Identification of Surface Heterogeneity Effects in Cyclic

Anal. Chem. 2008, 80, 3873–3881

Identification of Surface Heterogeneity Effects in Cyclic Voltammograms Derived from Analysis of an Individually Addressable Gold Array Electrode Chong-Yong Lee, Yong-Jun Tan,† and Alan M. Bond* School of Chemistry, Monash University, Clayton, Victoria 3800, Australia Voltammetric behavior at gold electrodes in aqueous media is known to be strongly dependent on electrode polishing and history. In this study, an electrode array consisting of 100 nominally identical and individually addressable gold disks electrodes, each with a radius of 127 µm, has been fabricated. The ability to analyze both individual electrode and total array performance enables microscopic aspects of the overall voltammetric response arising from variable levels of inhomogeneity in each electrode to be identified. The array configuration was initially employed with the reversible and hence relatively surface insensitive [Ru(NH3)6]3+/2+ reaction and then with the more highly surface sensitive quasi-reversible [Fe(CN)6]3-/4- process. In both these cases, the reactants and products are solution soluble and, at a scan rate of 50 mV s-1, each electrode in the array is assumed to behave independently, since no evidence of overlapping of the diffusion layers was detected. As would be expected, the variability of the individual electrodes’ responses was significantly larger than found for the summed electrode behavior. In the case of cytochrome c voltammetry at a 4,4′-dipyridyl disulfide modified electrode, a far greater dependence on electrode history and electrode heterogeneity was detected. In this case, voltammograms derived from individual electrodes in the gold array electrode exhibit shape variations ranging from peak to sigmoidal. However, again the total response was always found to be well-defined. This voltammetry is consistent with a microscopic model of heterogeneity where some parts of each chemically modified electrode surface are electroactive while other parts are less active. The findings are consistent with the common existence of electrode heterogeneity in cyclic voltammetric responses at gold electrodes, that are normally difficult to detect, but fundamentally important, as electrode nonuniformity can give rise to subtle forms of kinetic and other forms of dispersion. Classically, theories describing the Faradaic electron transfer processes in dynamic electrochemistry are based on the assump* Corresponding author. E-mail: [email protected]. Tel: +61 3 9905 1338. Fax: +61 3 9905 4597. † Current address: Guangzhou MatterSave Surface Technologies Co. Ltd, Guangzhou Technology Innovation Base, No. 80 Lan Yue Road, Guangzhou Science City, P. R. China 510663. 10.1021/ac8002227 CCC: $40.75  2008 American Chemical Society Published on Web 04/16/2008

tion that the electrode surface is homogeneous.1 Initial experimental studies used to validate the theory commonly employed liquid mercury electrodes where a truly homogeneous surface is indeed likely to be obtained. However, mercury electrodes are no longer widely employed in voltammetry. Rather, studies now commonly utilize solid metal, carbon or other conducting or semiconducting materials, which may even be chemically modified. Problems of electrode heterogeneity are likely to be even more acute when a chemically modified electrode is used instead of a bare electrode.2–4 Similarly, boron doped diamond and other doped electrodes also may have significant heterogeneities and polishing or other forms of pretreatment may induce further levels of nonuniformity. A challenge in using these modern electrode materials is to identify the presence of heterogeneity and establish to what extent available theories can be used to accommodate predictions derived from these nonhomogeneous surfaces,5–7 under circumstances where significant levels of kinetic, thermodynamic or other forms of dispersion may be present when two or more reactions may take place simultaneously at different parts of an electrode surface. The presence of heterogeneity on a solid electrode surface is usually detected by employing imaging techniques such as scanning electron microscopy, scanning tunneling microscopy, atomic force microscopy and optical microscopy. These methods can detect physical or surface topography differences at different parts of an electrode surface that might arise from the presence of surface oxide or the preferential adsorption of chemical species at defect sites. In principle, the results deduced from these images may be used to correlate the voltammetric response with the state of the surface. However, imaging techniques are commonly operated in a relatively specific and localized area, and most are applied ex situ rather than in situ during the course of the actual experiment. Thus, in many circumstances, the image does not necessarily represent the full details of the surface coverage present when the voltammetric experiment is undertaken. In the present study, heterogeneity associated with gold electrodes is of interest. The presence of heterogeneity at gold surfaces in contact with aqueous electrolytes has been studied for many years. Ertl8 and Somorjai9 have pointed out that atomic defects are present on such surfaces, including those formed with polycrystalline gold. The atoms at such defect sites exhibit unusual coordination numbers and have energetic and kinetic properties that differ from terrace atoms. The abnormal properties of surface (1) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamental and applications; John Wiley & Sons: New York, 2001.

