondary effluent. The doses of ozone 

required for a given degree of removal of the most representative 

individual pollutants has been assessed. For certain 

pollutants, the determination of second order kinetic 

constants for the ozonation reaction in a real wastewater 

matrix was also performed. 


2.1. Materials and plant description 

Wastewater samples were taken every month over a one-year 

period from the input and output of the secondary clarifier of 

a STP located in Alcalá de Henares (Madrid). This plant treats 

a mixture of domestic and industrial wastewater from some 

facilities located near the city with a nominal capacity of 

3000 m 3 /h of raw wastewater. The Council Directive 91/271/ 

EEC concerning urban wastewater treatment established 

strict limits for the wastewater discharged to sensitive areas 

particularly by plants serving an equivalent population of 

10 000 or more. In the period prior to the sampling campaign, 

the need to adapt STP to the new conditions forced, the 

implementation of a biological nutrient removal process 

aimed to the elimination of phosphorous and nitrogen. The

580 

biological treatment worked under a traditional A2O multistage 

configuration with nitrification–denitrification and 

enhanced phosphorus removal by phosphorus-accumulating 

microorganisms. The treatment takes place in three zones: 

anaerobic, anoxic and oxic. The nitrate produced in the oxic 

zone is recycled, with mixed liquor to the anoxic zone where 

denitrification takes place. The return sludge from the settler 

is recycled to the anaerobic zone where the influent and 

sludge are mixed under anaerobic conditions. All samples 

were immediately processed or stored in a refrigerator (

2.5. Ozonation 

The ozonation runs were carried out in a 5-L glass jacketed 

reactor operating in semi-batch mode. A temperature of 25 C, 

chosen to be close to average ambient temperature, was kept 

using a Huber Polystat cc2 and monitored throughout the runs 

bymeansofaPt100ResistanceTemperatureDetector(RTD). 

Ozone was produced by a corona discharge ozonator (Ozomatic, 

119 SWO100) fed by an AirSep AS-12 PSA oxygen 

generation unit. The gas containing about 9.7 g/Nm 3 ozone 

was bubbled by means of a porous glass disk with a gas flow of 

0.36 Nm 3 /h. The reaction vessel was agitated with a Teflon 

four-blade impeller at 1000 rpm. The mass transfer coefficient 

was determined in transient runs with pure water with 

a value of kLa ¼ 0.010 0.005 s 1 . Additional details on the 

experimental set-up and procedure are given elsewhere 

(Rosal et al., 2008a,b). Throughout the runs, certain samples 

were withdrawn for analysis at prescribed intervals. Residual 

ozone was removed by bubbling nitrogen in order to prevent 

oxidation reactions to continue. For the desorption conditions 

used, residual ozone fell down below 10% in less than 20 s. 

During the ozonation reactions a moderate increase of pH 

took place that can be attributed to the stripping of CO2 from 

solution. In conditions that favour the generation of hydroxyl 

radicals, this effect is not observed and pH tend to decrease 

during ozonation due to the accumulation of carboxylic acids. 

The choice of a pH higher than that of the raw wastewater not 

only favours hydroxyl radical mediated reactions, but allowed 

to keep an almost constant pH value of 8.5 0.1 by using the 

feed-back control procedure described above. 


3.1. Occurrence 

The pollutants analyzed were mainly PPCP and their metabolites 

together with some agrochemicals. A list of the 

72 anthropogenic emerging pollutants detected in at least one 

wastewater sample from the influent to the biological treatment 

is detailed in the Appendix together with the analytical 

method applied, the molecular formula and the octanol–water 

partition coefficient, when available. The limit of quantification 

for the biological effluent (LOQ), for most compounds in 

the tens of ng/L, is also shown in the Appendix that also lists 

the compounds checked but not detected with their limits of 

detection (LOD). Several pesticides like chlorfenvinphos and 

the herbicide isoproturon were never detected, as expected 

considering the urban origin of wastewater. Atrazine, 

however, was found in all samples with an average concentration 

at the inlet of the biological treatment of 109 ng/L. 

Diuron was encountered in two samples and simazine in 

three with maximum concentrations of 196 and 32 ng/L, 

respectively. Some drugs like paroxetine, the antibiotic cefotaxime 

and the antiinflammatory fenoprofen were not 

detected in any of the analyzed samples. A similar negative 

result was reported for paroxetine by Terzić et al. (2008), 

whereas the occurrence of fenoprofen was reported in 

concentrations as high as 0.759 mg/L in Canadian wastewater 

facilities (Metcalfe et al., 2004). 


The detailed data for the concentrations of the most significant 

individual pollutants are shown in Table 2. Itincludes 

maximum and minimum values of those compounds encountered 

over their quantification limit in at least 4 samples in the 

influent of the biological treatment. The compounds excluded 

for not complying with this criterion comprise, among other, 

carbamazepine-10,11-epoxide that was detected in 2 samples 

with a maximum concentration of 63 ng/L, sotalol, detected 

three times in the influent with concentrations in the 25–30 ng/L 

range or celestolide with a maximum influent concentration of 

30 ng/L. In addition, fenofibrate was encountered in one sample 

with 101 ng/L, whereas its human metabolite, fenofibric acid 

has been detected in all samples at concentrations as high as 

117 ng/L. Table 2 also excludes some compounds not systematically 

analyzed during the sampling campaign due to 

improvements in the analytical procedure. Certain compounds 

such as terbutaline or simazine have been encountered in 

several samples, but these data have not been considered 

statistically significant because they were found at concentrations 

so close to the LOQ that removal efficiency in STP could not 

be properly assessed and, therefore, they were not included in 

Table 2. Among them, diazepam was detected in half of the 

samples with a maximum concentration of 8 ng/L in the 

influent and 5 ng/L in the effluent, the average in both cases 

being near 3 ng/L (that equals LOQ). Also, mepivicaine was 

found in 7 samples with an average concentration of 8 ng/L in 

the influent and 7 ng/L in the effluent and maximum concentration 

in the influent of 14 ng/L. 

Paraxanthine, caffeine and acetaminophen were the individual 

pollutants usually found in higher concentration, with 

averages near 20 ppb in the influent. N-formyl-4-amino-antipiryne 

(4-FAA) and galaxolide also exhibited averages in the 

mg/L level. The fact that caffeine is the dominant micropollutant 

in STP is not new. Caffeine has been detected in 

many surface streams and STP effluents in concentration as 

high as 230 mg/L (Ternes et al., 2001; Heberer et al., 2002). 

Recently, Wilcox et al. (2009) also report caffeine, paraxanthine 

and acetaminophen as the most frequently detected 

compounds, in the influent of conventional septic systems. 

Muñoz et al. (2009) found several substances at mg/L level, with 

the highest concentrations corresponding to caffeine, acetaminophen, 

atenolol, and paraxanthine, that exceeded 40 mg/L 

each. In this work, atenolol was part of the group of 

compounds found with averages over the ppb limit that 

included ciprofloxacin, hydrochlorthiazide, ibuprofen, 

N-acetyl-4-amino-antipiryne (4-AAA), naproxen, nicotine, 

and ofloxacin. The occurrence of dipyrone (metamizol) residues 

in STP effluents has been less frequently assessed, but 

Feldmann et al. (2008) have recently reported concentrations 

up to 7 mgL 1 in influents and effluents of STP in Berlin. These 

dipyrone residues have been attributed to effluents originated 

in hospitals more than to private households. Nicotine was 

found in concentrations as high as 12 mg/L in the influent 

and 148 ng/L in the effluent. Cotinine, its major urinary 

metabolite, was detected in the biological effluent with 

a concentration of 100 ng/L. These values agree with those 

published by Buerge et al. (2008) who found concentrations of 

cotinine of approximately 1–10 mg/L in untreated wastewater, 

and 0.01–0.6 mg/L in the treated effluent, corresponding to 

elimination efficiencies of over 90%.

582 

Table 2 – Concentrations of pollutants in the influent and effluent of the studied STP calculated for compounds detected 

over LOQ in at least four influent samples along the monitoring programme. Averages and removal efficiencies have been 

calculated excluding concentrations below LOQ. 

Compound pKa a 

Influent (ng/L) Effluent (ng/L) Removal 

Maximum Minimum Average Maximum Minimum Average 

efficiency (%) 

4-aminoantipyrine (4-AA) 4.3 3325 262 1517 2253 127 676 55.4 

4-methylaminoantipyrine (4-MAA) 4.3 1894 314 880 1098 34 291 66.9 

Acetaminophen 9.4 37 458 1571 23 202

elow LOQ. It is well known that during wastewater treatment 

in STP, PPCP as well as their metabolites partition into the 

solid phase or remain dissolved depending on their hydrophobicity. 

The hydrophobicity of a neutral compound can be 

expressed as its octanol–water partition coefficient, Kow, but in 

the case of compounds that can exist in ionized form, the 

acid–base equilibrium must also be taken into account. The 

pH-dependent or apparent octanol–water distribution coefficient, 

D ow, that considers both the dissociation constant of 

acidic of basic solutes, pK a, and the current pH of wastewater 

can be derived from the Herderson–Hasselbalch equations 

(Scheytt et al., 2005). For acidic compounds, like indomethacin 

or naproxen, that are dissociated at the pH of wastewater, the 

equation yields: 

Kow 

Dow ¼ 

1 þ 10pH pKa 

In the case of basic drugs, the apparent partition coefficient 

can be expressed using the pKa for the corresponding conjugate 

acid: 

Kow 

Dow ¼ 

1 þ 10pKa pH (2) 

Codeine or fluoxethine are among compounds that are 

substantially dissociated at ambient pH, whose average was 

7.61 0.39, with boundaries representing the 95% confidence 

interval. For neutral substances, D ow ¼ K ow. 

There is experimental evidence that the removal of organic 

pollutants in STP is largely controlled by sorption process with 

solid–water distribution coefficients being a function of the 

octanol–water distribution coefficient Dow (Carballa et al., 

Removal efficiency (%) 

100 

90 

80 

70 

60 

50 

40 

30 


6 

(1) 

1 2 

2008). The removal efficiency obtained in this work for the 

compounds indicated in Table 2 has been related to Dow and 

the results represented in Fig. 1. For a significant group of 

compounds, ranging from ketorolac (44%) to galaxolide (88%), 

a clear relationship is observed between removal efficiency 

and Dow. A group of five compounds were almost completely 

removed in the STP even they present relatively low Dow 

values. They are the metabolite of caffeine paraxanthine, 

caffeine itself, acetaminophen (paracetamol), ibuprofen and 

nicotine, compounds otherwise usually found in high 

concentrations in raw urban wastewater (Table 2). A similar 

result was reported by Muñoz et al. (2008) who reported 

concentrations of caffeine, acetaminophen, atenolol, and 

paraxanthine that exceeded 40 mg/L each but a substantial 

removal in an activated sludge STP (90%, >99%, 43%, and 67%, 

respectively). Joss et al. (2005) reported removal efficiencies for 

ibuprofen beyond its quantification limit (>90%) and also 

coincident are data corresponding to naproxen (50–80%). 

This work also identified 14 compounds with removal 

efficiencies fell below 20%, but exhibiting intermediate D ow 

values ( 2 < D ow < 2.5). They have been marked by the lower 

square in Fig. 1 and are the beta-blockers atenolol, metoprolol 

and propanolol; the lipid regulator bezafibrate and the 

metabolite of fenofibrate fenofibric acid; the antibiotics 

erythromycin, sulfamethoxazole and trimethoprim; the antiinflammatories 

diclofenac, indomethacin, ketoprofen and 

mefenamic acid; the antiepileptic carbamazepine and the 

antiacid omeprazole. Joss et al. (2005) reported no significant 

removal of sulfamethoxazole and carbamazepine and partial 

elimination of diclofenac, results generally coincident with 

this work. For musk fragrances, galaxolide and tonalide, Joss 

3 4 

5 

22 

17 

11 

18 

19 

8 

7 

10 

12 

9 14 

21 

16 15 

20 

28 

10 

0 

-8.00 -6.00 -4.00 

27 32 

35 

31 

39 

30 34 

29 

40 

36 37 

-2.00 0.00 2.00 4.00 6.00 8.00 

log D ow 

Fig. 1 – Removal efficiency during conventional activated sludge treatment: (1) paraxanthine, (2) caffeine, 

(3) acetaminophen,(4) nicotine, (5) ibuprofen,(6) ketorolac,(7) clofibric acid,(8) furosemide, (9) ciprofloxacin,(10) fluoxethine, 

(11) ofloxacin,(12) naproxen,(13) hydrochlorothiazide, (14) 4-aminoantipyrine, (15) metronidazole, (16) N-acetyl-4-aminoantipiryne, 

(17) codeine, (18) N-formyl-4-amino-antipiryne, (19) 4-methylaminoantipyrine, (20) ranitidine, (21) antipyrine, 

(22) gemfibrozil, (23) benzophenone-3,(24) triclosan,(25) tonalide, (26) galaxolide, (27) atenolol, (28) sulfamethoxazole, 

(29) fenofibric acid, (30) metoprolol, (31) bezafibrate, (32) ketoprofen, (33) trimethoprim, (34) diclofenac, (35) indomethacine, 

(36) propanolol, (37) mefenamic acid, (38) omeprazole, (39) carbamazepine, (40) erythromycin. 

23 

24 

26 

25

584 

et al. (2005) reported, however, removal efficiencies somewhat 

lower (>50%) than those found in this work (w85%). Carballa 

et al. (2004) also reported overall removal efficiencies of the 

ranging between 70% and 90% for fragrances in coincidence 

with this work and generally higher for antiinflammatories 

(40–60%) and sulfamethoxazole. In a further work, Carballa 

et al. (2007) indicated different removal efficiencies for several 

PPCP during the anaerobic digestion of sewage sludge with 

high efficiencies for naproxen or sulfamethoxazole. In 

general, a high variability can be expected as a consequence of 

different pollutant concentrations, changes in the processes 

of the STP as well as operational conditions (Liu et al., 2008). 

3.3. Removal by ozonation 

The ozonation of a dissolved organic compound may take 

place by direct reaction with ozone molecule and also through 

the action of secondary oxidants produced from ozone in 

aqueous medium. Among them, the strongest oxidant is 

hydroxyl radical, generally associated with the oxidation of 

the more refractory substances. A mass balance to a given 

compound in solution yields the following expression: 

dCi 

dt ¼ kO 3 CiCO 3 þ kHO$CiCHO$ 

Elovitz and von Gunten (1999) introduced a parameter R ct 

defined as the relationship between the concentrations of 

ozone and hydroxyl radical at any moment. Based on the 

observation that R ct can be considered constant at least 

though certain periods of the ozonation runs, the concentration 

of the two main oxidants involved in ozonation reactions 

can be related so that Eq. (3) can be solved without experimental 

determination of CHO$: 

ln Ci 

Z t 

Z t 

¼ kO CO dt þ kHO$ 

3 

3 

Cio 

0 

0 

Z t 

¼ kR 

0 

CO dt 3 

CHO$dt ¼ kO 3 þ RctkHO$ 

Z t 

CO dt 3 

0 

The kinetic parameter kR would behave as second order 

kinetic constant provided Rct is constant throughout the 

sampling period. As shown below, this work shows that Rct is 

constant during ozonation at least for the samples taken after 

ozone appeared in solution. The preceding findings also lie on 

the assumption of slow kinetic regime. The kinetics of 

a heterogeneous gas–liquid semicontinuous process is determined 

by the relative rates of absorption and chemical reaction 

and is characterized by Hatta number. It represents the 

maximum rate of chemical reaction relative to the maximum 

rate of mass transfer, yielding, for a second order reaction, the 

following expression: 

Ha ¼ 

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 

kO Ci;oDO 3 3 

kL 

In the case of wastewater treatment, ozone reacts with 

many compounds in a complex parallel reaction system so 

that kO Ci;o can be substituted by 3 P 

kO Ci;o. As raw wastewater 

3 

i 

contains a number of compounds whose second order direct 

reaction constants with ozone are very large (Huber et al., 


(3) 

(4) 

(5) 

2003) mass transfer is likely to limit the ozonation rate during 

the first reaction minutes. Once ozone appears in solution, the 

following inequality holds: 

kLa C O3 

CO 3 

> X 

i 

kO 3 Ci;oCO 3 

Therefore an upper limit can be obtained for Ha as follows: 

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 

Ha < 

kLa C O 3 

C O3 

kL 

1 DO 3 

where DO 3 is the diffusivity of ozone in water 

(1.77 10 9 m 2 s 1 ). The equilibrium concentration of ozone 

C O3 was calculated from Henry’s law using the correlation of 

Rischbieter et al. (2000) from the concentration in the gas phase 

that was 9.4 g/Nm 3 . The value of the mass transfer coefficient, 

kL ¼ 5.5 10 5 ms 1 , was calculated by using the correlation of 

Calderbank and Moo-Young (1961). At the beginning of the run, 

there was no ozone in solution and Ha is supposed to reach 

high values. After about three minutes, the concentration of 

ozone that increased during runs approaching equilibrium was 

over 2 10 3 mM. This, considering that C O3 ¼ 0:05 mM, 

ensures Ha < 0.3 and, therefore, the kinetic regime was slow and 

for the sample taken at 4 min and those taken thereafter. 

Second order kinetic constants could be obtained as indicated 

before for bezafibrate, cotinine, diuron, ketoprofen and 

metronidazole. The logarithmic concentration decay is represented 

against the integral ozone dose in Fig. 2 in which, the 

experimental points represent data from samples taken at 4, 6, 

10 and 15 min, the first being considered the initial concentration, 

Cio. The comparison with literature data shows that the 

transformation of ozone-resistant micropollutants takes place 

primarily via indirect radical oxidation inhibited by the wastewater 

matrix. The values reported in the literature for 

the second order direct and indirect rate constants are: 

bezafibrate, kO3 ¼ 590 M 1 s 1 and kHO$ ¼ 7:40 10 9 M 1 s 1 

in Huber et al. (2005); ketoprofen, kO3 ¼ 0:4 M 1 s 1 and 

kHO$ ¼ 8:40 10 9 M 1 s 1 in Real et al. (2009), diuron, 

ln(C io/C i) 

4 

3 

2 

1 

0 

0 0.2 0.4 0.6 0.8 1 

Fig. 2 – Logarithmic decay of the concentration of diuron 

(B), metronidazole (C), ketoprofen (D) bezafibrate (,) and 

cotinine (-) as a function of the integral ozone exposure. 

(Two y-axes have been used for clarity.) 

6 

5 

4 

3 

2 

1 

0 

(6) 

(7) 

ln(C io/C i)

kO3 ¼ 16:5 M 1 s 1 and kHO$ ¼ 4:6 109 M 1 s 1 in Benitez et al. 

(2007) and kO3 ¼ 14:7 M 1 s 1 and kHO$ ¼ 6:6 109 M 1 s 1 in 

De Laat et al. (1996) ; metronidazole, kO3 ¼ < 350 M 1 s 1 

in Sánchez-Polo et al. (2008) and kHO$ ¼ 1:98 109 M 1 s 1 in 

Johnson and Mehrvar (2008). No data have been published 

for cotinine. The ozonation of bezafibrate was also studied 

by Dantas et al. (2007) who reported a direct kinetic constant 

kO3 ¼ 4:24 103 0:66 103 M 1 s 1 for pH 7 and kR ¼ 1.0 

10 4 

1.07 10 3 M 1 s 1 for pH 8. The reported direct rate 

constant for ketoprofen is particularly low and allows to calculate 

Rct by dividing the value of kR obtained in this work by kHO$ 

reportedby Real et al. (2009). The value found, Rct ¼ 3.62 10 7 ,is 

very close to those that could be derived using Eq. (4) for bezafibrate 

and diuron. The partial contribution of the direct ozone 

and radical pathways can be calculated for a given organic 

compound as follows: 

kHO$CHO$Ci 

fOH ¼ 

kHO$CHO$Ci þ kO Ci 3 

¼ 

kHO$Rct 

kHO$Rct þ kO 3 

Using literature constants, the fraction degraded by hydroxyl 

radicals were 0.819 for bezafibrate, 0.996 for diuron, 1.000 for 

ketoprofen and >0.887 for metronidazole. The contribution of 

direct ozone reaction is obviously low because the five 

compounds studied are relatively refractory to ozonation. By 

means of Eq. (8) and taking into account that kR ¼ kO þ kHO$$Rct, 

3 

the second order rate constants, kR, can be derived from literature 

data with the following results: metronidazole 

1.07 10 3 M 1 s 1 , bezafibrate 3.27 10 3 M 1 s 1 and diuron 

2.41 10 3 M 1 s 1 (Benitez et al., 2007). The two last are in good 

agreement with experimental values. In the case of metronidazole, 

the low value obtained for Rct suggests that kHO could 

have been underestimated. 

The efficiency of ozonation for the removal of the main 

micropollutants whose concentration in biologically treated 

wastewater was >10 ng/L is indicated in Table 3. The table 

shows the evolution of the concentration for samples taken 

during ozonation up to a reaction time of 15 min and the 

amount of ozone required for a given degree of removal. Only 

concentrations above LOQ are shown. These LOQ apply for 

ozonated samples and are lower than those reported in the 

Appendix for untreated samples. They could be reached, 

because of the higher preconcentration factor applied to the 

ozonated samples (see experimental section) and the reduction 

in the matrix effects observed in these cleaner extracts. 

The amount of ozone transferred to the liquid at a certain 

reaction time, TOD, was determined, from the integration of 

the ozone absorption rate equation: 

0 

1 

Z t 

TODðtÞ ¼ kLa@C 

t CO dtA 

(9) 

O3 3 

0 

For the experimental conditions used in this work CO 3 was 

always less then 10% of C O3 and, therefore, TOD(t) was 

essentially linear with time. The last right column of Table 3 

shows the dose of ozone required for the complete removal 

(no detection) of a given compound or to achieve certain 

removal efficiency in cases where ozonation was unable to get 

complete oxidation in less than 15 min corresponding to 

a dose of ozone of 0.34 mmol/L of wastewater. It is to be noted 

that a change in the pH of wastewater modifies the ozone 


(8) 

doses required for a given effect. A pH increase causes 

a higher hydroxyl radical exposure increasing Rct and leading, 

for a similar integral radical exposure, to lower ozone doses. 

Also, it is important to point out that a change of pH may 

affect the direct ozonation rates of dissociating compounds. In 

our case, raising pH from that of raw wastewater to 8.5, at 

which the ozonation took place, could affect the ozonation 

rate of some proton-accepting compounds whose pKa w 8, 

particularly benzophenone-3, hydrochlorthiazide, nicotine, 

ofloxacin and triclosan. 

A group of 15 compounds rapidly disappear during the first 

120 s on stream, with ozone doses

586 

Table 3 – Removal of pollutants contained in wastewater during ozonation. The ozone doses are those required to reach 

concentrations below the limit of quantification (LOQ a ) in treated samples. 

Ozonation time (min) LOQ a 

0 2 4 6 10 15 Ozone doses 

for remotion 

3-(4-methylbenzylidene) camphor 39 55 50 65 39 72 54 Not removed 

4-aminoantipyrine 19 58 – – – – –


This research showed the regular presence of over seventy 

anthropogenic individual pollutants, some of which are 

encountered in relatively high amounts. In raw sewage, 

25 compounds were detected in the mg/L range, 15 of which 

exceeded this level in yearly averages. Acetaminophen (paracetamol) 

and caffeine were persistently detected over 1 ppb 

in untreated wastewater. Galaxolide was also encountered in 

high concentration, but its occurrence showed a significant 

variability. Two metabolites, paraxanthine from caffeine and 

4-FAA from metamizol were also found in high amounts in 

almost all samples. Other metabolites detected were 4-AA and 

4-MAA from dipyrone, fenofibric acid from fenofibrate and 

4-AAA also from metamizol. This findings stress the need for 

exploring not only the pattern PPCP, but also their metabolic 

or photodegradation intermediates. 

The efficiency of removal of PPCP in STP was roughly 

dependent on its hydrophobicity expressed as apparent 

octanol–water distribution coefficient, D ow, a parameter that 

takes into account the octanol–water partition coefficient, 

Kow, as well as the dissociation constant of acidic or basic 

compounds. For most compounds, the removal efficiency 

during biological treatment increased with hydrophobicity as 

expected considering the higher sorption of non-polar 

compounds on sludge. Certain compounds showed important 

deviations with a group of relatively polar substances formed 

by paraxanthine, caffeine, acetaminophen, ibuprofen and 

nicotine that were almost completely removed during biological 

treatment. Another group formed by 14 compounds 

exhibited removal efficiencies below 20%, even their octanol– 

water distribution coefficient was not particularly low. They 

are all important pharmaceuticals prescribed and delivered to 

sewage in high amounts and include the beta-blockers atenolol, 

metoprolol and propanolol; the lipid regulator bezafibrate, 

fenofibric acid; the antibiotics erythromycin, 

sulfamethoxazole and trimethoprim; the antiinflammatories 

diclofenac, indomethacin, ketoprofen and mefenamic acid; 

the antiepileptic carbamazepine and the antiacid omeprazole. 

An ozonation treatment yielded high removal efficiencies 

of most individual pollutants detected in treated wastewater. 

The kinetic analysis of the part of the run that takes place in 

the slow kinetic regime, allowed the determination of second 

order kinetic constants for the ozonation of bezafibrate, cotinine, 

diuron, ketoprofen and metronidazole in its wastewater 

matrix. This work shows that Rct, a concept developed for 

drinking water, can be applied for the ozonation of wastewater 

as it was constant at least for the samples taken after 

ozone appeared in solution. The ratio of the concentrations of 

ozone and hydroxyl radical, Rct, during this period was 

3.62 10 7 and the fraction degraded by hydroxyl radicals was 

in the 0.819–1.000 range for the aforementioned compounds. 

Even when most compounds disappear for doses lower than 

340 mM, and most for less than 90 mM, some pollutants, 

essentially personal care products were not significantly 

removed during ozonation. These were the UV filters 

3-(4-methylbenzylidene) camphor and Ethylkexyl methoxycinnamate, 

the sunscreen agent Benzophenone-3 and the 

aromatic nitro musk xylene. On the other hand, certain polar 


compounds like diclofenac, indomethacin or the betablockers 

atenolol, metoprolol and propanolol, which are 

poorly removed in the activated sludge conventional treatment, 

exhibit large ozonation rates and can be removed from 

treated wastewater using moderate ozone doses. 


The authors wish to express their gratitude to the Ministry of 

Education of Spain (Contract CONSOLIDER-INGENIO 2010 

CSD2006-00044), the Dirección General de Universidades e 

Investigación de la Comunidad de Madrid, Research network 

0505/AMB-0395. 

Appendix. 

Supplementary information 


in the doi: 10.1016/j.watres.2009.07.004. 

references 

Beltrán, F.J., 2004. Ozone Reaction Kinetics for Water and 

Wastewater Systems. CRC, Boca Raton, pp. 113–149. 

Benitez, F.J., Real, F.J., Acero, J.L., Garcia, C., 2007. Kinetics of the 

transformation of phenyl-urea herbicides during ozonation of 

natural waters: rate constants. Water Res. 41, 4073–4084. 

Bueno, M.J.M., Agüera, A., Gómez, M.J., Hernando, M.D., 

García, J.F., Fernández-Alba, A.R., 2007. Application of liquid 

chromatography/quadrupole-linear ion trap mass 

spectrometry and time-of-flight mass spectrometry to the 

determination of pharmaceuticals and related contaminants 

in wastewater. Anal. Chem. 79, 9372–9384. 

Buerge, I.J., Kahle, M., Buser, H.R., Müller, M.D., Poiger, T., 2008. 

Nicotine derivatives in wastewater and surface waters: 

application as chemical markers for domestic wastewater. 

Environ. Sci. Technol. 42, 6354–6360. 

Calderbank, P.H., Moo-Young, M.B., 1961. The continuous phase 

heat and mass transfer properties of dispersions. Chem. Eng. 

Sci. 16, 39–54. 

Carballa, M., Fink, G., Omil, F., Lema, J.M., Ternes, T., 2008. 

Determination of the solid–water distribution coefficient (K d) 

for pharmaceuticals, estrogens and musk fragrances in 

digested sludge. Water Res. 42, 287–295. 

Carballa, M., Omil, F., Lema, J.M., Llompart, M., García-Jares, C., 

Rodríguez, I., Gómez, M., Ternes, T., 2004. Behavior of 


treatment plant. Water Res. 38, 2918–2926. 

Carballa, M., Omil, F., Ternes, T., Lema, J.M., 2007. Fate of 

pharmaceutical and personal care products (PPCPs) during 

anaerobic digestion of sewage sludge. Water Res. 41, 

2139–2150. 



treatment plants in Italy. Environ. Sci. Technol. 40, 357–363. 

Comninellis, C., Kapalka, A., Malato, S., Parsons, S.A., Poulios, I., 

Mantzavinos, D., 2008. Advanced oxidation processes for 

water treatment: advances and trends for R&D. J. Chem. 

Technol. Biotechnol. 83, 769–776.

588 

Dantas, R.F., Centerino, M., Marotta, R., Sans, C., Esplugas, S., 

Andreozzi, A., 2007. Bezafibrate removal by means of 

ozonation: primary intermediates, kinetics, and toxicity 

assessment. Water Res. 41, 2525–2532. 



Environ. Health Perspect. 107, 907–938. 

De Laat, J., Maouala-Makata, P., Dore, M., 1996. Rate constants for 

reactions of ozone and hydroxyl radicals with several phenylureas 

and acetamides. Environ. Technol. 17, 707–716. 

Elovitz, M.S., von Gunten, U., 1999. Hydroxyl radical/ozone ratios 

during ozonation processes I. The R ct concept. Ozone Sci. Eng. 

21, 239–260. 

Ernst, M., Jekel, M., 1999. Advanced treatment combination for 

groundwater recharge of municipal wastewater by 

nanofiltration and ozonation. Water Sci. Technol. 40, 277–284. 

Feldmann, D.F., Zuehlke, S., Heberer, T., 2008. Occurrence, fate and 

assessment of polar metamizole (dipyrone) residues in hospital 

and municipal wastewater. Chemosphere 71, 1754–1764. 

Gogate, P.R., Pandit, A.B., 2004a. A review of imperative technologies 

for wastewater treatment I: oxidation technologies at ambient 

conditions. Adv. Environ. Res. 8, 501–551. 

Gogate, P.R., Pandit, A.B., 2004b. A review of imperative 

technologies for wastewater treatment II: hybrid methods. 

Adv. Environ. Res. 8, 553–597. 

Gómez, M.J., Gómez-Ramos, M.M., Agüera, A., Mezcua, M., 

Herrera, S., Fernández-Alba, A.R., 2009. A new gas 

chromatography/mass spectrometry method for the 

simultaneous analysis of target and non-target organic 

contaminants in waters. J. Chromatogr. A, 1216, 4071–4082. 

Halling-Sørensen, B., Nielsen, B.N., Lanzky, P.F., Ingerslev, F., 

Lützhoft, H.C.H., Jorgensen, S.E., 1998. Occurrence, fate and 

effects of pharmaceutical substances in the environment, 

a review. Chemosphere 36, 357–394. 

Heberer, T., Reddersen, K., Mechlinski, A., 2002. From municipal 

sewage to drinking water: fate and removal of pharmaceutical 

residues in the aquatic environment in urban areas. Water Sci. 

Technol. 46, 81–88. 

Huber, M.M., Canonica, S., Park, G.Y., Gunten, U., 2003. Oxidation 

of pharmaceuticals during ozonation and advanced oxidation 

processes. Environ. Sci. Technol. 37, 1016–1024. 

Huber, M.M., Göbel, A., Joss, A., Hermann, N., Löffler, D., 

Mcardell, C.S., Ried, A., Siegrist, H., Ternes, T.A., von 




Ikehata, K., Naghashkar, N.J., El-Din, M.G., 2006. Degradation of 

aqueous pharmaceuticals by ozonation and advanced 

ozonation processes: a review. Ozone Sci. Eng. 28, 353–414. 

Johnson, M.B., Mehrvar, M., 2008. Aqueous metronidazole 

degradation by UV/H 2O 2 process in single- and multi-lamp 

tubular photoreactors: kinetics and reactor design. Ind. Eng. 

Chem. Res. 47, 6525–6537. 

Joss, A., Keller, E., Alder, A.C., Göbel, A., McArdell, C.S., Ternes, T., 

Siegrist, H., 2005. Removal of pharmaceuticals and fragrances 

in biological wastewater treatment. Water Res. 39, 3139–3152. 

Kupper, T., Plagellat, C., Brändli, R.C., de Alencastro, L.F., 

Grandjean, D., Tarradellas, J., 2006. Fate and removal of 

polycyclic musks, UV filters and biocides during wastewater 

treatment. Water Res. 40, 2603–2612. 

Liu, Z., Kanjo, Y., Mizutani, S., 2008. Removal mechanisms for 

endocrine disrupting compounds (EDCs) in wastewater 

treatment d physical means, biodegradation, and chemical 

advanced oxidation: a review. Sci. Total Environ.. doi:10.1016/ 

j.scitotenv.2008.08.039. 


Metcalfe, C., Miao, X.S., Hua, W., Letcher, R., Servos, M., 2004. 

Pharmaceuticals in the Canadian environment. In: 

Kümmerer, K. (Ed.), Pharmaceuticals in the Environment, 

Sources, Fate Effects and Risks. Springer, pp. 67–90. 

Muñoz, I., Gómez, M.J., Molina-Díaz, A., Huijbregts, M.A.J., 

Fernández-Alba, A.R., García-Calvo, E., 2008. Ranking potential 

impacts of priority and emerging pollutants in urban 

wastewater through life cycle impact assessment. 


Muñoz, I., Rodríguez, A., Rosal, R., Fernández-Alba, A.R., 2009. Life 

cycle assessment of urban wastewater reuse with ozonation 

as tertiary treatment. A focus on toxicity-related impacts. Sci. 

Total Environ. 407, 1245–1256. 

Radjenovic, J., Petrovic, M., Barceló, D., 2009. Fate and distribution 

of pharmaceuticals in wastewater and sewage sludge of the 

conventional activated sludge (CAS) and advanced membrane 

bioreactor (MBR) treatment. Water Res. 43, 831–841. 

Raknes, K.L., 2005. Ozone in Drinking Water Treatment: Process 

Design, Operation, and Optimization. American Water Works 

Association, Denver. 

Real, F.J., Benitez, F.J., Acero, J.L., Sagasti, J.J.P., Casas, F., 2009. 

Kinetics of the chemical oxidation of the pharmaceuticals 

primidone, ketoprofen, and diatrizoate in ultrapure and 

natural waters. Ind. Eng. Chem. Res. 48, 3380–3388. 

Rischbieter, E., Stein, H., Shumpe, A., 2000. Ozone solubilities in 

water and aqueous salt solutions. J. Chem. Eng. Data 45, 

338–340. 

Rosal, R., Rodríguez, A., Perdigón-Melón, J.A., Mezcua, M., 

Hernando, M.D., Letón, P., García-Calvo, E., Agüera, A., 

Fernández-Alba, A.R., 2008a. Removal of pharmaceuticals and 

kinetics of mineralization by O 3/H 2O 2 in a biotreated 

municipal wastewater. Water Res. 42, 3719–3728. 

Rosal, R., Rodríguez, A., Perdigón-Melón, J.A., Petre, A., García- 

Calvo, E., 2008b. Oxidation of dissolved organic matter in the 

effluent of a sewage treatment plant by ozone combined with 

hydrogen peroxide (O 3/H 2O 2). Chem. Eng. J. 149, 311–318. 

Sánchez-Polo, M., Rivera-Utrilla, J., Prados-Joya, G., Ferro- 

Garcıa, M.A., Bautista-Toledo, I., 2008. Removal of 

pharmaceutical compounds, nitroimidazoles, from waters by 

using the ozone/carbon system. Water Res. 42, 4163–4171. 

Scheytt, T., Mersmann, P., Lindstädt, R., Heberer, T., 2005. 

1-Octanol/water partition coefficients of 5 pharmaceuticals 

from human medical care: carbamezepine, clofibric acid, 

diclofenac, ibuprofen and propyphenazone. Water Air Soil 

Pollut. 165, 3–11. 

Tauxe-Wuersch, A., De Alencastro, L.F., Grandjean, D., 

Tarradellas, J., 2005. Occurrence of several acidic drugs in 

sewage treatment plants in Switzerland and risk assessment. 

Water Res. 39, 1761–1772. 


treatment plants and rivers. Water Res. 32, 3245–3260. 

Ternes, T., Bonerz, M., Schmidt, T., 2001. Determination of 

neutral pharmaceuticals in wastewater and rivers by liquid 

chromatography–electrospray tandem mass spectrometry. 

J. Chromatogr. A 938, 175–185. 

Terzić, S., Senta, I., Ahel, M., Gros, M., Petrović, M., Barcelo, D., 

Müller, J., Knepper, T., Martí, I., Ventura, F., Jovančić, P., 

Jabučar, D., 2008. Occurrence and fate of emerging wastewater 

contaminants in Western Balkan Region. Sci. Total Environ. 

399, 66–77. 

Wilcox, J.D., Bahr, J.M., Hedman, C.J., Hemming, J.D.C., 

Barman, M.A.E., Bradbury, K.R., 2009. Removal of organic 

wastewater contaminants in septic systems using 

advanced treatment technologies. J. Environ. Qual. 38, 

149–156.

Degradation of the emerging contaminant ibuprofen 

in water by photo-Fenton 

Fabiola Méndez-Arriaga*, Santiago Esplugas, Jaime Giménez 

Departamento de Ingeniería Química, Facultad de Química, Universidad de Barcelona, C/ Martí i Franquès 1, 08028 Barcelona, Spain 

article info 




3 July 2009 



Keywords: 

AOP 

Photo-Fenton 

Non-steroidal anti-inflammatory 

drug ibuprofen 


abstract 

The recent public interest regarding the presence of pharmaceutical 

pollutants in water has raised important concerns due 

to the unknown environmental impact and possible damages to 

the botany and fauna present in aquatic systems (Halling-Sorensen 

et al., 1998; Ternes et al., 2004). One of the most consumed 

medications corresponds to the classification of the Non- 

Steroidal Anti-Inflammatory Drugs (NSAIDs) with more than 70 

million annual prescriptions in the world (Takagi et al., 2006). 

Moreover, several recent reports have confirmed the presence of 





E-mail address: fmendeza@ub.edu (F. Méndez-Arriaga). 


doi:10.1016/j.watres.2009.07.009 

In this study the degradation of the worldwide Non-Steroidal Anti-Inflammatory Drug 

(NSAID) ibuprofen (IBP) by photo-Fenton reaction by use of solar artificial irradiation was 

carried out. Non-photocatalytic experiments (complex formation, photolysis and UV/Vis- 

H2O2 oxidation) were executed to evaluate the isolated effects and additional differentiated 

degradation pathways of IBP. The solar photolysis cleavage of H2O2 generates hydroxylated-IBP 

byproducts without mineralization. Fenton reaction, however promotes hydroxylation 

with a 10% contamination in form of a mineralization. In contrast photo-Fenton in 

addition promotes the decarboxylation of IBP and its total depletion is observed. In absence 

of H2O2 a decrease of IBP was observed in the Fe(II)/UV–Vis process due to the complex 

formation between iron and the IBP-carboxylic moiety. The degradation pathway can be 

described as an interconnected and successive principal decarboxylation and hydroxylation 

steps. TOC depletion of 40% was observed in photo-Fenton degradation. The iron-IBP 

binding was the key-point of the decarboxylation pathway. Both decarboxylation and 

hydroxylation mechanisms, as individual or parallel process are responsible for IBP 

removal in Fenton and photo-Fenton systems. An increase in the biodegradability of the 

final effluent after photo-Fenton treatment was observed. Final BOD5 of 25 mg L 1 was 

reached in contrast to the initial BOD5 shown by the untreated IBP solution (BOD5 < 1mgL 1 ). 

The increase in the biodegradability of the photo-Fenton degradation byproducts 

opens the possibility for a complete remediation with a final post-biological treatment. 


the NSAID ibuprofen (IBP), in effluents of wastewater treatment 

plants (WTPs) (Alonso et al., 2006; Gómez et al., 2008). Concentrations 

of IBP in the environment are reported between 

10 ng L 1 and 169 mgL 1 (Santos et al., 2007). A significant 

removal rate of this drug is reached by biological oxidation, in 

some cases more than 70% (Carballa et al., 2006). In spite of, its 

apparently high biodegradation rate, the ecological risk remains 

still high, due to the main byproducts generated during the 

biological oxidation. These byproducts, hydroxyl-IBP and carboxy-IBP, 

have shown quite similar toxicological consequences 

in aquatic environment (Roberts and Thomas, 2006).

590 

On the other hand, recent research has shown promising 

results in the removal of pharmaceutical pollutants through 

the application of Advanced Oxidation Processes (AOPs). The 

AOPs are oxidative processes applied for treatment of 

contaminants in water, soils and air, based on the presence 

and reactivity of the hydroxyl radicals ( OH) which are 

generated in atmospheric or subcritical conditions of 

temperature and pressure with or without catalyst and/or 

reactive energy (electrochemical, UV–Vis or ultrasounds) 

(Méndez-Arriaga et al., 2009). The final common byproducts 

after AOPs have been reported to be easily degraded by biological 

oxidation (Torres et al., 2003). Although AOPs use 

different reacting systems (homogeneous or heterogeneous 

phases, irradiated or dark conditions, etc.), they are mainly 

characterized by the production of hydroxyl radicals with 

consecutive unselective attack on the organic material. 

Several AOPs have been employed to remove NSAIDs (as 

diclofenac, naproxen, IBP, ketoprofen, etc.) but special focus 

was placed on the heterogeneous photocatalyst with TiO2, 

photo-Fenton solar process, ozonation process, electrochemistry 

(Sirés Sadornil Ignacio, 2006) and more recently on 

ultrasound energy (Hartmann et al., 2008; Méndez-Arriaga 

et al., 2008b). Photo-Fenton reaction is one of the most widely 

used AOPs due to its environment friendly application of solar 

energy. The extremely complex chemistry of the oxidation 

system based on the reaction between iron and hydrogen 

peroxide, the called Fenton reaction, has become of great 

interest for environmental applications. 

Well documented information, in a critical compilation 

about the basis of Fenton chemistry as well as its applications 

to wastewater treatment, can be founded in Venkatadri and 

Peters (1993), Bigda (1995), Chen and Pignatello (1999), Tarr 

(2003), Burkitt (2003), Neyens and Baeyens (2003), Safarzadeh- 

Amiri et al. (1996) and more recently in a widely recommended 

critical review, written by Pignatello et al. (2006). Photo-Fenton, 

an attractive alternative for degradation of emerging 

contaminants, is based on the redox reaction between Fe(II) 

and H2O2 with the concomitant generation of OH and 

consecutive catalytic role of iron (Fe(II)/Fe(III)) and it is 

improved under UV–Vis irradiation. 

Even the Fenton and photo-Fenton reactions have been 

applied for wide types of organic contaminants, their use for 

the degradation of pharmaceutical compounds has been 

applied to a smaller extent than other AOPs like photocatalysis 

or O 3. 

The Fenton and photo-Fenton processes have been 

employed in the treatment of wastewater containing pharmaceutical 

pollutants such as azathriopine, asparginase, DCF, 

penicillin, a-methyl-phenylglycine, metronidazole, lincomicyn 

metabolites, sulfamethazine, sulfamethoxazole, amoxicillin, 

bezafibrate, paracetamol, biguanides, guanidine, 

triamines, pharmaceutical industry and hospital wastewaters 

among others (Klaviaroti et al., 2008; González et al., 2007; 

Pérez-Moya et al., 2007; Pérez-Estrada et al., 2005). 

Degradation of IBP by photo-Fenton reaction has not been 

reported yet. Therefore, the goals of this work were to evaluate 

the photo-Fenton ability to degrade IBP in an aqueous 

solution with a simulated solar irradiation source as well as to 

assess the biodegradability of the identified byproducts in the 

effluent after the treatment. 


2. Reagents, equipments and 

analytical procedures 

IBP (Sigma–Aldrich), FeSO4$7H2O (Panreac) and H2O2 (30% w/v) 

(Panreac) were used without pre-treatment. The experimental 

device has been already described in earlier communications 

(Méndez-Arriaga et al., 2008a). Initial IBP aqueous solution 

(0.87 mM, pH 6.25 0.25) was pumped from the feeding stirred 

tank to the Duran-photoreactor (0.08 L) placed inside the 

Solarbox (Co.Fo.Me.Gra, Italy). The Xe lamp (Phillips OP 1 kW, 

6.9 mEinsteins s 1 in 290–400 nm wavelength range), placed 

top-inside of the Solarbox, was switched on, and samples at 

different irradiation times were withdrawn. In all experiments, 

fresh H2O2 and Fe(II)-acid dilutions were daily 

prepared. For photo-Fenton experiments a peristaltic pump 

was used to dose solutions of H2O2 and H2SO4 in the entrance 

of the photoreactor. Several concentrations of H2O2 were 

chosen, from 0 to 0.32 mM (0.04, 0.08, 0.16 and 0.32 mM), with 

a dose rate of 0.032 mM min 1 . pH adjust was carried out with 

H 2SO 4 diluted solution reaching a final pH 3. IBP concentration 

was monitored by HPLC by using an RP-C 18 Trace Extrasil 

OD52-5 Micromet 25 0.46 Teknockroma column in isocratic 

mode with acetonitrile (99.8% Panreac) and acetic acid (0.25 M) 

in 75–25 proportion. Aliquots were also used to measure the 

residual concentration of H2O2 by spectrophotometric method 

(Nogueira et al., 2005) with a Perkin Elmer UV/VIS Lambda 20. 

Total Organic Carbon (TOC) measurements were obtained by 

a Shimadzu TOC-V CNS auto sampler instrument. BOD5 

determinations were obtained according to Standard Methods 

(5120) by respirometric single measuring system OxiTop 

procedure. A Liquid Chromatography–Electrospray Ionization 

Jasco AS-2050 plus IS mass spectrometer (LC/ESI-MS, TOF 

Mariner) was employed to identify several byproducts. 


3.1. Photolysis, H 2O 2 and UV–Vis/H 2O 2 

photolysis effects 

Non-photocatalytic assays were undertaken to evaluate their 

isolated influence on the degradation of IBP. For the photolysis 

experiment, the IBP solution (initial concentration of 0.87 mM) 

was irradiated throughout the photoreactor during 2 h. On the 

other hand, several ratios of IBP-H2O2 (ranging from 1:0.001 to 

1:10) were mixed in 30 mL vials at free pH conditions (pH 

6.25 0.25), constant controlled temperature (30 C) and 

continuous dark stirring during 24 h. Furthermore, H2O2 

photolysis process was carried out irradiating a solution of 

0.87 mM of IBP with 0.32 mM of H 2O 2 during 2 h. 

Photolysis let IBP and TOC concentrations unchanged. The 

negligible degradation attained by photolysis was not unexpected, 

due to the low IBP molar absorption coefficient above 

300 nm. Similar results were observed for all ratios IBP-H2O2 

after 24 h darkness. Any difference in the original concentration 

of IBP was observed and TOC conversion showed an 

insignificant change, thus addition of peroxide in the dark 

gave a quite stable IBP solution. In contrast, UV–Vis/H2O2 

process showed an evident decrease in the concentration of

IBP. When solution was irradiated in presence of H 2O 2, almost 

40% of degradation took place after 2 h. The degradation of IBP 

can be attributed to the photochemical cleavage of hydrogen 

peroxide to yield hydroxyl radicals by light absorption (Tuhkanen, 

2004): 

H 2O 2 þ hv / 2 OH (1) 

In the configuration of our system (Xe lamp and Duranglass 

photoreactor) there is an overlap between the emitted 

light and the irradiation yet available to be absorbed for H 2O 2 

approximately from 290 to 350 nm. 

Thus, the elimination of IBP has the unique contribution of 

the hydroxyl radical attack produced by H2O2 photolysis but 

not due to their separate elements (irradiation or H2O2). An 

important remark is the fact that TOC did not decrease. Thus, 

the composition of the remained organic solution after 2 h of 

UV–Vis/H 2O 2 process could be addressed either to the 

hydroxylated byproducts of IBP with weak acidic properties 

according to the pH evolution observed from 6.5 to 5.5. The 

hydroxylation process of IBP is a consequence of its scavenger 

ability in favor of OH. Several indications of the OH scavenger 

property in NSAID have been widely reported (Hamburger and 

McCay, 1990; Bilodeau et al., 1995; Aruoma and Halliwell, 1988; 

Halliwell et al., 1995; Patricò et al., 1999). 

3.2. Fenton dark and photo-Fenton reaction 

For Fenton dark experiments, several Fe(II) concentrations 

(0.15 mM, 0.29 mM; 0.60 mM and 1.2 mM) were selected. Iron 

acid solutions were added to 1 L IBP solution (0.87 mM, pH 

6.25 0.25) under stirring and controlled temperature (30 C). 

