18081
2006, 110, 18081-18087 Published on Web 08/30/2006
Assembly and Electrochemical Characterization of Nanometer-Scale Electrode|Solid Electrolyte Interfaces Matthias Loster,† K. Andreas Friedrich,‡ and Daniel A. Scherson*,§ Center for Solar Energy and Hydrogen Research, DiVision 3, Helmholtzstrasse 8, Ulm, Germany, German Aerospace Center, Electrochemical Energy Technology, Pfaffenwaldring 38-40, 70569 Stuttgart, Germany, and Department of Chemistry, Case Western ReserVe UniVersity, CleVeland, Ohio 44106-7078 ReceiVed: May 26, 2006; In Final Form: July 8, 2006
A technique is herein described for the assembly and characterization of nanometer-scale metal electrode|solid electrolyte interfaces of variable dimensions. The specific system examined in this work involves a sharp Pt tip attached to the piezo-driven head of a scanning tunneling microscope (STM) allowing the tip to be inserted into (or retrieved from) a Nafion membrane placed normal to the direction of tip travel. The actual Pt|Nafion area of contact was determined by coulometric analysis of the characteristic voltammetric features of Pt, using the tip as the working electrode and a much larger Pt gauze attached to the other side of the Nafion as a counter-reference electrode, yielding for some of the interfaces examined values equivalent to as low as 35 000 Pt surface atoms. This rather versatile arrangement allows experiments to be performed in both inert (Ar) and reactive atmospheres, such as oxygen or hydrogen on either or both sides of the membrane, under controlled humidity conditions, and thus sheds light into such phenomena as changes in the overall faradaic currents induced by plastic deformations of the Nafion as well as fundamental aspects of mass transport at reactant gas|Pt|Nafion three-boundary interfaces of relevance to polymer electrolyte fuel cells (PEFCs).
Introduction The design of model systems aimed at the study of electrochemical reactivity at electrode|solid electrolyte|reactant gas three-phase interfaces is expected to provide much needed insight into the nature of the factors that control key aspects of the operation of electrodes in polymer electrolyte fuel cells (PEFCs). The basic catalytic material for both the anode and the cathode in these energy conversion devices consists of platinum (or Pt alloy) particles a few nanometers in diameter supported on high area carbon dispersed in a semi-hydrophobic polymeric matrix. The approach described in this work involves the use of a Pt wire ending in a sharp, conically shaped tip attached to the vertical scanner of a conventional scanning tunneling microscope (STM) head to mimic the Pt nanoparticle|solid electrolyte interface. This arrangement allows the tip to be inserted or retrieved from a solid electrolyte membrane, such as Nafion, and, hence, modulate with exquisite control the area of the tip/Nafion contact, which can be measured by electrochemical techniques. Attempts have been reported in the literature involving the use of both micro- and ultramicroelectrodes to examine O2 reduction and H2 oxidation at Pt|Nafion interfaces;1-6 however, the experiments herein described are, to the best of our knowledge, the first focusing on gaining fundamental information of Pt|Nafion|gas interfaces approaching conditions similar to those found in fuel cells without problems * Corresponding author. E-mail:
[email protected]. Phone: +1 216-368-5186. Fax: +1 216-368-3006. † Center for Solar Energy and Hydrogen Research. ‡ German Aerospace Center. § Case Western Reserve University.
