DISSERTATION
resolver
resolver
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
Jambrec Daliborka – 2016<br />
<strong>DISSERTATION</strong><br />
Understanding Potential-Assisted Surface<br />
Modification – From Self-Assembled<br />
Monolayers to DNA Chips<br />
Dissertation<br />
Submitted for the degree of<br />
Doctor of Natural Sciences (Dr. Rer. Nat.)<br />
Faculty of Chemistry and Biochemistry<br />
Ruhr University Bochum<br />
Daliborka Jambrec<br />
Bochum, October 2016
This work was carried out in the period from April 2013 to October 2016 in the Department of<br />
Analytical Chemistry under the supervision of Prof. Dr. Wolfgang Schuhmann, Ruhr<br />
University Bochum, Germany.<br />
First Examiner: Prof. Dr. Wolfgang Schuhmann<br />
Second Examiner: Prof. Dr. Axel Rosenhahn<br />
Chair of the Examination Board: Prof. Dr. Christof Hättig<br />
Date of the PhD Defense: 2 nd December 2016
Acknowledgements<br />
For me, the PhD studies were a period of learning – learning about science, about<br />
different cultures, languages and cuisines, about responsibilities, about growing up and making<br />
important decisions.<br />
During this period, many people passed through my life and helped me on my journey,<br />
professionally and privately. The best example for this is my mentor and friend, Prof. Dr.<br />
Wolfgang Schuhmann, who I owe my deepest gratitude for constantly helping me become a<br />
better researcher and a stronger person, encouraging and supporting me in every step of my<br />
path.<br />
Furthermore, I am grateful to Prof. Dr. Axel Rosenhahn, for being interested in my research<br />
and accepting to be my second examiner.<br />
I would like to thank Anna Lauks, who helped me make the first steps in the PhD, during the<br />
initial month of my stay in the group. It was a challenging period, starting from the fact that<br />
she did not speak English nor did I speak German, thus I am grateful for her patience and<br />
willingness to transfer all her experience to me. Also, I would like to thank Dr. Kirill Sliozberg<br />
who helped me throughout the thesis with technical aspects alongside Dr. Thomas Erichsen.<br />
I would like to express special thanks to Bettina Stetzka, for her big help with the<br />
administration, which was not always only work related, and for having patience with my<br />
German. Her constant smile, calmness and physical fitness are very inspirational for me.<br />
Furthermore, I would like to thank all the people who cooperated with me during this period. I<br />
am especially thankful to Dr. Magdalena Gebala, Prof. Dr. Fabio La Mantia and Dr. Arturo<br />
Estrada-Vargas for their help in understanding the fundamentals of electrochemistry and DNA<br />
sensing. I owe my thanks to Bianca Ciui for the work we started on miRNA detection, Dr.<br />
Adrian Ruff for his help in the work with the glucose-oxidase-acridine orange intercalator, Dr.<br />
Ĺubomír Švorc for helping me expand my knowledge about DNA to boron-doped-diamond<br />
electrodes, Dr. Sanaz Pilehvar and Prof. Dr. Karolien De Wael for cooperation in the research<br />
with DNA aptasensors and Vera Eβmann for the help in discovering the field of bipolar<br />
chemistry.
I would like to thank all colleagues from my group, from older members like Lutz (whose<br />
cheerful spirit and singing in our office always made me very happy), Piyanut and Fangyuan,<br />
to the new ones like Danea and Nergis, for the pleasant and always dynamic intercultural<br />
atmosphere, and for making long days in the lab not difficult. During this period I developed<br />
an especial fondness for a few people from the group. I would like to thank Jan and his wife<br />
Kathi for the great moments spent together, usually by cooking good food and listening to good<br />
records. My biggest appreciation goes to the members of my “fantastic 4” – Grecia, Felipe and<br />
Uğur, with who I shared the most beautiful and the most difficult moments of this period of<br />
my life. We became brothers and sisters.<br />
In the end, there are not enough words to show gratitude to my family, old and new, for their<br />
unconditional love and support during the ups and downs throughout my stay in Germany.<br />
Even though it was difficult to be separated from my grandma, my mom and dad, and especially<br />
my sister, they encouraged me and cheered for me in every step of my chosen way. Hvala vam<br />
na svemu, volim vas! My warmest thanks go to my husband – my best teammate, for giving<br />
me peace, stability and the biggest support in pursuing my goals. VTNNSC!
“A structure this pretty just had to exist”<br />
James Watson, “The Double Helix”, 1986<br />
“Science and everyday life cannot and should not be separated…<br />
it is based on fact, experience and experiment.”<br />
Rosalind Franklin, 1940
Table of Contents<br />
1. Introduction …..………………………………………………………………… 1<br />
1.1 Thiol self-assembling …..………………………………………………………… 3<br />
1.2 DNA ……………………………………………………………………………...…… 6<br />
1.2.1 DNA in solution …….……………………………………………………………...… 9<br />
1.2.2 DNA in front of a polarized electrode ……………………………………………..... 13<br />
1.3 DNA immobilization …………………………………………………………….. 16<br />
1.3.1 DNA immobilization approaches …...……………………………………………..... 16<br />
1.3.2 Techniques for DNA microarray fabrication .……………………………………..... 18<br />
1.4 Hybridization detection ..……………………………………………………….. 21<br />
2. Aims of the Work ….………………………………………..…………...… 26<br />
3. Results and Discussion ………………………………..……………...… 29<br />
3.1 Importance of preparing the surface. Criteria for cleanliness .…...….. 30<br />
3.2 Importance of knowing the surface ……………………………………...….. 34<br />
3.2.1 Electrochemical impedance spectroscopy. DNA assay build-up …...…………...….. 35<br />
3.2.2 Potential of zero charge of bare and DNA-modified electrodes ………………...….. 41<br />
3.3 Importance of controlling the surface ...………………………………...….. 50<br />
3.3.1 Fast and controlled formation of DNA surfaces. Optimization of ssDNA<br />
immobilization procedure …..…………………………………………………...….. 51<br />
3.3.2 Formation of compact thiol SAMs within minutes ……………………………...….. 67<br />
3.3.3 Reproducible recycling of Au modified surfaces within seconds …..…………...….. 81<br />
3.4 Potential-assisted preparation of DNA sensors …..…………………...….. 85<br />
3.4.1 Optimization of the potential-pulse assisted immobilization method …………...….. 85<br />
3.4.2 DNA microchip fabrication ……………………………………………………...….. 91<br />
3.5 Intercalation as a DNA detection technique …………………………...….. 96<br />
4. Conclusions ………………………………………………………………...… 104<br />
5. Experimental Work ...………………………………………………....... 108
5.1 Materials and consumables …..……………………………...…….......… 109<br />
5.2 Electrochemical setup and instrumentation ……………...……….…. 110<br />
5.3 Preparation of gold surfaces ….………………………………………. 111<br />
5.4 Determination of the potential of zero charge …...………………….. 113<br />
5.5 Potential-assisted formation of self-assembled monolayers ….…….. 114<br />
5.6 Potential-assisted desorption …………………………………………. 115<br />
5.7 Preparation of DNA sensors ….………………………………………. 116<br />
5.7.1 ssDNA immobilization via incubation ...…………………………………………… 116<br />
5.7.2 Potential-assisted ssDNA immobilization ….……………………………………… 116<br />
5.7.3 Passivation by means of incubation ...……………………………………………… 117<br />
5.7.4 Potential-assisted passivation .……………………………………………………… 117<br />
5.8 Preparation of DNA chips ..…………………………………………………… 117<br />
5.9 Characterization of DNA sensors …..………………………………... 118<br />
5.9.1 Electrochemical impedance spectroscopy ……………………………………… 118<br />
5.9.2 Cyclic voltammetry ……………………………………………………………… 118<br />
5.9.3 Chronocoulometry for determination of DNA coverage …………………………… 118<br />
5.10 Hybridization and dehybridization ..……………………………………… 118<br />
5.10.1 Detection of hybridization ………………………………………………………… 119<br />
5.10.2 DNA coverage determination by means of FSCV ...…………………....………… 119<br />
5.11 Intercalation .…………………………………………………………………… 120<br />
5.12 Methods …..……………………………………………………………………… 120<br />
5.12.1 Electrochemical impedance spectroscopy ………………………………………… 120<br />
5.12.2 Chronocoulometry for the determination of DNA coverage ...…………………… 124<br />
6. References ...………………………………………………………………….... 127<br />
7. Appendix ..……………………………………………………………………… 133<br />
7.1 List of symbols and abbreviations …...……………………………………… 133<br />
7.2 Publications list …..……………………………………………………………… 136<br />
7.3 Conference contributions ...…………………………………………………… 138
1. Introduction
______________________________________________________________________ Introduction<br />
Surface modification is the process of altering of the material surface by implementing<br />
new and desired characteristics to it. Molecular self-assembly, a process of high-level intermolecular<br />
orientation and organization without an outer force, is a powerful tool for surface<br />
modification. Self-assembled monolayers (SAMs) have a wide range of applications 1 , but<br />
whether they are used for regulation of the surface wettability, as a protection layer in corrosion<br />
inhibition, or surface functionalization for binding proteins, DNA or cells, the main requirement<br />
in all their applications is the control of the modification process. For example, depending on<br />
the envisaged application, the desired coverage of the layer can significantly differ.<br />
Furthermore, the choice of numerous parameters determines both the rate and the duration of<br />
the modification process resulting in the desired coverage.<br />
One of the biggest applications of SAMs is in the development of biosensors. DNA<br />
hybridization detection and DNA sensors are becoming tremendously important in diagnostics<br />
as a result of the continuous advancement of the Human Genome Project 2,3 . A DNA sensor<br />
usually consists of a single stranded DNA grafted on an electrode surface, essential for the<br />
recognition of the complementary target DNA present in a sample under investigation. The<br />
recognition process occurs by hybridization between these two DNA strands, which is<br />
converted into a signal by the transducer part of the DNA biosensor. Transduction of the signal<br />
can be done by different techniques – optically, piezoelectrically and electrochemically. In<br />
recent years the development of electrochemical DNA biosensors is in the spotlight, due to their<br />
operating simplicity, possibility of miniaturization, and thus portability, and low cost, which<br />
makes them very attractive for mass production. Continuous investigation of fundamental<br />
issues, such as surface characterization, tailoring of interfaces and DNA recognition, along with<br />
the progress in the field of microfabrication will lead to powerful, yet easy-to-use DNA<br />
diagnostic products 4 .<br />
The sensitivity and selectivity of DNA biosensors is highly dependent on the quality of the<br />
prepared DNA sensing surface. In order to tailor the desired DNA sensing surfaces for<br />
envisaged sensing platforms it is of utmost importance to understand processes occurring at the<br />
electrode during the surface modification 5 . Only in this way properties of the surface can be<br />
controlled in a reproducible manner on a desired time scale.<br />
The following sections will present the current state of the art in the field of self-assembly and<br />
DNA sensing, including some historical aspects related to the “molecule of life”.<br />
2
______________________________________________________________________ Introduction<br />
1.1 Thiol self-assembling<br />
The formation of self-assembled monolayers (SAMs) was first reported by Zisman et al. in the<br />
late 1940s 6,7 . Later, in the 1980s, Nuzzo and Allara made initial studies on alkylthiols selfassembly<br />
on gold 8 . Since then, a tremendous number of studies were devoted to understanding<br />
and improving the formation of alkylthiol SAMs for numerous applications 1 – tailoring surface<br />
wetting behavior (hydrophilicity/hydrophobicity), protective coatings, corrosion inhibition and<br />
surface functionalization.<br />
Self-assembling occurs from either a vapor or liquid phase 9 , where head groups assemble close<br />
to the surface, while tail groups orientate away from the surface (Figure 1.1). Self-assembly of<br />
thiols on gold surfaces is characterized by its simplicity of preparation and a high number of<br />
available functional groups, even though it has a limited stability due to the relatively small safe<br />
potential region of the Au-S bond.<br />
Alkylthiols consist of an alkyl chain with a thiol head group at one end, which has a high affinity<br />
towards gold, and a functional group of choice at the other end. Adsorbed thiols lose their<br />
hydrogen atom from the head group and form a strong thiolate-gold bond at the interface. The<br />
formed chains are in a trans-conformation obtaining an angle of 60-70° with respect to the<br />
surface. The sulfur atom is in sp3 hybridization and therefore bound to three Au atoms. Au-S<br />
bond formation occurs through the reduction of alkylthiols to alkylthiolates 10 and the following<br />
equation presents one of the proposed mechanisms for this reaction:<br />
R − SH + Au(0) n → R − S − Au(I) • Au(0) n + 1 2 H 2<br />
(1.1)<br />
Disulfides are believed to chemisorb on gold via S-S bond cleavage:<br />
RS − SH + 2Au → 2RS − Au (1.2)<br />
The thickness of the monolayer is controlled by the length of the alkyl chains. During selfassembly,<br />
alkyl chains align themselves by van der Waals forces. Therefore, longer alkylthiols<br />
result in more stable and densely packed monolayers 7 . Furthermore, longer alkylthiols adsorb<br />
preferentially over shorter ones 11 .<br />
1.1 Thiol self-assembling 3
______________________________________________________________________ Introduction<br />
Figure 1.1. Self-assembly on an electrode surface.<br />
In the beginning of the self-assembly process, while the density of molecules at the surface is<br />
small, molecules either form a disordered mass or an ordered lying phase at the surface 1 (Figure<br />
1.2, a). When the density increases, molecules start forming ordered three-dimensional<br />
structures. Even though the Au-S bond is reasonably strong, the adsorbed thiols can still move<br />
around on the electrode to heal gaps of exposed gold 12 . This so-called healing process occurs<br />
in a second phase, where thiols slowly reorganize and order on the surface 13-15 (Figure 1.2, b).<br />
Various parameters affect alkylthiol SAM formation such as the architecture and cleanliness of<br />
the gold surface, temperature, thiol concentration, functional groups and solvent composition.<br />
Commonly, organic solvents such as ethanol or DMSO are used for adsorption of alkylthiols<br />
on gold.<br />
Preparation of SAMs is standardly performed by immersing a gold substrate into an alkylthiol<br />
solution. Thus, the adsorption occurs at open circuit potential (OCP) and the kinetics of the<br />
adsorption roughly follows the Langmuir adsorption curve 1,9 . According to the Langmuir<br />
kinetic model, the rate of adsorption is proportional to the free space on the surface:<br />
dθ<br />
dt = k(1 − θ) (1.3)<br />
where θ is the partial coverage and k is the rate constant. Nevertheless, this model makes strong<br />
assumptions such as the homogeneity of the surface (perfectly flat surface) and the absence of<br />
interactions between adjacent molecules. In reality, surface roughness of polycrystalline gold<br />
and thiol island formation affect the SAM formation kinetics.<br />
1.1 Thiol self-assembling 4
______________________________________________________________________ Introduction<br />
Externally applied potentials affect already formed SAMs, causing the change of their<br />
structure 16 , change of wettability 17 or desorption 18,19 . The desorption of SAMs can occur by<br />
applying rather high positive or negative potentials. Nevertheless, the potential value required<br />
to invoke SAM desorption depends on the type of alkylthiols used (length of the alkyl chain,<br />
head group repulsion) and the coverage, since they influence the stability of the SAM 12 . Longer<br />
alkylthiols and more compact layers are more stable SAMs and therefore, more difficult to<br />
desorb. Furthermore, the possibility of controlling SAM formation by applying an external<br />
potential was observed in the 1990s 20 . Application of anodic constant potentials seems to<br />
accelerate the kinetics of SAM formation 13,21-24 . However, this phenomenon is still poorly<br />
understood.<br />
Figure 1.2. Self-assembly on an electrode surface occurs through two phases: a) in the first<br />
phase (lying phase) molecules randomly lie on the surface, and in b) the second phase<br />
(healing phase) they slowly reorganize to form highly compact monolayers.<br />
1.1 Thiol self-assembling 5
______________________________________________________________________ Introduction<br />
1.2 DNA<br />
Deoxyribonucleic acid (DNA) is the carrier of genetic instructions used for the proper<br />
development and functioning of all living organisms. It was isolated for the first time by<br />
Friedrich Miescher in 1869. Since then, the “molecule of life” has evoked curiosity from many<br />
researchers around the world. The defining moment in nucleic acid research was a paper<br />
published in The Journal of Experimental Medicine in 1944 by Oswald Avery, Colin MacLeod<br />
and Maclyn McCarty, where they show for the first time that DNA, and not proteins, is the<br />
material of inheritance 25 .<br />
The controversy about the discovery of the DNA structure still remains, as it is debatable who<br />
should get the credit for it. In 1953 five papers were published in the journal Nature, describing<br />
and providing evidence for the double helix structure of the DNA. James Watson and Francis<br />
Crick first suggested the correct double helix model 26,27 , based on an X-ray diffraction image<br />
taken by Rosalind Franklin and her student Raymond Gosling. Previously, J. Watson listened<br />
R. Franklin’s lecture about the DNA structure in 1951, and got access to her progress report<br />
without her knowledge. In the same issue of the journal, Maurice Wilkins with his colleagues 28<br />
and R. Franklin with R. Gosling 29,30 provided experimental evidence supporting the Watson<br />
and Crick model. In 1962, after R. Franklin’s death, Watson, Crick and Wilkins received<br />
together the Nobel Prize in Physiology or Medicine. Nobel Prizes are awarded only to living<br />
recipients and it remains unknown whether Rosalind Franklin should have received the prize<br />
as well.<br />
DNA consists of two strands, where each strand is composed of subunits called nucleotides. A<br />
nucleotide consists of three components – a nucleobase, a sugar (deoxyribose) and a phosphate<br />
group (Figure 1.3). Within the same strand, nucleotides are covalently bound to each other in a<br />
chain by connecting the sugar of one nucleotide to the phosphate group of the next one via a<br />
phosphodiester bond, creating a sugar-phosphate backbone. Each phosphate group is bound to<br />
the 3' carbon of the previous deoxyribose, and to the 5' carbon of the following sugar ring. The<br />
ends of a strand are then commonly designated as 3' end or 5' end if the chain finishes with an<br />
unbound 3' or 5' carbon, respectively. Moreover, the connection between nucleotides of two<br />
complementary DNA strands occurs through the nucleobases via formation of hydrogen bonds.<br />
The nucleobases are divided in two groups – pyrimidines (thymine and cytosine) and purines<br />
1.2 DNA 6
______________________________________________________________________ Introduction<br />
(adenine and guanine). Adenine binds only to thymine, forming two hydrogen bonds, and<br />
cytosine binds exclusively to guanine, forming three hydrogen bonds.<br />
Figure 1.3. Scheme of the primary DNA structure.<br />
Two complementary DNA strands (single-stranded DNA, ssDNA) coil around the same axis<br />
forming a double helix (double-stranded DNA, dsDNA). This way hydrophobic bases are inside<br />
the helix and the hydrophilic sugar-phosphate backbone is orientated outside (Figure 1.4). Both<br />
DNA strands store the same biological information, that is, the nucleobase sequence. However,<br />
they can bind only in an antiparallel disposition, which means that the strands are orientated in<br />
opposite directions, one in the direction 3' to 5' and the other from 5' to 3'. The 5' end has a<br />
terminal phosphate group, while the 3' end, a terminal hydroxyl group. DNA can form three<br />
conformations – A, B and Z. B-DNA is a right-handed conformation, where each full turn in<br />
the helix is 3.4 nm long and consists of ten bases (thus with 0.34 nm between adjacent bases).<br />
A-DNA is also right-handed and it occurs when the humidity is lower than 75 %. In this<br />
conformation the base pairs are not perpendicular to the helix axis, resulting in a wider and<br />
1.2 DNA 7
______________________________________________________________________ Introduction<br />
shorter structure than B-DNA. On the other hand, Z-DNA is left-handed and it is stable only at<br />
high salt concentrations in order to minimize electrostatic repulsion, since the phosphate groups<br />
in the backbone are closer to each other.<br />
The genetic information is stored within the double helix structure of the DNA molecule by the<br />
sequence of nucleobases. Besides encoding the formation of cell organelles and proteins, DNA<br />
sequences are responsible for all physical traits and disease susceptibility. Therefore, the<br />
knowledge of DNA sequences is priceless for biological research and fields like medical<br />
diagnosis, genetic screening, forensic biology, environmental control and biotechnology,<br />
among others.<br />
Figure 1.4. Scheme of the secondary DNA structure.<br />
1.2 DNA 8
______________________________________________________________________ Introduction<br />
1.2.1 DNA in solution<br />
DNA is a highly charged polymer with charge density equal to two elementary charges per base<br />
pair. Therefore, in solution DNA interacts strongly with surrounding ions that counterbalance<br />
its charge. These ions can be grouped into different zones 31 . In the region closest to the DNA<br />
there are so called site-bound or inner-sphere ions that share water molecules with the DNA. In<br />
the following zone are outer-sphere ions (territorial ions) that keep their inner hydration layer<br />
and are free to move along the DNA molecule but are kept close to it due to the electrostatic<br />
field. And in the last zone are free ions that form an ionic cloud around the DNA 3 .<br />
The two most known approaches employed to investigate the extent of DNA-ion interaction are<br />
the Manning-Oosawa (MO) counterion condensation theory 32 and the Poisson-Boltzmann (PB)<br />
equation. MO theory addresses the DNA charge compensation by focusing on the counterions<br />
that form a so-called condensed layer around the DNA (outer-sphere ions) in order to reduce<br />
the charge density below a certain critical value 33 . According to the theory, counterion<br />
accumulation at the DNA surface forming a condensed layer occurs under the condition that<br />
the charge density parameter η is > 1:<br />
η = zl B<br />
b<br />
(1.4)<br />
where z is the valence of counterions, lB is the Bjerrum length and b is the charge separation.<br />
The Bjerrum length represents the distance between charges at which their electrostatic<br />
interaction energy equals the thermal energy and is defined as:<br />
l B =<br />
e 2<br />
4πεε 0 kT<br />
(1.5)<br />
where e is the elementary charge, ε is the dielectric constant of the solvent, ε0 is the vacuum<br />
permeability and kT is thermal energy scale. This means that the counterion condensation (for<br />
monovalent ions) occurs when the charge separation b is smaller than the Bjerrum length, which<br />
is true for both ss- and dsDNA. Namely, lB = 0.71 nm in aqueous solutions, while the charge<br />
separation is 0.43 nm in ssDNA and 0.34 nm in dsDNA 31 . According to the theory, condensed<br />
counterions are still assumed to be mobile 34 .<br />
1.2 DNA 9
______________________________________________________________________ Introduction<br />
Accumulation of counterions occurs to a certain extent, until the charge density parameter<br />
decreases to 1, that is, until b increases to lB. The remaining effective charge of DNA is reduced<br />
by a factor r:<br />
r = 1 − 1 zη<br />
(1.6)<br />
Thus, it is predicted that the charge of DNA in a solution with monovalent counterions is<br />
reduced by 76 % 31 . An important condition of the MO theory is that condensation occurs only<br />
if the relation κ -1 >> a is satisfied (where κ -1 is the Debye length and a is the polymer radius),<br />
which means either for highly diluted solutions or an infinitely thin DNA strand. Furthermore,<br />
the theory is valid only for vanishingly small DNA concentrations.<br />
For finite DNA concentrations a more complicated model needs to be used that is based on<br />
solving of the PB equation. The PB equation describes the Gouy-Chapman (GC) model, where<br />
a charged solid comes into contact with an ionic solution creating a double layer. Due to the<br />
thermal motion of ions the counterion layer is a diffuse layer. The remaining charge around the<br />
DNA molecules (ions in the third zone) is described by a linearized form of the PB equation –<br />
the Debye-Hückel equation 35 that explains the relationship between the electrostatic behavior<br />
of DNA and the ionic strength. In the equation the screening of DNA is quantified by the Debye<br />
length κ -1 :<br />
κ 2 = 8πl B N A I (1.7)<br />
κ 2 = 2e2 N A I<br />
εε 0 kT<br />
(1.8)<br />
where NA is the Avogadro number (6.022 × 10 23 1/mol) and I is the ionic strength.<br />
The first quantitative experimental studies about conformational properties of DNA as a<br />
function of salt concentration were done by Harrington 36 , who measured the DNA radius of<br />
gyration in dilute DNA solutions. The DNA radius of gyration (Rg) depends on the DNA<br />
persistence length (lp):<br />
R g = √ Ll p<br />
3<br />
(1.9)<br />
1.2 DNA 10
______________________________________________________________________ Introduction<br />
where L is the contour length of DNA. The persistence length is a basic mechanical property of<br />
a polymer and it is a measure of its stiffness. It is defined as a characteristic length over which<br />
the chain maintains a certain direction 37 . When the contour length is smaller than the persistence<br />
length a polymer behaves as a rigid rod, while a flexible coil behavior is observed for contour<br />
lengths much higher than the persistence length. The persistence length of dsDNA is considered<br />
to be around 50 nm 38 , while the persistence length of ssDNA is only 1-2 nm 39,40 . Thus, dsDNA<br />
is expected to behave as a rigid rod and ssDNA as a flexible chain (Figure 1.5).<br />
dsDNA<br />
ssDNA<br />
L = N bp × b = 20 × 0.34 nm<br />
L = 6.8 nm<br />
l p ≈ 50 nm<br />
L ≪ l p (rigid rod)<br />
L = N b × b = 20 × 0.43 nm<br />
L = 8.6 nm<br />
l p ≈ 1 to 2 nm<br />
L > l p (flexible chain)<br />
Figure 1.5. Dependence of the DNA mechanical properties on the persistence length<br />
shown for an example of a 20-mer DNA strand. Nbp and Nb represent the number of base<br />
pairs and number of bases, respectively, while b represents the charge separation.<br />
Nevertheless, the persistence length consists of two contributions:<br />
l p = l 0 + l el (1.10)<br />
where l0 is an intrinsic stiffness due to chain properties and lel is the electrostatic repulsion<br />
within the chain, which depends on the ionic strength:<br />
l el =<br />
l B<br />
(2bκ) 2 (1.11)<br />
Through the Debye length, the ionic strength influences the flexibility of charged polymers,<br />
which should be considered especially while investigating the behavior of ssDNA in solution.<br />
In solutions of increased ionic strength, the Debye length decreases leading to a decrease in the<br />
1.2 DNA 11
______________________________________________________________________ Introduction<br />
electrostatic component of the persistence length. In this case, the polymer behaves more as a<br />
flexible coil (Figure 1.6). On the other hand, decreasing the ionic strength, that is, increasing<br />
the Debye length, leads to an increase in the rigidity of a polymer 41 . Therefore, depending on<br />
the ionic strength, ssDNA may also manifest high rigidity.<br />
Figure 1.6. Dependence of the polymer rigidity on the ionic strength. Figure adapted from<br />
ref. 41 .<br />
1.2 DNA 12
______________________________________________________________________ Introduction<br />
1.2.2 DNA in front of a polarized electrode<br />
In addition to the DNA-ion interaction, it is important to understand the influence of a charged<br />
electrode on the behavior of DNA on its surface for the application in DNA sensing at electrified<br />
surfaces. The GC model of the double layer describes how the ionic strength and the<br />
polarization of the electrode influence the double layer structure and the potential drop in front<br />
of the electrode. The GC equation:<br />
Φ = 2kT<br />
e<br />
ln 1 + γexp ( −d<br />
κ −1)<br />
1 − γexp ( −d<br />
κ −1) (1.12)<br />
γ = tanh ( eΦ 0<br />
4kT ) (1.13)<br />
where Φ0 and Φ are the potentials at the electrode surface and at a distance d from the surface,<br />
respectively, reveals that the potential distribution strongly depends on the ionic strength, where<br />
an increase of the ionic strength leads to a steeper drop of the potential (Figure 1.7, a). Thus, a<br />
few nm away from the surface, Brownian motion prevails over electric forces and dominates<br />
the system response 42 . Furthermore, the model predicts a sharp potential drop for highly<br />
charged electrodes (high Φ0), while the decline is more gradual for lower Φ0 values 43 (Figure<br />
1.7, b).<br />
The DNA conformation on the electrode surface can be manipulated by externally applied<br />
potentials 44,45 . However, this is true only under certain conditions 42 . Namely, as in solutions of<br />
high ionic strength the applied potential decays within a nm distance, the range is too short to<br />
significantly affect grafted DNA molecules. Therefore, both negative and positive potentials do<br />
not affect the conformation of neither ds- nor ssDNA. dsDNA exhibits a rigid conformation,<br />
while ssDNA coils on the electrode surface (Figure 1.8, a). This remarkable difference in<br />
conformation originates from the difference in the persistence length of dsDNA and ssDNA, as<br />
explained in Section 1.2.1. Furthermore, in presence of filler molecules with height comparable<br />
to the length of the DNA spacer, the dsDNA conformation is almost perpendicular with respect<br />
to the electrode surface 42,45 . The reason for the upright conformation is the steric repulsion<br />
between the lowest base pairs and a monolayer that backfills the gold electrode between DNA<br />
strands. In contrast, ssDNA remains lying on the surface since, besides the very weak electrical<br />
interactions, self-repulsion along the DNA strand is suppressed by the high ionic strength.<br />
1.2 DNA 13
______________________________________________________________________ Introduction<br />
Figure 1.7. a) Potential distribution in relation to distance from the electrode for different<br />
ionic strengths. Φ0 of 100 mV was used for the calculation. Figure adapted with<br />
permission from ref. 42 . Copyright (2010) American Chemical Society. b) Potential profile<br />
calculated for different Φ0 values and an ionic strength of 10 mM. Figure adapted from<br />
ref. 43 .<br />
In solutions of intermediate ionic strength, where the Debye length spans over few DNA base<br />
pairs the influence of the applied potential is more significant, scaling with the distance from<br />
the electrode surface (Figure 1.8, b). In this case, it is possible to manipulate the conformation<br />
of dsDNA by applying positive (invoking a lying conformation) or negative potentials<br />
(invoking up-right conformation). Still, control of dsDNA is achieved with less difficulty than<br />