Analytical Chemistry, Vol. 80, No. 10, May 15, 2008

3873

active sites are assumed to be a major contributor to the catalytic properties of oxide-supported gold nanoparticles.10,11 Burke et al.12 have explored the electrocatalytic properties of gold electrode in detail and developed a model of the electrode surface properties that explain why electrode performance is strongly dependent on electrode history, including polishing and cathodic pretreatment methods. These studies were performed with dc voltammetry, but more details of the surface and its relationship to electrode pretreatment have been recently elucidated by an ac method.13 Direct voltammetric detection of heterogeneity from a single macrodisk electrode experiment is difficult. However, in this paper it is proposed that a mimic of this situation will become available if an array electrode consisting of individually addressable macroelectrodes (diffusion overlap avoided) is fabricated. Under these conditions, where each electrode is nominally identical and pretreated in the same manner, a series of experiments may be effectively performed to examine the deviation in behavior of each individual electrode relative to the summed response obtained when all electrodes are simultaneously used in an experiment. Arrays of microelectrodes rather than macroelectrodes have been of interest for a number of years. These provide circumstances where a large and easily measured total current output is obtained, while retaining many of the advantageous features possessed by the individual microelectrodes. Photolithography, sputtering or screen printing methods are commonly employed to fabricate microarray electrodes.14–18 However, only in a few cases17,18 is individually addressable response available, as is needed to probe the details of the behavior of an array electrode. The major origin of heterogeneity at gold electrodes is probably introduced by polishing.13 In microelectrode arrays fabricated by lithographic procedures, prolonged use can result in a deterioration of the voltammogram due to the electrode fouling. Electrode surface renewal by polishing can readily damage the thin layer. Thus, commonly in order to achieve a reproducible (2) (3) (4) (5) (6) (7) (8)

(9) (10) (11) (12)

(13) (14) (15) (16) (17) (18)

Cox, J. A.; Tess, M. E.; Cummings, T. E. Anal. Chem. 1996, 15, 173. Anderson, J. L.; Coury, L. A.; Leddy, J. Anal. Chem. 2000, 72, 4497. Zen, J. M.; Kumar, A. S.; Tsai, D. M. Electroanalysis 2003, 15, 1073. Compton, R. G.; Banks, C. E. Understanding Voltammetry; World Scientific Publishing Co.: London, 2007. Davies, T. J.; Moore, R. R.; Banks, C. E.; Compton, R. G. J. Solid State Electrochem. 2005, 9, 797. Davies, T. J.; Moore, R. R.; Banks, C. E.; Compton, R. G. J. Electroanal. Chem. 2004, 574, 123. Ertl, G. Advances in Catalysis. In Impact of Surface Science on Catalysis; Gates, B. C., Knozinger, H., Eds.; Academic Press: New York, 2001; Vol. 45. Somorjai, G. A. Chem. Rev. 1996, 96, 1223. Haruta, M.; Yamada, N.; Kobayashi, T.; Iijima, S. J. Catal. 1989, 115, 301. Bond, G. C.; Thompson, D. T. Catal. Rev.sSci. Eng. 1999, 41, 319. (a) Burke, L. D. Gold Bull. 2004, 37, 125. (b) Burke, L. D.; Nugent, P. F. Gold Bull. 1997, 30, 43. (c) Burke, L. D.; Nugent, P. F. Gold Bull. 1998, 31, 39. (d) Burke, L. D.; Cunnane, V. J. J. Electroanal. Chem. 1986, 210, 69. Lertanantawong, B.; O’Mullane, A. P.; Surareungchai, W.; Somasundrum, M.; Burke, L. D.; Bond, A. M. Langmuir 2008, 24, 2856. Dobson, P. J.; Jiang, L.; Leigh, P. A.; Hill, H. A. O.; Kaneko, S. Adv. Mater. Opt. Electron. 1992, 1, 133. Thormann, W.; Van den Bosch, P.; Bond, A. M. Anal. Chem. 1985, 57, 2764. Zoski, C. G.; Simjee, N.; Guenat, O.; Koudelka-Hep, M. Anal. Chem. 2004, 76, 62. Priano, G.; González, G.; Günther, M.; Battaglini, F. Electronalysis 2008, 20, 91. Nagale, M. P.; Fritsch, Anal. Chem. 1998, 70, 2908.

