Metal Oxide Electrodes

Anal. Chem. 1999, 71, 3166-3170

Scanning Electrochemical Microscopy of Metal/ Metal Oxide Electrodes. Analysis of Spatially Localized Electron-Transfer Reactions during Oxide Growth Solomon B. Basame and Henry S. White*

Department of Chemistry, University of Utah, Salt Lake City, Utah 84112

Scanning electrochemical microscopy (SECM) has been used to study the oxidation of iodide at Ta electrodes covered by a thin (∼2.5 nm) film of Ta2O5. SECM images of surface activity reveal that the voltammetric response of a macroscopic Ta electrode comprises the individual responses of a large number of microscopic sites, each with its own unique electrochemical behavior. Oxide film growth and metal dissolution occur simultaneously with iodide oxidation, resulting in a complex voltammetric response. The component of the voltammetric current due to iodide oxidation can be separated from the total current by SECM analysis. The growth of nanometer-thick oxide films can also be studied using SECM by monitoring the rate at which iodide is oxidized at the electrode surface. Metal electrodes coated by a thin oxide film (e.g., PtOx on Pt and TiO2 on Ti) are frequently encountered in electrochemical measurements.1 The oxide film, which ranges in thickness from molecular dimensions to micrometers, presents analytical challenges in determining its thickness, chemical composition, crystallinity, and electronic properties. In electrochemistry, small variations in the thickness of insulating oxide films may determine the difference between facile and sluggish electron-transfer kinetics.2,3 Other interfacial chemistries and properties, such as the chemisorption of ions, reflectivity, and surface tension, are dependent on whether an oxide film is present. The characterization of electrodes covered by thin oxide films presents several difficulties that are not encountered with bare metal electrodes. First, and analogous to the study of molecular films on electrodes, it is difficult to ensure that a perfectly uniform oxide film is present across the entire surface. The occurrence of a small number of defects in the oxide film, e.g., local variations (1) (a) A recent and extensive listing of literature describing fundamental studies of very thin oxide films on commonly employed electrode materials (Pt, Au, Ir, etc.) is found in: Tremiliosi-Filho, G.; Dall’Antonia, L. H.; Jerkiewicz, G. J. Electroanal. Chem. 1997, 422, 149. (b) Henrich, V. E.; Cox, P. A. The Surface Science of Metal Oxides; Cambridge University Press: Cambridge, U.K., 1994. (c) Young, L. Anodic Oxide Films; Academic Press: New York, 1961. (d) Oxides and Oxide Films; Diggle, J. W., Ed.; Marcel Dekker: New York, 1972. (2) (a) Morisaki, H.; Ono H.; Yazawa, K. J. Electrochem. Soc. 1988, 135, 381. (b) Morisaki, H.; Ono H.; Yazawa, K. J. Electrochem. Soc. 1989, 136, 1710. (3) Casillas, N.; Charlebois, S.; Smyrl, W. H.; White, H. S. J. Electrochem. Soc. 1994, 141, 636.

