Presentation - National Water Research Institute

Characterisation of Organic Matter in Alum Treated
Drinking Water using Reverse Phase‐High
Performance Liquid Chromatography and
Ion‐Exchange Resins
Dongsheng Wang 1,2 , Linan Xing 1 , Mohamad Fared Murshed 2 , Theodore W.C. Lo 3 ,
Rolando Fabris 4 , Christopher W.K. Chow 1,2,4 , John van Leeuwen 1,2 , Mary Drikas 1,2,4
1 State Key Lab of Environmental Aquatic Chemistry, RCEES, CAS, Beijing, China.
2 SA Water Centre for Water Management and Reuse, University of South Australia, Mawson Lakes,
South Australia, Australia.
3. Laboratory Lb Animal lS Service Centre, The Chinese University i of Hong Kong, Prince Pi of Wl Wales Hospital,
Shatin, Hong Kong.
4
Australian Water Quality Centre, SA Water, Adelaide, South Australia, Australia.

Outlines
o Resin fractionation‐hydrophobicity & hydrophilicity
o Characterization of NOM using RP‐HPLC
o Key component of our RP‐HPLC approach
o Comparison of resin fractionation method and
RPHPLC for NOM
o Impact of coagulation on chromatographic behaviour

Resin fractionation
~ Resin fractionation has been introduced to
determine hydrophobicity of NOM in order to
optimize coagulation performance
~ Resin fractionation is based on adsorption onto
different absorbent resins by calculated the organic
concentration before and after contact with resins.
• An early case study in Australia has demonstrated dthe link
between source water hydrophobic fraction and
coagulant dose
• By characterising organics based on their hydrophobicity
and hydrophilicity, the treatability of NOM can be
determined

Earlier work on resin fractionation
Chow, C.W.K., Fabris R. and Drikas, M. (2004) A Rapid
Fractionation Technique To Characterise Natural Organic
Matter For The Optimisation Of Water Treatment Processes. J
Water SRT – Aqua 53(2) 85‐92.
Chow, C.W.K., Fabris R., Drikas, M. and Holmes, M. (2005) A
Case Study of Treatment Performance and Organic Character.
JWaterSRT– Aqua 54(6) 385‐395.
However, resin fractionation was recognized as a time
consuming technique and new alternative technique
using HPLC has been introduced

Characterization of NOM using RP HPLC
→ Since 80’s, researchers use RP HPLC to characterize HA and FA.
→ In recent years, people are working for finding standard materials to compare
with the polarity of NOM. Moreover, they combine with other advanced
techniques, such as:
i. pyrolysis‐gas chromatography‐mass spectrometry (Py‐GC‐MS)
ii. solid state 13 C‐NMR spectroscopy
iii. diffuse reflectance infrared Fourier transform (DRIFT).
→ However, most of them considered to consume considerable time and effort,
and require relatively expensive and complex instrumentation.
→ Therefore, this RP HPLC has been introduced which is:
o The detections all based on concentration NOM samples
o Recovery tends out between 70% to 200‐300%
o there are size exclusion effect during the detection

Mechanism of RPHPLC
‣ The principle of operation is that organic material in an
aqueous solution is introduced to a hydrophobic bonded HPLC
column (eg. C18) which adsorbs the hydrophobic materials
while the hydrophilic materials remain in the carrier and exit
immediately to the detector.
‣ After a set time period, the phase of the eluent is changed to a
predominantly polar organic solvent (ie. reversing the phase) and
the adsorbed hydrophobic material is desorbed from the column
and flows to the detector.

Key component of our RP HPLC approach
a. Optimized Reverse Phase HPLC Liquid System Program
b. Deterioration and reproducibility
c. Background Correction
b)
0.08
3 Repeat tInjections
UV absorbance @ 254nm (cm -1 )
0.07
0.06
0.05
0.04
0.03
0.02
0.01
pure water
Morgan raw
Middle River raw
MR
MG
MilliQ
0.00
0 5 10 15 20
Retention time (min)

Key component of our RP HPLC approach
d. Quantifying chromatography results using peak fitting
• Peak fitting software (Version 4, Systat Software Inc.) was used to resolve peak
components based on statistically ideal fits using various peak shapes
• Total peak area removal for the RPHPLC profile was calculated by using the
difference between the peak area of the resolved peaks (after peak fitting) of
raw and treated water from the RPHPLC scans
Peak 1 to 4 with low retention time were defined as the
RPHPLC‐hydrophilic proportion
Peak 5 to 7, detected at longer retention time, were
defined as the RPHPLC‐hydrophobic proportion

Natural water sample selection‐High SUVA water
Morgan – River Murray (Water
Source for South Australia)
Middle River Reservoir –
Kangaroo Island off the coast
of South Australia
Colour: 58 HU
DOC: 8.8 88mg/L
UV 254 : 0.325 cm ‐1
SUVA: 3.7 m ‐1 mg ‐1 L
Colour: 145 HU
DOC: 12.4 mg/L
UV ‐1
254 : 0.660 cm
SUVA: 5.0 m ‐1 mg ‐1 L