3874


signal the potential of the working electrode is pulsed.19–21 Nevertheless, even after this electrochemical activation step, it is not always possible to recover the initial behavior.22 If the array electrode is derived from gold wires, physical polishing procedures to renew the electrode surface can be used in the usual manner. Thus, the gold array electrode of the kind developed in this study can be repetitively and rapidly repolished and used for numerous experiments. However, each individual electrode in the array configuration will not be subject to exactly identical levels of polishing and scratching, so variability in each element of the array is predicted that will parallel the situation prevailing in different regions of a single large area electrode. In this study, the assembly of an array of 100 gold electrodes, each with a radius of 127 µm, has been prepared in order to explore the role of electrode heterogeneity in the performance of a macrodisk gold electrode. The fabricated individual electrodes are sufficiently large that they exhibit close to linear diffusion, and each is sufficiently separated so that, with a suitable scan rate, overlap of diffusion layers can be essentially avoided. Furthermore, the individual electrodes are sufficiently small so that ohmic (iR) drop is minimal in studies in aqueous media.23,24 In principle, under these circumstances, the sum of each individual response should equal that produced when all elements in the array electrode are operational. Systems that have been studied to probe the influence of electrode heterogeneity at the gold array electrode are the reversible and hence relatively surface insensitive [Ru(NH3)6]3+/2+ process, the quasi-reversible and more highly surface sensitive [Fe(CN)6]3-/4- process and cytochrome c voltammetry at a 4,4′-dipyridyl disulfide chemically modified electrode, which is extremely surface sensitive. EXPERIMENTAL SECTION Chemicals. Cytochrome c (Sigma, 95% purity), 4,4′-dipyridyl disulfide (Sigma), poly-L-lysine (Sigma, 0.1% w/v in water, M.W. 150000–300000), [Ru(NH3)6]Cl3 (Strem Chemicals), K3[Fe(CN)6] (Aldrich) and KCl (AR. BDH) were used as received from the manufacturer. Deionized water (resistivity 18 MΩ cm) from a MilliQ-MilliRho purification system was used to prepare aqueous solutions. Procedures. Fabrication of the Gold Array Electrode. The individually addressable array electrode used in this work was made from 100 gold wires (Surepure Chemetals) of radius of 127 µm. Each electrode was embedded in epoxy resin (Ciba-Geigy Australia Limited) and separated by a thin epoxy layer. The fabrication of the array electrode was commenced by applying a thin layer of epoxy resin to fully coat 10 approximately 50 cm long clean gold wires. After drying at room temperature for 48 h, the coated gold wires were cut into 100 small wires, each having an (19) Welch, L. E.; Lacourse, W. R.; Mead, D. A.; Johnson, D. C. Anal. Chem. 1989, 61, 555. (20) Strein, T. G.; Ewing, A. G. Anal. Chem. 1993, 65, 1203. (21) Johnson, D. C.; LaCourse, W. R. Anal. Chem. 1990, 62, 589A. (22) González-García, O.; Ariño, C.; Díaz-Cruz, J. M.; Esteban, M. Electronalysis 2007, 19, 429. (23) Wipf, D. O.; Kristensen, E. W.; Deakin, M. R.; Wightman, R. W. Anal. Chem. 1988, 60, 306. (24) Howell, J. O.; Wightman, R. M. Anal. Chem. 1984, 56, 524. (25) (a) Tan, Y.-J. Corros. Sci. 1999, 41, 229. (b) Tan, Y.-J.; Bailey, S.; Kinsella, B. Corros. Sci. 2001, 43, 1905. (c) Tan, Y.-J.; Bailey, S.; Kinsella, B. Corros. Sci. 2001, 43, 1919. (d) Tan, Y.-J.; Bailey, S.; Kinsella, B. Corros. Sci. 2001, 43, 1931. (e) Aung, N. N.; Tan, Y.-J. Corros. Sci. 2004, 46, 3057.

exposed end with a disk-shape. Electrical contact to the potentiostat was made possible by soldering a copper wire to one of the ends of each wire and then insulating by covering the solder area with epoxy. Ten of the dried gold wires, each connected with copper wire, were arranged in a 1 × 10 linear array on a piece of glass slide covered with the sellotape. The gold wires’ positions were temporarily fixed with quick drying Araldite (Selleys, Australia). Epoxy resin was applied to surround the linear array electrode. Another glass slide was then carefully pressed against the first one, so as to sandwich the array of gold wires. The same procedure was applied to the other 9 sets of 1 × 10 linear array electrodes. After slow drying for 48 h, the glass slides were removed and excess epoxy on the edge of the array was carefully removed with a sharp blade. The 10 sets of 1 × 10 array electrodes were then brought together with epoxy resin. The fully dried 10 × 10 array electrode was nestled into a cylindrical container, and a large amount of epoxy resin was introduced to cover the entire electrode, including the array electrode surface. The surface of the dried array electrode was then polished extensively with sand paper to ensure that all the disk shaped gold surfaces were exposed. A conductivity meter was used to monitor the extent of insulation, and hence confirm the removal of epoxy. Finally, each electrode was connected in a known position to the potentiostatted instrumention to give the arrangement shown in Figure 1a,b. The