of ssDNA, due to their persistence lengths, while ssDNA manipulation is only partial depending<br />
on the Debye length. While at a positively polarized electrode ssDNA will remain in the lying<br />
conformation, negative potentials evoke at least partially an up-right orientation. The upper part<br />
of the ssDNA is not exposed to the electric field and it exhibits a randomized conformation.<br />
In solutions of low ionic strength charge screening is weak resulting in high charge repulsion.<br />
Therefore, dsDNA is not stable under these conditions. Furthermore, applied electric potentials<br />
are very long ranged, which leads to an efficient repulsion of the ssDNA from the surface<br />
(Figure 1.8, c). Since the persistence length depends on the ionic strength, in solutions of low<br />
ionic strength ssDNA exhibits a rigid conformation. Thus, combining these two effects, ssDNA<br />
1.2 DNA 14
______________________________________________________________________ Introduction<br />
has an up-right position when negative potentials are applied and a lying conformation at<br />
positive potentials.<br />
Figure 1.8. Schematic representation of DNA conformation on negatively and positively<br />
charged surfaces for a) high, b) intermediate, and c) low ionic strengths. Figure adapted<br />
with permission from ref. 42 . Copyright (2010) American Chemical Society.<br />
1.2 DNA 15
______________________________________________________________________ Introduction<br />
1.3 DNA immobilization<br />
1.3.1 DNA immobilization approaches<br />
A DNA immobilization strategy is determined by the substrate material used for the attachment.<br />
Over the years, various surfaces were investigated for immobilization of DNA such as among<br />
others the hanging mercury drop electrode, carbonaceous materials, boron-doped diamond,<br />
silver, platinum and gold. Initially, research on DNA was conducted solely on mercury drop<br />
electrodes (in the beginning of the 1970s) and carbon electrodes (since the middle of the<br />
1970s 46 ). Later, gold electrodes became popular, as the chemisorption of thiol-tethered DNA<br />
showed to be a very promising method for the preparation of DNA sensors.<br />
With respect to the type of bond formed during immobilization, DNA immobilization methods<br />
are characterized by three main mechanisms 47 : physisorption, covalent immobilization and<br />
chemisorption.<br />
Physisorption is the simplest immobilization method, which is based on the adsorption of<br />
unmodified oligonucleotides on an electrode through electrostatic forces, van der Waals<br />
interactions, hydrogen bonds and hydrophobic interactions. The immobilization occurs either<br />
through nucleic bases (immobilization of ssDNA, Figure 1.9, a) or the phosphate backbone<br />
(dsDNA, Figure 1.9, b). It is characterized by a multiple site attachment, which allows for the<br />
investigation of direct DNA oxidation and reduction. The main drawback of this method is its<br />
sensitivity on environmental changes (pH value, temperature, ionic strength) due to the weak<br />
attachment. Furthermore, attachment of the DNA occurs via multiple points, which prevents<br />
further hybridization due to the restricted configurational freedom of physisorbed DNA 48 .<br />
Carbonaceous materials and the mercury drop electrode were mostly utilized for this<br />
immobilization technique.<br />
Covalent immobilization results in a much stronger binding between the surface and DNA<br />
(Figure 1.9, c). Another advantage of this method is the appropriate orientation of the probe<br />
DNA due to the end-point attachment of ssDNA, which facilitates hybridization. In order for<br />
the immobilization to occur, the immobilization surface (chemically or electrochemically) and<br />
the DNA itself need to be activated, which presents a drawback of this method. Activation of<br />
1.3 DNA immobilization 16
______________________________________________________________________ Introduction<br />
carbonaceous materials and glass substrates was extensively investigated for the covalent<br />
immobilization of DNA.<br />
Figure 1.9. Schematic representation of DNA immobilization by physisorption via a)<br />
bases, b) the phosphate backbone, c) covalent immobilization, d) chemisorption, and e)<br />
biotin-streptavidin immobilization.<br />
Chemisorption of thiol tethered DNA on noble metals (typically gold) occurs by the selfassembly<br />
process that is explained in Section 1.1. This method is widely used for the<br />
preparation of DNA sensors due to its simplicity and the formation of a relatively strong Au-S<br />
bond (Figure 1.9, d). Biotin-streptavidin immobilization was also employed for grafting of<br />
DNA. It occurs through two steps, namely biotinylation of either the surface or the DNA<br />
followed by binding of the streptavidin-modified counterpart (DNA or surface, respectively)<br />
and formation of a very strong biotin-streptavidin conjugate (Figure 1.9, e). Nevertheless, since<br />
the surface is usually biotinylated by chemisorption of biotin-terminated SAMs, the strength of<br />
the modification is determined by the Au-S bond strength.<br />
1.3 DNA immobilization 17
______________________________________________________________________ Introduction<br />
1.3.2 Techniques for DNA microarray fabrication<br />
A practical way of detecting and diagnosing various diseases is through the detection of nucleic<br />
acid sequences, specific for any living organism 49 . One of the possible ways to gain insight into<br />
a DNA sequence is by using DNA sensors. The sensitivity and selectivity of DNA biosensors<br />
is highly dependent on the quality of the prepared DNA sensing surface. Good immobilization<br />
technique ensures high reactivity, appropriate orientation, accessibility and the stability of the<br />
grafted probe DNA and prevent unspecific binding 47 .<br />
DNA immobilization methods can be divided into two groups 47,50 :<br />
- Base-by-base synthesis (light-directed synthesis), which represents a bottom-up synthesis<br />
of DNA sequences at the surface<br />
- Direct attachment of already synthesized DNA sequences to the surface<br />
Both strategies are used for DNA microarray fabrication. Several base-by-base synthesis<br />
strategies were developed, including light-directed synthesis using photolithographic masks by<br />
Affymetrix (GeneChips), photo-mediated synthesis by Roche (NimbleGen) that used digital<br />
masks instead, and inkjet base-by-base manufacturing (printing) of DNA probes on the surface<br />
developed by Agilent Technologies 51 . Among these, the Affymetrix strategy is the most known<br />
and it will be shortly described here. On the other hand, among strategies for DNA array<br />
fabrication by direct attachment of pre-synthesized DNA sequences, spotting (printing) is by<br />
far the most known approach. However, new technologies for DNA array production are arising<br />
on the market, including “electronic microarrays”.<br />
Light-directed synthesis is a complex method requiring specialized equipment, however<br />
attractive for DNA microarray fabrication. The principle of in situ DNA synthesis by<br />
Affymetrix consists of UV masking and light-directed chemical synthesis of DNA sequences<br />
directly at the array, one nucleotide at the time per spot for many spots simultaneously (Figure<br />
1.10). The surface is initially modified with a covalent linker containing a protection group that<br />
is easily removed upon irradiation. Using a photolithographic mask, desired spots on the surface<br />
are irradiated, removing locally the protecting groups. Subsequently, these spots are modified<br />
with the desired protected nucleotides. This experimental sequence is repeated as many times<br />
as necessary to obtain the desired DNA sequences on the surface and this is determined by the<br />
1.3 DNA immobilization 18
______________________________________________________________________ Introduction<br />
length of the DNA. On average, for sequences with 25 bases, around 100 masks are needed per<br />
chip. Nevertheless, the obtained chips can have an extremely high density (> 10 6 spots per<br />
array). Main drawbacks of this technique are the complex nature of chemical synthesis and a<br />
very expensive production 51 . Furthermore, difficult customization and the possibility of<br />
synthesizing only relatively short DNA sequences are limiting the application of the technique<br />
for production of point-of-care devices.<br />
Figure 1.10. Schematic representation of the photolithographic synthesis of DNA<br />
sequences on the electrode surface. a) surface modified with a covalent linker with a<br />
protective group, b) irradiation of desired spots on the surface, c) modification of<br />
deprotected spots with desired nucleotides, d-h) repetition of the process of irradiation<br />
and modification with nucleotides to fabricate spots with desired DNA sequences.<br />
The principle of spotting (printing) of already synthesized DNA sequences consists of spotting<br />
nano- to picoliter volume drops of DNA-containing solution onto a predefined grid by a<br />
1.3 DNA immobilization 19
______________________________________________________________________ Introduction<br />
specialized ink-jet printer ran by a robot (Figure 1.11). By this, spots with a diameter of 100-<br />
150 µm are created on the surface. The number of spots is limited to prevent crosscontamination.<br />
Therefore, the density of these microarrays is moderate with 10,000 to 30,000<br />
spots per array. Furthermore, this technique requires strict monitoring of the production<br />
reproducibility and quality control. Difficulties with efficiency and accuracy are another<br />
drawback of this technique 47,51 . Nevertheless, an advantage is that the content of the<br />
microarrays is flexible.<br />
Figure 1.11. Schematic representation of DNA solution spotting on the electrode surface.<br />
Electronic microarrays use an electric field to control the immobilization of DNA. The<br />
Company Nanogen developed a Nanochip with 12 connectors controlling 400 individual sites<br />
on a chip. The principle of immobilization consists of the transport of negatively charged DNA,<br />
modified with biotin, through an agarose permeation layer towards specific sites, modified with<br />
streptavidin, on the chip where a positive current is applied 52 . Even though the density of these<br />
chips is limited to 400 spots, this is sufficient for the majority of diagnostic applications 51 .<br />
Furthermore, the content of the microarray can be specified by the user, which decreases<br />
microarray manufacturing costs. Biotin-streptavidin chemistry offers a very strong bonding but<br />
comes with some limitations. The synthesis of streptavidin-modified surfaces consists of<br />
several steps, such as the activation of the surface, immobilization of streptavidin and blocking,<br />
which increases the production costs and time 47 . Furthermore, one of the drawbacks of using<br />
streptavidin is the problem with unspecific interactions.<br />
1.3 DNA immobilization 20
______________________________________________________________________ Introduction<br />
1.4 Hybridization detection<br />
Compared to the traditional methods for the detection of DNA sequences, such as electrophoresis<br />
or membrane blots, DNA sensors are faster, simpler and less expensive 53 , and<br />
therefore, present a very active research field. Presently, a wide range of DNA sensor<br />
technologies are in development or being commercialized.<br />
DNA sensors are based on the specific detection of a DNA sequence using a so-called probe<br />
DNA immobilized on the surface 49 (Figure 1.12). Therefore, two main parts of a DNA sensor<br />
are 48 :<br />
- biorecognition interface (immobilized specific DNA sequence) that allows the detection<br />
of a target element (complementary DNA sequence)<br />
- transducer that transforms a detected signal into a readable output<br />
Figure 1.12. Principle of a DNA sensor.<br />
The selective binding, followed by conformational and structural changes in the recognition<br />
layer, can be detected by various techniques. Depending on the transduction approach, DNA<br />
biosensors can be optical, piezoelectric and electrochemical.<br />
Optical DNA sensors based on fluorescence are very sensitive and DNA chips with this<br />
detection scheme have already been commercialized 51 . One of the optical readout strategies is<br />
the detection of an intercalating dye (e.g., ethidium bromide) upon hybridization. Furthermore,<br />
the use of molecular beacons was explored extensively (Figure 1.13). The ends of a molecular<br />
1.4 Hybridization detection 21
______________________________________________________________________ Introduction<br />
beacon are complementary and form a loop through self-hybridization. Coupling of a<br />
fluorophore and a quencher on opposite ends of a beacon results in quenching of fluorescence<br />
when the beacon is closed and a fluorescence signal when the beacon is opened due to<br />
hybridization with a target DNA. Another optical technique commonly used for DNA detection<br />
is surface plasmon resonance, based on the change in the refractive index of a thin metal<br />
substrate modified with a biorecognition interface upon hybridization. Nevertheless, the<br />
required equipment for optical detection is sophisticated and relatively expensive and therefore<br />
not suitable for clinical purposes and point-of-care diagnostics but rather for laboratory<br />
applications. Furthermore, inconsistent yields of target synthesis and labelling as well as nonuniform<br />
rates of photobleaching can result in insufficient readout accuracy required for patient<br />
diagnosis 54 .<br />
Figure 1.13. Optical DNA detection using molecular beacons.<br />
Piezoelectric DNA sensors rely on the mass change upon hybridization with a target DNA that<br />
is correlated with an increase in the fundamental resonance frequency of a crystal 55 . Quartz<br />
crystal microbalance is the most commonly used technique for the real-time monitoring of the<br />
hybridization process 56 . Different strategies to increase the mass change upon hybridization<br />
were used in order to increase the sensitivity, usually by using bulky labels 57,58 .<br />
Point-of-care diagnostics requires fast response, high sensitivity and selectivity, and operation<br />
simplicity. These, coupled with portability, low cost and possibility of miniaturization, are<br />
characteristics of electrochemical DNA sensors, which makes them very attractive for mass<br />
production. Unlike bulky optical readout systems, electrochemical detection can be integrated<br />
1.4 Hybridization detection 22
______________________________________________________________________ Introduction<br />
on a chip 4 . Since electrochemical detection results directly into an electric signal, there is no<br />
need for an expensive transduction equipment 56 . Electrochemical transduction of the<br />
hybridization event can be divided into direct and indirect DNA detection.<br />
Direct DNA detection relies either on changes in the electrical properties of the interface caused<br />
by hybridization or a direct oxidation of nucleic bases 48 . In the 1960s Paleček and coworkers<br />
pioneered the work on direct reduction and oxidation of DNA at a mercury electrode 46 (Figure<br />
1.14). Later, the oxidation of purine bases of DNA was achieved using carbon, gold, indium tin<br />
oxide and polymer-coated electrodes 56 . Since guanine is the most redox active from all DNA<br />
bases, its oxidation was studied the most. Even though this method is quite sensitive, its main<br />
drawback are high background currents due to the high potentials required for direct DNA<br />
oxidation.<br />
Figure 1.14. Reduction (red) and oxidation (blue) sites of DNA bases for direct DNA<br />
electrochemical detection. Figure adapted with permission from ref. 59 . Copyright (2012)<br />
American Chemical Society.<br />
Electrochemical impedance spectroscopy (EIS) was extensively used to follow changes in the<br />
interface properties upon hybridization. The method measures the change in the faradaic<br />
impedance in the presence of redox species resulting from the hybridization event. Different<br />
strategies have been used for the amplification of the signal.<br />
Indirect DNA detection can be achieved through the use of electrochemical mediators, redox<br />
active indicators that non-covalently interact with dsDNA, labelling of the probe or target DNA<br />
or using sandwich type assays (Figure 1.15).<br />
1.4 Hybridization detection 23
______________________________________________________________________ Introduction<br />
DNA detection via redox mediators was employed using various mediators such as<br />
[Ru(bpy)3] 2+ , [Co(phen)3] 3+ , ferrocene or [Fe(CN)6] 3-/4- (Figure 1.15, a). Catalytic oxidation of<br />
guanine using [Ru(bpy)3] 2+ on indium tin oxide (ITO) electrodes is a well-known example 60 .<br />
The approach involves addition of DNA into a solution containing the ruthenium complex,<br />
while the electrode is held at a potential suitable for the oxidation of the reduced form of the<br />
complex. The complex is regenerated by the oxidation of guanine from DNA. Consequently,<br />
the signal is enhanced proportionally to the amount of guanine available for oxidation, since<br />
the direct guanine oxidation is not possible at ITO electrodes.<br />
The characteristics of non-covalently interacting indicators are a different affinity towards dsand<br />
ssDNA. They can interact with DNA either by electrostatic binding, binding to the groove<br />
of dsDNA or intercalate into dsDNA (Figure 1.15, b). The most widely used indicators are<br />
daunomycin, proflavine, antraquinone, methylene blue, [Co(phen)3] 3+ and [Ru(NH3)6] 3+ . One<br />
of the initial studies on non-covalent interaction of compounds with DNA was done by<br />
Mikkelsen et al., using [Co(phen)3] 3+ as a dsDNA minor groove binder. The metal complex is<br />
positively charged and attracted by negatively charged DNA, resulting in a higher current for<br />
the more negatively charged dsDNA 61 . Furthermore, Barton and coworkers pioneered the work<br />
on long-range charge transfer resistance through DNA, using electrochemically active<br />
intercalators 62 (methylene blue and daunomycin). Interaction of non-electrochemically active<br />
intercalating compounds was investigated using EIS 63 .<br />
First covalently bound DNA markers were investigated in the beginning of the 1980s 46 .<br />
Labelling of the DNA can be performed using the probe DNA, where the label is positioned at<br />
the distant end of the probe (Figure 1.15, c). Due to the flexibility of the ssDNA, without the<br />
presence of target DNA, the label is close enough to the surface and the signal is detected. Upon<br />
hybridization the signal switches off due to the rigidity of the dsDNA. Namely, after the<br />
hybridization the label is too far away from the surface for the electron transfer to occur.<br />
Labelling of the target DNA is also possible, and it is usually done at the end that is close to the<br />
surface upon hybridization (Figure 1.15, d). Therefore, after hybridization the signal is switched<br />
on. However, this approach requires a preparation step in which the target DNA from the sample<br />
needs to be labelled, which prolongs the assay time. In a sandwich-type assay, immobilized<br />
probe DNA is initially hybridized with a portion of the non-labelled target DNA from the<br />
sample, and subsequently a labelled signal DNA is hybridized with the overhang of the target<br />
DNA (Figure 1.15, e). Enzymes are readily used as labels for this approach 64 . In this case, DNA<br />
1.4 Hybridization detection 24
______________________________________________________________________ Introduction<br />
detection can be performed using a redox mediator in solution. An alternative approach is to<br />
hybridize a signal DNA with an overhang from the target DNA that is placed close to the<br />
surface 65 . This way there is no need for the redox mediator to obtain the signal.<br />
Figure 1.15. Indirect DNA detection using a) electrochemical mediators, b) non-covalently<br />
interacting indicators, c-d) labelling of probe or target DNA and e) sandwich type<br />
detection.<br />
1.4 Hybridization detection 25
2. Aims of the Work
_________________________________________________________________ Aims of the Work<br />
Self-assembly finds its application in many different fields of research, yet the most<br />
used protocol is the same in all of them – simple immersion of the material of choice into a<br />
solution containing the molecule to be assembled, followed by a spontaneous adsorption and<br />
formation of appropriate bonds 66 . Among SAMs self-assembly of thiolated molecules on gold<br />
is the most known.<br />
While, for example, the aim of SAM formation of alkylthiols is generally to reach highly<br />
compact monolayers, the immobilization of thiolated DNA is performed aiming at a wide range<br />
of DNA coverages, depending on the envisaged sensing strategy. Nevertheless, in order to<br />
obtain high coverages of thiolated molecules a long incubation time is required, ranging from<br />
several hours to days 13-15 . In contrast, low coverages can be obtained in a shorter time, however,<br />
with a significant variation of densities and substantially decreased reproducibility 67 .<br />
To meet the demands of the market, future point-of-care devices have to link high-quality<br />
performance with speed, simplicity and low production costs 4 . Despite its simplicity of<br />
formation of comparably stable films, thiol chemisorption on gold is not yet the chemistry of<br />
choice for fabrication of commercial DNA biosensors, partly due to issues with reproducibility<br />
and the time required for modification. In order to profit from possibilities that self-assembly<br />
offers, the dependence on the spontaneous adsorption process, that is very slow and lacks<br />
reproducibility, needs to be eliminated. A new and simple strategy that will allow to<br />
reproducibly control the surface modification in a desired manner, and what is equally<br />
important, in a very short time, would significantly decrease the production costs and with this<br />
the cost of a final point-of-care device.<br />
Base-by-base DNA sequence synthesis made a tremendous contribution to fabrication of highly<br />
dense DNA chips for genome sequencing. However, the complex nature of chemical synthesis<br />
and very expensive production, coupled with a limited flexibility for customization and length<br />
of probe sequences, makes this technique less suitable for point-of-care devices 51 . Moreover,<br />
the very high number of individual test sites is often not necessary for specific measurements<br />
of point-of-care applications. Furthermore, the more flexible and cheaper approach, namely<br />
spotting of pre-synthesized DNA sequences, still suffers from several drawbacks, such as<br />
special working conditions and problems with accuracy and efficiency. Nevertheless, new<br />
approaches for the production of DNA chips are arising, including electrochemically driven<br />
surface modification, demonstrating the evolution of microarray technology to more practical<br />
platforms for diagnostic applications.<br />
27
_________________________________________________________________ Aims of the Work<br />
This thesis focuses on the development of a new strategy for the gold surface modification by<br />
thiolated molecules that will enable for a controlled immobilization of molecules, regardless of<br />
their charge in a fast manner. The main envisaged application of the proposed strategy is in the<br />
production of DNA chips as means for overcoming limitations in the development of point-ofcare<br />
devices.<br />
The initial part of this work focuses on understanding the processes occurring at the interface<br />
during surface modification, taking into consideration the interdependence of physico-chemical<br />
properties of the investigated molecules, electrode polarization and the surrounding solution,<br />
with the goal of finding strategies to improve the kinetics and the reproducibility of immobilization<br />
of both intrinsically charged molecules, such as DNA, and uncharged molecules, such<br />
as alkylthiols. The influence of potential pulses on the surface modification is explored as a<br />
new strategy to obtain high-quality DNA sensing platforms, and the advantages of this approach<br />
over the standard passive approach are investigated. Moreover, this work focuses on<br />
implementing the developed strategy into the fabrication of a DNA array for multiple probe<br />
detection. The potential pulse-assisted cleaning of Au modified surfaces is also investigated<br />
with the aim to regenerate the Au surfaces within a very short time, while not causing any<br />
damage to the electrode surface. The last part of the thesis focuses on the implementation of the<br />
newly established surface modification strategy into the development of a new DNA sensing<br />
platform based on the signal amplification via enzyme-conjugated intercalating compound as a<br />
hybridization indicator. Challenges related to the use of intercalators are mainly reflected in the<br />
need to prevent non-specific adsorption to obtain a desired contrast between ss- and dsDNA.<br />
Therefore, the ability of the developed surface modification strategy to create high-quality DNA<br />
sensing surfaces for this application is evaluated.<br />
28
3. Results and<br />
Discussion
_____________________________________________________________<br />
Results and Discussion<br />
3.1 Importance of preparing the surface. Criteria for cleanliness<br />
Building of the automatic polishing machine was the author’s idea. The machine was<br />
constructed by Dr. Kirill Sliozberg. Optimization of the mechanical cleaning procedure was<br />
done by the author.<br />
Even though the surface preparation is cardinally important for the quality and reproducibility<br />
of the measurements especially in the field of biosensors and self-assembly, it is often<br />
underestimated and neglected. Due to the nanoscale size of the DNA strands and its charge,<br />
surface architecture is a parameter that can significantly increase the sensitivity of a DNA<br />
sensor 68 . Depending on the surface pDNA loading can be higher or lower.<br />
Even though the interest of the thesis was not to investigate the influence of the surface<br />
roughness and the size of the furrows on the biosensor performance, special attention was given<br />
to the surface cleaning with the aim to improve the reproducibility of measurements. The<br />
preparation of electrodes consisted of several steps:<br />
- Mechanical cleaning using an automatic polishing machine<br />
- Electrochemical cleaning in H2SO4<br />
- Characterization of the surface by electrochemical impedance spectroscopy (EIS)<br />
- Characterization of the surface by cyclic voltammetry (CV)<br />
By reproducibly preparing the surfaces using an automatic polishing machine (polishing setup<br />
and procedure are explained in Section 5.3) the quality of the prepared electrodes was<br />
significantly improved in comparison with a previously used manual polishing procedure<br />
(Figures 3.1 and 3.2). Moreover, a reproducible roughness was achieved as one of prerequisites<br />
for minimal signal deviation between experiments. This allowed exclusion of the surface<br />
properties as one of the possible parameters affecting the already complex results.<br />
3.1 Importance of preparing the surface. Criteria for cleanliness 30
_____________________________________________________________<br />
Results and Discussion<br />
Figure 3.1. Optical microscope images of the electrode surface for a) a new electrode as<br />
received from a company, b) an electrode after polishing by hand with 3 µm, and c) 3 µm<br />
and 1 µm diamond polishing pastes.<br />
Figure 3.2. Optical microscope images of an electrode surface after cleaning with an<br />
automatic polishing machine following each cleaning step using: a) 3 µm, b) 1 µm, c) 0.5<br />
µm and finally d) 0.1 µm diamond polishing pastes.<br />
3.1 Importance of preparing the surface. Criteria for cleanliness 31
_____________________________________________________________<br />
Results and Discussion<br />
In addition to the roughness factor (for its determination see Section 5.3), the criterion for<br />
efficiency of electrochemical cleaning in H2SO4 was the shape of the CV. The electrode surface<br />
was considered clean only if the oxidation peaks of individual gold facets 69 were observed upon<br />
electrochemical cleaning as shown in Figure 3.3.<br />
Figure 3.3. Representative CV of a clean gold electrode. Electrochemical cleaning was<br />
performed in 0.5 M H2SO4 at a scan rate of 100 mV/s.<br />
Following the mechanical and electrochemical cleaning of the electrode surface, EIS and CV<br />
using the ferro/ferricyanide couple as redox probe were used to verify the electrode cleanliness<br />
and to allow for a comparison of separate electrodes. Afterwards, criteria were selected for the<br />
desired charge transfer resistance (Rct) derived from EIS and the peak current in the CV as a<br />
prerequisite for the use of prepared electrodes for subsequent surface modification.<br />
Representative EIS and CV of clean gold electrodes are shown in Figure 3.4.<br />
Modification of electrodes was performed immediately after mechanical and electrochemical<br />
cleaning and characterization of the bare electrodes since it was observed by EIS that upon<br />
storage of bare gold electrodes in air, H2SO4 or HClO4 unknown contaminants adsorb on the<br />
electrode surface (data not shown).<br />
3.1 Importance of preparing the surface. Criteria for cleanliness 32
_____________________________________________________________<br />
Results and Discussion<br />
Figure 3.4. a) EIS and b) CV of a clean 2 mm diameter gold electrode performed in<br />
K3[Fe(CN)6]/K3[Fe(CN)6] (5 mM each) in 10 mM phosphate buffer (PB), 20 mM K2SO4.<br />
EIS measurements were performed by applying a DC potential of 220 mV superimposed<br />
by a 5 mV AC perturbation. The frequency range was scanned from 30 kHz to 10 mHz.<br />
CV was performed at 100 mV/s scan rate.<br />
3.1 Importance of preparing the surface. Criteria for cleanliness 33
_____________________________________________________________<br />
Results and Discussion<br />
3.2 Importance of knowing the surface<br />
In order to tailor optimal DNA-modified surfaces for the envisaged sensing platforms it is of<br />
utmost importance to understand processes occurring at the electrode surface during the DNA<br />
assay build-up. Only in this way, the properties of the surface can be controlled in a reproducible<br />
manner. Electrochemical impedance spectroscopy is an excellent technique for this due to<br />
which it gained a wide popularity in the field of bioelectroanalysis 70,71 . It is a non-destructive<br />
technique allowing one to monitor surface modification without altering the system’s response.<br />
Moreover, due to its sensitivity it is a great tool for following very subtle changes in the surface<br />
architecture. EIS is a very informative technique that allows in depth investigation of processes<br />
occurring at the electrified interface by sequentially following each step of the build-up of DNA<br />
assays 5,72 .<br />
Furthermore, to understand the behavior of DNA strands in front of an electrode surface we<br />
need to investigate not only physico-chemical properties of the DNA itself but it is essential to<br />
also observe the electrode and the surrounding solution as important components of the system.<br />
Since the DNA is essentially a negatively charged polyelectrolyte (depending on the ionic<br />
strength of the surrounding solution), the charge of the surface has a significant impact on the<br />
behavior of DNA at the interface. Therefore, it is of great importance to know how the surface<br />
is polarized (positive or negative) upon application of different potentials. Consequently, the<br />
