Dynamics of Oxidation of Well-Defined Adsorbed CO Phases on Pt

Nov 4, 2014 - On this basis, the oxidation of the fully assembled (2 × 2)-3CO phase occurs at the edge of the domain and propagates through the entir...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCC

Dynamics of Oxidation of Well-Defined Adsorbed CO Phases on Pt(111) in Aqueous Acidic Electrolytes: Simultaneous in Situ Second Harmonic Generation and Differential Reflectance Spectroscopy Iosif Fromondi, Huanfeng Zhu, Zhange Feng, and Daniel Scherson* Ernest B Yeager Center for Electrochemical Sciences and The Department of Chemistry Case, Western Reserve University, Cleveland, Ohio 44106, United States ABSTRACT: The oxidation dynamics of the well-defined (2 × 2)-3CO (θCO = 0.75) and √19 × √19R23.4°-13CO (θCO = 0.68) CO adlayer phases on quasi-perfect Pt(111) facets have been examined in COsaturated 0.1 M H2SO4 aqueous solutions via a combination of simultaneous, in situ, time-resolved, second harmonic generation (SHG, λinc = 590 nm), normalized differential reflectance spectroscopy (ΔR/R, λinc = 633 nm), and potential step techniques. For both of these phases, the onset of oxidation of COads was found to be delayed with respect to the time at which the potential was stepped to values positive enough, Eox, to promote oxidation of the entire CO adlayers. This induction time, τind, was much longer for the (2 × 2)-3CO (few tens of milliseconds) compared to the √19 × √19R23.4°-13CO phase (few milliseconds) and decreased monotonically for both of the phases as Eox was increased. Furthermore, for a fixed Eox, the oxidation rates, as measured by the optical techniques, were higher for the √19 × √19R23.4°-13CO compared to the (2 × 2)-3CO phase. In the case of the (2 × 2)-3CO phase, the transient in situ optical data collected for Eox = 0.98 V vs RHE could be quantitatively accounted for by a model that assumes all sites on the surface are occupied either by CO or bisulfate. In fact, excellent statistical fits to the θCO transients for this latter phase derived from the SHG data could be obtained using Avrami’s nucleation and growth model. On this basis, the oxidation of the fully assembled (2 × 2)-3CO phase occurs at the edge of the domain and propagates through the entire island until the entire process is completed. Similar experiments involving CO adlayers on Pt(111) facets below saturation coverages (i.e., θCO < 0.75) formed by oxidative stripping of the fully assembled (2 × 2)-3CO phase yielded much longer τind values than surfaces with similar θCO prepared by dosing the otherwise bare Pt(111) surfaces. These findings strongly suggest that the rates of oxidation of COads increase with the number of non-CO-covered neighboring sites on the surface.

■

INTRODUCTION The electrochemical oxidation of adsorbed CO (COads) on well-defined metal surfaces in aqueous electrolytes has received extraordinary attention over the past two to three decades.1,2 Much of the interest in this specific area stems from the ability of COads to block active sites on highly catalytic metals, such as Pt, for reactions of technological importance, including hydrogen3 and methanol oxidation.4,5 Considerable progress toward gaining an understanding of the nuances associated with the mode of adsorption and reactivity of COads has been made based on information derived from in situ infrared reflection absorption spectroscopy, IRAS,6−11 and structural techniques.12−20 In particular, COads on a given metal elicits bands unique to the nature of the adsorption site, e.g., on top, bridge, and hollow. Such remarkable specificity has made it possible to examine site occupations as a function of the applied potential, E,21 and CO coverage, θCO,22 in a highly quantitative manner. It has also enabled to unveil subtle differences in the nature of CO adlayers on well-defined Pt single crystal surfaces induced by the presence of CO in solution10 formed via different protocols © 2014 American Chemical Society

as well as by the way in which submonolayers are formed, i.e., dosing and oxidative stripping.23 Furthermore, the relatively large infrared cross sections for COads on Pt and other metals enables detection of very small θCO and also of minute shifts in the COads stretching frequencies induced by changes in the field across the interface as E is varied.22,23 Additional insights into structural aspects of CO adlayers on Pt surfaces have been obtained from studies involving Pt(111) in acidic aqueous electrolytes using in situ scanning tunneling microscopy (STM),12−15 surface X-ray scattering (SXS)16−20 and, more recently, sum frequency generation (SFG).24 Based on the information so far collected, CO adsorption from COsaturated acidic aqueous solutions can yield two distinct and well-defined COads phases denoted as (2 × 2)-3CO (θCO = 0.75) and √19 × √19R23.4°-13CO (θCO = 0.685), over regions of lower and higher potentials, respectively.13 In fact, Received: September 5, 2014 Revised: October 31, 2014 Published: November 4, 2014 27901