inset to Figure 1b shows that the average electrode spacing was typically close to the diameter of a single gold wire, although the spacing was not fully uniform. Experimental Arrangement. The full experimental arrangement used in this study, consists of a three-electrode potentiostatted (BAS model 100B electrochemical workstation) electrochemical cell containing the array working electrode, a Ag/AgCl (3 M NaCl) reference electrode and a platinum wire auxiliary electrode (Figure 1a). The reference electrode was positioned directly below the middle of the array electrode to provide as uniform a distribution of ohmic (iRu) drop as possible. Full details of the instrumentation, which was originally designed for corrosion monitoring purposes, is available in ref 25. For the purposes of this study, the autoswitching system could be used to control and activate any individual electrode or any combination of electrodes. Voltammograms were initially obtained for the full array electrode and then for each individual electrode, and finally again for the full array to ensure that the array electrode response has remained stable for the duration of an experiment. The areas of the gold wire electrodes were determined by comparison of equivalent voltammograms for the reversible oneelectron reduction of 1.00 mM [Ru(NH3)6]3+ in 0.50 M aqueous KCl solution with simulations derived from the use of DigiSim software,26 which is based on the hemispherical diffusion approximation model.27,28 The diffusion coefficients used in these and other simulations were as follows: 7.6 × 10-6 cm2 s-1 and 7.8 × 10-6 cm2 s-1 for [Ru(NH3)6]3+ and [Ru(NH3)6]2+,29,30 respectively ; 7.6 × 10-6 cm2 s-1 and 6.3 × 10-6 cm2 s-1 for [Fe(CN)6]3and [Fe(CN)6]4- species,1 respectively; and 9.4 × 10-7 cm2 s-1 for cytochrome c in both oxidized and reduced forms.31 The total geometric area (including insulation) occupied by the full array electrode was approximately 0.071 cm2, while the total gold area was approximately 0.051 cm2 (100 × 0.00051 cm2). The geometric area of an individual 127 µm radius electrode is 5.07 × 10-4 cm2. All voltammetric experiments were undertaken at 20 ± 1 °C, and in order to remove oxygen, the solutions were degassed with highpurity nitrogen for at least 5 min prior to commencing electrochemical experiments. Before each voltammetric experiment, the gold array electrode was polished several times with 0.3 µm alumina slurry on a polishing cloth (Buehler). After each polish, the electrode was rinsed with water and ultrasonicated for 2–3 min. The surface was then cleaned electrochemically by cycling the potential between –0.2 and 1.5 V vs Ag/AgCl in 0.1 M H2SO4 solution until a reproducible voltammogram for the total response was obtained over this potential range. The electrochemically pretreated electrode was extensively rinsed with water, dried and immediately transferred into the solution to be studied. To prepare the chemically modified form, the electrode freshly prepared as described above was immersed for 1 min into a 1 mM 4,4′dipyridyl disulfide nitrogen degassed aqueous solution.32 The modified electrode was rinsed with water and transferred into a 400 µM cytochrome solution, prepared in a 20 mM phosphate

(26) Rudolph, M.; Reddy, D. P.; Feldberg, S. W. Anal. Chem. 1994, 66, 589A. (27) Bond, A. M.; Oldham, K. B.; Zoski, C. G. J. Electroanal. Chem. 1988, 245, 71. (28) Bond, A. M.; Oldham, K. B.; Zoski, C. G. Anal. Chim. Acta 1989, 216, 177.

(29) Engblom, S. O.; Myland, J. C.; Oldham, K. B.; Taylor, A. L. Electronalysis 2001, 13, 626. (30) Licht, S.; Cammarata, V.; Wrighton, M. S. J. Phys. Chem. 1990, 94, 6133. (31) Edowes, M. J.; Hill, H. A. O. J. Am. Chem. Soc. 1979, 101, 4661. (32) Zhou, W.; Baunach, T.; Ivanova, V.; Kolb, D. M. Langmuir 2004, 20, 4590.

Figure 1. Schematic representation of (a) side view and (b) plane view of the experimental arrangement used in this study consisting of an autoswitch device, potentiostat, array electrode and electrochemical cell.


3875

Figure 2. Comparison of an experimental (s) cyclic voltammogram (ν ) 50 mV s-1, T ) 20 °C) for reduction of 1.00 mM [Ru(NH3)6]3+ in 0.5 M KCl at a 127 µm radius gold electrode and simulated data based on planar (b) or hemispherical (O) diffusion models. Parameters used in simulation are as follows: process assumed to be reversible, diffusion coefficients of [Ru(NH3)6]3+ ) 7.6 × 10-6 cm2 s-1 and [Ru(NH3)6]2+ ) 7.8 × 10-6 cm2 s-1 respectively, concentration of [Ru(NH3)6]3+ ) 1.00 mM, electrode area ) 0.000507 cm2, double layer capacitance ) 0.25 µF cm-2, scan rate ) 50 mV s-1, temperature ) 20 °C.

buffer, pH 7.0 containing 100 mM NaCl. Before measurements were made, a 10 min pre-equilibrium period was employed.33 RESULTS AND DISCUSSION Basic Voltammetric Features of the Gold Array Electrode. The three major considerations that determine the nature of voltammograms obtained at an array electrode are the size of the individual electrodes, the spacing between each electrode and the scan rate. If the voltammograms of individual electrodes are to conform to simple theory, then the distance between each single electrode must be large enough to avoid overlap of diffusion layers. If this condition can be achieved, as is required in this study, the array electrode voltammetric response should correspond to the response of the individual microdisk multiplied by the number of electrodes, provided each electrode behaves in an identical manner. In this study, the geometric radius of each electrode in the array is 127 µm. Thus we have an array of electrodes that should approach macrodisk rather than microdisk behavior, which implies less radial diffusion and hence less overlap of diffusion layers than usually encountered in array electrodes which are typically based on individual elements of radius e20 µm. Figure 2 shows the peak-shaped voltammogram derived from a single gold element in the gold array obtained at a scan rate of 50 mV s-1 for reduction of 1.00 mM [Ru(NH3)6]3+ in the presence of aqueous 0.5 M KCl electrolyte and the corresponding simulated data derived from either fully planar diffusion expected at a macrodisk electrode or treating the element as a microdisk electrode and using the hemispherical diffusion approximation.27,28 The simulation-experiment comparison shows that while there is a contribution from radial diffusion at a scan rate of 50 mV s-1, the peak current magnitude is still well in excess of the steady state limiting predicted from the relationship27 IL ) 4nFcbDa (33) Fedurco, M. Coord. Chem. Rev. 2000, 209, 263.