potential of zero charge (pzc) of the bare polycrystalline electrode was determined.<br />
Additionally, the influence of the surface modification with DNA on the pzc was also<br />
investigated.<br />
3.2 Importance of knowing the surface 34
_____________________________________________________________<br />
Results and Discussion<br />
3.2.1 Electrochemical impedance spectroscopy. DNA assay build-up<br />
Electrochemical impedance spectroscopy is based on applying a DC potential that is commonly<br />
an open circuit potential superimposed with an AC potential of small amplitude. Measuring of<br />
the resulting AC current signal allows sampling of modulus and phase of the response. The<br />
impedance is usually presented by plotting the real and imaginary components in a Nyquist<br />
plot. The principle of the method is explained in detail in Section 5.13.1.<br />
In this study, a modified Randles equivalent electric circuit was used for modelling the behavior<br />
of the electrode surface during each step of the DNA assay build-up (Figure 3.5). In the circuit,<br />
a solution resistance (Rs) is connected in series with a constant phase element (CPE), which<br />
represents the double layer by taking into account the roughness of a polycrystalline gold<br />
electrode, and an impedance of a faradaic reaction (consisting of a charge transfer resistance,<br />
Rct, and a Warburg element representing the semi-infinite linear diffusion of electroactive<br />
species to a flat electrode, W). Since the alteration of Rct is most pronounced as compared to<br />
other electric circuit elements, the change of Rct was followed during surface preparation.<br />
Figure 3.5. Randles equivalent electric circuit used for fitting of Nyquist plots obtained<br />
during the DNA assay preparation. Rs represents the solution resistance, Rct is the charge<br />
transfer resistance, CPE is a constant phase element representing the double layer and W<br />
is the Warburg element representing diffusion.<br />
The interpretation of the properties of DNA-modified gold electrode surfaces by means of EIS<br />
was based on the repulsion of a negatively charged free-diffusing redox mediator, namely<br />
[Fe(CN)6] 3-/4- , from the DNA-modified electrode in a solution of moderate ionic strength 3 . The<br />
extent of the DNA charge screening depends on the ionic strength of the solution above a certain<br />
3.2 Importance of knowing the surface 35
_____________________________________________________________<br />
Results and Discussion<br />
threshold 34 . Thus, DNA is screened more in solutions of higher ionic strength, decreasing by<br />
this the ability of the DNA to repel molecules of the redox mediator. Therefore, in order to<br />
achieve the desired sensitivity, the ionic strength of the working solution should be low enough<br />
to allow for DNA to manifest high enough effective negative charge to significantly block the<br />
redox mediator. On the other hand, the ionic strength needs to be high enough to not affect the<br />
stability of the double helix during measurements with dsDNA-modified electrodes.<br />
Figure 3.6 represents a schematic view of the electrode surface during the preparation of a DNA<br />
sensor via a two-step immobilization method. Initially, a thiol-tethered ssDNA is immobilized<br />
on the electrode creating a negatively charged interface (Figure 3.6, a and b). Consequently, the<br />
approach of the redox mediator is hindered and the electron transfer rate decreases. In EIS this<br />
is observed as an increase in Rct (Figure 3.7). The increase in Rct depends on the amount of<br />
immobilized DNA, where a higher increase is observed for a higher ssDNA coverage. It should<br />
be noted that the obtained EIS response is a result of the repulsion of the redox mediator by<br />
both negatively charged immobilized ssDNA and unspecifically adsorbed DNA strands 73 , as<br />
well as steric hindrance caused by lying ssDNA that physically blocks the access of the redox<br />
mediator. Due to the fact that ssDNA behaves as a flexible coil and that it orientates randomly<br />
on the electrode surface, especially at lower ssDNA coverage this response lacks in<br />
reproducibility.<br />
Figure 3.6. Scheme of the electrode surface during the build-up of the DNA sensor: a)<br />
bare electrode, b) ssDNA-modified electrode, c) ssDNA/thiol-modified electrode.<br />
Therefore, upon immobilization the electrode surface is covered with a mixture of DNA strands<br />
that are chemisorbed via a Au-S bond and DNA strands that are bound to the surface through<br />
the DNA backbone or bases. Additionally, grafted DNA strands also adsorb to the electrode by<br />
3.2 Importance of knowing the surface 36
_____________________________________________________________<br />
Results and Discussion<br />
coiling on the surface due to their persistence length and flexible behavior. This leads to a steric<br />
hindrance towards the hybridization with target DNA. Thus, in the next step of the DNA sensor<br />
the build-up the ssDNA-modified surface is passivated using self-assembling properties of thiol<br />
molecules (Figure 3.6, c), as originally proposed by Tarlov et al. 74,75 . The formed thiol<br />
monolayer removes unspecifically adsorbed ssDNA from the electrode surface and forces<br />
grafted DNA to obtain an orientation that is more favorable for hybridization. It should be noted,<br />
that the DNA still does not obtain a perpendicular position with respect to the surface but rather<br />
lies down on the thiol layer due to its flexibility. Consequently, since the flux of the redox<br />
mediator to the electrode surface is facilitated, a decrease in Rct is commonly observed in EIS<br />
measurements upon the passivation step (Figure 3.8). On the other hand, tDNA can also<br />
unspecifically bind on the bare electrode surface (data not shown) and therefore this step is very<br />
important for preventing the unspecific adsorption of the target DNA and with this a false<br />
signal.<br />
Figure 3.7. Typical Nyquist plots representing the Rct increase upon ssDNA immobilization.<br />
ssDNA immobilization was performed in 10 mM PB with 450 mM K2SO4<br />
containing 1 µM ssDNA by incubation at 37 °C. EIS was measured in 10 mM PB with 20<br />
mM K2SO4 containing equimolar concentrations (5 mM) of K3[Fe(CN)6] and K4[Fe(CN)6].<br />
A DC potential of 220 mV was applied superimposed with a 5 mV AC perturbation. The<br />
frequency range was from 30 kHz to 10 mHz.<br />
3.2 Importance of knowing the surface 37
_____________________________________________________________<br />
Results and Discussion<br />
Taking into account that the reproducibility of the immobilization step as detected by EIS is<br />
quite low, it is more informative to observe the ssDNA/thiol-modified surface. However, since<br />
this response contains the contribution of both passivation and DNA immobilization, it should<br />
be compared with the Rct obtained for an only thiol-modified electrode. Namely, Rct of the<br />
ssDNA/thiol-modified electrode needs to be higher than the Rct of the electrode modified only<br />
with the passivating thiol implying that a detectable amount of DNA was immobilized.<br />
Figure 3.8. Change in the Rct value upon surface passivation with MCH. ssDNA<br />
immobilization was performed for 15 min as stated in Figure 3.7. MCH passivation was<br />
done by incubation for 19 h in 10 mM PB containing 20 mM K2SO4 and 10 mM MCH.<br />
EIS measurements were performed as stated in Figure 3.7.<br />
Therefore, after ssDNA immobilization by incubation for 15 min and subsequent passivation,<br />
the obtained Rct is negligibly higher as compared to the Rct of the electrode modified only with<br />
the mercaptohexanol (MCH) for a time equal to the passivation duration (Figure 3.9). This<br />
implies that the amount of immobilized DNA is very low and not detectable by means of EIS.<br />
On the other hand, for longer immobilization times (2 h and 8 h) a significant increase in the<br />
Rct value is observed as compared to the thiol-modified electrode reflecting higher DNA<br />
coverages.<br />
3.2 Importance of knowing the surface 38
_____________________________________________________________<br />
Results and Discussion<br />
One of advantages of the electrochemical impedance spectroscopy is that it enables label-free<br />
detection of DNA hybridization, circumventing the need for dye, enzyme or redox labelling of<br />
the target DNA 71 . However, a careful design of the electrode architecture needs to be achieved<br />
to improve the sensitivity of the overall system. Upon DNA hybridization an increase in Rct is<br />
generally observed. However, there are different interfacial phenomena caused by the<br />
hybridization process that influence the access of the redox mediator to the surface.<br />
Figure 3.9. Comparison of the Rct obtained for a MCH-modified electrode with the Rct<br />
obtained for ssDNA/MCH-modified electrodes using different incubation times. The<br />
MCH-modified electrode was obtained by incubation of the bare gold electrode in 10 mM<br />
PB containing 20 mM K2SO4 and 10 mM MCH for 19 h. The ssDNA/MCH-modified<br />
electrode was prepared by initial ssDNA immobilization followed by subsequent MCH<br />
passivation (see Figures 3.7 and 3.8, respectively). EIS measurements were performed as<br />
stated in Figure 3.7.<br />
EIS measurements are performed in an intermediate ionic strength to assure that the DNA is<br />
not fully screened by counterions and that it still manifests some effective negative charge.<br />
Therefore, as a result of the hybridization, the amount of negative charge in front of the<br />
electrode surface increases leading to a stronger repulsion of the redox mediator and an increase<br />
3.2 Importance of knowing the surface 39
_____________________________________________________________<br />
Results and Discussion<br />
of the Rct. Moreover, the presence of additional DNA strands leads to an additional steric<br />
hindrance decreasing the access of the redox mediator and contributes to an increase in Rct with<br />
a strong dependence on ssDNA coverage. When the coverage is very low, additional DNA<br />
strands do not notably alter the surface architecture and the redox mediator access is negligibly<br />
hampered. At high ssDNA coverage which is however low enough to allow for maximum<br />
hybridization efficiency, the approach of the redox mediator to the electrode surface is more<br />
likely affected.<br />
Nevertheless, due to the low persistence length of ssDNA it behaves as random coil lying on<br />
top of the thiol passivation layer (Figure 3.10) still preventing the access of the redox mediator<br />
to the electrode surface. The structure of the formed dsDNA is more rigid as compared to the<br />
one of the ssDNA, leading to a more upright orientation of the dsDNA and facilitating the<br />
approach of the redox species. The degree of this influence depends on the length of DNA<br />
strands. Short DNA is expected to exhibit only a little impact while longer DNA causes a more<br />
significant change of the electrode architecture.<br />
Figure 3.10. Interfacial changes upon DNA hybridization. DNA obtains a more upright<br />
orientation due to the rigidity of the dsDNA. Furthermore, the amount of the negative<br />
charge increases and an additional steric hindrance arises due to an increased number of<br />
strands. Thiol passivation layer was intentionally not shown for better clarity.<br />
3.2 Importance of knowing the surface 40
_____________________________________________________________<br />
Results and Discussion<br />
3.2.2 Potential of zero charge of bare and DNA-modified electrodes<br />
Theoretical discussions were done with Dr. Fabio La Mantia. Parts of this section were<br />
published in ref. 5 : “D. Jambrec, M. Gebala, F. La Mantia, W. Schuhmann, Angew. Chem. Int.<br />
Ed. 2015, 54, 15064-15068; Angew. Chem. 2015, 127, 15278-15283.”.<br />
The potential of zero charge is a fundamental property of an electrode-electrolyte interface.<br />
Therefore, for the in-depth understanding of the surface behavior it is of high importance to<br />
know the value of the pzc. At the pzc the excess charge density on the electrode surface equals<br />
zero 76,77 . The term “potential of zero free charge” (pzfc) is also used for the same value in order<br />
to differentiate it from the so-called potential of zero total charge (pztc) that refers to the case<br />
when a specific adsorption occurs on the interface involving charge transfer. The pztc is defined<br />
as a potential at which the sum of free electronic excess charge density and charge density<br />
transferred in the adsorption processes equals zero 76 . In order to better understand these terms,<br />
three different situations at the electrode-electrolyte interface are shown in Figure 3.11, when<br />
the applied potential equals pz(f)c, pztc or any other value.<br />
In the literature there is a substantial variation of reported values for the pzc of gold electrodes 78 .<br />
Due to the polycrystalline nature of the Au surface, the use of different electrolyte solutions and<br />
the fact that the pzc is very sensitive to surface impurities, many different values were proposed.<br />
Therefore, we determined the pzc for our system with thoroughly cleaned Au electrodes in the<br />
same solution that was later used for the modification of the surface.<br />
The most common way to determine the pzc is to find the minimum of the differential<br />
capacitance in the Cd-E curve, predicted by the GC theory 77 . It should be noted that for systems<br />
in which specific adsorption occurs the capacitance minimum provides the value of the pztc.<br />
Since we used a solution that contains sulfates, the obtained value evidently represents the pztc.<br />
However, for convenience of the readers, the general term pzc is used in the following text.<br />
Based on the definition, the pzc can be described by equation 3.1:<br />
σ m = dC d<br />
dE = 0 (3.1)<br />
3.2 Importance of knowing the surface 41
_____________________________________________________________<br />
Results and Discussion<br />
where σm represents the excess charge of the metal and Cd the differential capacitance. This<br />
equation explains that at the pzc the Cd-E curve shows a minimum. In order to find this potential,<br />
we need to evaluate the nature of the differential capacitance and on what parameters it depends.<br />
surface at the potential of zero (free)<br />
charge<br />
σ m = 0<br />
σ i = 0<br />
σ ddl = 0<br />
surface at the potential of zero total<br />
charge<br />
σ m + σ i = 0<br />
σ ddl = 0<br />
surface at any other potential<br />
σ m + σ i + σ ddl = 0<br />
Figure 3.11. Schematic representation of the electrode-electrolyte interface upon<br />
polarization at different potentials. σm, σi and σddl stand for the excess charge in the metal,<br />
compact layer and diffuse double layer, respectively.<br />
3.2 Importance of knowing the surface 42
_____________________________________________________________<br />
Results and Discussion<br />
According to the GC model of the double layer 43 , Cd consists of two contributions (Figure 3.12,<br />
a), namely the capacitance of the compact layer (Ci) and the capacitance of the diffuse double<br />
layer (Cddl):<br />
1<br />
= 1 + 1<br />
(3.2)<br />
C d C i C ddl<br />
Figure 3.12. Schematic representation of a) potential profile at the electrode surface, and<br />
b) expected behavior of Cd in dependence on the applied potential and the ionic strength<br />
according to the GCS theory. Φ represents the potential distribution and d distance from<br />
the electrode surface. Figures adapted from 43 .<br />
This means that Cd is determined by the lower value of these two components. In theory, Ci<br />
does not depend on the ionic strength nor on the potential. On the other hand, Cddl depends on<br />
both of these values under certain conditions. When the applied potential is far away from the<br />
pzc, Cd is determined by Ci regardless on the ionic strength. However, when the potential<br />
approaches the pzc (Figure 3.12, b), Cddl becomes significant at high ionic strength and the term<br />
1/Cddl in Equation (3.2) becomes negligible. In that case, Cd is again determined by Ci and only<br />
a constant capacitance of Ci is observed. However, at very low ionic strength Cddl is small and<br />
the term 1/Cddl becomes significant as compared with 1/Ci. Cd is then determined by Cddl, which<br />
is represented as a minimum in the Cd-E curve. It should be noted, that this model is based on<br />
certain assumptions like the independence of Ci on the ionic strength and the potential, that<br />
result in some discrepancies in real systems. In reality, a Cd-E curve does not necessarily follow<br />
the predicted ideal U-shape.<br />
3.2 Importance of knowing the surface 43
_____________________________________________________________<br />
Results and Discussion<br />
In order to find the capacitance minimum in the Cd-E curve, potentiodynamic electrochemical<br />
impedance spectroscopy (PDEIS) was used. This technique consists of a set of EIS<br />
measurements performed for a desired range of potentials. Initially, EIS was performed for all<br />
solutions to identify the suitable frequency range for the determination of the pzc. This was<br />
done to verify whether the system can be fitted to the RC equivalent circuit and that there are<br />
no artifacts in the frequency range of interest. Very high frequencies were excluded since a high<br />
frequency region of the impedance spectra is very sensitive to the cell geometry 79 . To be<br />
specific, the solution resistance is smaller for the electrode edge than for the center of the<br />
electrode, leading to a mixed behavior of the system which cannot be fitted with a simple RC<br />
circuit. In order to minimize this behavior, a setup that assures homogeneous distribution of<br />
current lines was constructed, where the counter electrode completely surrounds the working<br />
electrode and the reference electrode is placed below them (Chapter 5.4, Figure 5.4). Moreover,<br />
very low frequencies were also avoided and with this any possible faradaic side-reactions.<br />
Furthermore, a capacitive bridge (shunt capacitor 80 ) was implemented and a Pb/PbF2 reference<br />
electrode was used that shows high impedance 81 in the measuring setup to eliminate possible<br />
instrumental artifacts due to very low ionic strengths and high frequencies. After these<br />
optimization steps Nyquist plots with the characteristic behavior for a system with the solution<br />
resistance and the double layer capacitance connected in series (Figure 3.13) were obtained. In<br />
an ideal case, a Nyquist plot of this system should consist of a vertical line with the intercepting<br />
the real axis at the solution resistance. However, as shown in Figure 3.13 a frequency dispersion<br />
which is common for polycrystalline electrodes is observed. There are various theories<br />
discussing the reasons for this phenomena, among which are the roughness of polycrystalline<br />
electrodes, that creates inhomogeneities in the current density along the surface, or adsorption<br />
effects 77,82 .<br />
After performing PDEIS for all solutions in the chosen frequency range and the potential range<br />
from -0.2 to 0.7 V (vs. Ag/AgCl/3 M KCl), the determination of Cd was done by extracting the<br />
imaginary impedance component from the Nyquist plots for each sampled potential value and<br />
calculating Cd values using the following equations for the RC equivalent electric circuit:<br />
Z = R −<br />
j<br />
ωC d<br />
(3.3)<br />
3.2 Importance of knowing the surface 44
_____________________________________________________________<br />
Results and Discussion<br />
C d = 1<br />
(3.4)<br />
2πfZ<br />
′′<br />
where ω is the angular frequency.<br />
Figure 3.13. Example of a Nyquist plot used for the system characterization and<br />
calculation of Cd. The presented curve was recorded for a potential of 0.8 V vs. Pb/PbF2/5<br />
M KF (0.2 V vs. Ag/AgCl/3 M KCl) in the frequency range from 4 kHz to 100 Hz. The<br />
measurement was performed in 10 mM PB with 20 mM K2SO4. A capacitive bridge (2 µF)<br />
was used. The beige line represents the fit to the RC equivalent circuit.<br />
Figure 3.14 shows the obtained Cd values for three electrolyte solutions of different ionic<br />
strengths calculated at three different frequencies. Comparing different ionic strengths, no clear<br />
minimum is observed (Figure 3.14, a-c). On the contrary, for the highest ionic strength a<br />
maximum was observed in one part of the curve. This behavior was previously observed for<br />
high ionic strength electrolytes 83 . Furthermore, looking at the individual ionic strength, the<br />
same trend can be observed for all three chosen frequencies. However, when different ionic<br />
strengths for the same frequency are compared (Figure 3.14, d-f), the biggest difference is<br />
observed between different ionic strengths for a frequency of 100 Hz (Figure 3.14, d).<br />
3.2 Importance of knowing the surface 45
_____________________________________________________________<br />
Results and Discussion<br />
Figure 3.14. Determination of Cd in at varying potentials of the bare gold electrode. Cd<br />
was measured from the imaginary impedance component for different ionic strengths: a)<br />
10 mM PB with 20 mM K2SO4, b) 1 mM PB with 2 mM K2SO4 and c) 0.1 mM PB with 0.2<br />
mM K2SO4 and for different frequencies: d) 100 Hz, e) 200 Hz and f) 500 Hz.<br />
3.2 Importance of knowing the surface 46
_____________________________________________________________<br />
Results and Discussion<br />
Taking into account Equation (3.2), it is assumed that the investigated ionic strengths are not<br />
low enough to result in the desired capacitance minimum from the Cd-E dependence alone.<br />
Namely, for the highest ionic strength (10 mM PB, 20 mM K2SO4) Ci dominates the overall<br />
capacitance response, however, at the lowest investigated ionic strength (0.1 mM PB, 0.2 mM<br />
K2SO4) obviously contributions from both Ci and Cddl are significant. Therefore, by subtraction<br />
of these two values Cddl can be obtained:<br />
1<br />
C ddl =<br />
1<br />
C<br />
− 1 (3.5)<br />
d C i<br />
Finally, by plotting the Cddl-E dependence (Figure 3.15), the pzc can be determined from the<br />
capacitance minimum at a potential of 0.5 V (vs. Ag/AgCl/3 M KCl), where the most<br />
pronounced minimum is observed for a frequency of 100 Hz. The non-ideal U-shape of the<br />
obtained curves is probably a consequence of the polycrystalline nature of the gold surfaces and<br />
the contribution of all crystallographic facets.<br />
Figure 3.15. Determination of the pzc of the bare gold electrode. Cddl values were<br />
calculated using the data from Figure 3.14 for different frequencies.<br />
To investigate the effect of the modification of the surface with DNA on the pzc value,<br />
determination of the pzc (pztc) of the DNA-modified electrode (pzc (DNA)) was performed.<br />
This was done in the same manner as for the pzc determination of bare gold, except that Ci was<br />
3.2 Importance of knowing the surface 47
_____________________________________________________________<br />
Results and Discussion<br />
replaced by the capacitance of the DNA monolayer (CDNA). Cd is represented by the following<br />
equation:<br />
1<br />
= 1 + 1<br />
(3.6)<br />
C d C DNA C ddl<br />
Upon DNA immobilization, Cd (Figure 3.16) and subsequently Cddl values (Figure 3.17) were<br />
obtained in relation to the applied potential using PDEIS. In this case, the investigated potential<br />
range was shifted to -0.4 V to 0.5 V (vs. Ag/AgCl/3 M KCl) to avoid the cleavage of the Au-S<br />
bond at higher potentials. The pzc (DNA) was determined to be around 0.1 V (vs. Ag/AgCl/3<br />
M KCl) showing that the pzc shifts to more negative values due to the surface modification<br />
with DNA. Unlike in the case of the bare polycrystalline gold electrode, for the DNA-modified<br />
electrode a well-defined minimum was observed. A possible explanation is a more<br />
homogeneous surface after the modification with the DNA layer. Furthermore, the broad<br />
minimum is in agreement with the GC double layer model that assumes peak broadening with<br />
decreasing ionic strength 43 .<br />
Figure 3.16. Determination of Cd at varying applied potentials at the DNA-modified<br />
electrode. Impedance was measured for two ionic strengths: 10 mM PB with 20 mM<br />
K2SO4 and 0.1 mM PB with 0.2 mM K2SO4. Cd was calculated at a frequency of 10 Hz.<br />
DNA immobilization: 0.5/-0.2 V pulse profile (10 ms pulse duration, vs. Ag/AgCl/3 M<br />
KCl), 5 min in 1 µM DNA solution prepared in 10 mM PB with 450 mM K2SO4. Procedure<br />
explained in the Section 3.3.1.<br />
3.2 Importance of knowing the surface 48
_____________________________________________________________<br />
Results and Discussion<br />
Figure 3.17. Determination of the pzc of the DNA-modified electrode. Cddl values were<br />
calculated using the data from Figure 3.16 obtained at a frequency of 10 Hz.<br />
3.2 Importance of knowing the surface 49
_____________________________________________________________ Results and Discussion<br />
3.3 Importance of controlling the surface<br />
The most popular strategy for the immobilization of thiolated molecules on gold surfaces is a<br />
spontaneous formation of Au-S bonds via self-assembly achieved by a simple immersion of the<br />
gold electrode into a solution containing a desired thiol 66 . This strategy is widely used for both<br />
SAM formation of alkylthiols and immobilization of thiolated DNA. The aim of SAM<br />
formation of alkylthiols is generally to achieve the highest possible coverage and to obtain<br />
compact layers with a high blocking ability. On the other hand, desired DNA coverage depends<br />
on the envisaged DNA detection strategy and it can range from low to high coverages 66,84,85 .<br />
Nevertheless, in order to obtain high coverage of thiolated molecules long incubation times are<br />
required, ranging from several hours to days 13-15 . In contrast, low DNA coverages can be<br />
obtained in a short time, but with the drawback of significant variation of densities 67 .<br />
Therefore, a new immobilization strategy needs to be introduced that allows to reproducibly<br />
control the surface modification in a desired manner and what is equally important, in a very<br />
short time. The new approach needs to eliminate the dependence on the spontaneous selfassembly<br />
that is very long and lacks reproducibility. Thus, the possibility of surface control by<br />
potential-assisted surface modification was investigated using both thiolated DNA and<br />
alkylthiols as examples of intrinsically charged and uncharged molecules, respectively.<br />
Furthermore, with the aim of using the envisaged potential pulse-assisted immobilization<br />
method for the preparation of DNA arrays, this approach has to allow array modification with<br />
multiple DNA probes. Due to the need for an electrochemical system (reference and counter<br />
electrodes in addition to the chip working electrode) to perform potential pulsing and the size<br />
of the individual electrodes on an array (usually μm dimensions) it is obvious that more than<br />
one electrode of the array needs to be exposed to the solution used for modification. Therefore,<br />
to prevent crosstalk between electrodes, each electrode needs to be cleaned prior to the<br />
modification. Thus, potential pulse-assisted cleaning of Au modified surfaces was investigated<br />
with the aim to regenerate Au surfaces within a very short time, while not causing any damage<br />
to the surface.<br />
3.3 Importance of controlling the surface 50
_____________________________________________________________ Results and Discussion<br />
3.3.1 Fast and controlled formation of DNA surfaces. Optimization of ssDNA<br />
immobilization procedure<br />
Theoretical discussions were done with Dr. Magdalena Gebala and Prof. Dr. Fabio La Mantia.<br />
Parts of this section were published in ref. 5 : “D. Jambrec, M. Gebala, F. La Mantia, W.<br />
Schuhmann, Angew. Chem. Int. Ed. 2015, 54, 15064-15068; Angew. Chem. 2015, 127, 15278-<br />
15283.” written by the author. Figure adapted from ref. 5 .<br />
The amount and accessibility of immobilized ssDNA on the electrode surface greatly influences<br />
the later hybridization process. Therefore, well-defined reproducible and controlled DNAmodified<br />
surfaces are a prerequisite for the development of optimized DNA sensors 3 . Even<br />
though ssDNA-modified surfaces are the foundation for all envisaged applications, the<br />
mechanism of ssDNA immobilization has not been fully elucidated yet. DNA immobilization<br />
is a complex process depending on many parameters such as ionic strength, strand length, and<br />
the valence of the ions screening the charge of the DNA strands 86 . By introducing varying<br />
applied potentials, this process becomes even more complex. While it is known that certain<br />
potentials (positive or negative with respect of the pzc) induce the bending or lifting of DNA at<br />
the electrified interfaces, the reasoning that the DNA itself is attracted by a positively charged<br />
3.3 Importance of controlling the surface 51
_____________________________________________________________ Results and Discussion<br />
electrode or repelled by a negatively charged electrode is not justified for a set of parameters<br />
usually employed for surface modification with DNA.<br />
Figure 3.18. Nyquist plots with typical Rct values obtained for different immobilization<br />
times where the immobilization was performed at OCP (incubation method). The MCHmodified<br />
electrode was prepared by incubation of the bare gold electrode in a solution of<br />
10 mM MCH with 10 mM PB and 20 mM K2SO4 for 19 h at 37 °C. ssDNA immobilization,<br />
MCH passivation and EIS measurements were performed as stated in Figures 3.7 and 3.8.<br />
Experiments were performed using different electrodes.<br />
DNA immobilization performed at OCP by simple immersion of the gold electrode into a DNA<br />
containing solution is diffusion controlled. Initial immobilization of DNA strands is relatively<br />
fast, while with an increasing amount of immobilized ssDNA grafting of additional strands<br />
becomes energetically unfavorable and the diffusion of new strands is hindered. Since DNA<br />
strands can unspecifically adsorb onto the electrode, lying DNA strands sterically hinder the<br />
approach of new strands and with this the formation of Au-S bonds. Figure 3.18 presents the<br />
immobilization kinetics obtained by performing ssDNA immobilization at OCP followed by<br />
EIS. The immobilization efficiency was investigated by studying the surface after the<br />
passivation step (for reasons explained earlier, Section 3.2.1) and evaluated by comparing the<br />
3.3 Importance of controlling the surface 52
_____________________________________________________________ Results and Discussion<br />
Rct of ssDNA/MCH-modified electrodes with the Rct of an electrode modified only with MCH<br />
under the same conditions that were used for the passivation of the surface. When the incubation<br />
is done only for 15 min a minor increase in Rct is observed as compared with the MCH-modified<br />
electrode implying that a negligible amount of DNA is immobilized. Rct increases notably after<br />
2 h of incubation, however, by prolonging the incubation time to 4 and 8 h the immobilization<br />
kinetics drastically slows down, which is manifested in the EIS plots as a small additional<br />
increase in Rct.<br />
In order to design a potential-assisted immobilization method that will lead to a significantly<br />
improved immobilization kinetics and the formation of well-defined and reproducible DNAmodified<br />
surfaces, the influence of several parameters on the behavior of DNA in the vicinity<br />
of the electrode surface needs to be taken into consideration (Figure 3.19). As it was shown<br />
earlier (Section 3.2.2), the pzc shifts to more negative potential values due to the modification<br />
