Degenerate Si as an Electrode Material for Electrochemical Biosensors

Journal of The Electrochemical Society, 155 共12兲 J350-J354 共2008兲

J350

0013-4651/2008/155共12兲/J350/5/$23.00 © The Electrochemical Society

Degenerate Si as an Electrode Material for Electrochemical Biosensors Yin Huang and Ian Ivar Suni*,z Department of Chemical and Biomolecular Engineering, Center for Advanced Materials Processing, Clarkson University, Potsdam, New York 13699-5705, USA The use of degenerate 共heavily doped兲 Si as an electrode material for electrochemical biosensors is demonstrated by a series of voltammetry and impedance studies. Cyclic voltammetry in solutions of NH4F and K3Fe共CN兲6 /K4Fe共CN兲6 exhibits reversible electrochemistry without illumination, which to the best of the authors’ knowledge has not been previously reported for Si electrodes in an aqueous environment. The presence of F− prevents formation of the Si native oxide, while the use of a degenerate Si substrate makes its electrochemistry more like that of a metal. In addition, impedance detection of peanut protein Ara h 1 is demonstrated on a degenerate Si electrode following immobilization of the monoclonal mouse antibody to this protein. The variation in charge-transfer resistance with protein concentration is employed to estimate the dissociation constant of the surfaceimmobilized antibody-antigen complex as 0.25 nM. The advantages of degenerate Si relative to n- and p-type Si include the lack of photoeffects, a simpler equivalent circuit, easier electrical connection to the working electrode, and knowledge that the applied potential applied is dropped within the electrical double layer, not within the semiconductor space-charge layer. Most important, degenerate Si electrodes are easily incorporated into Si-based semiconductor devices. © 2008 The Electrochemical Society. 关DOI: 10.1149/1.2988047兴 All rights reserved. Manuscript submitted June 5, 2008; revised manuscript received August 25, 2008. Published October 7, 2008.

Due to the selectivity associated with biorecognition processes, biosensors for detection of numerous analytes have recently been an area of intense research focus. Biosensor applications include realtime in vivo monitoring of biological processes, clinical measurements of biomedical disease markers, real-time industrial and environmental process monitoring, detection of biological warfare agents, and detection of food allergens. While optical biosensors have been widely studied, electrochemical biosensors have some unique advantages, including the ability to incorporate optically opaque materials, including porous materials, avoiding false positives arising from contaminating chromophores, and easier miniaturization.1-3 In addition, for portable and implantable applications, they have the additional advantages of simplicity and lower noise levels, because room-temperature optical detectors have relatively high noise levels. The electrode materials in electrochemical biosensors are constrained by the requirements for both high electrical conductivity and biocompatibility, because biomolecules often denature with prolonged exposure to metal surfaces. To date, this has limited the electrode materials in electrochemical biosensors primarily to Au, Pt, and some form of carbon.4,5 Si has been widely employed as a biosensor substrate material, but in the vast majority of applications, biomolecules have been immobilized onto SiO2 rather than Si, preventing electrochemical interrogation. Recently, immobilization of biomolecules onto n- or p-type Si has been reported through direct formation of a Si–C bond by either alkene/alkyne insertion into Si–H bonds6,7 or cathodic electrografting.8,9 However, electrochemistry on n- and p-type Si is complicated by the partitioning of the applied potential between the semiconductor space-charge layer and the electrical double layer at the electrode/electrolyte interface.10 In many such cases, electrochemical processes may be irreversible due to Schottky barrier formation at the semiconductor-electrolyte interface.10 Here we demonstrate an electrochemical impedance biosensor that employs degenerate Si as the electrode material. Degenerate Si contains extremely high dopant concentrations and behaves like an electrically conducting, not semiconducting, electrode. This impedance biosensor is demonstrated for detection of peanut protein Ara h 1 following surface immobilization of its monoclonal mouse antibody. This was also recently demonstrated in our laboratory using a

* Electrochemical Society Active Member. z

E-mail: [email protected]