3166 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

in film thickness or stoichiometry, is difficult to prevent during film growth and even more difficult to detect by conventional surface analytical techniques, e.g., scanning Auger electron spectroscopy. Such defects, however, may dominate the electrode response by acting as active sites for electron transfer or electrochemical dissolution.3,4 Second, the structure and properties of thin oxide films in electrochemical environments may change as a function of the applied potential and time during the course of an experiment. In these cases, it is difficult to quantify the thickness and structure of the oxide film during the course of an electrochemical experiment with any degree of certainty. In situ X-ray diffraction-based methodologies yield information about the average properties of the interfacial oxide films but are generally not useful for studying localized structure or chemistry. Highresolution microscopies (e.g., STM, AFM), while widely successful in the study of single-crystal metal electrodes, have had limited application in the study of oxide-covered metals. In the present study, we describe an investigation of the electrochemical kinetics of Ta electrodes that are covered by very thin (∼2.5 nm) films of Ta2O5 (these electrodes are hereafter referred to as Ta/Ta2O5). Ta2O5 is a large band gap semiconductor (Eg ∼ 4 eV),5 and, thus, its presence as a film is expected to passivate the surface toward interfacial electron transfer. Using scanning electrochemical microscopy (SECM6,7) to map the activity at the oxide film/solution interface, we have recently demonstrated that the electrochemical behavior of Ta/Ta2O5 electrodes in solutions containing simple redox molecules, e.g., Ru(NH3)63+, is dominated by a small number of defect sites of micrometer dimensions that behave as metal-like microdisks embedded in an insulating plane.4a These sites act as efficient shunts for electron transfer across the oxide film, compromising (4) (a) Basame, S. B.; White, H. S. Langmuir 1999, 15, 819. (b) Basame, S. B.; White, H. S. J. Phys. Chem. 1998, 102, 9812. (c) Basame, S. B.; White, H. S. J. Phys. Chem. 1995, 99, 16430. (d) Casillas, N.; Charlebois, S.; Smyrl, W. H.; White, H. S. J. Electrochem. Soc. 1993, 140, L142. (5) (a) Clechet, P.; Martin, J. R.; Ollies, R.; Vallouy, C. R. Acad. Sci. 1976, 282C, 887. (b) Vijh, A. K. In Oxides and Oxide Films, Vol. 2; Diggle, J. W., Ed.; Marcel Dekker: New York, 1972. (6) (a) Bard, A. J.; Fan, F.-R. F.; Kwak, J.; Lev, O. Anal. Chem. 1989, 61, 132. (b) Bard, A. J.; Fan, F.-R. F.; Mirkin, M. V. In Electroanalytical Chemistry, Vol. 18; Bard, A. J., Ed.; Marcel Dekker: New York, 1994; p 243. (7) (a) Engstrom, R. C.; Weber, M.; Wunder, D. J.; Burgess, R.; Winquist, S. Anal. Chem. 1986, 58, 844. (b) Engstrom, R. C.; Meaney, T.; Tople, R.; Wightman, R. M. Anal. Chem. 1987, 59, 2005. (c) Engstrom, R. C. Anal. Chem. 1984, 56, 890. 10.1021/ac9902897 CCC: $18.00

© 1999 American Chemical Society Published on Web 06/10/1999

the passivity afforded by the oxide film. Recent developments in SECM methodology provide a means for quantifying the dimensions and current density at randomly dispersed redox-active sites during the course of an electrochemical experiment.4b Herein, we demonstrate the analytical capabilities of SECM for unraveling electrochemical phenomena at oxide-covered metal electrodes. Specifically, we report SECM images of complex electrochemical behavior that occurs during the oxidation of Iat oxide-covered Ta electrodes. In addition to being spatially localized, the oxidation of I- occurs simultaneously with metal dissolution and oxide film growth. All of these processes contribute to the net current measured in a conventional electrochemical experiment. We show that SECM can be used to selectively measure the contribution to the net current resulting from oxidation of I-. EXPERIMENTAL SECTION Chemicals. Potassium iodide and potassium sulfate (Aldrich) were used as received. All aqueous solutions were prepared using 18 MΩ‚cm water purified from a Barnstead “E-pure” system. Ta/Ta2O5 Electrodes. Ta electrodes were prepared from 25 µm thick 99.997% Ta foil (Johnson Matthey). Cu wire was contacted to an ∼1.5 cm2 piece of Ta foil sample using conductive Ag epoxy. The Ag epoxy was covered by an inert nonconductive epoxy, and the entire assembly was mounted at the end of a glass tube that served as a sample holder for voltammetry and SECM imaging. Immediately prior to each experiment, the electrode was etched for 5 min in an aqueous solution containing 6% HF and 6% HNO3 and then thoroughly rinsed with water. This procedure removes surface contamination and the previously air-formed oxide. The thickness of the native oxide is reported to be 2.47 nm.8 Scanning Electrochemical Microscopy. The in-house built SECM used in these studies has been previously described in detail.9 Briefly, the substrate potential (Es) and the SECM tip potential (Et) are independently controlled using two separate potentiostatic circuits. In the studies reported here, the tip is used to detect I3- that is electrogenerated at the Ta/Ta2O5 surface. This mode of SECM operation is referred to as the substrate generation/tip collection mode.10 A 4 µm radius, carbon microdisk-shaped SECM tip is used for SECM imaging. Details of tip preparation have been previously described.4b To acquire an SECM image, the C microdisk tip is scanned across the sample surface at a fixed height (∼5 µm) and the tip current is recorded as a function of its spatial position. The tip is scanned during imaging at a velocity between 2 and 10 µm/s. SECM tip movement and data acquisition are controlled with a PC using several virtual instrument subroutines written in LabView (National Instruments). RESULTS AND DISCUSSION Figure 1 shows two 300 × 300 µm SECM images of Ta/Ta2O5 electrodes immersed in an aqueous solution containing 10 mM KI and 0.1 M K2SO4. To acquire these images, the Ta/Ta2O5 (8) Sung, Y.-E.; Bard, A. J. J. Phys. Chem. 1998, 102, 9807. (9) (a) Scott, E. R.; White, H. S.; Phipps, J. B. Anal. Chem. 1993, 65, 1537. (b) Scott, E. R.; Laplaza, A. I.; White, H. S.; Phipps, J. B. J. Pharm. Res. 1993, 10, 1699. (10) Zhou, F.; Unwin, P. R.; Bard, A. J. J. Phys. Chem. 1992, 96, 4917.