of the electrode with DNA. This shift requires a careful selection of the applied pulse potentials<br />
to obtain control of the immobilization process.<br />
Figure 3.19. Scheme representing dominating parameters influencing the behavior of<br />
DNA at the electrode surface.<br />
Furthermore, in order to understand the behavior of DNA at the surface of a polarized electrode<br />
we need to observe the potential profile at the electrode surface developed upon applying a<br />
certain potential. By applying a potential more positive or negative with respect to the pzc, an<br />
excess charge is created on the metal side of the interface. As a response a double layer is<br />
3.3 Importance of controlling the surface 53
_____________________________________________________________ Results and Discussion<br />
formed in the electrolyte in order to compensate for this excess charge. A potential profile is<br />
formed and the potential decays with the distance from the electrode as it is explained in Section<br />
1.2.2. Even though the Gouy-Chapman double layer theory does not take into account the linear<br />
potential drop within the Helmholtz plane, it shows the difference in the system response in<br />
relation to parameters such as the ionic strength of the solution or the applied potential. In order<br />
to increase the immobilization kinetics and DNA coverage, DNA immobilization is often<br />
performed in solutions of high ionic strength. However, according to the GC model, the<br />
potential drop is steeper with increasing ionic strength. Therefore, under conditions of high<br />
ionic strength a significant potential drop is observed in the immediate proximity of the<br />
electrode surface. Thus, only a small fraction of a DNA strand in close vicinity of the electrode<br />
can be affected by the applied potential.<br />
Moreover, DNA is a highly negatively charged polyelectrolyte that strongly interacts with<br />
surrounding ions resulting in charge compensation, i.e., DNA screening (described in Section<br />
1.2.1). In the case of monovalent cations, the charge at a DNA strand is screened by counterions<br />
accumulating around the DNA strand in two layers, namely a condensed layer and additional<br />
ions in a second sphere 34 . Therefore, due to the absence of an effective net charge, a DNA strand<br />
cannot be directly affected by the applied potential as it is generally suggested.<br />
These observations imply that the polarized electrode neither attracts nor repels DNA strands<br />
directly, but rather affects the ions in the vicinity of the electrified interface. Namely, during<br />
the charging of the electrochemical double layer, ions have to rearrange in both Helmholtz<br />
planes and the diffuse layer. Thus, when the electrode is polarized to negative values with<br />
respect to the pzc, cations move towards the electrified interface while anions move towards<br />
the bulk of the solution and vice versa. This suggests that switching fast enough between these<br />
two situations creates a “stirring effect” that effectively exceeds the Debye length in front of<br />
the electrified interface. Furthermore, efficient stirring should also move DNA strands present<br />
in close proximity to the electrode surface including their condensed ion cloud. This way the<br />
immobilization will not be diffusion controlled but driven by the migration of ions in front of<br />
the electrode. Based on this hypothesis, we created a potential pulse-assisted immobilization<br />
method that consists of fast switching between potentials more positive and more negative with<br />
respect to the pzc. Figure 3.20 demonstrates the principle of the measurements conducted<br />
during this study.<br />
3.3 Importance of controlling the surface 54
_____________________________________________________________ Results and Discussion<br />
Figure 3.20. a) Potential-time dependence representing potential pulses during potentialassisted<br />
immobilization and b) corresponding current-time response.<br />
It becomes evident that in order to find an efficient pulse profile the following conditions need<br />
to be satisfied:<br />
- to find appropriate pulse intensities<br />
- to find an appropriate pulse duration<br />
Applied potential intensities need to be on the one hand within the stable potential window of<br />
the Au-S bond and on the other hand high enough to evoke an efficient stirring to bring the<br />
DNA towards the surface. Namely, even though according to the G-C theory the potential drop<br />
in front of the electrode is steeper for higher applied potentials, the absolute value of the<br />
potential Φ at a fixed distance from the electrode is higher for a higher applied potential (i.e.,<br />
potential at the OHP, Φ0). This difference is more pronounced in close proximity to the<br />
electrode surface. To demonstrate the influence of potential intensities in a pulse profile on the<br />
efficiency of the immobilization process, different pulse profiles were compared (Figures 3.21<br />
and 3.22). For the purpose of this study the total duration of the immobilization was kept<br />
constant to 15 min and the pulse duration was kept at 10 ms. Using the same upper potential<br />
(0.5 V vs. Ag/AgCl/3 M KCl) and varying the lower potential, three potential differences were<br />
defined: 300 mV (0.5/0.2 V pulse profile), 500 mV (0.5/0 V pulse profile) and 700 mV (0.5/-<br />
0.2 V pulse profile). Figure 3.21 shows how the chosen potential differences relate to both pzc<br />
3.3 Importance of controlling the surface 55
_____________________________________________________________ Results and Discussion<br />
(bare) and pzc (DNA). Namely, all pulse profiles have an upper potential equal to the pzc (bare).<br />
Furthermore, they all relate differently to the pzc (DNA). For the 300 mV pulse profile both<br />
potentials are always positive with respect to the pzc (DNA). For the 500 mV pulse profile<br />
pulsing occurs between positive and slightly negative potentials with respect to the pzc (DNA)<br />
and in the 700 mV pulse profile both positive and negative potentials with similar amplitude<br />
with respect to the pzc (DNA) are applied.<br />
Figure 3.21. Selected potential pulse profiles with respect to the pzc (bare) and pzc (DNA):<br />
ΔE = 300 mV (pulse profile 0.5/0.2 V), ΔE = 500 mV (pulse profile 0.5/0 V) and ΔE = 700<br />
mV (pulse profile 0.5/-0.2 V). Figure adapted from ref. 5 .<br />
By applying a 300 mV pulse profile a small increase in Rct as compared to the Rct of the MCHmodified<br />
electrode is observed showing that obviously only a very small amount of DNA was<br />
immobilized (Figure 3.22, blue curve). However, the amount of immobilized DNA still exceeds<br />
the amount of DNA immobilized by means of the incubation method after the same time (Figure<br />
3.22, black curve). This shows that the investigated pulse profile accelerates the immobilization<br />
of DNA. After the initial immobilization of a minor amount of DNA, the pzc shifts from pzc<br />
(bare) to a more negative value (pzc (DNA)). Therefore, since the pulsing occurs only between<br />
positive potentials relative to pzc (DNA) and the potential difference is small, only a minor ion<br />
stirring effect is achieved and upon initial DNA immobilization, the DNA strands remain lying<br />
on the electrode surface. The electrode is physically blocked by the initially immobilized DNA<br />
strands and the access of additional DNA molecules is impeded.<br />
3.3 Importance of controlling the surface 56
_____________________________________________________________ Results and Discussion<br />
A similar result is observed for the pulse profile with 500 mV potential difference (Figure 3.22,<br />
yellow curve). Although the potential difference is higher than for the 0.5/0.2 V pulse profile,<br />
the applied negative potential is apparently not sufficiently low to induce effective enough ion<br />
stirring and push up already immobilized DNA strands that are lying on the surface as a random<br />
coil due to their low persistence length. Therefore, the access of new strands is hindered and a<br />
poor immobilization yield is observed. On the other hand, further increase of the potential<br />
difference to 700 mV (pulse profile 0.5/-0.2 V) leads to a significantly improved immobilization<br />
yield (Figure 3.22, green curve). Evidently, this pulse profile is vigorous enough to increase the<br />
ion flux and to induce improved ion stirring. Thus, already immobilized DNA strands can be<br />
lifted from the electrode surface and can create space for new strands to approach the surface.<br />
As explained earlier, even though for higher applied potentials the potential drop is steeper, at<br />
a certain distance from the electrode, the DNA is affected by an absolute higher potential.<br />
Figure 3.22. Comparison of different pulse profiles used for potential-assisted DNA<br />
immobilization. MCH-modified electrode was prepared as stated in Figure 3.8. ssDNA<br />
immobilization was performed for 15 min by incubation at OCP or using different pulse<br />
profiles (0.5/0.2 V, 0.5/0 V and 0.5/-0.2 V) with 10 ms pulse duration. MCH passivation<br />
and EIS measurements were performed as stated in Figures 3.7 and 3.8. Figure adapted<br />
from ref. 5 .<br />
3.3 Importance of controlling the surface 57
_____________________________________________________________ Results and Discussion<br />
However, the selected potential values are not the only parameter that determine the efficiency<br />
of the potential-assisted immobilization. Namely, after increasing the potential difference to<br />
900 mV (pulse profile 0.5/-0.4 V) while keeping the pulse time at 10 ms no significant increase<br />
in Rct is observed (data not shown) suggesting that the pulse duration needs to be adjusted with<br />
respect to the selected pulse profile.<br />
In order to demonstrate the effect of the pulse duration on the efficiency of immobilization,<br />
different pulse times were investigated (1, 10, and 100 ms) keeping the total immobilization<br />
time and the pulse profile constant (Figure 3.23). The total duration was 5 min and the 0.5/-0.2<br />
V pulse profile was used. Decreasing the pulse time to 1 ms significantly increases the number<br />
of cycles to 150,000 for the total immobilization time of 5 min. However, at 1ms pulse time the<br />
lowest Rct was obtained among all tested pulse times. When the pulse duration is prolonged to<br />
10 ms and the number of cycles is 15,000, a significant increase in the immobilization efficiency<br />
is observed as shown in a substantial increase in Rct. By prolonging the pulse duration even<br />
more to 100 ms, only 1,500 cycles are completed and Rct again drops to lower values.<br />
Figure 3.23. Comparison of different pulse durations used for potential-assisted DNA<br />
immobilization. ssDNA immobilization was performed for 5 min using the 0.5/-0.2 V pulse<br />
profile with 1, 10 and 100 ms pulse duration. MCH passivation and EIS measurements<br />
were performed as stated in Figures 3.7 and 3.8. Figure adapted from ref. 5 .<br />
3.3 Importance of controlling the surface 58
_____________________________________________________________ Results and Discussion<br />
The obtained results suggest that when the pulse duration is too short to allow for a whole DNA<br />
strand to be pulled down via ion stirring during a single pulse (e.g. 1 ms), only a small fraction<br />
of the DNA in the vicinity of the electrode that is already oriented with the anchor group towards<br />
the electrode surface is grafted (Figure 3.24). On the other hand, if the pulse duration is too long<br />
for the investigated molecule length (e.g. 100 ms), a DNA strand will be brought completely to<br />
the surface during a fraction of a single pulse and it will remain lying on it until a negative pulse<br />
is applied. This creates less efficient ion stirring and, in addition, reduces the number of pulse<br />
cycles. For the investigated system a pulse time of 10 ms is apparently long enough for complete<br />
DNA strands to be pulled down, hence allowing for the formation of the Au-S bond. Moreover,<br />
the 10 ms pulse time is short enough to allow for a high number of potential pulse cycles.<br />
Figure 3.24. Schematic representation of the influence of the pulse duration on the DNA<br />
immobilization efficiency.<br />
After presenting the effect of applied potential intensities and the duration of a single pulse on<br />
the efficiency of potential-assisted DNA immobilization, a model that explains the processes<br />
occurring during grafting of DNA strands supported by applied potential pulses can be proposed<br />
(Figure 3.25). Due to the short distance to which applied potentials have an effect at a given<br />
ionic strength during the application of a single positive potential pulse, cations that are in the<br />
vicinity of the electrode are repelled from the electrode surface including counterions<br />
surrounding DNA strands that are located within the layer of influence. By this, an increased<br />
effective charge at the DNA strand is achieved. The extent of the layer of influence depends on<br />
the intensity of applied potentials and pulse duration. If the pulse is long enough, a wider layer<br />
3.3 Importance of controlling the surface 59
_____________________________________________________________ Results and Discussion<br />
of influence is attained with a higher applied potential. Conversely, anions in the same layer are<br />
attracted towards the electrode and with this a small fraction of a DNA strand that is in<br />
immediate proximity to the electrode is pulled towards its surface. As a result, the remaining<br />
part of a DNA strand that was outside of the area of influence is brought closer to the electrode<br />
surface. If the same potential is still applied the next part of the DNA strand undergoes the same<br />
process; DNA counterions are repelled together with cations from the solution and the rest of<br />
the DNA strand is brought closer to the surface. This process continues as long as the positive<br />
potential pulse is applied. If the duration of a pulse is long enough, the complete DNA strand<br />
is sequentially confined on the electrode surface. Therefore, regardless of the orientation of the<br />
DNA strand, the anchor group is brought close enough to the electrode surface for the formation<br />
of the Au-S bond.<br />
Figure 3.25. Scheme presenting the zipper-like pulling of a DNA strand towards the<br />
electrode surface during potential-assisted DNA immobilization. Figure adapted from<br />
ref. 5 .<br />
The positive potential pulse is followed by a potential jump to a negative potential with respect<br />
to the pzc (DNA). If an appropriate negative potential is applied, anions in the vicinity of the<br />
electrode are repelled from the surface together with the remaining negatively charged DNA<br />
backbone that was left unscreened. By this, the grafted DNA strand are lifted towards the bulk<br />
of the solution. Thus, additional space is created for new DNA strands to approach to the<br />
3.3 Importance of controlling the surface 60
_____________________________________________________________ Results and Discussion<br />
surface. Repetitive switching between positive and negative potentials leads to an increase in<br />
the amount of immobilized DNA due to the created ion stirring that facilitates the approach of<br />
DNA to the surface and the formation of the Au-S bond.<br />
As it was shown previously, that DNA immobilization at OCP exhibits a very slow kinetics.<br />
The proposed potential-assisted method seems to overcome diffusion limitations of the<br />
incubation method by controlling the immobilization kinetics via migration of ions causing ion<br />
stirring. In order to compare the developed method with the immobilization of DNA at OCP,<br />
potential-assisted DNA immobilization was performed using the optimal pulse profile (0.5/-0.2<br />
V with 10 ms pulse duration) while varying the total immobilization time and subsequently<br />
passivating the surface by incubation in MCH solution for the same duration as for DNA<br />
immobilization at OCP (19 h).<br />
Comparing the kinetics of these two methods it is clear that the efficiency of potential-assisted<br />
DNA immobilization is tremendously higher as compared to the incubation method (Figure<br />
3.26). A significant increase of the Rct value is obtained after immobilization for only 15 min<br />
as compared to the MCH-modified electrode. The DNA coverage determined for this<br />
immobilization time using the chronocoulometric method proposed by Steel et al. 75 was<br />
(6.85±0.47) × 10 12 DNA molecules/cm 2 , which is considered to be within an optimal DNA<br />
coverage range for application in DNA sensors 67,75,87,88 . On the other hand, using the incubation<br />
method only a negligible Rct increase is observed after 15 min of immobilization as compared<br />
to the MCH-modified electrode. For 2 h of immobilization at OCP, a coverage of (4.65±0.26)<br />
× 10 12 DNA molecules/cm 2 was determined, which is 47 % lower than the amount of<br />
immobilized DNA achieved within 15 min by the potential-assisted method. Increasing the<br />
incubation time results in a significant decrease in the immobilization kinetics. In order to<br />
obtain the same Rct value that is achieved in only 15 min using the potential-assisted method,<br />
immobilization for about 8 h is needed when the electrode is modified using the commonly<br />
used incubation method.<br />
Figures 3.26b and 3.27 show that using the potential-assisted DNA immobilization method<br />
much higher Rct values, that is, higher DNA coverages, can be achieved within the investigated<br />
immobilization duration. This may be valuable for applications such as the investigation of<br />
DNA repair proteins 66,84 .<br />
3.3 Importance of controlling the surface 61
_____________________________________________________________ Results and Discussion<br />
Figure 3.26. Comparison of DNA immobilization methods. a) Nyquist plots with Rct values<br />
obtained for different immobilization times with the immobilization being performed by<br />
incubation and the potential-assisted method. b) Comparison of the immobilization<br />
kinetics using these two methods. ssDNA immobilization, MCH passivation and EIS<br />
measurements were performed as stated in Figures 3.7 and 3.8. Rct values are obtained<br />
from EIS measurements for ssDNA/MCH-modified electrodes by fitting Nyquist plots to<br />
the [R(Q[RW])] equivalent circuit. Figure adapted from ref. 5 .<br />
3.3 Importance of controlling the surface 62
_____________________________________________________________ Results and Discussion<br />
Figure 3.27. Change of DNA coverage with the duration of potential-assisted DNA<br />
immobilization. DNA coverage was determined as discussed in Section 5.9.3.<br />
In order to prove that the accelerated immobilization kinetics is indeed due to a generated ion<br />
stirring, for which the application of alternating between positive and negative potentials is<br />
important, DNA immobilization was also investigated applying constant potentials. During the<br />
course of the immobilization, a constant potential of 0.5 V or -0.2 V was applied and the total<br />
immobilization time was kept constant (15 min).<br />
As expected, applying a constant negative potential of -0.2 V (vs. Ag/AgCl/3 M KCl) during<br />
DNA immobilization leads to a very small increase in Rct upon ssDNA immobilization (Figure<br />
3.28, a). Under these conditions, charging of the double layer at the electrode surface occurs<br />
relatively fast and the equilibrium of the system is reached. Therefore, no significant effect of<br />
the applied potential is expected if the DNA is considered as a screened molecule. Similar<br />
results are expected as in the case of immobilization at OCP. If a simple electrostatic<br />
attraction/repulsion model is considered, no change of Rct with respect to the bare electrode<br />
should be observed since the negatively charged electrode would repel the DNA strands and<br />
fully prevent DNA immobilization. However, applying a constant positive potential would be<br />
expected to make a high impact on the immobilization efficiency, since a positively charged<br />
electrode is assumed to attract negatively charged DNA and facilitate the grafting process. This<br />
is not observed and applying a constant potential of 0.5 V (vs. Ag/AgCl/3 M KCl) during DNA<br />
immobilization leads to a similar result as in the case of the negatively charged electrode, where<br />
3.3 Importance of controlling the surface 63
_____________________________________________________________ Results and Discussion<br />
only a slightly higher Rct value is observed. Furthermore, the resulting blocking of the electrode<br />
surface is much lower than in the case of potential pulse-assisted DNA immobilization (Figure<br />
3.28).<br />
Figure 3.28. Comparison of the potential pulse-assisted immobilization method with<br />
immobilization at constant potentials. ssDNA-modified electrodes were characterized<br />
using a) EIS and b) CV. ssDNA-modified electrodes were obtained by potential-assisted<br />
DNA immobilization using the 0.5/-0.2 V pulse profile with 10 ms pulse duration and by<br />
applying a constant potential of -0.2 V or 0.5 V vs. Ag/AgCl/3 M KCl for 15 min. ssDNA<br />
immobilization was performed in 10 mM PB with 450 mM K2SO4 and 1 µM ssDNA. EIS<br />
measurements were performed as stated in Figure 3.7. CVs were performed in 10 mM<br />
PB, 20 mM K2SO4 containing 5 mM of K3[Fe(CN)6] and K4[Fe(CN)6] at 100 mV/s scan<br />
rate.<br />
Finally, ssDNA/MCH-modified electrodes obtained by DNA immobilization using the<br />
potential pulse-assisted method (0.5/-0.2 V vs. Ag/AgCl/3 M KCl, 10 ms pulse duration) or<br />
immobilization supported by a constant positive potential (0.5 V vs. Ag/AgCl/3 M KCl) were<br />
compared (Figure 3.29). The total immobilization and passivation times were kept constant. As<br />
expected, the Rct value obtained by applying 0.5 V during immobilization is comparable to the<br />
Rct value obtained for immobilization performed by a simple incubation. Furthermore, this<br />
amount is significantly lower as compared to the potential-pulse assisted immobilization, which<br />
additionally supports the suggested mechanism behind the developed approach.<br />
3.3 Importance of controlling the surface 64
_____________________________________________________________ Results and Discussion<br />
Figure 3.29. Nyquist plots with typical Rct values obtained by comparing the potentialassisted<br />
immobilization method with immobilization performed by simple incubation or<br />
supported by applying a constant potential. MCH passivation was done by incubation for<br />
19 h at 37 °C.<br />
Besides the ability of the potential-assisted immobilization method to tremendously accelerate<br />
the immobilization kinetics and to achieve much higher DNA coverages in a shorter time as<br />
compared to the incubation method or applications of constant potentials, a high reproducibility<br />
of the surface modification needs to be attained. Figure 3.30 presents average Rct values<br />
obtained with the potential-assisted immobilization method using different pulse profiles. In all<br />
cases Rct values are obtained with a standard deviation below 5 %, showing that the developed<br />
immobilization protocol leads to a highly reproducible surface modification. Thus, the<br />
envisaged ssDNA surface coverage can be pre-selected by choosing the number of applied<br />
potential pulse cycles.<br />
Looking at the results presented in this chapter it can be concluded that understanding the<br />
behavior of ssDNA in front of the electrode surface affected by the surrounding electrolyte and<br />
the polarization of the electrode is essential for the development of highly reproducible DNAmodified<br />
surfaces. The developed potential pulse-assisted immobilization strategy leads to<br />
high-quality DNA-modified surfaces helping by this to overcome difficulties in DNA sensor<br />
3.3 Importance of controlling the surface 65
_____________________________________________________________ Results and Discussion<br />
preparation. Therefore, the proposed method may become a new standard DNA immobilization<br />
procedure to obtain desired surface coverages in a very short time as a prerequisite for the<br />
development of highly sensitive and reproducible DNA hybridization assays with<br />
electrochemical read-out.<br />
Figure 3.30. Reproducibility of the potential-assisted immobilization method. Data<br />
obtained from EIS measurements for ssDNA/MCH-modified electrodes by fitting Nyquist<br />
plots to a [R(Q[RW])] equivalent circuit. The pulse profile for DNA immobilization was<br />
performed for 15 min using a 10 ms pulse duration. The black dashed line represents the<br />
average fitted Rct value for the MCH-modified electrode. Error bars represent the<br />
standard deviation between measurements (n > 3). Figure adapted from ref. 5 .<br />
3.3 Importance of controlling the surface 66
_____________________________________________________________ Results and Discussion<br />
3.3.2 Formation of compact thiol SAMs within minutes<br />
Data analysis was done based on discussions with Dr. Felipe Conzuelo and Dr. Arturo Estrada-<br />
Vargas. Parts of this section were published in ref. 89 : “D. Jambrec, F. Conzuelo, A. Estrada-<br />
Vargas, W. Schuhmann, ChemElectroChem 2016, 3, 1484-1489.” written by the author. Figure<br />
adapted from ref. 89 .<br />
A typical approach for a DNA sensor buildup is a two-step immobilization strategy, in which<br />
initially a probe DNA is grafted on the surface and subsequently the electrode is covered by a<br />
thiol SAM. Thiol passivation forces the grafted DNA to lift up and to obtain a more favorable<br />
orientation for the hybridization process concomitantly removing unspecifically bound DNA<br />
strands. Additionally, the blocking ability of the passivation step in the DNA sensor preparation<br />
plays an important role in the efficiency of an envisaged DNA sensing scheme. It prevents<br />
unspecific adsorption of undesired molecules employed in the detection scheme and determines<br />
the signal of the negative control. Therefore, controlling of the blocking ability of the modified<br />
surface is crucial for construction of very sensitive and reproducible DNA sensors.<br />
3.3 Importance of controlling the surface 67
_____________________________________________________________ Results and Discussion<br />
Even though SAM formation has been an extensively investigated topic for decades until now,<br />
the implementation of procedures for a fast and controlled SAM deposition is of great interest,<br />
particularly when looking at future of point-of-care diagnostics devices, for which fast and<br />
cheap techniques for sensor fabrication are becoming more and more important in order to<br />
decrease production costs. The process of self-assembly supposedly consists of two<br />
phases 22,90,91 , during which in the initial stage thiol molecules randomly cover the surface, and<br />
in a second phase they slowly organize on the surface to form a monolayer. Although the twophase<br />
process is widely accepted, there are conflicting reports on the kinetics of SAM<br />
formation 92 .<br />
Most approaches for SAM formation involve a self-assembly process at open circuit potential.<br />
This is a simple method with, however, quite low reproducibility 24 and slow kinetics. It has<br />
been reported in the literature that the application of constant positive potentials accelerates the<br />
immobilization kinetics of long thiol chains providing compact SAMs in a shorter time 13,21,92 .<br />
Nevertheless, further improvement of the immobilization process is still needed, especially<br />
focusing on short-chain thiols commonly used for sensor fabrication 22 . Therefore, the<br />
application of the potential pulse-assisted immobilization method for alkylthiol derivatives as<br />
examples for uncharged molecules was investigated.<br />
One of the most reliable methods for the investigation of the self-assembly process is the<br />
measurement of the double-layer capacitance since it precisely describes the SAM adsorption<br />
properties. The interfacial capacitance of a gold-SAM-electrolyte interface consists of the<br />
capacitance of the SAM and the capacitance of the diffuse layer connected in series. The overall<br />
capacitance is determined by the smaller one, which is the capacitance of the SAM. EIS is a<br />
commonly used technique for the determination of the capacitance of a system. Real time<br />
measurement of electrochemical impedance allows the continuous determination of the<br />
capacitance with time and hence a detailed investigation of the SAM formation kinetics.<br />
Subramanian and Lakshminarayanan 92 used real time impedance monitoring to study the selfassembly<br />
mechanism of thiols by applying a constant DC potential superimposed with an AC<br />
signal at a single frequency. This measuring principle was modified in that way, that a pulsetype<br />
potential modulation employed for the above described potential-pulse assisted<br />
immobilization method was applied as DC potential.<br />
3.3 Importance of controlling the surface 68
_____________________________________________________________ Results and Discussion<br />
Figure 3.31. Schematic representation of the setup used for real-time impedance<br />
measurements during potential-pulse assisted acceleration of SAM formation: A – signal<br />
in, R – AC current magnitude, θ – AC current phase, AD/DA – analog-digital/digitalanalog<br />
conversion for potential application and data acquisition, RE – reference electrode,<br />
CE – counter electrode, WE – working electrode. The potentiostat receives the DC<br />
potential signal as potential pulses from the function generator and superimposes an AC<br />
signal with a single high frequency generated by the oscillator in the lock-in amplifier.<br />
The recorded current is fed back to the lock-in amplifier and the AC current at the<br />
excitation frequency is amplified providing the magnitude and the phase of the resulting<br />
AC response. Figure adapted from ref. 89 .<br />
The real-time impedance measuring setup consists of several components: a potentiostat, a<br />
function generator, a lock-in amplifier, an AD/DA card and an electrochemical cell (Figure<br />
3.31). The function generator is used to create a pulse-type DC potential modulation and to<br />
apply it to the external potential input of the potentiostat. In order to be able to apply potential<br />
pulses for accelerated SAM formation while simultaneously applying an AC frequency for<br />
impedance measurements, a summing amplifier is used to superimpose the square wave DC<br />
signal with a high-frequency AC signal. It should be noted that the AC signal needs to be of a<br />
significantly higher frequency with respect to the pulse time of the DC signal to prevent any<br />
influence of the small AC perturbation on the behavior of the investigated system. The current<br />
3.3 Importance of controlling the surface 69
_____________________________________________________________ Results and Discussion<br />
measured by the potentiostat is sent to the lock-in amplifier input after current-to-voltage<br />
conversion and the magnitude and the phase of the AC current response at the frequency of the<br />
AC excitation signal are obtained. Using these values, the imaginary impedance component and<br />
subsequently the interfacial capacitance can be calculated using the following expression,<br />
derived from an RC series equivalent electrical circuit (for detailed calculation see Section 5.5):<br />
C = −1<br />
ωZ ′′ (3.7)<br />
where ω is angular frequency and -Z" is the imaginary component of impedance.<br />
Figure 3.32. Mercaptoundecanol (MCU) SAM formation kinetics performed at constant<br />
potential equal to the previously measured OCP of the measuring solution (0 V vs.<br />
Ag/AgCl/3 M KCl). Measurement was performed in 10 mM PB with 20 mM K2SO4<br />
containing 1 mM MCU (30 % ethanol).<br />