10.1021/jp063255t CCC: $33.50
associated with the presence of other Pt particles which would compromise rigorous analyses.7-11 Of special relevance is the work of Watanabe, who examined correlations between interparticle distance and overall activity due to the extent of overlap between neighboring diffusion layers.11,12 Particularly noteworthy are the reports of O’Hayre et al.,13 who attempted to measure the area of contact for a single Pt nanoparticle|Nafion interface using impedance spectroscopy and contact force atomic force microscopy (AFM), as well as those of Chen et al. who studied the hydrogen oxidation and oxygen reduction reactions in sulfuric acid on single Pt particles of a geometric diameter (estimated from diffusion measurements) down to 100 nm.14 Well-defined structures were studied on a more macroscopic scale by Katakura et al., who contacted Nafion with a Pt tip6 and by Paulus et al. who brought a specially designed carbon structure that was partly covered with Pt in contact with Nafion to evaluate the extension of the reaction zone into areas not contacted by electrolyte and not covered by Pt.15 It is expected that systematic studies involving the use of the novel arrangement described in this work will afford insight into mass transport/reactivity relationships at these nanointerfaces including the role of structural modifications induced by local changes in the extent of hydration associated with faradaic processes of direct relevance to the operation of PEFC electrodes. Experimental Section Cell Design and Assembly. The electrochemical cell designed for the experiments herein described is comprised of an © 2006 American Chemical Society
18082 J. Phys. Chem. B, Vol. 110, No. 37, 2006
Letters
Figure 1. Schematic of the electrochemical solid-state cell showing the arrangement that allows for the tip of the platinum working electrode to be displaced along the z-axis normal to the surface of the electrolyte membrane.
upper (Teflon) and a lower (stainless steel) cylinder which are used to compress a Nafion 117 solid electrolyte membrane (3 × 3 mm2, see below) placed between the ends of the cylinders (see Figure 1). A set of holes drilled both along and normal to the main axis of the cylinders allow access of gas at the desired degree of humidification to either side of the Nafion membrane. As shown in Figure 1, the working electrode is a platinum wire ending in a very sharp tip (see below) attached to the piezodriven head of a Molecular Imaging PicoScan scanning tunneling microscope (STM). This wire is introduced through the vertical hole in the Teflon cylinder and can be lowered and raised along that axis using the z-scanner of the STM, allowing the tip to be inserted into or retrieved from the Nafion with a high degree of control to yield the desired gas|Pt|Nafion threephase interface. A circular piece of Pt gauze of a much larger area compared to that of the tip placed on a recess machined in the lower cylinder is used as a counter-reference electrode in a twoelectrode cell configuration. Although not used in the present studies, the electrochemical cell incorporates, in addition, a thick Pt wire inserted through an off-center hole along the z-direction, as shown in the figure. The potential of this wire could be poised to the value of a reversible hydrogen electrode (RHE) in the same solution by passing a small negative current using the gauze as a cathode until sufficient hydrogen is evolved. It is envisioned that, once the current is interrupted, the potential of the wire will remain practically fixed at the RHE for the duration of the experiments and serve as a reliable reference electrode in a three-electrode cell configuration. However, since for the experiments described in this work the currents flowing through the tip were for the most part very low, it was found to be unnecessary to implement this option. It is expected that this will not be the case for studies involving long time measurements for which the net charge passing through the working electrode tip could be considerably larger and thus lead to polarization of the gauze. All potentials in this work are referenced to a reversible hydrogen electrode in the same solution (RHE), which is about 50 mV negative to the onset potential for hydrogen evolution, as measured using the Pt tip as a working electrode. During the experiments, the cell is isolated from the ambient laboratory using a commercially available environmental chamber (Molecular Imaging) which allows the desired gases to flow into the cathode and anode chambers. The Pt tip electrode was formed by holding a 0.25 mm Pt wire (Alfa Aesar, 99.95% metals basis) at about 20 mm from its end with a wire cutter at a 45° angle and then gently increasing the cutting pressure while pulling the wire apart with enough force to break it. A scanning electron microscope (SEM) image of a tip created in this fashion, which had been used as a working electrode in the cell, is shown in Figure 2.
Figure 2. Scanning electron microscopy images of a platinum tip recorded after the experiments were completed. Shown in the bottom image is an enlarged version of the tip region, where the solid line represents the shape of the end of the tip, as determined from coulometric analyses of voltammetric curves obtained as a function of the depth of immersion (see text for details and caption, Figure 8).