H 2O 2 (0.32 mM) was immediately added in the tank and 

samples were periodically withdrawn. On the other hand, 

photo-Fenton experiments were carried out with similar 

conditions in 1.5 L IBP solution as described in Section 2. Iron 

acid solution (1.2 mM) was added in the continuously flowed 

solution. Several concentrations of H2O2 were chosen, from 

0 to 0.32 mM, with a dose rate of 0.032 mM min 1 . pH adjust 

was carried out with H2SO4 diluted solution reaching a final 

pH 3. 

Fig. 1 depicts the IBP conversion after 2 h of Fenton dark 

reaction. As higher Fe(II) concentration, higher IBP degradation 

is reached. IBP depletion by Fenton involves the chemical 

reaction with H 2O 2 as free OH generator (Haber and Weiss, 

1934): 

Fe(II) þ H2O2 / Fe(III) þ OH þ OH (2) 

Fe(III) þ H2O2 / Fe(II) þ O2H þ H þ 

Iron acts as catalyst by the simultaneous species Fe(II)/ 

Fe(III) present during the reaction meanwhile H2O2 is continually 

consumed forming OH. The increase of Fe(II) promotes 

IBP decrease, in the maximum case ca. of 60% corresponding 

to 10% of TOC reduction. Early report (Kennedy et al., 1990) 

suggests that for ratios equal or higher to 0.13 of iron/IBP all 


(3) 

IBP/ IBPo 

1.00 

0.75 

0.50 

0.25 

0.00 

0 30 60 90 120 

Time (min) 

Fig. 1 – Effect of Fe(II) concentration on removal of IBP 

(0.87 mM) by Fenton dark reaction.(B) 0.15 mM 

Fe(II),(:)0.29 mM Fe(II); (,)0.60 mM Fe(II) and (A)1.2 mM 

Fe(II). In all cases 0.32 mM of H2O2. 

the coordination sites of the metal are occupied by IBP and the 

OH production is hampered. However, in this work under the 

ratios tested – between 0.12 and 0.33 – the catalyst has shown 

to be reactive with H2O2 promoting the degradation of IBP. 

On the other hand, in photo-Fenton experiments more 

than 80% removal of IBP and almost 40% decrease of the initial 

TOC were achieved by using 0.32 mM H2O2 and 1.2 mM Fe(II). 

These values are four folds higher than the observed under 

dark conditions. Fig. 2 shows the IBP/IBP o for photo-Fenton 

process together with photolysis, H 2O 2/UV–Vis, Fenton and 

Fe(II)/UV–Vis process, and Fig. 3 shows the concomitant TOC 

removal for all the above described processes. 

Fig. 2 shows an evident change in the kinetic degradation 

of IBP by means of photo-Fenton reaction in comparison with 

the dark Fenton series above described. The increase in the 

reactivity of the photo-Fenton reaction is attributed to the 

simultaneous OH generation from the H2O2/UV–Vis and 

photo-Fenton systems. The photo-assisted reaction (photo- 

Fenton) gives typically faster rates of degradation and 

mineralization than the thermal (‘‘dark’’) reaction. In photo- 

Fenton reaction the production of OH is significantly 

increased under illuminated conditions due to the stimulating 

photoreduction of Fe(III) to Fe(II): 

IBP/ IBPo 

1.00 

0.75 

Photolysis 

H 2 O 2 /UV-Vis 

0.50 

Fenton 

Fe(II)/UV-Vis 

0.25 

Photo-Fenton (1.2mM Fe(II)) 

0.00 

0 30 60 90 

0.04mM H2O2 0.08mM H2O2 0.16mM H2O2 0.32mM H 

120 

2O2 Irradiation Time (min) 

Fig. 2 – Removal of IBP (0.87 mM) by (-) Photolysis; (,) 

H 2O 2/UV–Vis (0.32 mM H 2O 2); (A) Fenton (1.2 mM Fe(II), 

0.32 mM H 2O 2); (>)Fe(II)/UV–Vis (1.2 mM Fe(II)); Photo- 

Fenton in all cases 1.2 mM Fe(II) and (:) 0.04 mM H 2O 2;(6) 

0.08 mM H2O2; (C) 0.16 mM H2O2; (B) 0.32 mM H2O2.

592 

% TOC removal 

50 

40 

30 

20 

10 

0 

Photo-Fenton 1.2mM Fe(II) 

Fig. 3 – TOC removal for H 2O 2/UV–Vis (0.32 mM H 2O 2); 

Fenton (1.2 mM Fe(II), 0.32 mM H 2O 2); Fe(II)/UV–Vis (1.2 mM 

Fe(II)); Photo-Fenton with 1.2 mM Fe(II) and 0.04 mM H 2O 2, 

0.08 mM H 2O 2, 0.16 mM H 2O 2, 0.32 mM H 2O 2. In all cases 

0.87 mM initial IBP concentration. 

Fe(III) þ H2O þ hv / Fe(II) þ H þ þ OH (4) 

Ferrous ion is continuously recycled by irradiation and 

therefore it is not depleted during the course of the oxidation, 

as stated by Zepp et al. (1992). Additionally Pignatello et al. 

(1999) have shown the formation of iron complexes, as 

(FeOH) 2þ – predominant in acidic conditions – which have 

adsorption bands in the UV–Vis range (between 180 and 

410 nm) (Sirés Sadornil Ignacio, 2006), and act as additional 

hydroxyl radical source in photo-Fenton reaction. 

Fe(II) þ H2O2 / Fe(OH) 2þ þ OH (5) 

Degradation of IBP was directly proportional to the amount 

of H2O2 employed. However, H2O2 was not in all cases full 

consumed. For example, almost 60% of 0.32 mM of H2O2 

remained unconsumed. The same proportion is observed by 

using 0.16 mM however a clear decrease tendency was 

observed between 90 and 120 min. In contrast 0.08 and 

0.04 mM of H 2O 2 were almost full consumed after 120 min. 

Fenton reaction shows a rapid degradation phase resulting 

from a burst of $OH by Eq. (2). However when H2O2 is in large 

excess the extent of this burst phase will depend on the Fe/ 

contaminate molar ratio because the ratio determines the 

OH/contaminant ratio in the burst phase (Pignatello et al., 

2006). 

Moreover due to the pH was adjusted into the first 30 min, 

an early parallel formation of Fe(III) could promote in strong 

manner the Fenton-like reaction (Eq. (3)) rather than Fenton 

reaction. Since reaction in Eq. (3) is so much slower than 

reaction in Eq. (2), the parallel formation of Fe(III) also results 

in a slower initial rate of IBP and similar slow consume of 

H2O2. 

In the dark series more than 90% of the H2O2 remained 

after 2 h of reaction. TOC depletion was enhanced under 

irradiated conditions in the photo-Fenton reaction. As noted 

in Fig. 3 the maximum TOC removal is observed around 40%. 

The elimination of organic carbon is also due to the formation 


of volatile compounds (see also Section 3.5). Thus IBP oxidation 

and TOC depletion depend on experimental conditions 

and ratios between reactants, in a rather complex way due to 

the large number of reactions involved. 

3.3. Iron effect: IBP as organic ligand of iron 

and its photolysis 

An additional important query in Fenton process is the evaluation 

of possible complex formation between the organic 

compound and the metallic catalyst. Normally in Fenton 

applications the iron-complex formation with organic ligands 

is employed in order to widen the pH range, to avoid precipitation 

of Fe(III)-oxyhydroxide and/or to increase the ironphotolysis 

process (Andreozzi et al., 2006). However, undesirable 

effects can be also observed if the active coordination 

sites in the metal are not able to generate hydroxyl radicals in 

the Fenton reaction. 

The experiment carried out with 1.2 mM Fe(II) without 

H 2O 2, under irradiated condition (> doted in Fig. 2), gave an 

unexpected 60% of IBP removal as shown in Fig. 2. The 

removal of IBP using Fe(II)/UV–Vis configuration (without 

H 2O 2) was in deep studied by additional experiments varying 

the iron concentration under irradiated and free pH 5 conditions. 

Fig. 4 shows the IBP evolution, for 0.012, 1.2 and 5 mM 

FeSO4.7H2O concentrations, during 2 h of irradiation. Elimination 

of IBP is directly proportional to the amount of the 

irradiated-iron employed. IBP binds iron and under UV–Vis 

irradiated conditions this iron–IBP complex is photoactive. At 

pH 5, a large fraction of iron is in the form of Fe(III). Coordination 

of IBP with Fe(II) is probably negligible and would not be 

photolabile in the wavelength range used. Therefore, the 

reaction was due to the Fe(III)-complex photolysis. 

Iron forms complexes with organic compounds, especially 

those acting as polydentate ligands like Fe(III)–carboxylate or 

Fe(III)–polycarboxylate complexes (Pignatello et al., 2006). In 

the case of NSAID – which are characterized for their 

carboxylic acid moiety – some authors have shown their 

capacity to bind iron to form coordination complex. For 

instance, Kennedy et al. (1990) suggest that the complex 

formation between IBP and iron can be reached through the 

carboxylic moiety. Under our experimental conditions the 

possible production of decarboxylated byproducts and acidic 

IBP (mM) 

1.00 

0.75 

0.50 

0.25 

0.00 

0 30 60 

Time (min) 

90 

Fig. 4 – Effect of Fe(III)/IBP coordinated complex photolysis 

on IBP degradation during 120 min of UV–Vis irradiation. 

Initial Fe(II) acid ion added (C) 0.012 mM; (:) 1.2 mM; (-) 

5 mM and 0.87 mM IBP initial concentration. 

120

Table 1 – Isolated and combined effects of Fenton and 

photo-Fenton reagents on the IBP conversion and TOC 

removal observed at 2 h. 

Process A þ B / C XIBP TOC 

removal 

Fe(III) IBP þ hv Decarboxylated byproducts 0.70 20% 

H2O2 þ IBP Stable solution without reaction 0.00 0% 

H2O2 þ IBP þ hv Hydoxylated bycompounds 0.30 2% 

Fe(II) þ H2O2 þ IBP Fe(III)-IBP coordinated complex 

þ hydroxylated byproducts 

0.50 10% 

Fe(II) þ H2O2 þ hv 

þ IBP 

Decarboxylated þ hydroxylated 

þ biodegradable compounds 

1.00 40% 

aliphatic compounds are related with the change of the pH 

during the reaction from initial 5 to final 3.5. Higher iron–IBP 

complex amount, higher depletion of IBP through the 

carboxylic moiety, in agreement with reports of oxidation in 

iron-complex with organic functional group RCOO (Pignatello 

et al., 2006). The ligand-to-metal-charge-transfer reduction of 

the metal center promotes decarboxylation of IBP, at least 

partially, and further degradation of decarboxylated bycompounds 

is also driven under whole photo-Fenton system. 

3.4. Comparison between photo-Fenton processes and 

its isolated or combined reagents 

The photo-Fenton degradation of IBP is strongly marked for 

the hydroxyl attack and decarboxylation process. In dark 


H 3 C 

CH 3 

CH3 CH3 CH3 OH 

CH3 CH3 CH3 H3C 4 

O 

HO 

5 

CH3 H3C 6 OH 

H3C 7 

O 

conditions, the catalytic reaction is less favorable and low TOC 

removal is observed. In Table 1 all defined stages of photo- 

Fenton process are described together with their simultaneous 

contributions of isolated and combined reagents. The 

degradation of IBP by iron irradiated process seems to be the 

principal way as initial mechanism of degradation even if 

H2O2 is not present (0.7 and 20% of IBP conversion and TOC 

removal respectively). In photo-Fenton process, the OH yield 

improves and simultaneous scavenges the isobutyl moiety of 

IBP. Thus the carboxylic moiety of IBP binds the iron while the 

isobutyl moiety scavenges the OH radical generated. 

3.5. Byproducts from photo-Fenton process 

IBP byproducts generated by photo-Fenton treatment (1.2 mM 

Fe(II), 0.32 mM H 2O 2, 2 h of irradiation) were identified by an 

LC/ESI-TOF-MS, into the m/z range of 65–1000 in negative 

ionization mode. 

Fig. 5 depicts the proposed reaction pathway with the 

corresponding intermediates identified. The remaining TOC 

after treatment of IBP (m/z ¼ 206, TR 28) consisted mostly of 

two principal types of byproducts: decarboxylated and 

hydroxylated metabolites of IBP. Hydroxy-IBP byproducts 

identified were the three structures 1 (m/z ¼ 222, TR 16), 2 (m/ 

z ¼ 222, TR 23) and 3 (m/z ¼ 222, TR 18). The dihydroxylated 

compound 10 (m/z ¼ 238, TR 19) was founded at low intensity. 

Compound 1 suffers either the cleavage of C 1–C 2 bond of the 

isobutyl moiety, forming 4 (m/z ¼ 164, TR 13) and isopropanol 

or ketone acetone; or the decarboxylation forming 5 

CH 3 

CH3 CH3 CH3 OH 

CH3 OH 

CH3 HO 

1 CH3 O 

H3C 2 OH 

O 

H3C 3 

CH 3 

IBP 

O 

OH 

H3C H3C 8 OH 

9 

O 

CH3 CH3 CH 3 

HO 

H3C 10 

CH 3 

CH 3 

O 

OH 

CH3 HO 

O 

O 

OH 

Fig. 5 – Byproducts identified and possible degradation pathway by photo-Fenton reaction (1.2 mM Fe(II), 0.32 mM H2O2, 2h 

of irradiation, 0.87 mM IBP initial concentration). ID Name, Molecular Mass [Fragments m/z] IBP Ibuprofen, 206 [205 L ; 161 L ]. 

1 2-HydroxyIBP, 222 [221 L ; 177 L ]. 2 1-HydroxyIBP-, 222 [221 L ; 177 L ]. 3 2-hydroxy-2-[4-(2methylpropyl)phenyl]propanoic 

acid, 222 [221 L ; 191 L ]. 4 (2RS)-2-(4-Methylphenyl)propanoic acid, 164 [163 L ; 133 L ]. 5 1-ethyl-4-(2-hydroxy)isobutylbenzene, 

178 [177 L , 149 L ]. 6 1-ethyl-4-(1-Hydroxy)isobutylbenzene, 176 [175 L ; 133 L ]. 7 1-[4-(2-Methylpropyl)phenyl]ethanone, 176 

[175 L ; 133 L ]. 8 4-(1-hydroxy-2-methylpropyl)acetophenone, 192 [191 L ; 177 L ]. 9 2-Methyl-1-phenylpropane, 134[133 L ; 

119 L ]. 10 2-hydroxy-2-[4-(2methylpropyl)phenyl]peroxic acid, 238[237 L ; 221 L ].

594 

(m/z ¼ 178, TR 15.5) and formic acid. In the case of 2, the 

evidence of its decarboxylation could be represented by the 

byproduct 6 (m/z ¼ 178, TR 15.5). However undifferentiated 

retention times by MS analysis with its analogous compound 

5 were obtained. 

In the case of 3, the degradation process follows two 

possible pathways by either decarboxylation and deprotonation 

to reach 7 (m/z ¼ 176, TR 21) and formic acid, and by 

dihydroxylation to obtain the byproduct 10. However 

byproduct 10 can be also formed by a non-free radical 

pathway involving a nucleophilic exchange of OH of the 

carboxylic acid with H 2O 2. 7 can be again hydroxylated in 

position 1 of the isobutyl to reach 8 (m/z ¼ 191, TR 25) and also 

can be decarboxylated to form 9 (m/z ¼ 133, TR 15) and 

acetaldehyde. 

An approximation of the main byproducts remained after 

photo-Fenton treatment could be assigned to compounds 

7 and 9 with 23 and 36% of the total intensities. Thus the 

degradation of IBP by photo-Fenton goes through the simultaneous 

cleavage of the carboxyl acid moiety to produce formic 

acid or acetaldehyde and scavenger of hydroxyl radicals. As 

described before the formation of high soluble bycompounds 

isopropanol or formic acid and acetone promotes stable 

dissolution. The presence of those soluble compounds after 2 h 

of reaction improves the biodegradability in the effluent 

treated. Initial BOD5 for 0.87 mM of untreated-IBP was

and effects of pharmaceutical substances in the environment - 


Hartmann, J., Bartels, P., Mau, U., Witter, M., Tumpling, W.V., 

Hofmann, J., Nietzschmann, E., 2008. Degradation of the drug 

diclofenac in water by sonolysis in presence of catalysts. 


Kennedy, T.P., Rao, N.V., Noah, W., Michael, J.R., Jafri, M.H., 

Gurtner, G.H., Hoidal, J.R., 1990. Ibuprofen prevents oxidant 

lung injury and in vitro lipid-peroxidation by chelating iron. 

J. Clin. Invest. 86, 1565–1573. 

Klavarioti, M., Mantzavinos, D., Kassinos, D., 2008. Removal of 

residual pharmaceuticals from aqueous systems by advanced 

oxydation processes. Review article. Environment 

International 35 (2), pp. 402–417. 

Méndez-Arriaga, F., Esplugas, S., Gimenez, J., 2008a. 

Photocatalytic degradation of non-steroidal antiinflammatory 

drugs with TiO 2 and simulated solar irradiation. 

Water Res. 42, 585–594. 

Méndez-Arriaga, F., Torres Palma, R., Pétrier, C., Esplugas, S., 

Gimenez, J., Pulgarin, C., 2008b. Ultrasonic treatment of water 

contaminated with ibuprofen. Water Res. 42, 4243–4248. 

Méndez-Arriaga F., March 2009. Advanced Oxidation Processes 

(Photocatalysis, Photo-Fenton and Sonolysis) for Removal of 

Pharmaceutical Pollutants in Water. Doctoral thesis, 

Barcelona University. 

Neyens, E., Baeyens, J., 2003. A review of the classic Fenton’s 

peroxidation as an advanced oxidation technique. J. Hazard. 

Mater. 98, 33–50. 

Nogueira, R.F.P., Oliveira, M.C., Paterlini, W.C., 2005. Simple and 

fast spectrophotometric determination of H2O2 in photo- 

Fenton reactions using metavanadate. Talanta 66, 86–91. 

Patricò, D., Pasin, M., Barry, O.P., Ghiselli, A., Sabatino, G.A., 

Iuliano, L., FiztGerald, G., Violi, F., 1999. Iron-dependent 

human platelet activation and hydroxyl radical formation: 

involvement of protein kinase C. Circulation 99, 3118–3124. 

Pérez-Estrada, L., Malato, S., Gernjak, W., Agüera, A., 

Thurman, M., Ferrer, I., Fernandez-Alba, A., 2005. Photo 

Fenton degradation of diclofenac: identification of main 

intermediates and degradation pathway. Environ. Sci. 

Technol. 39 (21), 8300–8306. 

Pérez-Moya, M., Graells, M., del Valle, L., Centelles, E., Mansilla, H. 

, 2007. Fenton and photo-Fenton degradation of 

2-chlorophenol: Multivariate analysis and toxicity monitoring. 

Catalysis Today, 124, 3-4, 30 2007, 163–171. 

Pignatello, J.J., Liu, D., Huston, P., 1999. Evidence for an additional 

oxidant in the photoassisted Fenton reaction. Environ. Sci. 

Technol. 33, 1832–1839. 


Pignatello, J.J., Oliveros, E., MacKay, A., 2006. Advanced oxidation 

processes for organic contaminant destruction based on the 

Fenton reaction and related chemistry. Crit. Rev. Environ. Sci. 

Technol. 36, 1–84. 

Roberts, P.H., Thomas, K.V., 2006. The occurrence of selected 

pharmaceuticals in wastewater effluent and surface 

waters of the lower Tyne catchment. Sci. Total Environ. 356, 

143–153. 

Safarzadeh-Amiri, A., Bolton, J.R., Cater, S.R., 1996. The use of iron 

in advanced oxidation technologies. J. Adv. Oxid. Technol. 1, 

18–26. 

Santos, J.L., Aparicio, I., Alonso, E., 2007. Occurrence and risk 

assessment of pharmaceutically active compounds in 

wastewater treatment plants. A case study: Seville city 

(Spain). Environ. Int. 33, 596–601. 

Sirés Sadornil Ignacio, 2006. Electrochemical Advanced Oxidation 

Processes for the Removal of the Drugs Paracetamol, Clofibric 

Acid and Chlorophene from Waters. Doctoral thesis, 

Universidad de Barcelona. 

Takagi, T., Ramachandran, C., Bermejo, M., Yamashita, S., Yu, L.X., 

Amidon, G.L., 2006. A provisional biopharmaceutical 

classification of the top 200 oral drug products in the United 

States, Great Britain, Spain, and Japan. Mol. Pharmacol. 3, 

631–643. 

Tarr, M.A., 2003. Chemical degradation methods for wastes and 

pollutants. Chemical Degradation Methods for Wastes and 

Pollutants. In: Environmental Science and Pollution Control 

Series, vol. 26 165–200. 

Ternes, T.A., Herrmann, N., Bonerz, M., Knacker, T., Siegrist, H., 

Joss, A., 2004. A rapid method to measure the solid-water 

distribution coefficient (K–d) for pharmaceuticals and musk 

fragrances in sewage sludge. Water Res. 38, 4075–4084. 

Torres, R., Sarria, V., Torres, W., Peringer, P., Pulgarin, C., 2003. 

Electrochemical treatment of industrial wastewater 

containing 5-amino-6-methyl-2-benzimidazolone: toward and 

electrochemical–biological coupling. Water Res. 37, 7–13. 

Tuhkanen, T., 2004. UV/H 2O 2 processes. In: Parsons, S. (Ed.), 

Advanced Oxidation Processes for Water and Wastewater 

Treatment. IWA, London UK, pp. 86–109. 

Venkatadri, R., Peters, R., 1993. Chemical oxidation technologies: 

ultraviolet light/hydrogen peroxide, Fenton’s reagent and 

titanium dioxide-assisted photocatalysis, hazard. Waste 

Hazard. Mater. 10, 107–149. 

Zepp, R.G., Faust, B.C., Hoigne, J., 1992. Hydroxyl radical 

formation in aqueous reactions (pH 3–8) of iron(II) with 

hydrogen-peroxide - the photo-Fenton reaction. Environ. Sci. 

Technol. 26, 313–319.

Polar pollutants in municipal wastewater and the water cycle: 

Occurrence and removal of benzotriazoles 

Thorsten Reemtsma a, *, Ulf Miehe b , Uwe Duennbier c , Martin Jekel b 

a 

Federal Institute for Risk Assessment, Chemicals Safety, Thielallee 88–92, 14195 Berlin, Germany 

b 

Technical University of Berlin, Department of Water Quality Control, Sekr KF 4, Strasse des 17 Juni 135, 10623 Berlin, Germany 

c 

Berliner Wasserbetriebe, Laboratory, Motardstrasse 35, 13629 Berlin, Germany 

article info 




9 July 2009 



Keywords: 

Municipal wastewater 

Wastewater 

Treatment 

Benzotriazole 

Bank filtration 

Ozonation 

Activated carbon 

Groundwater 

Corrosion inhibitor 

Rivers 

Rhine 

Elbe 


abstract 

Benzotriazoles, in which a 1,2,3-triazole ring is condensed to 

a benzene ring, are a class of high production volume 

chemicals (HPVC) used in a wide variety of applications. The 

basic benzotriazole components are 1H-benzotriazole (BTri, 

also abbreviated as BTZ, BTA or BTAH) and the 1H-methylbenzotriazoles 

(tolyltriazoles; TTri, available as technical 




1H-benzo-1,2,3-triazole (BTri) and its methylated analogues (tolyltriazole, TTri) are corrosion 

inhibitors used in many industrial applications, but also in households in dishwashing 

agents and in deicing fluids at airports and elsewhere. BTri and one of the TTri-isomers 

(4-TTri) are typical examples of polar and poorly degradable trace pollutants. Benzotriazole 

elimination in four wastewater treatment plants (WWTP) in Berlin ranged from 20 to 70% 

for 5-TTRi over 30 to 55% for BTri to insignificant for 4-TTri. WWTP effluent concentrations 

were in the range of 7–18 mg/L of BTri, 1–5 mg/L of 4-TTri and 0.8–1.2 mg/L of 5-TTri. BTri and 

4-TTri proved to be omnipresent in surface waters of the rivers Rhine and Elbe with 

concentrations increasing from

enzotriazoles may dissociate and occur as anions at slightly 

basic pH or in contact with metal cations. At neutral pH in 

aqueous solution benzotriazoles are expected to be non-ionic 

rather than cationic. 

Owing to their NH acidity, benzotriazoles show metal 

complexing properties. BTri and TTri prevent corrosion of 

metals and steels, especially of copper and brass (Hart et al., 

2004), by forming a thin complexing film on the metal surface 

(Chadwick and Hashemi, 1978). For this reason BTri or TTri are 

used in the metal finishing industry, in semiconductor 

industry (Hollingsworth et al., 2005), in milk processing (Verheyen 

et al., 2009), and added to various fluids that come in 

contact with metals, such as aircraft deicing and antiicing 

fluids (Gruden et al., 2001), cooling liquids, brake fluids, metalworking 

fluids, etc. BTri is also used for silver protection in 

dishwashing machines (Ort et al., 2005). The annual production 

of benzotriazoles was reported to be in the range of 

9000 tons/year worldwide (Reemtsma et al., 2006). Benzotriazole 

can also be employed in the developing process in the 

photographic industry (Bergthaller, 2002) and is a biocide 

included in Annex I of Council Directive 98/8/EC (European 

Commission, 2007). 

Structurally more complex hydroxiphenyl-benzotriazoles 

are found in personal care products, where they act as stabilizers 

for UVA filter components (De Orsi et al., 2006). In the 

European Union some benzotriazole derivatives are included 

in Annex VII, part I of Council Directive 76/768/EEC as registered 

cosmetic ingredients (European Commission, 2000). 

Structurally related and partly chlorinated benzotriazole 

derivatives are used as stabilizers in plastic material such as 

polyethylene terephthalate (PET) (Bentayeb et al., 2007). 

First reports on the occurrence of benzotriazoles in the 

environment were related to their use in open systems as 

aircraft deicing fluids at airports (Cancilla et al., 1998, 2003). 

A few years later it was recognized that BTri and TTri are 

much more widely distributed in surface waters (Reemtsma 

et al., 2006; Voutsa et al., 2006; Giger et al., 2006) in 0.1–1 mg/L 

concentrations and are regularly discharged with municipal 

wastewater even after state-of-the art biological treatment 

in mg/L concentrations (Weiss and Reemtsma, 2005; Voutsa 

et al., 2006; Weiss et al., 2006). Only one of these studies 

differentiated between the two TTri-isomers (Weiss et al., 

2006) by liquid-chromatography tandem mass spectrometry 

(LC-MS/MS) (Weiss and Reemtsma, 2005). With this method 

it could be shown that the two isomers behave quite differently 

in the environment. While 5-TTri was mineralized 

within 15 days in an aerobic biodegradation test with activated 

sludge, the 4-TTri isomer remained stable (Weiss et al., 

2006). The much higher stability of 4-TTri as compared to 

5-TTri has also been observed in wastewater treatment 

and in surface waters (Weiss and Reemtsma, 2005; Weiss 

et al., 2006). 

Unlike the N-substituted 1,2,4-triazoles, which are used as 

herbicides, the benzo-1,2,3-triazoles have only limited biological 

activity. EC50 values for the acute toxicity to Vibrio 

fisheri were in the range of 7–21 mg/L and NOAEC levels for 

invertebrates and vertebrates in the range of 10–100 mg/L 

(Pillard et al., 2001). BTri has been shown to exhibit antiestrogenic 

activity in vitro in a yeast assay but not in vivo to 

fish (Harris et al., 2007). 


Despite these favourable ecotoxicity data the widespread 

occurrence of benzotriazoles in the environment should be 

a reason to improve our knowledge on the environmental 

behaviour of these compounds. Inter alia, the fate of benzotriazoles 

in surface waters is not well studied, especially for 

the TTri-isomers. It could also be worthwhile to become aware 

of those processes that are suitable for the removal of benzotriazoles 

from wastewater as well as from raw waters used 

for drinking water production. 


2.1. Wastewater treatment plants 

Samples are taken from the four largest wastewater treatment 

plants (WWTPs) in Berlin with treated water flows between 

40,000 and 200,000 m 3 /day. All plants are equipped with 

primary sedimentation, conventional activated sludge (CAS) 

treatment with nutrient removal (N/P removal) and secondary 

clarification. The plants differ in the P removal process: 

WWTP-R and W use biological P-elimination (Bio-P), WWTP-S 

with Bio-P and chemical precipitation and WWTP-M with 

chemical precipitation only. Hydraulic retention time (HRT) in 

the plants ranged from 11 to 24 h for the biological stage and 

the sludge retention time (SRT) from 8 to 25 days. The influent 

to WWTP-R had an organic load of 300–900 mg/L chemical 

oxygen demand (COD), while it ranged from 700 to 1100 mg/L 

COD for the other three plants. 

Samples of the influent (raw sewage) and the effluent of the 

secondary clarifier were taken as 24 h mixed samples between 

June and December 2006. Sampling was performed only at dry 

weather flow, and the sampling of effluent was corrected for 

hydraulic retention time in each of the plants. 

2.1.1. Flocculation-rapid filtration (F-RF) 

A polishing of the effluent of WWTP-R was investigated by 

rapid filtration on two experimental dual-media filters (0.8 m 

sand/1 m anthracite layer) that treated 100–180 L/h. Particle 

sizes were 0.7–1.25 mm for sand and 1.4–2.5 mm for anthracite 

in Filter 1 and 1.1–1.6 mm for sand and 2.5–4.0 mm for 

anthracite in Filter 2. Filter velocities ranged from 6 to 10 m/h. 

Filter backwash was performed every 1–2 days. 

Before filtration the effluent was treated by flocculation 

with either Al 3þ (1 mg/L) or Fe 3þ (2 mg/L). During some phases 

10–20 mg/L of powdered activated carbon (PAC; Norit Super 

SAE) were added during flocculation (flocculation, PAC, rapid 

filtration: F-PAC-RF). 

2.2. Field sites and field samples 

2.2.1. Rivers Rhine and Elbe 

Single grab samples were taken with a bucket in January 2007 

at sites used for sampling for the environmental specimen 

bank of the Federal Environment Agency (UBA). 

2.2.2. River Havel 

Samples of the river Havel were taken at the surface water 

treatment plants Spandau (upstream, n ¼ 5) and Beelitzhof 

(downstream; n ¼ 8) from July to November 2006.

598 

2.2.3. Bank filtration 

Bank filtration was studied at two sites in Berlin (Germany), at 

Lake Tegel and Lake Wannsee/Havel. Details of the bank 

filtration, the hydrology of the sites and the transects installed 

have been published before (Gruenheid et al., 2005; Massmann 

et al., 2008). Samples were withdrawn from several observation 

wells and from the production well at a distance of about 

100 m from the lakeshore. Samples, taken in April 2008 from 

the Tegel site and in May 2008, from both sites, were analyzed 

for benzotriazoles. 

2.2.4. Surface water treatment plants 

Two surface water treatment plants (Tegel and Beelitzhof) 

were investigated. These plants are designed to reduce the 

total phosphorus load of the surface water (median of 265 mg/L 

for Tegel and 129 mg/L for Beelitzhof) to values between 5 and 

30 mg/L by coagulation/flocculation, sedimentation and dualmedia 

filtration. The influent of the Tegel plant consists of 

25–80% effluent of WWTP-S, diluted with surface water. The 

Tegel plant treats 130,000–300,000 m 3 /day, the final filters 

have a height of 1.8 m and an average filter velocity of 6.6 m/h. 

The Beelitzhof plant treats surface water of Lake Wannsee 

(part of the River Havel system) downstream of the City of 

Berlin. A volume of 12,000 m 3 /day is treated in this plant and 

filtration occurs on filters of 1.6 m height and at a filter velocity 

of 5 m/h. 

2.3. Lab experiment 

2.3.1. PAC sorption studies 

Batch experiments for isotherm calculation were performed 

in either ultrapure water or effluent of WWTP-R with start 

concentrations of 100 mg/L of BTri or TTri (technical mix consisting 

of 40% 4-TTri and 60% 5-TTri (Weiss and Reemtsma, 

2005). Between 1 and 25 mg/L of PAC (Norit Super SAE) were 

added from suspensions (0.1–1 g/L) stirred at 600 rpm. The 

bottles were continuously mixed on a horizontal shaker 

(120 rpm) for 24 h. All experiments were performed in duplicate. 

Stability of the compounds was recorded in a control 

experiment without PAC addition. Dissolved concentrations 

of BTri and TTri (sum of both isomers) were determined after 

membrane filtration (0.45 mm cellulose-nitrate membrane 

filters) by LC-MS/MS analysis. A Freundlich model was fitted to 

the data. 


2.4. Analysis of benzotriazoles 

All field samples were stored in glass bottles, transported to 

the lab in ice chests, filtered over 0.45 mm membrane filters 

and stored in a refrigerator at 4 C pending analysis, usually 

within the next 24 h. 

Benzotriazoles were analyzed by LC-MS/MS with an isocratic 

elution by methanolic eluents on a diphenyl column 

(Pursuit Diphenyl, 3 mm, 150 2 mm) that allowed separation 

of the two TTri-isomers. Details have been published before 

(Weiss and Reemtsma, 2005). Samples of surface water and 

WTTP influents and effluents were analyzed by direct injection 

(20 mL for surface water and WWTP effluent; 10 mL for 

WWTP influent) into the LC-MS/MS system, while groundwater 

samples were analyzed after enrichment of a volume of 

50 mL by solid phase extraction (SPE) on a Oasis HLB 200 mg 

cartridge. With this procedure an LOQ of 0.15 mg/L (surface 

water) and 0.25 mg/L (wastewater) was obtained for direct 

injection and 0.05 mg/L after SPE (groundwater) for each of the 

three analytes. 


3.1. Wastewater treatment plants 

Four WWTP in Berlin, treating municipal wastewater as a mix 

of household and industrial wastewater were investigated for 

the occurrence and removal of benzotriazoles over a 6 month 

period (Fig. 1, Table S1). Mean influent concentration for the 

four plants ranged from 17 to 44 mg/L for BTri and from 1.1 to 

4.9 mg/L for each of the TTri-isomers. BTri and TTri concentrations 

were both higher in WWTP-M as compared to the 

other three plants. This was not due to a generally more 

concentrated wastewater in this plant, because the BTri/DOC 

ratio of 0.6& was also twice as high as the value for the other 

three plants (0.2–0.3&). The source of this elevated input of 

benzotriazoles to this WWTP is unknown. 

These influent concentrations are comparable to those 

reported in previous studies: mean concentrations of 6.6 mg/L 

BTri and 0.7 mg/L for TTri were found in a monitoring of eight 

WWTP in Western Europe (Reemtsma et al., 2006). In a study 

carried out in Switzerland on 10 WWTP even higher influent 

concentrations of 15–73 mg/L for BTri and 1.1–5.6 mg/L for 

Fig. 1 – Concentrations of (a) BTri, (b) 5-TTri and (c) 4-TTri in the influent and effluent of the four largest WWTP in Berlin (R 

(n [ 5), S (n [ 8), W (n [ 9): 05/2006–04/2007; M (n [ 8): 10/2005–04/2007).

TTri-isomers were recorded (Voutsa et al., 2006). In a previous 

study on WWTP-R in Berlin 12.0 mg/L for BTri and 1.3 mg/L for 

5-TTri and 2.1 mg/L for 4-TTri were reported as the 2-year 

average (Weiss et al., 2006), which is at the lower end of 

concentrations determined in this study. 

Benzotriazoles in municipal wastewater may originate 

from household due to their use as corrosion inhibitors in 

dishwashing agents. This could be proven for a residential 

area by modeling the loads in its sewer system (Ort et al., 

2005). When wastewater of a larger urban area is collected in 

one sewer system, industrial effluents, e.g. from metal finishing 

or semiconductor industry, may also be of relevance. 

Milligram per litre concentrations of TTri have only recently 

been reported in wastewater from milk processing (Verheyen 

et al., 2009). In mixed sewer systems runoff from streets 

(traffic) may also contribute. Airports, which are a well studied 

point source of benzotriazole emission (Cancilla et al., 1998, 

2003), may also contribute to the load in municipal wastewater 

if their surface runoff is collected and directed to the 

municipal sewer system (Weiss et al., 2006). 

Of the three benzotriazoles, none was eliminated 

completely during wastewater treatment. Generally relative 

removal was best for 5-TTri, but highly variable between the 

plants; it ranged from 19 to 69%. BTri showed moderate 

removal (29–58%), whereas for 4-TTri only one plant showed 

a significant but still limited removal (34%) (Fig. 1, Table S1). 

Only moderate removal of BTri and TTri (sum of both isomers) 

has been reported before in a monitoring of eight WWTP in 

Western Europe (Reemtsma et al., 2006) and for a variety of 

WWTP in Switzerland (Voutsa et al., 2006). 

This different removal of the two TTri-isomers observed 

in this study is in agreement with lab experiments, which 

showed that 5-TTri can be mineralized by activated sludge 

while 4-TTri is persistent (Weiss et al., 2006). These differences 

are attributed to the position of the methyl substituent 

at the ring system. Previous studies reported that the aerobic 

biodegradability of positional isomers of bicyclic aromatic 

compounds depends upon the substitution pattern and that 

isomers with the ortho-substituent (as in 4-TTri) were not 

biodegradable, while the isomer with the meta-substituent (as 

in 5-TTri) were degraded (Weiss and Reemtsma, 2005). 

The different removal behaviour of the three benzotriazoles 

can be highlighted by calculating the concentration 

ratios of either BTri or 5-TTri to the most persistent 4-TTri. 

Both these ratios tend to decrease during biological treatment 

from 6–7 to 4–5 for BTri/4-TTri and from 0.7–0.9 to 0.3–0.8 for 

5-TTri/4-TTri (Table S1). These ratios are especially useful to 

differentiate between dilution (decrease in concentrations but 

no change in the concentration ratios) and biodegradation 

(decrease in concentrations and in concentration ratios). This 

underlines that an appropriate analytical method that separates 

the isomers is essential to adequately study the environmental 

fate of TTri (Weiss and Reemtsma, 2005). 

Owing to the incomplete removal BTri median values of 

7–18 mg/L were found for the effluents of the four WWTP. 

Single values up to 38 mg/L were recorded in the effluent of 

WWTP-M. In a Swiss study BTri effluent concentrations 

ranging from 10 to 100 mg/L have been found (Voutsa et al., 

2006). These are remarkably high concentrations for single 

compounds in WWTP effluents. In a study of eight European 


WWTP, in which 28 trace pollutants have been monitored, 

BTri ranked second in effluent concentration; it was outcompeted 

only by ethylenediamine tetraacetic acid (EDTA) 

with 100 mg/L (Reemtsma et al., 2006). 

As shown previously the elimination of 5-TTri and BTri in 

biological wastewater treatment can be improved to about 

60% by using membrane bioreactors. But even with this 

upgraded biological treatment system the 4-TTri concentration 

was not reduced (Weiss et al., 2006). 

As a measure to prioritize trace pollutants in WWTP effluents 

with regard to their putative concentration in further 

compartments of the water cycle a so-called water cycle 

spreading index (WCSI) has been proposed (Reemtsma et al., 

2006). The WCSI is the ratio of the effluent concentration and 

the degree of removal in a WWTP. In previous studies the order 

of the three benzotriazoles in terms of increasing WCSI was 

5-TTri (16 mg/L), BTri (22 mg/L) and 4-TTri (46 mg/L) (Reemtsma 

et al., 2006; Weiss et al., 2006). From the influent and effluent 

data of these four WWTP the order would be 5-TTri (1–13 mg/L), 

4-TTri (4–27 mg/L, 74 mg/L) and BTri (10–54 mg/L). Accordingly, 

BTri would be expected to be the benzotriazole occurring in 

highest concentration in surface waters and also in later 

compartment of a water cycle, closely followed by 4-TTri. From 

its WCSI (lower effluent concentration and better removal) 

5-TTri would be the least critical components. 

3.2. Polishing of WWTP effluents 

If the biological processes in WWTP are not suited to remove 

polar pollutants from the wastewater, physico-chemical posttreatment 

may be used to avoid discharge of these pollutants 

into surface waters and their spreading along partially closed 

water cycles. 

For benzotriazoles, ozonation has been shown to be 

a suitable post-treatment process. An ozone dose of 1 mg/mg 

DOC in wastewater was sufficient for complete removal of 

BTri and TTri (Weiss et al., 2006). No difference in the reactivity 

of the three benzotriazoles towards ozone could be 

determined. 

3.2.1. Flocculation, sand filtration and powdered 

activated carbon 

In the present study we investigated whether flocculation 

followed by rapid sand filtration, with and without addition of 

powdered activated carbon (PAC) can also be used for such 

a polishing step. 

A flocculation and rapid filtration (F-RF) unit (100–180 L/h) 

was operated for more than 2 years on site of the WWTP-R 

with different flocculants (Al 3þ 1 mg/L, Fe 3þ 2 mg/L) and 

different filter velocities (6–7.5 m/h). Based on experiences in 

the deironing step of water treatment plants it was hypothesized 

that the biomass establishing in the subsequent (bio-) 

filters may remove some of the trace pollutants. The flocculation 

with either Al 3þ or Fe 3þ had no effect on the concentration 

of any of the benzotriazoles. However, 5-TTri showed 

an average removal of 50% in the subsequent rapid filtration 

(RF) step (Table S1). During some periods the removal was 

almost complete (Fig. 2). Thus, biodegradation of 5-TTri on the 

filter column must have been comparatively fast, as the residence 

time in the filter was about 20 min, only. These data

600 

Fig. 2 – Mean concentration of BTri, 4-TTri and 5-TTri in the 

feed water and the effluent of a rapid filtration column 

(filled with 0.8 m sand and 1 m anthracite) after flocculation 

with 1–2 mg/L Fe 3D (F-RF) with and without addition of PAC 

(10–20 mg/L) The error bars indicate standard deviation of 

three samplings at independent days. 

confirm the much better biodegradability of 5-TTri as 

compared to BTri and 4-TTri. 

The whole process was much more effective when PAC 

was added during flocculation (F-PAC-RF) (Fig. 2, Table S1). 

With the addition of 20 mg/L of PAC BTri could be reduced by 

81% to filter effluent concentrations around 3 mg/L and 5-TTri 

as well as 4-TTri were removed to levels below the LOQ of 

0.25 mg/L. A PAC concentration of 10 mg/L proved insufficient 

in effluent polishing (Fig. 2). In batch experiments with WWTP 

effluent, however, 10 mg/L of PAC were able to reduce the BTri 

and TTri concentration by 40 and 80%, respectively, in 24 h. 

This indicates that the contact time in the rapid filtration 

process was too short to make full use of the sorption capacity 

of the PAC. 

3.3. Occurrence in surface water 

3.3.1. Rhine and Elbe 

The two major rivers of Germany and also important European 

rivers, Rhine (catchment area around 200,000 km 2 ) and Elbe 

(catchment area around 150,000 km 2 ), were analyzed for their 

benzotriazole concentrations. BTri was quantified (LOQ ¼ 

0.05 mg/L) in all but one sample with concentrations ranging 

from 0.09 to 0.68 mg/L and 4-TTri in six of the nine samples (0.05– 

0.46 mg/L; Table 1). Clear trends are seen along both rivers, such 

as the increasing concentration of BTri (from 0.13 to 0.35 mg/L 

over the 700 km in the Rhine and from

discharged directly into trenches and creeks. However, the 

concentration of benzotriazoles in such surface runoffs has 

not been analyzed, yet. 

3.3.2. River Havel 

Also, water of the river Havel that passes the City of Berlin 

from the north to the south was analyzed for a period of 

5 months, because it is an important component of the 

partially closed water cycle of the city (Gruenheid et al., 2005). 

While the upstream samples did not show measurable 

concentrations of the triazoles, the downstream samples 

contained an average 1.6 mg/L of BTri, 0.34 mg/L of 5-TTri and 

2.1 mg/L of 4-TTri (Table 1). Here 4-TTri is the most prominent 

benzotriazole, which is consistent with our view on the relative 

stability of these three compounds. 

These surface water concentrations in river Havel downstream 

of the city of Berlin are about a factor of 5–10 higher 

than those found in Rhine and Elbe (Table 1). This is due to the 

high portion of WWTP effluent in the rivers of Berlin with their 

low water flow. Increased surface water concentration of BTri 

with increasing wastewater portion has been found also for 

different Swiss rivers (Giger et al., 2006). 

The widespread occurrence of BTri and 4-TTri in surface 

waters may also have consequences for the quality of raw 

waters used for drinking water production. Along all three 

rivers, Rhine, Elbe and Havel, water is abstracted via bank 

filtration for drinking water production. 

3.4. Treatment of surface water 

3.4.1. Flocculation and rapid filtration 

At several locations in Berlin (Fig. S1) surface water is treated 

by flocculation with Fe 3þ , sedimentation and rapid filtration to 

remove dissolved phosphorous and to avoid eutrophication in 

critical areas. Because removal of some trace pollutants, such 

as para-toluenesulfonamide, has been recorded in these 

surface water treatment plants (Richter et al., 2008) we 

investigated whether benzotriazoles could also be removed. 

However, no significant removal was obtained for any of the 

three compounds in each plant. For BTri and 4-TTri, these 

results agreed to those obtained for effluent polishing by F-RF 

(see above; Table S1). For 5-TTri, a 50% removal had been 

obtained in the RF step of the effluent, presumably by 

biodegradation. The lack of removal in the corresponding step 

with surface water, at the same contact time of 20 min, may 

indicate, that either microorganisms from the activated 

sludge may be required for this degradation or that the 

concentration of 5-TTri in the surface water (0.3–0.5 mg/L) was 

too low to induce this biodegradation process. 

3.4.2. (River) bank filtration 

River bank filtration is an important source of raw water for 

drinking water production and the passage of the littoral and 

the underground is an important process to improve and 

ensure water quality. This is true on a global scale, for 

Germany (16% of bank filtration; Hiscock and Grischek, 2002) 

but especially for the City of Berlin that gains more than 50% of 

the drinking water for its 3.5 Mio inhabitants from bank 

filtrate (Gruenheid et al., 2005). 


The concentrations of the three benzotriazoles were 

determined along two bank filtration transects towards the 

production wells at two different sites (Lake Tegel and Lake 

Wannsee) in Berlin (Fig. S1) At both sites the distance of the 

production well from the lakeshore is about 100 m, corresponding 

to travel times of the water in the range of 

4–5 months (Gruenheid et al., 2005; Massmann et al., 2008). 

Surface water in Lake Tegel consists of 15–30% effluent of 

WWTP-S (Gruenheid et al., 2005). The water abstracted in the 

production well consists of approx. 75% lake water and 25% 

land-sided groundwater (inland aquifer), corresponding to 

a volumetric contribution of previous wastewater of 11–22% 

(Gruenheid et al., 2005). Redox conditions are not stable but 

fluctuate between anoxic and aerobic depending on water 

temperature and the mode of operation of production wells 

(Gruenheid et al., 2005). At the second site the hydrological 

situation is largely similar. 

For two reasons the concentration data (Fig. 3) are a bit 

scattered. (i) Not all observation wells sampled at the same 

depth and, thus, redox conditions are not identical. (ii) All 

samples were taken within 2 days while the water needs 

months to travel along the transects from the lakeshores to 

the production wells. Thus, not the same body of water was 

sampled along one transect. Long-term investigations at these 

two sites have shown that seasonal fluctuations can be 

recorded in the groundwater body (Massmann et al., 2008). 

Despite these difficulties both transects show an ongoing 

continuous decrease in BTri concentration from the lakeshore 

(distance 0) to the respective production well (Figs. 3a,d). At 

Lake Tegel the lakewater concentration of around 2 mg/L is 

diminished to 0.2 mg/L in the production well, which corresponds 

to a 90% concentration decrease due to dilution and 

removal (Fig. 3a). If one considers the dilution by land-sided 

groundwater with no BTri contamination, the removal of BTri 

during this underground passage would be 86%. At the bank 

filtration site at Lake Wannsee the BTri concentration was 

reduced by 75% to 0.3 mg/L in the production well (Fig. 3d). The 

concentration gradient of BTri at the Lake Tegel transect is 

quite parallel to the DOC gradient (Fig. 3a), which may suggest 

a cometabolic degradation of BTri. 

Removal of the least stable 5-TTri from surface water was 

obviously much faster and, likely, complete (Figs. 3c,f). This is 

well visible for the Wannsee site, where the first observation 

well was placed right in the colmation zone at the lakeshore. 

Here the 5-TTri concentration was already diminished to 

0.1 mg/L (Fig. 3f), while the concentrations of 4-TTri was only 

slightly reduced as compared to the lake water (Fig. 3e). 

However, also 4-TTri was removed significantly along both 

transects, but the concentration gradient was less steep 

(Figs. 3b,e). At Lake Wannsee, where the 4-TTri concentration 

in surface water tends to be higher than in Lake Tegel the 

concentration in the production well in 100 m distance 

remained at 0.13 mg/L (Fig. 3e). 

It is unclear by which mechanism 4-TTri was removed. 

