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

Degenerate Si as an Electrode Material for ElectrochemicalBiosensorsYin 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 heavily doped Si as an electrode material for electrochemical biosensors is demonstrated by a series ofvoltammetry and impedance studies. Cyclic voltammetry in solutions of NH4F and K3FeCN6/K4FeCN6 exhibits reversibleelectrochemistry without illumination, which to the best of the authors’ knowledge has not been previously reported for Sielectrodes in an aqueous environment. The presence of F− prevents formation of the Si native oxide, while the use of a degenerateSi substrate makes its electrochemistry more like that of a metal. In addition, impedance detection of peanut protein Ara h 1 isdemonstrated on a degenerate Si electrode following immobilization of the monoclonal mouse antibody to this protein. Thevariation in charge-transfer resistance with protein concentration is employed to estimate the dissociation constant of the surface-immobilized antibody-antigen complex as 0.25 nM. The advantages of degenerate Si relative to n- and p-type Si include the lackof photoeffects, a simpler equivalent circuit, easier electrical connection to the working electrode, and knowledge that the appliedpotential 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. DOI: 10.1149/1.2988047 All rights reserved.

Manuscript submitted June 5, 2008; revised manuscript received August 25, 2008. Published October 7, 2008.

0013-4651/2008/15512/J350/5/$23.00 © The Electrochemical Society

Due to the selectivity associated with biorecognition processes,biosensors for detection of numerous analytes have recently been anarea of intense research focus. Biosensor applications include real-time in vivo monitoring of biological processes, clinical measure-ments of biomedical disease markers, real-time industrial and envi-ronmental process monitoring, detection of biological warfareagents, and detection of food allergens. While optical biosensorshave been widely studied, electrochemical biosensors have someunique advantages, including the ability to incorporate opticallyopaque materials, including porous materials, avoiding false posi-tives arising from contaminating chromophores, and easierminiaturization.1-3 In addition, for portable and implantable applica-tions, they have the additional advantages of simplicity and lowernoise levels, because room-temperature optical detectors have rela-tively high noise levels.

The electrode materials in electrochemical biosensors are con-strained by the requirements for both high electrical conductivityand biocompatibility, because biomolecules often denature with pro-longed exposure to metal surfaces. To date, this has limited theelectrode materials in electrochemical biosensors primarily to Au,Pt, and some form of carbon.4,5 Si has been widely employed as abiosensor substrate material, but in the vast majority of applications,biomolecules have been immobilized onto SiO2 rather than Si, pre-venting electrochemical interrogation. Recently, immobilization ofbiomolecules onto n- or p-type Si has been reported through directformation of a Si–C bond by either alkene/alkyne insertion intoSi–H bonds6,7 or cathodic electrografting.8,9 However, electrochem-istry on n- and p-type Si is complicated by the partitioning of theapplied potential between the semiconductor space-charge layer andthe electrical double layer at the electrode/electrolyte interface.10 Inmany such cases, electrochemical processes may be irreversible dueto Schottky barrier formation at the semiconductor-electrolyteinterface.10

Here we demonstrate an electrochemical impedance biosensorthat employs degenerate Si as the electrode material. Degenerate Sicontains extremely high dopant concentrations and behaves like anelectrically conducting, not semiconducting, electrode. This imped-ance biosensor is demonstrated for detection of peanut protein Ara h1 following surface immobilization of its monoclonal mouse anti-body. This was also recently demonstrated in our laboratory using a

* Electrochemical Society Active Member.z E-mail: isuni@clarkson.edu

Downloaded 15 Oct 2008 to 128.153.14.49. Redistribution subject to E

Au electrode.11 Peanut protein Ara h 1 is a common food allergen,so detection of this species can address a serious public health prob-lem.