The Nafion 117 membrane was cleaned by boiling in a 1:1:1 mixture of concentrated sulfuric acid, 30% hydrogen peroxide, and ultrapure water (Barnstead, 18.3 MΩ) for 1 h to remove impurities, including organic contaminants. It was then rinsed and boiled first in ultrapure water and subsequently in 0.5 M sulfuric acid to achieve full protonation. The Teflon cylinder was boiled first in concentrated sulfuric acid and then soaked three times with ultrapure water for 30 min each. Other cell components were rinsed and boiled in ultrapure water. Immediately before assembling the cell and beginning an experiment, the tip, gauze, and auxiliary electrode were annealed in a hydrogen flame for ∼2 s and then quenched in ultrapure water. Shortly thereafter, the tip was attached to the STM head and carefully lowered into the center hole of the upper cylinder to ∼100 µm above the Nafion surface. The environmental chamber was then attached and the purging with desired gases initiated. After ∼1 h, the final approach to the surface was started by engaging the stepper motor of the STM head while the Pt tip was held at a potential well within the hydrogen evolution region. In this fashion, a high current would be expected to flow immediately after mechanical (and thus) ionic contact was established, allowing the piezo to stop automatically to avoid piercing the membrane. Humidification of the purge gases to ∼70% relative humidity, as measured with a Testo Hygrotest monitor, was achieved (when required) by bubbling through a 5 cm column of ultrapure water. All experiments were conducted at room temperature under ambient pressure. Electronics. The potential of the tip, which, by design, is always held at ground in order to improve the signal-to-noise ratio when measuring currents, was set by applying an opposite value to the counter-reference electrode, that is, Pt gauze. Cyclic voltammetry curves were recorded using an Agilent 33220A arbitrary waveform generator capable of producing a smooth triangular voltage ramp. The small currents measured were
Letters
Figure 3. Cyclic voltammograms recorded for a tip just above (dashed curve) and after insertion into the Nafion surface (solid curve) recorded at a scan rate of 100 V/s in a humidified Ar atmosphere.
J. Phys. Chem. B, Vol. 110, No. 37, 2006 18083
Figure 4. Series of cyclic voltammograms in the potential range 0.051.35 V vs RHE obtained by slowly varying the immersion depth of the tip electrode. Each voltammogram was recorded as soon as a prescribed area of contact was reached.
preamplified and converted to a voltage signal with a ratio of 1 V/nA in the microscope head and captured by the A/D converter input of the data acquisition board (National Instruments NI 6014) used to interface the custom-made software with the STM. The software was also used to output an analogue signal to the microscope controller to set the piezo-controlled position of the tip along the z-axis, allowing automated acquisition of data as a function of z. Results and Discussion Determination of the Area of Contact of the Nanoelectrode|Nafion Interface. In the absence of a reactant, the electrical behavior of the Pt tip|Nafion interface will be determined by different contributions: stray capacitances from, for example, metal|air interfaces, double layer (Pt|Nafion) capacitance, as well as pseudocapacitances from, for example, hydrogen adsorption/ desorption and oxide formation/reduction. For voltammetric measurements, all of these components will be directly proportional to the potential scan rate. Whereas stray and to a large extent double layer capacitances are independent of the applied potential, pseudocapacitive contributions are often strongly potential dependent and characteristic of the specific surface of a given metal in contact with the electrolyte, and in the case of Pt provide means of determining real electrode areas in contact with the electrolyte. Because of the very small dimensions of the Pt tip, it is thus highly desirable to record voltammetric curves at very high scan rates to enhance pseudocapacitive currents that would otherwise not be detectable. Shown in dashed lines in Figure 3 is a cyclic voltammogram of a sharp Pt tip just above the Nafion surface recorded in the cell depicted in Figure 1 purged on both sides with humidified Ar over the range 0.05-1.6 V at a scan rate of 100 V/s displaying a box-type voltammogram as expected for a (stray) capacitor. Also shown in this figure (see solid line) is the cyclic voltammetry of the tip recorded once contact with the Nafion was established. The features in this case are characteristic of clean polycrystalline Pt in, for example, sulfuric acid16 or Nafion1 devoid of impurities superimposed on the stray capacitance contribution. Coulometric analysis of the pseudocapacitive voltammetric features associated with hydrogen desorption using a nominal value of 220 µC/cm2 for a monolayer of hydrogen desorbed from the Pt surface yielded an actual area of contact in this case of ∼4.6 × 10-10 cm2 or, equivalently, approximately 630 000 Pt surface atoms. It was assumed, for this calculation, that 85% of a hydrogen monolayer was adsorbed at the lower reversal potential.17
Figure 5. Plots of the integral under the hydrogen desorption (solid circles, integration limits (IL): 0.05-0.55 V vs RHE) and hydrogen adsorption (empty circles, IL: 0.07-0.4 V vs RHE) vs the charge associated with oxide formation (IL: 1.1-1.3 V vs RHE) features determined from a coulometric analysis of the data in Figure 4 (see text for details). The right y-axis shows the corresponding number of surface sites.