For sandy aquifer material a retardation factor of 1.1 has 

been experimentally determined (Jia et al., 2007). As sandy 

material is predominant also at these bank filtration sites, 

a removal by sorption appears unlikely. Rather a very slow 

microbial degradation may occur. Slow biodegradation of 

trace concentrations of organic contaminants taking place in

602 

Fig. 3 – Concentration of (a, d) BTri, (b, e) 4-TTri and (c, f) 5-TTri in bank filtration transects at two sites in Berlin. (left) at Lake 

Tegel (04/2008 and 05/2008); DOC concentration also shown in (a); (right) at Lake Wannsee (04/2008). PW: production well. 

Shaded area below LOQ. 

time periods of months are, however, difficult to verify 

experimentally. 

Whatever process may be responsible, bank filtration was 

the most effective natural process for removal of benzotriazoles 

from waters. This may also be true for other poorly 

degradable polar substances. However, one has to consider 

the duration of this treatment process of several months. The 

concentration profiles of the transects also show that these 

long residence times in the subsurface are essential to remove 

poorly degradable polar pollutants such as BTri and 4-TTri, 

while for 5-TTri much shorter residence times would have 

been sufficient. 

These findings consistently show that neither BTri nor 

4-TTri are completely eliminated during bank filtration, even 

after a travel time of 4 months (120 days). 

3.5. Drinking water treatment 

Benzotriazoles are widespread in surface waters and 

‘‘natural’’ processes such as biological wastewater treatment, 


biologically active filters and bank filtration are not sufficient 

to remove them completely. Thus, BTri and 4-TTri may occur 

in raw waters used for drinking water production and technical 

treatment processes should be available to remove such 

polar pollutants during drinking water treatment. While the 

suitability of ozonation for the removal of BTri and TTri with 

moderate dosages of ozone has been shown before (Weiss 

et al., 2006; Legube et al., 1987), it was investigated here 

whether activated carbon treatment could also be used to 

remove trace concentrations of benzotriazoles. 

Batch experiments with deionized water were performed 

to record isotherms for BTri and TTri (as the sum of both 

isomers; Fig. 4). The Freundlich parameters K F ¼ 84 (mg/g)/ 

(mg/L) n for BTri and 187 (mg/g)/(mg/L) n for TTri and n ¼ 0.44 

for BTri and 0.52 for TTri indicate a moderate tendency to 

adsorb onto activated carbon. It is stronger for the methylated 

benzotriazoles (TTri) than for the non-substituted analogue 

(BTri). 

Based on these isotherm data one can assume that trace 

concentrations of TTri that may occur in raw waters are well

Fig. 4 – Isotherms for the adsorption of BTri and TTri (sum 

of both isomers) onto PAC (Norit Super SAE) from deionized 

water with a start concentration C 0 [ 100 mg/L (mean of 

two experiments). 

removable by GAC filtrations, as it has been found for the 

pharmaceutical diclofenac with a comparable K F values of 140 

(mg/g)/(mg/L) n in a previous study (Ternes et al., 2002). For BTri 

with its lower KF value, however, the suitability of GAC filtration 

is questionable. Rather breakthrough in GAC filtration 

may occur, as found for clofibric acid with a comparable KF 

value of 71 (mg/g)/(mg/L) n (Ternes et al., 2002). A final evaluation 

of the suitability of GAC for the removal of the triazoles 

would have to be performed site-specific, as the sorption 

tendency and the sorption kinetics in a certain water may be 

influenced by its matrix constituents and its pH. 

4. Conclusion 

Benzotriazoles are typical examples of polar and poorly 

degradable pollutants and they behave characteristically in 

partially closed water cycles. Biological treatment is insufficient 

for their complete removal from water, as shown for 

wastewater treatment, surface waters, bank filtration and 

biologically active filters. Owing to its wide use and biological 

resistance BTri appears to be one of the most prominent 

single trace pollutants in WWTP effluents (mg/L concentrations), 

and it is omnipresent also in surface waters. Concentrations 

of 4-TTri tend to increase along the large European 

river systems Rhine and Elbe, indicating the continuous 

discharge of this compound into the rivers and its high 

stability also in surface waters. Bank filtration with residence 

times of several months in the underground proved the most 

effective natural process for removing these polar and poorly 

degradable trace pollutants. 

But neither BTri nor 4-TTri were removed completely and 

both compounds may occur at reduced concentration in raw 

waters used for drinking water production. Fortunately, these 

aromatic compounds of high electron density can be removed 

by ozonation and, at least, TTri also by activated carbon 

filtration, either in wastewater effluent polishing or in 

drinking water treatment. Thus, an intrusion of BTri and TTri 

from wastewater into drinking water in partially closed water 


cycles with indirect potable reuse can be avoided by these 

technical treatment steps. 

Data presented here clearly show an increasing resistance 

to biodegradation from 5-TTri over BTri to 4-TTri. Thus, the 

widespread contamination of the water cycle with benzotriazoles 

could be avoided to a large extent if solely 5-TTri 

would be used instead of BTri and of the technical mixture of 

5-TTri and 4-TTri. 


Funding of this research by Berliner Wasserbetriebe (project 

‘‘Barrieren’’) is gratefully acknowledged. We are grateful to 

Jutta Jakobs for assistance in the laboratory and to Ina Engel 

for performing the activated carbon study. The river water 

samples were received from the Department of Hydrogeology 

of the Free University of Berlin. 

references 

Bentayeb, K., Batlle, R., Romero, J., Nerin, C., 2007. UPLC-MS as 

a powerful technique for screening the nonvolatile 

contaminants in recycled PET. Anal. Bioanal. Chem. 388, 

1031–1038. 

Bergthaller, P., 2002. Couplers in color photography – chemistry 

and function – Part 3. Imaging Sci. J. 50, 233–276. 

Cancilla, D.A., Martinez, J., Van Aggelen, G.C., 1998. Detection of 

aircraft deicing/antiicing fluids in perched water monitoring well 

at an international airport. Environ. Sci. Technol. 32, 3834–3835. 

Cancilla, D.A., Baird, J.C., Rosa, R., 2003. Detection of aircraft 

deicing additives in groundwater and soil samples from 

Fairchild Air Force base, a small to moderate user of deicing 

fluids. Bull. Environ. Contam. Toxicol. 70, 868–875. 

Chadwick, D., Hashemi, T., 1978. Adsorbed corrosion inhibitors 

studied by electron spectroscopy: benzotriazoles on copper 

and copper alloys. Corros. Sci. 18, 39–51. 

Corsi, S.R., Zitomer, D.H., Field, J.A., Cancilla, D.A., 2003. 

Nonylphenol ethoxylates and other additives in aircraft 

deicers, antiicers and waters receiving airport runoff. Environ. 

Sci. Technol. 37, 4031–4037. 

De Orsi, D., Gianni, G., Gogliardi, L., Porra, R., Berri, S., Bolasco, A., 

Carpani, I., Tonelli, D., 2006. Simple extraction and HPLC 

determination of UV-A and UV-B filters in sunscreen products. 

Chromatographia 64, 509–515. 

European Commission, 2000. Twenty-Fourth Commission 

Directive 2000/6/EC Adapting to Technical Progress Annexes 

II, III, VI and VII to Council Directive 76/768/EEC on the 

Approximation of the Laws of the Member States Relating to 

Cosmetic Products. 

European Commission, 2007. Commission Regulation (EC) No. 

1451/2007 of Directive 98/8/EC Concerning the Placing of 

Biocidal Products on the Market. 

Giger, W., Schaffner, C., Kohler, H.-P.E., 2006. Benzotriazole and 

tolyltriazole as aquatic contaminants. 1. Input and occurrence 

in rivers and lakes. Environ. Sci. Technol. 40, 7186–7192. 

Gruden, C.L., Dow, S.M., Hernandez, M.T., 2001. Fate and toxicity 

of aircraft deicing fluids additives through anaerobic 

digestion. Water Environ. Res. 73, 72–79. 

Gruenheid, S., Amy, G., Jekel, M., 2005. Removal of bulk dissolved 

organic carbon (DOC) and trace organic compounds by bank 

filtration and artificial recharge. Water Res. 39, 3219–3228.

604 

Harris, C.A., Routledge, E.J., Schaffner, C., Brian, J.V., Giger, W., 

Sumpter, J.P., 2007. Benzotriazole is antiestrogenic in vitro but 

not in vivo. Environ. Toxicol. Chem. 26, 2367–2372. 

Hart, D.S., Davis, L.C., Erickson, L.E., Callender, T.M., 2004. 

Sorption and partitioning of benzotriazole compounds. 

Microchem. J. 77, 9–17. 

Hiscock, K.M., Grischek, T., 2002. Attenuation of groundwater 

pollution by bank filtration. J. Hydrol. 266, 139–144. 

Hollingsworth, J., Sierra-Alvarez, R., Zhou, M., Ogden, K.L., Field, J.A., 

2005. Anaerobic biodegradability and methanogenic toxicity of 

key constituents in copper chemical mechanical planarization 

effluent of the semiconductor industry. Chemosphere 59, 

1219–1228. 

Jia, Y., Breedveld, G.D., Aagaard, P., 2007. Column studies on 

transport of deicing additive benzotriazole in a sandy aquifer 

and a zerovalent iron barrier. Chemosphere 69, 1409–1418. 

Legube, B., Guyon, S., Dore, M., 1987. Ozonation of aqueous 

solutions of nitrogen heterocyclic compounds: benzotriazoles, 

atrazine and amitrole. Ozone Sci. Eng. 9, 233–246. 

Loos, R., Gawlik, B.M., Locoro, G., Rimaviciute, E., Contini, S., 

Bidoglio, G., 2009. EU-wide survey of polar organic persistent 

pollutants in European river waters. Environ. Pollut. 157, 

561–568. 

Massmann, G., Sueltenfuss, J., Duennbier, U., Knappe, A., 

Taute, T., Pekdeger, A., 2008. Investigation of groundwater 

residence times during bank filtration in Berlin: a multi-tracer 

approach. Hydrol. Process 22, 788–801. 

Ort, C., Schaffner, C., Giger, W., Gujer, W., 2005. Modeling 

stochastic load variations in a sewer system. Water Sci. 

Technol. 52, 113–122. 


Pillard, D.A., Cornell, J.S., Dufresne, D.L., Hernandez, M.T., 2001. 

Toxicity of benzotriazole and benzotriazole derivatives to 

three aquatic species. Water Res. 35, 557–560. 

Reemtsma, T., Weiss, S., Mueller, J., Petrovic, M., Gonzalez, S., 

Barcelo, D., Ventura, F., Knepper, T.P., 2006. Polar pollutant 

entry into the water cycle by municipal wastewater: 

a European perspective. Environ. Sci. Technol. 40, 5451–5458. 

Richter, D., Massmann, G., Duennbier, U., 2008. Identification and 

significance of sulphonamides (p-TSA, o-TSA, BSA) in an 

urban water cycle (Berlin, Germany). Water Res. 42, 1368–1378. 

Ternes, T.A., Meisenheimer, M., McDowell, D., Sacher, F., 

Brauch, H.-J., Haist-Gulde, B., Preuss, G., Wilme, U., Zulei- 

Seibert, N., 2002. Removal of pharmaceuticals during drinking 

water treatment. Environ. Sci. Technol. 36, 3855–3863. 

Verheyen, V., Cruickshank, A., Wild, K., Heaven, M.W., McGee, R., 

Watkins, M., Nash, D., 2009. Soluble, semivolatile phenol and 

nitrogen compounds in milk-processing wastewaters. J. Dairy 

Sci. 92, 3484–3493. 

Voutsa, D., Hartmann, P., Schaffner, C., Giger, W., 2006. 

Benzotriazoles, alkylphenols and bisphenol A in municipal 

wastewaters and in the Glatt River, Switzerland. Environ. Sci. 

Pollut. Res. 13, 333–341. 

Weiss, S., Reemtsma, T., 2005. Determination of benzotriazoles 

corrosion inhibitors from aqueous environmental samples by 

liquid chromatography-electrospray ionization-tandem mass 

spectrometry. Anal. Chem. 77, 7415–7420. 

Weiss, S., Jakobs, J., Reemtsma, T., 2006. Discharge of three 

benzotriazole corrosion inhibitors with municipal wastewater 

and improvements by membrane bioreactor treatment and 

ozonation. Environ. Sci. Technol. 40, 7193–7199.

Determining the fraction of pharmaceutical residues 

in wastewater originating from a hospital 

Christoph Ort a, *, Michael G. Lawrence a , Julien Reungoat a , Geoff Eaglesham b , 

Steve Carter b , Jurg Keller a 

a 

The University of Queensland, Advanced Water Management Centre (AWMC), Brisbane, QLD 4072, Australia 

b 

Queensland Health Forensic and Scientific Services, Organics Laboratory, QLD 4108, Australia 

article info 




21 July 2009 



Keywords: 

Experimental design 

Audit data 

Quantification 

Prediction 

Loads 

Consumption 


1.1. Brief overview 

abstract 

Hospital wastewater (HWW) is normally discharged directly, 

without pre-treatment, to sewers. Despite mostly being only 





E-mail address: c.ort@awmc.uq.edu.au (C. Ort). 


doi:10.1016/j.watres.2009.08.002 

Pharmaceutical residues in water are frequently analysed and discussed in connection with 

sewage treatment, ecotoxicity and, natural and drinking water quality. Among different 

localities hospitals are suspected, or implied, to be a major and highly variable source of 

pharmaceuticals that substantially contribute to the total wastewater load. In this study, the 

contribution of pharmaceuticals from a hospital to a sewage treatment plant (STP) serving 

around 45,000 inhabitants was evaluated. Approximately 200 hospital beds result in a hospital 

bed density of 4.4 beds per 1000 inhabitants, which is a typical value for developed world 

countries. Prior to sampling, a sound systems analysis was performed, and a sophisticated 

continuous flow-proportional sampling regime was applied. Hence, overall experimental 

uncertainty was reduced to a minimum, and measurements provide clear evidence that, for 28 

of 59 investigated substances, over 85% of the pharmaceutical residue loads do not originate 

from the hospital when applying a conservative error estimation. Only for 2 substances, 

trimethoprim (18%) and roxithromycin (56%), was the maximum observed contribution of the 

hospital >15%. On average, the contribution of the hospital for the compounds detected in 

both, hospital effluent and sewage treatment plant influent was small and fairly constant. Five 

compounds were only detected in hospital wastewater, and 24 neither in the hospital wastewater 

nor in the total wastewater at the influent of the STP. For these compounds no experimental 

contribution could be calculated. For the compounds where audit data for both the 

national consumption and the specific hospital under investigation were available, a prediction 

of the fraction of pharmaceuticals originating from the hospital was performed. Three 

quarters of the compounds, classified with the existing audit data, were in the same ‘‘hospital 

contribution category’’ as determined by measurements. For most of the other compounds, 

plausible reasons could be identified to explain the observed deviations. 


a small fraction of the total wastewater volume in the influent 

of a sewage treatment plant (STP), HWW has gained increasing 

scientific and public attention in the last decade. This is, in part 

due to the observation and expectation that HWW is a source 

for undesirable constituents, such as (multi-)antibiotic-resistant 

bacteria (Baquero et al., 2008; Kummerer, 2004). In other

606 

publications, the emission from hospitals was estimated for 

antibiotics, anaesthetics, disinfectants, heavy metals, AOX 

(Adsorbable Organic Halogens), iodised X-ray contrast media 

and cytostatic agents (e.g. Kummerer, 2001). The latter were 

also investigated in detail by Lenz et al. (2007). Furthermore, 

a number of toxicity assays were performed (Boillot et al., 2008; 

Ferk et al., 2009; Hartmann et al., 1998). As a result, it has been 

suggested in some studies that pre-treatment of HWW prior to 

discharge into the sewers provides a reasonable solution 

(Gautam et al., 2007; Lenz et al., 2007; Pauwels and Verstraete, 

2006). However, this view is not unanimously supported. The 

separate treatment of HWW to reduce the development of 

resistant bacteria was questioned (Kummerer, 2009): the 

substantial amount of antibiotics used outside of hospitals (in 

Germany more than 75%) seems to be a plausible reason that 

resistant bacteria are also abundant in wastewater not 

receiving any HWW. Additionally, Boillot et al. (2008) found 

quantitatively far fewer microorganisms in the effluents of 

hospitals than in urban wastewaters which is consistent with 

other studies. With regard to pharmaceuticals, Lenz et al. 

(2007) report that 1) for some pharmaceuticals merely a small 

fraction of the amounts administered in the hospital were 

actually found in its effluent (i.e. 0.1–0.2% for doxorubicin, 0.5– 

4.5% for 5-fluorouracil and 27–34% for total platinum) and 2) 

a complete onsite wastewater treatment process is needed to 

significantly remove targeted pharmaceuticals. This includes 

full physical and biological treatment steps, not only advanced 

processes. Capturing all sources within a hospital (wards, 

laboratories) may be further complicated by the fact that 

different facilities discharge through different pipes to the 

common sewer. This particularly holds true for large existing 

hospital complexes. 

Therefore, local circumstances need to be considered and 

the contribution of an individual hospital needs to be assessed 

in relation to the total load in a STP catchment. To our 

knowledge, only a few publications explicitly quantify pharmaceutical 

residues (subsequently referred to as ‘pharmaceuticals’) 

excreted within hospitals compared to the total 

pharmaceutical load in the corresponding STP influents 

(Feldmann et al., 2008; Heberer and Feldmann, 2005; Thomas 

et al., 2007). However, these studies are limited to a small 

number of pharmaceuticals, or make an assumption on the 

water flow instead of measuring the wastewater flow onsite to 

determine actual loads. 

In view of the local situation in South East Queensland 

where it is proposed to recycle wastewater for indirect potable 

reuse, it is sensible to consider whether pre-treatment of 

HWW will provide a significant benefit. From two previous 

research papers relevant for the region of interest also dealing 

with pharmaceuticals the contribution of hospitals cannot be 

derived (Khan and Ongerth, 2004; Watkinson et al., 2009). 

Therefore, the goal of our study is to determine accurately 

the contribution of a hospital to the total pharmaceutical load 

found at the inlet of the corresponding STP by means of 

measurements. Additionally, this experimentally data 

obtained from a limited time period is then compared with 

readily available audit data. It shall be assessed whether the 

contribution of a hospital can be predicted reliably without 

any additional administrative effort, i.e. without extra surveys 

on the hospital wards for day-specific consumptions. If 


measurements matched with the prediction, the same kind 

(comprehensiveness and quality) of information can be used 

at other locations to make a prediction, a priori without 

laborious measurements. 

The focus of this research is on dissolved pollutants which 

cannot be eliminated in conventional wastewater treatment. 

Pollutants showing poor to moderate biological removal need 

to be transformed by chemical reactions (e.g. oxidation) or 

separated by physical processes (e.g. adsorption onto activated 

carbon). 

1.2. Systems analysis 

The prediction and experimental quantification of pharmaceutical 

mass fluxes in the wastewater of a specific STP catchment 

are laborious. A sound understanding of the whole system is 

required prior to setting up a predictive model, and performing 

a confirmative sampling campaign. This particularly holds true 

when attempting to attribute different fractions to a multitude 

of individual sources, for example if there are several hospitals 

and multiple smaller health care facilities in a catchment. Due to 

the lack of generally accessible consumption data at sufficiently 

high spatial and temporal resolution, models often provide only 

a prediction of an average load. Additionally, the latter is prone 

to uncertainty due to varying transformations of pharmaceuticals 

during human metabolism. 

While it would be ideal to have a list of all health care 

facilities with size, services provided and precise pharmaceutical 

consumption, just obtaining generally available 

consumption data is a tedious task in itself. The ‘‘institutional 

resolution’’ is often not sufficient without additional administrative 

effort, i.e. temporary surveys of the wards in the 

hospital(s) under investigation (Feldmann et al., 2008; Kummerer, 

2001). Furthermore, the (average) household pharmaceutical 

consumption needs to be estimated from national or 

state-wide sales and/or prescription data if regional data is not 

available. 

Moreover, collecting representative samples requires 

a thorough knowledge of the sewer layout and awareness of 

potentially highly variable concentrations and loads in the 

course of a day. Clearly, accurate chemical analysis of a nonrepresentative 

sample is not adequate to characterise a real 

full-scale system. 

1.3. Sampling issues 

Accurately quantifying pharmaceutical loads in hospital effluents 

or sewers close to any source (sub-catchments, households 

or industry) is a demanding undertaking. It requires a substantial 

experimental effort and is still prone to uncertainties. The 

latter are extremely hard to quantify if sampling is carried out 

with conventional (unsophisticated) devices, i.e. auto-samplers 

operated in a discrete sampling mode with (too) long time 

intervals, or grab samples. Rarely are fluctuations of concentrations 

and loads assessed in separate experiments at high 

temporal resolution prior to the ‘‘real’’ measuring campaigns. 

These pre-experiments are very expensive and may not provide 

the data to answer the actual research question. However, if the 

applied sampling protocol does not result in the collection of 

a representative sample, then the care taken in the following

processes of transport, storage, preparation and chemical 

analyses with a sophisticated method cannot make up for this 

deficiency (de Gruijter et al., 2006). Subsequent (even sophisticated) 

statistical analyses of non-representative samples are 

unreliable and the resulting conclusions will therefore be of 

limited value. In some cases, the large variation observed in 

previous studies may not be ‘‘true natural variation’’ but 

instead, may simply be an artefact caused by inadequate 

sampling (Ort et al., in preparation). 

Therefore, strong emphasis has been put on obtaining 

representative samples for this study. In Ort and Gujer (2006) 

a method was presented to estimate the required sampling 

frequency in order to not exceed a certain sampling error. In 

gravity sewers this results in fairly short time intervals if the 

substance of interest is contained in a small number of 

‘‘wastewater pulses’’ per day (e.g. toilet flushes containing 

a specific excreted pharmaceutically active compound). 

Sampling frequencies that are too low result in large sampling 

uncertainties, especially in the case of only a few patients per 

day (Weissbrodt et al., 2009). The often claimed problem of 

‘‘limited storage capacity in an auto-sampler’’ can be easily 

solved by replacing the glass bottles more than once per day. 

This may be more laborious, but it is a much better solution than 

using a time-proportional sampling mode, which does not take 

samples weightedaccording to the flow inthe sewer. In contrast, 

physical boundary conditions such as deep sewers resulting in 

long dead times for purging the sampling hose or limited access 

to pressurised sewers are more difficult to overcome. 


2.1. Sewage treatment plant and catchment 

characteristics 

A total of approximately 45,000 inhabitants in two geographically 

separated sub-catchments, Morayfield and Caboolture, 

are connected to the South Caboolture STP (subsequently only 

referred to as STP) which is operated with two sequencing 

batch reactors (SBRs). It treats a daily dry weather flow of 

approximately 10,000 m 3 . During long dry periods with high 

level water restrictions, this value can drop to 7500 m 3 day 1 . 

Morayfield is drained by gravity sewers and contributes 

two thirds of the total wastewater. It is only pumped once, at 

the STP itself. Caboolture makes up for one third of the total 

influent and is a largely pressurised sewer system with 

numerous pumping stations. At specific times of the day the 

flow is diverted at the influent of the STP and stored in two 

large buffer tanks (800 m 3 each) before being pumped to the 

SBRs. This combination of sewers and the complex influent 

layout of the STP results in very high hydraulic fluctuations 

(see Fig. 1). Hours with almost zero flow contrast with hours 

around 250–300 L s 1 and in between, the flow varies rapidly 

and significantly. During wet weather the relative flow variations 

are less significant due to higher base flow. 

2.2. Hospital characteristics 

Caboolture Public Hospital has 190 beds and offers all services 

of a modern regional hospital (listed in Table SI 1, see 


supporting information). A small private hospital providing 

mainly day surgery (only around 10 beds) and a small dental 

surgery also drain into the same sewer. The wastewater from 

the private hospital cannot be accessed separately. Other 

small health care facilities within this sewer catchment make 

consultations to out-patients, and therefore, the wastewater 

from these facilities are not expected to significantly add to 

the pharmaceutical load of the STP. The hospital bed density 

for the whole STP catchment is 4.4 beds per 1000 inhabitants. 

All HWW is collected in a sewage pumping station (SPS CT-51, 

subsequently referred to as SPS) before being pumped to the 

primary rising main. There is no residential wastewater 

contributing to this SPS and the hydraulic residence time in 

the main sewer to the STP is approximately 30 min to 1 h 

(hydraulic calculations provided by the Regional Council for 

the decisive time in the morning when samples at the SPS and 

the STP needed to be coordinated). The average daily volume 

during dry periods pumped at the SPS is approximately 75 m 3 

which is 1% of the total wastewater volume discharged to the 

STP. The occupancy of hospital beds in Caboolture during the 

sampling period was close to 100% which is representative for 

the year to date average. 

Unfortunately no comprehensive database exists with 

regard to other health care facilities in the catchment of the 

STP. Hence, an internet search was performed. Four aged care 

facilities with a total capacity of 443 beds were found (297 high 

care and 146 low care) with an unknown occupancy rate. 

Furthermore, a total of 14 addresses for doctors plus 12 

dentists were found. If mass fluxes at the influent of the STP 

were significantly higher than expected from average national 

consumption and hospital usage, further investigations of 

these facilities would be warranted. 

2.3. Sampling 

Continuous flow-proportional sampling modes were applied 

in this study to minimise sampling error. Continuously 

diverting a small flow-proportional side stream is conceptually 

the best solution to obtain representative samples for 

dissolved compounds. However, low velocities in the side 

stream prevent proper sampling of solids and long-term 

operation may lead to biofilm growth. Due to the limited time 

of sampling biofilm growth is not considered problematic in 

this instance. 

Sampling over consecutive days was preferred to the 

alternative option of collecting samples on single days 

distributed over a longer period. This drastically reduces the 

effect of unknown system behaviour: missing a ‘‘decisive’’ 

HWW packet at the STP is then limited to the first hour of the 

first day and the last hour of the last day. All other water 

packets are captured, although they might be attributed to the 

STP sample a day later. However, this would merely lead to 

higher variability of the hospital’s contribution and not to 

a non-quantifiable effect. 

2.3.1. Sampling protocol for Caboolture Hospital (SPS CT-51) 

The HWW is not easily accessible before it enters the SPS. 

Furthermore it would have been very difficult to set up an 

accurate flow meter to measure flow in a small open channel 

with intermittent, partially very low flows and to use this data

608 

Flow [L s -1 ] 

500 

400 

300 

200 

100 

0 

Dry weather flow (cv = 0.72) 

Wet weather flow (cv = 0.44) 

8am 10am 12pm 2pm 4pm 6pm 8pm 10pm 12am 2am 4am 6am 8am 

to control the speed of the sampling pump. Instead, plumbers 

from the Regional Council fitted a tap in the rising main of the 

SPS (see Fig. 2). The tap is upstream of the non-return valve 

before the HWW enters the primary rising main leading to the 

STP. Electricians from the Regional Council installed an 

actuator after the tap which only opens when the pump of the 

SPS empties the wet well. Water runs without a sampling 

pump due to the pressure in the rising main. Under normal 

operating conditions there are about 24 pumping cycles per 

day, triggered automatically based on to the water level in the 

SPS. While it was found that the flow during one cycle is fairly 

constant, it can vary significantly among cycles due to variable 

hydraulic conditions in the primary rising main. Therefore, 

a manual operating mode was adopted, disabling the auto 

level control. This allowed for using the full storage capacity of 

the wet well. Starting at 7 AM it was emptied again at 12 PM, 6 

PM and 7 AM the following day which required personnel to be 

present three times per day (confined space). The pump 

Time [h] 

Fig. 1 – Two examples for typical flow patterns at the influent of the sewage treatment plant; cv [ coefficient of variation 

(standard deviation/mean). 

wet well 

SPS CT-51 


pump 

control 

unit 

operates at about 2500 L min 1 and the sampling side stream 

was adjusted with the tap to approximately 1 L min 1 , 

resulting in a sampling volume of about 10 L per pump cycle. 

In comparison, the dead volume of the tap installation 

including hose was 0.5 L (ca. 5% of the sampling volume). 

The three samples were collected in separate glass bottles, 

and analysed separately. The concentrations of the individual 

samples were multiplied with the flow for the corresponding 

pump cycle, and summed to obtain a 24-h load. Rough diurnal 

variations could also be determined with this sampling 

procedure, but they are not relevant for the system and time 

scales under investigation, and hence they are not further 

discussed in this paper. 

2.3.2. Sampling protocol at the sewage treatment plant 

To sample for the same ‘‘water packets’’ as at the SPS, sampling 

started at 7:45 AM in the influent of the STP. The storage tanks 

start filling at 8 AM and are emptied completely during night 

dry pit 

hose 

actuator 

tap 

non-return 

valve 

effluent Caboolture Hospital 

sample 

bottle 

to STP 

Fig. 2 – Schematic drawing of the sampling point at the sewage pumping station (SPS) CT-51 (not to scale): All hospital 

wastewater is discharged to the wet well of the SPS and intermittently pumped to the primary rising main leading to the 

sewage treatment plant (STP). Upstream of the non-return valve a stand pipe with a tap and an actuator was fitted. This 

allows for taking flow-proportional samples during individual pump cycles. 

primary rising main

time, and in the early morning hours. This guarantees that 

wastewater is not stored and dragged on over different 24-h 

sampling periods. Flow rates in the influent are routinely 

measured at high temporal resolution. A wire connected to an 

analogue digital converter provides a 4–20 mA signal from the 

PLC (programmable logic controller) linear to the flow in the 

sewer to control the speed of the sampling pump. The peristaltic 

pump (Watson Marlow 520UN, programmable interface, water 

proof casing, equipped with a 520R2 pump head and 3.2 mm 

tube bore) was tested in the lab to ensure its linear behaviour 

over the full speed range under similar physical boundary 

conditions (suction height approximately 2 m, pressure height 

negligible). The pump speed was set to 0 rpm (revolutions per 

minute) for 0 L s 1 in the sewer (pumping 0 mL min 1 )andto 

34 rpm for 1000 L s 1 (pumping 69.4 mL min 1 ). The finest 

increment of the pump is 0.1 rpm equivalent to 2.9 L s 1 

wastewater flow in the influent of the STP. With this setup 

approximately 15 L of wastewater were collected in a 20 L glass 

bottle which was located in a refrigerated container. Two field 

blanks were collected: to this end 0.5 L of MilliQ water was used 

to rinse the sampling tube and subsequently 0.5 L MilliQ water 

was pumped through the tube to be analysed in the laboratory. 

No substances were detected above the limit of quantification. 

2.4. Chemical analyses 

After collection, the continuously refrigerated samples were 

transported to the laboratory where they were filtered the same 

day and preserved before analysis. All samples were analysed 

for 59 substances by Queensland Health Forensic and Scientific 

Services (QHFSS). A detailed description of the method consisting 

of solid phase extraction followed by concentration prior 

to quantification by LC–MS/MS (liquid chromatography coupled 

with tandem mass spectrometry) is given in the supplementary 

information SI 2, accompanied with an alphabetical list of all 

compounds (Tables SI 2.2 and SI 2.3). 

As the method does not allow for correction of absolute 

analytical extraction recoveries in raw wastewater samples, 

we report relative loads. In order to compare hospital effluent 

samples with samples from the influent of the STP, it is 

necessary to assume that matrix effects between these 

sample types are similar. Any systematic error in recovery is 

therefore cancelled out when calculating ratios of loads, i.e. 

contribution of HWW to the total influent of the STP. 

2.5. Uncertainty assessment 

Flows in completely filled pressurised pipes can be measured 

more accurately than flows in open water channels (gravity 

flow). For this study a maximum error of 10% was assumed, 

which equals to 6% (¼10/3 0.5 ) as single standard deviation of 

a normal distribution. For chemical analysis a random uncertainty 

(reproducibility) of 20% for all compounds was chosen 

(see Tables SI 2.1–2.3). The two errors are independent, and 

Gaussian error propagation results in an overall uncertainty 

estimate for calculated loads of 21% (¼[6 2 þ 20 2 ] 0.5 ). 

The flow-proportional continuous sampling procedure 

covers all fluctuations in the wastewater over time. Since it is 

a reasonable assumption that dissolved compounds are 

completely mixed over the whole pipe cross section in the 


influent works, no additional errors need to be taken into 

account due to sampling. 

2.6. Pharmaceutical audit data 

2.6.1. National consumption 

An extract from the DUSC database (Drug Utilisation Sub- 

Committee) for the year 2008 is listed for the compounds 

investigated in this study (see supporting information, Table SI 

3). It comprises the sum of subsidised drugs (subsidised under 

the Pharmaceutical Benefits Scheme (PBS) and the Repatriation 

Pharmaceutical Benefits Scheme (RPBS) and processed by 

Medicare Australia) and non-subsidised drugs (under PBS copayment 

and private prescriptions). The amounts of non-subsidised 

drugs were estimated from continuous data on all 

prescriptions dispensed from a validated sample of 370 

community based pharmacies. The available data do not 

include drugs dispensed to public hospital in-patients, pharmacy 

over-the-counter drugs (i.e. non-prescription) and drugs 

supplied by supermarkets. 

2.6.2. Amounts administered to in-patients in Caboolture 

Public Hospital 

No specific survey was carried out during the sampling period on 

the wards. Routinely stored audit data for a current 12-month 

period (2007–2008) was made available by the pharmacy of the 

Caboolture Public Hospital. For each pharmaceutical, a specific 

database query was performed to derive the amounts exclusively 

used for hospitalised in-patients; pharmaceuticals given 

to out-patients (in consultations and pharmacy) were not 

considered, since they will be taken and excreted at home. The 

total annual hospital load was determined after summing the 

contributions of all medications containing the pharmaceutically 

active compound of interest. 

It has to be noted that the amounts derived from this database 

are amounts supplied by the pharmacy to the individual 

wards and not the amounts effectively administered. However, 

it is generally not the hospital’s policy to discard drugs to the 

(solid or liquid) waste system, both from a financial and environmental 

point of view. Nevertheless, some unused drugs for 

in-patients may be collected on the wards and returned to the 

pharmacy for reuse or proper disposal. Hence, these drugs do 

not contribute to the load in the HWW. However, in discussion 

with relevant hospital staff these amounts are considered to be 

very limited and are not assessed within this study. 


3.1. Evaluation of wastewater volumes 

The four consecutive weekdays, mid-February 2009 when 

sampling took place, were during a wet period, with flows 1.5– 

2 times higher than normal dry weather flow (i.e. surface 

runoff in catchments and infiltration to sewage pumping 

stations). In Table 1 the flows at the two sampling locations 

over the corresponding 24-h periods are summarised. During 

the sampling period, the hospital contributed less than 1% of 

the total wastewater flow to the STP.

610 

Table 1 – Wastewater volumes over 24 h at the SPS CT-51 (hospital wastewater) and the influent to the STP. 

Influent STP [m 3 ] Hospital wastewater 

(flow at SPS) 

7:45 AM – 7:45 AM of the following day 7 AM–7 AM of the following 

day [m 3 ] 

3.2. Evaluation of relative pharmaceutical loads 

To obtain relative pharmaceutical loads, measured concentrations 

were multiplied with the corresponding 24-h flow at 

each sampling location and normalised by the highest STP 

influent load. Four examples representing four different 

groups of pharmaceuticals are charted in Fig. 3. Absolute 

concentration values are not reported because they are difficult 

to compare among different studies; they highly depend 

on the sewer system (separate or combined) and on the 

hydraulic conditions (dry or wet weather flow). The key figures 

chosen for statistical evaluation are presented in Table 2, and 

discussed subsequently in detail for one example (atenolol, 

a beta-blocker, see also Fig. 3A). 

The numbers in black circles ( ) refer to the corresponding 

column in Table 2: 

Concentration values for atenolol in the influent of the 

STP were, on average, 10 times higher than the limit of 

quantification (LOQ). 

Fraction of influent STP [%] 

Day 1 16/2/09 14,064 109 0.8 

Day 2 17/2/09 16,921 129 0.8 

Day 3 18/2/09 19,059 138 0.7 

Day 4 19/2/09 14,347 127 0.9 

A 

Loads (normalised to 

max. STP influent load) 

Atenolol 

1-3% 

1-2% 

2-4% 

1-2% 

fraction of STP influent 

C Paracetamol 

D 



1.2 

1.0 

0.8 

0.6 

0.4 

0.2 

0.0 

1.2 

1.0 

0.8 

0.6 

0.4 

0.2 

0.0 

day 1 

day 1 

day 2 

day 3 

STP influent 

day 2 

day 3 

STP influent 

day 4 

day 4 

day 1 

day 1 


Hospital effluent 

day 2 

day 3 

day 4 


day 2 

day 3 

day 4 

3-7% 

3-7% 

3-8% 

4-10% 


B 

The concentrations in the hospital effluent were on 

average 2 times higher than in the STP influent. 

The STP influent loads show only little day-to-day variation 

(cv ¼ 0.06, cv ¼ coefficient of variation ¼ standard 

deviation/mean). Day-to-day variation is smaller than 

the estimated overall uncertainty. 

The loads in the hospital effluent varied more (cv ¼ 0.27). 

On average the hospital contributed only 1.8% to the 

total atenolol load in the influent of the STP. 

For a conservative error estimation, a maximum 

contribution of the hospital was calculated by dividing 

the upper uncertainty value of the hospital effluent by 

the lower uncertainty value of the STP influent for each 

day (see Fig. 3). Over all four days, the highest maximum 

contribution for atenolol was 3.5%. 

Over all four days, the smallest minimum contribution 

for atenolol was 0.9% (analogue procedure as in ). 

The prediction for an average contribution of the 

hospital based on audit data is 0.6% (see more details in 

Section 3.4). 

0.8 

0.6 

0.4 

0.2 

0.0 

Gabapentin 

1.20 1.2 

1.20 


0.03 

1.0 

0.04 

0.02 

0.01 

0.00 

Trimethoprim 

1-2% 

1-2% 

1-3% 

3-7% 


1.20 1.2 

1.20 


0.06 

0.15 

1.0 

0.04 

0.02 

0.00 





0.8 

0.6 

0.4 

0.2 

0.0 

day 1 

day 1 

day 2 

day 2 

day 3 

STP influent 

day 3 

STP influent 

day 4 

day 4 

day 1 

day 1 

day 2 

day 2 

day 3 

day 3 

day 4 

day 4 

8-18% 

6-15% 

6-13% 

7-16% 


Fig. 3 – Measured, relative pharmaceutical loads over 24-h periods in the influent of the STP and effluent of the hospital for 

four consecutive weekdays. Error bars include uncertainty of flow measurements (±6%) and chemical analysis (±20%), 

resulting in an overall uncertainty of ±21% (single standard deviation). Note the different scales for the y-axis of STP influent 

and hospital effluent. 

0.03 

0.02 

0.01 

0.00 

0.10 

0.05 

0.00

Table 2 – Classification of substances according to the contribution of the hospital to the total load in the influent of the STP 

(see Section 3.3 for more explanations of key figures marked with black circles ). LOQs for all compounds are between 0.1 

and 2 mgL L1 . 

Classification Substance Therapeutic 

group f 

CSTP 

LOQ 

C Hospital 

CSTP 

Coefficient of variation 

for loads 

Influent 

STP 

Contribution of hospital 

wastewater [% of total STP influent] 

Measured average 

predicted 

with audit 

data d 

Hospital Min Mean Max 

Max 5% Atenolol BB 10.7 2.0 0.06 0.27 0.9 1.8 3.5 0.6 

Atorvastatin HL

612 

Table 3 – Comparison with other hospital wastewater studies. 

Number of hospitals Number of beds 

per 1000 inhabitants 

Classification of all substances according to maximum 

observed contribution from the hospital ( ). 

The consistent results for atenolol are reflected across 

most of the 30 detected substances. Representatives of other 

pharmaceutical groups show also fairly constant loads over 

the four-day period: gabapentin (an anticonvulsant), paracetamol 

(an analgesic) and trimethoprim (an antibiotic, see 

Fig. 3B–D). 

From the 59 substances, 5 were detected only in the HWW 

but not in the influent of the STP and 24 substances were not 

detected above the LOQ in any of the samples. The 30 

substances detected at both locations were classified for the 

hospital’s contribution to the total influent of the STP. To this 

end, the maximum observed contribution including uncertainty 

as a conservative estimate was used (see description 

before in ). The hospital’s contribution for 17 substances was 

at all times ‘‘smaller than 5%’’, 11 additional substances fall in 

the category ‘‘smaller than 15%’’ and only 2 substances were 

‘‘above 15%’’ (trimethoprim and roxithromycin with a worst 

case estimate of 18% and 56% respectively). For most 

substances quantified in both STP influent and hospital 

effluent, the variations of the loads in the HWW were on 

average 2.4 times higher than in the influent to the STP. The 

small number of hospital patients compared to the potentially 

large number of individuals taking these pharmaceuticals at 

home is a valid explanation for this observed difference in 

variation. 


Investigated substances 

(% in influent of the corresponding 

STP originating from hospitals) 

2 4.4 diclofenac (1.4) a 

ibuprofen (0.7) a 

metoprolol (1.5) a 

paracetamol (12) a 

tetracycline (0.5) a 

trimethoprim (14) a 

1 3.6 5 X-ray contrast 

media (50) b 

cytostatics (max 7.5) b 

More than 5 12.1 carbamazepine (15) c 

diclofenac (10) c 

More than 5 12.1 metamizol (50) c 

Four out of the 5 substances only detected in the HWW 

were just above the LOQ. With the 100 fold dilution in the 

influent of the STP the LOQ would have to be at least three 

orders of magnitude lower to reliably quantify the hospital’s 

(high) contribution. When assuming that the concentrations 

in the influent of the STP were equivalent to the corresponding 

LOQ, only a one-sided estimation with regard to the hospital’s 

contributions from >5% up to >50% can be made. 

However, in some cases this deviates from the prediction 

based on audit data (see chapter 3.3). 

3.3. Comparison with audit data 

Location of study 

Oslo, Norway (Thomas et al., 2007) 

Winterthur, Switzerland (Weissbrodt et al., 2009) 

Berlin, Germany (Heberer and Feldmann, 2005) 

Berlin, Germany (Feldmann et al., 2008) 

a Concentrations measured over 12 weeks, loads estimated with water consumption, sum of the two major hospitals (an unknown number of 

other smaller hospitals/health care facilities are located in the catchment). 

b Influent STP was not measured in this study, percentage refers to loads of pharmaceuticals quantified in the hospital’s effluent compared to 

day-specific administered amounts. 

c Only measured in the effluent of one hospital and then extrapolated for the whole catchment based on audit data of the other hospitals. 

If the consumption of pharmaceuticals in a STP catchment 

can be estimated from existing national sales or prescription 

data, and audit data for the hospital are available, the 

contribution of the hospital can be calculated with the 

following equation 

ConsCab:Hosp:$excretion ratio 

contributionðhospitalÞ ¼ 

ConsCab:Pop:$excretion ratio þ ConsCab:Hosp:$excretion ratio 

measured loadðhospitalÞ$recovery$accuracy 

y 

measured loadðSTP catchmentÞ$recovery$accuracy with ConsCab:Pop: ¼ ConsAUS 

$45; 000 ð1Þ 

20; 000; 000 

where Cons is the consumption, Cab stands for Caboolture, 

Pop. for population, AUS for Australia and Hosp. for hospital. It 

becomes evident that the transformation due to human 

metabolism (excretion ratio) cancels out of the equation when 

assumed to be similar for patients in the hospital and for 

people at home. The consumption of pharmaceuticals in the 

STP catchment is estimated by calculating an average per 

capita consumption from the national consumption data 

multiplied with the number of inhabitants in the catchment. 

The consumption of in-patients in the hospital is added to the

domestic consumption to obtain an estimate for the total STP 

influent load (see also Table SI 3). 

The prediction for 27 compounds where both national and 

hospital audit data were available, in some cases deviated 

significantly from the experimentally determined values. 

However, only 8 substances would have been classified 

differently based on audit data when applying strict boundaries 

for the classification which does not change the overall 

picture substantially. 

Possible reasons for three examples are briefly discussed: 1) 

The overestimation in the case of ibuprofen may be reasonably 

explained by the fact that the national consumption is likely to 

be substantially underestimated because ibuprofen can also be 

obtained over the counter and in supermarkets without 

prescription. 2) A patient who regularly takes histamine 

blockers (at home) is likely to take them with him if he is being 

hospitalised (for any treatment not related or interfering with 

histamine blockers). This is one of the cases where patients 

may bring their own medication to the hospital and is also 

assumed to be valid for beta-blockers and diuretics. 3) In some 

countries trimethoprim is often applied together with sulphamethoxazole 

(combination item) and hence would be 

expected in a similar ratio. In Australia, the consumption 

pattern is different: 70% of trimethoprim is sold as single item 

(general public) and in the hospital under investigation even 

90% are administered as individual compound. 

In other cases the explanation may be sought in a higher or 

lower than average number of patients being treated during the 

sampling period in the hospital. However, if the number of 

treated patients shall be estimated from measurements, the 

excretion ratio and absolute recoveries for chemical analyses 

need to be taken in to account (see Eq. (1)). This makes it difficult 

to compare measured influent loads from a STP with audit data 

from an individual health care facility to reliably calculate the 

health care facility’s contribution to the total influent of the STP. 

3.4. Comparison with other studies 

The Caboolture catchment, with 4.4 beds per 1000 inhabitants, 

is comparable with two other studies (3.6 and 4.4 beds per 1000 

inhabitants, see Table 3). Without audit data for the hospitals 

and general public, the load estimations based on measured 

concentrations and an estimate for wastewater based on 

average water consumption in the study by Thomas et al. (2007) 

make a direct comparison difficult. However, higher contributions 

were also found for paracetamol and trimethoprim. In 

the study by Weissbrodt et al. (2009) the loads at the influent of 

the STP were not measured. The percentage determined in this 

study is the amounts measured in the sewer divided by the 

amounts administered on the corresponding days. The 

compounds investigated in the Swiss study are iodinated X-ray 

contrast media and cytostatics, both compounds almost 

exclusively administered in hospitals. Only 50% of the X-ray 

contrast media and a maximum of 7.5% of the cytostatics were 

quantified in the hospital’s effluent, implying that the 

remaining part is most likely ‘‘carried home’’ by patients and 

excreted in household toilets. In the studies by Heberer and 

Feldmann (2005) and Feldmann et al. (2008) the hospital bed 

density is significantly higher (12.1 beds per 1000 inhabitants) 

with a sub-catchment bed density of 24. Pharmaceutical loads 


were measured in the influent of the STP and in selected 

hospital effluents. With day-specific hospital consumption 

data the contribution of the other hospitals was estimated, 

resulting in a total hospital contribution of 15% (carbamazepine), 

10% (diclofenac) and 50% (metamizole, not measured in 

our study). Although the results seem to be in good agreement 

with our study, the limited number of compounds, the various 

approaches used, and the different catchment characteristics 

preclude a comprehensive comparison. 

3.5. Hospital wastewater treatment and catchments in 

South East Queensland 

Over 800 pharmaceuticals, disinfectants and other substances 

are recorded in the DUSC and the hospital database. Whilst 

the 59 substances analysed for in this study presents one of 

the more comprehensive studies of the relative contribution 

of a hospital to total load in wastewater, we do not claim that 

these results can be extrapolated for each of these 800 

substances, at all hospitals, or medical research activities in 

general. As is often the case, the selection of these 59 

substances was based upon the availability of a validated 

analytical method. Despite this ‘‘limitation’’, even if there 

were substances that originate almost exclusively from 

hospital wastewater, or if measures were taken to prevent 

pharmaceutical residues entering hospital wastewater 

(source control, separate collection of urine and faeces 

(Heinzmann et al., 2008)) or if hospital wastewater was treated 

on site, over 85% of the total load for the majority of the 

pharmaceuticals investigated in this study would still reach to 

the STP because they are excreted by the public at home in 

their households. Even for very specific compounds, almost 

exclusively administered in hospitals, the trends in many 

health care systems are moving towards shorter hospitalisations 

or even treatment of out-patients (particularly diagnostics). 

Two examples are the iodinated X-ray media and 

cytostatics: although administered in high amounts in 

hospitals, they cannot be recovered to 100% and hence solely 

attributed to hospital effluent (Weissbrodt et al., 2009). 

Relevance to other catchments in South East Queensland 

(SEQ): The three catchments of main interest within SEQ, the 

ones with advanced water treatment plants for providing 

purified recycled water to the region (for planned indirect 

potable reuse scheme), have approximately 8 hospital beds 

per 1000 inhabitants (Luggage Point, eleven hospitals), 0.4 

(Gibson Island, one hospital) and 1.7 (Bundamba, five hospitals). 

While the hospital in Caboolture (this study) contributes 

4.2 beds per 1000 inhabitants (total in the catchment 4.4), the 

biggest individual hospital in the catchment of Luggage Point 

accounts for only 1.5. A desktop exercise analysing audit data 

from the sum of all hospitals in these catchments is proposed 

to evaluate if further steps are required. This includes the 

planning of future sampling campaigns and the potential 

benefit of treating some hospitals’ wastewater at the source. 


Measurements: For several, widely applied pharmaceuticals, 

an individual hospital seems to be a small additional point

614 

source in the catchment of a sewage treatment plant. In 

this study a hospital with 4.4 hospital beds per 1000 

inhabitants contributed less than 15% to the total load in 

the influent of the sewage treatment plant for 28 

substances, detected in both hospital effluent and STP 

influent, which is in good agreement with estimates from 

other studies. Considering a conservative worst case 

uncertainty estimation, the hospital contribution only 

exceeded 15% for two substances, roxithromycin (max. 