Experimental

As-doped n-type degenerate Si111 wafers with a thickness of300 m and a diameter of 50 mm were purchased from VirginiaSemiconductor, Inc. The resistivity of these Si wafers is less than0.005 cm. 10-Undecenoic acid was purchased from Alfa;N-3-dimethylaminopropyl-N-ethylcarbodiimide hydrochlorideEDC, potassium dihydrogen phosphate, and dipotassium hydrogenphosphate were purchased from Sigma; N-hydroxysulfosuccinimidesodium salt NHSS was purchased from Pierce Biotechnology; pea-nut protein Ara h1 and anti-Ara h1 mouse monoclonal antibodywere purchased from Indoor Biotechnologies. All reagents wereused as received. The Ara h1 protein was labeled as the monomericform.

The degenerate Si electrode was cleaned in ethanol and waterand etched in 40 wt % NH4F to remove the native oxide. The de-generate Si electrode was then immersed immediately in 10% 10-undecenoic acid in deaerated toluene solution in a photochemicalreactor for 19 h where it was exposed to UV light.12 The UV lightsource used was a Way Too Cool model WTC 72L-110 that providesa long wavelength output at 352 nm. Photochemical activation isnecessary to increase the kinetics of alkene insertion into Si–Hbonds, which forms a covalent Si–C bond. The exposed carboxylicacid groups were then activated by immersion for 1 h into 75 mMEDC, 15 mM NHSS, and 50 mM phosphate buffer solution PBSat pH 7.3. Following carboxylate group activation, the antibody topeanut protein Ara h 1 was immobilized by immersing the electrodefor 1 h into a solution containing 50 g/mL antibody and 50 mMPBS at pH 7.3. This procedure immobilizes the antibody onto thedegenerate Si electrode by amide bond formation to amine groupson the protein surface.13 This sensor electrode can then be used todetect the peanut protein Ara h1 in solution. Figure 1 illustrates theinterface construction and detection scheme.

The degenerate Si electrode was used as the working electrode ina virgin Teflon three-electrode cell with a Pt spiral counter electrodeand an Ag/AgCl 1.0 M KCl reference electrode. Connection to theSi electrode was made to the back side using Cu tape. This is con-siderably easier than making an ohmic electrical contact to a Sisemiconductor electrode, which typically requires a Ga–In eutecticor other special alloy.14 Voltammetry measurements were performedusing an EG&G PAR 263A potentiostat. Impedance measurementswere performed using the same potentiostat coupled to a Solartron1250 frequency response analyzer over the frequency range

CS license or copyright; see http://www.ecsdl.org/terms_use.jsp

J351Journal of The Electrochemical Society, 155 12 J350-J354 2008 J351

0.05 Hz to 15 kHz with an ac probe amplitude of 5 mV. Each im-pedance spectrum takes about 2.5 min to acquire. The impedanceresults were obtained at a dc potential of 0 mV vs Ag/AgCl, whichis slightly cathodic to the open-circuit potential OCP about+50 mV in the electrolytes of interest. This procedure was fol-lowed to minimize the possibility of Si oxidation. The 10-undecenoic self-assembled monolayer SAM reached steady stateimmediately, in contrast to the 11-mercaptoundecanoic acid SAMformed on Au, which takes about 30 min to reach steady state, asmeasured by changes in the OCP.11

Results and Discussion

Before considering impedance studies of protein immobilizationon degenerate Si, this new substrate material was characterized bycyclic voltammetry CV. Figure 2 shows the voltammogram ob-tained for bare not oxidized degenerate Si in 50 mM PBS buffersolution and 5 mM K3FeCN6/K4FeCN6 at pH 7.3 in 10 wt %NH4F. This voltammogram shows reversible oxidation/reductionpeaks for K3FeCN6/K4FeCN6 with a peak separation of about340 mV, suggesting metal-like behavior of this electrode. To thebest of the authors’ knowledge, reversible hexacyanoferrateoxidation/reduction peaks have not been previously reported for anaqueous electrolyte on a Si electrode of any kind. For comparison,Fig. 2 also presents a voltammogram in the same electrolyte on ann-type Si electrode, showing no FeCN6

3−/4− oxidation/reductionpeaks.