3876


(1)

where D and cb are the diffusivity and bulk concentration of the electroactive species, n is the number of electrons in the charge transfer process, F is Faraday’s constant and a is the radius (127 µm) of the gold electrode. However, as the scan rate is lowered, the voltammogram of an individual gold electrode is expected to approach more closely to steady state response, which also would enhance overlap of diffusion layers when the electrodes are used in the array mode. As shown in Figure 3a, in the case of an individual 127 µm radius gold electrode, a closer to steady state voltammogram is obtained at a scan rate of 2 mV s-1. This is deduced by noting the approach to a sigmoidal shape and convergence of current values in forward and reverse potential sweep directions. Simulation data derived from the hemispherical diffusion approximation model agree well with experimental data obtain at a single element for all scan rates studied. The total array electrode voltammograms (Figure 3b) are expected to display shape characteristics similar to those for an individual electrode if the response is additive, and this is found at scan rates g50 mV s-1. However, the array voltammogram is more peak shaped and less close to steady state at lower scan rate, which implies that a small level of overlap of diffusions layers occurs under these conditions.34 In summary, at a scan rate of 50 mV s-1, virtually no evidence of overlap of diffusion layers is found, which is the condition needed to conveniently evaluate the effect of electrode heterogeneity. To further verify that this required diffusion behavior is achieved at a scan rate of 50 mV s-1, the individual gold electrodes were connected together in either regular or random distributions. The peak currents were found to increase linearly with the total number of active gold wires, irrespective of their location (spacing). Heterogeneity Effects at a Polished and Electrochemically Pretreated Gold Array Electrode. In order to probe the influence of electrode heterogeneity, the voltammetric responses of the array electrode at a scan rate of 50 mV s-1 were evaluated using three redox systems that were expected to exhibit distinctly different properties. Initially, the voltammetry derived from the reversible outer sphere [Ru(NH3)6]3+/2+ process, which should be insensitive to the surface conditions, except for electrode area, was examined. Figure 4a shows 98 individual voltammograms obtained from the gold array electrode. While there is some variation in the individual response, all peak current values lie in the range of 0.2 ± 0.02 µA. Furthermore, for each electrode the peak to peak separations of 81.3 ± 2 mV and the wave shape are almost identical. In this and all other series of experiments, 2 of the 100 electrodes were found to be electrochemically inactive (no electrical contact) and hence data were reported from the 98 electroactive electrodes. The variability at each individual electrode for the [Ru(NH3)6]3+/2+ process is predominantly attributed to the variation in the surface area achieved by the electrode pretreatment. In particular, electrode polishing generates scratches of variable length and depth. If a single electrode equal to the summed area were used, it is not possible to ascertain the contributions of localized sections of an electrode, but the electrode array data suggest that electrode area variability is undoubtedly present. When repetitive experiments were undertaken with a repolished and electrochemically (34) Chevallier, F. G.; Compton, R. G. Electroanalysis 2006, 18, 2369.

Figure 3. Cyclic voltammograms obtained as a function of scan rate for reduction of 1.00 mM [Ru(NH3)6]3+ in 0.5 M KCl from (a) a single 127 µm radius gold element and (b) an array electrode consisting of 98 electrochemically active elements. In the case of the single element, simulations obtained using parameters given in the caption to Figure 2, based on a hemispherical approximation diffusion model, are included (O).

Figure 4. Cyclic voltammograms obtained for the reduction of 1.00 mM [Ru(NH3)6]3+ in 0.5 M KCl at (a) each individual 127 µm radius gold element (total of 98), (b) the full array electrode with all elements connected simultaneously (black trace) compared to voltammograms derived from arithmetic summation of each individual element (red trace) derived from experiment a.

pretreated array electrode, the peak current values at each element varied randomly, implying that deviations in electrode area (roughness) associated with polishing account for the variability of current magnitudes from individual electrodes. Figure 4b confirms that, as required, a voltammogram obtained from an experiment derived from 98 simultaneously connected electrodes is indistinguishable from that found when the individual responses of the 98 electrodes are combined. This result confirms that when a scan rate of 50 mV s-1 is used, each electrode in the array format behaves independently, since overlapping of the diffusion layers would have led to an enhanced current. Thus, direct relationships between each individual voltammogram with those obtained from the array electrode can be deduced in a straightforward manner. The electrochemistry of the quasi-reversible inner sphere [Fe(CN)6]3-/4- redox couple in aqueous media has been the subject of extensive research interest and is widely applied as a benchmark system for studying electrode kinetics and mass transport behavior. This couple shows considerably more electrode dependent behavior than found with the reversible [Ru(NH3)6]3+/2+ model system. Importantly, the kinetics of the [Fe(CN)6]3-/4- redox couple is well-known to be significantly dependent on the pretreatment of the electrode surface (polishing, potential cycling) and also on the nature and concentration of the supporting electrolyte.35–37 Some studies have suggested that variability is associated with the formation of a passivating (35) Peter, L. M.; Duerr, W.; Bindra, P.; Gerisher, H. J. Electroanal. Chem. 1976, 71, 31.