56%) and trimethoprim (max. 18%). 

Audit data: The contribution of the hospital calculated with 

audit data and the chosen classification reveals good 

agreement with actual measurements for three quarters of 

the substances. National audit data to calculate the 

consumption by the general public in a catchment and 

hospital data for in-patients appear to be good predictors. 

This approach can be used with some confidence for 

substances where no analytical method exists to experimentally 

determine concentrations and loads or where the 

LOQ is not low enough. This needs to be tested for other 

countries (dependant upon the comprehensiveness and 

quality of national and hospital audit data). 

Sampling in general: Sampling campaigns in hospital wastewater 

are prone to high uncertainty due to a highly dynamic 

system (flow and concentrations). All effort should be 

undertaken to understand the system (behaviour) prior to 

setting up a sound sampling protocol to ensure that representative 

samples can be obtained. 

Other catchments in South East Queensland: The preliminary 

analysis based on hospital bed densities suggests focusing 

on the catchment of the STP at Luggage Point (approximately 

8 hospital beds per 1000 inhabitants). However it has 

to be noted that this hospital bed density consists of 3 

major public hospitals and a series of private hospitals. 

Since measurements will be very expensive to assess all 

hospitals’ contributions. A detailed desktop analysis of all 

audit data is planned to identify if there are major sources 

and if measurements at selected locations may be 

appropriate. 

Hospital wastewater treatment: If, for whatever motivation, 

hospital wastewater shall be treated separately onsite, it 

must be noted, that for many substances no major overall 

reduction can be achieved since many pharmaceuticals are 

taken on a regular basis at home. With the current trend to 

shorter hospitalisations and treatments (diagnostics) of outpatients, 

this also holds true for compounds mainly 

administered in hospitals. 


We would like to acknowledge the following institutions and 

individuals who contributed to this study: the South East 

Queensland Urban Water Security Research Alliance (financial 

support); David Fillmore, Rick Jones and Wayne Batchler 

from Moreton Bay Water (assistance in setting up proper 

sampling points and support during the campaign); 

Benjamin Tan and Mary Hodge from Queensland Health 


Forensic and Scientific Services (logistics and processing 

samples); John Doonan, Greg Jackson and Daniel Field from 

Queensland Health (making contacts and providing hospital 

data); Maxine Robinson and Chris Raymond from The Drug 

Utilisation Sub-Committee of the Pharmaceutical Benefits 

Advisory Committee, Commonwealth Department of Health 

and Aged Care (providing Australian drug-use statistics); 

Christa McArdell from Eawag (valuable discussion and 

review of manuscript); and the Swiss National Science 

Foundation (Grant PBEZP2-122958 awarded to the first 

author). 

Appendix. 

Supplementary information 

Information about the supplementary material can be found 

at doi:10.1016/j.watres.2009.08.002. 

references 

Baquero, F., Martínez, J.-L., Cantón, R., 2008. Antibiotics and 

antibiotic resistance in water environments. Current Opinion 

in Biotechnology 19, 260–265. 

Boillot, C., Bazin, C., Tissot-Guerraz, F., Droguet, J., Perraud, M., 

Cetre, J.C., Trepo, D., Perrodin, Y., 2008. Daily physicochemical, 

microbiological and ecotoxicological fluctuations of a hospital 

effluent according to technical and care activities. Science of 

the Total Environment 403 (1–3), 113–129. 

de Gruijter, J., Brus, D., Bierkens, M., Knotters, M., 2006. Sampling 

for Natural Resource Monitoring. Springer. 

Feldmann, D.F., Zuehlke, S., Heberer, T., 2008. Occurrence, fate 

and assessment of polar metamizole (dipyrone) residues in 

hospital and municipal wastewater. Chemosphere 71 (9), 

1754–1764. 

Ferk, F., Misik, M., Grummt, T., Majer, B., Fuerhacker, M., 

Buchmann, C., Vital, M., Uhl, M., Lenz, K., Grillitsch, B., 

Parzefall, W., Nersesyan, A., Knasmuller, S., 2009. Genotoxic 

effects of wastewater from an oncological ward. Mutation 

Research – Genetic Toxicology and Environmental 

Mutagenesis 672 (2), 69–75. 

Gautam, A.K., Kumar, S., Sabumon, P.C., 2007. Preliminary study 

of physico-chemical treatment options for hospital 

wastewater. Journal of Environmental Management 83 (3), 

298–306. 

Hartmann, A., Alder, A.C., Koller, T., Widmer, R.M., 1998. 

Identification of fluoroquinolone antibiotics as the main 

source of umuC genotoxicity in native hospital wastewater. 


Heberer, T., Feldmann, D., 2005. Contribution of effluents from 

hospitals and private households to the total loads of 

diclofenac and carbamazepine in municipal sewage effluents – 

modeling versus measurements. Journal of Hazardous 

Materials 122 (3), 211–218. 

Heinzmann, B., Schwarz, R.-J., Schuster, P., Pineau, C., 2008. 

Decentralized collection of iodinated X-ray contrast media in 

hospitals – results of the feasibility study and the practice test 

phase. Water Science and Technology 57 (2). 

Khan, S.J., Ongerth, J.E., 2004. Modelling of pharmaceutical 

residues in Australian sewage by quantities of use and 

fugacity calculations. Chemosphere 54 (3), 355–367.

Kummerer, K., 2001. Drugs in the environment: emission of 

drugs, diagnostic aids and disinfectants into wastewater by 

hospitals in relation to other sources – a review. Chemosphere 

45 (6–7), 957–969. 

Kummerer, K., 2004. Resistance in the environment. Journal of 

Antimicrobial Chemotherapy 54 (2), 311–320. 

Kummerer, K., 2009. Antibiotics in the aquatic environment – 

a review – part II. Chemosphere 75, 435–441. 

Lenz, K., Mahnik, S.N., Weissenbacher, N., Mader, R.M., Krenn, P., 

Hann, S., Koellensperger, G., Uhl, M., Knasmuller, S., Ferk, F., 

Bursch, W., Fuerhacker, M., 2007. Monitoring, removal and 

risk assessment of cytostatic drugs in hospital wastewater. 


Ort, C., Gujer, W., 2006. Sampling for representative 

micropollutant loads in sewer systems. Water Science and 

Technology 54 (6–7), 169–176. 

Ort, C., Lawrence, M.G., Joss, A. Sampling for pharmaceuticals 

and personal care products to assess representative 


environmental loads: sewers and wastewater treatment 

plants, in preparation. 

Pauwels, B., Verstraete, W., 2006. The treatment of hospital 

wastewater: an appraisal. Journal of Water and Health 4 (4), 

405–416. 

Thomas, K.V., Dye, C., Schlabach, M., Langford, K.H., 2007. 

Source to sink tracking of selected human pharmaceuticals 

from two Oslo city hospitals and a wastewater treatment 

works. Journal of Environmental Monitoring 9 (12), 1410– 

1418. 

Watkinson, A.J., Murby, E.J., Kolpin, D.W., Costanzo, S.D., 2009. 

The occurrence of antibiotics in an urban watershed: from 

wastewater to drinking water. Science of the Total 

Environment 407, 2711–2723. 

Weissbrodt, D., Kovalova, L., Ort, C., Pazhepurackel, V., Moser, R., 

Hollender, J., Siegrist, H., McArdell, C., 2009. Mass flows of Xray 

contrast media and cytostatics in hospital wastewater. 

Environmental Science & Technology 43 (13), 4810–4817.

A three-compartment model for micropollutants sorption 

in sludge: Methodological approach and insights 

Maialen Barret, Dominique Patureau, Eric Latrille, Hélène Carrère* 

INRA, UR050, Laboratoire de Biotechnologie de l’Environnement, Avenue des Etangs, 11100 Narbonne, France 

article info 







Keywords: 

Activated sludge 

Bioavailability 

Isotherm 

Sorption model 

Xenobiotic 


abstract 

Organic micropollutants such as Polycyclic Aromatic Hydrocarbons 

(PAHs) present very low water solubility and are 

highly hydrophobic. Due to these properties, they tend to sorb 

to organic matter. In the context of wastewater treatment, 

sorption phenomena firstly impact on physical fate of these 

compounds, determining the fluxes distribution throughout 

the physical separation steps (eg. settling, centrifugation). But 

these physical mechanisms only determine micropollutants 

dispersion, whereas biological mechanisms are mainly 

responsible for their removal. Biological mechanisms may 

also be influenced by sorption, throughout bioavailability 

limitation towards biodegradation. Bioavailability might be 

the main limiting factor of PAHs biodegradation in sludge. 





E-mail address: carrere@supagro.inra.fr (H. Carrère). 


doi:10.1016/j.watres.2009.08.029 

In sludge resulting from wastewater treatment, organic micropollutants sorb to particles 

and to dissolved/colloidal matter (DCM). Both interactions may influence their physical and 

biological fate throughout the wastewater treatment processes. To our knowledge, sludge 

has never been considered as a three-compartment matrix, in which micropollutants 

coexist in three states: freely dissolved, sorbed-to-particles and sorbed-to-DCM. A methodology 

is proposed to concomitantly determine equilibrium constants of sorption to 

particles (Kpart) and to DCM (KDCM). Polycyclic Aromatic Hydrocarbons (PAHs) were chosen 

as model compounds for the experiments. The logarithm of estimated equilibrium 

constants ranged from 3.1 to 4.3 and their usual correlation to PAH hydrophobicity was 

verified. Moreover, PAH affinities for particles and for DCM could be compared. Affinity for 

particles was found to be stronger, probably due to their physical and chemical characteristics. 

This work provided a useful tool to assess the freely dissolved, sorbed-to-particles 

and sorbed-to-DCM concentrations of contaminants, which are necessary to accurately 

predict their fate. Besides, guidelines to investigate the link between sorption and the 

fundamental concept of bioavailability were proposed. 


Micropollutants bioavailability has been widely investigated 

in soils and many tools were subsequently designed to quantify 

it (Reid et al., 2000; Tang et al., 2002; van Straalen et al., 2005; 

Oleszczuk, 2007; Hickman et al., 2008; Oleszczuk, 2008). 

However, these studies referred to different definitions of 

bioavailability. Moreover, bioavailability and bioaccessibility 

were sometimes confused. Semple et al. (2004, 2007) gathered 

most of the encountered concepts into two definitions: 

a bioavailable compound is ‘‘freely available to cross an 

organism’s cellular membrane from the medium the organism 

inhabits at a given time’’; a bioaccessible compound is ‘‘available 

to cross an organism’s cellular membrane from the environment, 

if the organism has access to the chemical’’. These 

definitions imply that a bioaccessible compound can become 

bioavailable after some time.

Nomenclature 

DCM Dissolved and Colloidal Matter 

DM Dry Matter of the sludge, which includes particles 

and DCM 

PAH Polycyclic Aromatic Hydrocarbon 

Cfree Concentration of freely dissolved PAHs (mg/mL) 

CDCM Concentration of sorbed-to-DCM PAHs (mg/gDCM) Cpart Concentration of sorbed-to-particles PAHs 

(mg/gPART) On the contrary, micropollutants bioavailability in sludge 

has been little studied. Related concepts and mechanisms are 

less precisely described than in soil. Some of them may be 

similar between soil and sludge, although matrix characteristics 

and micropollutants/matrix contact time completely differ. 

From a general point of view, micropollutants are usually 

assumed to be bioavailable when they are located in the 

aqueous phase and bioaccessible when they are initially sorbed-to-particles 

and can be transferred to the aqueous phase 

during the process (Byrns, 2001; Artola-Garicano et al., 2003; 

Langford et al., 2005; Urase and Kikuta, 2005; Dionisi et al., 2006). 

Otherwise, the assumption that a fraction of sorbed-to-particles 

compounds would be directly bioavailable can also be 

encountered (Fountoulakis et al., 2006). Anyway, even if the 

nature of the link between sorption phenomena and bioavailability 

was never demonstrated and is still subject to controversy 

when studying sludge, the existence of a link is generally 

admitted. Thus, to study sorption phenomena appears to be the 

first step towards bioavailability investigation. 

In published studies dealing with micropollutants sorption 

in sludge, sludge have been considered as a two-compartment 

system, corresponding to two states for micropollutants: sorbed-to-particles 

and aqueous (Byrns, 2001; Artola-Garicano 

et al., 2003; Ternes et al., 2004; Dionisi et al., 2006; Carballa 

et al., 2008). However, sorption phenomena also occur inside 

aqueous phase. Indeed, micropollutants have been shown to 

interact with dissolved and colloidal matter (DCM) of various 

origins (Chiou et al., 1986; Perminova et al., 2001) including 

sludge from wastewater treatment plant (Holbrook et al., 2004). 

The sorption to DCM is likely to influence micropollutants 

bioavailability in sludge as well as sorption to particles 

(Mackay and Fraser, 2000; Vinken et al., 2004). Thus, the freely 

dissolved and sorbed-to-DCM micropollutants states should be 

differentiated. Sludge should then be considered as a threecompartment 

matrix, with three states of micropollutants: 

freely dissolved, sorbed-to-DCM and sorbed-to-particles. To 

our knowledge, such a model has been used to study sorption 

phenomena in aquatic environmental systems such as sediments 

(Schrap and Opperhuizen, 1992; Lee and Kuo, 1999; 

Amiri et al., 2005) but it has never been applied to sludge. 

Micropollutants sorption equilibria in environmental 

samples including sludge are often modelled by Freundlich 

isotherms with a Freundlich coefficient close to 1, which is 

equivalent to a linear isotherm (Hung et al., 2004; Ivashechkin 

et al., 2004; Arias-Estevez et al., 2007). As a consequence, the 

equilibrium of sorption to particles is usually assumed to fit 

linear equations (Ternes et al., 2004; Carballa et al., 2006; 

Dionisi et al., 2006; Katsoyiannis et al., 2006) as well as the 


C aqu 

K DCM 

Kpart 

Kglobal 

Concentration of apparently dissolved 

micropollutant (mg/mL) 

Equilibrium constant of PAHs sorption to DCM 

(mL/gDCM) Equilibrium constant of PAHs sorption to particles 

(mL/gPART) 

Coefficient of PAHs partition between aqueous 

phase and particles (mL/gPART) 

equilibrium of sorption to DCM (Yamamoto et al., 2003; Zhou 

et al., 2007). 

The objective of this study was to propose and to validate 

a methodological approach to study micropollutants sorption 

equilibria in sludge, considering sludge as a three-compartment 

matrix. Even though this investigation focused on PAHs, 

the methodology and its implications may be transposed to 

a wider range of micropollutants. 


2.1. Sludge source 

All experiments were performed using activated sludge from 

a urban wastewater treatment plant. The influent was entirely 

domestic. The plant consisted in preliminary treatments, 

primary settling, aerated tank, secondary settling and thermophilic 

anaerobic digestion of sewage sludge. Hydraulic 

retention time in the aerated tank was 0.36 day. After collection, 

sludge was divided into several samples and stored at 

20 C in order to guaranty that all following experiments 

were performed with exactly the same matrix. DCM fraction 

represented the dissolved and colloidal matters which were 

not differentiated, defined as the centrifugation supernatant 

(12 000 g, 20 min, 35 C) filtered at 1.2 mm (Whatman GF/C 

glass microfiber filter). As some particles were unwillingly 

suspended back after centrifugation because of little pellet 

cohesiveness, the filtration step was performed to remove 

these particles. The DCM mass retained by the filter was 

negligible, as the DCM fraction beyond the size of 1.2 mm was 

insignificant (data not shown). On total sludge and on DCM 

fraction, the dry matter (DM) was measured by weighing the 

samples after heating at 105 C during 24 h and the organic 

matter after heating at 550 C during 2 h. The defrosted activated 

sludge contained 65.4 1.2 g DM/L, constituted of 7 1% 

of DCM and 93 2% of particles. Organic matter respectively 

accounted for 72 and 80% of particles and DCM dry matter. 


All solvents were purchased from J.T.Baker. Mixtures are 

indicated in volume percentage. Fluorene, phenanthrene, 

anthracene, fluoranthene, pyrene, benzo(a)anthracene, 

chrysene, benzo(b)fluoranthene, benzo(k)fluoranthene, 

benzo(a)pyrene, dibenzo(a,h)anthracene, benzo(g,h,i)perylene 

and indeno(1,2,3,c,d)pyrene powders were obtained from 

Dr Ehrenstorfer GmbH. Each PAH was dissolved in

618 

dichloromethane at 1 g/L. The spiking mix was prepared from 

these individual concentrated solutions, adding 5 mL of each, 

evaporating solvent under gentle nitrogen flow and dissolving 

in 50 mL of acetonitrile. Final concentrations were 100 mg/L 

for each PAH. This mixture was used in equilibrium experiments, 

as competitive phenomena are often assumed not to 

occur at low concentration (Krauss and Wilcke, 2005; Dionisi 

et al., 2006; Zhou et al., 2007). 

The 10 mg/L standard solution of PAHs in acetonitrile was 

provided by Dr Ehrenstorfer GmbH. Dilution factors from 10 to 

1000 were applied to obtain 6 calibration levels. Standards 

were stored at –20 C. 

2.3. PAHs quantification in sludge 

400 mL of sludge were centrifuged during 20 min at 

12 000 g, at 35 C. The supernatant was filtered at 1.2 mm 

with glass microfiber filters (GF/C Whatman) to obtain the 

aqueous phase containing the free and sorbed-to-DCM 

micropollutants, to which 1% of formaldehyde at 37% was 

added for conservation. The particulate phase (pellet) was 

frozen at 20 C and freeze-dried. 

Extraction from aqueous phase was performed using SPE 

cartridges with 1 g C18 sorbent from Macherey-Nagel. 

Cartridges were washed and conditioned with 2 5mL of 

methanol:toluene:acetonitrile (33:33:33), 2 5 mL of methanol 

and 2 5 mL of ultra pure water. 100 mL of sample were 

percolated through the extraction columns at a flow rate of 

2 mL/min. After the sample percolated, the sorbent was dried 

in a stream of atmospheric air during 30 min. Adsorbates were 

eluted with 2 5 mL of methanol:toluene (50:50). Extracts 

were dried under nitrogen flow and dissolved into 1 mL of 

acetonitrile. Each sample was extracted in duplicate and corrected 

by the recovery performance determined on the same 

aqueous sample (same DCM concentration) spiked at 

approximately the same PAHs concentration. 

Extraction from solid phase was carried out as described by 

Trably et al. (2004). 

PAHs quantification in extracts from aqueous and solid 

phases was performed according to the method optimised by 

the same authors (Trably et al., 2004). The sum of PAHs 

concentration in aqueous and in solid phase was compared to 

the spiking concentration to ensure that no significant losses 

occurred during the experiment. 

2.4. The three-compartment model 

A micropollutant present in sludge can be located in one 

among three physical compartments (Fig. 1): the freely dissolved 

one (concentration Cfree, mg/mL), the sorbed-to-DCM 

one (concentration CDCM, mg/gDCM) and the sorbed-to-particles 

one (concentration Cpart, mg/gPART). 

At equilibrium, the three-compartment system can be 

described by the two following equations: 

Kpart ¼ Cpart 

and 

Cfree 


(1) 

Micropollutant 

KDCM ¼ CDCM 

Cfree 

where K part is the equilibrium constant of micropollutant 

sorption to particles (mL/g PART) andK DCM is the equilibrium 

constant of sorption to DCM (mL/g DCM). C free is very difficult to 

obtain experimentally whereas Caqu (concentration of apparently 

dissolved micropollutant, sum of the freely dissolved and 

sorbed-to-DCM states, mg/mL) and Cpart are easily measurable. 

From Caqu and Cpart measurement, Kglobal (mL/gPART) can be 

estimated: 

Kglobal ¼ Cpart 

Caqu 

KpartCfree 

Kpart 

¼ 

¼ 

KDCMCfree½DCMŠþCfree KDCM½DCMŠþ1 

where [DCM] represents the DCM concentration (g/mL). K global 

is not a thermodynamical equilibrium constant but a partition 

coefficient, which is system-dependant. Equation (3) was 

previously used in several studies dealing with the effect of 

DCM on micropollutants sorption in environmental aquatic 

systems (Schrap and Opperhuizen, 1992; Amiri et al., 2005). 

2.5. Equilibration time 

The kinetics of PAHs sorption was monitored in order to 

determine the equilibration time. For this, eight sludge 

samples were firstly mixed at 100 rpm at 35 C in centrifugation 

flasks in order to stabilize the matrix. Then, each sample 

was spiked at 5 mg PAH/g DM to start with the sorption kinetics. 

The aqueous PAHs concentration was measured after 1, 3, 4, 6, 

8, 10, 14 and 24 h of incubation (extractions were carried out in 

duplicate). 

2.6. Linearity verification 

Particle 

K part KDCM 

Dissolved and 

colloidal matter 

(DCM) 

Fig. 1 – Representation of the three-compartment model of 

a micropollutant in sludge. 

At constant DCM concentration, K global is constant. As 

a consequence, the linearity of C part ¼ f(C aqu) is a linearity 

indicator of equilibria (1) and (2). Five samples at 0.7 g DCM/L 

and 4.9 g PART/L were therefore directly put in centrifugation 

flasks and equilibrated at 35 C on a mixing table (100 rpm). 

After 30 min of equilibration, they were spiked with the 

spiking mix at different concentrations ranging from 1 to 5 

mgHAP/gDM and incubated again at 35 C for 4 h. The samples 

were then centrifuged at 35 C and both Cpart and Caqu were 

measured according to quantification section. 

(2) 

(3)

2.7. Determination of sorption equilibrium constants 

Kpart and KDCM can be extracted from equation (3) by assessing 

Kglobal with various DCM concentrations for each PAH. Variations 

of DCM concentrations from 0.6 to 4.5 gDCM/L were 

obtained by diluting sludge with its own supernatant previously 

separated and with water in different proportions. To 

ensure that this dilution method did not modify pH and that 

no particulate matter was released into DCM, the pH and the 

chemical oxygen demand (COD) were measured in 10 samples 

diluted by 1.2 (3.7 gDCM/L) to 200 (0.02 gDCM/L) fold factor. The 

COD was chosen here because it is a fast and little volume 

consuming method. 

The same procedure previously detailed was applied to 

equilibrate, spike, incubate and quantify Cpart and Caqu in the 

samples from 0.6 to 4.5 gDCM/L. Hence, Kglobal dependence to 

DCM concentration was measured. A non linear regression 

algorithm of Levenberg-Marquardt type (Marquardt, 1963) was 

used to minimize the sum of square errors and to estimate the 

two parameters of the K global model: K part and K DCM. In some 

cases, the algorithm did not converge, due to the measurement 

noise and to high K DCM[DCM] values compared to 1. 

Indeed, when the ‘‘þ1’’ term becomes negligible in the 

denominator of equation (3), this leads to Equation (4): 

Kglobalz 

Kpart 

KDCM½DCMŠ 

¼ a 

½DCMŠ 

An infinity of couples (Kpart;KDCM) can be solution of this 

equation, defined by the ratio a¼Kpart/KDCM. a could be 

precisely determined by the algorithm. Afterwards, introduction 

of a into equation (3) allowed us to estimate Kpart for 

each Kglobal measurement and its corresponding [DCM]: 

Kglobal 

Kpart ¼ 

1 Kglobal ½DCMŠ 

a 

The average of calculated Kpart values and the corresponding 

KDCM were used with the Levenberg-Marquardt algorithm 

in order to evaluate standard deviations and correlation 

coefficient. This procedure was applied to dibenzo(a,h)anthracene, 

benzo(g,h,i)perylene and indeno(1,2,3,c,d)pyrene. 


3.1. Hypotheses verification 

From the initial DCM concentration in sludge (12 700 mgO2/L), 

the DCM concentration in diluted samples (from 1.2 to 200 fold 

factor) could be predicted, assuming that no matter was 

transferred from particles into aqueous phase. These predicted 

concentrations were compared to measured values. The ratio 

between them was very close to 1 (Fig. 2). Only for dilution 

factors higher than 20, a little fraction of particulate matter 

(15%) was found to be released, since the ratio reached 1.15. 

The highest dilution factor applied in sorption experiments 

was 7.5, which ensures that no particulate matter release 

occurred during these experiments. As a consequence, the 

DCM concentrations estimated thanks to the dilution factor 

were reliable. Moreover the pH was not modify by the dilution 


(4) 

(5) 

Measured COD (mgO2/L) / 

theoretical COD (mgO2/L) 

1,3 

1,2 

1,1 

1 

0,9 

0,8 

1 10 100 

5,4 

1000 

Dilution factor 

Fig. 2 – Ratio between measured and theoretical aqueous 

COD (mg/L) (black circles) and pH (grey triangle) as 

a function of sludge dilution factor. The dotted frame 

indicates the range of dilution factors used for sorption 

experiments. 

(Fig. 2), implying that the physicochemical properties of DCM 

can be assumed to be unchanged. 

Besides, the absence of particulate matter release in 

diluted samples indicates that sludge DCM does not behave 

like sorbable/desorbable matter. Indeed, matter is likely to be 

either dissolved/colloidal or particulate according to intrinsic 

parameters. This implies that PAHs sorption to particles via 

the sorption of DCM on which PAHs are sorbed does not occur, 

which validates the ‘‘three-compartment model / two equilibria’’ 

description (Fig. 1). 

The kinetics of fluorene and benzo(a)pyrene sorption in 

sludge are presented in Fig. 3. The sorption equilibrium state 

for this two PAHs was achieved within 1 h shaking, and it was 

the same for all other PAHs (data not shown). This value of 1 h 

for equilibrium achievement is consistent with literature 

(Dionisi et al., 2006) and suggests that during wastewater 

treatment processes, bioavailability is not limited by dynamic 

phenomena, as contact between xenobiotics and sludge lasts 

for several hours or days. Furthermore, a 4 h incubation time 

was selected for equilibrium experiments. 

The linear regressions performed to extract K global from 

C part ¼ f(C aqu) produced regression coefficients R 2 ranging 

from 0.85 to 0.95 (Table 1). This is consistent with the 

hypothesis of linearity assumed in the model. Moreover, 

Kglobal was plotted as a function of DCM concentration (Fig. 4) 

for each PAH. The obtained models fitted very well the 

experimental data with R 2 ranging from 0.68 to 0.95, which 

again reinforced the model. 

3.2. K global dependence to DCM concentration 

According to the considered PAH, the measured partition 

coefficients varied of a ten fold factor in the set of tested 

conditions (Fig. 4). Indeed, a little shift of DCM concentration 

was shown to imply an important variation of Kglobal, above all 

at low concentrations. Moreover, when sludge is not considered 

as a three-compartment system and measured Kglobal is 

equated to an equilibrium constant, this so-called constant 

can be largely underestimated. For example, measured Kglobal 

6,4 

6,2 

6 

5,8 

5,6 

pH

620 

Aqueous fraction (%) 

100 

80 

60 

40 

20 

Fluorene 

for benzo(b)fluoranthene at 0.6 g DCM/L was about 2000 mL/ 

g PART whereas the real equilibrium constant K part was estimated 

at 6920 mL/gPART (Fig. 4). In this case, assimilating 

Kglobal to the equilibrium constant would lead to an underestimation 

of 71%. Overall, this underestimation varied from 57 

to 94% within the PAHs family and the more hydrophobic the 

PAH, the higher the error. Such an underestimation prevents 

from accurately predicting micropollutants sorption in sludge. 

All these results highlighted the necessity to consider the 

third phase to precisely assess the micropollutants fate. 

The obtained partition coefficients could not be strictly 

compared to literature data. Indeed, the studies published 

until now about PAHs sorption in sludge are very few, and 

authors neither take into account the three compartments nor 

measure DCM concentration. Moreover, our results underlined 

the strong Kglobal dependence to this concentration, 

which obliges to compare Kglobal values at equivalent DCM 

concentrations. 

3.3. Comparison between K part and K DCM 

0 

0 10 20 30 

0 

0 10 20 30 

t (h) 

t (h) 

The logarithm of thermodynamical equilibrium constants 

Kpart and KDCM extracted from Kglobal dependence to DCM 

concentration are presented in Table 1. Standard deviation 


Aqueous fraction (%) 

100 

80 

60 

40 

20 

Benzo(a)pyrene 

Fig. 3 – Percentage of initially spiked PAHs that remain in aqueous phase as a function of equilibration time. Since PAHs are 

spiked in liquid form, at time 0, the aqueous fraction is assumed to be 100%. 

was higher for highly hydrophobic compounds than for 

weakly hydrophobic ones, due to mathematical difficulties 

mentioned in the paragraph 2.7. The first way to reduce 

standard deviations would be to multiply measurements in 

order to decrease experimental uncertainty due to (i) sludge 

heterogeneity and (ii) quantification procedure. The second 

way would be to optimize measurements distribution on the 

DCM concentration scale. However, this option is limited for 

practical reasons. Indeed, the upper DCM concentration limit 

comes from the DCM concentration in raw sludge that cannot 

be passed over. Besides, to minimize DCM concentration, no 

supernatant volume should be added and a small sludge 

volume should be diluted with water. Nevertheless, particles 

concentration would concomitantly decrease and particles 

quantity needed for PAH quantification would constitute the 

lower DCM concentration limit. Nonetheless, standard deviation 

values for most PAHs were acceptable in comparison 

with the literature data on adsorption topic, since standard 

deviations, although very seldom estimated, are in this range 

(Kordel et al., 1997; Dionisi et al., 2006; Carballa et al., 2008). 

Again, K part comparison with previously published data 

was hardly feasible. However, at low particulate matter 

concentration (1 g/L), which indicates a probable very low 

DCM concentration, Dionisi et al. (2006) measured a K global of 

Table 1 – log Kow, partition coefficient Kglobal and coefficient R 2 obtained from linear regression of Cpart [ f(Caqu) at 0.7 gDCM/L, 

and equilibrium constants extracted from Kglobal dependence to DCM concentration in the range 0.6–4.5 gDCM/L. 

Compound log Kow Partition coefficient Equilibrium constants 

K global (mL/ g part) R 2 

log K part 

log K DCM 

Fluorene 4.18 638 0.89 3.21 0.23 3.24 0.37 

Phenanthrene 4.46 2142 0.92 3.56 0.24 3.35 0.36 

Anthracene 4.5 1493 0.92 3.45 0.17 3.11 0.29 

Fluoranthene 4.9 2357 0.85 4.01 0.44 3.46 0.61 

Pyrene 4.88 3196 0.95 3.71 0.37 3.20 0.62 

Benzo(a)anthracene 5.63 2712 0.92 3.78 0.21 3.33 0.32 

Chrysene 5.63 1833 0.92 3.86 0.40 3.47 0.56 

Benzo(b)fluoranthene 6.04 2381 0.89 3.84 0.38 3.49 0.52 

Benzo(k)fluoranthene 6.21 1857 0.90 4.27 0.76 4.02 0.85 

Benzo(a)pyrene 6.06 566 0.89 4.01 0.55 3.83 0.65 

Dibenzo(a,h)anthracene 6.86 1701 0.90 4.34 1.57 4.12 1.71 

Benzo(g,h,i)perylene 6.78 1542 0.89 4.11 1.10 4.00 1.20 

Indeno(1,2,3,c,d)pyrene 6.58 1035 0.87 4.03 0.83 3.81 0.97

) 

( mL/ 

gP 

ART 

mL/ 

gP 

AR 

) Kglobal ( T 

mL/ 

gP 

AR 

) Kglobal ( T 

P ) Kglobal Kglobal ( mL/ 

g ART 

) 

Kglobal ( mL/ 

gP 

ART 

1200 

800 

400 

Fluorene 

K part = 1 606 

K DCM = 1 717 

R 2 = 0.857 

0 

0 1 2 3 4 5 

8000 

Fluoranthene 

Kpart = 10 225 

6000 

4000 

KDCM = 2 894 

R2 = 0.823 

2000 

0 

0 1 2 3 4 5 

Chrysene 

5000 

4000 

3000 

Kpart = 7 251 

KDCM = 2 940 

R2 = 0.854 

2000 

1000 

3000 

2000 

1000 

9700 5000 mL/g PART for pyrene in an activated sludge. This 

value might be compared to our K part, asK part corresponds to 

K global limit when DCM concentration tends to zero. We 

calculated a K part in the range 2200–12 000 mL/g PART for pyrene 

(Table 1), which is in good agreement with Dionisi et al.’s 

(2006) values. 

KDCM is easier to compare with literature data than Kpart, 

since several studies were performed in absence of particles. As 

an illustration, Holbrook et al. (2004) studied pyrene adsorption 

onto sludge DCM. Depending on sample origin and size 

Phenanthrene 

K part = 3 659 

K DCM = 2 244 

R 2 = 0.902 

0 

0 1 2 3 4 5 

4000 

Pyrene 

Kpart = 5 089 

3000 

2000 

KDCM = 1 594 

R2 = 0.680 

1000 

0 

0 1 2 3 4 5 

Benzo(b)fluoranthene 

5000 

4000 

3000 

Kpart = 6 920 

KDCM = 3 073 

R2 = 0.877 

2000 

1000 

3000 

2000 

1000 

Anthracene 

K part = 3 636 

K DCM= 2 117 

R 2 = 0.833 

0 

0 1 2 3 4 5 

Benzo(a)anthracene 

5000 

4000 

3000 

Kpart = 7 941 

KDCM = 3 337 

R2 = 0.876 

2000 

1000 

0 

0 1 2 3 4 5 

8000 

Benzo(k)fluoranthene 

Kpart = 18 604 

6000 

4000 

KDCM = 10 429 

R2 = 0.950 

0 

0 1 2 3 4 5 

0 

0 1 2 3 4 5 

0 

0 1 2 3 4 5 

5000 

4000 

3000 

Kpart = 10 283 

KDCM = 6 745 

R 

2000 

1000 

2 8000 

= 0.940 

6000 

4000 

Kpart = 22 076 

KDCM = 13 202 

R 

2000 

2 5000 

= 0.863 

4000 

3000 

Kpart = 12 770 

KDCM = 9 992 

R 

2000 

1000 

2 Benzo(a)pyrene Dibenzo(a,h)anthracene Benzo(g,h,i)perylene 

α = 1.672 

α = 1.278 

= 0.895 

5000 

4000 

3000 

2000 

1000 

0 

0 1 2 3 4 5 

Indeno(1,2,3,c,d)pyrene 

Kpart = 10 599 

KDCM = 6 465 

R2 α = 1.639 

= 0.823 

0 

0 1 2 3 4 5 

DCM (g/L) 


0 

0 1 2 3 4 5 

DCM (g/L) 

2000 

0 

0 1 2 3 4 5 

DCM (g/L) 

Fig. 4 – Measured (mark) and modelled (line) K global (mL/g PART) as a function of DCM concentration (g/L). The a coefficient (for 

dibenzo(a,h)anthracene, benzo(g,h,i)perylene and indeno(1,2,3,c,d)pyrene), Kpart,KDCM and regression coefficients estimated 

thanks to the algorithm are presented on each graph. The corresponding logarithmic values and standard deviation for Kpart 

and KDCM are presented in Table 1. 

fractionation, the authors measured some K DCM values ranging 

from below 1000–80 000 mL/g OC (g of organic carbon). Since 

organic carbon usually represents approximately 25% of the 

total matter in sludge (Dignac et al., 2000), these values are 

widespread from below 250 to 20 000 mL/gDCM. This range is in 

accordance with our estimation of 380–6607 mL/gDCM (Table 1). 

The common and general correlation between the logarithm 

of equilibrium constants and hydrophobicity of PAHs 

(Karickhoff et al., 1979; Poerschmann and Kopinke, 2001) was 

observed for both Kpart and KDCM (Fig. 5). This demonstrates

622 

log K part , log K DCM 

7 

6 

5 

4 

3 

2 

y = 0.308x + 2.13 

y = 0.348x + 1.61 

1 

4 4.5 5 5.5 6 6.5 7 

log K ow 

Fig. 5 – log Kpart (black rounds) and log KDCM (grey rounds) 

as a function of PAHs log K ow. 

that PAHs sorption to particles and to DCM may be driven by 

hydrophobic interactions. Moreover, the linear regressions 

revealed very similar slopes (0.31 and 0.35) but different 

origins (2.13 and 1.61), resulting in different values (Fig. 5). 

Indeed, the PAHs affinity for particles was found to be stronger 

than for DCM. This finding can be related to the compartments 

characteristics since equilibrium constants reflect the 

affinity of PAHs for the corresponding compartment. In this 

work, the compartments were differentiated according to 

their physical criteria, as defined in the 2.1. Section: the 

density criteria imposed by centrifugation, followed by the 

size criteria of filtration. Other physical properties such as 

hydrophobicity, surface charge and specific surface area 

differentiate particles and DCM. In addition of physical properties, 

Bougrier et al. (2008) demonstrated that biochemical 

composition (chemical oxygen demand, carbohydrates and 

proteins content) is different in particles and DCM. These 

properties are influenced by sludge history in terms of 

substrate (wastewater), involved microbiological population, 

and process operating parameters. For example, high sludge 

retention time in activated sludge process reduces particles 

specific surface area (Liss et al., 2002) and leads to more 

hydrophobic particles (Liao et al., 2001) with higher lipid 

content (Liss et al., 2002). Quantifying the influence of physical 

and biochemical sludge characteristics on the sorption equilibrium 

constants of PAHs would be very interesting. 

3.4. Assessment of PAHs compartment distribution 

Kpart and KDCM determination and dependence to hydrophobicity 

provided the parameters needed to simulate PAHs 

distribution between the three sludge compartments. 

Aqueous and freely dissolved percentages were simulated in 

two theoretical and extreme cases: for a weakly hydrophobic 

PAH and for a strongly hydrophobic one (Fig. 6). 

The simulation demonstrated that freely dissolved fraction 

does not exceed 0.4% of the total concentration for a very 

hydrophobic compound and 2.5% for a weakly hydrophobic 

one. This fraction is slightly influenced by DCM concentration 

in both cases. 

On the contrary, aqueous fraction of both PAHs appeared 

to strongly depend on DCM concentration. This result is of 


% 

35 

30 

25 

20 

15 

10 

5 

0 

0 2 4 

6 8 

DCM (g/L) 

Fig. 6 – Percentage of free (plots) and aqueous (linked plots) 

concentrations modelled for a theoretical weakly 

hydrophobic PAH (log K ow [ 4.2, square) and for 

a theoretical very hydrophobic PAH (log Kow [ 6.8, triangle) 

as a function of dissolved and colloidal matter 

concentration, particulate concentration being fixed at 

10 g/L. For this simulation, Kpart and KDCM of both PAHs 

were estimated thanks to the relationship presented in 

Fig. 5. The reversion point is located at 6.01 gDCM/L, which 

corresponds to 25.5% of aqueous PAHs. Over this value, 

aqueous fraction is slightly higher for the very 

hydrophobic PAH than for the weakly hydrophobic one. 

great concern since aqueous fraction is usually equated to 

bioavailable fraction, as previously mentioned. It means that 

low DCM concentrations favour higher percentage in aqueous 

phase for weakly hydrophobic PAHs than for very hydrophobic 

ones, whereas reasonable high concentrations 

discriminate less PAHs. The extrapolation of the model to 

extremely concentrated conditions ([DCM] > 6 g/L) shows that 

the aqueous concentration of a very hydrophobic PAH could 

equal and even exceed a weakly hydrophobic one (Fig. 6). In 

the usual bioavailability hypothesis ‘‘aqueous is bioavailable’’, 

this would mean that a very hydrophobic PAH could be more 

bioavailable than a weakly hydrophobic one. This case is 

really atypical, since micropollutants are usually assumed to 

be less bioavailable when they are more hydrophobic (Mackay 

and Fraser, 2000). It is an important insight derived from the 

three-compartment approach. The reversion point would be 

located at [DCM] ¼ 6.0 g/L for the studied activated sludge, as 

particulate matter concentration was fixed at 10 g/L for the 

simulation. However, it has to be reminded that the model 

was established for DCM concentrations up to 4.5 g/L. Moreover, 

DCM at the reversion point would represent 37.5% of the 

total matter, which is much higher than common values 

(Bougrier et al., 2008). As a conclusion, the latter case is little 

probable in real systems, but it could be interesting to create it 

in laboratory investigations. 

Apart from the descriptive issue previously discussed, such 

simulations constitute a new tool for micropollutants physical 

fate assessment. They will allow to predict the fate of PAHs 

throughout the separation steps encountered in wastewater

treatment plants. In the example of a settler, the flux of sorbedto-particles 

PAHs in solid effluent (for agricultural disposal) 

can be quantified as well as the flux of sorbed-to-DCM and 

dissolved PAHs carried by clarified water (discharged into 

surface water). Indeed, Katsoyiannis and Samara (2007) 

observed an influence of DCM concentration on micropollutants 

fate, without quantifying this effect. The threecompartment 

approach could provide the missing data to 

quantitatively take this interaction between micropollutants 

and DCM into account. As a result, the methodology presented 

here constitutes a real improvement in the assessment of 

micropollutants fate throughout treatment plants and of their 

environmental discharges. 

3.5. Towards bioavailability 

Quantifying PAHs distribution in sludge samples has an 

important outcome concerning its consequences on 

bioavailability topic. Indeed, it may be useful to verify the 

usual hypothesis of identity between aqueous and bioavailable 

fractions (Artola-Garicano et al., 2003; Dionisi et al., 2006). 

This assumption may be verified by measuring biodegradation 

at different aqueous concentrations. Initial biodegradation 

rate in batch experiments could be an appropriate 

indicator for biodegradation. The assumption would be 

confirmed if biodegradation rate correlates with aqueous 

concentration. Guidelines for such an experiment are 

provided by the present work: it suggests that different 

aqueous concentrations of one PAH might be achieved thanks 

to different DCM concentrations. 

On the other hand, investigation on the link between freely 

dissolved and bioavailable fractions of one PAH appears more 

difficult, as freely dissolved concentration varied little in the 

presented conditions. It means that this link might be investigated 

through either the comparison between different 

compounds in one matrix or the comparison of one 

compound in various matrixes for which this compound has 

different affinities. 

These guidelines are of great importance for the modeling 

of micropollutants biological fate in sludge, since until now, 

models have been built on hypotheses which have never been 

clearly demonstrated. 


The three-compartment methodology proposed to investigate 

micropollutants sorption in sludge was validated. The results 

underlined the fact that a reliable tool for micropollutants 

behaviour prediction in sludge should integrate a threecompartment 

model since DCM concentration influences 

micropollutants distribution in a significant way. Indeed, the 

usual biphasic approach was shown to lead to an important 

underestimation of equilibrium constants. Despite the lack of 

published data about PAH sorption in sludge, the results could 

be confronted to a few studies. The confrontation showed the 

consistency of equilibrium constant values and of dependency 

to PAH hydrophobicity. 

Besides, PAH affinities for particles and for DCM could be 

compared. Affinity for particles was found to be stronger. This 


may be due to the differences between particles and DCM in 

term of physical and chemical characteristics. Further investigation 

is necessary to include the influence of both sludge 

and micropollutant characteristics in the three-compartment 

model. This research, being presently undergone by the 

authors, will enable to predict PAHs behaviour in any sludge, 

as a function of its characteristics. The future predictive 

model would allow to estimate the PAH distribution in any 

sludge in order (i) to determine fluxes throughout physical 

separation steps and (ii) to confront PAHs distribution 

between compartments to their biodegradation, which might 

help to elucidate bioavailability issue. 


This study was funded by the Institut National de Recherche 

Agronomique and the Ministère de l’Enseignement Supérieur 

et de la Recherche du Gouvernement Français, France, and the 

authors gratefully acknowledge them. 

references 

Amiri, F., Bornick, H., Worch, E., 2005. Sorption of phenols onto 

sandy aquifer material: the effect of dissolved organic matter 

(DOM). Water Research 39, 933–941. 

Arias-Estevez, M., Fernandez-Gandara, D., Garcia-Falcon, M.S., 

Garcia-Rio, L., Mejuto, J.C., Simal-Gandara, J., 2007. Sorption of 

PAHs to colloid dispersions of humic substances in water. 

Bulletin of Environmental Contamination and Toxicology 79, 

251–254. 

Artola-Garicano, E., Borkent, I., Damen, K., Jager, T., Vaes, W.H.J., 

2003. Sorption kinetics and microbial biodegradation activity 

of hydrophobic chemicals in sewage sludge: model and 

measurements based on free concentrations. Environmental 


Bougrier, C., Delgenès, J.P., Carrère, H., 2008. Effects of thermal 

treatments on five different waste activated sludge samples 

solubilisation, physical properties and anaerobic digestion. 

Chemical Engineering Journal 139, 236–244. 

Byrns, G., 2001. The fate of xenobiotic organic compounds in 

wastewater treatment plants. Water Research 35, 2523–2533. 

Carballa, M., Fink, G., Omil, F., Lema, J.M., Ternes, T., 2008. 

Determination of the solid-water distribution coefficient (K-d) 

for pharmaceuticals, estrogens and musk fragrances in 

digested sludge. Water Research 42, 287–295. 

Carballa, M., Omil, F., Alder, A.C., Lema, J.M., 2006. Comparison 

between the conventional anaerobic digestion of sewage 

sludge and its combination with a chemical or thermal pretreatment 

concerning the removal of pharmaceuticals and 

personal care products. Water Science and Technology 53, 

109–117. 

Chiou, C.T., Malcolm, R.L., Brinton, T.I., Kile, D.E., 1986. Water 

solubility enhancement of some organic pollutants and 

pesticides by dissolved humic and fulvic acids. Environmental 


Dignac, M.F., Ginestet, P., Rybacki, D., Bruchet, A., Urbain, V., 

Scribe, P., 2000. Fate of wastewater organic pollution during 

activated sludge treatment: nature of residual organic matter. 


Dionisi, D., Bertin, L., Bornoroni, L., Capodicasa, S., Papini, M.P., 

Fava, F., 2006. Removal of organic xenobiotics in activated 

sludges under aerobic conditions and anaerobic digestion of

624 

the adsorbed species. Journal of Chemical Technology and 

Biotechnology 81, 1496–1505. 

Fountoulakis, M.S., Stamatelatou, K., Batstone, D.J., Lyberatos, G., 

2006. Simulation of DEHP biodegradation and sorption during 

the anaerobic digestion of secondary sludge. Water Science 

and Technology 54, 119–128. 

Hickman, Z.A., Swindell, A.L., Allan, I.J., Rhodes, A.H., Hare, R., 

Semple, K.T., Reid, B.J., 2008. Assessing biodegradation 

potential of PAHs in complex multi-contaminant matrices. 

Environmental Pollution 156, 1041–1045. 

Holbrook, R.D., Love, N.G., Novak, J.T., 2004. Investigation of 

sorption behavior between pyrene and colloidal organic 

carbon from activated sludge processes. Environmental 


Hung, N.V., Tateda, M., Ike, M., Fujita, M., Tsunoi, S., Tanaka, M., 

2004. Sorption of biodegradation end products of nonylphenol 

polyethoxylates onto activated sludge. Journal of 

Environmental Sciences-China 16, 564–569. 

Ivashechkin, P., Corvini, P.F.X., Dohmann, M., 2004. Behaviour of 

endocrine disrupting chemicals during the treatment of 

municipal sewage sludge. Water Science and Technology 50, 

133–140. 

Karickhoff, S.W., Brown, D.S., Scott, T.A., 1979. Sorption of 

hydrophobic pollutants on natural sediments. Water Research 

13, 241–248. 

Katsoyiannis, A., Samara, C., 2007. The fate of dissolved 

organic carbon (DOC) in the wastewater treatment process 

and its importance in the removal of wastewater 

contaminants. Environmental Science and Pollution 

Research 14, 284–292. 

Katsoyiannis, A., Zouboulis, A., Samara, C., 2006. Persistent 

organic pollutants (POPs) in the conventional activated sludge 

treatment process: model predictions against experimental 

values. Chemosphere 65, 1634–1641. 

Kordel, W., Hennecke, D., Franke, C., 1997. Determination of the 

adsorption-coefficients of organic substances on sewage 

sludges. Chemosphere 35, 107–119. 

Krauss, M., Wilcke, W., 2005. Persistent organic pollutants in soil 

density fractions: distribution and sorption strength. 


Langford, K.H., Scrimshaw, M.D., Birkett, J.W., Lester, J.N., 2005. 

The partitioning of alkylphenolic surfactants and 

polybrominated diphenyl ether flame retardants in activated 

sludge batch tests. Chemosphere 61, 1221–1230. 

Lee, C., Kuo, L., 1999. Quantification of the dissolved organic 

matter effect on the sorption of hydrophobic organic 

pollutant: application of an overall mechanistic sorption 

model. Chemosphere 38, 807–821. 

Liao, B.Q., Allen, D.G., Droppo, I.G., Leppard, G.G., Liss, S.N., 2001. 

Surface properties of sludge and their role in bioflocculation 

and settleability. Water Research 35, 339–350. 

Liss, S.N., Liao, B.Q., Droppo, I.G., Allen, D.G., Leppard, G.G., 2002. 