Comparison of resin fractionation method and
RPHPLC for NOM
RPHPLC profile of isolated resin sample…..
Resin fractionation
Percentage of peak area result using Peak fitting software
the resin‐hydrophobic fraction
not only contains RPHPLChydrophobic
components but
also RPHPLC‐hydrophilic
components

Impact of coagulation on chromatographic behaviour
A treatment simulation test, the ‘Jar test’ was conducted
on Morgan and Middle river raw water
Jar testing was performed at pH 5.5, 6, 7 and 8.5 at dose
from low to high (60 – 200 mg/L)
According to the profiles, it
can be seen that with the
increase in alum dose, both
the hydrophobic and
hydrophilic proportions
decreased gradually

Impact of coagulation on chromatographic behaviour
Peak 1 easily can be removed by
coagulation
Hydrophilic peak (2‐4) can be
removed better than
hydrophobic peak (5‐7)

Comparison of organic character of the recalcitrant organic at
maximum DOC removal coagulation
maximum DOC removal coagulation
Molecular weight
distribution
‐ Max alum coagulation dose at 200 mg/L
‐ the removal efficiency of the hydrophilic proportion (Peak 2‐4) is lower than that of
the hydrophobic proportion (Peak 5‐7)
‐ RPHPLC‐hydrophilic contains components of the resin‐hydrophobic fraction which is
expected to be removed by coagulation.

Comparison with resin fractionation –Treated waters
UV abs sorbance @ 254 nm
(cm -1 )
0.018
0.015
Middle River treated
Hydrophpbic of Middle River treated
Hydrophilic of Middle River treated
The peak areas 0.012 of each peak for the treated Middle River water were higher than that of
0.009
treated Morgan 0.009 water.
0.015
0.012
Morgan treated
Hydrophobic of Morgan treated
Hydrophilic of Morgan treated
0.006
However, the peak area percentages of resin hydrophobic and resin hydrophilic for the two
0.003
0.003
treated t waters were similar
il
0.000
10 100 1000 10000 100000
Apparent molecular weight (Da)
0.006
0.000
10 100 1000 10000 100000
Apparent molecular weight (Da)

Comparison with resin fractionation –Treated waters
Precentage Pea ak Area (%)
100
90
80
70
60
50
40
30
20
10
0
18 15 17 19
16
28
24
26
33 28 37
7
26
5
10
17
21 12
16 22 20
14
11 10
9
10 10
8
5
5
25 26
12 13 11 13
MG treated MR treated MG treated MR treated MG treated MR treated
resin resin resin resin
hydrophobic hydrophobic hydrophilic hydrophilic
Peak 7
Peak 6
Peak 5
Peak 4
Peak 3
Peak 2
Peak 1
‣ the DOC of the resin
hydrophobic proportion
isolated decreased from 7.2 72
mg L ‐1 to 3.6 mg L ‐1 for
Morgan water, while Middle
River water decreased from
11.2 mg L ‐1 to 5.6 mg L ‐1 .
Their DOC removal
efficiencies were both 50%.
‣ The DOC resin hydrophilic value decreased from 1.6 mg L ‐1 to 0.8 mg L ‐1 for Morgan
water, while Middle River water decreased from 1.3 mg L ‐1 to 1.1 mg L ‐
‣ RPHPLC profiles of resin hydrophilic proportions of were compared and showed that,
for the same water, the profiles before and after coagulation were very similar. This
confirmed that resin hydrophilic compounds were recalcitrant to alum coagulation
removal

Further statistical analysis using
multiple regression analysis
• % UV 254 removal = 0.012012 × % removal of Peak 1
+ 0.97 × % removal of Peak 2‐4
+ 0.026 × % removal of Peak 5‐7
• % DOC removal = 0.0072 × % removal of Peak 1
+ 0.79 × % removal of Peak 2‐4
+ 021× 0.21 × % removal of Peak 5‐7
• The R 2 values of both equations were 0.99.
• This confirmed that it is possible to establish a relationship between
RPHPLC peaks (resolved by the proposed peak fitting model) with
treatability.

Conclusion
The RPHPLC method developed and reported in this s paper was found to be reliable and
convenient for characterizing detecting NOM raw and coagulated waters. It is important to
be able to identify and evaluate the removable components of NOM during water
treatment.
This study confirmed that the hydrophilic portion was recalcitrant to conventional
treatment . By combining with HPSEC method, this study also confirmed that the hydrophilic
proportion was considered as low molecular weight compounds.
It was found that the hydrophilic proportions isolated by the XAD procedure are all of small
molecular weight and were recalcitrant to alum coagulation.
The RPHPLC method should be a useful tool for the optimization of water treatment
procedure by allowing assessment of the efficiencies of combinations of pre‐ozonation,
enhanced coagulation and advanced oxidation processes.

Acknowledgement
NSF of China under 51025830 and 50921064, National Basic
Research Program of China under 2011CB9337005108
South Australian Premier’s Science and Research Fund Project
“Development of materials engineering solutions for treatment of
Murray‐Darling Basin sourced water supplies”
Special fund from the State Key Laboratory of Environmental
Aquatic Chemistry, Project 08K08ESPCR
SA Water and Australian Water Quality Centre, University of South
Australia, Australia

Thank you for your attention….