This cell arrangement makes it possible to displace the electrode vertically (along the z-axis, normal to the Nafion surface) using the STM electronics, allowing the tip to be inserted into and/or retrieved from the membrane and thus modulate the Pt tip|Nafion area of contact. Shown in Figure 4 is a series of cyclic voltammograms in the potential range 0.051.35 V versus RHE obtained by slowly varying the position of the tip along the negative z-axis (into the Nafion), where each curve was recorded once a prescribed area of contact was reached using an on-line integrating routine. Plots of the charge under the hydrogen desorption (empty circles, integration limits (IL): 0.05-0.55 V vs RHE) and hydrogen adsorption (solid circles, IL: 0.07-0.4 V vs RHE) versus the charge associated with oxide formation (IL: 1.11.3 V vs RHE) determined from coulometric analyses are shown in Figure 5. The charge attributed to the formation of a partial, although fixed, coverage of PtO was multiplied by an empirically determined constant factor of 1.85, assuming a baseline of 21 pA for all integrations. The upper integration limit for H adsorption was somewhat arbitrary, because the separation between the PtO reduction and the first H adsorption peaks is not clearly delineated. As is evident from these data, PtO formation charges are largely proportional to those obtained from H desorption. Shown on the right y-axis, Figure 5, is the corresponding number of surface sites determined from the
18084 J. Phys. Chem. B, Vol. 110, No. 37, 2006
Figure 6. Cyclic voltammograms for a Pt tip inserted into Nafion under less humidified conditions (see text for details) for two different positive potential limits, 1.4 (solid curve) and 1.6 V vs RHE (dashed curve). The latter trace was recorded following 12 s of continuous cycling. An additional oxide reduction peak appears at 0.16 V vs RHE. The current axis of the curve in dashed lines was divided by 1.13 to allow an area normalized comparison with the curve in solid lines.