As mentioned earlier, the kinetics of SAM formation at OCP is quite slow after the initial fast<br />
phase (Figure 3.32). During the first hour a relatively fast decrease in capacitance is manifested<br />
after which a significant deceleration of the immobilization kinetics is observed. Even after<br />
more than 12 h a compact thiol monolayer did not form, which is evident in the figure from the<br />
absence of a capacitance plateau. The measurement was conducted by applying a constant<br />
potential equal to the previously measured OCP. After determining the OCP, the measurement<br />
was performed applying a constant potential equal to the OCP superimposing an AC signal of<br />
3.3 Importance of controlling the surface 70
_____________________________________________________________ Results and Discussion<br />
high frequency. The fact that the OCP does not significantly change over the course of the SAM<br />
formation by simple incubation (data not shown) justifies this approach.<br />
As in the previous section, the influence of different parameters on the potential-pulse assisted<br />
SAM formation is investigated: potential intensities in the potential-pulse cycle and duration of<br />
the potential pulse (Figure 3.33). Emphasis in this chapter is on the dependence of these<br />
parameters on the length of a alkyl chain of the thiol derivative.<br />
Figure 3.33. Scheme of investigated potential pulse profiles (ΔE = 400 mV, 0.3/-0.1 V; ΔE<br />
= 700 mV, 0.5/-0.2 V; ΔE = 900 mV, 0.5/-0.4 V, all vs. Ag/AgCl/3 M KCl) and pulse<br />
durations (1 ms, 10 ms, 100 ms and 10 s). Figure adapted from ref. 89 .<br />
It was previously shown that the pzc shifts from 0.5 V vs. Ag/AgCl/3 M KCl (pzc of the bare<br />
electrode) towards more negative values due to surface modification with DNA strands. The<br />
same behavior is expected in the case of thiol SAM formation, since it was reported that the<br />
pzc of a thiol-modified electrode is around 0.1 V 93 (vs. Ag/AgCl/3 M KCl). Therefore, this shift<br />
in the pzc value was taken into account during the selection of the potential pulse intensities.<br />
Using the ΔE = 400 mV pulse profile (0.3/-0.1 V vs. Ag/AgCl/3 M KCl) potentials of +200 mV<br />
and -200 mV relative to the pzc value of the thiol-modified surface are applied, while for ΔE =<br />
700 mV (0.5/-0.2 V vs. Ag/AgCl/3 M KCl) +400 mV and -300 mV are applied and for ΔE =<br />
900 mV (0.5/-0.4 V vs. Ag/AgCl/3 M KCl) +400 mV and -500 mV are applied.<br />
Figure 3.34 shows capacitance curves obtained during the formation of a MCU SAM using<br />
different potential-pulse profiles. It presents the influence of potential intensities on the kinetics<br />
3.3 Importance of controlling the surface 71
_____________________________________________________________ Results and Discussion<br />
of SAM formation. All experiments were performed with a pulse duration of 10 ms and only<br />
the intensities of the applied potentials within a pulse profile were varied. The capacitance<br />
values are normalized by the real electrode surface area derived from CVs in sulfuric acid.<br />
Using the ΔE = 400 mV pulse profile (0.3/-0.1 V vs. Ag/AgCl/3 M KCl) a tremendous<br />
acceleration of the immobilization kinetics is observed as compared to SAM formation at OCP.<br />
The monitored capacitance reaches a plateau after about 90 min implying that a fully SAM<br />
covered electrode surface was obtained (Figure 3.34, blue curve). Increasing the potential<br />
difference to 700 mV (0.5/-0.2 V pulse profile) an even faster immobilization kinetics is<br />
achieved, and the capacitance plateau is reached within minutes (Figure 3.34, orange curve).<br />
With a higher potential intensity within the applied pulse profile, the potential drop in the<br />
vicinity of the electrode is faster generating a higher concentration gradient and thus a more<br />
efficient ion stirring. This obviously results in a much faster immobilization kinetics of the<br />
alkylthiol molecules. However, the efficiency of an applied pulse profile does not only depend<br />
on the selected potential values relative to the pzc, but also on the pulse duration and the<br />
migration of ions in solution as it was above described for the potential-assisted immobilization<br />
of DNA. If the potential pulse is too short to allow the formation of a sufficient concentration<br />
gradient, increasing of potential intensities while keeping the pulse time constant results in a<br />
slower SAM immobilization kinetics. This is shown in Figure 3.34 (yellow curve) for the case<br />
when increasing the potential difference to 900 mV (0.5/-0.4 V pulse profile) while keeping the<br />
same pulse time of 10 ms.<br />
In contrast to DNA immobilization, the evaluated alkanethiols are uncharged molecules, the<br />
principle of potential pulse-assisted SAM formation is similar to the potential pulse-assisted<br />
DNA immobilization. Namely, during a single potential pulse, a certain potential drop is<br />
generated in the vicinity of the electrode surface. By switching to a potential with opposite sign<br />
with respect to the pzc, the excess of ions in front of the electrode moves away from the surface<br />
and exchanges with ions of opposite charge. Fast switching between positive and negative<br />
potentials creates an ion stirring in the vicinity of the surface, moving along thiol molecules<br />
that are within this layer. If a potential pulse is sufficiently long, a whole thiol molecule can be<br />
pulled to the surface during a single pulse regardless of its orientation, facilitating the formation<br />
of the Au-S bond. Thus, potential pulse-assisted thiol immobilization, like in the case of charged<br />
molecules, is driven by the migration of ions in the vicinity of the electrode rather than the<br />
diffusion of thiols, what is the case during of self-assembly at OCP 94 .<br />
3.3 Importance of controlling the surface 72
_____________________________________________________________ Results and Discussion<br />
Figure 3.34. Capacitance kinetics curves obtained for potential-pulse assisted SAM<br />
formation of MCU using different potential pulse intensities. Pulse time was kept constant<br />
at 10 ms for all experiments. All measurements were performed in 10 mM PB, 20 mM<br />
K2SO4 (30 % ethanol) containing 1 mM MCU. Figure adapted from ref. 89 .<br />
Figure 3.35. Capacitance kinetics curves obtained for potential-pulse assisted SAM<br />
formation of MCU using the 0.3/-0.1 V pulse profile while varying the pulse time. All<br />
measurements were performed as explained in Figure 3.34. Figure adapted from ref. 89 .<br />
3.3 Importance of controlling the surface 73
_____________________________________________________________ Results and Discussion<br />
The duration of a pulse has to allow the system to respond to the perturbation invoked by<br />
applying a certain pulse profile. In order to achieve the highest SAM formation efficiency, the<br />
duration of a given potential pulse should be long enough to allow for an appropriate<br />
concentration gradient to form and a whole thiol molecule to be brought to the electrode surface.<br />
By this the formation of the Au-S bond is achieved regardless of the orientation of a thiol<br />
molecule. However, the pulse time should be also short enough to allow for a high number of<br />
pulsing cycles. Figure 3.35 demonstrates the influence of the pulse duration on the SAM<br />
formation kinetics. For the investigated system (MCU, ΔE = 400 mV) 10 ms pulse duration<br />
manifests as the optimal pulse time. Prolonging the pulse time leads to a less efficient ion<br />
stirring and with this a slower immobilization kinetics (Figure 3.35, yellow and green curves).<br />
On the other hand, decreasing the pulse time prolongs the time necessary to obtain a compact<br />
monolayer (Figure 3.35, blue curve) even though the number of stirring cycles during the<br />
overall immobilization time is increased.<br />
In order to demonstrate the influence of the length of the investigated thiol derivative on the<br />
optimization of the potential-pulse profile for potential-assisted immobilization, Figure 3.36<br />
compares alkylthiols of three different lengths: MCH (6 carbon atoms), MCU (11 carbon atoms)<br />
and MCHD (mercaptohexadecanol, 16 carbon atoms). In all three cases, a more efficient ion<br />
stirring and faster SAM formation kinetics are achieved using a higher pulse potential<br />
difference, suggesting that the optimal potential difference does not depend on the molecule<br />
length as long as the chosen pulse duration is appropriate. In contrast, the optimal pulse duration<br />
depends on the molecule length, more precisely, on its diffusion coefficient. For the<br />
immobilization of short and intermediately long molecules (Figure 3.36, a and b) the 10 ms<br />
pulse duration is sufficiently long. Prolonging the pulse time leads to a lower number of pulse<br />
cycles per time and slower adsorption kinetics. The optimal potential-pulse assisted<br />
immobilization procedure for short and intermediate thiols employs the 0.5/-0.2 V pulse profile<br />
with 10 ms pulse duration. In case of longer thiols such as MCHD (Figure 3.36, c) 10 ms is too<br />
short to allow for the whole alkyl chain to be brought to the interface, which results in a slower<br />
adsorption kinetics. By prolonging the pulse duration to 10 s a much faster SAM formation rate<br />
is observed. Therefore, the optimal immobilization kinetics of longer thiols is achieved using<br />
the 0.5/-0.2 V pulse profile with 10 s pulse duration. At these conditions the capacitance reaches<br />
a plateau within only 5 min.<br />
3.3 Importance of controlling the surface 74
_____________________________________________________________ Results and Discussion<br />
Figure 3.36. Optimization of the potential-pulse assisted SAM formation kinetics for three<br />
alkylthiol derivatives of different length: a) MCH, b) MCU, and c) MCHD. All<br />
measurements were performed as explained in Figure 3.34. Figure adapted from ref. 89 .<br />
According to the Helmholtz model of the double layer the interfacial capacitance of a SAMmodified<br />
electrode is inversely proportional to the thickness of the SAM, that is, the length of<br />
the alkyl chain of the thiol derivative 43 :<br />
C = εε 0<br />
δ<br />
(3.8)<br />
3.3 Importance of controlling the surface 75
_____________________________________________________________ Results and Discussion<br />
where ε has the value 2 for compact chains and δ is the thickness increment (0.11 nm per carbon<br />
atom for a 30° tilted alkyl chain 92 ). This supports the results from Figure 3.37, in which it is<br />
shown that the capacitance of the MCH SAM is the highest and the one of a MCHD SAM the<br />
lowest. Furthermore, it was demonstrated previously 95 that for a low number of carbon atoms<br />
(less than 8) the relation between thiols of different length may not be linear, which is seen also<br />
in Figure 3.37. Considering the relationship derived from the Helmholtz model, capacitance<br />
values of 2.7, 1.5 and 1.0 µF/cm 2 are expected for SAMs of MCH, MCU and MCHD,<br />
respectively. The experimentally determined capacitance values are smaller, however, they are<br />
in the same order of magnitude.<br />
Figure 3.37. Capacitance values of fully covered thiol monolayers in dependence on the<br />
inverse number of carbon atoms of the investigated thiol derivatives. Data obtained from<br />
Figure 3.36. Figure adapted from ref. 89 .<br />
From the capacitance curves it is possible to calculate the coverage kinetics of the investigated<br />
alkylthiols using the following equation:<br />
θ = (C − C 0)<br />
(C f − C 0 )<br />
(3.9)<br />
where C0 is the initial capacitance of the bare electrode, Cf is the capacitance of a fully covered<br />
monolayer and C is the capacitance at any time. The initial capacitance was calculated to be<br />
6.77 µF/cm 2 as determined in background electrolyte in absence of any thiol. Calculated<br />
coverage curves are presented in Figure 3.38, where it is shown that using the optimal potential-<br />
3.3 Importance of controlling the surface 76
_____________________________________________________________ Results and Discussion<br />
pulse assisted immobilization procedures, tailored for the specific molecule length, compact<br />
SAMs can be obtained within minutes.<br />
Figure 3.38. Coverage curves calculated from optimal capacitance curves for all three<br />
thiol derivatives: MCH 0.5/-0.2 V (10 ms pulse duration), MCU 0.5/-0.2 V (10 ms pulse<br />
duration), MCHD 0.5/-0.2 V (10 s pulse duration) according to Equation (3.9). Figure<br />
adapted from ref. 89 .<br />
It was previously reported that by applying constant potentials an acceleration of the SAM<br />
formation kinetics of aliphatic thiols (C12-C14) can be achieved 21,92 . For example, Ma and<br />
Lennox reported to obtain a complete chemisorption in only 15 min upon applying a constant<br />
potential 13 . They assessed the SAM integrity by evaluating the blocking ability towards electron<br />
transfer of a redox mediator and observed a full blocking of electron transfer for SAM formation<br />
supported by applying constant potentials from 0.2 V up to even 1 V. The obtained result is<br />
surprising since it is known that relatively high anodic potentials of above 600 mV induce Au-<br />
S bond cleavage 19 . Nevertheless, for the purpose of comparison, our potential-assisted SAM<br />
formation method was tested using this approach. A typical CV obtained by using the optimized<br />
procedure for potential-assisted SAM formation of MCU is presented in Figure 3.39, where it<br />
can be seen that a complete blocking of electron transfer is achieved already within seconds.<br />
3.3 Importance of controlling the surface 77
_____________________________________________________________ Results and Discussion<br />
Figure 3.39. CVs for the evaluation of blocking of the electrode surface. a) Before and<br />
after MCU SAM formation using the potential-pulse assisted method. b) Rescaled CVs<br />
obtained after different times of surface modification. 0.5/-0.2 V pulse profile with 10 ms<br />
pulse duration was used. Immobilization was performed in 10 mM PB, 20 mM K2SO4<br />
containing 1 mM MCU (30 % ethanol). CVs were performed in 10 mM PB, 20 mM K2SO4<br />
containing 5 mM of K3[Fe(CN)6] and K4[Fe(CN)6] at 100 mV/s scan rate.<br />
To evaluate how constant potentials influence the self-assembly of thiols with an –OH<br />
functional group in aqueous solutions, a control experiment was performed by applying a<br />
relatively high constant potential of 0.5 V (vs. Ag/AgCl/3 M KCl) during thiol chemisorption,<br />
that is still below the potential range of Au-S bond cleavage. As can be seen in Figure 3.40, a<br />
significant acceleration of the immobilization kinetics was observed as compared with the SAM<br />
formation at OCP. However, the SAM formation kinetics is still substantially slower than the<br />
observed kinetics obtained by the potential pulse-assisted immobilization method. A constant<br />
capacitance value is reached after 3 h (Figure 3.40, grey curve), while with our approach<br />
compact monolayers are obtained within only 5 min (Figure 3.40, orange curve). It should be<br />
noted that, in contrast to the vast majority of reported procedures for electrochemical-assisted<br />
thiol immobilization 13,21,22,92 , the proposed technique for SAM deposition is performed in<br />
aqueous solutions ensuring high compatibility with biomolecules used for further surface<br />
modification.<br />
3.3 Importance of controlling the surface 78
_____________________________________________________________ Results and Discussion<br />
Figure 3.40. Comparison of the formation kinetics of MCU SAMs performed at a constant<br />
potential equal to the previously measured OCP (0 V vs. Ag/AgCl/3 M KCl, black curve),<br />
a constant potential of 0.5 V (grey curve) and by using the potential-assisted method with<br />
a 0.5/-0.2 V potential-pulse profile with 10 ms pulse duration (orange curve).<br />
Measurements were performed in 10 mM PB with 20 mM K2SO4 containing 1 mM MCU<br />
(30 % ethanol). Inset: Immobilization kinetics of MCU at a potential equal to the OCP<br />
shown for 12 h. Figure adapted from ref. 89 .<br />
The obtained results for the kinetics of potential-assisted SAM formation suggest that the<br />
stirring process generated by pulsing enhances both, the approach of thiol molecules towards<br />
the electrified interface to allow formation of the Au-S bond and the packing and reorganization<br />
of adsorbed molecules. It is important to understand that this process is not likely driven by<br />
diffusion of thiols, but rather by migration of ions from the solution that carry along thiol<br />
molecules towards the surface. Furthermore, a Langmuir isotherm cannot be used to fit our<br />
experimental data, which supports the assumption that the process is not diffusion but rather<br />
migration driven. The developed potential pulse-assisted SAM formation method is promising<br />
for a fast and reproducible surface modification that may be implemented in diverse<br />
applications especially related to the production of point-of-care devices.<br />
3.3 Importance of controlling the surface 79
_____________________________________________________________ Results and Discussion<br />
Moreover, the fact that the potential-assisted immobilization method tremendously accelerates<br />
the immobilization kinetics of uncharged molecules as well, and not only of intrinsically<br />
charged molecules such as DNA, proves that simple DNA attraction/repulsion by the polarized<br />
surface is a far too simple model to explain the mechanism of DNA immobilization supported<br />
by potential application.<br />
3.3 Importance of controlling the surface 80
_____________________________________________________________ Results and Discussion<br />
3.3.3 Reproducible recycling of Au modified surfaces within seconds<br />
Parts of this section are in the manuscript that is in preparation: “Fully potential-controlled<br />
preparation of DNA chips” with coauthors: D. Jambrec, Y. U. Kayran and W. Schuhmann.<br />
Electrochemical SAM desorption has a wide range of applications including cleaning of gold<br />
electrodes to allow reusability of the surface (i.e. sensor recycling) 18 or modifying surface<br />
wettability and fabrication of mixed SAM layers 96,97 . The interest for electrochemical<br />
desorption of SAMs in this thesis is the development of a very fast and efficient desorption<br />
approach that allows cleaning of individual electrodes on a chip prior to their modification with<br />
appropriate DNA strands. Since DNA and alkylthiols that are used for surface modification<br />
have different anchoring molecules, where alkylthiols form only a single Au-S bond and the<br />
DNA forms six bonds Au-S from three disulfides within the anchoring site (see Section 3.4.1),<br />
we tested the efficiency of the developed potential pulse-assisted desorption approach on both<br />
types of molecules.<br />
Figure 3.41. Removal of DNA from the electrode surface (2 mm gold electrode). a) EIS<br />
and b) CV were used for the characterization of the surface at each step. DNA<br />
immobilization: incubation for 2 min in 1 µM DNA solution (10 mM PB, 450 mM K2SO4).<br />
Potential pulse-desorption: 0.9/-0.9 V vs. Ag/AgCl/3 M KCl, 10 ms pulse duration, 30 s.<br />
Measurements were performed in 10 mM PB, 20 mM K2SO4 containing 5 mM of<br />
K3[Fe(CN)6] and K4[Fe(CN)6]; CV at 100 mV/s scan rate.<br />
3.3 Importance of controlling the surface 81
_____________________________________________________________ Results and Discussion<br />
The typical approach for SAM desorption is the polarization of SAM-modified Au surfaces to<br />
potentials that are either too positive or too negative, leading to the cleavage of the Au-S bond,<br />
that is, oxidative or reductive desorption, respectively. However, the exact potential values<br />
needed in order to provoke SAM desorption are under discussion for decades and a substantial<br />
variety of values is reported. According to our knowledge, applying both positive and negative<br />
potentials within the same desorption process by pulsing between these potentials was not yet<br />
reported. Even though fairly fast desorption times are reported in the literature 18 (down to 30 s)<br />
the aim of the study was to further improve the efficiency of desorption.<br />
Applying very high potentials seems a reasonably simple approach to achieve SAM desorption,<br />
however, another important condition for the development of a desorption process is to leave<br />
the electrode surface undamaged after polarization, especially in the case of multi-electrode<br />
chips where the Au layer thickness is in the order of nm. Therefore, the potential pulse-assisted<br />
desorption approach was developed by exposing the electrode to very short potential pulses<br />
within the range of ms. The modified electrode was exposed to potential pulsing between<br />
potential values of 0.9 V and -0.9 V (vs. Ag/AgCl/3 M KCl) with a pulse time of 10 ms. The<br />
influence of the applied pulse profile on a DNA-modified electrode is presented in Figure 3.41.<br />
Prior to desorption, a bare electrode was modified with ssDNA by incubation for 2 min to mimic<br />
the conditions that would be observed on the chip. Namely, during potential-assisted DNA<br />
immobilization on a selected chip electrode, the rest of the electrodes would be exposed to the<br />
same solution and the DNA molecules would immobilize at OCP. In order to prevent later<br />
cross-talk between electrodes, these surfaces need to be cleaned prior to further modification<br />
with a desired probe DNA. Even though short DNA immobilization at OCP leads to a minor<br />
Rct increase (Figure 3.41, a) and consequently to a small DNA coverage, it can behave as an<br />
impurity significantly affecting the surface modification. Subsequent modification with a DNA<br />
probe of interest is then leading to a less efficient hybridization since the probe DNA coverage<br />
is higher than the optimal one. Even though the optimized immobilization procedure is<br />
employed the hybridization will be hindered due to the contribution from the leftover DNA.<br />
After the treatment of the DNA-modified electrode with the developed desorption method for<br />
30 s Rct decreases to the value of the bare electrode and the voltammogram obtains the same<br />
shape as for the bare electrode, showing that all DNA was efficiently removed from the surface.<br />
3.3 Importance of controlling the surface 82
_____________________________________________________________ Results and Discussion<br />
Figure 3.42. Recycling treatment of an Au electrode. Three cycles of surface modification<br />
with a MCU SAM and subsequent desorption are shown. a) bare electrode; b) after 1 min<br />
of potential-pulse assisted SAM formation (0.5/-0.2 V vs. Ag/AgCl/3 M KCl, 10 ms); c)<br />
after 30 s potential-pulse desorption (0.9/-0.9 V vs. Ag/AgCl/3 M KCl, 10 ms); d) after<br />
SAM formation; e) after desorption (30 s); f) after SAM formation; g) after desorption (5<br />
s). Measurements were performed in 10 mM PB, 20 mM K2SO4 containing 5 mM of<br />
K3[Fe(CN)6] and K4[Fe(CN)6] at 100 mV/s scan rate.<br />
Furthermore, to investigate whether the developed method can desorb also highly compact<br />
SAM layers, desorption of MCU SAMs was investigated (Figure 3.42 and Figure 3.43). A MCU<br />
SAM was formed by the potential-pulse assisted method (0.5/-0.2 V, 10 ms pulse duration, total<br />
time 1 min) leading to a full blocking of electron transfer from the free-diffusing redox mediator<br />
(Figure 3.42, a and b). Then, potential-pulse assisted desorption was performed for 30 s<br />
resulting in a complete regeneration of the gold surface (Figure 3.42, c), showing that the<br />
employed desorption procedure efficiently removes also compact thiol layers from the electrode<br />
surface. To verify whether the electrode surface remains undamaged after the used desorption<br />
procedure, immobilization/desorption cycles were repeated 3 times on the same electrode. The<br />
passivation of the surface by the MCU SAM is very reproducible and the surface becomes fully<br />
blocked (Figure 3.43, a). By using the developed desorption method the surface gets fully<br />
regenerated after each cycle (Figure 3.43, b), which proves that the employed desorption<br />
approach is harmless for the Au surface. Finally, the first two desorption cycles were performed<br />
3.3 Importance of controlling the surface 83
_____________________________________________________________ Results and Discussion<br />
for 30 s, while the last desorption cycle was performed for only 5 s, resulting in the same effect,<br />
namely the full regeneration of the surface. Shorter desorption times were not investigated.<br />
The potential pulse-assisted desorption allows regeneration of Au surfaces within only 5 s while<br />
preserving the surface quality for further modification, which makes this method suitable for<br />
the application in DNA microarrays.<br />
Figure 3.43. CV characterization during the removal of a MCU SAM from the electrode<br />
surface. Data was taken from Figure 3.42 and plotted to compare each cycle of a) SAM<br />
formation by the potential-pulse assisted immobilization method; b) surface<br />
characterization before and after regeneration by potential-pulse desorption.<br />
3.3 Importance of controlling the surface 84
_____________________________________________________________ Results and Discussion<br />
3.4 Potential-assisted preparation of DNA sensors<br />
Experiments for the optimization of the DNA sensing surface were done by Bianca Ciui under<br />
the supervision of the author. Experiments with multi-electrode chips were performed together<br />
with Yasin Uğur Kayran. Parts of the chapter are included in the manuscript that is in<br />
preparation: “Fully potential-controlled preparation of DNA chips” with coauthors: D.<br />
Jambrec, Y. U. Kayran and W. Schuhmann.<br />
3.4.1 Optimization of the potential-pulse assisted immobilization method<br />
Optimization of the potential-pulse assisted immobilization method to achieve highly<br />
reproducible and very fast immobilization of intrinsically charged (DNA) or uncharged<br />
molecules (thiols) provides a very efficient technique for preparation of DNA sensing<br />
platforms, for which the desired ssDNA coverage and blocking ability can be reached within<br />
minutes. Moreover, due to a very short modification time while maintaining a high level of<br />
reproducibility and the ability to electrochemically remove undesired layers from the electrode<br />
surface within seconds, the developed protocol has a potential for application in DNA array<br />
production.<br />
Figure 3.44. Scheme of the anchoring site used for immobilization of ssDNA.<br />
Nevertheless, implementation of the developed technique into a DNA sensor or a DNA array<br />
production requires further investigation. It is important to know whether the stability of a<br />
3.4 Potential-assisted preparation of DNA sensors 85
_____________________________________________________________ Results and Discussion<br />
formed DNA film is affected by the subsequent potential-pulse assisted passivation steps,<br />
whether the passivating thiol replaces the grafted DNA from the surface, and whether the<br />
application of potential pulses within a subsequent passivation step causes a desorption of<br />
previously grafted DNA molecules.<br />
Figure 3.45. Nyquist plots of ssDNA-modified electrode before and after pulsing in 10 mM<br />
PB (450 mM K2SO4) using the 0.5/-0.2 V (vs. Ag/AgCl/3 M KCl) pulse profile with 10 ms<br />
pulse duration. EIS measurements were performed as explained in Figure 3.7.<br />
To assure that the passivating thiol does not remove the immobilized DNA, a special anchoring<br />
molecule with six binding sites is used for the immobilization of ssDNA, where six Au-S bonds<br />
are formed per individual DNA strand (Figure 3.44). Since the thiol molecules form only one<br />
Au-S bond they cannot replace ssDNA that is much more stable 98 .<br />
The effect of the pulsing itself on the stability of the DNA film was investigated by exposing a<br />
ssDNA-modified electrode to potential pulsing performed in buffer solution alone (10 mM PB,<br />
450 mM K2SO4). The same potential profile (0.5/-0.2 V, 10 ms pulse duration) that is employed<br />
in the passivation step of short or intermediate length thiol was used. Figure 3.45 shows that<br />
upon pulsing for 1 min the DNA layer remains unaltered, i.e., no visible desorption occurs.<br />
After 2 min, a small decrease in Rct is observed suggesting that the DNA film starts slightly to<br />
desorb even though it cannot be excluded that the change in Rct may originate from<br />
3.4 Potential-assisted preparation of DNA sensors 86
_____________________________________________________________ Results and Discussion<br />
reorganization of the DNA at the surface. Pulsing for 5 min creates a much stronger effect on<br />
the DNA film inducing a significant decrease of Rct, which is most likely connected to DNA<br />
desorption.<br />
Figure 3.46. CV measurements made before and after surface modification with a) MCH<br />
and b) MCU and subsequent FSCV measurements made with c) MCH- and d) MCUmodified<br />
electrodes before and after incubation in Fc-tDNA solution. SAM formation was<br />
done using the potential-pulse assisted method (0.5/-0.2 V vs. Ag/AgCl/3 M KCl, 10 ms<br />
pulse duration) for 1 min in 1 mM thiol solution (10 mM PB, 20 mM K2SO4, 30 % ethanol).<br />
CV measurements were performed in 10 mM PB, 20 mM K2SO4 containing 5 mM of<br />
K3[Fe(CN)6] and K4[Fe(CN)6] at 100 mV/s scan rate. FSCV measurements were<br />
performed in 10 mM PB with 450 mM K2SO4 at 1 V/s scan rate.<br />
3.4 Potential-assisted preparation of DNA sensors 87
_____________________________________________________________ Results and Discussion<br />
Even though the employed pulse profile is within a safe potential window of the Au-S bond,<br />
fast pulsing still provokes a partial desorption of DNA molecules, albeit at a much slower rate<br />
than with the pulse profile employed for potential-pulse assisted desorption (0.9/-0.9 V, 10 ms).<br />
This raises the question about why the same pulse profile results in a significantly accelerated<br />
DNA and thiol immobilization. The reason is that the immobilization rate is apparently much<br />
higher than the desorption rate and it prevails, leading to a very fast immobilization kinetics.<br />
After DNA immobilization, the following passivation step should be performed within a time<br />
window that provokes a negligible desorption of the previously formed DNA layer, which is in<br />
this case 1-2 min. In order to find the optimal passivation procedure, potential-assisted<br />
immobilization of MCH and MCU was performed for a duration of 1 min and the obtained<br />
layers were compared with respect to their integrity using CV. Furthermore, the ability of<br />
formed layers to block the unspecific adsorption of tDNA was investigated by measuring the<br />
ferrocene signal from ferrocene labelled tDNA (Fc-tDNA) using fast scan cyclic voltammetry<br />
(FSCV) upon incubation of modified electrodes into the Fc-tDNA solution. Since the SAM<br />
formation of MCU is more efficient as compared with MCH in the investigated time, which is<br />
observed as a better blocking of the redox mediator (Figure 3.46, a) and absence of unspecific<br />
adsorption of the Fc-tDNA (Figure 3.46, b), MCU was chosen for the potential-pulse assisted<br />
passivation step in the build-up of the DNA assay. This finding is expected since longer<br />
alkylthiols produce a more ordered SAM with less defects 12,99 .<br />
A crucial step in designing a DNA sensor is the optimization of the ssDNA coverage, since it<br />
determines both the hybridization efficiency and its kinetics. To demonstrate the use of the<br />
potential-pulse assisted immobilization method for the preparation of a DNA sensor, the ssDNA<br />
coverage was optimized for hybridization detection based on the measurement of the ferrocene<br />