Au electrode.11 Peanut protein Ara h 1 is a common food allergen, so detection of this species can address a serious public health problem. Experimental As-doped 共n-type兲 degenerate Si共111兲 wafers with a thickness of 300 ␮m and a diameter of 50 mm were purchased from Virginia Semiconductor, Inc. The resistivity of these Si wafers is less than 0.005 ⍀ cm. 10-Undecenoic acid was purchased from Alfa; hydrochloride N-共3-dimethylaminopropyl兲-N⬘-ethylcarbodiimide 共EDC兲, potassium dihydrogen phosphate, and dipotassium hydrogen phosphate were purchased from Sigma; N-hydroxysulfosuccinimide sodium salt 共NHSS兲 was purchased from Pierce Biotechnology; peanut protein Ara h1 and anti-Ara h1 mouse monoclonal antibody were purchased from Indoor Biotechnologies. All reagents were used as received. The Ara h1 protein was labeled as the monomeric form. The degenerate Si electrode was cleaned in ethanol and water and etched in 40 wt % NH4F to remove the native oxide. The degenerate Si electrode was then immersed immediately in 10% 10undecenoic acid in deaerated toluene solution in a photochemical reactor for 19 h where it was exposed to UV light.12 The UV light source used was a Way Too Cool model WTC 72L-110 that provides a long wavelength output at 352 nm. Photochemical activation is necessary to increase the kinetics of alkene insertion into Si–H bonds, which forms a covalent Si–C bond. The exposed carboxylic acid groups were then activated by immersion for 1 h into 75 mM EDC, 15 mM NHSS, and 50 mM phosphate buffer solution 共PBS兲 at pH 7.3. Following carboxylate group activation, the antibody to peanut protein Ara h 1 was immobilized by immersing the electrode for 1 h into a solution containing 50 ␮g/mL antibody and 50 mM PBS at pH 7.3. This procedure immobilizes the antibody onto the degenerate Si electrode by amide bond formation to amine groups on the protein surface.13 This sensor electrode can then be used to detect the peanut protein Ara h1 in solution. Figure 1 illustrates the interface construction and detection scheme. The degenerate Si electrode was used as the working electrode in a virgin Teflon three-electrode cell with a Pt spiral counter electrode and an Ag/AgCl 共1.0 M KCl兲 reference electrode. Connection to the Si electrode was made to the back side using Cu tape. This is considerably easier than making an ohmic electrical contact to a Si semiconductor electrode, which typically requires a Ga–In eutectic or other special alloy.14 Voltammetry measurements were performed using an EG&G PAR 263A potentiostat. Impedance measurements were performed using the same potentiostat coupled to a Solartron 1250 frequency response analyzer over the frequency range

Downloaded 15 Oct 2008 to 128.153.14.49. Redistribution subject to ECS license or copyright; see http://www.ecsdl.org/terms_use.jsp


J351

0.20 0.15 0.10

2

j (mA/cm )

0.05 0.00 -0.05 -0.10

(a)

-0.15 -0.20 -0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1

0.0

0.1

0.2

0.3

E (V vs. Ag/AgCl) Figure 2. Voltammetric behavior of bare degenerate Si in 50 mM PBS buffer solution and 5 mM K3Fe共CN兲6 /K4Fe共CN兲6 in 10 wt % NH4F at a scan rate of 50 mV/s. The dashed line corresponds to voltammetry of n-type Si 共␳ = 2–35 ⍀ cm兲 for comparison.

(b) Figure 1. Schematic illustration of the interface construction and detection scheme for peanut protein Ara h 1: 共a兲 sensor electrode with self-assembly film and antibody immobilized and 共b兲 detection of the protein Ara h1 by the antibody. The Y-shaped molecule represents the antibody and the orbital shape represents protein Ara h1.