film derived from adsorption of ferricyanide and/or ferrocyanide,34,35 although there is not universal agreement on the origin of the surface dependence.38,39 Pharr and Griffiths concluded that complications in KCl electrolyte used in the present studies are emphasized when the potential is scanned over a large range of potential (e.g., –0.33 to 0.80 V).40,41 Thus, to minimize this contribution, we used a narrower potential range of 0 to 0.5 V for cycling the potential. In the case of the [Fe(CN)6]3-/4- process, individual electrode activity also can be assessed in terms of the heterogeneous electron transfer rate via the reproducibility of the peak-to-peak separation, ∆Ep, derived from results at individual electrodes. Figure 5a displays voltammograms obtained for reduction of [Fe(CN)6]3-/4- in 0.5 M KCl at 98 individual electrodes. Figure 5b gives the experimental result for the array electrode and also the arithmetically summed response of all the individual voltammetric response for reduction of [Fe(CN)6]3-/4-. As for the [Ru(NH3)6]3+/2+ case, almost complete agreement is evident, again implying that little overlap of diffusion layers occurs at a scan rate of 50 mV s-1. However, the inner sphere [Fe(CN)6]3-/4- system is sensitive to the electrode surface. In this case, ∆Ep values are less reproducible than for the ideal Huang, W.; McCreery, R. J. Electroanal. Chem. 1992, 326, 1. Wieckowski, A.; Szklarczyk, J. J. Electroanal. Chem. 1982, 142, 157. Niwa, K.; Doblhofer, K. Electrochim. Acta 1986, 31, 439. Christensen, P. A.; Hamnett, A.; Trevellick, P. R. J. Electroanal. Chem. 1988, 242, 23. (40) Pharr, C. M.; Griffiths, P. R. Anal. Chem. 1997, 69, 4665. (41) Pharr, C. M.; Griffiths, P. R. Anal. Chem. 1997, 69, 4673.

(36) (37) (38) (39)


3877

Figure 5. Cyclic voltammograms obtained for the reduction of 1.00 mM [Fe(CN)6]3+ in 0.5 M KCl at (a) each individual 127 µm radius gold element (total of 98), (b) the full array electrode with all elements connected simultaneously (black trace) compared to the voltammograms derived from arithmetic summation of each individual element (red trace) derived from experiment a. Table 1. Comparison of Mean, Standard Deviation and Percentage Deviation of Parameters Obtained from Voltammograms for the [Ru(NH3)6]3+/2+ and [Fe(CN)6]3-/4- Processes Derived from 98 Individual Gold Electrodes and from Repetitive Experiments When All 98 Electrodes Are Connected Together in an Array Format analysis based on cyclic voltammograms obtained from 98 active individual gold electrodes

analysis based on cyclic voltammogram(s) obtained from a gold array electrodea

redox system

parameter

mean (deviation)

% deviation

mean (N ) 1)

mean (N ) 20)

% deviation (N ) 20)

1 mM [Ru(NH3)6]Cl3 in 0.5 M KCl

Ip (µA) ∆Ep (mV) Ip (µA) ∆Ep (mV)

0.201 (±0.017) 81.3 (±2.0) 0.202 (±0.021) 87.8 (±6.6)

8.46 2.46 10.40 7.52

0.213 81.9 0.209 87.5

0.218 (±0.003) 80.9 (±1.5) 0.207 (±0.003) 86.2 (±1.4)

1.38 1.85 1.45 1.62

1 mM K3[Fe(CN)6] in 0.5 M KCl a

N ) number of repetitive experiments when all elements of array electrode are connected.

[Ru(NH3)6]3+/2+ process, as expected if the electron transfer rate constant is dependent on the surface pretreatment. Table 1 contains results of an analysis of peak height and peak potential separations obtained from voltammograms from the [Ru(NH3)6]3+/2+ and [Fe(CN)6]3-/4- processes with 98 active individual gold electrodes and for multiple experiments with an array gold electrode, polished and pretreated between each experiment. The individual electrodes exhibit variation in peak height of around 8% and 10% for the [Ru(NH3)6]3+/2+ and [Fe(CN)6]3-/4- processes, respectively. In contrast, 20 repetitive experiments with an array electrode give less than 2% variation. The ∆Ep values obtained for the reversible [Ru(NH3)6]3+/2+ process are smaller than for the [Fe(CN)6]3-/4- system, which confirms that the kinetics of the electrode process for the former redox couple is faster. The increased variability of 7.5% in ∆Ep versus 2.5% for the [Ru(NH3)6]3+/2+ studies (Table 1) reflects the sensitivity of the rate of the [Fe(CN)6]3-/4- process to the state of the electrode surface. Theoretical ∆Ep values were calculated from simulations of the microdisk electrode, based on the hemispherical approximation. For a reversible one electron transfer process, ∆Ep is predicted to be 79 mV at a scan rate of 50 mV s-1 for a 127 µm radius electrode at 20 °C. The ∆Ep value of 81 mV obtained experimentally for the [Ru(NH3)6]3+/2+ redox system is therefore close to that expected for a reversible one electron transfer process. For [Fe(CN)6]3-/4-, a ∆Ep value of 87.8 ± 6.6 mV implies that the electrode processes from 98 active electrodes range from reversible to quasi-reversible. This is consistent with values of the 3878