Effect of solids retention time on floc structure. Water Science 


Mackay, D., Fraser, A., 2000. Bioaccumulation of persistent 

organic chemicals: mechanisms and models. Environmental 

Pollution 110, 375–391. 

Marquardt, D., 1963. An algorithm for least-squares estimation of 

nonlinear parameters. SIAM Journal of Applied Mathematics 

11, 431–441. 

Oleszczuk, P., 2007. Investigation of potentially bioavailable 

and sequestrated forms of polycyclic aromatic 


hydrocarbons during sewagesludgecomposting. 


Oleszczuk, P., 2008. Tenax-TA extraction as predictor for free 

available content of polycyclic aromatic hydrocarbons (PAHs) 

in composted sewage sludges. Journal of Environmental 

Monitoring 10, 883–888. 

Perminova, I.V., Grechishcheva, N.Y., Kovalevskii, D.V., 

Kudryavtsev, A.V., Petrosyan, V.S., Matorin, D.N., 2001. 

Quantification and prediction of the detoxifying properties of 

humic substances related to their chemical binding to 

polycyclic aromatic hydrocarbons. Environmental Science and 


Poerschmann, J., Kopinke, F.D., 2001. Sorption of very 

hydrophobic organic compounds (VHOCs) on dissolved humic 

organic matter (DOM). 2. Measurement of sorption and 

application of a Flory–Huggins concept to interpret the data. 


Reid, B.J., Stokes, J.D., Jones, K.C., Semple, K.T., 2000. 

Nonexhaustive cyclodextrin-based extraction technique for 

the evaluation of PAH bioavailability. Environmental Science 


Schrap, S.M., Opperhuizen, A., 1992. On the contradictions 

between experimental sorption data and the sorption 

partitioning model. Chemosphere 24, 1259–1282. 

Semple, K.T., Doick, K.J., Jones, K.C., Burauel, P., Craven, A., 

Harms, H., 2004. Defining bioavailability and bioaccessibility 

of contaminated soil and sediment is complicated. 

Environmental Science and Technology 38, 228A–231A. 

Semple, K.T., Doick, K.J., Wick, L.Y., Harms, H., 2007. Microbial 

interactions with organic contaminants in soil: definitions, 

processes and measurement. Environmental Pollution 150, 

166–176. 

Tang, J.X., Liste, H.H., Alexander, M., 2002. Chemical assays of 

availability to earthworms of polycyclic aromatic 

hydrocarbons in soil. Chemosphere 48, 35–42. 


Joss, A., 2004. A rapid method to measure the solid-water 

distribution coefficient (K-d) for pharmaceuticals and musk 

fragrances in sewage sludge. Water Research 38, 4075–4084. 

Trably, E., Delgenes, N., Patureau, D., Delgenes, J.P., 2004. 

Statistical tools for the optimization of a highly reproducible 

method for the analysis of polycyclic aromatic hydrocarbons 

in sludge samples. International Journal of Environmental 

Analytical Chemistry 84, 995–1008. 

Urase, T., Kikuta, T., 2005. Separate estimation of adsorption and 

degradation of pharmaceutical substances and estrogens in 

the activated sludge process. Water Research 39, 1289–1300. 

van Straalen, N.M., Donker, M.H., Vijver, M.G., van Gestel, C.A.M., 

2005. Bioavailability of contaminants estimated from uptake rates 

into soil invertebrates. Environmental Pollution 136, 409–417. 

Vinken, R., Hollrigl-Rosta, A., Schmidt, B., Schaffer, A., Corvini, P.F.X., 

2004. Bioavailability of a nonylphenol in dependence on the 

association to dissolved humic substances. Water Science and 


Yamamoto,H.,Liljestrand,H.M.,Shimizu,Y.,Morita,M.,2003.Effects 

of physical–chemical characteristics on the sorption of selected 

endocrine disruptors by dissolved organic matter surrogates. 


Zhou,J.L.,Liu,R.,Wilding,A.,Hibberd,A.,2007.Sorptionof 

selected endocrine disrupting chemicals to different aquatic 

colloids. Environmental Science and Technology 41, 

206–213.

Removal of micropollutants and reduction of biological 

activity in a full scale reclamation plant using ozonation 

and activated carbon filtration 

J. Reungoat a, *, M. Macova b , B.I. Escher b , S. Carswell b,c , J.F. Mueller b , J. Keller a 

a 

The University of Queensland, Advanced Water Management Centre (AWMC), Qld 4072, Australia 

b 

The University of Queensland, National Research Centre for Environmental Toxicology (EnTox), QLD 4108, Australia 

c 

Queensland Health Forensic and Scientific Services, Organics Laboratory, QLD 4108, Australia 

article info 







Keywords: 

Water reuse 

Oxidation 

Adsorption 

Pesticides 


Bioassays 




abstract 


E-mail address: j.reungoat@awmc.uq.edu.au (J. Reungoat). 


doi:10.1016/j.watres.2009.09.048 

Pharmaceutical compounds are found in secondary treated effluents up to mgL -1 levels and 

therefore discharged into surface waters. Since the long term effects of these compounds 

on the environment and human health are, to date, largely unknown, implementation of 

advanced treatment of wastewaters is envisaged to reduce their discharge. This is of 

particular relevance where surface waters are used as drinking water sources and when 

considering indirect potable reuse. This study aimed at assessing the removal of organic 

micropollutants and the concurrent reduction of their biological activity in a full scale 

reclamation plant treating secondary effluent. The treatment consists of 6 stages: denitrification, 

pre-ozonation, coagulation/flocculation/dissolved air flotation and filtration 

(DAFF), main ozonation, activated carbon filtration and final ozonation for disinfection. For 

that purpose, representative 24-hour composite samples were collected after each stage. 

The occurrence of 85 compounds was monitored by LC/MS-MS. A battery of 6 bioassays 

was also used as a complementary tool to evaluate non-specific toxicity and 5 specific toxic 

modes of action. Results show that, among the 54 micropollutants quantified in the 

influent water, 50 were removed to below their limit of quantification representing more 

than 90% of concentration reduction. Biological activity was reduced, depending on the 

specific response that was assessed, from a minimum of 62% (AhR response) to more than 

99% (estrogenicity). The key processes responsible for the plant’s performances were the 

coagulation/flocculation/DAFF, main ozonation and activated carbon filtration. The effect 

of these 3 processes varied from one compound or bioassay to another but their combination 

was almost totally responsible for the overall observed reduction. Bioassays yielded 

complementary information, e.g. estrogenic compounds were not detected in the 

secondary effluent by chemical analysis, but the samples had an estrogenic effect. The 

main ozonation formed oxidation by-products of the organic micropollutants but 

decreased the level of non-specific toxicity and other specific toxic modes of action, 

demonstrating that the mixture of oxidation by-products was less potent than the mixture 

of the parent compounds for the considered effects. 


626 


In the last decade, numerous studies around the world have 

demonstrated the presence of pharmaceutical compounds in 

domestic wastewaters. These compounds are removed to 

different degrees by the commonly used biological activated 

sludge processes. While some compounds (e.g. ibuprofen, 

paracetamol) are effectively removed, others (e.g. carbamazepine, 

diclofenac) are barely affected by the biological treatment 

(Onesios et al., 2009). As a result, some pharmaceutical 

compounds are released into the environment, contaminating 

surface waters. This situation is of concern as these 

compounds have been designed to affect biological systems 

and long term exposure effects on the biological elements in 

the environment are still largely unknown. Human health is 

also at risk when treated wastewater is discharged to water 

bodies that are used as drinking water sources or considered 

for indirect potable reuse. Therefore, advanced treatment of 

wastewaters has to be envisaged to reduce the load of these 

compounds in discharged effluents. 

Several technologies have proven to be effective in 

removing pharmaceutical compounds from water: activated 

carbon adsorption (Nowotny et al., 2007; Snyder et al., 2007; 

Ternes et al., 2002; Westerhoff et al., 2005; Yu et al., 2008), 

ozonation and advanced oxidation processes (Esplugas et al., 

2007; Huber et al., 2003; Huber et al., 2005; Kim et al., 2008; 

Nakada et al., 2007; Ternes et al., 2003; Zwiener and Frimmel, 

2000) and membrane filtration (Kimura et al., 2004; Snyder 

et al., 2007; Yoon et al., 2007). Activated carbon adsorption and 

ozonation are considered to be economically feasible for 

tertiary treatment of wastewaters by Joss et al. (2008). Nevertheless, 

none of these processes can remove all the 

compounds of concern and to date, no extensive study has 

been carried out on the efficiency of the sequential combination 

of these two processes for the removal of micropollutants 

from secondary treated wastewater. 

Chemical analysis of micropollutants in water has gone 

through major developments in recent years and today’s 

analytical techniques can measure some compounds at 

concentrations as low as a few ng L 1 . Chemical analysis is 

powerful to measure the concentration of targeted 

compounds but does not cover the whole range of chemicals 

that might be present in the effluent of a wastewater treatment 

plant. Analytical techniques usually focus on pharmaceutically 

active parent compounds whereas a variable 

fraction is excreted from the human body as metabolites. 

Moreover, biological treatment of wastewater may form yet 

unidentified compounds through biodegradation reactions; 

e.g. 17b-estradiol is degraded to estrone (Ternes et al., 1999). 

Subsequent chemical treatments such as ozonation can also 

form by-products. Chemical analysis of a limited suite of 

compounds does not allow assessment of the potential biological 

adverse effects of the wastewater as the cumulative 

effect of the mixture of chemicals that may be present cannot 

be easily integrated. Bioassays have been utilised as complementary 

monitoring tools for the assessment of possible biological 

effects of chemicals that are present in a particular 

water sample (Escher et al., 2008b; Muller et al., 2007). These 

bioanalytical tools are designed to quantify non-specific 


toxicity or particular toxic modes of action (e.g. estrogenicity, 

genotoxicity, phytotoxicity) induced by a sample on a biological 

organism or a biological process. To date, bioassays have 

not been widely used in the evaluation of water treatment 

processes and have rarely been accompanied with chemical 

analysis (Macova et al., 2010). 

The present study assessed the removal of selected 

micropollutants (pharmaceuticals and pesticides) and the 

decrease of biological activity along a full scale water reclamation 

plant using ozonation and activated carbon adsorption 

to treat a secondary effluent. This aimed at identifying the 

key steps in the treatment process as well as the key operating 

parameters. Results of chemical analysis and bioassays were 

compared to determine whether the use of both techniques 

was redundant or allowed a deeper understanding of the 

performances of individual treatment processes. The performance 

and relevance of the treatment train is further discussed 

in the context of indirect potable reuse. 


2.1. South Caboolture Water Reclamation Plant 

The South Caboolture Water Reclamation Plant was designed 

to reduce riverine pollution from the 40,000 population 

equivalent wastewater treatment plant and to provide recycled 

water to industry and community consumers. Whilst the 

plant provides water for non-potable applications, it has been 

designed to meet drinking water standards. The treatment 

process detailed in Fig. 1 incorporates biological denitrification, 

pre-ozonation, coagulation/flocculation/dissolved air 

flotation-sand filtration (DAFF), main ozonation, biological 

activated carbon filtration and final ozonation for disinfection. 

The activated carbon was renewed in March 2008 after 9 years 

of service, its adsorption capacity was assumed not to be 

exhausted at the time sampling took place, 4 months later. 

van Leeuwen et al. (2003) published more details on the 

process and its performances. 

2.2. Sample collection 

Four sets of samples were collected over winter 2008 under 

dry weather conditions including three during week days and 

one during the weekend (11-07-08, 22-07-08, 27-07-08 and 06- 

08-08). Water temperature across the plant was 22 2 C and 

pH was 7.0 0.5. Samples were collected at 7 sampling points 

along the treatment train, labelled S1 to S7 on Fig. 1, in order to 

evaluate the performance of individual treatment steps. As 

the flow rate in the reclamation plant is constant (8 megalitres 

per day), representative samples were collected as time 

proportional 24-h composites. For S1 and S7, refrigerated 

auto-samplers collected 200 mL every 15 minutes; for the 

remaining sampling locations, a continuous flow pump 

collected samples at a flow rate of 7 mL min -1 . At each point, 

samples were collected into a glass bottle pre-washed with 

MilliQ water and HPLC grade acetone. The samples were 

protected from light and refrigerated during collection. For 

micropollutant analysis, 1 L was transferred into solvent

washed amber glass bottles and preserved with sodium thiosulfate. 

For the bioassays, 2 L were transferred into solvent 

washed amber glass bottles and hydrochloric acid (36%) was 

added to a final concentration of 5 mM for preservation. For 

dissolved organic carbon (DOC) measurements, 100 mL were 

filtered through a 0.45 mm PTFE membrane and collected in 

plastic (HDPE) bottles rinsed with the sample beforehand. All 

samples were transported on ice and stored frozen or at 4 C. 

2.3. Chemical analyses 

2.3.1. Dissolved organic carbon 

The dissolved organic carbon (DOC) was measured as nonpurgable 

organic carbon (NPOC) with an Analytik Jena multi 

N/C 3100 instrument at the Advanced Water Management 

Centre (AWMC). For each sample, 2d3 replicates were 

measured, giving a relative standard deviation of less than 3%. 

2.3.2. Micropollutants 

Micropollutant analysis was carried out by Queensland Health 

Forensic and Scientific Services (QHFSS, accredited by the 

National Association of Testing Authorities, Australia, and ISO 

9001 certified). The method consisted of solid phase extraction 

(SPE), concentration and quantification by liquid chromatography 

coupled with tandem mass spectrometry (LC/MSdMS). 

This method, described in details in the supporting information 

SI 1, allowed the quantification of 85 compounds (listed in 

Table SI 3 in the supporting information SI 2) selected on the 


Fig. 1 – South Caboolture Water Reclamation Plant process (adapted from van Leeuwen et al. (2003)). S1 to S7: sampling 

locations. HRT: hydraulic retention time. SRT: sludge retention time. 

basis of quantity of usage of the particular compounds, their 

potential toxicity and their resistance to degradation. Their 

limit of quantification (LOQ) was 0.01 mgL -1 in most cases. 

Concentrations were calculated using an internal calibration 

method. Results were not corrected for recovery efficiency of 

the extraction method. Recoveries of single compounds are 

assumed to be consistent along the treatment as the matrix is 

similar therefore; calculation of the removal efficiency is not 

affected. Indeed, Gros et al. (2009) determined the recoveries 

of 73 pharmaceutical compounds by SPE in raw wastewater 

and treated effluent and their results show that the recovery 

for a given compound generally vary by less than 20% from 

one matrix to the other despite the fact that these matrices are 

very different. The standard deviation of the method is 20% 

(Table SI 1 in the supporting information SI 1). When the 

reported outlet concentration was below the LOQ of the 

compound, removal efficiency was calculated as a minimum 

value using the LOQ as outlet concentration. 

2.3.3. Bioassays 

Six bioassays, discussed in details in Macova et al. (2010), were 

applied to the samples collected in the study. Baseline toxicity 

was determined with the non-specific Vibrio fischeri bioluminescence 

inhibition test. Five additional specific toxic modes 

of action were also evaluated: estrogenic activity (E-SCREEN 

assay), arylhydrocarbon receptor response (CAFLUX assay), 

neurotoxicity (acetylcholinesterase inhibition assay), phytotoxicity 

(PSII inhibition I-PAM assay) and genotoxicity (umuC

628 

assay). Water samples were extracted by SPE using Oasis HLB 

cartridges. Full dose response curves were determined for 

a serial dilution of the extract for each bioassay. Results were 

expressed as toxic equivalent concentrations (TEQ) except for 

the umuC assay. The TEQ represents the concentration of 

a given reference compound that would be required to 

produce the same effect as the mixture of the various 

compounds in the sample. When the outlet TEQ was below 

the LOQ of the bioassay, removal efficiency was calculated as 

a minimum value using the LOQ as outlet TEQ. In the umuC 

assay, the response is determined as an induction ration (IR), 

an IR 1.5 is considered genotoxic. For genotoxic samples, 

ECIR1.5 corresponds to how many times the sample must be 

concentrated or diluted to elicit an IR of 1.5. Results are 

expressed as 1/ECIR1.5 therefore a higher number represents 

a higher genotoxic effect. 


3.1. Micropollutants 

In the reclamation plant’s influent, 54 of the 85 targeted 

compounds had a median concentration above their LOQ 

confirming that conventional activated sludge treatment does 

not completely remove these micropollutants from wastewater 

(Fig. 2). The concentrations ranged from 0.01 to 

2.10 mgL -1 with the exception of gabapentin, which was 

consistently found at higher concentrations ranging from 

5.60–6.50 mgL -1 (data provided in the supporting information 

in Table SI 3). The factor between the minimum and the 

maximum concentrations measured for each individual 

compound was generally close to or lower than 2, with 

a maximum of 3.6 observed for iopromide. No clear pattern 

Fig. 2 – Number of compounds quantified and dissolved 

organic carbon (DOC) after indicated stage along the 

treatment train. Bars represent the number of compounds 

(left y-axis) with a median concentration above the LOQ 

(4 samples). Dots represent DOC on two different sampling 

days (right y-axis). Denit: denitrification; Pre-O3: preozonation; 

DAFF: dissolved air flotation-sand filtration; 

Main O3: main ozonation; AC: activated carbon; Fin O3: 

final ozonation. 


could be distinguished between the different sampling days. 

The increase or decrease of single compound concentrations 

from one day to another appeared to be random, even when 

comparing the sample collected during the weekend to 

samples collected during week days. Twenty-five compounds 

had an influent median concentration above 0.10 mgL –1 , their 

fate is further detailed hereafter. The removal efficiency of 

these compounds was determined in each treatment step 

except when the concentration before treatment was lower 

than ten times the LOQ and below LOQ after treatment. This 

criterion was used to allow the determination of removals up 

to 90% in any case. The DOC was measured for two sets of 

samples (22-07-08 and 06-08-08), and varied from 14.2 to 

19.7 mg L -1 in the influent water. The following paragraphs 

discuss the performance of each individual stage of the 

treatment by comparing concentrations of the selected 

compounds in the inlet and outlet water for the considered 

process. 

3.1.1. Denitrification 

The number of compounds above LOQ was unchanged after 

denitrification and the concentrations of the 25 selected 

compounds decreased by less than 20% (with the exception of 

atenolol, 38%). To our knowledge, the removal of micropollutants 

by denitrifying bacteria has not been specifically 

studied to date. In the present case, methanol is added to the 

water prior to the denitrification reactor and the bacteria use it 

as the primary electron donor which may prevent or reduce 

the degradation of micropollutants. The observed increase in 

DOC after the denitrification was probably due to residual 

methanol. 

3.1.2. Pre-ozonation 

The number of compounds quantified was stable through the 

pre-ozonation process (Fig. 2) and the concentration of the 25 

selected compounds decreased by less than 30%. The ozone 

dose applied in the pre-ozonation step is approximately 

2mgL -1 and the DOC was approximately 20 mg L -1 corre- 

-1 

sponding to an ozone dose of 0.1 mgO3 mgDOC. 

In such conditions, 

ozone is rapidly consumed by the organic matter (Buffle 

et al., 2006a,b) therefore micropollutants oxidation is limited. 

Ozonation has been proved to be very effective for oxidising 

various micropollutants in secondary treated wastewaters 

(Hollender et al., 2009; Huber et al., 2005; Snyder et al., 2006; 

Ternes et al., 2003; Wert et al., 2009) but with higher ozone 

-1 

doses of at least 0.25 to 0.50 mgO3 mgDOC. Snyder et al. (2006) 

also investigated a full scale wastewater treatment plant 

using pre-ozonation with a low ozone dose (not specified) 

prior to biological activated carbon filtration, and similarly, 

observed removal of micropollutants inferior to 30%. 

3.1.3. Coagulation/flocculation/DAFF 

In the coagulation/flocculation/DAFF process, the negatively 

charged colloids are neutralised by the addition of a chemical 

agent (aluminium sulphate) which causes them to aggregate 

in floccs (flocculation) which are subsequently removed by 

dissolved air flotation and sand filtration. The main purpose of 

this stage is to decrease the DOC, which was achieved effectively 

as demonstrated by the 40–50% DOC reduction (Fig. 2). 

This reduction of the DOC is important to achieve maximal

performance of the main ozonation process as discussed 

above. After the DAFF, 53 compounds were quantified and the 

concentrations of the 25 selected compounds decreased by 

less than 20% except for atenolol (42%), caffeine (29%), gabapentin 

(44%), gemfibrozil (32%) and roxithromycin (37%). It is 

also interesting to note that paracetamol concentration 

increased by 44% in the process, which could not be explained. 

No literature reporting investigations of the removal of 

micropollutants from treated wastewaters by coagulationflocculation 

could be identified. However, several studies on 

lab-scale and full scale drinking water treatment systems 

have reported removals lower than 50% for several pharmaceuticals 

and pesticides by such solids removal processes 

(Adams et al., 2002; Ternes et al., 2002; Thuy et al., 2008; Vieno 

et al., 2006; Westerhoff et al., 2005). Suarez et al. (2009) also 

showed similar limited removal in hospital wastewaters. 

Nevertheless, Suarez et al. (2009) and Westerhoff et al. (2005), 

showed that removals of up to 80% can be achieved for highly 

hydrophobic compounds (i.e logKow > 6) suggesting that the 

micropollutants removal during coagulation-flocculation 

occurs via hydrophobic interactions with neutral particles. 

Most of the pharmaceuticals and pesticides measured in this 

study are hydrophilic or of moderate hydrophobicity with 

logKow < 4(Table SI 3 in supporting information SI 2); therefore 

little removal can be expected in the coagulation/flocculation/ 

DAFF stage. Among the exceptions observed, the removal of 

gemfibrozil can be explained by its more hydrophobic nature 

(logKow ¼ 4.77, EPI SUITE 4.0) but other compounds are very 

hydrophilic and no explanation for their removal could be 

advanced. 

3.1.4. Main ozonation 

-1 

The main ozonation stage (approximately 0.5 mgO3 mgDOC) decreased the concentration of 26 compounds below LOQ 

(Fig. 2). Among the 25 selected compounds, 9 showed 


a reduction of more than 90% and 13 others were reduced by 

more than 70% while iopromide and gabapentin were reduced 

by 55% (Fig. 3). Removal of naproxen was not determined 

because its concentration was lower than ten times its LOQ 

before treatment and below LOQ after. The higher removal 

efficiency obtained by the main ozonation stage compared 

with the pre-ozonation is due to the higher ozone dosage 

relative to the DOC content. These results are in agreement 

with previous findings (Buffle et al., 2006b; Hollender et al., 

2009; Huber et al., 2005; Snyder et al., 2006; Ternes et al., 2003; 

Wert et al., 2009). The reaction of organic compounds with 

molecular ozone is selective and only certain groups of 

compounds react rapidly, e.g. aliphatic molecules with double 

bonds, deprotonated amines and aromatics with an activating 

group. Most of the pharmaceuticals and pesticides possess 

one of these moieties and would be expected to react with 

molecular ozone with high second order rate constants (kO3). 

The results show that the removal of micropollutants 

increased with increasing value of kO3 but even compounds 

weakly reactive with ozone (i.e. iopromide and MCPA; Table 1) 

were oxidised. Oxidation in ozonation processes can also 

occur via the attack of hydroxyl radicals formed by ozone 

decomposition in water. Hydroxyl radicals are highly reactive 

with almost any organic molecule; the second order rate 

constants of oxidation by hydroxyl radicals are typically in the 

10 9 M -1 s -1 range (Table 1). Therefore, the observed removals of 

iopromide and MCPA suggest a substantial exposure to 

hydroxyl radicals. Indeed, Buffle et al. (2006a,b) showed that 

the ozonation of wastewaters leads to high exposure to 

hydroxyl radicals; in the first instants of the reaction, the 

hydroxyl radical production is equivalent to, or greater than 

what can be obtained with advanced oxidation processes. To 

our knowledge, gabapentin oxidation by ozone has not been 

reported in the literature and no reaction rate data could be 

found. Gabapentin possesses both carboxylic and amine 

Fig. 3 – Median removal of selected compounds (median of influent concentration > 0.10 mgL -1 ) by the main ozonation stage 

and the combination of the main ozonation and the activated carbon (AC) filtration stages. Error bars represent minimum 

and maximum removal, no error bar means that the compound was below LOQ after treatment; therefore removal was 

calculated as a minimum using the LOQ. C S4: concentration before main ozonation; C S5: concentration after main ozonation; 

CS6: concentration after activated carbon filtration.

630 

Table 1 – Removal of selected pharmaceuticals in the main ozonation stage and their first order reaction rate constants with 

ozone (kO3) and hydroxyl radicals (kHO ). The temperature (T ) and pH given are the experimental conditions used for the 

determination of the reaction rates. 

Compound Removal kO3 (M 1 s 1 ) pH T ( C) kHO (109.M 1 s 1 ) pH T ( C) Ref. 

Iopromide 55% 91% 3.0 10 4 –1.0 10 6 

2 20 

Trimethoprim >93% 2.7 10 5 

7 20 6.5 0.2 7 25 

Sulphamethoxazole >93% ~2.5 10 6 

7 20 5.5 0.7 

5.5 10 5 

7 20 7 25 

Diclofenac >94% ~10 6 

7 20 7.5 1.5 

(1.84 0.15) 10 4 

6 25 7 25 

Paracetamol >96% 4.11 

6 k 

10 7 25 2.2 0.017 5.5 25 

Carbamazepine >98% ~3 10 5 

7 20 8.8 1.2 7 25 

(7.81 1.31) 10 4 

25 

2.05 0.14 5 room 

a Huber et al. (2003). 

b Benitez et al. (2004). 

c Benner et al. (2008). 

d Song et al. (2008). 

e Dodd et al. (2006). 

f Rivas et al. (2009). 

g Vogna et al. (2004b). 

h Andreozzi et al. (2003). 

i Andreozzi et al. (2002). 

j Vogna et al. (2004a). 

k Calculated from Andreozzi et al (2003). 

groups, their pKa are 3.89 and 9.56 respectively (Zour et al., 

1992). Therefore gabapentin is present as a zwitterion under 

usual wastewater pH conditions. Protonated amines and 

deprotonated carboxylic acids are typically not highly reactive 

with ozone (Hoigné and Bader, 1983) and gabapentin oxidation 

can be expected to be controlled by hydroxyl radical reactions. 

As a primary amine, gabapentin has an expected second order 

rate constant for the reaction with ozone of about 50 M -1 s -1 

which is significantly higher than for iopromide. However, the 

similar behaviour of the two substances shows that their 

decrease is controlled by hydroxyl radical reactions. The DOC 

decreased by less than 10% in the process, showing that 

ozonation does not lead to substantial mineralization of dissolved 

organic matter. 

3.1.5. Activated carbon filtration 

Activated carbon filtration further removed the micropollutants 

and only 2 compounds were quantified above their 

LOQ downstream: roxithromycin (0.01 mgL -1 ) and gabapentin 

(0.70 mgL -1 ). The removal efficiency of the compounds having 

a median concentration of at least ten times their LOQ prior to 

activated carbon filtration was calculated: oxazepam, tramadol 

and venlafaxine were removed by more than 90% and 

gabapentin was removed by 53%. Adsorption onto activated 

carbon also removed 20–30% of the DOC showing a less 

effective adsorption of organic matter than of micropollutants. 

Previous studies (Ormad et al., 2008; Snyder et al., 

2007; Ternes et al., 2002; Westerhoff et al., 2005) demonstrated 


that powdered and granular activated carbon can efficiently 

remove micropollutants from natural water sources used for 

drinking water. To date, the removal of micropollutants from 

wastewaters by activated carbon has not been extensively 

studied; Snyder et al. (2007) found that a water reuse facility 

using granular activated carbon without regular replacement/ 

regeneration provided little removal. In the present work, the 

activated carbon had been replaced about 4 months before the 

collection of the samples so it is assumed that its adsorption 

capacity was not yet exhausted. Adsorption onto activated 

carbon is driven mainly by hydrophobic interactions and 

Westerhoff et al. (2005) observed that the trend in the removal 

efficiency of micropollutants by activated carbon can be predicted 

from logKow values. Gabapentin is zwitterionic at 

neutral pH as shown above and much more hydrophilic 

(logKow ¼ -1.37) compared to oxazepam, tramadol and venlafaxine 

(logKow ¼ 2.32, 3.01 and 3.28 respectively). This 

explains the difference observed in the removal efficiencies of 

these compounds. It can be expected that, with time, the 

adsorption capacity of the activated carbon filter will decrease 

and eventually be exhausted while biological activity will 

develop in the filter and contribute to biodegradation of the 

by-products formed by ozonation (Simpson, 2008). 

3.1.6. Final ozonation 

In the final effluent, after final ozonation, 4 compounds had 

a median concentration above the LOQ: gabapentin (0.45 mgL -1 ), 

roxithromycin (0.01 mgL -1 ), DEET (0.03 mgL -1 ) and caffeine 

a 

b 

c 

d 

a 

e 

c 

d 

f 

e 

a 

e 

a 

g 

h 

a 

i 

j

(0.02 mgL -1 ). The median concentrations of caffeine and DEET 

are very close to the LOQ and are below the LOQ prior to the 

final ozonation stage; therefore one can assume that their 

quantification is due to analytical variations rather than an 

actual concentration variation. It is difficult to evaluate the 

efficiency of the final ozonation in the removal of micropollutants 

because they all had concentrations equal or below 

the LOQ before the treatment except gabapentin. The median 

removal of gabapentin in the final ozonation was only 20%, 

which is low compared to the main ozonation, and is due to 

-1 

the lower ozone dose employed (approximately 0.3 mgO3 mgDOC). The DOC was not removed in this treatment step. 

3.1.7. Overall treatment 

The full treatment decreased the concentration of 50 of the 54 

compounds quantified in the influent water to levels below 

LOQ (Fig. 2). Concomitantly, DOC was also reduced by 55–60% 

in the effluent compared to the influent water. Overall, among 

the 25 selected compounds, 11 were removed by more than 

95% and 11 by more than 89%. The median removal of gabapentin 

was 86% and the removals of naproxen and iopromide 

were not calculated because their concentration was lower 

than 10 times their LOQ in the influent and below their LOQ in 

the effluent. The 4 remaining compounds were gabapentin 

(0.45 mgL -1 ), roxithromycin (0.01 mgL -1 ), DEET (0.03 mgL -1 ) and 

caffeine (0.02 mgL -1 ). 

The first 3 stages of the treatment train (i.e. denitrification, 

pre-ozonation and coagulation/flocculation/DAFF) did not 

decrease the number of micropollutants quantified in the 

water (Fig. 2). After these 3 stages, the concentrations of the 25 

compounds that had an influent median concentration of at 

least 0.10 mgL -1 were generally still greater than 50% compared 

to the influent concentration (Fig. 4). The main ozonation and 


the activated carbon adsorption played a key role in the 

treatment bringing the concentration of 26 and 25 compounds 

below LOQ respectively (Fig. 2). The main ozonation generally 

decreased the micropollutants to less than 20% of their 

influent concentration and activated carbon filtration further 

removed the compounds to levels below LOQ except for 

gabapentin and roxythromycin (Fig. 4). The combined effects 

of the main ozonation and the activated carbon filtration 

processes decreased the concentration of 10 of the 25 selected 

micropollutants by more than 95% and by more than 89% for 

12 of the 15 remaining compounds compared to their 

concentration prior to the main ozonation (Fig. 3). Gabapentin 

concentration was reduced by 79%. These results show that 

ozonation followed by activated carbon filtration is a very 

effective combination of processes to remove micropollutants 

from secondary treated wastewater. The key steps in the 

removal of the DOC were the DAFF and the activated carbon 

adsorption. Although the DAFF reduced the concentration of 

micropollutants by less than 30%, it also played a key role 

indirectly by reducing the DOC which enhanced the performances 

of the main ozonation. 

3.2. Bioanalytical results 

The influent biological activity was higher than the blank 

(MilliQ water) in all the bioassays (Table 2). The effect of the 

treatment train on each bioassay is discussed individually 

hereafter. 

3.2.1. Baseline toxicity 

The Vibrio fischeri bioluminescence inhibition test is a nonspecific 

bacterial toxicity test widely recognised in the field of 

ecotoxicology as the standard assay for acute cytotoxicity. The 

Fig. 4 – Median relative concentrations of selected compounds (median of influent concentration > 0.10 g L -1 ) after the 

dissolved air flotation-sand filtration (DAFF, S4), the main ozonation (S5) and the activated carbon (S6) stages (error bars 

represent maximum and minimum values). C is the concentration after the specified treatment step and the reference 

concentration, C0, is the concentration in the influent water of the reclamation plant.

632 

Table 2 – Maximum, median and minimum biological activity in the influent and effluent of the reclamation plant and 

overall maximum, median and minimum decrease observed. 

Bioassay Result expression Blank Influent Effluent Decrease (%) 

assay reflects the general ‘‘energy status’’ of the bacteria and 

can indicate the toxic potency of a broad spectrum of 

compounds with different modes of action. Denitrification and 

pre-ozonation did have a slight stimulatory effect on the 

baseline toxicity whereas the following treatment steps were 

able to substantially decrease baseline toxicity (Fig. 5). The TEQ 

was reduced to 62, 37 and 21% of the influent water level after 

the coagulation/flocculation/DAFF, the main ozonation and 

the activated carbon filtration respectively. As is discussed in 

more details in Macova et al. (2010), an almost linear correlation 

exists between DOC level and TEQ. The slight increase in 

TEQ after denitrification cannot be related to the residual 

methanol added before this process, which is still partially 

present after denitrification as is discussed above, but is likely 

to be related to some non-volatile organic chemicals. The 52% 

decrease of TEQ in the DAFF stage is accompanied by a 40–50% 

reduction in DOC, whereas micropollutants concentrations 

were generally reduced by less than 30% in the meantime. 

Although the SPE that is performed prior to toxicity testing 

should be able to remove a substantial fraction of the DOC and 

will remove all of the residual methanol, some DOC, most 

likely smaller breakdown products that have similar 

Max Med Min Max Med Min Max Med Min 

Baseline toxicity Baseline toxicity EqC a (TEQ, mg L 1 ) 0.21 2.9 2.1 2.0 0.72 0.52 0.31 84 78 67 

Estrogenicity Estradiol EqC (EEQ, ng L 1 )

Activated carbon filtration reduced the baseline toxicity by 

50% and the DOC by 30 to 35%. Activated carbon effectively 

adsorbs the more hydrophobic compounds, which is again 

consistent with the general trend discussed above that the 

more hydrophobic compounds have a higher toxic activity 

than the more hydrophilic ones. Based on this fact, identification 

of the compounds exhibiting a high toxic activity could 

start with the identification of the more hydrophobic 

compounds. 

The final ozonation did not further reduce the baseline 

toxicity compared to after activated carbon filtration. The 

effluent TEQ was approximately 80% lower than the 

influent TEQ (Fig. 5) and only 2.5 times higher than 

the blank (Table 2). This , indicates that the residual toxicity 

is of no concern, unless the residual organic chemicals and 

organic matter inducing this effect were of very specific 

potency. This latter question was tested with a series of 

specific endpoints that respond to environmentally relevant 

modes of toxic action. 

3.2.2. Estrogenic activity 

The E-SCREEN assay specifically responds to natural 

hormones and other compounds that can mimic the activity 

of the female sex hormone estradiol. The estrogenic activity of 

the samples is expressed as an estradiol equivalent concentration 

(EEQ). The median influent EEQ was 5.8 ng L -1 ; higher 

than levels previously reported in South East Queensland. 

Most of the effluents from 12 activated sludge wastewater 

treatment plants tested by Leusch et al. (2006) had EEQs below 

4ngL -1 and sometimes below 1 ng L -1 . Hormones and endocrine 

disrupting compounds (EDCs) were not investigated in 

this study but an earlier sampling campaign of the secondary 

treated wastewater showed that the concentrations of 

measured estrogenic compounds (17b-estradiol, 17a-ethynylestradiol, 

estrone, estriol, bisphenol A, nonylphenol) were all 

below the LOQ of 1 ng L -1 (Table SI 4 in supporting information 

SI 3). This demonstrates the relevance of using bioassays as 

complementary tools to chemical analysis for the assessment 

of water quality and process performances. 

Denitrification did not affect the estrogenicity (Fig. 5). Preozonation 

with an ozone dose of approximately 0.10 mgO3 

-1 

mgDOC reduced the EEQ by 34% compared to the influent. This 

is higher than the removal previously observed by Snyder 

et al. (2006) who measured the EEQ reduction induced by 

various ozone doses in treated wastewater with a DOC of 

6.38 mg L -1 . They found that an ozone dose of 2.1 mg L -1 

1 

(0.33 mgO3 mgDOC) 

only removed 18% of the EEQ but, with 

ozone doses of 3.6 mg L -1 1 

(0.56 mgO3 mgDOC) 

and above, 90% 

or more removal could be achieved. In a recent study on a full 

scale ozonation in a Swiss sewage treatment plant (STP), the 

dose-dependency of removal of micropollutants yielded 

similar results (Escher et al., 2009). While most endpoints 

showed a clear dose-dependency of reduction of effects, the 

reduction of estrogenicity was already large at low ozone 

doses and depended more on the EEQ than on the ozone 

dose. When estrogenicity was already below a certain level, 

which was very close to the detection limit, the quantification 

of further reduction became difficult and prone to large 

uncertainty. For the remaining samples, ozone doses of 1.6– 

5.3 mg L -1 in the presence of 4.2–6.0 mg L -1 DOC lead to more 


than 90% reduction of estrogenicity. This is consistent with 

laboratory experiments that demonstrated that almost all 

first-generation transformation products of estrogenic 

chemicals had severely decreased estrogenic potency (Lee 

et al., 2008). Thus ozonation can be considered as a fairly 

selective oxidation, where even low doses selectively target 

one of the most environmentally relevant modes of toxic 

action, namely estrogenicity. 

After the coagulation/flocculation/DAFF stage the EEQ 

increased drastically by a median factor of 3.3 compared to the 

level prior to treatment. At this treatment step, the concentration 

of DOC is greatly reduced (by 40–50%), and there is 

a likelihood that the estrogenic chemicals that were bound to 

DOC were released during this treatment step. We have 

previously observed with another estrogenicity assay that 

DOC appears to reduce the bioavailability of estrogens 

(B. Escher, unpublished results). Estrogenic chemicals are 

typically relatively hydrophobic and bind well to DOC (Neale 

et al., 2008). In general DOC is not bioavailable in bioassays 

(the discussion on the small breakdown products above is an 

exception of this general paradigm) and micropollutants sorbed 

to DOC would not be bioavailable either. A large fraction of 

the matrix and also the DOC is supposed to be removed by SPE 

but, given the color of the extracts, it is possible that 

a substantial fraction of larger DOC is co-extracted. In addition, 

for the E-SCREEN test, it was demonstrated that the 

presence of serum proteins modulates the free and bioavailable 

concentration of estrogenic chemicals (Heringa et al., 

2004). This effect was also hydrophobicity dependent and was 

much more pronounced for the more hydrophobic octylphenol 

than for the less hydrophobic estradiol. Protein binding is 

generally less important than binding to DOC or lipids, 

therefore while the effect on bioavailability was not very large 

for estradiol in the study of Heringa et al. (2004); it might well 

be relevant under the conditions of the present study. This 

hypothesis needs to be evaluated in the future by exploring 

the correlation between size distribution of naturally occurring 

DOC and effect on bioavailability, estrogenicity and 

toxicity. 

The main ozonation reduced the EEQ by a median value of 

92 and 95% compared to the level of the reclamation plant’s 

influent and to the level before treatment respectively 

whereas DOC was not affected. It can be concluded that the 

mixture of by-products formed by the oxidation of the estrogenic 

compounds by ozone and hydroxyl radicals have 

a much lower estrogenic activity than the mixture of parent 

compounds, which is consistent with expectations as discussed 

above and in Lee et al. (2008). 

Activated carbon filtration was able to efficiently adsorb 

residual estrogenic compounds and further reduced the EEQ 

by another 95% to levels below the detection limit of 0.02 ng L -1 

and the final effluent concentration was below the quantification 

limit of 0.06 ng L -1 . The overall treatment efficiency for 

the removal of estrogenic activity was greater than 99%. This 

is in good agreement with observations on a full scale ozonation 

in a Swiss STP (Escher et al., 2009). As discussed above 

the analytically determined concentrations of (xeno)estrogens 

were below the quantification limit, therefore for this 

endpoint the very sensitive bioassay poses a great advantage 

despite the observed limitations due to matrix effects.

634 

3.2.3. Ah-receptor response 

The CAFLUX assay targets dioxins and dioxin-like compounds 

such as polychlorinated biphenyls (PCBs) but can also respond 

to other chemicals such as polycyclic aromatic hydrocarbons 

(PAHs) (Macova et al., 2010). The results of the test are 

expressed as 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalent 

concentration (TCDDEQ). The median TCDDEQ of the influent 

water was 0.82 ng L -1 and there was no significant variation 

along the first three steps of the treatment process; i.e denitrification, 

pre-ozonation and coagulation/flocculation/DAFF 

(Fig. 5). The main ozonation removed about 50% of the TCDDEQ 

but subsequent activated carbon filtration and final ozonation 

did not show further important removal and the median 

TCDDEQ of the final effluent was approximately 3.9 times 

higher than the blank (Table 2). Two sets of samples were 

submitted to a sulphuric acid silica gel clean up procedure that 

aims at removing organic chemicals except those that are not 

oxidised such as polychlorinated dibenzodioxins, -furans and 

PCBs. The samples were then tested again with the CAFLUX 

assay to evaluate the contribution of these very persistent 

chemicals (i.e dioxins, furans and dioxin-like PCBs). Results 

showed that after clean up the TCDDEQ was not significantly 

different from the blank (values ranged from 0.09 to 0.11 ng L -1 ). 

This shows that the effect induced by the samples without 

sulphuric acid silica gel clean up is not due to the presence of 

dioxins, furans or dioxin-like PCBs but was caused by other 

chemicals. Since none of these groups of chemicals were 

quantified by chemical analysis in this study, no comparison 

between chemical and biological analysis is possible. 

3.2.4. Genotoxicity 

The umuC assay responds specifically to genotoxic compounds 

that cause DNA damages. To detect genotoxic effects caused by 

metabolites, the test is also performed in presence of a rat liver 

extract that can transform indirect genotoxicants to metabolites 

that are DNA damaging compounds. The median influent 

1/ECIR1.5 were 0.19 and 0.060 in the absence and presence of the 

rat liver extract respectively, showing that the sample was less 

genotoxic after metabolisation. This is what one would 

commonly expect, an exception would be PAHs that are activated 

by metabolism. Denitrification and pre-ozonation did not 

have a substantial influence on genotoxicity (Fig. 5). The coagulation/flocculation/DAFF 

stage decreased 1/EC IR1.5 by 59% 

compared to the influent. The main ozonation drastically 

reduced the genotoxicity, 1/EC IR1.5 was reduced by 80 and 93% 

compared to the DAFF effluent and to the influent of the plant 

respectively. After activated carbon filtration as well as in the 

final effluent, 1/ECIR1.5 was below the LOQ of the bioassay (Table 2). 

In every case, the genotoxicity of the metabolised sample was 

lower than the non-metabolised sample, indicating that the 

type of chemicals inducing the genotoxic effect did not change 

over the treatment. 

3.2.5. Neurotoxicity 

Neurotoxicity is measured by the inhibition of the enzyme 

acetylcholinesterase (AChE). Organophosphate and carbamate 

pesticides specifically bind to this enzyme and the 

results are expressed as parathion equivalent concentration 

(PTEQ). The median PTEQ in the secondary treated wastewater 

was 3.1 mgL -1 ; denitrification and pre-ozonation did not 


reduce the PTEQ whereas DAFF decreased it by 31% compared 

to influent (Fig. 5). Unlike the other bioassays, the effect of the 

main ozonation on PTEQ was not significant but activated 

carbon filtration reduced it drastically to level below the 

quantification limit of the bioassay (0.30 mgL -1 ) which represents 

more than an 80% and 90% decrease compared to the 

main ozonation effluent and the plant influent water respectively. 

This observation is consistent with theoretical expectation, 

as it is known that compounds like diazinon and 

chlorpyrifos, which often constitute a large fraction of the 

acetylcholinesterase inhibitors, are not well oxidised by 

ozone. In contrast, these compounds are fairly hydrophobic 

(logKow ¼ 3.96 and 4.66 respectively), therefore sorption to 

activated carbon can be expected. A similar removal pattern 

has been observed for acetylcholinesterase inhibitors in the 

above-mentioned Swiss STP: none of the single removal steps 

(biological treatment, ozonation, sand filtration) had a high 

removal efficiency but all steps taken together produced 

a satisfactory overall removal (Escher et al., 2009). 

3.2.6. Phytotoxicity 

The I-PAM assay is sensitive to herbicides that directly inhibit 

photosynthesis; the results are reported as a diuron equivalent 

concentration (DEQ). The DEQ of the influent water ranged from 

0.05 to 0.22 mgL -1 with a median value of 0.10 mgL -1 (Table 2). 

Diuron concentrations were measured by chemical analysis; it 

was reported in every sample of the influent water from 0.02 to 

0.04 mgL -1 , suggesting that its contribution to the effect observed 

was limited. Among the other herbicides quantified, only 

simazine is also a photosystem II inhibitor with a relative 

potency of 0.15 (Muller et al., 2008). Simazine concentrations in 

the influent ranged from 0.05 to 0.19 mgL -1 . These two 

compounds considered together accounted for 17–93% of the 

measured DEQ demonstrating the interest of this bioassay to 

take into account non-measured compounds. The DEQ 

increased by factors of 2.2 and 3.5 after denitrification and preozonation 

respectively but variation from one day to another 

was large therefore it is difficult to draw a conclusion (Fig. 5). This 

increase was accompanied with a slight increase in baseline 

toxicity and could therefore be caused by baseline toxicants 

interfering with the measurement of the photosynthesis yield 

(Macova et al., 2010). The coagulation/filtration/DAFF stage 

reduced DEQ by 67% and 88% compared to the plant’s influent 

water and to the pre-ozonated water respectively. After the 

main ozonation the DEQ was not further significantly reduced; 

Diuron’s concentrations were equal to or below the LOQ of 

0.01 mgL -1 and simazine was removed by approximately 50%, 

their contribution accounting for 16–38% of the observed DEQ. 

Subsequent activated carbon filtration and final ozonation did 

not further affect the DEQ but diuron and simazine were 

removed below their LOQ. The overall treatment achieved 75% 

median decrease of DEQ, the effluent median DEQ was 0.03 mgL -1 

(Table 2). 

3.2.7. Overall treatment 

The biological activity of the effluent was lower compared to 

the influent and close or equal to the blank level showing that 

the treatment process could effectively decrease the biological 

adverse effects observed with the bioassays; from 62% for the 

AhR response to more than 99% for estrogenicity (Table 2). The

key treatment steps responsible for the decrease of biological 

activity are the DAFF stage, the main ozonation and the activated 

carbon adsorption (Fig. 5). 

3.3. Indirect potable reuse considerations 

Micropollutants reported concentrations were compared to the 

guidelines values for indirect potable reuse given in the 

Australian Guidelines for Water Recycling: Augmentation of 

Drinking Water Supplies (Table SI 3 in supporting information 

SI 2). The reported concentrations of the measured compounds 

were found to be below the guideline values in the influent 

water of the reclamation plant before any treatment. After 

removal by the advanced treatment process, concentrations 

were several orders of magnitude below the guideline values. 

For information purpose, median equivalent concentrations 

obtained with the bioassays were compared to the corresponding 

reference compound’s guideline value when 

available. Note however, that the effect caused by a mixture 

cannot be compared directly to a guideline value of a single 

compound. Moreover, the bioassays used here are acute tests 

and no conclusions can be drawn about chronic effects. 

Nevertheless such a comparison gives an impression on the 

expected hazard of the mixture but must be communicated 

with caution to a lay audience. For estrogenicity, neurotoxicity 

and phytotoxicity the reference compounds were estradiol, 

parathion and diuron and the guidelines values were 175 ng L -1 , 

10 mgL -1 and 30 mgL -1 respectively. Similarly to individual 

compounds concentrations, the bioassays equivalent 

concentrations were already below the guidelines values in 

the water entering the reclamation plant. Final effluent 

median equivalent concentrations were also several orders of 

magnitude below the corresponding guideline values, i.e. 

more than 2900, 33 and 428 fold for estrogenicity, neurotoxicity 

and phytotoxicity respectively. 

For the parameters considered, the water quality complies 

with the requirements of the Australian Guidelines for Water 

Recycling: Augmentation of Drinking Water Supplies. This 

suggests that such a treatment train could be considered as an 

alternative to the combination of microfiltration and reverse 

osmosis for indirect potable reuse schemes. It has the advantage 

of not producing a waste stream and would be certainly 

less energy intensive. Nevertheless, before this process can be 

recommended for indirect potable reuse, additional consideration 

needs to be given to the overall risk management 

strategies of the treatment train. Moreover, the removal of 

pathogens such as viruses and bacteria has to be assessed as 

well as the potential to form disinfection by-products due to 

the remaining DOC levels. 