0.05 Hz to 15 kHz with an ac probe amplitude of 5 mV. Each impedance spectrum takes about 2.5 min to acquire. The impedance results were obtained at a dc potential of 0 mV vs Ag/AgCl, which is slightly cathodic to the open-circuit potential 共OCP兲共about + 50 mV兲 in the electrolytes of interest. This procedure was followed to minimize the possibility of Si oxidation. The 10undecenoic self-assembled monolayer 共SAM兲 reached steady state immediately, in contrast to the 11-mercaptoundecanoic acid SAM formed on Au, which takes about 30 min to reach steady state, as measured by changes in the OCP.11 Results and Discussion Before considering impedance studies of protein immobilization on degenerate Si, this new substrate material was characterized by cyclic voltammetry 共CV兲. Figure 2 shows the voltammogram obtained for bare 共not oxidized兲 degenerate Si in 50 mM PBS buffer solution and 5 mM K3Fe共CN兲6 /K4Fe共CN兲6 at pH 7.3 in 10 wt % NH4F. This voltammogram shows reversible oxidation/reduction peaks for K3Fe共CN兲6 /K4Fe共CN兲6 with a peak separation of about 340 mV, suggesting metal-like behavior of this electrode. To the best of the authors’ knowledge, reversible hexacyanoferrate oxidation/reduction peaks have not been previously reported for an aqueous electrolyte on a Si electrode of any kind. For comparison, Fig. 2 also presents a voltammogram in the same electrolyte on an n-type Si electrode, showing no Fe共CN兲3−/4− oxidation/reduction 6 peaks. Si electrochemistry has been recently reviewed by Zhang.15 While many redox reactions on Si electrodes have been studied, most are irreversible, often involving Si oxidation and dissolution, metal deposition, or hydrogen evolution. For redox couples such as

Fe共CN兲3−/4− , whose electrochemistry is typically reversible on metal 6 electrodes, most reported voltammograms do not exhibit both an oxidation peak 共anodic scan兲 and a reduction peak 共cathodic scan兲, as seen in Fig. 2. Reversible aqueous voltammetry on Si has typically been observed primarily on illuminated Si electrodes.15 Because illumination is a significant constraint for sensing applications, reversible electrochemistry that does not depend on photoeffects is highly desirable. Few reports exist of reversible aqueous electrochemistry on Si electrodes that are not illuminated. Two-electron oxidation/reduction peaks have been reported for methyl viologen in an aqueous electrolyte for n-type Si electrode following oxide removal in HF, and this was explained by arguing that the Si electrode was under accumulation conditions. 16 Reversible oxidation/reduction peaks have been reported for ferrocene and its derivates on n- and p-type Si, but only through incorporation of the electroactive species into a surface film by direct Si–C bond formation.17-20 Metal electrodeposition onto n-type Si has been widely studied by Searson and others, but electrodissolution of noble metals 共Au, Pt, Ag, Cu兲 from these substrates has not been observed due to Schottky barrier formation.10 Generally, during the cathodic sweep, the metal deposition peak overlaps with hydrogen evolution, but no current peak is observed during the anodic sweep. However, it should be noted that Penner’s group has reported at least partially reversible Ag electrodeposition on degenerate n-type Si共100兲, although degenerate Si electrochemistry was not thoroughly investigated.21 In addition, at least partially reversible deposition of less noble metals such as Pb has been demonstrated on n-type Si.22-24 In summary, reversible aqueous electrochemistry such as that illustrated in Fig. 2 has not often been reported. The impedance spectra obtained at different stages of degenerate Si electrode preparation are shown in Fig. 3. The best complex nonlinear least square 共CNLS兲 fitting results for these spectra to a standard Randles equivalent circuit are given in Table I. Here Rs is the solution phase resistance, Rct is the charge-transfer resistance, and Cd is the differential capacitance. The introduction of a constant phase element was unnecessary. Initial formation of a SAM from 10-undecenoic acid results in an extremely high charge-transfer resistance 共Rct兲. However, the charge-transfer resistance is reduced by immobilization of the NHSS layer, because its terminal carboxylate groups are electrochemically active. The charge-transfer resistance increases again with immobilization of the antibody to peanut protein Ara h1. The change in the impedance spectra upon introduction of in-



J352

700 650

500

600 550 500

-ZIm (kΩ-cm )

450 2

2

-ZIm (kΩ-cm )