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

Figure 1. Plots of Ipp(2ω) (panel A) and ΔR/R (Eref = Eads; panel B) vs time recorded simultaneously for a Pt(111) microfacet in CO-saturated 0.1 M H2SO4 following a potential step at t = 0, from Eads = 0.1 V, i.e. (2 × 2)-3CO phase, to Eox = 0.98 (curve a), 0.99 (b), 1.00 (c), and 1.02 V (d) for τads = 1 s and τox = 0.2 s (see potential−time protocol in the inset of panel D). The black solid lines in panel B are the derivatives of the best fits to the ΔR/R data. Panels B and D display corresponding Ipp(2ω) and ΔR/R data recorded under otherwise identical conditions for Eads = 0.6 V, i.e., √19 × √19R23.4°-13CO phase, to Eox = 0.98 (a), 1.00 (b), and 1.02 V (c). Each of the curves in these panels represents the average of ca. 600 Acq. Photon counter: SR430, bin width = 163.84 μs, 1000 pts/record.

SXS experiments performed both in gas phase25 and also in situ in perchloric and sulfuric acid aqueous solutions17 have afforded evidence for a marked Pt surface relaxation induced by the √19 × √19R23.4°-13CO phase involving both a vertical expansion for Pt atoms predominantly beneath near top sites as well as lateral displacements. Studies aimed at elucidating the factors that control the dynamics of COads electrooxidation have relied mostly on electrochemical techniques, particularly cyclic voltammetry and potential step chronocoulometry on massive polycrystalline26,27 and single crystal surfaces.9,28−34 In the case of Pt(111), the rates of COads oxidation have been found to be sensitive among other factors to the cooling atmosphere following surface flame annealing35 and the temperature at which the oxidation process is carried out.32 From a mechanistic viewpoint, the consensus appears to have been reached regarding the need for an oxygenated adsorbed species neighboring COads for the oxidation of the latter to ensue, i.e., Langmuir−Hinshelwood mechanism.36 However, analyses of chronoamperometric experiments for COads oxidation on Pt(111) and to a lesser extent on the other low index faces in solutions devoid of CO, for which θCO has been estimated to be 0.68 and thus lower that the corresponding saturation coverage for the (2 × 2)3CO phase,37 appear consistent with two widely different pathways: nucleation and growth (NG)28 and the mean-field approximation (MFA).33,34 In particular, NG regards the adsorbed reactant to be immobile, i.e., no surface diffusion,

where the reaction takes place only at the interface between the CO covered and nominally uncovered Pt areas, causing a progressive reduction in the size of the CO islands and their final disappearance, whereas MFA assumes a high COads mobility leading to surface randomization at rates higher than COads oxidation. In fact, the MFA model has been claimed to yield better agreement with the experimental data than NG for CO on Pt(111) in neat 0.5 M H2SO4.3,33 More recently, however, experiments involving in situ, spatially resolved, dualbeam coincidence SHG in CO-saturated solutions, under conditions that would lead to the formation of a well-defined (2 × 2)-3CO,38 were found to be qualitatively inconsistent with the MFA model. The present contribution examines various aspects of the dynamics of oxidation of COads on quasi-perfect Pt(111) microfacets in CO-saturated 0.1 M H2SO4 using a combination of simultaneous, time-resolved, second harmonic generation (SHG), normalized differential reflectance spectroscopy (ΔR/ R), and potential step techniques.39 This approach exploits the specificity of the optical probes to the presence of COads (SHG) and bisulfate ions (ΔR/R) for monitoring their temporal evolution at a constant applied potential. The latter offers a great advantage in that complex effects due to the convolution of time and potential, as e.g., in linear scan voltammetry, can be effectively eliminated. Of primary interest is to compare the oxidation dynamics of the well-defined hexagonal close-packed (2 × 2)-3CO and √19 × √19R23.4°27902