Figure 1. Representative SECM images (300 × 300 µm) of a Ta/ Ta2O5 electrode in a solution containing 10 mM KI and 0.1 M K2SO4. The dark regions correspond to sites of high activity for the oxidation of I- at the Ta/Ta2O5 electrode: (A) an isolated site; (B) a cluster of sites. A 4 µm radius SECM tip was scanned at 10 µm/s at a tip-tosurface separation, d ∼ 5 µm. ET ) 0.0 V, and Es ) 1.0 V vs Ag/ AgCl.

potential (Es) was set to 1.0 V vs Ag/AgCl in order to oxidize I(eq 1), while the tip potential (Et) was held at 0.0 V to detect electrogenerated I3-.

3I- h I3- + 2e-

(1)

The darker areas in the SECM images correspond to spatial regions where a relatively high concentration of I3- is detected above the Ta/Ta2O5 surface. Thus, the SECM tip allows visualization of the localized oxidation of I-. Redox-active sites at the Ta/ Ta2O5 surface are observed as isolated sites (Figure 1A) or as a cluster of sites (Figure 1b). The redox-active sites are roughly circular in shape; the vertical streaks apparent in the image of the cluster of sites (Figure 1B) are due to local stirring caused by the moving tip. A previously reported quantitative analysis of the Ta/Ta2O5 surface yields a site number density of 259 ( 51 cm-2 and site radii ranging from 2 to 46 µm.4a The chemical and physical properties of the redox-active sites that are responsible for their metal-like behavior are currently under investigation. We speculate that the facile electron-transfer kinetics at these sites is due to a local thinning of the oxide film (resulting in an increase in the electron tunneling across the oxide film) or to a local variation in the oxide composition (resulting in a increase in the electronic conductivity of the oxide film). Either type of defect might lead to increased electrochemical activity. Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

3167

Figure 2. SECM images of a 300 × 300 µm area of a Ta/Ta2O5 electrode in a solution containing 10 mM KI and 0.1 M K2SO4 as a function of the Ta/Ta2O5 potential. The SECM tip potential (Et) was held constant at 0.0 V, while the substrate potential (Es) was varied as indicated on the voltammogram. Other conditions are the same as in Figure 1. The peaks in the images correspond to active sites for the oxidation of I- on the Ta/Ta2O5 surface. The voltammetric response of the Ta/Ta2O5 electrode in a solution containing only the supporting electrolyte, 0.1 M K2SO4, is also shown.