coulometric analysis, which yielded values that ranged from approximately 2.6 × 105 to 7 × 106. Although not as noticeable in the data displayed in Figure 4, small traces of oxygen could not be avoided in other similar experiments, yielding a charge associated with hydrogen adsorption that is slightly larger than that of hydrogen desorption. Inspection of this voltammogram in solid lines in the Figure 3 (tip inserted into the Nafion) curve reveals that the peak potential associated with the reduction of platinum oxide, Ep, is shifted negative by ∼0.1 V relative to that typically found for Pt under more conventional conditions, that is, 0.8 V versus RHE. This behavior persisted upon a decrease of the scan rate to values of only a few volts per second, as well as following extensive potential cycling. Shifts in the oxide reduction potential were also found in similar experiments upon purging the side of the membrane where the Pt tip was located with dry Ar, while retaining full humidification of its underside (see Figure 6). Following 12 s of continuous cycling, the positive scan limit was increased from 1.45 (see solid line) to 1.6 V versus RHE (see dashed line). For ease of comparison, the curves in this figure were normalized to account for an increase in the total area caused by changes in the interfacial area of contact or other factors (see caption for details). Particularly noteworthy is the emergence of a feature centered at 0.2 V, which overlaps the hydrogen adsorption region, leaving, otherwise, the feature at 0.30 V versus RHE almost unchanged. Despite these changes, however, the total charge determined from the scan during the oxidation processes was in very good agreement with that determined from the corresponding reduction processes. On this basis, the pronounced peak at 0.2 V in the scan in the negative direction can be attributed to the overlap of oxide reduction and hydrogen adsorption. Similar effects have been reported for conventional Pt electrodes in aqueous acidic electrolytes.18 In particular, oxide formation has been found to become increasingly more irreversible as the potential at which it is formed is increased above 0.93 V versus RHE, leading to a corresponding shift in Ep toward more negative potentials as the oxide is reduced.19 An analogous phenomenon has also been noted for Pt electrodes polarized at, for example, 1.2 V for several tens of hours. Rather interestingly, the emergence of the more negative peak in the hydrogen adsorption region has been reported upon reduction
Letters
Figure 7. Series of cyclic voltammograms recorded following insertion of the Pt tip into the Nafion in 2 nm steps (along the z-axis) every 70 ms until a total displacement of 100 nm was reached. The curves shown represent each sixth step, i.e., z ) 2, 14, 26, 38, 50, 62, 74, 86, and 98 nm.
of Pt oxides formed at potentials as high as 1.9 V.20 In other experiments, shifts in Ep of ∼0.2 V in the negative direction have been reported for small platinum particles, a few nanometers in diameter, supported on carbon.21,22 Although many of the conditions used for the experiments described in this work are indeed much different than those employed in some of the reported studies, the similarities between the results obtained are indeed suggestive of a common underlying mechanism responsible for the effects observed. Undoubtedly, more experiments will be required before these phenomena are better understood. Estimates of the Tip Geometry. A series of cyclic voltammograms were recorded following insertion of the Pt tip into the Nafion in 2 nm steps (along the z-axis) every 70 ms until a total displacement of 100 nm was reached. Shown in Figure 7 are curves collected every sixth step, that is, z ) 2, 14, 26, 38, 50, 62, 74, 86, and 98 nm. As expected, all pseudocapacitive features increase with the depth of insertion, a behavior consistent with a monotonic increase in the Pt|Nafion area of contact. The slightly larger charge associated with hydrogen adsorption compared to its reduction in this case is in all likelihood related to the presence of a small amount of oxygen which complicates a reliable coulometric analysis using those features as a basis for estimating the area of contact. A much better approach is to use the oxide formation charge, which is significantly less compromised by the presence of oxygen (see below). It must be emphasized that the charges obtained from the coulometric analysis based on the oxide formation region are proportional to the actual Pt|Nafion area of contact. If one assumes that the shape of the tip is (rotationally) symmetric about the z-axis and that insertion of the tip does not deform the surface of the Nafion, one can deduce the shape of the tip from the coulometric data as a function of the depth of immersion. Since the shape is not known a priori, it is a fairly good approximation to assume that the incremental area is that of the curved wall of a cylinder plus the area of the bottom annulus not covered by the cylinder below. For each immersion step, i, the cylinder radius, ri, and height, h, are assumed constant, as indicated in the inset of Figure 8, that is,