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

(a)

(b)

Figure 1. Schematic illustration of the interface construction and detectionscheme for peanut protein Ara h 1: a sensor electrode with self-assemblyfilm and antibody immobilized and b detection of the protein Ara h1 by theantibody. The Y-shaped molecule represents the antibody and the orbitalshape represents protein Ara h1.

Downloaded 15 Oct 2008 to 128.153.14.49. Redistribution subject to E

FeCN63−/4−, whose electrochemistry is typically reversible on metal

electrodes, most reported voltammograms do not exhibit both anoxidation peak anodic scan and a reduction peak cathodic scan,as seen in Fig. 2. Reversible aqueous voltammetry on Si has typi-cally been observed primarily on illuminated Si electrodes.15 Be-cause illumination is a significant constraint for sensing applica-tions, reversible electrochemistry that does not depend onphotoeffects is highly desirable.

Few reports exist of reversible aqueous electrochemistry on Sielectrodes that are not illuminated. Two-electron oxidation/reductionpeaks have been reported for methyl viologen in an aqueous elec-trolyte for n-type Si electrode following oxide removal in HF, andthis was explained by arguing that the Si electrode was under accu-mulation conditions. 16 Reversible oxidation/reduction peaks havebeen reported for ferrocene and its derivates on n- and p-type Si, butonly through incorporation of the electroactive species into a surfacefilm by direct Si–C bond formation.17-20 Metal electrodepositiononto n-type Si has been widely studied by Searson and others, butelectrodissolution of noble metals Au, Pt, Ag, Cu from these sub-strates has not been observed due to Schottky barrier formation.10

Generally, during the cathodic sweep, the metal deposition peakoverlaps with hydrogen evolution, but no current peak is observedduring the anodic sweep. However, it should be noted that Penner’sgroup has reported at least partially reversible Ag electrodepositionon degenerate n-type Si100, although degenerate Si electrochem-istry was not thoroughly investigated.21 In addition, at least partiallyreversible deposition of less noble metals such as Pb has been dem-onstrated on n-type Si.22-24 In summary, reversible aqueous electro-chemistry such as that illustrated in Fig. 2 has not often been re-ported.

The impedance spectra obtained at different stages of degenerateSi electrode preparation are shown in Fig. 3. The best complexnonlinear least square CNLS fitting results for these spectra to astandard Randles equivalent circuit are given in Table I. Here Rs isthe solution phase resistance, Rct is the charge-transfer resistance,and Cd is the differential capacitance. The introduction of a constantphase element was unnecessary. Initial formation of a SAM from10-undecenoic acid results in an extremely high charge-transfer re-sistance Rct. However, the charge-transfer resistance is reduced byimmobilization of the NHSS layer, because its terminal carboxylategroups are electrochemically active. The charge-transfer resistanceincreases again with immobilization of the antibody to peanut pro-tein Ara h1.

The change in the impedance spectra upon introduction of in-

-0.8 -0.7 -0.6 -0.5 -0.4 -0.3 -0.2 -0.1 0.0 0.1 0.2 0.3-0.20

-0.15

-0.10

-0.05

0.00

0.05

0.10

0.15

0.20

j(mA/cm2 )

E (V vs. Ag/AgCl)

Figure 2. Voltammetric behavior of bare degenerate Si in 50 mM PBSbuffer solution and 5 mM K3FeCN6/K4FeCN6 in 10 wt % NH4F at ascan rate of 50 mV/s. The dashed line corresponds to voltammetry of n-typeSi = 2–35 cm for comparison.