Figure 2 shows the voltammetric response of a Ta/Ta2O5 electrode in 0.1 M K2SO4 solutions in the presence and absence of 10 mM I-. In the absence of I-, a current is observed beginning at ∼1.0 V that corresponds to growth of the oxide film (2Ta0 + 5 H2O f Ta2O5 + 10e- + 10H+) and metal dissolution (Ta0 f Ta5+ + 5e-).1c,11 This current reaches a limiting value at ∼1.5 V, reflecting the slow electric field driven transport of Ta5+ through the oxide film. When 10 mM KI is added to the solution, a current corresponding to the oxidation of I- (eq 1) is observed beginning at ∼0.5 V. Figure 2 also shows SECM images of the Ta/Ta2O5 electrode recorded during the voltammetric experiment. The SECM image is essentially featureless when the potential of the Ta/Ta2O5 electrode is held at the foot of the voltammetric wave for I(11) Udupa, H. V. K.; Venkatesan, V. K. In Standard Potentials in Aqueous Solutions; Bard, A. J., Parsons, R., Jordan, J., Eds.; Marcel Dekker: New York, 1985; p 507.

3168 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

oxidation, (image a). As the electrode potential is shifted positive in order to oxidize I- at a significant rate, image b shows that Iis oxidized at two sites (labeled as 1 and 2). Three new sites (35) appear when the Ta/Ta2O5 electrode potential is shifted slightly more positive (image c). The appearance of sites 3-5 at more positive potentials than sites 1 and 2 demonstrates that the electron-transfer kinetics at these two sets of sites are significantly different. The images shown in Figure 2 are stable with respect to time and, thus, represent the steady-state redox activity at the Ta/Ta2O5 surface. The redox activity is a function only of the electrode potential. When the electrode potential is increased to 1.2 V (image d), the electrochemical activity at sites 4 and 5 abruptly disappear. A similar irreversible loss of activity at sites 2 and 3 occur at 1.4 V (image e). The sudden loss of redox activity at these sites is most likely associated with Ta2O5 film growth, reducing the rate of electron transfer across the interface. The redox activity at site 1

Figure 4. Comparison of the sum of SECM tip currents (∑it) measured at 38 redox-active sites on different Ta/Ta2O5 electrodes with the substrate current (is) measured at one individual Ta/Ta2O5 electrode (15 mm2). All measurements were made in solutions containing 10 mM KI and 0.1 M K2SO4. Figure 3. Plots of the SECM tip current (it) vs the substrate potential (Es), measured over six active sites on six different Ta/Ta2O5 electrodes. The voltammograms were obtained at 5 mV/s in a solution containing 10 mM KI and 0.1 M K2SO4. The magnitude and shape of the voltammetric curves differ due to differences in the size and electron-transfer kinetics at each redox-active site. The sudden decrease in current at positive potentials is due to local oxide growth and passivation of the electroactive site. Et ) 0.0 V vs Ag/AgCl, and d ) 4 µm.

finally disappears when the electrode potential is increased to 1.8 V (image g). This featureless image suggests that the oxidation of I- is effectively blocked as the oxide film becomes thicker. Thus, the current observed in the voltammetric response at 1.8 V is due solely to oxide growth and metal dissolution, a fact that would be impossible to deduce based only on the voltammetric response of the Ta/Ta2O5 electrode. Interestingly, the voltammetric current is significantly larger in solutions containing I- than in the blank solution, Figure 2. A likely explanation for the increase in current is that the presence of I- increases the solubility of Ta5+, thus increasing the rate of metal dissolution. An interesting feature of the images in Figure 2 is the unusual dependence of the electrochemical activity of site 1 on the electrode potential. Inspection of the images shows that the activity of this site decreases when the potential of the Ta/Ta2O5 electrode is increased from 0.8 to 1.0 V (compare images b and c), but increases again beginning at ∼1.2 V (compare images c-e). The appearance and disappearance of site 1 does not reflect a change in the electron-transfer rate at site 1 but rather the competition of site 1 with neighboring site 3 for the electrochemical reactant (I-). As site 3 becomes active (beginning at 1.0 V, image c), Ithat was previously oxidized at site 1 is now captured by site 3. This competition for molecules reduces the activity at site 1. The electroactivity of site 1 reappears when site 3 is passivated, beginning at ∼1.4 V (image d). The passivation of redox-active sites by oxide film growth was further investigated by recording the SECM tip current, it, for I3reduction above individual sites as the Ta/Ta2O5 potential, Es, was scanned from 0.0 to 2.5 V at 5 mV/s. Figure 3 shows the it vs Es curves from six experiments in which the SECM tip was positioned ∼4 µm above the centers of redox-active sites on different electrodes. As in the imaging experiments, the tip