400

300

200

400 350 300 250 200 150

100

100 50 0

0 0

100

200

300

400

500

600

0

100

200

creasing concentrations of peanut protein Ara h 1 is shown in Fig. 4. The best-fit CNLS results for these results fit to a standard Randles equivalent circuit are given in Table II. The variations in both the differential capacitance 共Cd兲 and charge-transfer resistance 共Rct兲 with Ara h 1 concentration appear to be physically reasonable. As the average protein film thickness increases due to further protein binding, the effective capacitance decreases with the increase in polymer film thickness. In addition, as the polymer film thickness increases, the charge-transfer resistance 共Rct兲 increases as the degenerate Si electrode becomes more inaccessible to charge transfer. The relative change in Rct is much greater than the relative change in Cd, as is often seen in impedance studies of protein binding.25 Although the detection limit was not quantitatively determined, the lowest concentration studied here is 0.005 ␮g/mL, which corresponds to 0.08 nM. Several advantages are expected for degenerate Si electrodes relative to other electrode materials for electrochemical biosensors. For example, although direct comparison is difficult, the organic linker chemistry used here on degenerate Si should be more stable than the more popular thiol linker chemistry on Au. Oxidation of alkanethiol SAMs on Au has been shown to occur rapidly in ambient conditions.26,27 Because Si is directly beneath C in the periodic table, the stability of the Si–C–C bond chain is expected to be considerably higher than that of the Au–S–C bond chain. The organic linker chemistry employed here to form Si–C covalent bonds is stable for at least 1 week. The bond energy of the Si–C bond in Si–CH3 is about 520 kJ/mol, while that of the Au–S bond ranges from 125 to 150 kJ/mol, depending on the length of the hydrocarbon chain.28 In addition, recent atomic force microscopy studies report that the Si–C bond ruptures at an applied force of 2.0 ⫾ 0.3 nN while the Au–S bond ruptures at 1.4 ⫾ 0.3 nN.29 Immobilization of biomolecules onto carbon electrodes may provide excellent stability as well.30,31 However, carbon electrodes of various types have the well-known drawback of exhibiting complex electrochemistry that depends on the type of carbon, surface preparation, and on chemical treatment.32-34 Furthermore, the most critical advantage of degenerate Si electrodes relative to carbon electrodes is much easier incorporation into ultralarge-scale integrated technology, which is Si-based. Another advantage expected for Si electrodes relative to Au electrodes is that they should be self-passivating, because Si surface sites that do not form Si–C bonds should rapidly oxidize. In other words, exposed Si atoms, unlike Au, will rapidly oxidize and become electrochemically inert, except in the presence of both high

400

500

600

700

ZRe (kΩ-cm )

ZRe (kΩ-cm ) Figure 3. Impedance spectra obtained after modifying the degenerate Si electrode with 10-undecenoic acid 共䊐兲, 11 − MUA + NHSS 共⫹兲, and anti Ara h1 at 0 mV vs Ag/AgCl 共1 M KCl兲共䊊兲. The test solution also contains 50 mM PBS and 5 mM K3Fe共CN兲6 /K4Fe共CN兲6 at pH 7.3.

300

2

2

Figure 4. Impedance spectra obtained following addition of 0 共䊏兲, 0.005 ␮g/mL 共⽧兲, 0.010 ␮g/mL 共䉴兲, 0.015 ␮g/mL 共䉳兲, 0.020 ␮g/mL 共䉲兲, 0.040 ␮g/mL 共䉱兲, and 0.080 ␮g/mL 共쎲兲 peanut protein Ara h1 in a test solution that also contains 50 mM PBS and 5 mM K3Fe共CN兲6 /K4Fe共CN兲6 at pH 7.3.

fluoride concentrations and low pH, which would rapidly denature proteins anyway. Evidence for this argument is provided by the quite high charge-transfer resistance 共9.51 ⫻ 105 ⍀ cm2兲 reported here for 10-undeceneoic acid on degenerate Si. Such high values for the charge-transfer resistance are not typically observed for Au–S SAMs. The authors are aware of only one research group, Hamers and co-workers, that has reported the use of Si as a substrate for impedance biosensing.35,36 They have reported impedance detection of antibody–antigen interactions and DNA hybridization on both n- and p-type Si. They had to use more elements in their equivalent circuit than the current study to account for the resistance and capacitance of the space-charge layer. They also observed somewhat different sensor results on n- and p-type Si, complicating their understanding of sensor performance. The use of a degenerate Si electrode allows a simpler equivalent circuit analysis, easier electrical connection to the working electrode, and knowledge that the applied potential is dropped within the electrical double layer, not within the semiconductor space-charge layer. One might argue that impedance biosensors can be constructed on n- and p-type Si electrodes by operating them under accumulation conditions, which are characterized by injection of majority carriers into the space-charge region. While this is true, this causes serious complications, particularly for practitioners unfamiliar with semiconductor physics. Two recent studies that utilize electrochemical impedance spectroscopy to study n- and p-type Si illustrate some of the complexities involved.37,38 Studying p-type Si under accumulation conditions requires application of a significant anodic potential, which causes difficulties in many electrolytes given the strong reactivity of the Si electrode. Obtaining accumulation conditions on n-type Si requires applications of a significant cathodic potential, where hydrogen evolution is often the dominant reaction.