(Fc) signal from a Fc-tDNA. Both ssDNA immobilization and subsequent MCU passivation<br />
were performed using the potential-assisted method (0.5/-0.2 V, 10 ms pulse duration). The<br />
ssDNA immobilization duration was varied and passivation was kept at 1 min. Sensor<br />
preparation was characterized by CV, following each step of the build-up (an example is shown<br />
in Figure 3.47, a). The prepared sensors were subjected to hybridization with a fully<br />
complementary Fc-tDNA. FSCV was used for the detection of the ferrocene signal (Figure<br />
3.47, b) and the calculation of dsDNA coverage according to equations from Section 5.11.1.<br />
The obtained results are shown in Figure 3.48. When the ssDNA coverage on the surface is low<br />
that adjacent DNA strands do not interfere with each other neither sterically nor<br />
3.4 Potential-assisted preparation of DNA sensors 88
_____________________________________________________________ Results and Discussion<br />
electrostatically, a maximal hybridization efficiency of ideally 100 % is obtained providing a<br />
sufficient concentration of tDNA. Therefore, an increase of the ssDNA coverage leads to an<br />
increase in the hybridization yield and in the current signal obtained from labelled tDNA<br />
(observed for DNA immobilization up to 2 min). However, further increase of the ssDNA<br />
concentration causes a steric and electrostatic hindrance depending on the ionic strength<br />
towards the hybridization process and the efficiency of hybridization decreases leading to a<br />
lower signal. Therefore, for the employed detection scheme (hybridization with Fc-tDNA, 20-<br />
mer probe DNA) the optimal ssDNA coverage is around 2-3 × 10 12 molecules/cm 2 , achieved<br />
by immobilization of ssDNA for 2 min using the developed potential-assisted immobilization<br />
method.<br />
Figure 3.47. a) CV characterization of the surface during sensor preparation, and b)<br />
FSCV measurements before and after hybridization of the ssDNA/MCU-modified electrode<br />
with Fc-tDNA. ssDNA immobilization: 10 mM PB, 450 mM K2SO4, 1 µM DNA; p.a.<br />
immobilization (0.5/-0.2 V vs. Ag/AgCl/3 M KCl, 10 ms pulse duration), 2 min. MCU<br />
passivation: 10 mM PB, 20 mM K2SO4, 1 mM MCU (30 % ethanol), 0.5/-0.2 V vs.<br />
Ag/AgCl/3 M KCl (10 ms), 1 min. CV and FSCV measurements were performed as<br />
explained in Figure 3.46.<br />
This chapter demonstrates that using the developed potential-pulse assisted immobilization<br />
technique DNA sensor preparation can be done in as little as 3 min depending on the desired<br />
DNA coverage. Compared to the standard sensor preparation praxis where surface modification<br />
3.4 Potential-assisted preparation of DNA sensors 89
_____________________________________________________________ Results and Discussion<br />
usually occurs over a period of two days, the ability to reproducibly prepare DNA sensors<br />
within minutes presents a significant improvement.<br />
Figure 3.48. Dependence of dsDNA coverage on the total duration of ssDNA immobilization.<br />
After performing ssDNA immobilization (0.5/-0.2 V vs. Ag/AgCl/3 M KCl pulse<br />
profile, 10 ms pulse duration, 10 mM PB with 450 mM K2SO4 and 1 µM DNA) for a<br />
predefined time, all electrodes were passivated with MCU (1 min, 0.5/-0.2 V vs. Ag/AgCl/3<br />
M KCl pulse profile, 10 ms pulse duration, 10 mM PB with 20 mM K2SO4 and 1 mM<br />
MCU) and subjected to hybridization with Fc-tDNA (10 min incubation, 37 °C, 10 mM<br />
PB, 450 mM K2SO4, 1 µM DNA). FSCV:10 mM PB, 450 mM K2SO4, 1 V/s. Coverage<br />
determination as described in Section 5.9.3.<br />
3.4 Potential-assisted preparation of DNA sensors 90
_____________________________________________________________ Results and Discussion<br />
3.4.2 DNA microchip fabrication<br />
DNA chips are becoming tremendously important in diagnostics 2,3 . They allow for a<br />
simultaneous analysis of a large number of target molecules and have become a crucial tool in<br />
genomic analysis. The development of an electrochemically produced DNA microarrays may<br />
provide highly controlled surfaces complemented with low production costs and time.<br />
Therefore, the applicability of the developed potential-assisted immobilization technique for<br />
the production of DNA microarrays was investigated using a 32-electrode chip comprising of<br />
(70 × 70) µm 2 gold electrodes (Figure 3.49) by modifying the chip with two different DNA<br />
sequences.<br />
Figure 3.49. a-e) Schemes of the chip modification sequence: a) bare chip; b) after<br />
potential-assisted immobilization of columns (C) 1 and 2 with the FRIZ sequence; c)<br />
potential-assisted MCU passivation of C1 and C2; d) potential-assisted immobilization of<br />
C5 and C6 with an E. coli sequence performed after the potential-assisted cleaning; e)<br />
potential-assisted cleaning of C3 and C4 and subsequent potential-assisted MCU<br />
passivation of C3-C8; during each step the whole chip was exposed to each solution; f)<br />
scheme of the final chip modification with the region (a) representing ssFRIZ/MCUmodified<br />
electrodes, region (b) representing MCU-modified electrodes, region (c)<br />
representing E. coli/MCU-modified electrodes and region (d) representing ssFRIZ/E.<br />
coli/MCU-modified electrodes; picture of the chip with the zoomed picture of the 32<br />
electrodes (taken from ref. 100 ).<br />
3.4 Potential-assisted preparation of DNA sensors 91
_____________________________________________________________ Results and Discussion<br />
DNA chips were prepared employing potential-assisted immobilization and desorption methods<br />
by initially immobilizing a DNA sequence 1 (20-mer, “FRIZ sequence” in the further text) on<br />
the first two columns of the chip for 30 s with a subsequent passivation with MCU for 1 min<br />
(Figure 3.49, a-c). Since the whole chip was immersed in the same solution the other electrodes<br />
were subjected to immobilization at OCP. In order to prevent any undesired signal on these<br />
columns, prior to the immobilization with the sequence 2 (42-mer, “E. coli sequence”) columns<br />
5 and 6 were cleaned by the potential-assisted desorption for 5 s. Afterwards, they were<br />
modified with an E. coli sequence for 30 s (Figure 3.49, d). Subsequently, columns 3 and 4<br />
were cleaned with the potential-assisted desorption (5 s) and finally columns 3-8 were<br />
passivated with MCU (Figure 3.49, e). It should be noted that the aim of the experiment was<br />
not the optimization of the assay parameters but rather a proof of concept. Therefore, the<br />
employed immobilization and passivation times are not optimized. Independently it was<br />
observed that an immobilization time of only 5 s and a passivation time of 3 s are sufficient to<br />
achieve optimal DNA coverage for the employed system and the desired passivation of the<br />
electrode surface, respectively (data not shown). Compared to macro electrodes, the optimal<br />
modification time is much shorter, which can be explained by the improved diffusion profile of<br />
microelectrodes.<br />
By designing the experiment in the explained manner the obtained DNA chip is supposed to<br />
consist of 4 different regions: region a modified with FRIZ/MCU, region b modified only with<br />
MCU, region c modified with E. coli/MCU and region d where both FRIZ and E. coli<br />
sequences are immobilized by incubation at OCP and a subsequent passivation is done by the<br />
potential-assisted method (Figure 3.49, f). The whole chip was exposed to each solution leaving<br />
the possibility of contamination between different regions. Therefore, to verify the quality of<br />
preparation, the chip was subjected to hybridization using both target DNA sequences, FRIZ<br />
and E. coli tDNA, labelled with ferrocene. The results obtained by FSCV are shown in Figure<br />
3.50. In general, it can be observed that each region that was exposed to the same conditions<br />
shows reproducible results among individual electrodes (Figure 3.50, a).<br />
Figure 3.50, b and c compare representative electrodes from each region upon hybridization<br />
with FRIZ and E. coli tDNA sequences, respectively. In order to understand what these figures<br />
show, each region is discussed for both cases. Region a is the only region where FRIZ ssDNA<br />
immobilization was performed via the potential-assisted method and therefore should show the<br />
hybridization signal from Fc-labelled FRIZ-tDNA. Indeed, this is observed on each electrode<br />
3.4 Potential-assisted preparation of DNA sensors 92
_____________________________________________________________ Results and Discussion<br />
Figure 3.50. a) Response after hybridization of a complete chip with FRIZ (orange curve)<br />
or E. coli sequences (green curve) and baseline corrected FSCVs presenting cathodic<br />
peaks after hybridization with b) FRIZ and c) E. coli sequence.<br />
from region a (Figure 3.50, b, orange curve). The obtained signal is not very high since the<br />
amount of hybridized DNA is in the order of 10 11 molecules/cm 2 . The reason for this is that the<br />
3.4 Potential-assisted preparation of DNA sensors 93
_____________________________________________________________ Results and Discussion<br />
immobilization time employed in the experiment (30 s) is not optimal and already leads to a too<br />
high ssDNA coverage for the employed detection scheme resulting in a decreased hybridization<br />
efficiency. By doing this the rest of the electrodes were longer exposed to immobilization by<br />
incubation allowing to investigate, how a possible contamination influences the quality of the<br />
chip preparation. After potential-assisted immobilization of FRIZ ssDNA and passivation with<br />
MCU, region a was also exposed to E. coli ssDNA. There was no signal upon hybridization<br />
with E. coli tDNA showing the high integrity of the formed FRIZ/MCU films.<br />
Region b was exposed to FRIZ ssDNA solution, and subsequently to the MCU passivating<br />
solution and finally to the E. coli ssDNA solution, with immobilization occurring by incubation<br />
in all cases. Prior to the final potential-assisted passivation with MCU, the electrodes were<br />
cleaned using the potential-assisted desorption method. Thus, this region should be modified<br />
only with MCU. Figure 3.50, b and c confirm this by showing the absence of any signal from<br />
FRIZ and E. coli tDNA (grey curve). In order to prove that the absence of signals is due to a<br />
successful desorption and not due to an undetectable amount of immobilized DNA, region d<br />
was exposed to same conditions, without performing any desorption step. That region should<br />
be modified with both FRIZ and E. coli ssDNA sequences and passivated completely with<br />
MCU. In both cases (Figure 3.50, b and c, beige curve) a very small amount of tDNA is detected<br />
upon hybridization with both FRIZ and E. coli tDNA, however, this signal is negligible as<br />
compared to the signal obtained in the regions with potential-assisted immobilization.<br />
Nevertheless, the amount of contaminating DNA is measurable, proving that desorption in<br />
region b was done successfully.<br />
Finally, region c was initially exposed to FRIZ ssDNA solution at OCP, after which the<br />
electrodes were cleaned by potential-assisted desorption and subsequently immobilized with E.<br />
coli ssDNA and passivated with MCU by the potential-assisted method. Therefore, it is the only<br />
region that should show a signal upon hybridization with the E. coli tDNA sequence (Figure<br />
3.50, c, green curve). The absence of any peak upon hybridization with FRIZ tDNA (Figure<br />
3.50, b, green curve) also indicates that desorption was performed successfully in this region as<br />
well.<br />
Potential-assisted methods for immobilization and desorption allow for a fully<br />
electrochemically controlled preparation of DNA chips within a very short time. The big<br />
advantage of the potential-assisted immobilization method is that it provides the desired DNA<br />
coverage and blocking of the surface tremendously faster than the process at OCP and by this<br />
3.4 Potential-assisted preparation of DNA sensors 94
_____________________________________________________________ Results and Discussion<br />
it does not allow undesired contamination, which guarantees a high quality of the prepared<br />
DNA chips. Additionally, the construction of DNA chips in this way allows for the use of even<br />
smaller electrodes. The size of the electrode does not need to be limited by the size of the droplet<br />
used for modification or its evaporation rate since there is no need for covering individual<br />
electrodes for their modification with the solution of interest.<br />
3.4 Potential-assisted preparation of DNA sensors 95
_____________________________________________________________ Results and Discussion<br />
3.5 Intercalation as a DNA detection technique<br />
Synthesis of the intercalator was done by Dr. Adrian Ruff. Results from this chapter are<br />
included in the manuscript to be submitted: “Amperometric detection of dsDNA via acrydinium<br />
orange modified glucose oxidase” with coauthors: D. Jambrec, A. Ruff, W. Schuhmann, written<br />
by A. Ruff (synthesis part) and the author.<br />
In order to implement electrochemical sensing platforms into point-of-care diagnostic devices<br />
plenty of improvement still needs to be done. Besides the design of the sensing platform, the<br />
applied strategy for the sensing process itself is of tremendous importance defining the<br />
sensitivity of the chosen approach. Even though covalent labelling of target DNA using dyes<br />
or enzymes significantly enhances the sensitivity of DNA detection schemes, it demands<br />
sample preparation and lacks of simplicity 101 . In contrast, label-free electrochemical detection<br />
methods provide several advantages such as miniaturization, fabrication of low cost devices,<br />
operation simplicity and rapid detection time. Therefore, they are of particular interest in<br />
personal diagnostics using point-of-care devices 102 .<br />
Among label-free approaches, sandwich-type assays, in which a third labelled signaling<br />
sequence hybridizes the overhang of the target DNA sequence, have been extensively<br />
investigated. Nevertheless, this strategy requires the use of different sequences for each<br />
investigated target DNA, which increases the complexity and cost of the assay. Furthermore,<br />
3.5 Intercalation as a DNA detection technique 96
_____________________________________________________________ Results and Discussion<br />
EIS was employed for label-free detection of DNA hybridization. Even though this technique<br />
is very sensitive to surface modification, it is not yet sensitive enough to achieve the desired<br />
sensitivity. On the other hand, the use of compounds that are able to intercalate into the DNA<br />
helix is very promising. One of the main advantages of intercalators is that these molecules can<br />
be used universally with any given DNA sequence. This significantly decreases the complexity<br />
of the system and allows for the application in multiple probe DNA chips. Therefore, this<br />
chapter focuses on the investigation of a new intercalation compound, an acridine orange based<br />
intercalator (AO) conjugated with glucose oxidase (GOx) from Aspergillus niger.<br />
Synthesis of a ferrocenyl labelled intercalation compound for the detection of double stranded<br />
DNA (dsDNA) was reported recently 102 . However, this strategy is limited by low currents in<br />
case of low dsDNA coverages due to the oxidation of the 1e - redox-species, i.e., the ferrocene<br />
derivative. On the other hand, redox-enzyme conjugated intercalation compounds offer much<br />
higher efficiency at low dsDNA coverages due to the intrinsic catalytic amplification by<br />
catalytic redox-conversion of a specific substrate and thus the continuous recycling of a redox<br />
mediator/redox probe by the enzyme. DNA hybridization detection schemes based on redoxenzymes<br />
modified with an intercalation compound were previously reported 103,104 in which the<br />
authors used the well-known biotin-streptavidin chemistry. By constructing an intercalator that<br />
is covalently bound to an enzyme, one assay preparation step is avoided as compared to<br />
previously described intercalator-enzyme conjugates 103,104 ) in the DNA detection scheme.<br />
Figure 3.51. Scheme of the synthesized acridine orange-glucose oxidase intercalating<br />
compound.<br />
Since acridine orange is a well-established intercalator selective towards dsDNA 105,106 it was<br />
chosen as a basis to synthesize an intercalator-enzyme conjugate. Glucose oxidase is a widely<br />
3.5 Intercalation as a DNA detection technique 97
_____________________________________________________________ Results and Discussion<br />
used enzyme in amperometric biosensors. In the presence of glucose and a suitable redox<br />
mediator it provides enhanced catalytic currents due to the continuous recycling of the redox<br />
probe 107 . The modification of GOx is readily achieved by reductive amination of the<br />
deglycosylated enzyme. Therefore, the acridine orange-glucose oxidase conjugate (AO-GOx)<br />
was synthesized by binding the acridine orange moiety via a flexible 14-atom-long tether to<br />
deglycosylated GOx via reductive amination using NaBH4 (Figure 3.51).<br />
To assure that the enzyme remains active towards the oxidation of glucose even after the harsh<br />
conditions employed during synthesis, electrochemical characterization of AO-GOx was<br />
performed. CVs were recorded in PB containing ferrocene methanol as redox mediator and<br />
glucose as substrate for GOx before and after addition of AO-GOx. The observed catalytic<br />
current upon addition of the intercalating compound confirms that the enzyme remained intact<br />
and that it can still transfer electrons to the electrode via the redox mediator (Figure 3.52).<br />
Figure 3.52. Electrochemical characterization of the AO-GOx intercalating compound.<br />
CVs recorded at a bare gold electrode in a solution of 10 mM PB, 450 mM K2SO4<br />
containing 1 mM ferrocene methanol and 100 mM glucose without AO-GOx (grey line)<br />
and with increasing concentrations of AO-GOx (black lines). Measurements were made<br />
with a scan rate of 10 mV/s in a potential window of -0.1 to +0.45 V vs. Ag/AgCl/3 M KCl.<br />
3.5 Intercalation as a DNA detection technique 98
_____________________________________________________________ Results and Discussion<br />
Figure 3.53. Scheme of the characterization sequence: a) ssDNA immobilization (1 min,<br />
0.5/-0.2 V pulse profile with 10 ms pulse duration); b) passivation with MCU (1 min, 0.5/-<br />
0.2 V pulse profile with 10 ms pulse duration); c) hybridization with Fc-tDNA (10 min<br />
incubation); d) dehybridization (H2O, 10 min); e) hybridization with a non-labelled tDNA<br />
(10 min incubation); f) intercalation with AO-GOx (15 min, incubation) and detection of<br />
the glucose oxidation current; g) dehybridization and removal of AO-GOx (H2O, ethanol,<br />
~10 min); h) evaluation of the interaction of AO-GOx with ssDNA (15 min, incubation)<br />
by amperometric detection. Chronoamperometric measurements were conducted at an<br />
applied potential of +400 mV. FAD: flavin adenine dinucleotide. All potentials vs.<br />
Ag/AgCl/3 M KCl; drawing not to scale.<br />
In order to optimize DNA detection via the synthesized AO-GOx a thorough characterization<br />
sequence was employed, which is presented in Figure 3.53. The DNA assay build-up consisted<br />
of an initial ssDNA immobilization (Figure 3.53, a) followed by passivation of the surface with<br />
MCU (Figure 3.53, b). The created DNA sensing platform was then subjected to hybridization<br />
3.5 Intercalation as a DNA detection technique 99
_____________________________________________________________ Results and Discussion<br />
with Fc-tDNA to indirectly determine the ssDNA coverage which is possible due to a low<br />
ssDNA coverage, Figure 3.53, c. After subsequent dehybridization (Figure 3.53, d), free ssDNA<br />
was again hybridized with a fully complementary non-labelled tDNA sequence (Figure 3.53, e)<br />
and subsequent intercalation of AO-GOx was performed (Figure 3.53, f) followed by chronoamperometric<br />
measurement. Afterwards, tDNA and AO-GOx were removed from the electrode<br />
by H2O and ethanol (Figure 3.53, g), the remaining ssDNA/MCU-modified surface was again<br />
immersed into the AO-GOx solution (Figure 3.53, h) and the signal of the negative control was<br />
measured by subsequent amperometric experiments.<br />
Control of the ssDNA coverage and minimization of unspecific adsorption was achieved by<br />
careful surface preparation based on the developed potential-assisted surface modification<br />
method. The envisaged pDNA coverage is lower than in the case of a direct detection of<br />
hybridization (e.g., via detection of Fc-tDNA). Due to the large dimension of the enzyme<br />
(diameter of the native enzyme is around 7 nm 108,109 ) and with this the entire intercalation<br />
compound, ssDNA coverage must be low to prevent steric hindrance of neighboring DNA<br />
strands towards intercalation. Shortening the immobilization time to obtain ssDNA coverage in<br />
the order of 10 11 molecules/cm 2 provides enough space for the intercalating compound to<br />
interact with dsDNA.<br />
After the immobilization of ssDNA the electrode was passivated with MCU by applying the<br />
potential-assisted method. One of the main challenges in DNA detection schemes involving<br />
intercalators is to reduce the signal of the negative control, i.e. the interaction of the intercalator<br />
with the ssDNA/thiol-modified surface. This interaction can occur either with the ssDNA or<br />
from unspecific adsorption of the intercalator on the electrode surface. Since it is reported that<br />
acridine orange solely intercalates into dsDNA 105,106 interaction with the ssDNA is expected to<br />
be minimal. Thus, it is important to provide well-passivated surfaces that prevent significant<br />
unspecific adsorption of the intercalator to allow for a high signal-to-noise ratio. DNA detection<br />
via GOx amplification is based on the re-oxidation of the redox mediator at the electrode<br />
surface. Therefore, the electron transfer between the mediator and the electrode needs to be<br />
allowed. Previously it was reported that MCU provides the best compromise between electrode<br />
passivation and permeability 100 . Applying the developed approach for potential-assisted<br />
passivation of the surface with MCU for 1 min it was possible to clearly differentiate between<br />
dsDNA and ssDNA-modified electrodes and obtain a high signal-to-noise ratio.<br />
3.5 Intercalation as a DNA detection technique 100
_____________________________________________________________ Results and Discussion<br />
In order to perform the experimental sequence as explained in Figure 3.53, the possibility of<br />
removal of the intercalating compound was tested using the ssDNA/MCU-modified electrode.<br />
Upon incubation with AO-GOx, a chronoamperometric measurement was performed in<br />
phosphate buffer containing the redox mediator (ferricyanide, 1 mM) by applying a potential<br />
of +400 mV (vs. Ag/AgCl/3 M KCl). A small interaction of the intercalator with the<br />
ssDNA/MCU modified electrode was detected, shown in Figure 3.54 as a catalytic current upon<br />
addition of glucose. Afterwards the electrode was treated with ethanol and water to remove<br />
unspecifically bound intercalator and the chronoamperometric measurement was repeated. The<br />
absence of any catalytic current upon addition of glucose in this case indicates that the<br />
intercalator was completely removed from the surface.<br />
Figure 3.54. Chronoamperometry with a ssDNA/MCU-modified electrode a) after<br />
incubation with AO-GOx for 15 min, and b) after removal of AO-GOx. A potential of<br />
+400 mV (vs. Ag/AgCl/3 M KCl) was applied; 10 mM PB containing 450 mM K2SO4 and<br />
1 mM ferricyanide. Glucose was injected as shown on the figure.<br />
Furthermore, to investigate how the intercalation process influences the stability of the formed<br />
double helix, a control experiment was performed by following each step of the assay by means<br />
of FSCV. Namely, initial hybridization of the ssDNA/MCU electrode was performed using FctDNA<br />
and subsequent intercalation was done by leaving the labelled tDNA on the surface.<br />
FSCV does not show any significant change in the response after intercalation, except of a small<br />
shift of the oxidation peak. The DNA coverage remains unchanged suggesting that the<br />
intercalation process does not impede the stability of the double helix (Figure 3.55).<br />
3.5 Intercalation as a DNA detection technique 101
_____________________________________________________________ Results and Discussion<br />
Figure 3.55. FSCV recorded for a ssDNA/MCU (grey line) and dsDNA/MCU-modified<br />
electrode before (black line) and after intercalation with AO-GOx into the dsDNA (green<br />
line). FSCV measurements were conducted in 10 mM PB solution containing 450 mM<br />
K2SO4 with a 1 V/s scan rate. DNA immobilization and subsequent passivation were<br />
performed by potential-assisted immobilization (0.5/-0.2 V vs. Ag/AgCl/3 M KCl pulse<br />
profile with 10 ms pulse duration), each for 1 min. Immobilization was performed in 1<br />
µM DNA solution in 10 mM PB with 450 mM K2SO4. Passivation was performed with 1<br />
mM MCU in 10 mM PB with 20 mM K2SO4 (30 % ethanol). Hybridization was done by<br />
incubation for 10 min in 1 µM Fc-tDNA solution in 10 mM PB with 450 mM K2SO4 at 37<br />
°C. Intercalation was performed by drop coating of an aliquot of an AO-GOx solution (in<br />
100 mM PB) on the dsDNA/MCU-modified electrode (incubation time: 15 min).<br />
Figure 3.56 shows the possibility to clearly differentiate between dsDNA and ssDNA-modified<br />
electrodes via AO-GOx intercalation using the developed DNA detection scheme based on<br />
potential-assisted surface modification to control the ssDNA coverage and the passivation of<br />
the surface. Normalized steady state currents with respect to dsDNA-modified electrodes are<br />
extracted from I-t curves after the addition of glucose (40 mM) measured with ssDNA and<br />
dsDNA-modified electrodes. After removal of tDNA and AO-GOx the electrodes were again<br />
incubated in the AO-GOx solution and the amperometric measurement was repeated in order<br />
to investigate the interaction of the AO-GOx with ssDNA (negative control). Higher currents<br />
3.5 Intercalation as a DNA detection technique 102
_____________________________________________________________ Results and Discussion<br />
can be detected for dsDNA-modified electrodes (right column) than for the negative control<br />
(left column). This shows that the 14-atom-long and flexible tether between the AO-moiety and<br />
the enzyme allows for an at least partial intercalation of AO into dsDNA. AO-GOx is able to<br />
intercalate into dsDNA and upon addition of glucose, the enzyme catalyzes the oxidation of<br />
glucose to gluconolactone. The reduced enzyme transfers the electrons via a ferricyanide to the<br />
electrode surface. The obtained catalytic current proves the presence of AO-GOx and with this<br />
DNA hybridization.<br />
Figure 3.56. Background corrected and normalized currents measured with ssDNA/MCU<br />
(left) and dsDNA/MCU (right) electrodes exposed to AO-GOx solution upon addition of<br />
glucose (40 mM) into the electrolyte solution. Chronoamperometric detection was<br />
performed in 10 mM PB containing 450 mM K2SO4 and 1 mM ferricyanide at an applied<br />
potential of +400 mV (vs. Ag/AgCl/3 M KCl). Preparation of electrodes was performed as<br />
explained in Figure 4.55. Error bars represent standard deviation between measurements<br />
(n = 3).<br />
It should be noted that the absolute currents vary for different electrodes, however, the ratio<br />
between the currents obtained for dsDNA and ssDNA-modified electrodes is rather constant<br />
(see error bar for the left column). Therefore, following the developed procedure, well defined<br />
DNA-modified surfaces are obtained that ensure a high signal-to-noise ratio. The conversion of<br />
glucose by the enzyme and the continuous regeneration of the redox probe ensure a high signal<br />
amplification. Thus, the novel sensing platform allows for the clear differentiation between dsand<br />
ssDNA even at low surface coverages.<br />
3.5 Intercalation as a DNA detection technique 103
4. Conclusions
______________________________________________________________________ Conclusions<br />
Self-assembly is a very powerful tool for surface modification, providing solutions for a large<br />
number of applications from water repellent car windshields to the inhibition of corrosion in<br />
industry. This thesis presents the development of a new strategy to achieve controlled selfassembly<br />
of charged and uncharged thiolated molecules on gold surfaces in a very fast manner.<br />
The main envisioned application is the development of a low-cost, yet powerful electrochemical<br />
technique for the production of DNA chips as point-of-care devices. The explored strategy<br />
consists of the potential-pulse assisted acceleration of the surface modification process by<br />
applying a pulse-type potential modulation, using carefully selected potential pulse intensities<br />
and an optimized pulse duration. Once optimized, this strategy increases significantly the<br />
immobilization kinetics in a reproducible way as compared to passive self-assembly or<br />
immobilization supported by the application of constant potentials.<br />
The quality and cleanliness of the material used for surface modification is highly important for<br />
the reproducibility and sensitivity of the envisaged sensing devices. Therefore, special attention<br />
was given to the preparation of the surface used for modification. This resulted in a surface<br />
preparation protocol defining certain criteria as a prerequisite to achieve a reproducible surface<br />
architecture, reflected by the reproducible roughness factor and signals obtained from the<br />
characterization via electrochemical impedance spectroscopy and cyclic voltammetry.<br />
In order to tailor optimal DNA sensing surfaces, it is important to understand the processes<br />
occurring at the electrode surface during the DNA assay build-up, taking into consideration the<br />
physico-chemical properties of the investigated molecules, electrode polarization and the<br />
surrounding solution. EIS was used to sequentially follow each step of the build-up of DNA<br />
assays and understand how the variation of different parameters alters the quality of the surface<br />
modification. The influence of surface modification by DNA on the value of the potential of<br />
zero charge of polycrystalline gold was investigated by determining the pzc of bare gold before<br />
and upon its modification with DNA. It was observed that the pzc shifts towards more negative<br />
values due to the surface modification with DNA. The importance of this shift was evident in<br />
the study for the selection of appropriate potential-pulse profiles to achieve high immobilization<br />
rates of both DNA and alkylthiol derivatives as representatives of intrinsically charged and<br />
uncharged molecules, respectively.<br />
Based on previous findings showing that in solutions of high ionic strength charge screening of<br />
DNA is significant and the potential profile in front of an electrode surface upon electrode<br />
polarization is fairly steep, it is evident that the widely accepted model explaining the influence<br />
105
______________________________________________________________________ Conclusions<br />
of the polarized electrode on the behavior of DNA in its proximity is far too simple.<br />
Attraction/repulsion of the DNA by the electric field in front of the polarized electrode can very<br />
unlikely be responsible for an improved immobilization rate due to the shortness of the Debye<br />
length. Thus, a new model is proposed, suggesting that the polarized electrode rather affects the<br />