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

(i) The time elapsed between application of the potential step (t = 0) and the onset of COads oxidation, to be referred to hereafter as induction time, τind, decreases as Eox increases. This phenomenon is in all likelihood derived from an increase in the probability of forming a defect site (probably at the periphery of the facet), which triggers the oxidation of the CO adlayer. A somewhat related effect is found in metal electrodeposition, where the probability of creating new nuclei increases with the applied overpotential, η.43 Not surprisingly, τind was found to vary for (111) facets of the same or different Pt single crystal spheres. (ii) The time required for Ipp(2ω) and ΔR/R to achieve half their initial values, τ1/2, becomes shorter as Eox increases. A better illustration of this behavior is afforded by plots of the derivatives of the ΔR/R transients (see smooth curves in panel B in this figure). The decrease in the half-widths of these curves as E ox is increased is in qualitative agreement with chronocoulometric measurements for experiments performed for massive Pt(111) surfaces in aqueous acidic solutions devoid of CO (see e.g., refs 28 and 34). (iii) There is an initial increase in both optical signals immediately following application of the potential step, which can be attributed to the onset of the (2 × 2)-3CO to the √19 × √19R23.4°-13CO phase transition (vide inf ra). √19 × √19R23.4°-13CO|Pt(111). Virtually identical measurements were performed for Eads = 0.6 V under conditions leading to the full assembly of the √19 × √19R23.4°-13CO phase (see right panels, Figure 1) for Eox = 0.98, 1.00, and 1.02 V vs RHE. In this case, τind was found to be an order of magnitude smaller than the values observed for the (2 × 2)-3CO phase, which serves to prove that the time constant for the oxidation of the (2 × 2)-3CO layer is not controlled by the RC constant of the cell. This phenomenon can be attributed to the more open character of the √19 × √19R23.4°-13CO phase compared with the (2 × 2)-3CO phase and also to the possible presence of multiple domains. Both of these factors would increase the number of empty (non-CO-covered) sites neighboring COads sites, which serve as triggers for the full oxidation of the adlayer. This explanation is supported by the results obtained for CO adlayers involving θCO values below saturation prepared by dosing techniques, as will be discussed in the next section. In analogy with the results obtained for the (2 × 2)-3CO phase, the half-widths of the derivative curves for the √19 × √19R23.4°-13CO phase (not shown here) decreased as Eox increased. b. CO Adlayers below Saturation Coverage. Attention in the subsections below will focus on the results of experiments involving CO adlayers for θCO below saturation formed either by oxidative stripping or dosing (vide inf ra). Dosing. For these experiments, the protocol involved the initial full oxidation of a (2 × 2)-3CO layer at Eox = 0.98 V to generate a CO free surface (τox = 0.2 s), followed by a potential step to Eads = 0.1 V, allowing CO adsorption (or dosing) to ensue. After a certain period of time, τdos, the potential was stepped to Eox = 0.98 V, while monitoring the optical signals. As before, numerous runs were coadded to improve the statistics. Surfaces involving different θCO were prepared by adjusting τdos, as originally described by Chang and Weaver.23 Shown in Figure 2 are Ipp(2ω) (panel A) and ΔR/R vs time (panel B) plots recorded for four different values of τdos, i.e., 1 (curve a), 0.6 (b), 0.4 (c), and 0.3 s (d), where each of the curves represents the average of ca. Acq = 500. As clearly evident from these results, for the shorter τdos (see curves c and d) or,

13CO adsorbed phases. As mentioned above, most electrochemical experiments involving oxidation of COads on Pt(111) have been performed in solutions devoid of CO, for which the resulting CO adlayers are far less structurally understood.

■

EXPERIMENTAL SECTION Measurements were conducted in CO-saturated 0.1 M H2SO4 solutions prepared with ultrapure chemicals and (111) faceted single crystal Pt beads grown by methods specified elsewhere as working electrodes.40 The instrumental array involved in the dual optical monitoring of single Pt(111) facets has also been described in detail in previous communications.41,42 Values of the transient intensities of the second harmonic light, Ipp(2ω), where pp refers to p-input and p-output polarizations, were collected for an incident wavelength λinc = 590 nm and of ΔR/ R = [R(Esam) − R(Eref)]/R(Eref)], where R(Ei) is the intensity of the reflected light (λinc = 633 nm) measured at a potential Ei, i = ref (reference) and sam (sampling), respectively, at an angle of incidence of ca. 60°. The cross sections for SHG, as well as the changes in ΔR/R induced by the interfacial processes of relevance to this work, are indeed very small. It becomes therefore essential to average the results of hundreds to thousands of replicate experiments in order to increase the signal-to-noise and thereby generate meaningful time-resolved data. Such stringent requirements make it essential to assemble the CO adlayer under carefully controlled conditions so as to render identical interfaces prior to the application of the potential protocol (vide inf ra). It should be stressed that for each set of measurements involving a single adsorption or oxidation potentials, Eads or Eox, respectively, the faceted Pt single crystal sphere was flame-annealed to restore its microstructural characteristics, particularly at the edges of the facets. This tactic was implemented to mitigate problems associated with surface quality degradation described in a previous communication.39