Table I. Impedance parameters for a Randles equivalent circuit fit to the impedance data from Fig. 3. SAM

NHSS

Antibody

38.0 共0.5兲 37.3 共0.4兲 37.7 共0.5兲 Rs 共⍀ cm 兲 2.39 共0.05兲 2.35 共0.01兲 2.40 共0.02兲 Cd 共␮F/cm2兲 Rct 共⍀ cm2兲 9.51 共0.18兲 ⫻ 105 5.28 共0.07兲 ⫻ 105 6.72 共0.13兲 ⫻ 105 2



J353

Table II. Impedance parameters for a Randles equivalent circuit fit to the impedance data from Fig. 4. Ara h1 concentration

0 共␮g/mL兲

0.005 共␮g/mL兲

0.01 共␮g/mL兲

0.015 共␮g/mL兲

0.02 共␮g/mL兲

0.04 共␮g/mL兲

0.08 共␮g/mL兲

Rs 共⍀ cm2兲 Cd 共␮F/cm2兲 Rct 共⍀ cm2兲

37.7 共0.5兲 2.40 共0.02兲 6.72共0.13兲 ⫻ 105

38.8 共0.1兲 2.39 共0.01兲 7.51共0.11兲 ⫻ 105

38.8 共0.1兲 2.39 共0.01兲 8.06共0.13兲 ⫻ 105

38.6 共0.1兲 2.39 共0.02兲 8.51共0.20兲 ⫻ 105

38.8 共0.1兲 2.38 共0.02兲 8.97共0.21兲 ⫻ 105

38.9 共0.1兲 2.39 共0.02兲 9.34共0.23兲 ⫻ 105

38.9 共0.1兲 2.39 共0.02兲 10.2共0.3兲 ⫻ 105

As a result, one cannot guarantee a priori that a surfaceimmobilized biomolecule will be stable atop either an n- or p-type Si electrode under accumulation conditions, because biomolecules may denature with application of an extreme potential. In addition, one cannot always easily discern whether an n- or p-type Si electrode is under accumulation conditions due to ambiguities in analyzing capacitance measurements, or measurements using other analytical methods.38 Given the complications reported by Hamers and co-workers for impedance biosensors on n- and p-type Si electrodes, the current suggestion of using degenerate Si electrodes seems to be a highly significant advantage. Another advantage of using a degenerate Si electrode relative to a semiconducting Si electrode is that the electrode behavior is not light-sensitive. For n- and p-type Si electrodes under light illumination, minority charge carriers are injected and complicate the charge transfer at the interface. Also, photogenerated holes in n-type Si are highly energetic and might conceivably denature an immobilized protein; however, recombination in degenerate Si will be quite rapid. In addition, any requirement for light illumination during biosensor operation is highly undesirable because this constrains possible applications. Careful inspection of the results in Table II suggests that the surface-immobilized antibody film is becoming saturated with peanut protein Ara h 1 at high protein concentration. In other words, the incremental change in Rct and Cd per unit increase in peanut protein concentration is reduced at high protein concentration. This information can be used quantitatively to further understand the antigen– antibody interaction during immobilization onto degenerate Si by estimating the dissociation constant. The dissociation reaction for the allergen 共A兲 and its antibody 共Y兲 is given below for the equilibrium constant Kd AY → A + Y

关1兴

Kd =

Kd =

冉冊

1−␪ 关A兴 ␪

关3兴

The dissociation constant can then be obtained by relating either the charge-transfer resistance 共Rct兲 or differential capacitance 共Cd兲 to the surface coverage. For fundamental reasons, the capacitance is a better choice, because protein-bound and unbound surface sites can be considered to be parallel circuit elements, and only capacitance elements in parallel combine in an additive manner. However, this analysis does not appear to be meaningful due to the small capacitance changes seen in Table II and the scatter in these results. Instead, this analysis will use Rct, because these results from Table II show both a larger change with increasing peanut protein concentration and less scatter. Assuming a Langmuir adsorption isotherm and a linear relationship between the surface coverage 共␪兲 and R−1 ct −1 −1 ⌬Rct = ␪共⌬Rct 兲max