Ai - Ai-1 ) π[2hri + (ri2 - ri-12)] from which ri may be shown to be given by
(1)
Letters
J. Phys. Chem. B, Vol. 110, No. 37, 2006 18085
ri ) -h +
x
h2 + ri-12 +
Ai - Ai-1 π
(2)
Although largely qualitative and approximate, the shape of the symmetric tip inferred from the experimental data, using the recursive formulas above, is shown in Figure 8, where a common scaling of the x- and y-axes was used. The observed profile agrees well with what is expected from the SEM image of a typically used tip (see Figure 2). To initiate the recursive operation, it was assumed that the very first area determined was that of a flat disk, which yielded a radius, r1, of ∼25 nm, which is roughly the smallest surface area that can be measured with the measured signal to noise and corresponds to about 35 000 surface atoms. Although surprising, in view of all the assumptions made, a direct comparison between the coulometrically resolved tip shape and an actual shape as found from the SEM yielded very good agreement. Effects of Oxygen Purging. Shown in Figure 9 are cyclic voltammograms of a Pt tip in contact with Nafion under either an Ar atmosphere (solid line) or an O2 atmosphere (dashed line) for Pt|Nafion interfaces of approximately the same size as determined from the oxide formation feature (see above). As may have been expected, the curve recorded in O2 displays a much higher current in a potential region negative to 0.7 V versus RHE. Additional measurements were attempted to gain further insight into the reaction kinetics using chronoamperometric techniques. For these experiments, the potential was stepped from a value at which O2 is not reduced, Ein, to a value negative enough for the reaction to ensue, Efin, but positive to the onset of hydrogen evolution. Under these conditions, all pseudocapacitive effects due to hydrogen adsorption would disappear after a very short period of time following the step. The results of one of these such experiments for Ein ∼ 0.9 and Efin ∼ 0.13 V versus RHE are shown in Figure 10 (see right panels, where the left panels show the cyclic voltammetric curves at 100 V/s) obtained in an Ar atmosphere (upper panels) and in an O2 atmosphere (lower panels). Unlike the results found in Ar, for which the current remains negligible and unchanged following the potential step, a clear transient was observed in the presence of O2, yielding after a few seconds a relatively constant and finite value. The fact that oxygen is present both in the gas phase and dissolved in the Nafion introduces significant complications to the analysis of these data which will require that a more complicated modeling be involved. These problems are currently being examined in this laboratory and will be reported in due course. Nevertheless, although largely qualitative, the model system introduced in this work can provide interesting information regarding other aspects of electrocatalysis at the Pt|nanosurface interface. In particular, shown as a solid line in Figure 11 is the cyclic voltammetry of a Pt tip immersed in Nafion to form an interface involving approximately 7.5 × 106 atoms while both the upper and lower chambers were purged with humidified O2. The tip was then retracted by 100 nm while recording continuously voltammetric curves at 100 V/s, yielding after 0.1 s the voltammogram shown in dashed lines. Coulometric analysis of these two curves using oxide formation as a basis yielded a decrease in the area of contact of 19% to 6.1 × 106 sites; however, the current due to O2 reduction, iO2, using the value measured at 0.1 V versus RHE on the upward scan, virtually doubled. Shown in the left inset is a plot of the coulometrically evaluated transient interfacial areas (circles) and the current due to O2 reduction, iO2, at 0.1 V versus RHE measured in 0.1 s
Figure 8. Tip profile from radii calculated for every immersion step from the total calculated interfacial area (see text for details). By assumption, the cross section shown is simply a reflection of the curve along the axis z ) 0. For a better physical representation, the plot was drawn with a common scaling of the x- and y-axes. Inset: sketch of the tip end modeled as a stack of discrete cylinders with heights, h, and radii, ri, used to calculate the tip profile. The thick lines mark the surface Ai - Ai-1 that adds to the total surface when the tip is immersed from position i-1 to i.