ions in the vicinity of the electrified interface. When the electrode is polarized to negative values<br />
with respect to the pzc, cations move towards the electrified interface while anions move<br />
towards the bulk of the solution. In contrast, the opposite behavior is obtained when the<br />
electrode is polarized to more positive potentials with respect to the pzc. Switching fast enough<br />
between these two situations creates a “stirring effect” that effectively exceeds the Debye length<br />
in front of the electrified interface and pulls along DNA strands present in close proximity to<br />
the electrode surface including their condensed ion cloud. Hence, immobilization is not<br />
diffusion controlled but driven by the migration of ions in front of the electrode. Parameters<br />
that are playing a crucial role in achieving a significant improvement in the immobilization<br />
kinetics are the applied potential intensities and the duration of an individual pulse. The applied<br />
potential intensities need to be on the one hand within the stable potential window of the Au-S<br />
bond and on the other hand high enough to evoke an efficient stirring to bring the DNA towards<br />
the surface. The potential-pulse duration needs to be long enough to allow for an appropriate<br />
concentration gradient to form and for a whole molecule to be pulled down to the electrode<br />
surface, hence allowing for the formation of the Au-S bond regardless of the orientation of the<br />
molecule. On the other hand, it needs to be simultaneously short enough to allow a high number<br />
of potential pulse cycles.<br />
Using the developed strategy, alkylthiol SAM formation can also be significantly accelerated,<br />
leading to the formation of compact SAMs within minutes. It was shown that, besides the<br />
potential-pulse intensities and the pulse duration, the length of a molecule influences the<br />
efficiency of the immobilization. Optimization of the potential difference does not depend on<br />
the molecule length, as long as the appropriate pulse duration is chosen, since longer potential<br />
pulses are needed to bring down longer molecules to the electrode. The fact that this strategy<br />
tremendously accelerates the immobilization of uncharged molecules as well supports the<br />
conclusion that only DNA attraction/repulsion by the polarized electrode is a far too simple<br />
model to explain the mechanism of DNA immobilization supported by applied potentials.<br />
The developed strategy was subsequently implemented into the development of a DNA sensor,<br />
using the developed potential-pulse assisted method for both DNA immobilization and thiol<br />
106
______________________________________________________________________ Conclusions<br />
passivation steps to create a DNA sensing platform. The optimization of the protocol was<br />
performed by investigating the influence of subsequent passivation step and the potential<br />
pulsing itself on the stability of DNA-modified surfaces and by choosing the right thiol for<br />
passivation.<br />
Furthermore, the applicability of the developed method for the production of DNA chips was<br />
investigated using a 32-electrode array chip. Using this approach, the whole chip is exposed to<br />
each DNA solution used for the modification of a certain number of electrodes, avoiding the<br />
need for a localized positioning of DNA solutions on individual electrodes as in light-directed<br />
synthesis or spotting procedures that require expensive and sophisticated equipment. Due to the<br />
goal of making multiple probe DNA chips, it is clear that electrodes need to be cleaned prior to<br />
the modification with a selected DNA sequence. Thus, a method for potential-pulse assisted<br />
cleaning of Au-modified surfaces was developed, which allows a very fast and efficient<br />
regeneration of individual electrodes without causing any damage to the surface. The fact that<br />
by using our approach the Au surface can be easily regenerated without jeopardizing the quality<br />
of the following surface modification, points also to chip recycling as another advantage of the<br />
proposed approach. Until now research in this field did not address this topic, but relatively<br />
soon, in the world of private medicine, where everyone will own point-of-care devices, the<br />
question of reusability will become an important issue. Furthermore, construction of DNA chips<br />
in this manner opens the door towards the use of even smaller electrodes, since their size does<br />
not need to be limited by the size of the droplet used for modification or its evaporation rate.<br />
Finally, the developed technique was implemented into the development of a new DNA sensing<br />
platform based on signal amplification via an enzyme-conjugated intercalating compound<br />
(glucose oxidase-acridine orange) as hybridization indicator. Using the potential-pulse assisted<br />
surface modification strategy, DNA/thiol surfaces were fabricated to obtain lower probe DNA<br />
coverage and very efficient blocking of unspecific adsorption, which resulted in a significant<br />
contrast between ss- and dsDNA. The developed DNA sensing strategy finds its application in<br />
multiple probe DNA chips, since the synthesized intercalating compound can universally<br />
interact with all DNA sequences present on the chip and, by this, simultaneously increase the<br />
sensitivity on all electrodes.<br />
Even though this work is primarily orientated towards DNA chip development, the developed<br />
potential-pulse assisted surface modification method is an equally promising concept for<br />
applications in many other research fields, such as protein binding or investigation of cells.<br />
107
5. Experimental<br />
Work
________________________________________________________________ Experimental Work<br />
5.1 Materials and consumables<br />
The following materials were bought from Sigma-Aldrich Chemie (Germany): 6-mercapto-1-<br />
hexanol, 11-mercapto-1-undecanol, [Ru(NH3)6]Cl3, K3[Fe(CN)6] and K4[Fe(CN)6]. 16-<br />
mercapto-1-hexadecanol was from Frontiers Scientific (United States). KH2PO4, K2HPO4 and<br />
K2SO4 were from VWR International (Germany). Glucose was purchased from AppliChem<br />
(Germany) and was used one day after the preparation of the solution to allow for mutarotation.<br />
All reagents were of analytical grade and used as received. Polishing cloths were bought from<br />
Leco (USA) including polishing pastes with 3 µm, 1 µm and 0.5 µm particle size. 0.1 µm<br />
particle size polishing paste was from Struers (Germany). Non-modified (bare gold) 32-<br />
electrode chips were bought from FRIZ Biochem (Germany).<br />
Glucose oxidase from Aspergillus niger (type X-S, lyophilized powder, 100000–250000 U/g)<br />
was purchased from Sigma-Aldrich and stored at -22 °C. All DNA probes were purchased from<br />
FRIZ Biochem (Germany) and are listed in Table 5.1.<br />
Table 5.1. Summary of DNA sequences used in all experiments. Sequences marked with –<br />
SS were used as probe DNA, the rest of sequences were employed as target DNA. DTPA<br />
stands for dithiophosphoramidite. Fc stands for ferrocene. A – adenine, T – thymine, C –<br />
cytosine, G – guanine.<br />
Internal name<br />
Sequence<br />
FRIZ 12-SS<br />
5’ TGC GGA TAA CAC AGT CAC CT TTTTTTT (DTPA)3<br />
FRIZ 5 tDNA<br />
5’ AGG TGA CTG TGT TAT CCG CA<br />
FRIZ 5-Fc<br />
5’ (Fc)4 AGG TGA CTG TGT TAT CCG CA<br />
E. coli-SS<br />
E. coli-Fc<br />
5’ GTC AAT GAG CAA AGG TAT TAA CTT TAC TCC<br />
CTT CCT CCA TTA TTTTTTT (DTPA)3<br />
5’ (Fc)4 GGA GGA AGG GAG TAA AGT TAA TAC CTT<br />
TGC TCA TTG CG<br />
109
________________________________________________________________ Experimental Work<br />
Synthesis of the acridine orange-glucose oxidase intercalator was done by Dr. Adrian Ruff<br />
following procedures from Biver et al. 105 (for the modification of acridine orange) and<br />
Schuhmann et al. 110 (for deglycosylation of glucose oxidase).<br />
5.2 Electrochemical setup and instrumentation<br />
All measurements were done in an electrochemical cell consisting of a polycrystalline gold<br />
working electrode (2 mm diameter, CH Instruments, USA), a platinum auxiliary electrode<br />
(Goodfellow, Germany) and a homemade Ag/AgCl (3 M KCl) or Pb/PbF2 (5 M KF) reference<br />
electrode. Unless stated otherwise all electrochemical measurements were performed with an<br />
Autolab PGSTAT302N containing a frequency response analyzer (Metrohm-Autolab,<br />
Netherlands).<br />
Figure 5.1. a) CV representing the deposition of PbF2 (2 nd cycle) performed at 20 mV/s in<br />
0.5 M NaF solution. b) Picture of a home-made Pb/PbF2 reference electrode.<br />
Preparation of the lead-lead fluoride reference electrode was done according to 81 . A ceramic<br />
frit was introduced at the tip of a Pasteur pipette and the surrounding glass was melted to seal<br />
it and prevent later electrolyte leakage. Subsequently, the tip was polished to expose the frit and<br />
allow later ionic contact with the electrolyte solution. The pipette was filled with 5 M KF<br />
solution and a piece of cotton was placed in the pipette construction to prevent clogging of the<br />
electrode by detached PbF2. A lead wire was connected to a copper wire and the connection<br />
110
________________________________________________________________ Experimental Work<br />
was isolated with a shrink-tube. PbF2 was electrochemically deposited on the lead wire in a 0.5<br />
NaF solution by cyclic voltammetry (Figure 5.1, a). During the first scan the potential was<br />
initially decreased to negative values in order to remove the native oxide from the surface of<br />
the lead wire and by this improve the deposition of lead fluoride. The deposition was done for<br />
20 cycles at a scan rate of 20 mV/s. After deposition the electrode was kept in the same solution<br />
and chronoamperometry was performed for 2 h at 0.2 V vs. Ag/AgCl (3 M KCl) in order to<br />
improve the stability of the deposited PbF2 film. Finally, the lead wire with deposited PbF2 was<br />
inserted in the prepared pipette and fixed (Figure 5.1, b). The prepared electrode was usually<br />
left overnight to assure stabilization of the electrode potential. The electrode potential was<br />
controlled before use and it was around -600 mV vs. Ag/AgCl (3 M KCl).<br />
5.3 Preparation of gold surfaces<br />
Figure 5.2. Home-made polishing machine.<br />
Polycrystalline gold electrodes were mechanically and electrochemically cleaned before each<br />
experiment in the following manner. Electrodes were mechanically polished using a homemade<br />
automatic polishing machine built by Dr. Kirill Sliozberg (Figure 5.2). The polishing<br />
machine consisted of a rotating disk covered with an appropriate polishing cloth and a holder<br />
for two electrodes that was simultaneously moved in x direction keeping the electrodes<br />
perpendicular to the surface of the rotating disk. By this polishing was performed in all<br />
111
________________________________________________________________ Experimental Work<br />
directions equally, allowing the formation of homogenous surfaces. The procedure was<br />
controlled by a home-made software. Polishing clothes were soaked in water prior to polishing<br />
and covered homogeneously with an appropriate polishing paste. Electrodes were polished with<br />
diamond pastes of 3, 1, 0.5 and 0.1 µm particle size successively, using separate polishing cloths<br />
for each paste. Polishing time depended on the state of the electrode.<br />
Afterwards, the gold electrodes were electrochemically cleaned by cyclic voltammetry in 0.5<br />
M H2SO4 until reproducible voltammograms were obtained. Cycling was performed in the<br />
potential range of 0 V to 1.6 V vs. Ag/AgCl (3 M KCl) at 100 mV/s scan rate. The roughness<br />
factor was kept around 1.4 and its determination was done according to 111 . The method is based<br />
on the determination of oxygen adsorption on a gold surface. During a positive potential scan<br />
chemisorption of oxygen occurs, where different gold facets have different oxidation potentials<br />
(Figure 5.3). During the negative potential scan oxides are reduced, which can be seen as a<br />
single reduction peak. Assuming that oxygen atoms form a monolayer on the gold surface, the<br />
roughness factor of an electrode can be determined using the following equations:<br />
R = A real<br />
A g<br />
(5.1)<br />
A g = r 2 π (5.2)<br />
A real = Q exp<br />
Q theor<br />
(5.3)<br />
Q exp =<br />
peak area<br />
ν<br />
(5.4)<br />
where Ag is the geometrical surface area of the electrode (0.0314 cm 2 for a 2 mm diameter gold<br />
electrode), Areal is the real surface area, Qexp is the charge involved in the reduction of gold<br />
oxide, Qtheor is 482 μC/cm 2 and it is a calculated charge value for chemisorption of an oxygen<br />
monolayer on the surface of polycrystalline gold, based on the density and atomic weight of<br />
gold 112 , υ is the scan rate of the measurement and peak area is the area under the reduction peak<br />
from the CV measurement obtained by integration of the peak (Figure 5.3, area marked in<br />
green).<br />
After mechanical polishing and electrochemical cleaning, the electrodes were immediately<br />
characterized by EIS and CV to verify their cleanliness prior to modification (procedures<br />
explained in Sections 5.9.1 and 5.9.2, respectively).<br />
112
________________________________________________________________ Experimental Work<br />
Figure 5.3. Cyclic voltammogram (20 th cycle) of a bare polycrystalline gold electrode in<br />
0.5 M H2SO4 at 100 mV/s scan rate. The area under the reduction peak is marked in green.<br />
5.4 Determination of the potential of zero charge<br />
The determination of the potential of zero charge was done via potentiodynamic<br />
electrochemical impedance spectroscopy (PDEIS) in PB/K2SO4 solution of different ionic<br />
strength – 10 mM PB with 20 mM K2SO4, 1 mM PB with 2 mM K2SO4 and 0.1 mM PB with<br />
0.2 mM K2SO4. Prior to experiments, the solutions were purged with argon for at least half an<br />
hour and the electrodes were prepared as explained in Chapter 5.3. A capacitive bridge<br />
(capacitance of 2 µF) was used during measurements to avoid artifacts from the potentiostat<br />
(Figure 5.4). The electrochemical setup consisted of a big cylindrical Pt counter electrode<br />
surrounding the working electrode and a Pb/PbF2 reference electrode that was placed below the<br />
working electrode.<br />
Determination of the pzc of the bare electrode was done immediately after electrode preparation<br />
to avoid any contamination of the surface. PDEIS was done for a potential range of -0.2 V to<br />
0.7 V vs. Ag/AgCl (3 M KCl) with 30 mV potential steps. EIS was performed at each potential<br />
step for a different range of frequencies, depending on the ionic strength and each potential was<br />
superimposed with an AC perturbation of 10 mVpp amplitude.<br />
113
________________________________________________________________ Experimental Work<br />
Figure 5.4. Scheme of the electrochemical setup used for determination of the pzc. The<br />
implemented capacitive bridge (Cb) is shown in the figure. CE – counter electrode, RE –<br />
reference electrode, WE – working electrode.<br />
Determination of the pzc of the DNA modified surface was done by initially immobilizing<br />
ssDNA (from 1 µM solution in 10 mM PB with 450 mM K2SO4) on a clean electrode surface<br />
for 5 min via potential-assisted immobilization (0.5/-0.2 V (vs. Ag/AgCl/3 M KCl) following<br />
a pulse profile with 10 ms pulse duration (procedure explained in detail in Section 5.7.2).<br />
Afterwards the electrode was thoroughly rinsed with the immobilization buffer and water,<br />
respectively to remove any loosely bound DNA strands, and placed in an appropriate solution<br />
for pzc determination. PDEIS was performed for a potential range of -0.4 V to 0.5 V vs.<br />
Ag/AgCl (3 M KCl) with 30 mV potential steps. EIS measurements for discrete potential values<br />
were done as in the case of the bare electrode.<br />
5.5 Potential-assisted formation of self-assembled monolayers<br />
Preparation of gold electrodes for thiol self-assembling was done as explained in Chapter 5.3.<br />
After characterization with EIS and CV (see Sections 5.9.1 and 5.9.2) gold electrodes were<br />
immediately used for potential-assisted SAM formation of thiols of different length – MCH,<br />
MCU and MCHD (1 mM in phosphate buffer, 20 mM K2SO4 with 30 % ethanol), using several<br />
114
________________________________________________________________ Experimental Work<br />
pulse profiles: 0.3/-0.1 V, 0.5/-0.2 V and 0.5/-0.4 V (vs. Ag/AgCl/3 M KCl) with various pulse<br />
durations of 1 ms, 10 ms, 100 ms and 10 s.<br />
Real-time impedance measurements were performed during potential-assisted SAM formation<br />
to investigate the kinetics of self-assembly. This was done in a way that the applied pulse<br />
profiles were considered as DC potentials which were superimposed with an AC signal of 1<br />
kHz frequency and 5 mVrms amplitude. Experiments were performed using a setup that<br />
consisted of a potentiostat with a summing amplifier (IPS AJ, Germany), a function generator<br />
(Agilent 33120A, USA), a lock-in amplifier (EG&G Instruments 7265 DSP, USA) and a 16 bit<br />
CIO-DAS 1602/16 AD/DA board (Plug-In electronic, Germany).<br />
Determination of the capacitance was done by calculating initially the imaginary impedance<br />
component (-Z'') using the recorded magnitude (|Z|) and phase (θ) values:<br />
|Z| =<br />
A AC<br />
k|Z| meas<br />
(5.5)<br />
−Z ′′ = |Z|sinθ<br />
where AAC is AC amplitude, k is the correction coefficient taking into account the current range<br />
of the potentiostat (10 mA/10 V) and |Z|meas is the measured value of the magnitude. Using the<br />
value of the imaginary impedance component the capacitance was calculated using the<br />
following expression, derived from an RC series equivalent electrical circuit:<br />
C = −1<br />
ωZ ′′ (5.6)<br />
Data sampling was done at a rate of 100 ms per data point and recorded with a home-made<br />
software. Oscillations of the recorded signals coming from the applied pulse potential were<br />
suppressed by applying a digital filter to the experimental plots. Data analysis and curve fitting<br />
was performed with Origin.<br />
5.6 Potential-assisted desorption<br />
Potential-assisted desorption was performed using ssDNA-, MCU- or ssDNA/MCU-modified<br />
electrodes. The experiment was done in 10 mM PB with 450 mM K2SO4 by applying a 0.9/-0.9<br />
115
________________________________________________________________ Experimental Work<br />
V (vs. Ag/AgCl/3 M KCl) potential pulse profile with 10 ms pulse duration, for 5 or 30 s. Upon<br />
desorption, the electrode was rinsed with the same buffer and water.<br />
5.7 Preparation of DNA sensors<br />
5.7.1 ssDNA immobilization via incubation<br />
Immobilization of ssDNA via incubation was done immediately after cleaning and<br />
characterization of a gold electrode as explained in Chapter 5.3. The electrode was placed in an<br />
Eppendorf tube containing 200 μL of ssDNA solution (1 μM in 10 mM PB, 450 mM K2SO4)<br />
and kept in a thermo-mixer (HTC BioTech, Germany) at 37 ºC. The electrode was sealed within<br />
the tube using parafilm to prevent solution evaporation. The duration of the immobilization<br />
procedure depended on the desired DNA coverage. Afterwards, the electrode was thoroughly<br />
rinsed with the immobilization buffer (10 mM PB with 450 mM K2SO4) and water to remove<br />
any loosely bound ssDNA.<br />
5.7.2 Potential-assisted ssDNA immobilization<br />
Figure 5.5. Electrochemical setup used for potential-assisted sensor preparation.<br />
116
________________________________________________________________ Experimental Work<br />
Potential-assisted ssDNA immobilization was performed right after the electrode preparation,<br />
in an electrochemical setup using small volumes (100-200 μL) as shown in Figure 5.5.<br />
Measurements were performed in the same solution as the immobilization via incubation at<br />
room temperature using different potential pulse profiles with various pulse durations. The total<br />
duration of the immobilization process depended on the desired DNA coverage. In order to<br />
remove unspecifically bound DNA from the electrode surface, electrodes were rinsed after<br />
modification with the immobilization buffer.<br />
5.7.3 Passivation by means of incubation<br />
Passivation via incubation was done by placing the electrode in an Eppendorf tube containing<br />
500 μL of a thiol derivative solution (in 10 mM PB, 20 mM K2SO4). The electrode was sealed<br />
within the tube to prevent solution evaporation and kept in a thermo-mixer (HTC BioTech,<br />
Germany) at 37 ºC for 19 h unless stated otherwise. Afterwards, the electrode was thoroughly<br />
rinsed initially with absolute ethanol and then water to remove any loosely bound MCH.<br />
5.7.4 Potential-assisted passivation<br />
Potential-assisted passivation was performed using a 0.5/-0.2 V pulse profile with 10 ms pulse<br />
duration. It was done in the same electrochemical setup as used for potential-assisted<br />
immobilization (Figure 5.5). Passivation by MCH was done in a 10 mM solution with 10 mM<br />
PB and 20 mM K2SO4, while passivation by MCU was performed in a 1 mM solution with 10<br />
mM PB and 20 mM K2SO4. Measurements were performed at room temperature for 1 min<br />
unless specified otherwise. In order to remove loosely bound thiols from the electrode surface,<br />
the electrodes were rinsed with absolute ethanol and then water after modification.<br />
5.8 Preparation of DNA chips<br />
Prior to use, DNA chips were cleaned with piranha solution (cc. H2SO4 with 30 % H2O2, 3:1)<br />
for 10 min. After thoroughly rinsing with water, the chips were electrochemically cleaned in<br />
H2SO4 as explained in Chapter 5.3. Modification of chips was done using potential-assisted<br />
procedures for immobilization, passivation and desorption. The detailed experiment sequence<br />
is explained in Section 3.4.2.<br />
117
________________________________________________________________ Experimental Work<br />
5.9 Characterization of DNA sensors<br />
5.9.1 Electrochemical impedance spectroscopy<br />
EIS measurements were performed after each assay preparation step in 10 mM PB with 20 mM<br />
K2SO4 (pH 7.4) containing equimolar concentrations of K3[Fe(CN)6] and K4[Fe(CN)6] (5 mM<br />
each). Experiments were conducted at the equilibrium potential of the redox couple (DC<br />
potential, +220 mV vs. Ag/AgCl/3 M KCl) superimposed by an AC perturbation of 10 mVpp<br />
amplitude. The frequency range from 30 kHz to 10 mHz was scanned unless stated otherwise.<br />
The charge transfer resistance was determined by fitting the Nyquist plots to an [R(Q[RW])]<br />
Randles equivalent circuit.<br />
5.9.2 Cyclic voltammetry<br />
Characterization of the surface before and during the assay preparation was performed also via<br />
cyclic voltammetry in 10 mM PB with 20 mM K2SO4 (pH 7.4) containing equimolar<br />
concentrations of K3[Fe(CN)6] and K4[Fe(CN)6] (5 mM). Measurements were performed at 100<br />
mV/s scan rate in the potential window from -0.1 V to 0.5 V (vs. Ag/AgCl/3 M KCl).<br />
5.9.3 Chronocoulometry for determination of DNA coverage<br />
Determination of ssDNA coverage was done using the chronocoulometric method developed<br />
by Steel and coworkers 75 . A potential step from 0 V to -0.4 V (vs. Ag/AgCl/3 M KCl) was<br />
applied for 500 ms at ssDNA/MCH modified electrodes immersed in two different solutions,<br />
initially in 10 mM PB solution containing 20 mM K2SO4 and subsequently in the same buffer<br />
containing additionally 100 µM [Ru(NH3)6] 3+ . The resulting charge was measured in both cases.<br />
Prior to measurements solutions were purged with argon for at least 30 min.<br />
5.10 Hybridization and dehybridization<br />
DNA hybridization was performed via incubation with the target strand after characterization<br />
of ssDNA/thiol modified electrodes. The electrode was initially rinsed with the hybridization<br />
buffer (10 mM PB, 450 mM K2SO4) and then placed in an Eppendorf tube containing 200 μL<br />
118
________________________________________________________________ Experimental Work<br />
of the labeled or non-labeled target DNA solution (1 μM in 10 mM PB, 450 mM K2SO4) and<br />
kept in a thermo-mixer (HTC BioTech, Germany) at 37 ºC for 10 min or 1 h. The electrode was<br />
sealed within the tube to prevent solution evaporation. After hybridization the electrode was<br />
rinsed with the hybridization buffer to remove any unspecifically bound tDNA.<br />
Dehybridization of dsDNA was performed by incubation of the dsDNA/thiol electrode in water<br />
for 5-10 min. The efficiency of dehybridization was confirmed using Fc-tDNA by measuring<br />
the Fc redox peak before and after dehybridization using fast-scan cyclic voltammetry.<br />
5.10.1 Detection of hybridization<br />
Direct hybridization detection was done using different detection methods, depending whether<br />
the tDNA was labeled or not. Hybridization with non-labeled tDNA was detected by means of<br />
EIS following the change of Rct upon hybridization. EIS parameters were the same as in Section<br />
5.9.1.<br />
Hybridization with labeled tDNA was detected using FSCV in 10 mM PB with 450 mM K2SO4<br />
at a scan rate of 1 V/s in the potential window from -0.05 V to 0.55 V (vs. Ag/AgCl/3 M KCl)<br />
under argon atmosphere.<br />
5.10.2 DNA coverage determination by means of FSCV<br />
Direct determination of tDNA coverage or indirect determination of pDNA coverage (in case<br />
of low coverages) was done by integration of the cathodic peak of the Fc label from FSCV<br />
measurements. The cathodic peak area can be used to calculate the transferred charge:<br />
Q =<br />
peak area<br />
ν<br />
(5.7)<br />
where peak area is the background corrected charge under the cathodic wave and v is the scan<br />
rate used in the FSCV measurement. Then the surface coverage can be calculated:<br />
Γ =<br />
Q<br />
nFA<br />
(5.8)<br />
where n is the number of electrons transferred in the redox process (in this case 4 since four<br />
ferrocene moieties were used per DNA strand), F is the Faraday constant (96485 C/mol) and A<br />
119
________________________________________________________________ Experimental Work<br />
is the real surface area of the electrode. In all experiments with the exception of multi-electrode<br />
chip measurements 2 mm diameter gold electrodes were used having a geometrical surface area<br />
of 0.0314 cm 2 and a real surface area calculated as explained in Section 5.3. The surface<br />
coverage is obtained in mol/cm 2 .<br />
Finally, to determine the amount of DNA molecules the following equation should be used:<br />
where NDNA is DNA coverage obtained in molecules/cm 2 .<br />
N DNA = Γ × N A (5.9)<br />
5.11 Intercalation<br />
Intercalation of AO-GOx was carried out at room temperature for 15 min by drop casting 10<br />
µL of the intercalator solution (in 100 mM PB) on ssDNA/thiol or dsDNA/thiol-modified<br />
electrodes. The electrode was subsequently rinsed with 10 mM PB containing 450 mM K2SO4<br />
to remove unspecifically bound intercalator molecules.<br />
Chronoamperometric measurements were performed in a 10 mM PB with 450 mM K2SO4<br />
solution containing 1 mM ferricyanide that was purged with Ar for at least 30 min prior to<br />
experiments. During measurements a constant potential of +400 mV (vs. Ag/AgCl/3 M KCl)<br />
was applied and after stabilization of the background current 40 mM glucose was injected to<br />
assure enzyme saturation. Measurements were done under argon atmosphere.<br />
5.12 Methods<br />
5.12.1 Electrochemical impedance spectroscopy<br />
Impedance is a measure of the resistance of the system towards an alternating current and it<br />
extends the concept of resistance to AC circuits. It is a complex function that can be represented<br />
in two ways (Figure 5.6), namely by the real and the imaginary impedance components or by<br />
the modulus and the phase shift:<br />
Z = Z ′ + jZ ′′ (5.10)<br />
120
________________________________________________________________ Experimental Work<br />
Z = |Z|exp(jθ) (5.11)<br />
Z ′ = |Z|sinθ Z ′′ = |Z|cosθ (5.12)<br />
where j = √−1, Z ′ is the real impedance component, Z′ ′ is the imaginary impedance<br />
component, |Z| is the magnitude and θ is the phase shift between potential and current.<br />
Electrochemical impedance spectroscopy is based on the application of a DC potential that is<br />
superimposed with a small amplitude AC potential and measurement of the resulting AC<br />
current signal:<br />
E = E AC sin(ωt) (5.13)<br />
i = i AC sin(ωt + θ) (5.14)<br />
where E AC and i AC are potential and current amplitude, respectively. Except in the case of a pure<br />
resistor, the measured current has a phase shift as compared with the applied potential (Figure<br />
5.6). The impedance of the system is:<br />
Z = E i = E ACsin(ωt)<br />
i AC sin(ωt + θ) = |Z| sin(ωt)<br />
sin(ωt + θ)<br />
(5.15)<br />
Figure 5.6. a) Phase diagram of the applied potential and measured current; b) impedance<br />
vector diagram.<br />
121
________________________________________________________________ Experimental Work<br />
As a result of the measurement, modulus and phase are sampled and the impedance is presented<br />
usually in a Nyquist plot, since it provides fast information about system parameters, or<br />
alternatively in a Bode plot when the frequency dependence needs to be investigated (Figure<br />
5.7).<br />
Figure 5.7. Representation of EIS data: a) Nyquist plot, b) Bode plot.<br />
Figure 5.8. Current vs. potential dependence in electrochemical systems. A region<br />
exhibiting quasi-linear behavior is shown.<br />
For reliable data acquisition, two conditions need to be satisfied. Ideally, during EIS<br />
measurements, the system should remain stationary, that is, the system parameters should not<br />
122
________________________________________________________________ Experimental Work<br />
change during the experiment. Furthermore, since the interpretation becomes much more<br />
complicated for non-linear systems, it is important to perform EIS experiments in a quasi-linear<br />
regime for electrochemical systems (Figure 5.8). This condition determines the amplitude of<br />
the applied AC potential since a smaller amplitude results in a more linear system. Depending<br />
on the system, E AC is typically between 1 to 10 mV.<br />
Table 5.2. Common electrical elements and equivalent electrical circuits, their impedance<br />
and corresponding Nyquist plots.<br />
Equivalent electrical circuit Impedance Nyquist plot<br />
Z = R<br />
Z =<br />
1<br />
jωC<br />
Z = R −<br />
j<br />
ωC<br />
1<br />
Z = 1 R + jωC<br />
EIS data are generally analyzed by fitting to an equivalent electrical circuit. Electrical elements<br />
commonly used for the construction of equivalent electrical circuits are resistor, capacitor,<br />
constant phase element and Warburg impedance. When designing an equivalent electric circuit<br />
123
________________________________________________________________ Experimental Work<br />
it is important to understand the system under investigation, meaning that the electrical elements<br />
need to have a physico-chemical meaning. Table 5.2 shows basic circuit models and their<br />
impedance.<br />
The so-called Randles equivalent circuit is commonly used to model interfacial electrochemical<br />
reactions under semi-infinite linear diffusion control. A constant phase element (CPE) is<br />
commonly used instead of a real capacitor due to frequency capacitance dispersion (Figure 5.9).<br />
The CPE consists of two elements, Q0 that models the capacitance and n that represents the<br />
degree of frequency dispersion with values from 0 to 1, where 1 represents pure capacitive<br />
behavior and 0 pure ohmic resistance.<br />
Figure 5.9. Randles equivalent circuit and corresponding Nyquist plot.<br />
5.12.2 Chronocoulometry for the determination of DNA coverage<br />
Determination of DNA coverage by means of chronocoulometry is based on the determination<br />