The assessment of a tertiary treatment train regarding the 

removal of micropollutants and decrease of biological activity 

leads to the following conclusions: 

1) The treatment train reduced the concentration of the 54 

micropollutants quantified in the secondary treated 

wastewater as well as the biological activity observed in 


bioassays. Overall concentration reductions were typically 

higher than 90% and most of the compounds were removed 

to levels lower than 0.01 mgL 1 . The observed effects 

decreased by 62% to more than 90% depending on the assay 

and levels in the final effluent were close to or equivalent to 

the blank’s. The effect of individual processes varied from 

one compound and bioassay to another but the combination 

of the coagulation/flocculation/DAFF stage, the main 

ozonation and the activated carbon filtration was responsible 

for the major part of the observed reduction. 

2) The use of a battery of bioassays as complementary tools to 

chemical analysis yielded valuable additional information on 

the water quality and the process efficiency. The results 

showed the limitations of chemical analysis to assess potential 

biological adverse effects and the ability of bioassays to 

take into account the presence of non-measured compounds, 

formed transformation products and/or mixture effects. 

3) Concurrent measurement of DOC, micropollutants and 

biological activity showed that the DOC level can influence 

the effect observed in the bioassays. Baseline toxicity was 

shown to correlate well with the DOC levels. It is supposed 

that non-targeted organic matter is co-extracted with the 

targeted organic micropollutants in the SPE purification 

step. More understanding of the nature and quantity of 

these co-extracted compounds is needed. 

4) Ozone dose relative to DOC content is a key parameter for 

the ozonation performance. A low ozone dose of 0.1 mgO3 

-1 

mgDOC did not affect the concentration of micropollutants 

-1 

or biological activity whereas a dose of 0.5 mgO3 mgDOC was 

able to remove most of the micropollutants by more than 

70% and substantially decrease the biological activity. 

5) Degradation products are a major concern in ozonation 

processes. Here, the results from the battery of bioassays 

show that the main ozonation leads to lower baseline and 

specific toxic effects. It can be concluded that the mixture of 

degradation products formed have an overall less harmful 

potential than the mixture of parent compounds. 


This work was co-funded by the Urban Water Security 

Research Alliance under the Enhanced Treatment Project, the 

CRC Water Quality and Treatment Project No. 2.0.2.4.1.1 – 

Dissolved Organic Carbon Removal by Biological Treatment 

and by EnTox. The National Research Centre for Environmental 

Toxicology (EnTox) is a joint venture of the University 

of Queensland and Queensland Health Forensic and Scientific 

Services. The authors acknowledge the following institution 

and individuals who contributed to this study: Moreton Bay 

Water for access to the South Caboolture Water Reclamation 

Plant; Ray McSweeny and Paul McDonnell (Moretom Bay 

Water) for their help during sampling; Chris Pipe-Martin 

(Ecowise) for providing information on the South Caboolture 

Water Reclamation Plant; Geoff Eaglesham, Steve Carter and 

Mary Hodge (Queensland Health Forensic and Scientific 

Services) for the analysis of micropollutants and discussions 

on the analytical method; Wolfgang Gernjak, Michael G. 

Lawrence, Christoph Ort and Maria Jose Farre Olalla

636 

(Advanced Water Management Centre) for fruitful discussions 

and proof reading of the manuscript. 

Appendix. 



in the online version, at doi:10.1016/j.watres.2009.09.048. 

references 

Adams, C., Wang, Y., Loftin, K., Meyer, M., 2002. Removal of 

antibiotics from surface and distilled water in conventional 

water treatment processes. Journal of Environmental 

Engineering–ASCE 128 (3), 253–260. 

Andreozzi, R., Caprio, V., Marotta, R., Vogna, D., 2003. 

Paracetamol oxidation from aqueous solutions by means of 

ozonation and H2O2/UV system. Water Research 37 (5), 

993–1004. 

Andreozzi, R., Marotta, R., Pinto, G., Pollio, A., 2002. 

Carbamazepine in water: persistence in the environment, 

ozonation treatment and preliminary assessment on algal 

toxicity. Water Research 36 (11), 2869–2877. 

Benitez, F.J., Acero, J.L., Real, F.J., Roman, S., 2004. Oxidation of 

MCPA and 2,4-D by UV radiation, ozone, and the combinations 

UV/H2O2 and O-3/H2O2. Journal of Environmental Science 

and Health Part B-Pesticides Food Contaminants and 

Agricultural Wastes 39 (3), 393–409. 

Benner, J., Salhi, E., Ternes, T., von Gunten, U., 2008. Ozonation of 

reverse osmosis concentrate: Kinetics and efficiency of beta 

blocker oxidation. Water Research 42 (12), 3003–3012. 

Buffle, M.O., Schumacher, J., Meylan, S., Jekel, M., von Gunten, U., 

2006a. Ozonation and advanced oxidation of wastewater: 

effect of O-3 dose, pH, DOM and HO center dot-scavengers on 

ozone decomposition and HO center dot generation. Ozone: 

Science and Engineering 28 (4), 247–259. 

Buffle, M.O., Schumacher, J., Salhi, E., Jekel, M., von Gunten, U., 

2006b. Measurement of the initial phase of ozone decomposition 

in water and wastewater by means of a continuous quench-flow 

system: application to disinfection and pharmaceutical 

oxidation. Water Research 40 (9), 1884–1894. 

Dodd, M.C., Buffle, M.O., Von Gunten, U., 2006. Oxidation of 

antibacterial molecules by aqueous ozone: Moiety-specific 

reaction kinetics and application to ozone-based wastewater 

treatment. Environmental Science & Technology 40 (6), 

1969–1977. 

Escher, B.I., Bramaz, N., Mueller, J.F., Quayle, P., Rutishauser, S., 

Vermeirssen, E.L.M., 2008a. Toxic equivalent concentrations 

(TEQs) for baseline toxicity and specific modes of action as 

a tool to improve interpretation of ecotoxicity testing of 

environmental samples. Journal of Environmental Monitoring 

10 (5), 612–621. 

Escher, B.I., Bramaz, N., Ort, C., 2009. Monitoring the treatment 

efficiency of a full scale ozonation on a sewage treatment 

plant with a mode-of-action based test battery. Journal of 

Environmental Monitoring 11 (10), 1836–1846. 

Escher, B.I., Bramaz, N., Quayle, P., Rutishauser, S., 

Vermeirssen, E.L.M., 2008b. Monitoring of the ecotoxicological 

hazard potential by polar organic micropollutants in sewage 

treatment plants and surface waters using a mode-of-action 

based test battery. Journal of Environmental Monitoring 10 (5), 

622–631. 

Esplugas, S., Bila, D.M., Krause, L.G.T., Dezotti, M., 2007. 

Ozonation and advanced oxidation technologies to remove 


endocrine disrupting chemicals (EDCs) and pharmaceuticals 

and personal care products (PPCPs) in water effluents. Journal 

of Hazardous Materials 149 (3), 631–642. 

Gros, M., Petrovic, M., Barcelo, D., 2009. Tracing pharmaceutical 

residues of different therapeutic classes in environmental 

waters by using liquid chromatography/quadrupole-linear ion 

trap mass spectrometry and automated library searching. 

Analytical Chemistry 81 (3), 898–912. 

Heringa, M.B., Schreurs, R., Busser, F., Van Der Saag, P.T., Van Der 

Burg, B., Hermens, J.L.M., 2004. Toward more useful in vitro 

toxicity data with measured free concentrations. 


Hoigné, J., Bader, H., 1983. Rate constants of reactions of ozone with 

organic and inorganic compounds in water–II: dissociating 

organic compounds. Water Research 17 (2), 185–194. 




wastewater treatment plant upgraded with a full-scale post- 

Ozonation followed by sand filtration. Environmental Science 

and Technology 43 (20), 7862–7869. 

Huber, M.M., Canonica, S., Park, G.Y., von Gunten, U., 2003. 


oxidation processes. Environmental Science and Technology 

37 (5), 1016–1024. 






Joss,A.,Siegrist,H.,Ternes,T.A.,2008.Areweaboutto 

upgrade wastewater treatment for removing organic 

micropollutants? Environmental Science and Technology 

57 (2), 251–255. 

Kim, I.H., Tanaka, H., Iwasaki, T., Takubo, T., Morioka, T., Kato, Y., 

2008. Classification of the degradability of 30 pharmaceuticals 

in water with ozone, UV and H2O2. Water Science and 

Technology 57 (2), 195–200. 





Lee, Y., Escher, B.I., Von Gunten, U., 2008. Efficient removal of 


containing steroid estrogens. Environmental Science and 

Technology 42 (17), 6333–6339. 

Leusch, F.D.L., Chapman, H.F., van den Heuvel, M.R., Tan, B.L.L., 

Gooneratne, S.R., Tremblay, L.A., 2006. Bioassay-derived 

androgenic and estrogenic activity in municipal sewage in 

Australia and New Zealand. Ecotoxicology and Environmental 

Safety 65 (3), 403–411. 

Macova, M., Escher, B.I., Reungoat, J., Carswell, S., Lee Chue, K., 

Keller, J., Mueller, J.F., 2010. Monitoring the biological activity 

of micropollutants during enhanced wastewater treatment 

with ozonation and activated carbon filtration. Water 

Research 44 (2), 477–492. 

Muller, R., Schreiber, U., Escher, B.I., Quayle, P., Nash, S.M.B., 

Mueller, J.F., 2008. Rapid exposure assessment of PSII 

herbicides in surface water using a novel chlorophyll 

a fluorescence imaging assay. Science of the Total 

Environment 401 (1–3), 51–59. 

Muller, R., Tang, J.Y.M., Thierb, R., Mueller, J.F., 2007. Combining 

passive sampling and toxicity testing for evaluation of mixtures 

of polar organic chemicals in sewage treatment plant effluent. 

Journal of Environmental Monitoring 9 (1), 104–109. 



pharmaceuticals and personal care products (PPCPs) and



Research 41 (19), 4373–4382. 

Neale, P.A., Escher, B.I., Schafer, A.I., 2008. Quantification of 

solute–solute interactions using negligible-depletion solidphase 

microextraction: measuring the affinity of estradiol to 

bulk organic matter. Environmental Science and Technology 

42 (8), 2886–2892. 

Nowotny, N., Epp, B., vonSonntag, C., Fahlenkamp, H., 2007. 

Quantification and modeling of the elimination behavior of 

ecologically problematic wastewater micropollutants by 

adsorption on powdered and granulated activated carbon. 


Onesios, K.M., Yu, J.T., Bouwer, E.J., 2009. Biodegradation and 

removal of pharmaceuticals and personal care products in 

treatment systems: a review. Biodegradation 20 (4), 441–466. 

Ormad, M.P., Miguel, N., Claver, A., Matesanz, J.M., Ovelleiro, J.L., 

2008. Pesticides removal in the process of drinking water 

production. Chemosphere 71 (1), 97–106. 

Rivas, J., Gimeno, O., Encinas, A., Beltrán, F., 2009. Ozonation of 

the pharmaceutical compound ranitidine: Reactivity and 

kinetic aspects. Chemosphere 76 (5), 651–656. 

Simpson, D.R., 2008. Biofilm processes in biologically active 

carbon water purification. Water Research 42 (12), 2839–2848. 

Snyder, S.A., Adham, S., Redding, A.M., Cannon, F.S., DeCarolis, J., 

Oppenheimer, J., Wert, E.C., Yoon, Y., 2007. Role of membranes 

and activated carbon in the removal of endocrine disruptors 

and pharmaceuticals. Desalination 202 (1–3), 156–181. 

Snyder, S.A., Wert, E.C., Rexing, D.J., Zegers, R.E., Drury, D.D., 

2006. Ozone Oxidation of Endocrine Disruptors and 

Pharmaceuticals in Surface Water and Wastewater. Ozone: 

Science and Engineering 28 (6), 445–460. 

Song, W.H., Cooper, W.J., Mezyk, S.P., Greaves, J., Peake, B.M., 

2008. Free radical destruction of beta-blockers in aqueous 

solution. Environmental Science & Technology 42 (4), 

1256–1261. 

Suarez, S., Lerna, J.M., Omil, F., 2009. Pre-treatment of hospital 

wastewater by coagulation-flocculation and flotation. 

Bioresource Technology 100 (7), 2138–2146. 

Ternes, T.A., Kreckel, P., Mueller, J., 1999. Behaviour and 

occurrence of estrogens in municipal sewage treatment plants – 

II. Aerobic batch experiments with activated sludge. Science of 

the Total Environment 225 (1–2), 91–99. 

Ternes,T.A.,Meisenheimer,M.,McDowell,D.,Sacher,F.,Brauch,H.J., 

Gulde, B.H., Preuss, G., Wilme, U., Seibert, N.Z., 2002. Removal of 


pharmaceuticals during drinking water treatment. 


Ternes, T.A., Stuber, J., Herrmann, N., McDowell, D., Ried, A., 


of pharmaceuticals, contrast media and musk fragrances from 

wastewater? Water Research 37 (8), 1976–1982. 

Thuy, P.T., Moons, K., van Dijk, J.C., Anh, N.V., Van der Bruggen, B., 

2008. To what extent are pesticides removed from surface 

water during coagulation-flocculation? Water and Environment 

Journal 22 (3), 217–223. 

van Leeuwen, J., Pipe-Martin, C., Lehmann, R.M., 2003. Water 

reclamation at South Caboolture, Queensland, Australia. 

Ozone: Science and Engineering 25 (2), 107–120. 

Vieno, N., Tuhkanen, T., Kronberg, L., 2006. Removal of 

pharmaceuticals in drinking water treatment: effect of chemical 

coagulation. Environmental Technology 27 (2), 183–192. 

Vogna, D., Marotta, R., Andreozzi, R., Napolitano, A., d’Ischia, M., 

2004a. Kinetic and chemical assessment of the UV/H2O2 

treatment of antiepileptic drug carbamazepine. Chemosphere 

54 (4), 497–505. 

Vogna, D., Marotta, R., Napolitano, A., Andreozzi, R., d’Ischia, M., 

2004b. Advanced oxidation of the pharmaceutical drug 

diclofenac with UV/H2O2 and ozone. Water Research 38 (2), 

414–422. 

Wert, E.C., Rosario-Ortiz, F.L., Snyder, S.A., 2009. Effect of ozone 

exposure on the oxidation of trace organic contaminants in 





processes. Environmental Science and Technology 39 (17), 

6649–6663. 

Yoon, Y., Westerhoff, P., Snyder, S.A., Wert, E.C., Yoon, J., 2007. 

Removal of endocrine disrupting compounds and 

pharmaceuticals by nanofiltration and ultrafiltration 

membranes. Desalination 202 (1–3), 16–23. 

Yu, Z.R., Peldszus, S., Huck, P.M., 2008. Adsorption characteristics 

of selected pharmaceuticals and an endocrine disrupting 

compound – naproxen, carbamazepine and nonylphenol – on 

activated carbon. Water Research 42 (12), 2873–2882. 

Zour, E., Lodhi, S.A., Nesbitt, R.U., Silbering, S.B., Chaturvedi, P.R., 

1992. Stability studies of gabapentin in aqueous-solutions. 

Pharmaceutical Research 9 (5), 595–600. 

Zwiener, C., Frimmel, F.H., 2000. Oxidative treatment of 

pharmaceuticals in water. Water Research 34 (6), 1881–1885.

Removal of natural hormone estrone from secondary effluents 

using nanofiltration and reverse osmosis 

Xue Jin 1 , Jiangyong Hu*, Say Leong Ong 

Division of Environmental Science and Engineering, National University of Singapore, 10 Kent Ridge Crescent, 119260, Singapore 

article info 


Received 19 December 2008 


1 June 2009 



Keywords: 

Estrone 

Nanofiltration 

Reverse osmosis 

Rejection 

Treated effluent 

Binding 


abstract 

Recently, the use of treated municipal wastewater for 

groundwater recharge and indirect drinking water reuse has 

become promising worldwide. Nanofiltration (NF) and reverse 

osmosis (RO) membranes are widely used in this kind of 

environmental application (Schäfer et al., 2005). However, the 

presence of endocrine disrupting chemicals (EDCs) in the 

aquatic environment has raised great consumer concern due 

to their potential health risk (Tim, 1997; Adams, 1998; Croley 




The rejection of steroid hormone estrone by nanofiltration (NF) and reverse osmosis (RO) 

membranes in treated sewage effluent was investigated. Four NF/RO membranes with 

different materials and interfacial characteristics were utilized. To better understand 

hormone removal mechanisms in treated effluent, effluent organic matters (EfOM) were 

fractionated using column chromatographic method with resins XAD-8, AG MP-50 and IRA- 

96. The results indicate that the presence of EfOM in feed solution could enhance estrone 

rejection significantly. Hydrophobic acid (HpoA) organic fraction made a crucial contribution 

to this ‘‘enhancement effect’’. Hydrophobic base (HpoB) could also improve estrone 

rejection while hydrophobic neutral (HpoN) and hydrophilic acid (HpiA) with low aromaticity 

had little effects. The increment in estrone rejection was predominantly attributed to 

the binding of estrone by EfOM in feed solutions, which led to an increase in molecular 

weight and appearance of negative charge (for the HpoA case) and thus an increased level 

of estrone rejection. However, the improvement of estrone rejection by HpoA decreased 

with increasing calcium ion concentration. The important conclusion of this study is, first, 

hydrophobic acid macromolecules are recommended to be added into feed water to 

improve the rejection of trace hormone during NF/RO membrane process, and, second, 

removal of calcium ions via pretreatment and application of membrane with more negative 

charge at its interface can greatly intensify this ‘‘enhancement effect’’. 


et al., 2000; Baker, 2001; Iguchi et al., 2001). Amongst many 

kinds of EDCs, the impact of steroid hormones such as 

estrone, estradiol and ethinylestradiol is prominent because 

their potency can be several thousand times higher than that 

of others (Desbrow et al., 1998; Ternes et al., 1999; Johnson and 

Sumpter, 2001). It is noteworthy to emphasize that concentrations 

of some steroid hormones in secondary effluent are 

often high enough to harm wildlife although it is within ng/L 

range (Baronti et al., 2000; Williams et al., 2003; Soliman et al., 

2004). In the light of the magnitude of this problem, NF/RO 


E-mail addresses: jinxuesky@ucla.edu (X. Jin), esehujy@nus.edu.sg (J. Hu), cveongsl@nus.edu.sg (S.L. Ong). 

1 

Current address: Civil & Environmental Engineering Department, University of California, 5732 Boelter Hall, Los Angeles, CA 90095- 

1593, USA 


doi:10.1016/j.watres.2009.09.057

membranes and different post-treatment technologies are 

essential to remove EDCs from treated effluent before water 

reuse and ensure the safety of drinking water. 

Removal of EDCs using NF/RO membranes in single-organic 

solution (the experiments were conducted with target EDCs 

spiked in synthetic model solutions in the absence of other 

organics) has been reported in several recent studies (Kimura 

et al., 2003, 2004; Schäfer et al., 2003; Nghiem et al., 2004a, b). 

Many factors may influence the interaction between EDCs and 

membrane and thus their rejection including physicochemical 

properties of EDCs, such as molecular weight (MW), molecular 

size, charge (as a function of dissociation constant), hydrophobicity 

and polarity; membrane properties, such as material, 

surface charge (zeta potential), hydrophobicity, molecular 

weight cut-off (MWCO), porosity and desalting degree; and 

solution chemistry, such as pH, ionic strength and presence of 

divalent cations (Bellona et al., 2004, Schäfer et al., 2005). 

Another important factor influencing the rejection efficiency is 

initial concentration of target EDCs in feed solution. Kimura 

et al. (2003) reported that experiments conducted at a ng/l 

concentration range resulted in a lower rejection efficiency as 

compared to experiments conducted at a mg/l range. This 

observation substantiates that experiments conducted within 

the ‘‘realistic’’ concentration range are necessary to provide an 

accurate value of rejection efficiency. It should, however, be 

pointed out that the results achieved from a single-organic 

solution cannot be extrapolated to the mixtures typically 

encountered in a real aquatic environment. 

In previous works, we have investigated the influence of 

other dissolved organic matter (DOM) in feed solution on the 

removal of steroid hormone estrone during NF processes (Jin 

et al., 2007). The results indicated that estrone rejection depended 

on the type of DOM co-present. Neutral and highly hydrophilic 

dextran had little effect on estrone rejection. The presence 

of humic acid without phenolic groups but great aromaticity 

improved estrone rejection slightly. In contrast, estrone rejection 

was obviously increased by the presence of hydrophobic 

acid fraction (HpoA) derived from secondary effluent. Therefore, 

it is expected that the rejection mechanism of EDCs in real water 

matrix, such as secondary effluent, is very complicated. 

Organic matter in secondary effluent is referred to as 

effluent organic matter (EfOM) and can be categorized into 

fractions based on the difference in hydrophobicity and 

charge properties using a column chromatographic method 

(Leenheer, 1981; Thurman and Malcolm, 1981; Imai et al., 2002; 

Hu et al., 2003). Adsorption of some fractions in EfOM onto 

membranes would be expected to alter the membrane surface 

properties, which subsequently results in changes in the 

rejection of EDCs (Ng and Elimelech, 2004; Yoon et al., 2004; Xu 

et al., 2006; Steinle-Darling et al., 2007). Alternatively, EDCs 

can bind to some fractions in EfOM and be retained together, 

which can also affect their rejection characteristics (Agbekodo 

et al., 1996; Devitt et al., 1998; Majewska-Nowak et al., 2002; 

Nghiem et al., 2004a, b; Hu et al., 2007; Jin et al., 2007). Both 

interactions mentioned above (EfOM-membrane and EfOM- 

EDCs) are markedly influenced by the physicochemical characteristics 

of EfOM. Therefore, different fractions of EfOM 

would demonstrate diverse impacts on EDCs rejection by 

membranes. To our knowledge, however, very little information 

is available on the rejection of trace EDCs by NF/RO 


membranes with the presence of different fractions of EfOM. 

In order to better understand the removal mechanism of 

EDCs in secondary effluent by NF/RO membranes and optimize 

membrane performance, it is important to clarify this 

issue. 

It is known that divalent cations (especially calcium) affect 

the characteristics of EfOM, membrane charges and thus the 

two above-mentioned interactions (EfOM–membrane and 

EfOM–EDCs). For example, Devitt et al. (1998) reported that 

atrazine–NOM association decreased in the presence of 

calcium. Therefore, calcium ion is predicted to affect the 

rejection of EDCs with the presence of other EfOM. However, 

to date, the intricate relationship between physicochemical 

characteristics of EfOM, calcium ion concentration and 

membrane performance on EDC rejection remains poorly 

understood. In order to gain insight into the removal mechanism 

of EDCs in treated effluent, more research effort is 

needed to better understand the above relationship. 

In this study, we investigated the removal of steroid hormone 

estrone from secondary effluent using a bench-scale cross-flow 

NF/RO system. We examined the effect of various organic fractions 

isolated from secondary effluent and membrane characteristics 

on the removal of estrone by this process. After 

determining the key organic fraction which makes a crucial 

contribution to the enhanced estrone rejection by EfOM, the 

influence of calcium ion concentration on estrone rejection with 

the presence of the key fraction was investigated. 


2.1. NF/RO membranes and characterization 

Four NF/RO membranes, denoted as DL, CK, AK and CG (Sepa 

membranes, OSMONICS, USA), were used. DL and AK are polyamide 

(PA) composite membranes. CK and CG are made of 

cellulose acetate (CA) polymer. New membranes were soaked in 

ultrapure water for 24 h with the water being replaced regularly 

prior to use. Membrane properties were characterized as 

described elsewhere (Jin et al., 2007) and are shown in Table 1. 

Briefly, contact angle was determined with ultrapure water at 

pH 5.8 0.2 using a goniometer (Ramé-hart contact angle 

goniometer, Imaging System, Mountain Lakes, NJ, USA). Zeta 

potential of the membrane surface was measured in the electrolyte 

background solution (1 mM NaHCO3 and 8 mM NaCl, pH 

7) using an Electro Kinetic Analyzer (EKA) from Anton Paar 

(Austria). To characterize the membrane pore size, MWCO was 

estimated by filtering the electrolyte background solution containing 

polyethylene glycol (PEG) with MW of 200, 400, 600 and 

1000 Da, purchased from Fluka (Switzerland). Filtration tests 

were performed using the same cross-flow membrane test unit 

as used in the estrone filtration tests with clean membrane. 

A PEG concentration of approximately 6 ppm as dissolved 

organic carbon (DOC) was used for each measurement. 

2.2. Membrane filtration unit 

A laboratory-scale cross-flow membrane filtration unit was 

used (Fig. S1). Membrane coupon (13 cm by 13 cm) was housed 

in a stainless steel membrane cell. Feed solution was fed from

640 

Table 1 – Membrane properties. 

Membrane Membrane type Contact angle ( ) Zeta potential a 

a reservoir by a pump, capable of providing a maximum 

pressure of 4137 kPa and a maximum flow of 20 l/min. The 

temperature inside the feed reservoir was maintained 

constant (22 0.1 C) by a recirculating chiller. The pressure 

(1379 kPa with DL, CK and AK, and 2758 kPa with CG) inside 

the membrane cell was set with a back-pressure regulator and 

a bypass valve. 

Each experiment was conducted in two steps: (1) the 

membrane was compacted with ultrapure water until the 

water flux remained constant; (2) the reservoir was emptied 

and filled with test solution. The solution was then filtered at 

constant transmembrane pressure. All experiments were 

conducted for 24 h. During filtration, both permeate and 

concentrate were recirculated back to the feed tank. At specified 

intervals, permeate and feed samples were collected for 

analysis. Rejection (R) is defined as R ¼ 100 ð1 Cp=C fÞ, 

where C p and C f are the permeate and feed concentrations, 

respectively. The effect of EfOM on estrone rejection was 

determined based on the rejection measured at the end of 

filtration where stable rejection was achieved. 

2.3. EfOM fractionation method 

Secondary effluent was collected from a local municipal 

wastewater treatment plant. The water was first filtered through 

amicrofiltration(MF)unit(0.05mm, Desal/Osmonics, USA) to 

simulate pretreatment ahead of NF/RO processes in the real 

plant, which is commonly used to remove suspended solids and 

colloids. The total organic carbon (TOC) and total dissolved solid 

(TDS) of the water samples ranged from 9.2 to 10.8 ppm and 455 

to 618 ppm, respectively. Other characteristics of the MF treated 

secondary effluent are summarized in Table S1. 

The filtrates were subsequently fractionated into six 

subcomponents: HpoA, hydrophobic bases (HpoB), hydrophobic 

neutrals (HpoN), hydrophilic acids (HpiA), hydrophilic 

bases (HpiB) and hydrophilic neutrals (HpiN) using a method 

developed by Leenheer (1981). SupeliteÔ XAD-8 resin 

(SUPELCO, USA.), AG-MP-50 cation exchange resin (BioRad, 

USA) and Amberlite IRA-96 anion exchange resin (Rohm and 

Haas, France) were the resins used. Details on the instruments 

and fractionation methods have been published previously 

(Hu et al., 2003). The DOC fractionation results for the 

secondary effluent are presented in Table 2. HpoA predominated 

in secondary effluent and accounted for 36.6–40.3% of 

the total amount of dissolved organic matter. These results do 

not agree with an earlier study by Imai et al. (2002) who found 

different results on the fractions with HpiA as the most 


MWCO (Da) b 

abundant fraction for the samples taken from Japan. Thus, 

DOM composition in secondary effluent varies with the origin 

of wastewater and the type of treatment process used. In this 

study, the experimental results obtained in MF treated 

secondary effluent should be analyzed based on its specific 

DOM-fraction distribution as reported here. 

2.4. Chemical and solution chemistry 

Typical flux/pressure 

(m/s per Pa) c 

DL Polyamide NF 30.7 2.7 16.3 2.7 490 1:46 10 5 =6:89 10 5 

CK Cellulose acetate NF 54.2 1.4 6.0 0.3 560 1:32 10 5 =1:52 10 6 

AK Polyamide RO 41.7 1.3 19.9 0.2

In all feed waters, initial estrone concentration was maintained 

at 100 ng/L. Solution with different calcium ion 

concentrations (0, 0.3 and 0.6 mM) were accomplished by 

adding CaCl2 stock solution, while the total ionic strength was 

kept constant by adjusting NaCl concentration. 


2.5.1. Estrone detection 

Estrone was extracted from aqueous sample by OASIS HLB 

cartridge (500 mg) from Waters, Singapore. Before the analytes 

were extracted, 80 ng/l of E1-d4 was added to the sample 

water to increase accuracy of the analytical data. The 

cartridge was subsequently conditioned by 6 ml of diethylether, 

5 ml of methanol and 10 ml of ultrapure water. The 

sample water (400 ml) was loaded through the cartridge at 

flow rate of 3 ml/min with the aid of a vacuum pump. After 

sample loading, the cartridge was eluted with 10 ml of 

a methanol–diethylether solution (10:90, v/v). The eluant was 

collected in a brown glass vial and was allowed to dry under 

a gentle stream of nitrogen. After drying, the residuals were 

re-dissolved in 0.4 ml of methanol and then analyzed by liquid 

chromatography–tandem mass spectrometry (MDS Sciex API 

2000 tandem triple-quadropole mass spectrometer) from 

Applied Biosystems, USA. Analytes were chromatographed on 

a 2.1 150 mm column filled with 3.5 mm C18 reversed phase 

packing (Agilent, Singapore). The multiple reaction monitoring 

(MRM) mode was chosen for quantification. After 

observing collision-induced dissociation (CID) spectra 

obtained by full-scan production experiments, the following 

MRM pairs were chosen: E1: 269.2/145.0; 269.2/143.0; E1-d4: 

273.0/146.9. Estrone recovery was not matrix-dependent due 

to using internal standard. The mean recovery and relative 

standard deviation (RSD) was 99 7.4% (n ¼ 16). The limit of 

detection was 0.5 ng/l, calculated using a signal-to-noise ratio 

of 3. Details with respect to basic information on this method 

are provided separately (Jin, 2007). 

2.5.2. TOC and UV 254 analysis 

In all cases, the concentrations of EfOM were detected by 

a TOC analyzer (Shimadzu TOC-Vcsh, Japan). TOC and DOC 

are used interchangeably in this study because secondary 

effluent was filtered through 0.05 mm membrane before using. 

The specific ultraviolet absorption of selected organic fraction 

at 254 nm (SUVA254) was calculated by dividing the absorbance 

at 254 nm measured using the Shimadzu UV-1700 

UV-Visible Spectrophotometer (Japan) by the DOC level of the 

solution. 

2.5.3. Molecular Weight Measurement 

High performance size exclusion chromatography (HPSEC) 

was performed to determine both weight- and number-averaged 

molecular weights. Instrumentation consisted of a SHI- 

MADZU (Japan) HPLC instrument LC-10AT VP with a Polymer 

Laboratories aquagel-OH column and a SHIMADZU SPD-M10A 

diode array detector operating at 260 cm. The HPSEC system 

was calibrated using polystyrene sulfonates standards of 

known average MW between 210 and 13000, and acetone. 

Mobile phase was composed of ultrapure water buffered with 

phosphate to pH of 6.8. 



3.1. Rejection of estrone by NF/RO membranes in 

electrolyte background solution 

Filtration of estrone in electrolyte background solution (1 mM 

NaHCO 3 and 8 mM NaCl, pH 7) provides the baseline from 

which the effect of other organic matter on estrone rejection 

can be found. The open symbols in Fig. 1 present the observed 

rejections of estrone in electrolyte background solution by 

four kinds of NF/RO membranes. It appeared that estrone 

rejection was initially higher than 90% and decreased 

dramatically and then stabilized at later filtration stage. 

The excellent removal performances of all membranes at 

the initial filtration stage can be attributed to their adsorption 

capabilities and steric hindrance. However, the adsorption 

effect can only contribute to the short-term removal of 

estrone. As the feed solution is continuously filtered through 

the membrane, more and more available sites on the 

membrane are occupied by adsorbed estrone. When the 

partition of estrone between feed solution and membrane 

reaches equilibrium, there is no further net adsorption effect 

taking place and thus the contribution from adsorption would 

be negligible. Under this condition, size exclusion would 

become the overriding removal mechanism at the later 

filtration stage. Consistent with the MWCO values of the 

membranes, the ultimate rejection of estrone followed the 

order: AK > CG > DL > CK. AK membrane with MWCO value 

below the MW of estrone can remove estrone efficiently. 

However, for the other three kinds of membranes with MWCO 

values larger than the MW of estrone, ultimate rejection was 

extremely low. To avoid overestimation of estrone rejection, 

in this paper, all of the determined rejections were based on 

23 h of filtration when stable rejection was achieved. 

3.2. Rejection of estrone by NF/RO membranes in 

secondary effluent matrix 

Estrone rejection in MF treated secondary effluent, compared 

with that in electrolyte background solution, using four kinds 

of NF/RO membranes was monitored and the results are presented 

in Fig. 1, whereas the corresponding DOC rejection in 

secondary effluent is presented in Fig. 2. Estrone removal 

efficiencies in MF treated secondary effluent matrix for all 

NF/RO membranes used were consistently higher than the 

results obtained in electrolyte background solution. The 

improvements in ultimate estrone rejection were 6.5%, 32.5%, 

21.6% and 11.6% for AK, DL, CK and CG membranes, respectively. 

The results obtained here are consistent with an earlier 

study by Comerton et al. (2008), who investigated the influence 

of water matrix on EDC rejection by RO and NF membranes. 

They reported a significant increase in EDCs rejection by tight 

NF membranes in membrane bioreactor effluent and natural 

waters when compared to the ultrapure water. 

The improvement in estrone rejection could be due to the 

following reasons. Firstly, permeate flux decline was observed in 

MF treated secondary effluent matrix during separation 

processes (in electrolyte background solution, reduction in 

permeate flux was not observed for any membrane tested,

642 

a b 

Estrone Rejection (%) 


100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

20 

0 

Electrolyte background 

Secondary effluent 

0 5 10 15 20 25 

Filtration Time (h) 

which can be attributed to the very low concentration of estrone 

in the feed solution). After 23 h of filtration, the flux decreased by 

about 13.0%, 8.9%, 33.8% and 17.5% for DL, CK, AK and CG 

membrane, respectively. This suggests that the membranes 

were fouled by EfOM. The entrapment of EfOM onto the 

membrane active layer restricted the passage of contaminants, 

resulting in an increase in the observed rejection of both DOC 

(Fig. 2) and estrone at the ng/l level (Fig. 1). Similar increases in 

trace organic rejection as a result of organic fouling have been 

reported in previous studies (Agenson and Urase, 2007; Bellona 

and Drewes, 2007; Nghiem and Hawkes, 2007). However, simple 

EfOM fouling may not be the main reason, because the sequence 


c d 



0 5 10 15 20 25 



100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

20 



0 5 10 15 20 25 




0 

0 5 10 15 20 25 


Fig. 1 – Estrone rejection in electrolyte background solution and MF treated secondary effluent by (a) DL, (b) CK, (c) AK and (d) CG. 

DOC Rejection (%) 

100 

95 

90 

85 

80 

75 

70 

65 

60 

DL 

CK 

AK 

CG 

0 5 10 15 20 25 


Fig. 2 – DOC rejection in MF filtered secondary effluent by 

four kinds of NF/RO membranes. 


of the improvement in estrone rejection (DL > CK > CG > AK) 

was mismatched with the sequence based on permeate flux 

decline in secondary effluent matrix (AK > CG > DL > CK). 

Therefore the enhanced estrone rejection can be mainly attributed 

to the other reason, namely that, secondly, estrone molecules 

may bind to some fractions of EfOM in bulk solution and be 

retained together, resulting in a reduction in the passage of 

estrone molecules through membrane. The remainder of this 

paper is focused on proving this hypothesis. 

3.3. Rejection of estrone in the presence of specific 

organic fractions at same TOC level 

In order to optimize membrane performance, it is essential to 

identify which organic fraction in secondary effluent would 

make a crucial contribution to the ‘‘enhancement effect’’ on 

estrone removal. Amongst the six fractions derived from 

EfOM, HpiB was present in insignificant quantities and thus 

was expected to have a negligible influence on estrone 

removal. HpiN, as the last fraction derived from the resin 

fractionation system, inevitably contains some impurities of 

other fractions (Hu et al., 2003). Moreover, as shown in our 

previous work (Jin et al., 2007), dextran representing hydrophilic 

neutral organics had little effect on estrone rejection. 

Therefore, in this study, the research was focused on the 

influences of HpoA, HpoB, HpoN and HpiA on estrone rejection 

by NF/RO membranes. 

As shown earlier, DL and CK membranes demonstrated 

a greater improvement in estrone rejection in MF treated 

secondary effluent compared to AK and CG membranes. 

Investigations were firstly conducted to compare the capability 

of each isolated fraction to improve estrone removal

a 


b 


100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

20 

0 


With HpoA 

With HpoB 

With HpoN 

With HpiA 

0 5 10 15 20 25 



With HpoA 

With HpoB 

With HpoN 

With HpiA 

0 5 10 15 20 25 


Fig. 3 – Observed rejection of estrone with the presence of 

different organic fractions at same DOC level (1 ppm) by (a) 

DL and (b) CK. 

using DL and CK membranes. To avoid the influence of DOC 

values, the initial concentration of each organic fraction was 

adjusted to a DOC value of approximately 1 ppm, which was 

close to the original concentrations of HpoB, HpoN and HpiA 

present in secondary effluent (Table 2). 

Estrone rejection in the presence of each fraction over 

a 24 h filtration period is presented in Fig. 3. It is evident that 

estrone rejection was significantly improved by the presence 

of HpoA, rising from 15.3% to 41.9% with DL membrane and 

from 8.2% to 15.3% with CK membrane. The presence of HpoB 

also resulted in an improvement in estrone rejection of as 

high as 3.6% and 4.1% with DL and CK, respectively. In 

contrast, HpoN and HpiA did not exhibit any obvious increase 

in estrone rejection for any of the membranes tested, 

although HpoN and HpiA were removed more efficiently than 

HpoB (Fig. 4). The differences in the ‘‘enhancement effect’’ 

between HpoA and HpiA indicate that hydrophobic organic 

fractions have a greater contribution to the enhanced estrone 

rejection in secondary effluent than hydrophilic organic 

fractions. 

In all experiments, no obvious permeate flux decline (

644 

a 


b 


100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

0 5 10 15 20 25 


HpoA 

HpoB 

HpoN 

HpiA 

HpoA 

HpoB 

HpoN 

HpiA 

0 5 10 15 20 25 


Fig. 4 – Observed rejection of bulk organics (as DOC) with 

the presence of different organic fractions at same DOC 

level (1 ppm) by (a) DL and (b) CK. 

molecules can bind to bulk DOM in feed matrix and subsequently 

be retained together. In this case, the ‘‘enhancement 

effect’’ on estrone removal was controlled not only by the 

binding between estrone and DOM in feed solution, but also by 

the rejection efficiency of the DOM with which estrone 

molecules are associated. Therefore, the weak effect of HpoB 

on estrone rejection is due to the inefficient removal of HpoB 

by CK membrane (Fig. 7b). It is apparent that the HpoA fraction, 

for all membranes used, had a significantly greater 

‘‘enhancement effect’’ than the other two hydrophobic fractions 

on ultimate estrone removal. After 23 h of filtration, the 

presence of HpoA increased estrone rejection by 29.8%, 13.3% 

and 20.5% for DL, CK and CG membranes, respectively. This 

could be attributed to a combination of the following reasons. 

Firstly, the structural characteristics of HpoA (phenolic group) 


Fig. 5 – Schematic of possible hydrogen bonding between 

phenolic groups of estrone molecule and HpoA and 

between phenolic group in estrone molecule and aromatic 

amine group in HpoB (A–E stand for any possible functional 

groups within HpoB fraction). 

allowed estrone molecules to bind with HpoA in feed solution 

through hydrogen bonding, thus estrone could be removed 

together with HpoA. Secondly, HpoA with the highest 

concentration (Table 2) in feed solution would have the 

highest chance to associate with estrone molecules. Thirdly, 

HpoA with the largest MW of more than 4000 Da and a negative 

charge could be removed effectively (Fig. 7). The formation 

of estrone-HpoA complex in bulk solution led to 

a significant increase in MW (from 270 Da to more than 

4000 Da) and the appearance of negative charge and thus an 

increased level of estrone rejection by negative charged 

membranes based on the mechanisms of size exclusion and 

charge repulsion. Therefore, it could be concluded that the 

HpoA fraction was the greatest contributor to the ‘‘enhancement 

effect’’ on estrone rejection in treated effluent. 

It is worthwhile to note that the ‘‘enhancement effect’’ of 

HpoA followed the order of DL > CG > CK. This sequence 

agrees with the sequence based on the negative charge of the 

membrane surface (Table 1). According to the above discussion, 

the binding of estrone with HpoA in feed solution was 

critically important for the ‘‘enhancement effect’’. Less negative 

charge of membrane surface would result in more HpoA 

adsorption onto membrane due to the weaker electrostatic 

repulsion between HpoA and membrane. Consequently, less 

HpoA molecules would be available in feed solution for binding 

with estrone, resulting in a weaker improvement in estrone 

rejection. The results suggest that a more negatively-charged 

Table 3 – Characteristics of specific organic fractions derived from secondary effluent. 

Parameter HpoA HpoB HpoN HpiA 

MW (Da) 4192–4442 762–1751 1406–2658 3628–3825 

r 1.29–1.43 1.01–1.84 1.01–1.27 1.18–1.32 

SUVA254 (m 1 L/mg C) 2.58–4.25 2.06–3.81 1.19–1.57 0.75–0.81

a 


b 


c 


100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

20 

0 


With HpoA 

With HpoB 

With HpoN 

0 5 10 15 20 25 



With HpoA 

With HpoB 

With HpoN 

0 5 10 15 20 25 



With HpoA 

With HpoB 

With HpoN 

0 5 10 15 20 25 


Fig. 6 – Observed rejection of estrone with the presence of 

different hydrophobic organic fractions (at their original 

concentration in secondary effluent) by (a) DL, (b) CK and (c) CG. 

membrane would benefit the interaction between HpoA and 

estrone and thus estrone rejection. 

3.5. Effect of calcium ion concentration on estrone 

rejection in HpoA-containing solution 

Based on the discussion above, the estrone-binding capacity 

of HpoA is the major contributor to its significant ‘‘enhancement 

effect’’ on estrone rejection by NF/RO membranes. It is 

known that divalent cations, such as calcium, can influence 

the binding of trace contaminants by humic substances 


a 


b 


c 


100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

100 

90 

80 

70 

60 

50 

40 

30 

20 

10 

0 

HpoA 

HpoB 

HpoN 

0 5 10 15 20 25 


HpoA 

HpoB 

HpoN 

0 5 10 15 20 25 


HpoA 

HpoB 

HpoN 

0 5 10 15 20 25 


Fig. 7 – Observed rejection of bulk organics (as DOC) with the 

presence of different hydrophobic organic fractions (at their 

original concentration in SE) by (a) DL, (b) CK and (c) CG. 

(Schlautman and Morgan, 1993). Therefore, calcium ion 

concentration in feed solution is predicted to affect the 

estrone rejection in HpoA-containing solution. 

Fig. 8 presents the influence of HpoA on estrone rejection 

by DL membrane at different calcium concentrations. It is 

evident that while the presence of HpoA could increase the 

rejection of estrone, a higher calcium ion concentration tended 

to reverse this effect. In particular, the addition of 0.3 mM 

calcium ion in feed solution resulted in the ‘‘enhancement

646 

a 


b 


c 


100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

20 

0 

100 

80 

60 

40 

w/o HpoA 

with HpoA 

0 5 10 15 20 25 


w/o HpoA 

with HpoA 

0 5 10 15 20 25 


w/o HpoA 

with HpoA 

0 5 10 15 20 25 


Fig. 8 – Influence of HpoA on estrone rejection by DL 

membrane at different calcium ion concentration (a) 0 mM 

Ca 2D , (b) 0.3 mM Ca 2D and (c) 0.6 mM Ca 2D . 

effect’’ of HpoA on ultimate estrone rejection by DL 

membrane decreasing to 18.0% as compared to the improvement 

of 29.8% with the absence of calcium ion. When calcium 

ion concentration was increased to 0.6 mM, HpoA did not 

exhibit observable improvement in estrone rejection. 

In addition, negligible differences in permeate flux decline 

and DOC rejection were observed with increasing calcium 

concentration (data are not shown, but are available on 

request). This observation is unexpected and also differs 


from other published investigations (Hong and Elimelech, 

1997; Tang et al., 2007), in which a more serious permeate 

flux decline with an increased calcium concentration was 

reported. This phenomenon was possibly due to the relatively 

low calcium ion concentration added in this study. 

Such low calcium ion concentration might not be enough to 

overcome the electrostatic repulsive energy barrier between 

HpoA and membrane and thus cannot result in a severe 

permeate flux decline. In this case, the diminished effect of 

HpoA on estrone rejection by membrane with increasing 

calcium concentration can most probably be attributed to the 

lower estrone–HpoA complex formation. Calcium ions can 

interact specifically with acidic functional groups (predominantly 

carboxylic groups) of humic substances (Hong and 

Elimelech, 1997). Such interactions reduce the charge repulsion 

between negative functional groups on HpoA molecules 

due to charge shielding and neutralization. As a result, the 

structure of HpoA tends to change from a stretched and 

linear configuration to a coiled and compact configuration 

with increased calcium concentration (Clark and Lucas, 

1998). This would result in the interactive sites within HpoA 

molecules becoming less easily reachable by estrone molecules 

by restricting access to interior sorption sites. In this 

case, the increase in calcium concentration would decrease 

the ability of HpoA to bind estrone molecules in feed water. 

This observation is similar to the results of Schlautman and 

Morgan (1993) who reported that the presence of calcium at 

neutral pH decreased the ability of humic acid to bind perylene. 

They attributed the finding to the fact that the 

compression of the HA structure decreased the size of voids 

in humic acid into which target solute could partition. Since 

the presence of calcium ions was not conducive to the 

binding of estrone by HpoA, the effect of increasing calcium 

concentration, even at relatively low levels, was to push 

estrone rejection towards the ‘‘free’’ estrone limit of rejection 

obtained in the absence of HpoA. 


The results reported in this paper demonstrate that NF and 

loose RO membranes cannot remove trace steroid hormone 

efficiently in single-organic solution. However, the removal 

efficiency can be enhanced significantly in the presence of 

EfOM in feed solution. The HpoA fraction played a paramount 

role in the ‘‘enhancement effect’’. This enhancement is 

mainly attributed to the binding of estrone molecules to HpoA 

in feed solutions and being removed with the negatively 

charged HpoA macromolecules based on the mechanisms of 

size exclusion and charge repulsion. However, the presence 

of calcium ion tended to diminish this effect. Application of 

membrane with more negative-charge at its interface can 

greatly intensify this ‘‘enhancement effect’’. 



in online version at doi:10.1016/j.watres.2009.09.057.

eferences 

Adams, N.R., 1998. Clover phyto-oestrogens in sheep in Western 

Australia. Pure and Applied Chemistry 70 (9), 1855–1862. 

Agbekodo, K.M., Legube, B., Dard, S., 1996. Atrazine and simazine 

removal mechanisms by nanofiltration: influence of natural 

organic matter concentration. Water Research 30 (11), 2535–2542. 

Agenson, K.O., Urase, T., 2007. Change in membrane performance 

due to organic fouling in nanofiltration (NF)/reverse osmosis 

(RO) applications. Separation and Purification Technology 55 

(2), 147–156. 

Baker, V.A., 2001. Endocrine disrupters – testing strategies to 

assess human hazard. Toxicology in Vitro 15 (4-5), 413–419. 

Baronti, C., Curini, R., D’Ascenzo, G., Di Corcia, A., Gentili, A., 

Samperi, R., 2000. Monitoring natural and synthetic estrogens 

at activated sludge sewage treatment plants and in a receiving 

river water. Environmental Science & Technology 34 (24), 

5059–5066. 

Bellona, C., Drewes, J.E., 2007. Viability of a low-pressure nanofilter 

in treating recycled water for water reuse applications: a pilotscale 

study. Water Research 41 (17), 3948–3958. 

Bellona, C., Drewes, J.E., Xu, P., Amy, G., 2004. Factors affecting 

the rejection of organic solutes during NF/RO treatment – 

a literature review. Water Research 38 (12), 2795–2809. 

Clark, M.M., Lucas, P., 1998. Diffusion and partition of humic acid 

in a porous ultrafiltration membrane. Journal of Membrane 

Science 143 (1–2), 13–25. 

Comerton, M.A., Andrews, C.R., Bagley, M.D., Hao, C., 2008. The 

rejection of endocrine disrupting and pharmaceutically active 

compounds by NF and RO membranes as a function of 

compounds and water matrix properties. Journal of 

Membrane Science 313 (1-2), 323–335. 

Croley, T.R., Hughes, R.J., Koenig, B.G., Metcalfe, C.D., March, R.E., 

2000. Mass spectrometry applied to the analysis of estrogens 

in the environment. Rapid Communications in Mass 

Spectrometry 14 (13), 1087–1093. 