CS license or copyright; see http://www.ecsdl.org/terms_use.jsp

5

J352 Journal of The Electrochemical Society, 155 12 J350-J354 2008J352

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 Randlesequivalent circuit are given in Table II. The variations in both thedifferential capacitance Cd and charge-transfer resistance Rctwith Ara h 1 concentration appear to be physically reasonable. Asthe average protein film thickness increases due to further proteinbinding, the effective capacitance decreases with the increase inpolymer film thickness. In addition, as the polymer film thicknessincreases, the charge-transfer resistance Rct increases as the degen-erate Si electrode becomes more inaccessible to charge transfer. Therelative change in Rct is much greater than the relative change in Cd,as is often seen in impedance studies of protein binding.25 Althoughthe detection limit was not quantitatively determined, the lowestconcentration studied here is 0.005 g/mL, which corresponds to0.08 nM.

Several advantages are expected for degenerate Si electrodesrelative to other electrode materials for electrochemical biosensors.For example, although direct comparison is difficult, the organiclinker chemistry used here on degenerate Si should be more stablethan the more popular thiol linker chemistry on Au. Oxidation ofalkanethiol SAMs on Au has been shown to occur rapidly in ambi-ent conditions.26,27 Because Si is directly beneath C in the periodictable, the stability of the Si–C–C bond chain is expected to be con-siderably higher than that of the Au–S–C bond chain. The organiclinker chemistry employed here to form Si–C covalent bonds isstable for at least 1 week. The bond energy of the Si–C bond inSi–CH3 is about 520 kJ/mol, while that of the Au–S bond rangesfrom 125 to 150 kJ/mol, depending on the length of the hydrocar-bon chain.28 In addition, recent atomic force microscopy studiesreport that the Si–C bond ruptures at an applied force of2.0 0.3 nN while the Au–S bond ruptures at 1.4 0.3 nN.29

Immobilization of biomolecules onto carbon electrodes may pro-vide excellent stability as well.30,31 However, carbon electrodes ofvarious types have the well-known drawback of exhibiting complexelectrochemistry that depends on the type of carbon, surface prepa-ration, and on chemical treatment.32-34 Furthermore, the most criticaladvantage of degenerate Si electrodes relative to carbon electrodesis much easier incorporation into ultralarge-scale integrated technol-ogy, which is Si-based.

Another advantage expected for Si electrodes relative to Au elec-trodes is that they should be self-passivating, because Si surfacesites that do not form Si–C bonds should rapidly oxidize. In otherwords, exposed Si atoms, unlike Au, will rapidly oxidize and be-come electrochemically inert, except in the presence of both high

0 100 200 300 400 500 6000

100

200

300

400

500

-ZIm(kΩ-cm2 )

ZRe (kΩ-cm2)

Figure 3. Impedance spectra obtained after modifying the degenerate Sielectrode with 10-undecenoic acid , 11 − MUA + NHSS , and antiAra h1 at 0 mV vs Ag/AgCl 1 M KCl . The test solution also contains50 mM PBS and 5 mM K3FeCN6/K4FeCN6 at pH 7.3.

Downloaded 15 Oct 2008 to 128.153.14.49. Redistribution subject to E

fluoride concentrations and low pH, which would rapidly denatureproteins anyway. Evidence for this argument is provided by the quitehigh charge-transfer resistance 9.51 105 cm2 reported herefor 10-undeceneoic acid on degenerate Si. Such high values for thecharge-transfer resistance are not typically observed for Au–SSAMs.

The authors are aware of only one research group, Hamers andco-workers, that has reported the use of Si as a substrate for imped-ance biosensing.35,36 They have reported impedance detection ofantibody–antigen interactions and DNA hybridization on both n- andp-type Si. They had to use more elements in their equivalent circuitthan the current study to account for the resistance and capacitanceof the space-charge layer. They also observed somewhat differentsensor results on n- and p-type Si, complicating their understandingof sensor performance. The use of a degenerate Si electrode allowsa simpler equivalent circuit analysis, easier electrical connection tothe working electrode, and knowledge that the applied potential isdropped within the electrical double layer, not within the semicon-ductor space-charge layer.