Figure 9. Cyclic voltammograms recorded using the same Pt tip in an Ar atmosphere (solid line) and in an O2 atmosphere (dashed line) at 100 V/s. For each curve, the tip position was adjusted until the desired number of contact sites was reached, approximately 2.9 × 106 as determined from the PtO formation charge.
intervals (triangles) following tip retraction, t ) 0. As is evident from these data, the interfacial area decreases to reach a fairly constant value at about 0.3-0.4 s, whereas iO2 initially increases from 0.64 to 1.4 nA and later decreases to 0.77 nA, a value still 20% higher than that observed initially. In other words, the loss in interfacial area, from initially 7.5 to 5.6 × 106 sites after 2 s, is accompanied by a net increase in iO2. In fact, normalization of the current by the corresponding interfacial areas would make this effect even more marked, that is, a gain in specific activity of 61%. The most likely explanation of this seemingly peculiar phenomenon involves a plastic deformation induced by the Pt tip retraction, which brings about formation of an adherent Nafion meniscus, as depicted schematically in the right inset of Figure 11. Such a distortion would decrease the diffusion length through the thin meniscus and thereby enhance mass transport from the gas phase. As time elapses, however, water will be generated, leading to a swelling of the meniscus and to a net decrease in the observed iO2 value as the results show, an effect that would not affect the net interfacial area much, which is also consistent with the data collected. Dynamic Effects. The Pt tip|Nafion assembly described in Figure 1 allows changes in the membrane to be monitored by measuring the area of contact as a function of time while keeping the tip position fixed. In one such series of experiments, the tip was initially positioned so as to barely touch the Nafion (as determined by cyclic voltammetry) and the potential then cycled continuously at a fixed scan rate, that is, 100 V/s. Shown in
18086 J. Phys. Chem. B, Vol. 110, No. 37, 2006
Letters
Figure 10. Cyclic voltammograms (scan rate 100 V/s, left panels) and chronoamperometric plots for Ein ∼ 0.9 and Efin ∼ 0.13 V vs RHE (right panels, see text for details) obtained in an Ar atmosphere (upper panels) and an O2 atmosphere (lower panels). The number of contact sites determined from oxide formation was approximately 2.6 × 105 in both experiments.
Figure 11. Cyclic voltammograms recorded in the cell shown in Figure 1 at a scan rate of 100 V/s, before (solid line) and 0.1 s after retracting the tip by 100 nm (dashed line). Oxygen is present in the gas phase and in the membrane. The left inset shows the transient interfacial areas (see text for details, circles) and the oxygen reduction activities at 0.1 V vs RHE in 0.1 s intervals (triangles) following tip retraction, t ) 0. The inset on the right is a schematic representation of the proposed interfacial structure immediately after retraction (top) and after 2 s of continuous cycling (bottom).
Figure 12 is every other curve of a series of voltammetric cycles collected over a period of ∼60 s in 4 s intervals in a dry Ar atmosphere, while purging the opposite side of the Nafion membrane with humidified O2. Evidence for an increase in the area of contact is provided by the monotonic increase in the pseudocapacitive features. Also noticeable is a tilt toward negative currents at potentials more negative than 0.75 V due to oxygen reduction. Shown in the inset of Figure 12 is a plot of the interfacial area determined from the platinum oxide features based on all of the voltammetric curves recorded. Although a quantitative account for the shape of this curve is outside the scope of this contribution, this increase in area, as invoked in the previous section, is due to the swelling of the membrane induced by water generation during oxygen reduction.
Figure 12. Series of cyclic voltammograms recorded in a dry Ar atmosphere while purging the opposite side of the Nafion membrane with humidified O2 for a fixed Pt tip immersion depth (only every other scan is given). Inset: interfacial area determined from the platinum oxide features from voltammetric curves recorded at 4 s intervals.