of the charge corresponding to the amount of a cationic redox marker ([Ru(NH3)6] 3+ ) noncovalently<br />
bound to DNA strands 75 that is later used to calculate the amount of immobilized<br />
DNA on the electrode surface. Since measurements are performed at low ionic strength (10 mM<br />
PB with 20 mM K2SO4), trivalent [Ru(NH3)6] 3+ can exchange native counterions surrounding<br />
the DNA and non-covalently bind to phosphate residues on DNA strands. To ensure accurate<br />
determination of the amount of DNA strands on the electrode surface, the measurements need<br />
124
________________________________________________________________ Experimental Work<br />
to be conducted under saturation of the redox marker. That is, the method works under the<br />
assumption that [Ru(NH3)6] 3+ exchanges all counterions screening the DNA.<br />
Figure 5.10. Cyclic voltammogram of a bare gold electrode immersed in 10 mM PB<br />
containing 20 mM K2SO4 and 100 μM [Ru(NH3)6]Cl3 in the potential window -0.4 to 0.1<br />
V (vs. Ag/AgCl/3 M KCl) at 100 mV/s scan rate.<br />
In order to measure the charge, a potential step is applied from a potential at which a negligible<br />
reduction of the redox mediator is observed (0 V vs. Ag/AgCl/3 M KCl, Figure 5.10) to a<br />
potential that corresponds to the diffusion limited current for the reduction of all surface<br />
confined redox species (-0.4 V vs. Ag/AgCl/3 M KCl, Figure 5.10). Initially, the charge is<br />
measured in a background solution without the redox marker, to obtain the double layer charge.<br />
Subsequently, the measurement is repeated in the same solution containing additionally the<br />
redox molecule [Ru(NH3)6] 3+ . The charge measured in this solution is given by the integrated<br />
Cottrell equation:<br />
Q = 2nFAD 1<br />
2 C ∗<br />
π 1 2<br />
t 1 2 + Q dl + nFAΓ 0<br />
(5.16)<br />
where D is the diffusion coefficient of the redox molecule (cm 2 /s), C* is the bulk concentration<br />
of [Ru(NH3)6] 3+ (mol/mL), Qdl is the double layer charge and the term nFAΓ0 is the charge<br />
related to the reduction of surface confined species. Making an assumption that the double layer<br />
125
________________________________________________________________ Experimental Work<br />
charge remains the same regardless of the presence of [Ru(NH3)6] 3+ the term nFAΓ0 can be<br />
determined as the difference in the intercept at t = 0 from a Q-t 1/2 graph (Figure 5.11).<br />
Figure 5.11. DNA coverage determination via chronocoulometry. Curves are shown for a<br />
ssDNA/MCH-modified electrode in the absence (grey curve) or presence (green curve) of<br />
[Ru(NH3)6] 3+ in solution.<br />
Since it is assumed that each molecule of [Ru(NH3)6] 3+ binds to three phosphate groups from<br />
the DNA backbone, the amount of DNA strands present on the electrode surface is then<br />
calculated using the following equation:<br />
Γ DNA = Γ 0 ( z m ) N A (5.17)<br />
where z is the charge of the redox molecule and m is the number of phosphate groups at the<br />
DNA strand.<br />
126
6. References<br />
1. Schreiber F.: Structure and growth of self-assembling monolayers. Prog. Sur. Sci. 2000,<br />
65, 151–257.<br />
2. Rant U.; Arinaga K.; Scherer S.; Pringsheim E.; Fujita S.; Yokoyama N., et al.: Switchable<br />
DNA interfaces for the highly sensitive detection of label-free DNA targets. Proc. Natl.<br />
Acad. Sci. U.S.A. 2007, 104, 17364–17369.<br />
3. Gebala M.; Schuhmann W.: Understanding properties of electrified interfaces as a<br />
prerequisite for label-free DNA hybridization detection. Phys. Chem. Chem. Phys. 2012,<br />
14, 14933–14942.<br />
4. Wang J.: Towards Genoelectronics. Electrochemical Biosensing of DNA Hybridization.<br />
Chem. Eur. J. 1999, 5, 1681–1685.<br />
5. Jambrec D.; Gebala M.; La Mantia F.; Schuhmann W.: Potential-Assisted DNA<br />
Immobilization as a Prerequisite for Fast and Controlled Formation of DNA Monolayers.<br />
Angew. Chem. Int. Ed. 2015, 54, 15064–15068.<br />
6. Bigelow W. C.; Pickett D. L.; Zisman W. A.: Oleophobic monolayers. Films adsorbed<br />
from solution in non-polar liquids. J. Colloid. Sci. 1946, 1, 513–538.<br />
7. Porter M. D.; Bright T. B.; Allara D. L.; Chidsey C. E. D.: Spontaneously organized<br />
molecular assemblies. 4. Structural characterization of n-alkyl thiol monolayers on gold by<br />
optical ellipsometry, infrared spectroscopy, and electrochemistry. J. Am. Chem. Soc. 1987,<br />
109, 3559–3568.<br />
8. Nuzzo R. G.; Allara D. L.: Adsorption of bifunctional organic disulfides on gold surfaces.<br />
J. Am. Chem. Soc. 1983, 105, 4481–4483.<br />
9. Daniel K Schwartz: Mechanisms and kinetics of self-assembled monolayer formation.<br />
Annu. Rev. Phys. Chem. 2001, 52, 107–137.<br />
10. Tielens F.; Santos E.: Au-S and SH Bond Formation/Breaking during the Formation of<br />
Alkanethiol SAMs on Au (111). A Theoretical Study. J. Phys. Chem. C 2010, 114, 9444–<br />
9452.<br />
11. Tamada K.; Hara M.; Sasabe H.; Knoll W.: Surface Phase Behavior of n-Alkanethiol Self-<br />
Assembled Monolayers Adsorbed on Au (111). An Atomic Force Microscope Study.<br />
Langmuir 1997, 13, 1558–1566.<br />
12. Gooding J. J.; Mearns F.; Yang W.; Liu J.: Self-Assembled Monolayers into the 21 st<br />
Century. Recent Advances and Applications. Electroanal. 2003, 15, 81–96.<br />
13. Ma F.; Lennox R. B.: Potential-Assisted Deposition of Alkanethiols on Au: Controlled<br />
Preparation of Single- and Mixed-Component SAMs. Langmuir 2000, 16, 6188–6190.<br />
14. Voicu R.; Ellis T. H.; Ju H.; Leech D.: Adsorption and Desorption of Electroactive Self-<br />
Assembled Thiolate Monolayers on Gold. Langmuir 1999, 15, 8170–8177.<br />
15. Sahli R.; Fave C.; Raouafi N.; Boujlel K.; Schöllhorn B.; Limoges B.: Switching on/off<br />
the chemisorption of thioctic-based self-assembled monolayers on gold by applying a<br />
moderate cathodic/anodic potential. Langmuir 2013, 29, 5360–5368.<br />
16. Sondag-Huethorst J. A. M.; Fokkink L. G. J.: Electrochemical Characterization of<br />
Functionalized Alkanethiol Monolayers on Gold. Langmuir 1995, 11, 2237–2241.<br />
17. Gorman C. B.; Biebuyck H. A.; Whitesides G. M.: Control of the Shape of Liquid Lenses<br />
on a Modified Gold Surface Using an Applied Electrical Potential across a Self-<br />
Assembled Monolayer. Langmuir 1995, 11, 2242–2246.<br />
127
18. Canaria C. A.; So J.; Maloney J. R.; Yu C. J.; Smith J. O.; Roukes M. L., et al.: Formation<br />
and removal of alkylthiolate self-assembled monolayers on gold in aqueous solutions. Lab<br />
Chip 2006, 6, 289–295.<br />
19. Diao P.; Hou Q.; Guo M.; Xiang M.; Zhang Q.: Effect of substrate potentials on the<br />
structural disorders of alkanethiol monolayers prepared by electrochemically directed<br />
assembly. J. Electroanal. Chem. 2006, 597, 103–110.<br />
20. Lösch R., Stratmann M., Viefhaus H.: Structure and Properties of Mercaptan-LB Films<br />
prepared under electrochemical potential control. Electrochim. Acta, 39, 1215–1221.<br />
21. Dilimon V. S.; Rajalingam S.; Delhalle J.; Mekhalif Z.: Self-assembly mechanism of thiol,<br />
dithiol, dithiocarboxylic acid, disulfide and diselenide on gold: an electrochemical<br />
impedance study. Phys. Chem. Chem. Phys. 2013, 15, 16648–16656.<br />
22. Li Z.; Niu T.; Zhang Z.; Bi S.: Potential control characteristics of short-chain thiols of<br />
thioctic acid and mercaptohexanol self-assembled on gold. Electrochim. Acta 2010, 55,<br />
6907–6916.<br />
23. Brett C. M. A.; Kresak S.; Hianik T.; Oliveira Brett A. M.: Studies on Self-Assembled<br />
Alkanethiol Monolayers Formed at Applied Potential on Polycrystalline Gold Electrodes.<br />
Electroanal. 2003, 15, 557–565.<br />
24. Petrović Ž.; Metikoš-Huković M.; Babić R.: Potential-assisted assembly of 1-<br />
dodecanethiol on polycrystalline gold. J. Electroanal. Chem. 2008, 623, 54–60.<br />
25. Oswald Avery, Colin MacLeod, Maclyn McCarty: Studies on the chemical nature of the<br />
substance inducing transformation of Pneumococcal types. Induction of transformation by<br />
a desoxyribonucleic acid fraction isolated from Pneumococcus type III. J. Experim. Med.<br />
1944, 79, 137–159.<br />
26. Watson, J. D., Crick, F. H. C.: Molecular structure of nucleic acids. A structure for<br />
deoxyribose nucleic acid. Nature 1953, 171, 4356–4357.<br />
27. Watson, J. D., Crick, F. H. C.: Genetical implications of the structure of deoxyribonucleic<br />
acid. Nature 1953, 171, 964–967.<br />
28. Wilkins, M. H. F., Stokes, A. R., Wilson, H. R.: Molecular structure of deoxypentose<br />
nucleic acids. Nature 1953, 171, 738–740.<br />
29. Franklin, R., Gosling, R. G.: Molecular configuration in sodium thymonucleate. Nature<br />
1953, 171, 740–741.<br />
30. F. Rosalind R. G.: Evidence for 2-chain helix in crystalline structure of sodium<br />
deoxyribonucleate. Nature 1953, 172, 156–157.<br />
31. Wilson, W., Rau, D., Bloomfield, V.: Comparison of polyelectrolyte theories of the<br />
binding of cations to DNA. Biophys. J. 1980, 30, 317–326.<br />
32. G. Manning: Counterion binding in Polyelectrolyte theory. Acc. Chem. Res., 12, 443–449.<br />
33. Stigter D.: Evaluation of the counterion condensation theory of polyelectrolytes. Biophys.<br />
J. 1995, 69, 380–388.<br />
34. Tomić S.; Babić S.; Vuletić T.; Krča S.; Ivanković D.; Griparić L.; Podgornik R.:<br />
Dielectric relaxation of DNA aqueous solutions. Phys. Rev. E 2007, 75.<br />
35. J. Schellman; D. Stigter: Electrical double layer, zeta potential, and electrophoretic charge<br />
of double-stranded DNA. Biopolymers 1977, 16, 1415–1434.<br />
36. Harrington R.: Opticohydrodynamic properties of high-molecular-weight DNA. III. The<br />
effects of NaCl concentration. Biopolymers 1978, 17, 919–936.<br />
37. Chen H.; Meisburger S. P.; Pabit S. A.; Sutton J. L.; Webb W. W.; Pollack L.: Ionic<br />
strength-dependent persistence lengths of single-stranded RNA and DNA. Proc. Natl.<br />
Acad. Sci.USA 2012, 109, 799–804.<br />
38. Bustamante C.; Bryant Z.; Smith S. B.: Ten years of tension: single-molecule DNA<br />
mechanics. Nature 2003, 421, 423–427.<br />
128
39. Tinland, B., Pluen, A., Sturm, J., Weill, G.: Persistence length of single-stranded DNA.<br />
Macromolec. 1997, 30.<br />
40. Chi Q.; Wang G.; Jiang J.: The persistence length and length per base of single-stranded<br />
DNA obtained from fluorescence correlation spectroscopy measurements using mean field<br />
theory. Phys. A 2013, 392, 1072–1079.<br />
41. Netz R. R.; Andelman D.: Neutral and charged polymers at interfaces. Phys. Rep. 2003,<br />
380, 1–95.<br />
42. Kaiser W.; Rant U.: Conformations of end-tethered DNA molecules on gold surfaces:<br />
influences of applied electric potential, electrolyte screening, and temperature. J. Am.<br />
Chem. Soc. 2010, 132, 7935–7945.<br />
43. Bard A. J.; Faulkner L. R.: Electrochemical methods: Fundamentals and applications.<br />
New York: Wiley, 2001.<br />
44. Kelley S. O.; Barton J. K.; Jackson N. M.; McPherson L. D.; Potter A. B.; Spain E. M., et<br />
al.: Orienting DNA Helices on Gold Using Applied Electric Fields. Langmuir 1998, 14,<br />
6781–6784.<br />
45. Gorodetsky A. A.; Buzzeo M. C.; Barton J. K.: DNA-mediated electrochemistry.<br />
Bioconjugate Chem. 2008, 19, 2285–2296.<br />
46. Palecek E.: Past, present and future of nucleic acids electrochemistry. Talanta 2002, 56,<br />
809–819.<br />
47. Nimse S. B.; Song K.; Sonawane M. D.; Sayyed D. R.; Kim T.: Immobilization techniques<br />
for microarray: challenges and applications. Sensors 2014, 14, 22208–22229.<br />
48. Odenthal K. J.; Gooding J. J.: An introduction to electrochemical DNA biosensors.<br />
Analyst 2007, 132, 603–610.<br />
49. Liu A.; Wang K.; Weng S.; Lei Y.; Lin L.; Chen W., et al.: Development of<br />
electrochemical DNA biosensors. TrAC 2012, 37, 101–111.<br />
50. Cooper J.; Cass A. E. G. (eds): Biosensors: A practical approach. Oxford: Oxford<br />
University Press 2004.<br />
51. Miller M. B.; Tang Y.-W.: Basic concepts of microarrays and potential applications in<br />
clinical microbiology. Clin. Microbiol. Rev. 2009, 22, 611–633.<br />
52. Ronald G. Sosnowski, Eugene Tu, William F. Butler, James P. O’Connel, and Michael J.<br />
Heller: Rapid determination of single base mismatch mutations in DNA hybrids by direct<br />
electric field control. Proc. Natl. Acad. Sci. 1997, 94, 1119–1123.<br />
53. Wang J.: Electrochemical nucleic acid biosensors. Anal. Chim. Acta 2002, 469, 63–71.<br />
54. Ryan M. Fryer, Jeffrey Randall, Takumi Yoshida, Li-Li Hsiao: Global analysis of gene<br />
expression: methods, interpretation, and pitfalls. Exp. Nephrol. 2002, 10, 64–74.<br />
55. TELES F.; FONSECA L.: Trends in DNA biosensors. Talanta 2008, 77, 606–623.<br />
56. Drummond T. G.; Hill M. G.; Barton J. K.: Electrochemical DNA sensors. Nature<br />
Biotech. 2003, 21, 1192–1199.<br />
57. Zhou X. C.; Huang L. Q.; Li S. F. Y.: Microgravimetric DNA sensor based on quartz<br />
crystal microbalance. Comparison of oligonucleotide immobilization methods and the<br />
application in genetic diagnosis. Biosens. Bioelec. 2001, 16, 85–95.<br />
58. Dell'Atti D.; Zavaglia M.; Tombelli S.; Bertacca G.; Cavazzana A. O.; Bevilacqua G., et<br />
al.: Development of combined DNA-based piezoelectric biosensors for the simultaneous<br />
detection and genotyping of high risk Human Papilloma Virus strains. Clin. Chim. Acta<br />
2007, 383, 140–146.<br />
59. Palecek E.; Bartosik M.: Electrochemistry of nucleic acids. Chem. Rev. 2012, 112, 3427–<br />
3481.<br />
60. Armistead P. M.; Thorp H. H.: Modification of Indium Tin Oxide Electrodes with Nucleic<br />
Acids. Detection of Attomole Quantities of Immobilized DNA by Electrocatalysis. Anal.<br />
Chem. 2000, 72, 3764–3770.<br />
129
61. Mikkelsen S. R.: Electrochecmical biosensors for DNA sequence detection. Electroanal.<br />
1996, 8, 15–19.<br />
62. Furst A.; Hill M. G.; Barton J. K.: Electrocatalysis in DNA Sensors. Polyhedron 2014, 84,<br />
150–159.<br />
63. Gebala M.; Stoica L.; Neugebauer S.; Schuhmann W.: Label-Free Detection of DNA<br />
Hybridization in Presence of Intercalators Using Electrochemical Impedance<br />
Spectroscopy. Electroanal. 2009, 21, 325–331.<br />
64. Zhang Y.; Pothukuchy A.; Shin W.; Kim Y.; Heller A.: Detection of approximately 10(3)<br />
copies of DNA by an electrochemical enzyme-amplified sandwich assay with ambient<br />
O(2) as the substrate. Anal. Chem. 2004, 76, 4093–4097.<br />
65. Umek R. M.; Lin S. W.; Vielmetter J.; Terbrueggen R. H.; Irvine B.; Yu C. J., et al.:<br />
Electronic Detection of Nucleic Acids. J. Mol. Diag. 2001, 3, 74–84.<br />
66. Doneux T.; De Rache A.; Triffaux E.; Meunier A.; Steichen M.; Buess-Herman C.:<br />
Optimization of the Probe Coverage in DNA Biosensors by a One-Step Coadsorption<br />
Procedure. ChemElectroChem 2014, 1, 147–157.<br />
67. Keighley S. D.; Li P.; Estrela P.; Migliorato P.: Optimization of DNA immobilization on<br />
gold electrodes for label-free detection by electrochemical impedance spectroscopy.<br />
Biosens. Bioelec. 2008, 23, 1291–1297.<br />
68. Ravan H.; Kashanian S.; Sanadgol N.; Badoei-Dalfard A.; Karami Z.: Strategies for<br />
optimizing DNA hybridization on surfaces. Anal. Biochem. 2014, 444, 41–46.<br />
69. Dickertmann D.; Schultze J. W.; Vetter K. J.: Electrochemical formation and reduction of<br />
monomolecular oxide layers on (111) and (100) planes of gold single crystals. J.<br />
Electroanal. Chem. Inter. Electrochem. 1974, 55, 429–443.<br />
70. Bandarenka A. S.: Exploring the interfaces between metal electrodes and aqueous<br />
electrolytes with electrochemical impedance spectroscopy. Analyst 2013, 138, 5540.<br />
71. Gebala M.; Schuhmann W.: Controlled Orientation of DNA in a Binary SAM as a Key for<br />
the Successful Determination of DNA Hybridization by Means of Electrochemical<br />
Impedance Spectroscopy. ChemPhysChem 2010, 11, 2887–2895.<br />
72. Lisdat F.; Schäfer D.: The use of electrochemical impedance spectroscopy for biosensing.<br />
Anal. Bioanal. Chem. 2008, 391, 1555–1567.<br />
73. Mearns F. J.; Wong E. L. S.; Short K.; Hibbert D. B.; Gooding J. J.: DNA Biosensor<br />
Concepts Based on a Change in the DNA Persistence Length upon Hybridization.<br />
Electroanalysis 2006, 18, 1971–1981.<br />
74. Herne T. M.; Tarlov M. J.: Characterization of DNA Probes Immobilized on Gold<br />
Surfaces. J. Am. Chem. Soc. 1997, 119, 8916–8920.<br />
75. Steel A. B.; Herne T. M.; Tarlov M. J.: Electrochemical Quantitation of DNA<br />
Immobilized on Gold. Anal. Chem. 1998, 70, 4670–4677.<br />
76. Cuesta A.: Measurement of the surface charge density of CO-saturated Pt(111) electrodes<br />
as a function of potential. The potential of zero charge of Pt(111). Sur. Sci. 2004, 572, 11–<br />
22.<br />
77. Shao L.-H.; Biener J.; Kramer D.; Viswanath R. N.; Baumann T. F.; Hamza A. V.;<br />
Weissmüller J.: Electrocapillary maximum and potential of zero charge of carbon aerogel.<br />
Phys. Chem. Chem. Phys. 2010, 12, 7580–7587.<br />
78. Lorenz W. J.; Plieth W.: Electrochemical nanotechnology: In-situ local probe techniques<br />
at electrochemical interfaces. Weinheim, New York: Wiley-VCH, ©1998.<br />
79. Gnahm M.; Pajkossy T.; Kolb D. M.: The interface between Au(111) and an ionic liquid.<br />
Electrochim. Acta 2010, 55, 6212–6217.<br />
80. Čolić V.; Tymoczko J.; Maljusch A.; Ganassin A.; Schuhmann W.; Bandarenka A. S.:<br />
Experimental Aspects in Benchmarking of the Electrocatalytic Activity.<br />
ChemElectroChem 2015, 2, 143–149.<br />
130
81. Pasta M.; Battistel A.; La Mantia F.: Lead–lead fluoride reference electrode. Electrochem.<br />
Commun. 2012, 20, 145–148.<br />
82. Lockett V.; Sedev R.; Ralston J.; Horne M.; Rodopoulos T.: Differential Capacitance of<br />
the Electrical Double Layer in Imidazolium-Based Ionic Liquids. Influence of Potential,<br />
Cation Size, and Temperature. J. Phys. Chem. C 2008, 112, 7486–7495.<br />
83. Grahame D. C.: Differential Capacity of Mercury in Aqueous Sodium Fluoride Solutions.<br />
I. Effect of Concentration at 25°. J. Am. Chem. Soc. 1954, 76, 4819–4823.<br />
84. Pheeney C. G.; Arnold A. R.; Grodick M. A.; Barton J. K.: Multiplexed electrochemistry<br />
of DNA-bound metalloproteins. J. Am. Chem. Soc. 2013, 135, 11869–11878.<br />
85. Peterson A. W.: The effect of surface probe density on DNA hybridization. Nucleic Acids<br />
Res. 2001, 29, 5163–5168.<br />
86. Revenga-Parra M.; García T.; Pariente F.; Lorenzo E.; Alonso C.: Effects of Ionic<br />
Strength and Probe DNA Length on the Electrochemical Impedance Spectroscopic<br />
Response of Biosensors. Electroanal. 2011, 23, 100–107.<br />
87. Gong P.; Levicky R.: DNA surface hybridization regimes. Proc. Natl. Acad. Sci. 2008,<br />
105, 5301–5306.<br />
88. Irving D.; Gong P.; Levicky R.: DNA Surface Hybridization: Comparison of Theory and<br />
Experiment. J. Phys. Chem. B 2010, 114, 7631–7640.<br />
89. Jambrec D.; Conzuelo F.; Estrada-Vargas A.; Schuhmann W.: Potential-Pulse-Assisted<br />
Formation of Thiol Monolayers within Minutes for Fast and Controlled Electrode Surface<br />
Modification. ChemElectroChem 2016.<br />
90. Bain C. D.; Troughton E. B.: Formation of monolayer films by the spontaneous assembly<br />
of organic thiols from solution onto gold. J. Am. Chem. Soc. 1989, 111, 321–335.<br />
91. Qu D.; Morin M.: The effect of concentration on the oxidative deposition of a monolayer<br />
of alkylthiolate on gold. From island formation to random adsorption. J. Electroanal.<br />
Chem. 2004, 565, 235–242.<br />
92. Subramanian R.; Lakshminarayanan V.: A study of kinetics of adsorption of alkanethiols<br />
on gold using electrochemical impedance spectroscopy. Electrochim. Acta 2000, 45,<br />
4501–4509.<br />
93. Kuznetsov V.; Papastavrou G.: Ion Adsorption on Modified Electrodes as Determined by<br />
Direct Force Measurements under Potentiostatic Control. J. Phys. Chem. C 2014, 118,<br />
2673–2685.<br />
94. Peterlinz K. A.; Georgiadis R.: In Situ Kinetics of Self-Assembly by Surface Plasmon<br />
Resonance Spectroscopy. Langmuir 1996, 12, 4731–4740.<br />
95. Hill R. T.; Mock J. J.; Hucknall A.; Wolter S. D.; Jokerst N. M.; Smith D. R.; Chilkoti A.:<br />
Plasmon Ruler with Angstrom Length Resolution. ACS Nano 2012, 6, 9237–9246.<br />
96. Strutwolf J.; O'Sullivan C. K.: Microstructures by Selective Desorption of Self-Assembled<br />
Monolayer from Polycrystalline Gold Electrodes. Electroanal. 2007, 19, 1467–1475.<br />
97. Sun K.; Jiang B.; Jiang X.: Electrochemical desorption of self-assembled monolayers and<br />
its applications in surface chemistry and cell biology. J. Electroanal. Chem. 2011, 656,<br />
223–230.<br />
98. Liepold P.; Kratzmüller T.; Persike N.; Bandilla M.; Hinz M.; Wieder H., et al.:<br />
Electrically detected displacement assay (EDDA): a practical approach to nucleic acid<br />
testing in clinical or medical diagnosis. Anal. Bioanal. Chem. 2008, 391, 1759–1772.<br />
99. Rentsch S.; Siegenthaler H.; Papastavrou G.: Diffuse layer properties of thiol-modified<br />
gold electrodes probed by direct force measurements. Langmuir : the ACS journal of<br />
surfaces and colloids 2007, 23, 9083–9091.<br />
100. Zimdars A.; Gebala M.; Hartwich G.; Neugebauer S.; Schuhmann W.:<br />
Electrochemical detection of synthetic DNA and native 16S rRNA fragments on a<br />
131
microarray using a biotinylated intercalator as coupling site for an enzyme label. Talanta<br />
2015, 143, 19–26.<br />
101. Gebala M.; Hartwich G.; Schuhmann W.: Amplified detection of DNA hybridization<br />
using post-labelling with a biotin-modified intercalator. Faraday Discuss. 2011, 149, 11–<br />
22.<br />
102. Yang Z.; Anglès d'Auriac M.; Goggins S.; Kasprzyk-Hordern B.; Thomas K. V.; Frost<br />
C. G.; Estrela P.: A novel DNA biosensor using a ferrocenyl intercalator applied to the<br />
potential detection of human population biomarkers in wastewater. Environ. Sci. Technol.<br />
2015, 49, 5609–5617.<br />
103. Won B. Y.; Lee D. W.; Shin S. C.; Cho D. Y.; Lee S. S.; Yoon H. C.; Park H. G.: A<br />
DNA intercalation-based electrochemical method for detection of Chlamydia trachomatis<br />
utilizing peroxidase-catalyzed signal amplification. Biosens. Bioelec. 2008, 24, 665–669.<br />
104. Gebala M.; Stoica L.; Guschin D.; Stratmann L.; Hartwich G.; Schuhmann W.: A<br />
biotinylated intercalator for selective post-labeling of double-stranded DNA as a basis for<br />
high-sensitive DNA assays. Electrochem. Commun. 2010, 12, 684–688.<br />
105. Biver T.; Eltugral N.; Pucci A.; Ruggeri G.; Schena A.; Secco F.; Venturini M.:<br />
Synthesis, characterization, DNA interaction and potential applications of gold<br />
nanoparticles functionalized with Acridine Orange fluorophores. Dalton Trans. 2011, 40,<br />
4190–4199.<br />
106. Georghiou S.: Interaction of acridine drugs with DNA and nucleotides. Photochem.<br />
Photobiol. 1977, 26, 59–68.<br />
107. Won B. Y.; Choi H. G.; Kim K. H.; Byun S. Y.; Kim H. S.; Yoon H. C.:<br />
Bioelectrocatalytic signaling from immunosensors with back-filling immobilization of<br />
glucose oxidase on biorecognition surfaces. Biotechnol Bioeng 2005, 89, 815–821.<br />
108. Courjean O.; Gao F.; Mano N.: Deglycosylation of Glucose Oxidase for Direct and<br />
Efficient Glucose Electrooxidation on a Glassy Carbon Electrode. Angew. Chem. 2009,<br />
121, 6011–6013.<br />
109. Libertino S.; Aiello V.; Scandurra A.; Renis M.; Sinatra F.: Immobilization of the<br />
Enzyme Glucose Oxidase on Both Bulk and Porous SiO2 Surfaces. Sensors 2008, 8,<br />
5637–5648.<br />
110. Schuhmann W.; Ohara T. J.; Schmidt H. L.; Heller A.: Electron transfer between<br />
glucose oxidase and electrodes via redox mediators bound with flexible chains to the<br />
enzyme surface. J. Am. Chem. Soc. 1991, 113, 1394–1397.<br />
111. Hoogvliet J. C.; Dijksma M.; Kamp B.; van Bennekom, W. P.: Electrochemical<br />
Pretreatment of Polycrystalline Gold Electrodes To Produce a Reproducible Surface<br />
Roughness for Self-Assembly: A Study in Phosphate Buffer pH 7.4. Anal. Chem. 2000,<br />
72, 2016–2021.<br />
112. Oesch U.; Janata J.: Electrochemical study of gold electrodes with anodic oxide films -<br />
I. Formation and reduction behaviour of anodic oxides on gold. Electrochim. Acta 1983,<br />
28, 1237–1246.<br />
132
________________________________________________________________________ Appendix<br />
7. Appendix<br />
7.1 List of symbols and abbreviations<br />
A<br />
real surface area<br />
a<br />
polymer radius<br />
AC alternating current<br />
AD/DA analog-digital/digital-analog conversion<br />
AO acridine orange<br />
AO-GOx conjugate of acridine orange and glucose-oxidase<br />
b<br />
charge separation<br />
C<br />
interfacial capacitance<br />
C* bulk concentration<br />
C0<br />
Cd<br />
Cddl<br />
CE<br />
Cf<br />
Ci<br />
CPE<br />
CV<br />
d<br />
D<br />
DC<br />
dsDNA<br />
e<br />
EAC<br />
EIS<br />
F<br />
Fc<br />
Fc-tDNA<br />
initial capacitance<br />
differential capacitance<br />
capacitance of the diffuse double layer<br />
counter electrode<br />
capacitance of a fully covered monolayer<br />
capacitance of the compact layer<br />
constant phase element<br />
cyclic voltammetry<br />
distance from the surface<br />
diffusion coeficient<br />
direct current<br />
double stranded DNA<br />
elementary charge<br />
potential AC amplitude<br />
electrochemical impedance spectroscopy<br />
Faraday constant<br />
ferrocene<br />
ferrocene-labelled tDNA<br />
133
________________________________________________________________________ Appendix<br />
FSCV<br />
GC<br />
GOx<br />
I<br />
iAC<br />
k<br />
L<br />
lB<br />
lel<br />
lo<br />
lp<br />
m<br />
MCH<br />
MCHD<br />
MCU<br />
n<br />
NA<br />
Nbp<br />
Nbp<br />
NDNA<br />
OCP<br />
PB<br />
pzc<br />
pzc (DNA)<br />
pzfc<br />
pztc<br />
Q<br />
Qdl<br />
Rct<br />
RE<br />
Rg<br />
Rs<br />
fast-scan cyclic voltammetry<br />
Gouy-Chapman (theory)<br />
glucose-oxidase<br />
ionic strength<br />
current AC amplitude<br />
rate constant<br />
contour length of DNA<br />
Bjerrum length<br />
electrostatic repulsion within a polymer<br />
intrinsic stiffness of a polymer<br />
persistence length<br />
number of phosphate groups<br />
mercapto-1-hexanol<br />
mercapto-1-hexadecanol<br />
mercapto-1-undecanol<br />
number of electrons<br />
Avogadro number<br />
number of base pairs<br />
number of bases<br />
DNA coverage<br />
open circuit potential<br />
Poisson-Boltzmann (equation)<br />
potential of zero charge<br />
potential of zero charge of DNA-modified<br />
electrode<br />
potential of zero free charge<br />
potential of zero total charge<br />
charge<br />
double layer charge<br />
charge trasfter resistance<br />
reference electrode<br />
radius of gyration<br />
solution resistance<br />
134
________________________________________________________________________ Appendix<br />
SAM<br />
ssDNA<br />
tDNA<br />
W<br />
WE<br />
z<br />
|Z|<br />
Z'<br />
-Z"<br />
Γ<br />
δ<br />
ε<br />
ε0<br />
η<br />
θ<br />
κ -1<br />
ν<br />
σm<br />
φ<br />
φ0<br />
ω<br />
self-assembled monolayer<br />
single stranded DNA<br />
target DNA<br />
Warburg constant<br />
working electrode<br />
valence<br />
modulus<br />
real impedance component<br />
imaginary impedance component<br />
surface coverage<br />
thickness increment<br />
dielectric constant<br />
permeability of vacuum<br />
charge density<br />
partial coverage; phase shift<br />
Debye length<br />
scan rate<br />
excess charge of a metal<br />
potential at a distance d from the surface<br />
potential at the surface<br />
angular frequency<br />
135
________________________________________________________________________ Appendix<br />
7.2 Publications list<br />
Published papers:<br />
D. Jambrec, F. Conzuelo, A. Estrada-Vargas, W. Schuhmann, “Potential Pulse-Assisted<br />
Formation of Thiol Monolayers within Minutes as a Promising Technique for Fast and<br />
Controlled Electrode Surface Modification”. ChemElectroChem, 2016, 3, 1484-1489.<br />
D. Jambrec, R. Haddad, A. Lauks, M. Gebala, W. Schuhmann, M. Kokoschka, “DNA<br />
Intercalators for Detection of DNA Hybridization. SCS(MI)-MP2 Calculations and<br />
Electrochemical Impedance Spectroscopy”. ChemPlusChem, 2016, 81, 604-612.<br />
A. Estrada-Vargas, D. Jambrec, Y. U. Kayran, V. Kuznetsov, W. Schuhmann, “Differentiation<br />
Between Single and Double-Stranded DNA by Local Capacitance Measurements”.<br />
ChemElectroChem, 2016, 3, 855-857.<br />
D. Jambrec, M. Gebala, F. La Mantia, W. Schuhmann, “Potential-assisted DNA<br />
Immobilization as a Prerequisite for Fast and Controlled Formation of DNA Monolayers”.<br />
Angew. Chem., 2015, 127, 15278–15283; Angew. Chem. Int. Ed., 2015, 54, 15064–15068.<br />
L. Švorc, D. Jambrec, M. Vojs, S. Barwe, J. Clausmeyer, P. Michniak, M. Marton, W.<br />
Schuhmann, “Doping Level of Boron-Doped Diamond Electrodes Controls the Grafting<br />
Density of Functional Groups for DNA Assays”. ACS Appl. Mater. Interfaces, 2015, 7, 18949-<br />
18956.<br />
S. Pilehvar*, D. Jambrec*, M. Gebala, W. Shuhmann, K. De Wael, “Intercalation of Proflavine<br />
in ssDNA Aptamers: Effect on Binding of the Specific Target Chloramphenicol”. Electroanal.,<br />
2015, 27, 1836-1841.<br />
V. Eßmann, D. Jambrec, A. Kuhn, W. Schuhmann, “Linking Glucose Oxidation to Luminol-<br />
Based Electrochemiluminescence Using Bipolar Electrochemistry”. Electrochem. Comm.,<br />
2015, 50, 77-80.<br />
* these authors contributed equally to this work<br />
136
________________________________________________________________________ Appendix<br />
In preparation:<br />
D. Jambrec, A. Ruff, W. Schuhmann, “Amperometric Detection of dsDNA via Acrydinium<br />
Orange Modified Glucose Oxidase”.<br />
D. Jambrec, Y. Uğur Kayran, W. Schuhmann, “Fully potential-controlled preparation of DNA<br />
chips”.<br />
D. Jambrec, W. Schuhmann: “Potential-assisted DNA/thiol co-immobilization as a tool for<br />
DNA chip production”.<br />
D. Jambrec, B. Ciui, C. Cristea, W. Schuhmann: “Potential-assisted dehybridization as a DNA<br />
detection tool employing a double-stranded ligation strategy”.<br />
137
________________________________________________________________________ Appendix<br />
7.3 Conference contributions<br />
Oral presentations:<br />
D. Jambrec, M. Gebala, F. La Mantia, W. Schuhmann, “Diving into the Mechanism of<br />
Potential-Assisted ssDNA Immobilization”, 66th Annual Meeting of the International Society<br />
of Electrochemistry, 4 th -9 th October 2015, Taipei (Taiwan), best presentation prize.<br />
D. Jambrec, M. Gebala, F. La Mantia, W. Schuhmann, “Highly controlled and fast potentialassisted<br />
ssDNA immobilization”, Summer meeting on bio-electrochemistry (SMOBE 2015),<br />
17 th -20 th August 2015, Antwerp (Belgium), best presentation prize.<br />
D. Jambrec, M. Gebala, F. La Mantia, W. Schuhmann, “Highly Controlled and Fast Formation<br />
of ssDNA-Modified Surfaces Using a Potential-Assisted Immobilization Method”, XXIII<br />
International Symposium on Bioelectrochemistry and Bioenergetics, 14 th -18 th June 2015,<br />
Malmö (Sweden).<br />
D. Jambrec, M. Gebala, F. La Mantia, W. Schuhmann, “Potential assisted immobilization of<br />
DNA probe on Au electrode surfaces”, Bioenergy Summer School, 28 th September-4 th October<br />
2014, Ile d`Oleron (France).<br />
Poster presentations:<br />
D. Jambrec, F. Conzuelo, A. Ruff, A. Estrada-Vargas, W. Schuhmann, “Optimizing DNA<br />
Assays. DNA Sensor Preparation in Minutes”, 67th Annual Meeting of the International<br />
Society of Electrochemistry, 21 st -26 th August 2016, The Hague (Netherlands).<br />
D. Jambrec, F. Conzuelo, A. Estrada-Vargas, W. Schuhmann, “Potential pulse-assisted<br />
formation of thiol monolayers within minutes as a promising technique for fast and controlled<br />
electrode surface modification”, 16 th International Conference on Electroanalysis, 12 th -16 th<br />
June 2016, Bath (UK).<br />
138
________________________________________________________________________ Appendix<br />
D. Jambrec, M. Gebala, W. Schuhmann, “Fast Potential-Assisted Immobilization of<br />
ssDNA on Au Electrode Surfaces”, 65th Annual Meeting of the International Society of<br />
Electrochemistry, 31 st August-5 th September 2014, Lausanne (Switzerland), best poster award.<br />
D. Jambrec, M. Gebala, W. Schuhmann, “Potential assisted immobilization of<br />
DNA probe on Au electrode surfaces”, Surface Modification for Chemical and Biochemical<br />
Sensing 2013, 8 th November-12 th November 2013, Łochów (Polland).<br />
139