Desbrow, C., Routledge, E.J., Brighty, G.C., Sumpter, J.P., Waldock, M., 

1998. Identification of estrogenic chemicals in STW effluent. 1. 

Chemical fractionation and in vitro biological screening. 


Devitt, E.C., Ducellier, F., Cote, P., Wiesner, M.R., 1998. Effects of 

natural organic matter and the raw water matrix on the 

rejection of atrazine by pressure-driven membranes. Water 

Research 32 (9), 2563–2568. 

Hong, S., Elimelech, M., 1997. Chemical and physical aspects of 

natural organic matter (NOM) fouling of nanofiltration 

membranes. Journal of Membrane Science 132 (2), 159–181. 

Hu, J.Y., Ong, S.L., Shan, J.H., Kang, J.B., Ng, W.J., 2003. Treatability of 

organic fractions derived from secondary effluent by reverse 

osmosis membrane. Water Research 37 (19), 4801–4809. 

Hu, J.Y., Jin, X., Ong, S.L., 2007. Rejection of estrone by 

nanofiltration: influence of solution chemistry. Journal of 

Membrane Science 302 (1-2), 188–196. 

Iguchi, T., Watanabe, H., Katsu, Y., 2001. Developmental effects of 

estrogenic agents on mice, fish, and frogs: a mini-review. 

Hormones and Behavior 40 (2), 248–251. 

Imai, A., Fukushima, T., Matsushige, K., Kim, Y.H., Choi, K., 

2002. Characterization of dissolved organic matter in 

effluents from wastewater treatment plants. Water Research 

36 (4), 859–870. 

Jin, X., 2007. Rejection of steroid hormone estrone by NF/RO 

membranes. PhD thesis. Department of Civil Engineering, 

National University of Singapore, Singapore. 

Jin, X., Hu, J.Y., Ong, S.L., 2007. Influence of dissolved organic 

matter on estrone removal by NF membranes and the role of 

their structures. Water Research 41 (14), 3077–3088. 


Johnson, A.C., Sumpter, J.P., 2001. Removal of endocrinedisrupting 

chemicals in activated sludge treatment works. 


Kimura, K., Amy, G., Drewes, J.E., Heberer, T., Kim, T.U., 

Watanabe, Y., 2003. Rejection of organic micropollutants 

(disinfection by-products, endocrine disrupting compounds, 

and pharmaceutically active compounds) by NF/RO 

membranes. Journal of Membrane Science 227 (1–2), 113–121. 





Leenheer, J.A., 1981. Comprehensive approach to preparative 

isolation and fractionation of dissolved organic carbon from 

natural waters and wastewaters. Environmental Science & 

Technology 15 (5), 578–587. 

Majewska-Nowak, K., Kabsch-Korbutowicz, M., Dod, M., 

Winnicki, T., 2002. The influence of organic carbon 

concentration on atrazine removal by UF membranes. 

Desalination 147 (1–3), 117–122. 

Ng, H.Y., Elimelech, M., 2004. Influence of colloidal fouling on 

rejection of trace organic contaminants by reverse osmosis. 


Nghiem, L.D., Hawkes, S., 2007. Effects of membrane fouling 

on the nanofiltration of pharmaceutically active 

compounds (PhACs): Mechanisms and role of membrane 

pore size. Separation and Purification Technology 57 (1), 

176–184. 

Nghiem, L.D., Manis, A., Soldenhoff, K., Schäfer, A.I., 2004a. 

Estrogenic hormone removal from wastewater using NF/RO 

membranes. Journal of Membrane Science 242 (1–2), 

37–45. 

Nghiem, L.D., Schäfer, A.I., Elimelech, M., 2004b. Removal of 


measurement, modeling, and mechanisms. Environmental 


Schäfer, A.I., Fane, A.G., Waite, T.D., 2005. Nanofiltration: 

Principles & Applications. Elsevier Advanced Technology, 

Oxford, New York. 

Schäfer, A.I., Nghiem, L.D., Waite, T.D., 2003. Removal of the 

natural hormone estrone from aqueous solutions using 

nanofiltration and reverse osmosis. Environmental Science & 

Technology 37 (1), 182–188. 

Schlautman, M.A., Morgan, J.J., 1993. Effects of aqueous chemistry 

on the binding of polycyclic aromatic hydrocarbons by 

dissolved humic materials. Environmental Science & 

Technology 27 (5), 961–969. 

Soliman, M.A., Pedersen, J.A., Suffet, I.H., 2004. Rapid gas 

chromatography–mass spectrometry screening method for 

human pharmaceuticals, hormones, antioxidants and 

plasticizers in water. Journal of Chromatography A 1029 (1-2), 

223–237. 

Steinle-Darling, E., Zedda, M., Plumlee, H.M., Ridgway, F.H., 

Reinhard, M., 2007. Evaluating the impacts of membrane 

type, coating, fouling, chemical properties and water 

chemistry on reverse osmosis rejection of seven 

nitrosoalklyamines, including NDMA. Water Research 41 (17), 

3959–3967. 

Tang, Y.C., Kwon, Y.N., Leckie, O.J., 2007. Fouling of reverse 

osmosis and nanofiltration membranes by humic acid – 

effects of solution composition and hydrodynamic 

conditions. Journal of Membrane Science 290 (1–2), 

86–94. 

Ternes, T.A., Stumpf, M., Mueller, J., Haberer, K., Wilken, R.D., 

Servos, M., 1999. Behavior and occurrence of estrogens in 

municipal sewage treatment plants - I. Investigations in 

Germany, Canada and Brazil. The Science of the Total 

Environment 225 (1–2), 81–90.

648 

Thurman, E.M., Malcolm, R.L., 1981. Preparative isolation of 

aquatic humic substances. Environmental Science & 

Technology 15 (4), 463–466. 

Tim, Z., 1997. In vitro bioassays for assessing estrogenic substances. 


Williams, R.J., Johnson, A.C., Smith, J.J., Kanda, R., 2003. Steroid 

estrogens profiles along river stretches arising from sewage 

treatment works discharges. Environmental Science & 

Technology 37 (9), 1744–1750. 

Xu, P., Drewes, J.E., Kim, T.U., Bellona, C., Amy, G., 2006. Effect of 

membrane fouling on transport of organic contaminants in 


NF/RO membrane applications. Journal of Membrane Science 

279 (1–2), 165–175. 

Yamamoto, H., Liljestrand, H.M., Shimizu, Y., Morita, M., 2003. 

Effects of physical-chemical characteristics on the sorption of 

selected endocrine disruptors by dissolved organic matter 

surrogates. Environmental Science and Technology 37 (12), 

2646–2657. 

Yoon, Y., Westerhoff, P., Yoon, J., Snyder, S.A., 2004. Removal of 

17 beta estradiol and fluoranthene by nanofiltration and 

ultrafiltration. Journal of Environmental Engineering 130 (12), 

1460–1467.

Screening of antimycotics in Swedish sewage treatment 

plants – Waters and sludge 

Richard H. Lindberg*, Jerker Fick, Mats Tysklind 

Department of Chemistry, Umea˚ University, SE-901 87 Umea˚, Sweden 

article info 




19 October 2009 

Accepted 23 October 2009 


Keywords: 

Antimycotics 

Sewage water 

Sludge 

Mass flow 


abstract 

There are approximately 40 antimycotics, i.e. pharmaceuticals 

that can be used to treat fungal infections in humans, 

many of which are imidazoles, triazoles or allylamines. Antimycotics 

are usually administered topically, although a few, 

such as ketoconazole and fluconazole, are also used perorally. 

All antimycotics currently used are general cytochrome P450 

inhibitors that reportedly affect CYP families 1–3. For instance, 

the imidazoles and triazoles strongly inhibit CYP3A4 and 2C9, 

while the allylamines are known to affect CYP2D6 (Sweetman, 

2007). This has potentially significant implications for their 

use as pharmaceuticals since the P450s are responsible for the 

initial metabolism of ca. 75% of xenobiotics in humans, hence 





E-mail address: richard.lindberg@chem.umu.se (R.H. Lindberg). 


doi:10.1016/j.watres.2009.10.034 

Concentrations of six pharmaceutical antimycotics were determined in the sewage water, 

final effluent and sludge of five Swedish sewage treatment plants (STPs) by solid phase 

extraction, liquid/solid extraction, and liquid chromatography–electrospray tandem mass 

spectrometry. The antimycotics were quantified by internal standard calibration. The 

results were used to estimate national flows that were compared to predictions based on 

sales figures. Fluconazole was the only one of the six investigated antimycotics that was 

detected (at concentrations ranging from 90 to 140 ng L 1 ) in both raw sewage water and 

final effluent. Negligible amounts of this substance were removed from the aqueous phase, 

and its levels were below the limit of quantification in all of the analyzed sludge samples. 

In contrast, clotrimazole, ketoconazole and econazole were present in all of the sludge 

samples, at concentrations ranging between 200 and 1000 mgkg 1 , dry weight. There were 

close correlations between the national measured and predicted antimycotic mass flows. 

Antimycotic fate analysis, based on sales figures, indicated that 53% of the total amount of 

fluconazole sold appeared in the final effluents of the STPs, while 1, 155, 35, 209 and 41% of 

the terbinafine, clotrimazole, ketoconazole, econazole and miconazole sold appeared in the 

digested dewatered sludge. 


they could have undesirable effects on patients’ ability to 

eliminate other substances (Sweetman, 2007; Guengerich, 

2008). More specifically, the imidazole and triazole antimycotics 

inhibit the essential activity of CYP51 (also known as 

ERG11 in fungi) in ergosterol synthesis in fungal pathogens, 

thereby increasing the permeability of their membranes, 

inhibiting their growth and/or killing them (Graybill and 

Drutz, 1980; Yoshida et al., 2000; Lepesheva and Waterman, 

2007). The antimycotic activity of allylamine antimycotics is 

also based on interference with ergosterol biosynthesis, but by 

inhibiting squalene epoxidase (Ryder, 1992). 

Gunnarson et al. have recently shown that six species 

commonly used in aquatic ecotoxicology tests possess CYP51 

orthologs that have more than 70% homology to the drug

650 

targets (Gunnarsson et al., 2008). Further, imidazole and triazole 

antimycotics have been shown to inhibit CYP19, which 

participates in the conversion of androgens to estrogens 

(Trösken et al., 2004; Gyllenhammar et al., 2009), with serious 

potential consequences for sex differentiation in exposed 

vertebrates, including testicular development in female 

salmon (Piferrer et al., 1994). Gyllenhammar et al. (2009) have 

also shown that exposure to the imidazole clotrimazole can 

have complex effects on aromatase activity in the gonads and 

brain of Xenopus tropicalis larvae; decreasing brain aromatase 

activity during their gonadal differentiation and increasing 

gonadal aromatase activity during metamorphosis. 

Concentrations of antimycotics have been analyzed 

(generally by solid phase extraction, SPE, followed by liquid (or 

gas) chromatography mass spectrometry, LC(GC)–MS/MS) in 

various aquatic environments and sewage flows (Kolpin et al., 

2004; Thomas and Hilton, 2004; Roberts and Bersuder, 2006; 

Roberts and Thomas, 2006; Peschka et al., 2007; Kahle et al., 

2008; Van De Steene and Lambert, 2008). Clotrimazole levels in 

the Main and Rhine rivers in Germany, and both the raw 

sewage water and final effluent from a German sewage 

treatment plant (STP), have been reported to be in the low 

ng L 1 range (Peschka et al., 2007). Comparable concentrations 

of clotrimazole have also been reported in two UK studies, 

which assessed sewage water from the Howton STP and 

samples collected along the Tyne estuary (Thomas and Hilton, 

2004; Roberts and Thomas, 2006). Somewhat higher levels of 

clotrimazole and fluconazole, between 10 and 110 ng L 1 , have 

been found in Swiss STP waters, along with fluconazole at 

levels close to its limit of quantification (LOQ), 1 ng L 1 ,in 

Swiss lakes (Kahle et al., 2008). In addition, quantifiable levels 

of miconazole have been found in both surface and sewage 

waters in the UK (Roberts and Bersuder, 2006), although it was 

not found at levels exceeding its reporting limit (17.5 ng L 1 )in 

an analysis of the contribution of wastewater contaminants to 

US streams (Kolpin et al., 2004). However, no published studies 

to our knowledge have analyzed the concentrations of antimycotics 

in digested sewage sludge. 

The six substances examined here were chosen to represent 

a range, in terms of applications and physicochemical 

properties, of antimycotics commonly used in Sweden. 

Information on these six substances and the internal standard 

sulconazole (which is not used in this country) is presented in 

Table 1. Little is known about the fate and ecotoxic potential of 

these substances, thus the aim of the presented study was to 

increase information on their occurrence, levels and fate 

within STPs. In addition, the collected data were cross-validated 

against predictive equivalents. To our knowledge, this is 

the first published study to present national screening data on 

antimycotics in both sewage water and sludge. 

2. Experimental 


All of the antimycotics were bought from Sigma Aldrich 

(Stockholm, Sweden) except for terbinafine, which was 

obtained from LGC Standards (Bora˚s, Sweden). Formic acid 

and methanol (HPLC-grade) were purchased from JT Baker 


(Deventer, the Netherlands), acetonitrile (HPLC-grade) from 

Fischer chemicals (Zurich, Switzerland), and sulfuric acid 

from Merck (Darmstadt, Germany). Water was purified (to 

18.2 MU cm resistivity) using a MAXIMA HPLC ultra pure 

water system (ELGA, High Wycombe Bucks, England). Stock 

and working (including calibration standards) solutions of 

individual antimycotics were prepared in methanol and 

diluted in water, respectively. 

2.2. Sample sites, collection and pre-extraction 

procedure 

Raw sewage water, final effluent and sludge samples were 

taken from five Swedish STPs during October and November 

2007. The STPs surveyed were Stockholm (Henriksdal), Gothenburg 

(Ryaverken), Umea˚ , Alingsa˚s (Nolhaga), and Bollebygd, 

which were selected to represent a wide range of STPs in 

terms of size and catchment area. Sewage water from the 

Umea˚ county hospital (UCH) was also collected during the 

same time period. In addition, raw sewage water, raw sewage 

water particles, raw sludge and digested dewatered sludge 

were collected from the Umea˚ STP in April 2008 to assess the 

fate of antimycotics during the treatment process. The locations 

of the STPs and data on their waste flows and catchment 

areas are provided in Fig. 1. All of the STPs treat sewage water 

mechanically, chemically and biologically. Nitrogen is also 

actively removed at the Henriksdal, Ryaverken and Nolhaga 

STPs, and the final solid product of all the STPs is anaerobically 

digested sludge, except at Bollebygd, where only clarifiers 

are used. 

Flow-proportional samples of raw sewage water and final 

effluents were collected from each of the five STPs in each 

case over a full single day, while grab samples of the UCH 

sewage water and digested dewatered sludge were collected. 

After collection, the sewage water (2 L in brown glass bottles) 

and sludge samples (500 g in brown glass beakers) were 

immediately stored at 18 C prior to extraction. Grab 

sampling was also used in the antimycotic fate study at the 

Umea˚ STP, in which raw sewage water (1 þ 10 L), final effluent 

(2 L), primary sludge (2 L) and 500 g of digested dewatered 

sludge samples were collected between 8:30 and 9:00 on each 

of two consecutive days in April 2008 (except for the digested 

dewatered sludge, which was only collected on the first of the 

sampling days due to technical problems). The raw sewage 

water (1 L) and final effluent samples for SPE were filtered 

(through 0.45 mm filters) immediately after collection, while 

the remaining raw sewage water (10 L) and primary sludge 

samples were centrifuged at 4800 rpm for 20 min, to concentrate 

the particles sufficiently for analysis (Lindberg et al., 

2006), then air-dried for 24 h. All of the samples were then 

stored at 18 C awaiting extraction. 

2.3. Sample preparation – sewage waters 

The aqueous samples were thawed, filtered through 0.45 mm 

MFÔ membrane filters (Millipore, Sundbyberg, Sweden), 

acidified to pH 3 with sulfuric acid and 500 ng of the internal 

standard (IS) sulconazole was added to 500 mL portions of the 

resulting samples, which were then subjected to SPE using 

Evolute ABN (6 mL, 200 mg) solid sorbent columns (Biotage,

Table 1 – Information on the antimycotics investigated in this study. 

Terbinafine 

(allylamine) 

Clotrimazole 

(imidazole) 

Ketoconazole 

(imidazole) 

Fluconazole 

(triazole) 

Econazole 

(imidazole) 

Miconazole 

(imidazole) 

IS sulconazole 

(imidazole) 

O 

N 

N 

N 

Cl 

N N O 

Cl 

Cl 

Cl 

N 

Cl 

N 

HO 

N 

N 

N 

N 

N 

H 

O 

N 

O 

N 

Cl 

F F 

N 

O 

O 

N 

S 

Cl 


N 

Cl 

N 

Cl 

N 

Cl 

Cl 

Cl 

Cl 

Application Use a 

Topical 43 

Topical 11 

(kg year 1 ) 

EAS b Log K ow c 

g 

g 

PEC d 

(ng L 1 ) 

UD e BCF f 

5.81 6 m 5370 

6.26 2 m 13,183 

Topical/peroral 399/22 –/0% 4.45 (4.35) 60 r 447 

Peroral 121 80% u 0.25 20 r 3 

Topical 67 

Topical 134 

Not used 

in Sweden 

g 

g 

5.61 10 r 4169 

6.25 20 r 12,882 

a Estimated from Apoteket AB’s Swedish sales during 2007. 

b Excretion of active substance through urine and/or feces (LIF, 2008). 

c Log Kow values estimated by KowWin, EPI suiteÔ (EPI SuiteÔ) (experimental values in parentheses). 

d Predicted environmental concentrations in surface waters, calculated according to EMEA guidelines Phase 1, except that Swedish sales values 

were used rather than the supposed penetration factor (Fpen), in combination with the daily defined dose (European Medicines Agency, 2006). 

e Ultimate degradation (m ¼ months, r ¼ recalcitrant) estimated by BioWin, EPI suiteÔ (EPI SuiteÔ). 

f Estimated bioconcentration factor as estimated by BcfWin, EPI suiteÔ (EPI SuiteÔ). 

g No excretion due to topical application. 

6.46

652 

Uppsala, Sweden) that had been conditioned with 6 mL of 

methanol and 6 mL water adjusted to pH 3 with sulfuric acid. 

Each sample was applied at a flow rate of 5 mL min 1 , cleaned 

using 5 mL of aqueous methanol (5%), and eluted by two 

washes of 3 mL methanol with 0.5% formic acid. The eluates 

were evaporated almost to dryness and finally reconstituted 

with 0.7 mL aqueous acetonitrile (10%). 

2.4. Sample preparation – sludge 

The sludge samples were thawed, then 2.0 g portions from 

each sample were placed in separate glass centrifuge tubes. 

The IS was added, 500 ng, and the samples were subjected to 

liquid/solid extraction step by adding 10.0 mL of methanol to 

each tube and ultrasonically agitating them for 10 min. They 

were subsequently centrifuged at 4800 rpm for 10 min, and 

the supernatants were placed in separate 30 mL glass vials. 

The residues were then re-extracted in the same manner and 

the two supernatants were combined. The combined extracts 

(approximately 19 mL) were made up to 200 mL with purified 


Fig. 1 – Mass flows of the measured antimycotics (g day L1 ) in the raw sewage water, final effluent and digested dewatered 

sludge of five Swedish sewage treatment plants. 

water, acidified to pH 3, and finally subjected to SPE using the 

protocol described above for the aqueous samples. 

The dry weights of the samples were determined by 

weighing portions after drying them at 105 C for 24 h. 

2.5. Liquid chromatography–mass spectrometry 

Sample extracts and calibration solutions (10 mL) were injected 

using an AS 3000 autoinjector (Thermo Finnigan, San Jose, CA, 

US) into a Hydrosphere C18 guard column (10 4 mm i.d., 

5 mm particle size; YMC Inc., Wilmington, NC, US) followed by 

an analytical column of the same type (150 4.6 mm i.d., 5 mm 

particle size). The antimycotics were then eluted by a linear 

gradient of 90:10 to 70:30 A:B (v/v) over 17 min followed by 

a further 4-min linear gradient to 30:70 (v/v), where A and B are 

water and acetonitrile, respectively (both acidified with 0.1% 

formic acid), at a 0.8 mL min 1 flow rate, generated using 

a P4000 HPLC pump (Spectra system, Thermo Finnigan) at 

a constant temperature of 25 C. 

The eluting antimycotics were detected by an LCQ Duo ion 

trap mass spectrometer (Thermo Finnigan) with an

electrospray ion source operating in positive ion mode. The 

source voltage was maintained at a constant 6.0 kV and the 

heated capillary temperature was set to 250 C. The MS/MS 

parameters and the collision energies were optimized semiautomatically 

and manually, respectively. 

2.6. Extraction yield assessment 

Absolute extraction yields (including matrix suppression of 

precursor/product ions during instrumental analysis) were 

investigated by fortifying 500 ng of each analyte and the IS to 

raw sewage water (500 mL, n ¼ 4) and digested sludge samples 

(2.0 g, n ¼ 3). The SPE of the water samples began immediately 

but the sludge samples were left for an hour prior to the liquid/ 

solid extraction. The extraction yield was evaluated by 

a comparison of the analyte and IS peak areas of the samples 

to the corresponding peak areas in a calibration solution. The 

same experimental set up was used in order to assess corrected 

extraction yields with the exception that the IS was 

added post extraction. The corrected extraction yields were 

based on peak area ratios (analyte peak area/IS peak area) of 

the samples to the corresponding peak areas ratios in a calibration 

solution. 

2.7. Quantification 

The antimycotics were identified by retention time, precursor 

and product ion m/z. The most abundant transitions were 

monitored and the IS calibration method (peak area ratios of 

analytes and IS) was used for all substances except fluconazole 

(external calibration without recovery correction). The 

ten-point calibration curve, based on 500 ng of sulconazole 

and ranging from 1 ng mL 1 to 1000 ng mL 1 of the analytes, 

was prepared in aqueous acetonitrile (10%). 

2.8. Mass flow calculations 

Mass flows (g day 1 ) of each antimycotic substance at each 

STP were calculated using the obtained concentrations and 

the aqueous and solid flows during the sample collection 

periods, and the daily production of digested dewatered 

sludge (dw) obtained from monthly production figures 


supplied by staff of the STPs. The predicted mass flows 

(g day 1 ) were based on total sales of the active substances to 

both institutional and non-institutional care during October 

and November 2007 (Svensson, 2008), the number of people 

connected to each STP and the total population of Sweden 

(Statistics Sweden). 


3.1. Quantification 

At least two transitions were monitored for each of the antimycotics 

(see Table 2), except ketoconazole, for which only 

one detectable product ion was formed, high CEs only resulted 

in additional low m/z product ions with weak signal intensities. 

An approach that could have been used to improve the 

level of confirmation is the use of the other Cl-isotope of 

ketoconazole, unfortunately it was not considered in this 

work. In addition, although two product ions were recorded 

for fluconazole, signals from one of them were often severely 

suppressed by matrix components, so it was not used for 

quantification. In general, matrix suppression of the analyte 

and IS precursor/product ion signals was immense and in this 

context it should be noted that the Evolute ABN sorbent 

column used in the SPE process was designed specifically for 

quantifying analytes in dirty extracts, since it reduces interference 

by matrix components such as macromolecules, but 

the sewage water and sludge extracts were brownish in color, 

indicating that it failed to remove substantial proportions of 

low molecular weight substances from these samples. It 

should also be noted that IS calibration was not used to 

quantify fluconazole, since its extraction yields were much 

higher than those of the other antimycotics (including the IS), 

and it correlated most poorly with the IS in terms of retention 

time and Log Kow. In addition, the IS was not natively present 

in the samples, thus an underestimation of the antimycotics 

was avoided. Terconazole was also considered as an IS, 

however, traces of it were found in the sewage waters and 

sludge from the southern part of Sweden, most likely due to 

its use as a pesticide in the agricultural areas, and it was 

therefore not used. 

Table 2 – Monitored ions, retention times, method detection limits and extraction yields of the antimycotics in aqueous and 

solid matrices. 

MS/MS transitions CE a (AU) t r (min) MLOQs b 

Extraction 

yields (%) 

IS corrected 

extraction yields (%) 

Waters (ng L 1 ) Sludge (mgkg 1 ) Waters Solids Waters Solids 

Terbinafine 292.8 > 141.4, 205.4, 170.3 31 17.3 5 3 27 9 40 7 133 48 109 26 

Clotrimazole 277.2 > 165.5, 242.5 35 17.2 5 3 23 5 34 8 117 31 79 6 

Ketoconazole 532.9 > 490.7 33 14.9 20 10 9 5 48 15 104 21 108 9 

Fluconazole 307.3 > 238.3 23 13.3 50 40 104 12 64 6 – – 

Econazole 383.0 > 125.3, 193.1 34 18.6 80 40 10 6 53 16 115 21 115 10 

Miconazole 417.7 > 161.4, 229.4 38 19.8 100 50 5 3 41 12 69 28 95 17 

Sulconazole (IS) 399.3 > 330.8, 183.2 22 19.1 – 13 10 43 10 – – 

a Collision energy in LC–MS/MS (arbitrary units). 

b Method limits of quantification based on the second calibration curve point in terms of mass per liter (liquid samples) and dry weight (sludge 

samples).

654 

3.2. Concentrations and mass flows of the antimycotics 

in sewage water and sludge 

The results from the Umea˚ STP antimycotic fate and the 

national screening study are presented in Table 3. The 

concentrations of the antimycotics in both the aqueous and 

solid matrices were generally similar to, or slightly lower than, 

previously reported levels of antimycotics (Kahle et al., 2008) 

and other pharmaceuticals, such as antibiotics (Golet et al., 

2003; Carballa et al., 2004; Göbel et al., 2005a,b; Lindberg et al., 

2005, 2006). However, the only substances found at higher 

levels than their respective LOQs in the aqueous samples were 

fluconazole (in both the Umea˚ STP and hospital samples) and 

ketoconazole (in the hospital samples). The levels found in the 

final effluent were in close correlation with the raw sewage 

water (just below LOQ in raw sewage water of Stockholm but 

the chromatographic peak was identified and above three 

times the signal to noise ratio) hence elimination from the 

aqueous phase is minimal. Fluconazole was not present in 

any of the sludge samples thus it is the only one of the 

monitored antimycotics that will be directly released into 

recipient waters of the STPs. Kahle et al. came to the same 

conclusion regarding the fate of fluconazole in their study 

made at Swiss STPs (Kahle et al., 2008). Three of the other 

antimycotics (clotrimazole, ketoconazole and econazole) were 

detected in sludge samples collected from all of the studied 

STPs. However, miconazole was not detected in any samples 

Table 3 – Sewage water and sludge antimycotic concentrations. 

National screening study a,b 

STP Um STP St STP Al STP Go STP Bo UCH 

RSW, ng L 1 

from the Alingsa˚s or Bollebygd STPs, and terbinafine was not 

detected in samples from the Umea˚ and Alingsa˚s STPs. 

According to EMEA guidelines (see Table 1), predicted environmental 

surface water concentrations can only be considered 

acceptable for fluconazole, and at best misleading for the 

other substances, since sludge is not included in the equation. 

It should be noted here that the distribution of the antimycotics 

between the aqueous and solid phases seems to be 

consistent with their respective Log Kow values. However, 

models based on Log K ow are not suitable to predict the environmental 

fate of pharmaceuticals (Tolls, 2001), and 

substances with similar Log K ow values have been shown to 

sorb to sludge to widely differing degrees (Lindberg et al., 2005, 

2006). 

Antimycotic mass flows (g day 1 ) were established to: (a) 

investigate the absolute distribution of these substances 

between the aqueous and solid phases, and (b) cross-validate 

them with mass flows based on sales statistics. The 

recorded mass flows from the STP screening study are 

presented together with the predicted values in Fig. 1. The 

recorded antimycotic mass flows were similar (within an 

order of magnitude) to their respective predicted equivalents 

at each STP with two exceptions: a) the measured 

mass flows in Bollebygd STP were ca. 100-fold lower than 

predicted, possibly because it is the smallest of the investigated 

STPs, and that Bollebygd is a small town lacking 

both a hospital and large industries; and b) the much lower 

RSW, ng L 1 

Terbinafine d d d d d d d 

Clotrimazole d d d d d d d 

Ketoconazole d d d d d 30 d 

Fluconazole 90 d d d d 570 120 

Econazole d d d d d d d 

Miconazole d d d d d d d 

FE, ng L 1 


FE, ng L 1 

Terbinafine d d d d d d 

Clotrimazole d d d d d d 

Ketoconazole d d d d d d 

Fluconazole 140 100 d d d 100 

Econazole d d d d d d 

Miconazole d d d d d d 

Fate in Umea˚ STP study a,c 

DDS, mgkg 1 (dw) RSP, mgkg 1 (dw) RS, mgkg 1 (dw) DDS, mgkg 1 (dw) 

Terbinafine d 4 d 7 30 d 40 d 

Clotrimazole 30 120 120 110 50 10 310 60 

Ketoconazole 910 910 480 670 280 980 1300 1800 

Fluconazole d d d d d d d d 

Econazole 590 1000 660 680 210 470 240 550 

Miconazole 160 970 d 190 d d d 370 

a RSW ¼ raw sewage water, FE ¼ Final effluent, DDS ¼ digested dewatered sludge, RSP ¼ raw sewage particles, and RS ¼ raw sludge. 

b STP Um ¼ Umea˚ ; STP St ¼ Stockholm, Henriksdal; STP Al ¼ Alingsa˚ s; STP Go ¼ Gothenburg, Ryaverken; STP Bo ¼ Bollebygd; UCH ¼ Umea˚ 

county hospital (Norrlands Universitetssjukhus). 

c Mean concentrations from two days of sampling. 

d Below LOQ.

observed mass flow of terbinafine, which suggests that this 

topically used substance (of the ones included) is only to 

a small extent present in the sewage waters post application. 

In Umea˚, the use of ketoconazole and fluconazole in 

the institutional care accounts for ca. 23% and 56% of the 

total amounts sold, respectively. However, while the mass 

flow of fluconazole in the UCH sewage water represented 

31% of the total mass flow in the STP raw sewage water, 

a substantially smaller proportion of the ketoconazole of 

UCH (1%) was found in the solid flows of the STP (digested 

dewatered sludge). In addition, mean mass flow values 

normalized to the number of people connected to each STP 

indicated that 53% of the purchased fluconazole eventually 

appeared in the final effluents, while 1, 35, 41, 155 and 

209% of terbinafine, ketoconazole, miconazole, clotrimazole 

and econazole, respectively, appeared in digested dewatered 

sludge. In addition, the relative distribution between 

the antimycotics based on sales statistics (total sales nonand 

institutional care) in comparison to the observed 

results in dewatered digested sludge, with the exception of 

final effluents for fluconazole, can be seen in Fig. 2. 

Although the comparison would be more accurate by using 

raw sewage water and raw sewage particles and sales 

statistics of each catchment area of the STPs, the posttreatment 

findings at the STPs show similar distribution of 

the antimycotics as to the predicted equivalent. 


Fig. 2 – The relative distribution of antimycotics based on 

sales records compared to observed results. Abbreviations: 

STP Um [ Umea˚; STP St [ Stockholm, Henriksdal; STP 

Al [ Alingsa˚s; STP Go [ Gothenburg, Ryaverken; STP 

Bo [ Bollebygd; STP mean [ mean value of the 

distribution of the antimycotics. 

3.3. Fate analysis within Umea˚ sewage treatment plant 

The mass flow results from the fate analysis in Umea˚ STP are 

given in Fig. 3. Again, fluconazole was not detected in any of 

the solid matrices, and its mass flows in the raw sewage water 

Fig. 3 – Mass flows of antimycotics (g day L1 ) within the Umea˚ sewage treatment plant.

656 

and the final effluent were very similar (removal efficiency ca. 

13%). Thus, these results corroborate the indications from the 

screening study that fluconazole is at best marginally affected 

by current sewage water treatments. In contrast, clotrimazole, 

ketoconazole and econazole were all present in the solid flow 

of the raw sewage water, and they were also, together with 

terbinafine, detected in the raw sludge. The results for clotrimazole 

are consistent with data presented by Kahle et al. 

(2008), who found extractable amounts of this substance in 

the suspended solids of raw sewage water of Swiss STPs. 

Clotrimazole, ketoconazole, econazole and terbinafine were 

not strongly affected by sludge digestion either, and all of 

them (apart from terbinafine) were found at levels exceeding 

their respective LOQs in the resultant sludge. Traces of 

miconazole were also observed post-digestion. 

Our results show that fluconazole is released at significant 

levels into the aquatic environment, and although this 

substance is reportedly the least effective CYP19 aromatase 

inhibitor of the azole antimycotics (Trösken et al., 2004), it is 

a moderate Cyp3A4 and a potent Cyp2C9 inhibitor, and thus 

may have significant direct ecotoxicological effects. The 

possibility that it may also have adverse indirect environmental 

effects cannot be excluded, since (for example) ketoconazole 

has been shown to increase the sensitivity of 

rainbow trout to ethynylestradiol (Hasselberg et al., 2008). 

However, sludge is the major route for antimycotics into the 

environment, and thus their exposure levels and potential 

negative effects will depend on the fate of the sludge. Sludge is 

commonly used as a fertilizing agent (e.g. for golf courses and 

forests) or as landfill material, so we recommend that future 

environmental risk assessments of this group of pharmaceuticals, 

not only focus on the aqueous environment, but also 

consider their potential impact on terrestrial organisms. 


The levels of fluconazole in the raw sewage water were 

similar to those of the final effluent and in addition it was 

not observed in digested dewatered sludge. 

Terbinafine, clotrimazole, ketoconazole, econazole and 

miconazole were only present in sludge samples, in turn, 

not strongly affected by sludge digestion. 

Observed antimycotic mass flows were in general similar to 

their respective predicted equivalents at each STP except for 

terbinafine and for the smallest STP included in this study. 

Future environmental risk assessments of antimycotics 

should consider both the aqueous and the terrestrial 

environment. 


We gratefully acknowledge financial support from the 

Swedish Research Council for Environment, Agricultural 

Sciences and Spatial Planning (FORMAS). We also thank the 

personnel at the sewage treatment plants for their assistance. 


references 

Carballa, M., Omil, F., Lema, J.M., Llompart, M., Garcia-Jares, C., 



treatment plant. Water Res. 38, 2918–2926. 

EPI SuiteÔ was Developed by the EPA’s Office of Pollution 

Prevention Toxics and Syracuse Research Corporation (SRC), 

2000. U.S. Environmental Protection Agency. 

European Medicines Agency, 2006. Pre-authorisation Evaluation 

of Medicines for Human Use. http://www.emea.europa.int/ 

pdfs/human/swp/444700en.pdf (accessed 11.12.2008). 

Göbel, A., Thomsen, A., McArdell, C.S., Joss, A., Giger, W., 2005a. 

Occurrence and sorption behaviour of sulfonamides, 



Göbel, A., McArdell, C.S., Suter, M.J.-F., Giger, W., 2005b. Trace 

determination of macrolide and sulfonamide antimicrobials, 

a human sulfonamide metabolite, and trimethoprim in 

wastewater using liquid chromatography coupled to 

electrospray tandem mass spectrometry. Anal. Chem. 76, 

4756–4764. 

Golet, E.M., Xifra, I., Siegrist, H., Alder, A.C., Giger, W., 2003. 

Environmental exposure assessment of fluoroquinolone 

antibacterial agents from sewage to soil. Environ. Sci. Technol. 

37, 3243–3249. 

Graybill, J.R., Drutz, D.J., 1980. Ketoconazole – a major innovation 

for treatment of fungal disease. Ann. Int. Med. 93, 921–923. 

Guengerich, P.F., 2008. Cytochrome P450 and chemical toxicology. 

Chem. Res. Toxicol. 21, 70–83. 

Gunnarsson, L., Jauhiainen, A., Kristiansson, E., Nerman, O., 

Larsson, D.G.J., 2008. Evolutionary conservation of human 

drug targets in organisms used for environmental risk 

assessments. Environ. Sci. Technol. 42, 5807–5813. 

Gyllenhammar, I., Eriksson, H., Söderqvist, A., Lindberg, R.H., 

Fick, J., Berg, C., 2009. Clotrimazole exposure modulates 

aromatase activity in gonads and brain during gonadal 

differentiation in Xenopus tropicalis frogs. Aquat. Toxicol. 91, 

102–109. 

Hasselberg, L., Westerberg, S., Wassmur, B., Celander, M.C., 2008. 

Ketoconazole, an antifungal imidazole, increases the 

sensitivity of rainbow trout to 17a-ethynylestradiol exposure. 

Aquat. Toxicol. 86, 256–264. 

Kahle, M., Buerge, I.J., Hauser, A., Müller, M.D., Poiger, T., 2008. 

Azole fungicides: occurrence and fate in wastewaters and 

surface waters. Environ. Sci. Technol. 42, 7193–7200. 

Kolpin, D.W., Skopec, M., Meyer, M.T., Furlong, E.T., Zaugg, S.D., 

2004. Urban contribution of pharmaceuticals and other 

organic wastewater contaminants to streams during differing 

flow conditions. Sci. Total Environ. 328, 119–130. 

Lepesheva, G.I., Waterman, M.R., 2007. Sterol 14 

alpha-demethylase cytochrome P450 (CYP51), a P450 in all 

biological kingdoms. Biochim. Biophys. Acta-General Subjects 

1770, 467–477. 

LIF, 2008. Läkemedelsindustriföreningen (the Swedish 

association of the pharmaceutical industry). FASS. Elanders, 

Kungsbacka. 

Lindberg, R.H., Wennberg, P., Johansson, M.I., Tysklind, M., 

Andersson, B.A.V., 2005. Screening of human antibiotic 

substances and determination of weekly mass flows in five 

sewage treatment plants in Sweden. Environ. Sci. Technol. 39, 

3421–3429. 

Lindberg, R.H., Olofsson, U., Rendahl, P., Johansson, M.I., 

Tysklind, M., Andersson, B.A.V., 2006. Behavior of 

fluoroquinolones and trimethoprim during mechanical, 

chemical and active sludge treatment of sewage water and 

digestion of sludge. Environ. Sci. Technol. 40, 1042–1048.

Peschka, M., Roberts, P.H., Knepper, T.P., 2007. Analysis, fate 

studies and monitoring of the antifungal agent clotrimazole in 

the aquatic environment. Anal. Bioanal. Chem. 389, 959–968. 

Piferrer, F., Zanuy, S., Carillo, M., Solar, I.I., Devlin, R.H., 

Donaldson, E.M., 1994. Brief treatment with an aromatase 

inhibitor during sex differentiation causes chromosomally 

female salmon to develop as normal, functional males. J. Exp. 

Zool. 270, 255–262. 

Roberts, P.H., Bersuder, P., 2006. Analysis of OSPAR priority 

pharmaceuticals using high-performance liquid 

chromatography–electrospray ionisation tandem mass 



pharmaceuticals in wastewater effluent and surface waters of 

the lower Tyne catchment. Sci. Total Environ. 356, 143–153. 

Ryder, N.S., 1992. Terbinafine: mode of action and properties of the 

squalene epoxidase inhibition. Br. J. Dermatol. 126 (Suppl. 39), 2–7. 

Statistics Sweden. http://www.scb.se/templates/ 

tableOrChart____78315.asp (accessed 02.10.08). 

Svensson, Elin, 2008. Apoteket AB, Sweden. Personal 

communication. 


Sweetman, S.C. (Ed.), 2007. Martindale: The Complete Drug 

Reference. Pharmaceutical Press, London Electronic version, 

Edition 070214. 

Thomas, K.V., Hilton, M.J., 2004. The occurrence of selected 

human pharmaceutical compounds in UK estuaries. Mar. 

Pollut. Bull. 49, 436–444. 

Tolls, J., 2001. Sorption of veterinary pharmaceuticals in soils: 

a review. Environ. Sci. Technol. 35, 3397–3406. 

Trösken, E.R., Scholz, K., Lutz, R.W., Volkel, W., Zarn, J.A., Lutz, W. 

K., 2004. Comparative assessment of the inhibition of 

recombinant human CYP19 (aromatase) by azoles used in 

agriculture and as drugs for humans. Endocr. Res. 30, 387–394. 

Van De Steene, J.C., Lambert, W.E., 2008. Validation of a solidphase 

extraction and liquid chromatography–electrospray 

tandem mass spectrometric method for the determination of 

nine basic pharmaceuticals in wastewater and surface water 

samples. J. Chromatogr. A 1182, 153–160. 

Yoshida, Y., Aoyama, Y., Noshiro, M., Gotoh, O., 2000. Sterol 14demethylase 

P450 (CYP51) provides a breakthrough for the 

discussion on the evolution of cytochrome P450 gene 

superfamily. Biochem. Biophys. Res. Commun. 273 (3), 799–804.

Pharmaceuticals and personal care products in archived U.S. 

biosolids from the 2001 EPA national sewage sludge survey 

Kristin McClellan, Rolf U. Halden* 

Center for Environmental Biotechnology, The Biodesign Institute at Arizona State University, 1001 S. McAllister Avenue, Tempe, 

AZ 85287-5701, USA 

article info 


Received 1 July 2009 


15 December 2009 

Accepted 21 December 2009 

Available online 4 January 2010 

Keywords: 

Municipal sludge 

Organic wastewater contaminants 


Land application 


abstract 

Pharmaceuticals and personal care products (PPCPs) are 

common contaminants of the environment, and have been 

detected in surface water (Kolpin et al., 2002; Cahill et al., 

2004; Moldovan, 2006; Roberts and Thomas, 2006; Tamtam 

et al., 2008), groundwater (Heberer et al., 2000; Lindsey et al., 




In response to the U.S. National Academies’ call for a better assessment of chemical 

pollutants contained in the approximately 7 million dry tons of digested municipal sludge 

produced annually in the United States, the mean concentration of 72 pharmaceuticals and 

personal care products (PPCP) were determined in 110 biosolids samples collected by the 

U.S. Environmental Protection Agency (EPA) in its 2001 National Sewage Sludge Survey. 

Composite samples of archived biosolids, collected at 94 U.S. wastewater treatment plants 

from 32 states and the District of Columbia, were analyzed by liquid chromatography 

tandem mass spectrometry using EPA Method 1694. Thirty-eight (54%) of the 72 analytes 

were detected in at least one composite sample at concentrations ranging from 0.002 to 

48 mg kg 1 dry weight. Triclocarban and triclosan were the most abundant analytes with 

mean concentrations of 36 8 and 12.6 3.8 mg kg 1 (n ¼ 5), respectively, accounting for 

65% of the total PPCP mass found. The loading to U.S. soils from nationwide biosolids 

recycling was estimated at 210–250 metric tons per year for the sum of the 72 PPCPs 

investigated. The results of this nationwide reconnaissance of PPCPs in archived U.S. 

biosolids mirror in contaminant occurrences, frequencies and concentrations, those 

reported by the U.S. EPA for samples collected in 2006/2007. This demonstrates that PPCP 

releases in U.S. biosolids have been ongoing for many years and the most abundant PPCPs 

appear to show limited fluctuations in mass over time when assessed on a nationwide 

basis. The here demonstrated use of five mega composite samples holds promise for 

conducting cost-effective, routine monitoring on a regional and national basis. 


2001; Fick et al., 2009), and drinking water (Stackelberg et al., 

2004; Loraine and Pettigrove, 2006; Loos et al., 2007; Focazio 

et al., 2008), as well as in agricultural soils subject to land 

application of digested municipal sludge (Kinney et al., 2008; 

Kupper et al., 2004; Wu et al., 2009), also known as biosolids. 

Wastewater treatment plants were identified as one possible 

source for surface water contamination. Over-the-counter 

Abbreviations: DL, Detection limit; EPA, Environmental protection agency; NEBRA, North east biosolids and residuals association; 

NSSS, National sewage sludge survey; PPCP, Pharmaceuticals and personal care products; RPD, Relative percent difference; 

TNSSS, Targeted national sewage sludge survey. 


E-mail addresses: Kristin.McClellan@asu.edu (K. McClellan), Halden@asu.edu (R.U. Halden). 


doi:10.1016/j.watres.2009.12.032

and prescription drugs enter the wastewater via excretion of 

urine and feces containing parental drugs and their conjugates 

as well as other metabolites, or from disposal of 

unwanted or expired medications (Halling-Sorensen et al., 

1998; Fent et al., 2006). Similarly, chemical constituents of 

personal care products may be directly disposed of into 

domestic wastewater. Removal of PPCPs during municipal 

wastewater treatment is rarely complete, thereby creating 

a pathway for entry of these compounds into aquatic environments 

via wastewater reclamation (Halling-Sorensen 

et al., 1998; Ternes, 1998; Daughton and Ternes, 1999; Hirsch 

et al., 1999) and into terrestrial environments via land 

application of biosolids (Ternes et al., 2004a). 

Of the more than 7 million tons of sewage sludge produced 

in the United States in 2004, about 50% was applied to land as 

fertilizer or soil amendment, and 45% was disposed of in 

landfills or as landfill cover (NEBRA, 2007). Terrestrial environments 

can offer effective biological, physical, and chemical 

attenuation mechanisms for manmade pollutants. However, 

they also can act as a source term for chemical migration into 

surface and groundwater from biosolids runoff and leachate. 

Pharmaceutical compounds are designed to be biologically 

active and therefore may have effects on non-target organisms 

even at trace concentrations extant in terrestrial and 

aquatic environments. While acute toxic effects of pharmaceuticals 

on non-target organisms have been investigated for 

some compounds, chronic toxicity and potential subtle environmental 

effects are only scarcely known (Fent et al., 2006). 

Also insufficiently investigated is the effect of mixtures of 

pharmaceuticals on aquatic organisms, although biochemical 

interactions of drugs in humans are well known. Of additional 

concern is the possible uptake of contaminants into food 

crops grown on agricultural fields that were fertilized with 

biosolids (Kumar et al., 2005; Dolliver et al., 2007). Currently, 

no regulation exists in the United States for PPCPs contained 

in biosolids, and a need for more information on the occurrence 

of and risk from these compounds has been noted by 

the National Research Council of the National Academies of 

the United States (National Research Council, 2002). In the 

past, analytical methods were limited, especially for trace 

analyses of complex environmental samples. In 2007, the 

release of U.S. EPA method 1694 (USEPA, 2007a) for the analysis 

of PPCPs in various matrices afforded the opportunity to 

analyze biosolids samples using a standardized protocol. 

The U.S. EPA has performed national sewage sludge surveys 

(NSSS) in 1989, 2001, and 2007. The survey conducted in 2001 

served to evaluate the potential need for regulations of trace 

levels of dioxins and polychlorinated biphenyls (USEPA, 2007b). 

After the 2001 survey was completed, unused samples were 

released to a nationwide repository of biosolids samples now 

maintained at the Biodesign Institute at Arizona State University. 

This investigation evaluated occurrences and concentrations 

of PPCPs in biosolids from the year 2001 to enable risk 

assessments and to establish a national baseline for evaluating 

temporal trends of PPCPs in U.S. biosolids. The analysis 

of composite samples by EPA Method 1694 was employed to 

determine average concentrations of 72 PPCPs in archived 

biosolids collected by the EPA, as a representative sample of 

the more than 16,000 treatment plants located in the contiguous 

United States. 



2.1. Sampling procedure 

Biosolids samples with solid contents between 1% and 30% 

were obtained from 94 wastewater treatment plants in 32 

states and the District of Columbia for the 2001 National 

Sewage Sludge Survey (USEPA, 2007b). They were selected by 

the U.S. EPA to obtain a representative estimate of the 

occurrence of chemical contaminants in sewage sludge that is 

disposed of primarily by land application. Information on the 

exact sampling locations is available in Table S1 of the 

supplementary material (SM). This survey aimed to estimate 

levels of dioxins, dibenzofurans, and coplanar polychlorinated 

biphenyls in biosolids. The sampling was conducted 

by the U.S. EPA between February and March 2001, and 

samples were collected according to sampling procedures 

developed by the U.S. EPA (USEPA, 2001). Samples were only 

taken from fully processed sewage sludges intended for 

disposal. Eighty-nine of the 94 WWTPs had one single system 

for sludge treatment, therefore one sample was collected. Five 

facilities had two systems for treating their sludges, therefore 

two samples were taken from each of these plants. In addition, 

duplicate samples were collected from 15% of facilities 

(14 samples) for precision analysis. This amounted to 113 

samples overall. After completion of the 2001 NSSS, the 

samples were acquired by the Halden laboratory for further 

studies. For the 7-year period between acquisition and analysis 

in the spring of 2008, samples were stored at 20 C. 

2.2. Composite sample preparation 

From the 113 biosolids samples acquired from the EPA, three 

were excluded from analysis because the sample containers 

were broken or compromised; the remaining 110 samples were 

randomly grouped into five groups. Composite samples were 

prepared by weighing out approximately 1 g of dry weight from 

each sample and pooling it to obtain 5 composites each containing 

solids from between 21 and 24 individual samples. 