Conclusion A new technique has been herein described that allows the formation and characterization of single nanoscale Pt|Nafion interfaces with direct relevance to fuel cell applications. The ability to isolate the response of a single nanoscale electrode avoids problems associated with statistical averaging of ensembles of particles of different sizes, and with effects associated with the presence of other particles in the neighboring region, such as those derived from the overlap of diffusion layers.12 As illustrated, the interfacial areas can be determined quite accurately by coulometric analyses of voltammetric curves both in inert and oxygen-containing atmospheres using the platinum oxide formation as a basis. This technique opens up new prospects for gaining a better understanding of the factors involved in electrocatalysis at the gas|nanoparticle|solid electrolyte interfaces, including mass transport of reactant and products involving both the gas phase and the electrolyte. A better theoretical understanding of the interplay between the two forms of mass transport and its resulting effect on the measured
Letters currents will be required to extract reliable kinetic parameters from the data collected. This aspect, however, is beyond the scope of this contribution. Illustrations were given of effects derived from plastic deformations of Nafion induced by mechanical means as well as membrane swelling which accompanies oxygen reduction. Acknowledgment. This work was supported by grants from Forschungsallianz Brennstoffzelle Baden-Wuerttemberg, Deutsche Forschungsgemeinschaft, and the U.S. Department of Energy. Fruitful discussions with Dr Attila Palencsar are gratefully acknowledged. References and Notes (1) Parthasarathy, A.; Martin, C. R.; Srinivasan, S. J. Electrochem. Soc. 1991, 138, 916. (2) Mitsushima, S.; Araki, N.; Kamiya, N.; Ota, K. J. Electrochem. Soc. 2002, 149, A1370. (3) Beattie, P. D.; Basura, V. I.; Holdcroft, S. J. Electroanal. Chem. 1999, 468, 180. (4) Basura, V. I.; Beattie, P. D.; Holdcroft, S. J. Electroanal. Chem. 1998, 458, 1. (5) Buchi, F. N.; Wakizoe, M.; Srinivasan, S. J. Electrochem. Soc. 1996, 143, 927. (6) Katakura, K.; Hinatsu, J. T.; Inatomi, K.; Inaba, M.; Ogumi, Z.; Takehara, Z. Denki Kagaku 1996, 64, 711.
J. Phys. Chem. B, Vol. 110, No. 37, 2006 18087 (7) Watanabe, M.; Igarashi, H.; Yosioka, K. Electrochim. Acta 1995, 40, 329. (8) Gamez, A.; Richard, D.; Gallezot, P.; Gloaguen, F.; Faure, R.; Durand, R. Electrochim. Acta 1996, 41, 307. (9) Antoine, O.; Bultel, Y.; Durand, R. J. Electroanal. Chem. 2001, 499, 85. (10) Paulus, U. A.; Schmidt, T. J.; Gasteiger, H. A.; Behm, R. J. J. Electroanal. Chem. 2001, 495, 134. (11) Higuchi, E.; Uchida, H.; Watanabe, M. J. Electroanal. Chem. 2005, 583, 69. (12) Watanabe, M.; Sei, H.; Stonehart, P. J. Electroanal. Chem. 1989, 261, 375. (13) O’Hayre, R.; Feng, G.; Nix, W. D.; Prinz, F. B. J. Appl. Phys. 2004, 96, 3540. (14) Chen, S. L.; Kucernak, A. J. Phys. Chem. B 2004, 108, 3262. (15) Paulus, U. A.; Veziridis, Z.; Schnyder, B.; Kuhnke, M.; Scherer, G. G.; Wokaun, A. J. Electroanal. Chem. 2003, 541, 77. (16) Jerkiewicz, G.; Vatankhah, G.; Lessard, J.; Soriaga, M. P.; Park, Y. S. Electrochim. Acta 2004, 49, 1451. (17) Gilman, S. J. Electroanal. Chem. 1964, 7, 382. (18) Conway, B. E. Prog. Surf. Sci. 1995, 49, 331. (19) Angerstein-Kozlowska, H.; Conway, B. E.; Sharp, W. B. A. J. Electroanal. Chem. 1973, 43, 9. (20) Tremiliosi-Filho, G.; Jerkiewicz, G.; Conway, B. E. Langmuir 1992, 8, 658. (21) Takasu, Y.; Fujii, Y.; Yasuda, K.; Iwanaga, Y.; Matsuda, Y. Electrochim. Acta 1989, 34, 453. (22) Frelink, T.; Visscher, W.; Vanveen, J. A. R. J. Electroanal. Chem. 1995, 382, 65.