One might argue that impedance biosensors can be constructedon n- and p-type Si electrodes by operating them under accumula-tion conditions, which are characterized by injection of majoritycarriers into the space-charge region. While this is true, this causesserious complications, particularly for practitioners unfamiliar withsemiconductor physics. Two recent studies that utilize electrochemi-cal impedance spectroscopy to study n- and p-type Si illustrate someof the complexities involved.37,38 Studying p-type Si under accumu-lation conditions requires application of a significant anodic poten-tial, which causes difficulties in many electrolytes given the strongreactivity of the Si electrode. Obtaining accumulation conditions onn-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 circuitfit to the impedance data from Fig. 3.

SAM NHSS Antibody

Rs cm2 38.0 0.5 37.3 0.4 37.7 0.5Cd F/cm2 2.39 0.05 2.35 0.01 2.40 0.02R cm2 9.51 0.18 105 5.28 0.07 105 6.72 0.13 10

0 100 200 300 400 500 600 700050100150200250300350400450500550600650700

-ZIm(k

Ω-cm2 )

ZRe (kΩ-cm2)

Figure 4. Impedance spectra obtained following addition of 0 ,0.005 g/mL , 0.010 g/mL , 0.015 g/mL , 0.020 g/mL ,0.040 g/mL , and 0.080 g/mL peanut protein Ara h1 in a testsolution that also contains 50 mM PBS and 5 mM K3FeCN6/K4FeCN6 atpH 7.3.

ct

CS license or copyright; see http://www.ecsdl.org/terms_use.jsp

J353Journal of The Electrochemical Society, 155 12 J350-J354 2008 J353

As a result, one cannot guarantee a priori that a surface-immobilized biomolecule will be stable atop either an n- or p-typeSi electrode under accumulation conditions, because biomoleculesmay denature with application of an extreme potential. In addition,one cannot always easily discern whether an n- or p-type Si elec-trode is under accumulation conditions due to ambiguities in analyz-ing capacitance measurements, or measurements using other analyti-cal methods.38 Given the complications reported by Hamers andco-workers for impedance biosensors on n- and p-type Si electrodes,the current suggestion of using degenerate Si electrodes seems to bea highly significant advantage.

Another advantage of using a degenerate Si electrode relative toa semiconducting Si electrode is that the electrode behavior is notlight-sensitive. For n- and p-type Si electrodes under light illumina-tion, minority charge carriers are injected and complicate the chargetransfer at the interface. Also, photogenerated holes in n-type Si arehighly energetic and might conceivably denature an immobilizedprotein; however, recombination in degenerate Si will be quite rapid.In addition, any requirement for light illumination during biosensoroperation is highly undesirable because this constrains possible ap-plications.

Careful inspection of the results in Table II suggests that thesurface-immobilized antibody film is becoming saturated with pea-nut protein Ara h 1 at high protein concentration. In other words, theincremental change in Rct and Cd per unit increase in peanut proteinconcentration is reduced at high protein concentration. This infor-mation can be used quantitatively to further understand the antigen–antibody interaction during immobilization onto degenerate Si byestimating the dissociation constant. The dissociation reaction forthe allergen A and its antibody Y is given below for the equilib-rium constant Kd

AY → A + Y 1

Table II. Impedance parameters for a Randles equivalent circuit fit

Ara h1concentration

0g/mL

0.005g/mL

0.01g/mL

Rs cm2 37.7 0.5 38.8 0.1 38.8 0.1Cd F/cm2 2.40 0.02 2.39 0.01 2.39 0.01Rct cm2 6.720.13 105 7.510.11 105 8.060.13 10

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09

0.05

0.10

0.15

0.20

0.25

[A]/∆R-1 ct(µg/ml)

[A] (µg/ml)

Figure 5. Hanes–Woolf plot for determining the dissociation constant Kdbetween the antibody and antigen.