关4兴

−1 −1 −1 −1 = 共Rct 兲␪=0 − Rct /共Rct 兲␪=0 ⌬Rct

关5兴

−1 −1 −1 −1 ⌬共Rct 兲max = 共Rct 兲␪=0 − 共Rct 兲␪=1/共Rct 兲␪=0

关6兴

where and The unitless change in Rct should be put into the Hanes–Woolf form to avoid overweighting of the data at low protein concentrations39 关A兴

=

关A兴 −1 共⌬Rct 兲max

+

Kd −1 共⌬Rct 兲max

关7兴

The corresponding plot is given in Fig. 5. The dissociation constant is calculated by dividing the intercept by the slope, which is 0.014 ␮g/mL, which corresponds to 0.25 nM. This is reasonably close to the value 共0.52 nM兲 obtained on a Au electrode.11

0.25

Conclusions

0.20

0.15

-1

[A]/∆Rct (µg/ml)

关2兴

Assuming the surface coverage of the antibody–antigen complex is ␪, the surface coverage of unbound antibody will be 1 − ␪, so

−1 ⌬Rct

0.10

0.05 0.00

关A兴关Y兴关AY兴

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

[A] (µg/ml) Figure 5. Hanes–Woolf plot for determining the dissociation constant 共Kd兲 between the antibody and antigen.

CV of bare 共nonoxidized兲 degenerate Si in solutions of NH4F and K3Fe共CN兲6 /K4Fe共CN兲6 exhibits reversible electrochemistry, with a peak separation of about 340 mV, a phenomenon not previously reported for Si electrodes in an aqueous environment. The presence of F− prevents formation of the Si native oxide, while the use of a degenerate Si substrate makes its electrochemical behavior more like that of a metal. In other words, the applied potential is dropped across the electrical double layer, not across a space-charge layer. In addition, impedance detection of peanut protein Ara h 1 is demonstrated on a degenerate Si electrode following immobilization of the monoclonal mouse antibody to this protein. This immobilization procedure includes covalent Si–C bond formation through alkene insertion into a Si–H bond, forming a more stable biosensor interface than the commonly used Au–S covalent bond formation. In addition, the degenerate Si electrode is “self-passivating” due to


J354


oxidation of bare Si sites that do not react during electrode preparation. The concentration of peanut protein Ara h 1 causes a moderate decrease in the differential capacitance and a much larger relative increase in the charge-transfer resistance. The detection limit is estimated to be 0.08 nM. In addition, the variation in charge-transfer resistance with protein concentration is employed to estimate the dissociation constant of the surface-immobilized antibody–antigen complex as approximately 0.25 nM. Acknowledgment This research was supported by a U.S. Army grant no. W911NF05-1-0339. Clarkson University assisted in meeting the publication costs of this article.

References 1. A. L. Ghindilis, P. Atanasov, M. Wilkins, and E. Wilkins, Biosens. Bioelectron., 13, 113 共1998兲. 2. M. Mir, M. Vreeke, and I. Katakis, Electrochem. Commun., 8, 505 共2006兲. 3. R. H. Lai, K. W. Plaxco, and A. J. Heeger, Anal. Chem., 79, 229 共2007兲. 4. G. S. Wilson and Y. Hu, Chem. Rev. (Washington, D.C.), 100, 2693 共2000兲. 5. E. Katz and I. Willner, Electroanalysis, 15, 913 共2003兲. 6. J. M. Buriak, Chem. Rev. (Washington, D.C.), 102, 1271 共2002兲. 7. R. Boukherroub, Curr. Opin. Solid State Mater. Sci., 9, 66 共2005兲. 8. S. Palacin, C. Bureau, J. Charlier, G. Deniau, B. Mouanda, and P. Viel, ChemPhysChem, 5, 1468 共2004兲. 9. J. Pinson and F. Padvorica, Chem. Soc. Rev., 34, 429 共2005兲. 10. G. Oskam, J. G. Long, A. Natarajan, and P. C. Searson, J. Phys. D, 31, 1927 共1998兲. 11. Y. Huang, M. C. Bell, and I. I. Suni, Anal. Chem., Accepted. 12. R. Voicu, R. Boukherroub, V. Bartzoka, T. Ward, J. T. C. Wojtyk, and D. D. M. Wayner, Langmuir, 20, 11713 共2004兲. 13. B. Johnsson, S. Lofas, and G. Lindquist, Anal. Biochem., 198, 268 共1991兲. 14. S. M. Sze, Semiconductor Devices, Physics and Technology, 2nd ed., John Wiley & Sons, New York 共2001兲. 15. X. G. Zhang, Electrochemistry of Silicon and its Oxide, Chapter 6, Kluwer, Norwell, MA 共2001兲.