potential was set to 0.0 V in order to detect I3-. Since I3- that is detected at the SECM tip is electrogenerated by oxidation of I-, the it vs Es curves should reflect the component of the voltammetric response at an individual redox site that corresponds to the oxidation of I-. Not all of the electrogenerated I3- is detected at the tipsa fraction escapes detection due to radial diffusion of I3away from the electroactive sites. Each of the it vs Es curves in Figure 3 has its own unique shape, reflecting real differences in the electron-transfer kinetics at individual sites. The magnitude of the recorded current is different from site to site, in part, due to differences in the dimensions of the redox-active sites. This factor has not been corrected for in plotting the data. However, it is clear that the dependence of the rate of I- oxidation on electrode potential is significantly different at each site, presumably due to differences in the oxide thickness, composition, or stoichiometry. Interestingly, several of the it vs Es responses (curves a, c, and e in Figure 3) show a sudden loss of electrochemical activity at positive electrode potentials. We previously speculated that this behavior corresponds to sudden nucleation and growth of the oxide film at the redox-active site,4b thereby suddenly reducing the conductivity of the oxide or the rate of electron tunneling across the film. Multiple steplike decreases are observed in two of the responses (curves a and e), which may correspond to a layer-by-layer growth of the oxide film. Work is currently in progress to investigate this possibility in more detail. The unique electrochemical behavior exhibited by each individual redox site raises the interesting question of how individual responses (Figure 3) combine to give rise to the net voltammetric response of the Ta/Ta2O5 electrode (Figure 2). Ideally, one might expect that the summation of the it vs Es responses for many redox-active sites should yield a voltammetric curve that reflects the total rate of I- oxidation at Ta/Ta2O5 electrode. To make a statistically significant comparison, the it vs Es responses of 38 different redox-active sites (each on a different electrode) were recorded. The voltammograms at these sites were then summed and plotted as ∑it vs Es in Figure 4. A conventional voltammogram recorded at a Ta/Ta2O5 electrode, is vs Es, is also shown for comparison. The ∑it vs Es voltammogram in Figure 4 provides mechanistic and kinetic information not available in the conventional voltamAnalytical Chemistry, Vol. 71, No. 15, August 1, 1999

3169

mogram. First, the data clearly show that I- is oxidized between 0.5 and 2 V at the Ta/Ta2O5 electrode. The fact that I- oxidation ceases at ∼2 V can only be determined from the SECM datas this finding cannot be deduced from conventional voltammetric measurements alone, (e.g., by subtraction of the blank voltammogram recorded in the absence of I- from the voltammogram measured in the presence of I-). Second, neither of the two large voltammetric peaks in the is vs Es response are due to I- oxidation, but rather appear to be associated solely with oxide growth and/ or metal dissolution.12 Similarly, all of the current at electrode potentials beyond ∼2 V is due to oxide growth and/or metal dissolution. CONCLUSIONS The results presented above demonstrate that SECM can be used to deconvolute complex and spatially localized electrochemi(12) The second large peak at ∼2.0 V vs SCE in the voltammetric response is not reproducible from one electrode to another (compare voltammograms in Figures 2 and 4). We believe that this peak is due to oxide growth and is very sensitive to surface preparation.

3170 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

cal processes. This feature of SECM makes it an especially powerful analytical tool for investigating oxide-covered metal electrodes as well as other nonuniform surfaces. SECM imaging of the redox activity at Ta/Ta2O5 electrodes demonstrates that the voltammetric response of a macroscopic Ta/Ta2O5 electrode results from a large number of microscopic sites, each with is own unique electrochemical behavior. SECM has been used to qualitatively separate the Faradaic component of the net current corresponding to the oxidation of I- at Ta/Ta2O5 from that associated with oxide film growth and metal dissolution. ACKNOWLEDGMENT This work was supported by the Office of Naval Research.

Received for review March 16, 1999. Accepted April 27, 1999. AC9902897