electron transfer rate of [Fe(CN)6]3-/4- lying in the range of 0.5 cm s-1 to 0.01 cm s–1.34,42 The 3-fold greater percentage deviation in ∆Ep values derived from analysis of individual gold electrodes in the [Fe(CN)6]3-/4compared to [Ru(NH3)6]3+/2+ system is consistent with the greater sensitivity of the [Fe(CN)6]3-/4- process to the surface conditions. In contrast, an equivalent analysis for an array electrode based on 20 repetitive voltammograms produces about a 1.5% deviation in ∆Ep values for both systems. The fact that the average values of ∆Ep for both systems derived from analysis of the array electrode are almost identical implies that the presence of heterogeneity may remain virtually unnoticed unless analysis of localized parts of the electrode surface can be probed. Theoretical analyses of arrays of microelectrodes have been developed to explain how electrode heterogeneity may be related to a situation giving rise to partial blocking of a large electrode.43 The [Fe(CN)6]3-/4- case is related to rate constants being variable at different parts of the electrode surface to give what is usually referred to as kinetic dispersion, with the variable surface state predominantly being introduced by nonuniform polishing. A study of background current at a bare gold electrode using Fourier transform large amplitude ac voltammetry confirmed the presence of “hot spots” that generate Faradaic reactions.13 Polishing and (42) Goldstein, E. L.; Van De Mark, M. R. Electrochim. Acta 1982, 27, 1079. (43) (a) Gueshi, T.; Tokuda, K.; Matsuda, H. J. Electroanal. Chem. 1978, 89, 247. (b) Gueshi, T.; Tokuda, K.; Matsuda, H. J. Electroanal. Chem. 1979, 101, 29. (c) Tokuda, K.; Gueshi, T.; Matsuda, H. J. Electroanal. Chem. 1979, 102, 41. (44) Taniguchi, I.; Murakami, T.; Toyosawa, K.; Yamaguchi, H.; Yasukouchi, K. J. Electroanal. Chem. 1982, 131, 397.

Figure 6. Cyclic voltammograms obtained at a scan rate of 50 mV s-1 for 400 µM cytochrome c (0.1 M NaCl in 20 mM phosphate buffer) at pH 7 for (a) each individual 4,4′-dipyridyl disulfide modified, 127 µm radius gold element (total of 98), (b) arithmetic summation of voltammograms from the 98 individual elements (red trace) and when all electrodes are electrically connected (black trace). (c) and (d) contain representative examples of voltammograms that exhibit well-defined and poorly defined peaks, respectively.

electrochemical cleaning protocols contribute to the extent of their presence, and these can contribute to surface pretreatment dependent heterogeneity associated with gold electrode surfaces. Heterogeneity Effects at a Chemically Modified Gold Array Electrode. Chemically modified electrodes are widely employed to deliberately impart functionality that produces voltammetric behavior that is distinctly different from that formed at the bare electrode.3 In the case of cytochrome c, the voltammetry at a bare gold electrode is poorly defined and nonreproducible, whereas, at a thiol based chemically modified surface, classical diffusion controlled behavior has been reported.44 In this investigation, the heterogeneity effect of a thiol monolayer modified electrode surface is probed with respect to the diffusion controlled electrochemistry of cytochrome c. Figure 6a reveals that a wide range of shapes are detected at individual gold electrodes in the case of 400 µM cytochrome c voltammetry at a scan rate of 50 mV s-1 in pH 7 phosphate buffer (0.1 M NaCl) when the array electrode is modified with 4,4′dipyridyl disulfide. This is in contrast to the well-defined peak shaped voltammetry always found from the summation of the responses from the 98 individual electrodes (Figure 6a) or the voltammogram obtained when all 98 electrodes are electrically connected (Figure 6b). In this chemically modified electrode scenario, the overall voltammogram obtained from the array electrode represents the summation of a wide distribution of voltammetric responses. The level of variation is emphasized by perusal of selected individual voltammograms displayed in Figures 6c and 6d. The former provide well-defined peaked shaped voltammograms, while the latter provide examples of sigmoidal shaped Faradaic responses where background current is domi-

Table 2. Comparison of Mean, Standard Deviation and Percentage Deviation of Parameters Obtained from Voltammetry of Cytochrome c Derived from 88 Individual Gold Electrodes and When All Electrodes Are Connected Together in an Array Formata analysis based on voltammograms obtained from 88 individual analysis based on gold electrodesb cyclic voltammogram mean % obtained from a parameter (deviation) deviation gold array electrode Ip (µA) ∆Ep (mV)

4.03 (±1.06) 80.4 (±7.9)

26.30 9.83

4.33b,c 83.0b,c

a Redox system: 400 µM cytochrome c in 20 mM phosphate buffer + 0.1 M NaCl. b See text for further details. c Derived from voltammograms at 98 electrodes.

nant. The midpoint potential, Em, which should approximate the reversible potential, is estimated from the average of the forward and reverse peak potentials. For the summed and electrically full connected gold electrode configuration (Figure 6b), Em was 0.047 ± 0.004 V vs Ag/AgCl (3 M NaCl) or 0.256 V ± 0.004 vs NHE and hence is in excellent agreement with the published value of 0.255 V vs NHE.31,44 Table 2 summarized data gained from the 88 active individual gold electrodes that provided peaks that enable Ip and ∆Ep values to be measured. The deviation of Ip values at this modified gold electrode is much larger (∼25%) than obtained at a bare electrode surface (∼10%) for the surface sensitive [Fe(CN)6]3-/4- system (see Table 1). This enhanced deviation is consistent with significant variation in surface coverage of the 4,4′-dipyridyl disulfide modified electrode. Analytical Chemistry, Vol. 80, No. 10, May 15, 2008