Downloaded 15 Oct 2008 to 128.153.14.49. Redistribution subject to E

Kd =AYAY

2

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

Kd = 1 −

A 3

The dissociation constant can then be obtained by relating either thecharge-transfer resistance Rct or differential capacitance Cd tothe surface coverage. For fundamental reasons, the capacitance is abetter choice, because protein-bound and unbound surface sites canbe considered to be parallel circuit elements, and only capacitanceelements in parallel combine in an additive manner. However, thisanalysis does not appear to be meaningful due to the small capaci-tance changes seen in Table II and the scatter in these results. In-stead, this analysis will use Rct, because these results from Table IIshow both a larger change with increasing peanut protein concentra-tion and less scatter.

Assuming a Langmuir adsorption isotherm and a linear relation-ship between the surface coverage and Rct

−1

Rct−1 = Rct

−1max 4where

Rct−1 = Rct

−1=0 − Rct−1/Rct

−1=0 5and

Rct−1max = Rct

−1=0 − Rct−1=1/Rct

−1=0 6

The unitless change in Rct should be put into the Hanes–Woolf formto avoid overweighting of the data at low protein concentrations39

ARct

−1 =A

Rct−1max

+Kd

Rct−1max

7

The corresponding plot is given in Fig. 5. The dissociation constantis calculated by dividing the intercept by the slope, which is0.014 g/mL, which corresponds to 0.25 nM. This is reasonablyclose to the value 0.52 nM obtained on a Au electrode.11

Conclusions

CV of bare nonoxidized degenerate Si in solutions of NH4Fand K3FeCN6/K4FeCN6 exhibits reversible electrochemistry,with a peak separation of about 340 mV, a phenomenon not previ-ously reported for Si electrodes in an aqueous environment. Thepresence of F− prevents formation of the Si native oxide, while theuse of a degenerate Si substrate makes its electrochemical behaviormore like that of a metal. In other words, the applied potential isdropped across the electrical double layer, not across a space-chargelayer.

In addition, impedance detection of peanut protein Ara h 1 isdemonstrated on a degenerate Si electrode following immobilizationof the monoclonal mouse antibody to this protein. This immobiliza-tion procedure includes covalent Si–C bond formation through alk-ene insertion into a Si–H bond, forming a more stable biosensorinterface than the commonly used Au–S covalent bond formation. Inaddition, the degenerate Si electrode is “self-passivating” due to

impedance data from Fig. 4.

0.015g/mL

0.02g/mL

0.04g/mL

0.08g/mL

38.6 0.1 38.8 0.1 38.9 0.1 38.9 0.12.39 0.02 2.38 0.02 2.39 0.02 2.39 0.02

.510.20 105 8.970.21 105 9.340.23 105 10.20.3 105

to the

5 8

CS license or copyright; see http://www.ecsdl.org/terms_use.jsp

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

oxidation of bare Si sites that do not react during electrode prepara-tion. The concentration of peanut protein Ara h 1 causes a moderatedecrease in the differential capacitance and a much larger relativeincrease in the charge-transfer resistance. The detection limit is es-timated to be 0.08 nM. In addition, the variation in charge-transferresistance with protein concentration is employed to estimate thedissociation constant of the surface-immobilized antibody–antigencomplex as approximately 0.25 nM.

Acknowledgment

This research was supported by a U.S. Army grant no. W911NF-05-1-0339.

Clarkson University assisted in meeting the publication costs of this ar-ticle.

References1. A. L. Ghindilis, P. Atanasov, M. Wilkins, and E. Wilkins, Biosens. Bioelectron.,

13, 113 1998.2. M. Mir, M. Vreeke, and I. Katakis, Electrochem. Commun., 8, 505 2006.3. R. H. Lai, K. W. Plaxco, and A. J. Heeger, Anal. Chem., 79, 229 2007.4. G. S. Wilson and Y. Hu, Chem. Rev. (Washington, D.C.), 100, 2693 2000.5. E. Katz and I. Willner, Electroanalysis, 15, 913 2003.6. J. M. Buriak, Chem. Rev. (Washington, D.C.), 102, 1271 2002.7. R. Boukherroub, Curr. Opin. Solid State Mater. Sci., 9, 66 2005.8. S. Palacin, C. Bureau, J. Charlier, G. Deniau, B. Mouanda, and P. Viel,

ChemPhysChem, 5, 1468 2004.9. J. Pinson and F. Padvorica, Chem. Soc. Rev., 34, 429 2005.