16. E. S. Kooji, R. W. Despo, F. P. J. Mulders, and J. J. Kelly, J. Electroanal. Chem., 406, 139 共1996兲. 17. E. A. Dalchiele, A. Aurora, G. Bernardini, F. Cattaruzza, A. Flamini, P. Pallavicini, R. Zanoni, and F. Decker, J. Electroanal. Chem., 579, 133 共2005兲. 18. R. Zanoni, F. Cattaruzza, C. Coluzza, E. A. Dalchiele, F. Decker, G. DiSanto, A. Flamini, L. Funari, and A. G. Marrani, Surf. Sci., 575, 260 共2005兲. 19. R. Zanoni, A. Aurora, F. Cattaruzza, E. A. Dalchiele, F. Decker, G. DiSanto, A. Flamini, L. Funari, and A. G. Marrani, Mater. Sci. Eng., C, 26, 840 共2006兲. 20. A. G. Marrani, E. A. Dalchiele, R. Zanoni, F. Decker, F. Cattaruzza, D. Bonifazi, and M. Prato, Electrochim. Acta, 53, 3903 共2008兲. 21. R. M. Stiger, S. Gorer, B. Craft, and R. M. Penner, Langmuir, 15, 790 共1999兲. 22. B. Rashkova, B. Guel, R. T. Potzschke, G. Staikov, and W. J. Lorenz, Electrochim. Acta, 43, 3021 共1998兲. 23. Y. A. Ivanova, D. K. Ivanou, and E. A. Streltsov, Electrochem. Commun., 9, 599 共2007兲. 24. X. Li and I. S. Nandhakumar, Electrochem. Commun., 10, 363 共2008兲. 25. I. I. Suni, TrAC, Trends Anal. Chem., In press. 26. M. H. Schoenfisch and J. E. Pemberton, J. Am. Chem. Soc., 120, 4502 共1998兲. 27. T. M. Willey, A. L. Vance, T. van Buren, C. Bostedt, L. J. Terminello, and C. S. Fadley, Surf. Sci., 576, 188 共2005兲. 28. Y. R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press, Boca Raton, FL 共2007兲. 29. M. Grandbois, M. Beyer, M. Rief, H. C. Schaumann, and H. E. Gaub, Science, 283, 1727 共1999兲. 30. R. Blankespoor, B. Limoges, B. Schollhorn, J. L. Syssa-Magale, and D. Yazidi, Langmuir, 21, 3362 共2005兲. 31. H. F. Teh, H. Gong, X. D. Dong, X. Zeng, A. L. K. Tan, X. Yang, and S. N. Tan, Anal. Chim. Acta, 551, 23 共2005兲. 32. P. Chen and R. L. McCreery, Anal. Chem., 68, 3958 共1996兲. 33. P. Ramesh and S. Sampath, Anal. Chem., 75, 6949 共2003兲. 34. C. E. Banks and R. G. Compton, Anal. Sci., 21, 1263 共2005兲. 35. W. Cai, J. R. Peck, D. W. van der Weide, and R. J. Hamers, Biosens. Bioelectron., 19, 1013 共2004兲. 36. W. Yang, J. E. Butler, J. N. Russell, and R. J. Hamers, Analyst (Cambridge, U.K.), 132, 296 共2007兲. 37. M. Chemla, J. F. Dufreche, I. Darolles, F. Rouelle, D. Devilliers, S. Petitdidier, and D. Levy, Electrochim. Acta, 51, 665 共2005兲. 38. D. J. Michalak, F. Gstrein, and N. S. Lewis, J. Phys. Chem. C, 111, 16516 共2007兲. 39. M. L. Schuler and F. Kargi, Bioprocess Engineering, 2nd ed., p. 65, Prentice-Hall, Upper Saddle River, NJ 共2002兲.