3879

Figure 7. Cyclic voltammograms (scan rate ) 50 mV s-1) derived from (a) summation of 98 active gold elements for 400 µM cytochrome c in 0.1 M NaCl (20 mM phosphate buffer solution pH 7) before (black trace) and after (blue trace) the addition of 10 µg mL-1 poly-L-lysine. (b), (c) and (d) are representative examples of voltammograms obtained from individual gold elements.

Simulations, based on the hemispherical diffusion approximation of a disk electrode and with a diffusion coefficient for cytochrome c of 9.4 × 10-7 cm2 s-1,31 predict a ∆Ep value of 66 mV for a chemically and electrochemically reversible process at 20 °C. The average ∆Ep value obtained experimentally is 80.4 ± 7.9 mV. Notionally, these data suggest that the heterogeneous rate of electron transfer for cytochrome c electrochemistry at a modified gold electrode lies in the range of about 0.01 cm s-1 to 0.003 cm s-1. Furthermore, the ∆Ep value of 83.0 mV obtained from the array configuration is close to the mean value of 80.4 mV obtained from individual electrodes. These results could be used to imply that the process is quasi-reversible. However, the voltammetric behavior is better explained by a microscopic model of the modified gold electrode, where fast electron transfer occurs at suitably modified parts of the electrode with no electron transfer occurring at other parts of the heterogeneous surface.45–47 Nonuniform surface coverage with 4,4′-dipyridyl disulfide, surface heterogeneity, blockage of the surface by protein denaturation and impurities present in cytochrome c have been proposed to contribute to the electrode behavior of cytochrome c at a chemically modified electrode.46 Voltammetry at a 4,4′-dipyridyl disulfide modified gold electrode is known to be altered by addition of the positively charged poly-L-lysine, which displaces adsorbed 4,4′-dipyridyl disulfide from the electrode surface.45 Figure 7a shows the voltammogram at an array electrode before and after addition of poly-L-lysine (45) Hill, H. A. O.; Page, D. J.; Walton, N. J. J. Electroanal. Chem. 1987, 217, 141. (46) Bond, A. M.; Hill, H. A. O.; Page, D. J.; Psalti, I. S. M.; Walton, N. J. Eur. J. Biochem. 1990, 191, 737. (47) Hill, H. A. O.; Hunt, N. I.; Bond, A. M. J. Electroanal. Chem. 1997, 436, 17.

3880


(average molecular weight of 150000–300000) at a concentration of 10 µg mL-1. Displacement of 4,4′-dipyridyl disulfide leads to a change from peak to a more sigmoidal shaped voltammogram as significant parts of the surface become electroinactive. However, again, the influence is not uniform. Comparison of voltammograms in Figures 7b and 7c shows that the decrease in peak current in the former case is larger than the latter, even though the initial voltammetry before addition of poly-L-lysine was very similar. In the example depicted in Figure 7d, the response prior to addition of poly-L-lysine was already sigmoidal shaped. In this case, the poly-L-lysine acts to further reduce the current. Thus, variable strength of binding of 4,4′-dipyridyl disulfide to the gold electrode surface is assumed to contribute to microscopic differences in its displacement behavior, as detected via analysis of an array of 127 µm radius gold electrodes. Clearly, cytochrome c voltammetry on thiol-monolayer modified surfaces takes place at highly heterogeneous surfaces and relating ∆Ep values to a single rate of electron transfer via a planar diffusion model is not valid, even though the total voltammetric response is peak shaped. Use of a complex model based on a partially but randomly blocked surface is more appropriate. The studies of Compton et al. consider the various scenarios that may arise in these circumstances.5–7 CONCLUSION A voltammetric study with individually addressable gold electrodes that form part of an array configuration allow data to be analyzed that are summed from each individual response or obtained when all electrodes are electrically connected. Results imply that random levels of inhomogeneities in gold electrode surfaces may contribute to the overall voltammetric response

obtained from a gold electrode. In most cases, the influence caused by electrode heterogeneity will be subtle, although in the case of a chemically modified electrode surface, heterogeneity may drastically influence even the wave shape. Studies by Compton et al.5–7 on effects of heterogeneity at carbon electrodes and the present study with gold electrodes imply that electrochemists may need to more widely recognize the influence of surface inhomogeneities as a factor that introduces nonideal behaviors relative to those predicted on the basis of a uniform surface.

ACKNOWLEDGMENT We thank Xiaohu Qu and Chuan Zhao for generous and valuable assistance. The financial support of the Australian Research Council and award of Monash University postgraduate scholarships (MIPRS and MGS) to C.-Y.L. are also gratefully acknowledged. Received for review January 30, 2008. Accepted March 3, 2008. AC8002227


3881

Identification of Surface Heterogeneity Effects in Cyclic

Recommend Documents