10. G. Oskam, J. G. Long, A. Natarajan, and P. C. Searson, J. Phys. D, 31, 19271998.

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 2004.13. B. Johnsson, S. Lofas, and G. Lindquist, Anal. Biochem., 198, 268 1991.14. S. M. Sze, Semiconductor Devices, Physics and Technology, 2nd ed., John Wiley &

Sons, New York 2001.15. X. G. Zhang, Electrochemistry of Silicon and its Oxide, Chapter 6, Kluwer, Nor-

well, MA 2001.

Downloaded 15 Oct 2008 to 128.153.14.49. Redistribution subject to E

16. E. S. Kooji, R. W. Despo, F. P. J. Mulders, and J. J. Kelly, J. Electroanal. Chem.,406, 139 1996.

17. E. A. Dalchiele, A. Aurora, G. Bernardini, F. Cattaruzza, A. Flamini, P. Pallavicini,R. Zanoni, and F. Decker, J. Electroanal. Chem., 579, 133 2005.

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 2005.

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 2006.

20. A. G. Marrani, E. A. Dalchiele, R. Zanoni, F. Decker, F. Cattaruzza, D. Bonifazi,and M. Prato, Electrochim. Acta, 53, 3903 2008.

21. R. M. Stiger, S. Gorer, B. Craft, and R. M. Penner, Langmuir, 15, 790 1999.22. B. Rashkova, B. Guel, R. T. Potzschke, G. Staikov, and W. J. Lorenz, Electrochim.

Acta, 43, 3021 1998.23. Y. A. Ivanova, D. K. Ivanou, and E. A. Streltsov, Electrochem. Commun., 9, 599

2007.24. X. Li and I. S. Nandhakumar, Electrochem. Commun., 10, 363 2008.25. I. I. Suni, TrAC, Trends Anal. Chem., In press.26. M. H. Schoenfisch and J. E. Pemberton, J. Am. Chem. Soc., 120, 4502 1998.27. T. M. Willey, A. L. Vance, T. van Buren, C. Bostedt, L. J. Terminello, and C. S.

Fadley, Surf. Sci., 576, 188 2005.28. Y. R. Luo, Comprehensive Handbook of Chemical Bond Energies, CRC Press,

Boca Raton, FL 2007.29. M. Grandbois, M. Beyer, M. Rief, H. C. Schaumann, and H. E. Gaub, Science,

283, 1727 1999.30. R. Blankespoor, B. Limoges, B. Schollhorn, J. L. Syssa-Magale, and D. Yazidi,

Langmuir, 21, 3362 2005.31. 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 2005.32. P. Chen and R. L. McCreery, Anal. Chem., 68, 3958 1996.33. P. Ramesh and S. Sampath, Anal. Chem., 75, 6949 2003.34. C. E. Banks and R. G. Compton, Anal. Sci., 21, 1263 2005.35. W. Cai, J. R. Peck, D. W. van der Weide, and R. J. Hamers, Biosens. Bioelectron.,

19, 1013 2004.36. W. Yang, J. E. Butler, J. N. Russell, and R. J. Hamers, Analyst (Cambridge, U.K.),

132, 296 2007.37. M. Chemla, J. F. Dufreche, I. Darolles, F. Rouelle, D. Devilliers, S. Petitdidier, and

D. Levy, Electrochim. Acta, 51, 665 2005.38. D. J. Michalak, F. Gstrein, and N. S. Lewis, J. Phys. Chem. C, 111, 16516 2007.39. M. L. Schuler and F. Kargi, Bioprocess Engineering, 2nd ed., p. 65, Prentice-Hall,

Upper Saddle River, NJ 2002.

CS license or copyright; see http://www.ecsdl.org/terms_use.jsp