A duplicate of composite sample #3 was prepared to serve as 

a blind duplicate. 

2.3. Sample analysis 

The samples were analyzed by AXYS Analytical Services (2045 

Mills Road West, Sydney, British Columbia, Canada V8L 3S8) 

according to EPA method 1694 (USEPA, 2007a). For the purpose 

of compound detection, the 72 analytes were divided into four 

groups. All analytes were separated by liquid chromatography 

and detected by tandem mass spectrometry. For compounds 

with a respective labeled analog, the concentration was 

determined using the isotope dilution technique. The corresponding 

concentrations are deemed to be of high quality. For 

compounds where a labeled analog was not available, the 

concentration was determined using an external calibration. 

Quantitative data for analytes not determined by the isotope 

dilution method are judged to be less robust. More detailed 

information on the analysis method and accuracy and precision 

criteria for acceptance of analytical data is available in

660 

supplementary material (SM). Further information on study 

limitations can be found in the discussion section. 

2.4. Quality assurance 

To ensure system and laboratory performance, several tests 

were performed before sample analysis. Calibration accuracy 

was verified using a calibration standard solution with labeled 

and native analytes. Retention times of native and labeled 

compounds had to be within 15 seconds of the respective 

retention time established during the previous calibration. In 

addition, ongoing precision and recovery were ensured. Lab 

blanks were analyzed before each sample analysis. A duplicate 

sample analysis was performed by the lab for each batch 

between 7 and 20 samples. In addition to these standard 

procedures, a blind duplicate was included in the sample set to 

evaluate analysis precision. Precision is expressed as relative 

percent difference (RPD) between each pair of measured 

concentrations. It was calculated using the following equation, 

RPD½%Š ¼ jCsample Cduplicatej 100 

Csample þ Cduplicate 

2 

where C sample and C duplicate are, respectively, the concentration 

detected in the original sample and in its duplicate 

sample, and RPD is the relative percent difference. 

2.5. Modeling of soil and porewater concentration 

For assessment of their potential environmental impact, the 

concentrations of PPCPs after mixing with agricultural soil 

were calculated. The mixing ratio with soil was assumed 

based on the EPA-recommended rate of biosolids application 

of up to 4.5 dry kg per m 2 and by further assuming incorporation 

into soil to a depth of 10 cm (USEPA, 1994). The soil bulk 

density was assumed to equal 1.3 g cm 3 and the biosolids 

bulk density 1.6 g cm 3 . Though soil moisture content is 

known to vary widely depending on soil properties, for the 

purpose of this calculation, a soil moisture content of 22% (v/v) 

was used, as reported by others for agricultural soil (De Lannoy 

et al., 2006). A typical soil organic carbon fraction was 

assumed (Causarano et al., 2008), as was a biosolids organic 

carbon fraction of 0.4 (USEPA, 2007b). The concentration in soil 

and porewater at equilibrium was calculated based on the 

organic carbon fraction of the soil-biosolids mixture and the 

compound specific organic-carbon distribution coefficient 

(K OC) of each analyte (Chalew and Halden, 2008). Calculations 

took into consideration the soil moisture content, as well as 

the volume of biosolids added to the soil, and were conducted 

for an environmentally relevant pH range of 7–9. The 

concentrations of PPCPs in porewater of soil/biosolid mixtures 

were calculated using the equations below, 

mbiosolids 

Cbiosolids 

msoilþbiosolids 

Cporewater ¼ 

fporewater rporewater þ KOC fOC 


Csoil ¼ mbiosolids 

Cbiosolids 


fporewater rporewater Cporewater 


where m is the dry mass in kg m 3 of the solid matrix, C the 

concentration in mgkg 1 , r the density in kg m 3 , and fporewater 


(1) 

(2) 

(3) 

and f OC the dimensionless fractions of, respectively, porewater 

and organic carbon in the soil/biosolids mixture. 

2.6. Drug usage and ecotoxicity data 

Information of prescription drug consumption was obtained 

from Internet sources (www.drugtopics.com). Exotoxicity 

data were taken from EPA’s Ecotox database (www.epa.gov/ 

ecotox). 

2.7. Modeling of annual loading to agricultural soil 

The annual loading of PPCPs contained in biosolids was based 

on a production of 5.6–7 million dry tons of sewage sludge in 

the U.S., of which 50–60% is applied to land (National Research 

Council, 2002; Jones-Lepp and Stevens, 2007; NEBRA, 2007). 

3. Results 

3.1. Data quality assurance 

Lab blanks showed no detections above the detection limit for 

any of the analytes except for ciprofloxacin [61 mg kg 1 dry 

weight (dw); detection limit (DL) 32 mg kg 1 dw] and erythromycin-H2O 

(2.6 mg kg 1 dw; DL 1.5 mg kg 1 dw)]. However, 

concentrations detected in biosolids were 100- and 40-times 

greater than the background level detected in lab blanks. 

Therefore, measured concentrations for both analytes were 

accepted. 

Recovery for all analytes typically was good with an average 

of 112% but some notable outliers were observed leading to 

a range of 12–493% (See Table 1 and Table S2, SM). A total of 

10 analytes exceeded the methods’ lower and upper control 

limits. For 5 analytes (anhydrotetracycline, azithromycin, 

cimetidine, 4-epianhydrochlortetracycline, and 4-epianhydrotetracycline) 

recovery rates were below the method’s lower 

control limit of 40%. The concentrations reported for these 

analytes in biosolids may represent underestimates. The 

recovery rates of 5 analytes (clinafloxacin, enrofloxacin, 

lomefloxacin, ofloxacin, and sarafloxacin) were above the 

method’s upper control limits. The respective concentrations 

reported for these analytes may represent overestimations. All 

of the above mentioned analytes were quantified using labeled 

surrogate standards. In the case of clinafloxacin, enrofloxacin, 

lomefloxacin, ofloxacin, and sarafloxacin, 13 C3- 15 N-ciprofloxacin 

was used as an isotope-labeled surrogate standard. 

The recoveries were determined from spiked quality control 

samples, where recovery of 13 C3- 15 N-ciprofloxacin was below 

the method’s lower control limit. This led to extremely high 

recoveries of the above mentioned analytes. In the biosolids 

samples, recovery of 13 C 3- 15 N-ciprofloxacin was always within 

the method’s control limits. However, the lack of adequate 

recovery of 13 C 3- 15 N-ciprofloxacin in quality control samples 

suggests that concentrations reported in biosolids for clinafloxacin, 

enrofloxacin, lomefloxacin, ofloxacin, and sarafloxacin 

should be interpreted with caution. 

The duplicate analysis revealed 18% relative percent 

difference (RPD) for all analytes. The blind duplicate analysis 

of a subset of 10 analytes revealed 28% RPD. Both RPD values

Table 1 – Analytical results and summary statistics for pharmaceuticals and personal care products detected in U.S. biosolids collected in 2001 and reported on a dry weight basis. 

Substance name CAS RN Detection 

limit 

[mg kg 1 ] 

Anhydrotetracycline g 

Azithromycin g 

Standard 

deviation 

of detection 

limit [mgkg 1 ] 

n ¼ 5 

Frequency 

measured 

[%] 

Maximum 

concentration 

[mg kg 1 ] 

Mean detected 

concentration 

[mg kg 1 ] 

Standard 

deviation 

of mean 

concentration 

[mg kg 1 ] n ¼ 5 

Recovery 

[%] 

Isotope 

dilution 

quantification 

Use Lowest effect 

concentration 

for aquatic 

biota [mg L 1 ] 

Projected 

annual 

land application 

[kg yr 1 ] f 

4496-85-9 74.1 7.4 100 880 392 249 51 Antibiotic 1300–1600 

83905-01-5 5.6 1.6 100 1220 838 224 12 Antibiotic 2800–3500 

Caffeine 58-08-2 59.0 16.4 100 643 248 200 78 Stimulant 

Carbamazepine 298-46-4 5.6 1.6 100 238 163 56.4 139 Anticonvulsant 100 b (Jos 

Chlortetracycline 57-62-5 22.9 6.2 60 43.5 23.4 16.9 159 Antibiotic 36 c (Brain 

Cimetidine g 

0.05 a (Bantle 

et al., 1994) 

et al., 2003) 

et al., 2004) 

51481-61-9 6.4 2.0 100 893 504 208 41 Antacid 1700–2100 

Ciprofloxacin 85721-33-1 20.9 6.4 100 10800 6858 2348 98 Antibiotic 

5 d (Halling- 

Sorensen, 2001) 

Clarithromycin 81103-11-9 5.6 1.6 100 94.6 66.2 25.5 114 Antibiotic 220–270 

Codeine 76-57-3 11.2 3.2 20 29.7 n.d. 114 Analgesic 

Cotinine 486-56-6 9.9 3.3 100 38.6 28.1 8.3 103 

Nicotine 

metabolite 

Diltiazem 42399-41-7 1.3 0.1 100 109 45.2 34.2 110 Antianginal 150–190 

Diphenhydramine 58-73-1 2.3 0.6 100 1740 1166 516 101 Antihistamine 3900–4800 

Doxycycline 564-25-0 23.9 6.3 100 1780 966 436 85 Antibiotic 3200–4000 

Enrofloxacin h 

93106-60-6 19.5 6.5 60 28.6 n.d. 190 Antibiotic 49 d (Robinson 

et al., 2005) 

4-Epianhydrotetracycline 

g 

4465-65-0 77.1 20.2 100 399 261 71.7 39 Antibiotic 900–1100 

4-Epichlortetracycline 

14297-93-9 56.3 16.0 40 93.0 n.d. 161 Antibiotic 

4-Epitetracycline 23313-80-6 74.4 18.8 100 3040 2376 517 82 Antibiotic 8000–9800 

Erythromycin-H 2O 114-07-8 1.6 0.2 100 183 81.5 52.3 97 Antibiotic 

Fluoxetine 54910-89-3 8.2 1.0 100 258 171 46.6 89 Antidepressant 

Gemfibrozil 25812-30-0 12.2 12.3 100 159 152 13.2 107 

Antihyperlipidemic 

22 700 c (Williams 

et al., 1992) 

36 h, c (Flaherty 

and Dodson, 2005) 

30,040 c (Zurita 

et al., 2007) 

Ibuprofen 15687-27-1 122 123 80 359 246 121 109 Anti-inflammatory 830–1000 

Isochlortetracycline 514-53-4 22.5 6.4 60 36.0 n.d. 70 Antibiotic 

Lomefloxacin h 

98079-51-7 13.2 2.1 40 16.1 n.d. 388 Antibiotic 106 e (Robinson 

et al., 2005) 

Metformin 657-24-9 119 32.5 80 456 305 152 116 Antidiabetic 1000–1300 

Miconazole 22916-47-8 5.8 1.3 100 1100 777 266 86 Antifungal 2600–3200 

Minocycline 10118-90-8 1880 524 80 2630 1884 939 54 Antibiotic 6300–7800 

830–1000 

550–680 

80–100 

23,000–28,000 

100–120 

270–340 

580–710 

510–630 



Substance name CAS RN Detection 

limit 

[mg kg 1 ] 

Standard 

deviation 

of detection 

limit [mgkg 1 ] 

n ¼ 5 

Frequency 

measured 

[%] 

Maximum 

concentration 

[mg kg 1 ] 

Mean detected 

concentration 

[mg kg 1 ] 

Standard 

deviation 

of mean 

concentration 

[mg kg 1 ] n ¼ 5 

Naproxen 22204-53-1 24.3 24.6 100 273 119 79 106 

Recovery 

[%] 

Isotope 

dilution 

quantification 

Antiinflammatory 

Use Lowest effect 

concentration 

for aquatic 

biota [mg L 1 ] 

Projected 

annual 

land application 

[kg yr 1 ] f 

Norfloxacin 70458-96-7 63.5 6.3 100 418 289 74.0 136 Antibiotic 18 000 b (Yang 

et al., 2008) 

970–1200 

Ofloxacin h 

82419-36-1 7.7 4.2 100 8140 5446 1941 493 Antibiotic 21 d (Robinson 

et al., 2005) 

18000–23,000 

Oxytetracyclin 79-57-2 22.8 5.8 100 114 87.5 22.2 90 Antibiotic 50 d (Hanson 

et al., 2006) 

300–360 

Ranitidine 66357-35-5 6.5 1.7 100 30.1 21.0 8.0 51 Antacid 70–90 

Sulfamethoxazole 723-46-6 2.9 0.7 20 3.3 n.d. 98 Antibiotic 

9 e (Brain 

et al., 2008) 

Sulfanilamide 63-74-1 56.2 16.0 40 87.3 n.d. 101 Antibiotic 

Tetracycline 60-54-8 57.5 15.5 100 2790 1914 691 124 Antibiotic 47 e (Brain 

et al., 2004) 

Thiabendazole 148-79-8 5.9 1.2 100 370 110 131 103 Fungicide 310 c (EPA, 2000) 370–460 

Triclocarban 101-20-2 183 67.0 100 48100 36060 8049 96 Disinfectant 

Triclosan 3380-34-5 487 493 100 19700 12640 3816 105 Disinfectant 

Trimethoprim 738-70-5 10.6 3.8 60 60.5 26.0 21.5 97 Antibiotic 

0.101 c 

(EPA/OTS, 1992) 

0.12 b (Wilson 

et al., 2003) 

16 000 b (Luetzhoft 

et al., 1999) 

400–500 

6400–7900 

120,000–150,000 

42,000–52,000 

a Amphibium. 

b Green algae. 

c Crustacean. 

d Cyanobacteria. 

e Macrophyte. 

f Projected deposition rates of PPCPs on land based on 5.6–6.9 million dry tons of annual sewage sludge production and a land application rate of 60% (National Research Council, 2002; USEPA, 2003; USGS, 2006; 

Jones-Lepp and Stevens, 2007). 

g Concentration may be underestimated due to low recovery in quality control samples. 

h Concentration may be overestimated due to high recovery in quality control samples. 

90–110

were in control with respect to the target RPD of 30% or less. 

The RPD value improved to 11% for 9 analytes, when excluding 

results for metformin, whose lack of detection in the blind 

duplicate unfavorably increased the summary statistic for 

measurement precision. A non-blinded duplicate analysis, 

performed by the contract laboratory for these 10 analytes, 

showed an RPD value of 11%. 

3.2. Study representativeness and sample integrity 

The prolonged storage of samples between sampling event 

and analysis may have allowed for the chemical degradation 

of labile analytes to occur. Therefore, any results of this study 

are conservative with respect to the detection frequency and 

concentration of compounds found. In other words, due to 

pooling of a large number of samples, analytes occurring 

infrequently and at low concentrations may have been diluted 

out to below the detection limit. A comparison of the presented 

data with the EPA data (USEPA, 2009), reveals that the 

mean concentrations of all analytes show no statistically 

significant difference within the 95th percentile confidence 

interval. Therefore, the prolonged storage did not impair the 

detection of multiple analytes at elevated concentrations in 

archived samples. 

3.3. Occurrence of PPCPs in biosolids 

All composites tested positive for at least 26 analytes. Of the 72 

PPCPs targeted, 38 (54%) were detected at concentrations 

ranging from the low parts-per-billion (ppb) to the parts-permillion 

(ppm) range. These 38 PPCPs include 8 that have not 

previously been reported in biosolids in the peer-reviewed 

literature but that also were observed in the U.S. EPA’s TNSSS 

published online (USEPA, 2009). The remaining 34 PPCPs were 

Concentration in biosolids [µg kg -1 dry weight] 

100 

000 

10 

000 

1000 

100 

10 

1 

triclocarban 

triclosan 

‡ 

ciprofloxacin 

ofloxacin 

* 

4-epitetracycline 

tetracycline 

minocycline 


‡ 

diphenhydramine 

doxycycline 

azithromycin 

miconazole 

cimetidine 

‡ 

‡ 

* 

anhydrotetracycline 

metformin 

norfloxacin 

‡ 

* 

not detected in any of the composite biosolids samples. The 

mean total concentration of all targeted PPCPs combined in 

the five composite samples was 74.4 mg kg 1 dw of sewage 

sludge 21.4 mg kg 1 standard deviation (n ¼ 5). 

The two most abundant contaminants were the disinfectants 

triclocarban (48% of total detected PPCP mass) and triclosan 

(17%) (Fig. 1). Their mean concentrations were 36 8 

and 12.6 3.8 mg kg 1 dw (n ¼ 5), respectively (Table 1). The 

second most abundant class of PPCPs found was antibiotics. In 

order of decreasing concentration, ciprofloxacin, ofloxacin, 4epitetracycline, 

tetracycline, minocycline, doxycycline and 

azithromycin were found at concentrations between 6.8 2.3 

and 0.8 0.2 mg kg 1 dw (n ¼ 5) (Table 1). The combined mass 

of all antibiotics constituted about 29% of the total mass of 

PPCPs per sample. 

In addition to the 3 tetracyclines reported above, three 

tetracycline antibiotics were found for which peer-reviewed 

occurrence data in sewage sludge thus far were lacking. 

Anhydrotetracycline, 4-epianhydrotetracycline, and chlortetracycline 

(in order of decreasing concentration) together 

accounted for 6.9 2.5 mg kg 1 dw (n ¼ 5) (Table 1). Also not 

previously reported in biosolids were two prescription drugs, 

metformin and ranitidine. They together accounted for 0.4% 

of the total mass of PPCPs found in sewage sludge composites. 

4. Discussion 

4.1. Study limitations 

For this study, a relatively large number of individual samples 

were combined to form 5 composite pools or mega composites. 

This approach served to reduce the number of samples to be 

analyzed in order to obtain a defensible estimate of mean 

* 

4-epianhydrotetracycline 

caffeine 

ibuprofen 

fluoxetine 

carbamazepine 

gemfibrozil 

naproxen 

thiabendazole 

oxytetracycline 

erythromycin-H2O clarithromycin 

diltiazem 

sulfanilamide 

* 

* 

4-epichlortetracycline 

cotinine 

trimethoprim 

chlortetracycline 

ranitidine 

* 

‡* 

ppm 

ppb 

* 

‡* 

isochlortetracycline 

enrofloxacin 

codeine 

lomefloxacin 

Fig. 1 – Rank order of mean concentrations for 38 PPCPs detected in composites of a total of 110 U S biosolids samples from 

94 treatment plants in 32 states and the District of Columbia. Newly detected compounds are shown in darker hue. Error 

bars depict ± one standard deviation (n [ 5). Some concentrations represent estimates only (z) and some analytes were 

detected inconsistently (*). 

sulfamethoxazole *

664 

analyte concentrations across the various treatment facilities 

represented. While being efficient and economical for the 

intended purpose, this approach was not well suited to capture 

the full spectrum of concentrations of individual PPCPs as 

a function of plant type, treatment processes employed, populations 

served, as well as of geographical locations and 

climate zones represented. As a comparison with the EPA 

TNSSS data reveals (USEPA, 2009), the detection frequency of 

less abundant analytes was significantly reduced in composite 

samples compared to individual sample analysis. Therefore, 

analytes that were not detected will represent a conservative 

estimate, and may still occur at detectable concentrations in 

individual samples from specific plants. While the mega 

composite approach cannot serve to determine variability 

between the large numbers of WWTPs studied, it was found to 

be suitable for identifying major contaminants of concern as 

well as their average concentration in a large sample set. 

4.2. Sanitizing agents 

Triclocarban, which in previous U.S. studies had been found in 

concentrations ranging from 5.97 to 51 mg kg 1 dry weight 

(dw) (Heidler et al., 2006; Chu and Metcalfe, 2007; Sapkota et al., 

2007), was detected in every composite sample assayed. Also 

found in significant amounts was triclosan, which has been 

observed in a number of U.S. studies of sewage sludge with 

reported concentrations ranging from 0.53 to 30 mg kg 1 dw 

(Chu and Metcalfe, 2007; Heidler and Halden, 2007; Kinney 

et al., 2008; McAvoy et al., 2002; USEPA, 2003). The EPA TNSSS 

(USEPA, 2009) found triclocarban and triclosan at concentrations 

of up to 441 and 133 mg kg 1 dw, respectively, with mean 

concentrations at 38.7 59.7 and 12 18 mg kg 1 dw (n ¼ 74), 

respectively. The relatively high mean concentrations of triclocarban 

and triclosan reported here and by the U.S. EPA 

(USEPA, 2009) in U.S. biosolids are in line with the intense 

usage of these antimicrobials and their high octanol/water 

partitioning coefficient (log KOW) of 4.9 and 4.8, respectively 

(both at neutral pH), which indicates significant potential of 

both compounds for sorption to biosolids (Halden and Paull, 

2005). In addition, triclosan was found to be persistent in 

biosolids after aeration (Ying et al., 2007). Both triclosan and 

triclocarban concentrations found in the present study fall 

within the range of concentrations previously reported. 

Information on previously reported concentrations of PPCPs in 

biosolids is available in Table S2, SM. 

4.3. Antibiotics 

The most abundant antibiotic was ciprofloxacin (6.8 2.3 mg 

kg 1 dw; n ¼ 5), which is among the 30 most prescribed drugs 

in the United States, according to Internet sources. Ciprofloxacin, 

a metabolite of enrofloxacin, is polar and therefore 

prone to electrostatic interactions with the negatively charged 

surfaces of microbes that are found in high concentrations 

especially in secondary sludge (Ternes et al., 2004b). Ciprofloxacin 

has been detected in sewage sludge by several studies 

conducted in Sweden and Switzerland. Detected concentrations 

ranged from 6 10 5 –11 mg kg 1 dw (Golet et al., 2002; 

Lindberg et al., 2005, 2006, 2007). Therefore, the mean 

concentration found in the present study (6.8 2.3 mg kg 1 


dw; n ¼ 5) falls in the mid range of concentrations reported 

from outside of the U.S. The EPA TNSSS also identified ciprofloxacin 

as an abundant microcontaminant in sewage sludge, 

which was detected in every sample analyzed, yielding 

a mean concentration of 8.7 8.5 mg kg 1 dw. 

At a mean concentration of 5.4 1.9 mg kg 1 dw, ofloxacin 

was the fourth most abundant contaminant found in biosolids. 

It is fairly hydrophilic (log KOW of 0.2) and does not appear on 

the list of the top 200 prescription drugs. Its detection at 

elevated levels in every composite sample came as a surprise. 

However, ofloxacin had been found in concentrations of up to 

2mgkg 1 dw in biosolids from Sweden (Lindberg et al., 2005). It 

can be speculated that the carboxyl group contained in ofloxacin 

may form an ion complex with exchangeable cations 

associated with negatively charged surfaces. 

Among the 10 most abundant PPCPs of the present study 

were three tetracycline antibiotics, 4-epitetracycline, tetracycline 

and minocycline, which had not been reported in 

biosolids before in the peer-reviewed literature. In addition, 

doxycycline was found as the ninth most abundant PPCP at 

a mean concentration of 1 0.4 mg kg 1 dw. It had not been 

detected in biosolids from the U.S. before and only once in 

a Swedish study that found concentrations similar to those 

reported here (Lindberg et al., 2005). All tetracycline antibiotics 

are fairly hydrophilic (log KOW of 1.33 for tetracycline/4-epitetracycline, 

0.42 for minocycline, and 1.36 for doxycycline). 

Despite their hydrophilic character, the mean concentrations 

reported here are around 1–2 mg kg 1 dw. Tetracycline antibiotics 

are known to precipitate with ions of magnesium, calcium 

and ferric iron, and therefore accumulate in the solid fraction 

during wastewater treatment. Tetracycline, doxycycline and 

minocycline also rank among the 200 prescription drugs most 

widely used in the U.S. Other tetracycline antibiotics analyzed 

in this study were only found at mean concentrations below 

0.4 mg kg 1 dw, but the concentrations of anhydrotetracycline, 

4-epianhydrotetracycline and 4-epianhydrochlortetracycline 

were possibly underestimated due to low recoveries. The sum 

of tetracycline antibiotics found in sewage sludge constitute 

about 8 1.3mgkg 1 dw, which is similar to the findings of the 

EPA TNSSS that found about 5 mg kg 1 dw (USEPA, 2009). 

Azithromycin was found at 0.8 0.2 mg kg 1 dw. It ranked 

as the 6th most frequently prescribed drug in 2007 and is also 

fairly hydrophobic. Due to both these properties and the fact 

that biodegradation of azithromycin was found to be insignificant 

(Ericson, 2007), one would expect high concentrations 

in biosolids. Yet, there are 7 drugs (not counting triclocarban 

and triclosan) that were found at higher concentrations. 

However, a low recovery of azithromycin of only 12% may 

indicate that actual azithromycin concentrations in biosolids 

are much higher than the detected concentration. Azithromycin 

has been detected in previous studies at concentrations 

of up to 6.5 mg kg 1 dw in the U.S. (Jones-Lepp and 

Stevens, 2007) and at up to 0.16 mg kg 1 dw in sludge from 

Germany and Switzerland (Gobel et al., 2005a, 2005b). 

4.4. Bioavailability and soil/porewater equilibria 

To explore the importance of bioavailability of PPCPs 

sequestered in biosolids, the concentrations of individual 

PPCPs that are anticipated to occur in the solid and liquid

phases upon mixing of land applied biosolids into soil were 

calculated (Fig. 2). In fully equilibrated soil-biosolids mixtures 

(assuming an EPA-recommended mixing ratio of approximate 

25:1), concentrations of PPCPs on soil particles are expected to 

fall into the ppb range for most analytes, except for triclocarban, 

which was projected to be present at around 1 ppm(w/w) 

dw. Since most of the PPCPs found in biosolids are fairly 

hydrophobic, their calculated concentrations in porewater at 

equilibrium typically were quite low (Fig. 2 B) (Kinney et al., 

2008). A comparison of these estimated dissolved PPCP levels 

in porewater with the lowest effect concentrations for aquatic 

organisms (red circles in Fig. 2 B) suggests that the leaching of 

dissolved PPCPs into surface waters probably does not present 

an important mechanism for exposure of aquatic biota for the 

majority of analytes detected. Concentrations calculated for 

soil porewater typically were several orders of magnitude 

below the lowest effect concentration reported for aquatic 

organisms. Notable exceptions were 6 analytes (Fig. 2) that are 

expected to yield potentially problematic concentrations in 

porewater after land application and partitioning of biosolidsderived 

PPCPs. These include the antibiotics ciprofloxacin, 

ofloxacin and tetracycline, as well as the stimulant caffeine, 

Conc. in dry weight soil at 

pH 7-9 [µg kg-1 ] 

Conc. in porewater 

at pH 7-9 [µg L-1 ] 

1000 

100 

10 

1 

0. 

1 

0. 

01 

100 

000 

10 

000 

1000 

100 

10 

1 

0. 

1 

0. 

01 

0. 

001 

0. 

0001 

‡ 

* 

‡ 

‡ 

* 

‡ 

* 

‡* 

and the two sanitizing agents triclosan and triclocarban 

(Fig. 2). 

4.5. Risk assessment data gaps 

These results suggest that aside from the 7 notable exceptions 

discussed above, the majority of the PPCPs detected in 

biosolids in this study likely exert no acute effects on aquatic 

organisms, assuming that biosolids are applied as regulated 

by the EPA (USEPA, 1994) and that the migration of solids from 

agricultural land to surface water via soil erosion and runoff 

can be completely prevented. While the former assumption is 

plausible, the latter may not always apply, as soil erosion is 

a common phenomenon. 

However, chronic toxicity as well as effects from 

mixtures of PPCPs on non-target organisms cannot be 

assessed due to lack of appropriate toxicity data. It has been 

speculated that the presence of sub-therapeutic concentrations 

of antibiotics may adversely affect soil microbial 

community structures as well as induce spreading of resistance 

among bacterial pathogens. In addition to influencing 

microbial populations, it has been shown that some 

* 

* 

‡* 

* 




anhydrotetracycliine 

azithromycin 

caffeine 

carbamazepine 


cimetidine 

ciprofloxacin 

clarithromycin 

codeine 

cotinine 

diltiazem 


doxycycline 

enrofloxacin 

erythromycin-H2O fluoxetine 

gemfibrozil 

ibuprofen 


lomefloxacin 

metformin 

miconazole 

minocycline 

naproxen 

norfloxacin 

ofloxacin 


ranitidine 

sulfamethoxazole 

sulfanilamide 

tetracycline 

thiabendazole 

triclocarban 

triclosan 

trimethoprim 

‡ 

* 

‡ 

‡ 

* 

‡ 


* 

‡* 

* 

* 

‡* 

* 




anhydrotetracycliine 

azithromycin 

caffeine 

carbamazepine 


cimetidine 

ciprofloxacin 

clarithromycin 

codeine 

cotinine 

diltiazem 


doxycycline 

enrofloxacin 

erythromycin-H2O fluoxetine 

gemfibrozil 

ibuprofen 


lomefloxacin 

metformin 

miconazole 

minocycline 

naproxen 

norfloxacin 

ofloxacin 


ranitidine 

sulfamethoxazole 

sulfanilamide 

tetracycline 

thiabendazole 

triclocarban 

triclosan 

trimethoprim 

* 

* 

‡ 

‡ 

* 

* 

* 

* 

* 

* 

100 

000 

10 

000 

1000 

100 

10 

1 

0. 

1 

0. 

01 

0. 

001 

0. 

0001 

Fig. 2 – Predicted equilibrium concentrations of PPCPs associated with particulates of soil-biosolids mixtures (top) and 

dissolved in porewater (bottom) after land application of biosolids on agricultural soil. Data depict the environmentally 

relevant range between pH 7 and 9. Circles represent the lowest ecological effect concentrations contained in the EPA Ecotox 

database. Some concentrations were calculated based on estimates only (z) and some analytes were detected inconsistently (*). 

Effect concentration [µg L -1 ]

666 

antibiotics, specifically drugs belonging to the tetracyclines, 

fluoroquinolones and sulfonamides, may be taken up by 

crop plants (Migliore et al., 2003; Kumar et al., 2005; Dolliver 

et al., 2007). This presents a potential exposure pathway to 

humans through ingestion of contaminated food and may 

result in the promotion of resistant bacteria in humans 

(Shoemaker et al., 2001). 

Furthermore, information is lacking to determine risks to 

the health of agricultural soils and soil-dwelling organisms. 

Few studies have examined the half-lives in soils of PPCPs 

sequestered in biosolids. The effect of ppm levels of sanitizing 

agents on soil microbial communities has rarely been investigated 

to date (Liu et al., 2009). Similarly, the half-life of PPCPs 

in biosolids-amended soils will require additional research to 

inform risk assessment analyses. Passage through municipal 

digesters and chemical aging may reduce the bioavailability of 

PPCPs and with it the risk of chemical uptake of and exposure 

to soil-dwelling organisms. However, a reduced bioavailability 

also may imply a prolonged half-life of these compounds in 

the environment, with possible delayed release in sensitive 

compartments. Furthermore, studies are lacking on the 

potential of biosolids-derived antimicrobials and antibiotics to 

exert selective pressure for the enrichment of drug-resistant 

microorganisms, a scenario demonstrated in vitro (Braoudaki 

and Hilton, 2004). Also unavailable are threshold concentrations 

for toxic effects in terrestrial organisms of many PPCPs, 

including some that have been detected in biosolids at ppm 

levels. The EPA’s Ecotox database (www.epa.gov/ecotox) 

provides only some toxicity values in aquatic organisms for 

the PPCPs investigated here. Bioaccumulation and biomagnification 

are other aspects that will need to be considered 

for risk assessment purposes. Although PPCPs are not 

typically thought of as representing persistent hydrophobic 

pollutants, some may be subject to bioaccumulation and 

possibly biomagnification thereafter in both terrestrial and 

aquatic environments. Bioaccumulation was demonstrated 

for triclosan and triclocarban in lab and field studies that 

examined uptake of these compounds from soil, sediment 

and water (Coogan and La Point, 2008; Higgins et al., 2009; 

Kinney et al., 2008). 


Overall, this study reemphasizes the significance of biosolids 

recycling as a mechanism for the release of PPCPs into the 

environment. Based on the mean concentrations of all analytes 

detected, it is estimated that the total loading to U.S. soils 

from nationwide biosolids recycling is on the order of 

210–250 metric tons per year for the 72 PPCPs investigated 

here. 

It is concluded that, despite large variations found by the 

U.S. EPA between different treatment plants, mean concentrations 

of PPCPs in U.S. biosolids on a nationwide basis have 

remained fairly constant between 2001 and 2007 and possibly 

longer. Good agreement between this and the U.S. EPA study 

further suggests that the here demonstrated use of mega 

composite samples represents a cost-effective approach for 

collecting regional and nationwide information on average 

concentrations of contaminants in biosolids. 



We thank Rick Stevens and Harry B. McCarty from the U.S. EPA 

for providing samples from the 2001 National Sewage Sludge 

Survey. We also thank Randhir Deo and Jochen Heidler for 

their valuable input. This work was supported in part by the 

National Institute of Environmental Health Sciences by NIEHS 

grant 1R01ES015445 and by the Johns Hopkins University 

Center for a Livable Future. 

Appendix. 

Supplementary material 

Additional details regarding experimental procedures and 

results can be found in the Supplementary Material accompanying 

this article. This information is available free of 

charge via the Internet at www.iwaponline.com. 

Supplementary data associated with this article can be 

found in the online version at doi:10.1016/j.watres.2009.12. 

032. 

references 

Bantle, J.A., Burton, D.T., Dawson, D.A., Dumont, J.N., Finch, R.A., 

Fort, D.J., Linder, G., Rayburn, J.R., Buchwalter, D., 

Gaudethull, A.M., Maurice, M.A., Turley, S.D., 1994. Fetax 

interlaboratory validation study - Phase II testing. 


Brain, R.A., Johnson, D.J., Richards, S.M., Sanderson, H., Sibley, P.K., 

Solomon, K.R., 2004. Effects of 25 pharmaceutical compounds 

to Lemna gibba using a seven-day static-renewal test. 


Brain, R.A., Ramirez, A.J., Fulton, B.A., Chambliss, C.K., Brooks, B. 

W., 2008. Herbicidal Effects of Sulfamethoxazole in Lemna 

gibba: Using p-Aminobenzoic Acid As a Biomarker of Effect. 


Braoudaki, M., Hilton, A.C., 2004. Low level of cross-resistance 

between triclosan and antibiotics in Escherichia coli K-12 and 

E. coli O55 compared to E. coli O157. Fems Microbiology 

Letters 235, 305–309. 

Cahill, J.D., Furlong, E.T., Burkhardt, M.R., Kolpin, D., Anderson, L.G., 

2004. Determination of pharmaceutical compounds in surfaceand 

ground-water samples by solid-phase extraction and highperformance 

liquid chromatography-electrospray ionization 

mass spectrometry. Journal of Chromatography A 1041, 171–180. 

Causarano, H.J., Franzluebbers, A.J., Shaw, J.N., Reeves, D.W., 

Raper, R.L., Wood, C.W., 2008. Soil organic carbon fractions 

and aggregation in the Southern Piedmont and Coastal Plain. 

Soil Science Society of America Journal 72, 221–230. 

Chalew, T., Halden, R., 2008. Environmental exposure of aquatic 

and terrestrial biota to triclosan and triclocarban. Journal of 

the American Water Resources Association 45 (1), 4–13. 

Chu, S.G., Metcalfe, C.D., 2007. Simultaneous determination of 

triclocarban and triclosan in municipal biosolids by liquid 

chromatography tandem mass spectrometry. Journal of 

Chromatography A 1164, 212–218. 

Coogan, M.A., La Point, T.W., 2008. Snail bioaccumulation of 

triclocarban, triclosan, and methyltriclosan in a North Texas, 

USA, stream affected by wastewater treatment plant runoff. 

Environmental Toxicology and Chemistry 27, 1788–1793.



Environmental Health Perspectives 107, 907–938. 

De Lannoy, G.J.M., Verhoest, N.E.C., Houser, P.R., Gish, T.J., Van 

Meirvenne, M., 2006. Spatial and temporal characteristics of 

soil moisture in an intensively monitored agricultural field 

(OPE3). Journal of Hydrology 331, 719–730. 

Dolliver, H., Kumar, K., Gupta, S., 2007. Sulfamethazine uptake by 

plants from manure-amended soil. Journal of Environmental 

Quality 36, 1224–1230. 

Ericson, J.F., 2007. An evaluation of the OECD 308 water/sediment 

systems for investigating the biodegradation of 

pharmaceuticals. Environmental Science & Technology 41, 

5803–5811. 

Fent, K., Weston, A.A., Caminada, D., 2006. Ecotoxicology of 

human pharmaceuticals. Aquatic Toxicology 76, 122–159. 

Fick, J., Soderstrom, H., Lindberg, R.H., Phan, C., Tysklind, M., 

Larsson, D.G.J., 2009. Contamination of surface, 

ground, and drinking water from pharmaceutical 

production. Environmental Toxicology and Chemistry 28, 

2522–2527. 

Focazio, M.J., Kolpin, D.W., Barnes, K.K., Furlong, E.T., Meyer, M.T., 

Zaugg, S.D., Barber, L.B., Thurman, M.E., 2008. A national 

reconnaissance for pharmaceuticals and other organic 

wastewater contaminants in the United States – II untreated 

drinking water sources. Science of the Total Environment 402, 

201–216. 

Gobel, A., Thomsen, A., McArdell, C.S., Alder, A.C., Giger, W., 

Theiss, N., Loffler, D., Ternes, T.A., 2005a. Extraction and 

determination of sulfonamides, macrolides, and 

trimethoprim in sewage sludge. Journal of Chromatography A 

1085, 179–189. 

Gobel, A., Thomsen, A., McArdell, C.S., Joss, A., Giger, W., 2005b. 

Occurrence and sorption behavior of sulfonamides, 



Golet, E.M., Strehler, A., Alder, A.C., Giger, W., 2002. 

Determination of fluoroquinolone antibacterial agents in 

sewage sludge and sludge-treated soil using accelerated 

solvent extraction followed by solid-phase extraction. 

Analytical Chemistry 74, 5455–5462. 

Halden, R.U., Paull, D.H., 2005. Co-occurrence of triclocarban and 

triclosan in US water resources. Environmental Science & 

Technology 39, 1420–1426. 

Halling-Sorensen, B., Nielsen, S.N., Lanzky, P.F., Ingerslev, F., 

Lutzhoft, H.C.H., Jorgensen, S.E., 1998. Occurrence, fate and 

effects of pharmaceutical substances in the environment - 


Halling-Sorensen, B., 2001. Inhibition of aerobic growth and 

nitrification of bacteria in sewage sludge by antibacterial 

agents. Archives of Environmental Contamination and 

Toxicology 40 (4), 451–460. 

Hanson, M.L., Knapp, C.W., Graham, D.W., 2006. Field assessment 

of oxytetracycline exposure to the freshwater macrophytes 

Egeria densa Planch and Ceratophyllum demersum L. 

Environmental Pollution 141 (3), 434–442. 

Heberer, T., Fuhrmann, B., Schmidt-Baumler, K., Tsipi, D., 

Koutsouba, V., Hiskia, A., 2000. Occurrence of pharmaceutical 

residues in sewage, river, ground-, and drinking water in 

Greece and Germany. Abstracts of Papers of the American 

Chemical Society 219 U623. 

Heidler, J., Halden, R.U., 2007. Mass balance assessment of 

triclosan removal during conventional sewage treatment. 


Heidler, J., Sapkota, A., Halden, R.U., 2006. Partitioning, 

persistence, and accumulation in digested sludge of the 

topical antiseptic triclocarban during wastewater treatment. 



Higgins, C.P., Paesani, Z.J., Chalew, T.E.A., Halden, R.U., 2009. 

Bioaccumulation of triclocarban in Lumbriculus variegates. 

Environmental Toxicology and Chemistry 65, 141–148. 

Hirsch, R., Ternes, T., Haberer, K., Kratz, K.L., 1999. Occurrence of 

antibiotics in the aquatic environment. Science of the Total 


Jones-Lepp, T.L., Stevens, R., 2007. Pharmaceuticals and personal 

care products in biosolids/sewage sludge: the interface 

between analytical chemistry and regulation. Analytical and 

Bioanalytical Chemistry 387, 1173–1183. 

Jos, A., Repetto, G., Rios, J.C., Hazen, N., Molero, M.L., del Peso, A., 

Salguero, M., Fernandez-Freire, P., Perez-Martin, J.M., 

Camen, A., 2003. Ecotoxicological evaluation of 

carbamazepine using six different model systems with 

eighteen endpoints. Toxicology in Vitro 17 (5–6), 525–532. 

Kinney, C.A., Furlong, E.T., Kolpin, D.W., Burkhardt, M.R., 

Zaugg, S.D., Werner, S.L., Bossio, J.P., Benotti, M.J., 2008. 

Bioaccumulation of pharmaceuticals and other anthropogenic 

waste indicators in earthworms from agricultural soil 

amended with biosolid or swine manure. Environmental 

Science & Technology 42, 1863–1870. 


Barber, L.B., Buxton, H.T., 2002. Pharmaceuticals, hormones, and 

other organic wastewater contaminants in US streams, 

1999–2000: a national reconnaissance. Environmental Science & 

Technology 36, 1202–1211. 

Kumar, K., Gupta, S.C., Baidoo, S.K., Chander, Y., Rosen, C.J., 2005. 

Antibiotic uptake by plants from soil fertilized with animal 

manure. Journal of Environmental Quality 34, 2082–2085. 

Kupper, T., Berset, J.D., Etter-Holzer, R., Furrer, R., Tarradellas, J., 

2004. Concentraions and specific loads of polycyclic musks in 

sewage sludge originating from a monitoring network in 

Switzerland. Chemosphere 54, 1111–1120. 

Lindberg, R.H., Wennberg, P., Johansson, M.I., Tysklind, M., 

Andersson, B.A.V., 2005. Screening of human antibiotic 

substances and determination of weekly mass flows in five 

sewage treatment plants in Sweden. Environmental Science & 

Technology 39, 3421–3429. 

Lindberg, R.H., Olofsson, U., Rendahl, P., Johansson, M.I., Tysklind, M. 

, Andersson, B.A.V., 2006. Behavior of fluoroquinolones and 

trimethoprim during mechanical, chemical, and active sludge 

treatment of sewage water and digestion of sludge. 


Lindberg, R.H., Bjorklund, K., Rendahl, P., Johansson, M.I., 

Tysklind, M., Andersson, B.A.V., 2007. Environmental risk 

assessment of antibiotics in the Swedish environment with 

emphasis on sewage treatment plants. Water Research 41, 

613–619. 

Lindsey, M.E., Meyer, M., Thurman, E.M., 2001. Analysis of trace 

levels of sulfonamide and tetracycline antimicrobials, in 

groundwater and surface water using solid-phase extraction 

and liquid chromatography/mass spectrometry. Analytical 

Chemistry 73, 4640–4646. 

Liu, F., Ying, G.-G., Yang, L.-H., Zhou, Q.-X., 2009. Terrestrial 

ecotoxicological effects of the antimicrobial agent triclosan. 

Ecotoxicology and Environmental Safety 72, 86–92. 

Loos, R., Wollgast, J., Huber, T., Hanke, G., 2007. Polar herbicides, 

pharmaceutical products, perfluorooctanesulfonate (PFOS), 

perfluorooctanoate (PFOA), and nonylphenol and its 

carboxylates and ethoxylates in surface and tap waters 

around Lake Maggiore in Northern Italy. Analytical and 

Bioanalytical Chemistry 387, 1469–1478. 

Loraine, G.A., Pettigrove, M.E., 2006. Seasonal variations in 

concentrations of pharmaceuticals and personal care products 

in drinking water and reclaimed wastewater in Southern 

California. Environmental Science & Technology 40, 687–695. 

Luetzhoft, H.C.H., Halling-Sorensen, B., Jorgensen, S.E., 1999. 

Algal toxicity of antibacterial agents applied in Danish fish

668 

farming. Archives of Environmental Contamination and 

Toxicology 36 (1), 1–6. 

McAvoy, D.C., Schatowitz, B., Jacob, M., Hauk, A., Eckhoff, W.S., 

2002. Measurement of triclosan in wastewater treatment 

systems. Environmental Toxicology and Chemistry 21, 

1323–1329. 

Migliore, L., Cozzolino, S., Fiori, M., 2003. Phytotoxicity to and 

uptake of enrofloxacin in crop plants. Chemosphere 52, 

1233–1244. 

Moldovan, Z., 2006. Occurrences of pharmaceutical and personal 

care products as micropollutants in rivers from Romania. 


National Research Council, 2002. Biosolids Applied to Land: 

Advancing Standards and Practices Washington, D.C. 

North East Biosolids and Residuals Association, 2007. A National 

Biosolids Regulation, Quality, End Use & Disposal Survey 

Tamworth, NH. 


pharmaceuticals in wastewater effluent and surface waters of 

the lower Tyne catchment. Science of the Total Environment 

356, 143–153. 

Robinson, A.A., Belden, J.B., Lydy, M.J., 2005. Toxicity of 

fluoroquinolone antibiotics to aquatic organisms. 


Sapkota, A., Heidler, J., Halden, R.U., 2007. Detection of 

triclocarban and two co-contaminating chlorocarbanilides in 

US aquatic environments using isotope dilution liquid 

chromatography tandem mass spectrometry. Environmental 

Research 103, 21–29. 

Shoemaker, N.B., Vlamakis, H., Hayes, K., Salyers, A.A., 2001. 

Evidence for extensive resistance gene transfer among 

Bacteroides spp. and among Bacteroides and other genera in 

the human colon. Applied and Environmental Microbiology 

67, 561–568. 

Stackelberg, P.E., Furlong, E.T., Meyer, M.T., Zaugg, S.D., 

Henderson, A.K., Reissman, D.B., 2004. Persistence of 

pharmaceutical compounds and other organic wastewater 

contaminants in a conventional drinking-watertreatment 

plant. Science of the Total Environment 329, 99–113. 

Tamtam, F., Mercier, F., Le Bot, B., Eurin, J., Dinh, Q.T., 

Clement, M., Chevreuil, M., 2008. Occurrence and fate of 

antibiotics in the Seine river in various hydrological 

conditions. Science of the Total Environment 393, 84–95. 


treatment plants and rivers. Water Research 32, 3245–3260. 


Joss, A., 2004a. A rapid method to measure the solid-water 


distribution coefficient (K-d) for pharmaceuticals and musk 

fragrances in sewage sludge. Water Research 38, 4075–4084. 

Ternes, T.A., Joss, A., Siegrist, H., 2004b. Scrutinizing 

pharmaceuticals and personal care products in wastewater 

treatment. Environmental Science & Technology 38, 

392A–399A. 

USEPA, 1994. Biosolids Recycling: Beneficial Technologies for 

a Better Environment. Office of Water. U.S. Environmental 

Protection Agency, Washington, D.C. EPA 832-R-94–009. 

USEPA, 2001. Sampling Procedures for the 2001 National Sewage 

Sludge Survey. Office of Science and Technology, Washington, 

D.C. 

USEPA, 2003. Workshop on Emerging Pollutants. Region/ORD 

Workshop on Emerging Pollutants. U.S. EPA, Chicago, IL. 

USEPA, 2007a. 2001 National Sewage Sludge Survey Report 

Washington, D.C. EPA-822-R-07-006. 

USEPA, 2007b. Method 1694: Pharmaceuticals and Personal Care 

Products in Water, Soil, Sediment, and Biosolids by HPLC/MS/ 

MS Washington, D.C. EPA-821-R-08–002. 

USEPA, 2009. Targeted National Sewage Sludge Survey 

Washington, D.C. EPA-822-R-08–014. 

Williams, R.R., Bell, T.A., Lightner, D.V., 1992. Shrimp 

antimicrobial testing: II. Toxicity testing and safety 

determination for twelve antimicrobials with penaeid shrimp 

larvae. Journal of Aquatic Animal Health 4 (4), 262–270. 

Wilson, B.A., Smith, V.H., Denoyelles, F., Larive, C.K., 2003. Effects 

of three pharmaceutical and personal care products on 

natural freshwater algal assemblages. Environmental Science 

& Technology 37 (9), 1713–1719. 

Wu, C.X., Witter, J.D., Spongberg, A.L., Czajkowski, K.P., 2009. 

Occurrence of selected pharmaceuticals in an agricultural 

landscape, western Lake Erie basin. Water Research 43, 

3407–3416. 

Yang, L.H., Ying, G.G., Su, H.C., Stauber, J.L., Adams, M.S., 

Binet, M.T., 2008. Growth-inhibiting effects of 12 antibacterial 

agents and their mixtures on the freshwater microalga 

Pseudokirchneriella subcapitata. Environmental Toxicology and 

Chemistry 27 (5), 1201–1208. 

Ying, G.G., Yu, X.Y., Kookana, R.S., 2007. Biological 

degradationoftriclocarbanandtriclosaninasoilunder 

aerobic and anaerobic conditions and comparison with 

environmental fate modelling. Environmental Pollution 

150, 300–305. 

Zurita, J.L., Repetto, G., Jos, A., Salguero, M., Lopez-Artiguez, M., 

Camean, A.M., 2007. Toxicological effects of the lipid regulator 

gemfibrozil in four aquatic systems. Aquatic Toxicology 81 (1), 

106–115.

WATER RESEARCH

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?