■

RESULTS AND DISCUSSION One of the main objectives of this work is to compare the dynamics of oxidation of the (2 × 2)-3CO and √19 × √19R23.4°-13CO phases on Pt(111) microfacets in COsaturated 0.1 M H2SO4 solutions via a combination of in situ SHG and ΔR/R and potential step techniques. The following subsections summarize our findings. a. (2 × 2)-3CO|Pt(111). For these measurements, the welldefined (2 × 2)-3CO phase on the Pt(111) facet was assembled by polarizing the electrode at Eads = 0.1 V for a time, τads = 1 s. The potential was then stepped to a value Eox, sufficiently positive to promote full oxidation of the CO adlayer for a time, τox, long enough for the process to fully ensue, i.e., 0.1 s. Subsequently, the potential was stepped back to Eads and the entire sequence, shown as an inset in panel D, Figure 1, repeated numerous times until adequate signal-to-noise ratios were achieved. As mentioned in the previous section, the single crystal Pt bead was annealed before each individual run to minimize effects due to the formation of defect sites at the periphery of the (111) facets discussed in an earlier publication.33 Shown in Figure 1 are raw, averaged (ca. 600 acquisitions, Acq) Ipp(2ω) transients (panel A) and ΔR/R (Eref = Eads, and Esam (= Eox), panel B) for Eox = 0.98 (curve a), 0.99 (b), 1.0 (c), and 1.02 V vs RHE (d) . Careful inspection of these results revealed a number of interesting facts: 27903

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

= 0.1 V for τads ≥ 1 s and then stepped at t = 0 to Eox = 0.98 V, i.e., sufficiently positive for the adlayer to undergo oxidation. After a prescribed period of time, τh, the potential was stepped back to Eads for a time τst of 15−20 ms and then once again stepped to Eox and held there until the CO adlayer was fully oxidized. At this stage the potential was stepped to Eads, and the entire sequence repeated numerous times while monitoring ΔR/R. Shown in Figure 3 are averaged transient ΔR/R curves, where Eref in this case is 0.98 V, recorded following application of the potential protocol in thick black lines for two different values of τh, i.e., 105 (open circles) and 139 ms (solid circles). As evidenced by the data collected for τh = 105 ms, the overall intensity and shape was very similar to that observed for this same facet for τ < 139 ms, which is long enough for the layer to partially oxidize. The slight drop in ΔR/R immediately following the step from 0.1 to 0.98 V at t = 0 is related to the Stark shift as was discussed in a previous communication,26 whereas that at t = τh is consistent with the conversion of a small fraction of √19 × √19R23.4°-13CO produced at 0.98 V to the (2 × 2)-3CO phase (vide inf ra). A strikingly different response was observed for τh = 139 ms (see solid circles), a time long enough for the partial oxidation of the original fully assembled (2 × 2)-3CO layer. In this case, the step down to E = 0.1 V was accompanied by a sudden increase followed by a gradual decrease in ΔR/R attributed respectively to to the instantaneous adsorption of hydrogen and full desorption of bisulfate and the subsequent gradual adsorption of CO during τst. As the potential was then stepped up to E = 0.98 V, the optical signal suddenly decreased, consistent with the adsorption of bisulfate on the sites made available by the partial stripping of the CO adlayer. This explanation is also in accordance with the oxidation of CO molecules adsorbed during τst, which based on the results of the dosing experiments above would occupy random sites on the stripped surface and thus yield much faster rates of oxidation. Most importantly, however, following this sudden drop, the optical transient decayed in a more gradual fashion very similar

Figure 2. Plots of Ipp(2ω) (panel A) and ΔR/R (Eref = Eads, panel B) vs time recorded simultaneously for the oxidation of CO adlayers prepared by dosing (see text) following application of a potential step from Eads = 0.1 V at t = 0 to Eox = 0.98 V in CO-saturated 0.1 M H2SO4, τads = 1 (a), 0.6 (b), 0.4 (c), and 0.3 s (d). Each of the curves represents the average of ca. 500 Acq.

equivalently, smaller θCO, the decay occurs over times much shorter than those observed for the longer τdos (see curves a and b) or higher θCO. It can be surmised based on these data that the marked reduction in the time required for the full oxidation of the dosed layer is due to the presence of a large number of adsorbed CO species neighboring empty sites, which undergo more facile oxidation than those neighboring CO occupied sites. Oxidative Stripping. For these experiments, the (2 × 2)3CO phase was first assembled by holding the potential at Eads

Figure 3. Plots of ΔR/R (Eref = 0.98 V) vs time recorded in CO-saturated 0.1 M H2SO4 solutions, following application of a potential step at t = 0 from E = 0.1 V to E = 0.98 V followed by a potential pulse from Efin = 0.98 V down to E = 0.1 V, and back to E = 0.98 V applied in the region represented by the shaded areas as shown, i.e., τh = 105 ms, τst = 15 ms (open circles, light gray) and τh = 139, τst = 20 ms (solid blue circles, darker gray). Each of the curves represents the average of ca. 500 Acq. 27904

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

Figure 4. Plots of Ipp(2ω) vs time recorded for the forward (2 × 2) → √19 × √19R23.4° (panel A) and reverse √19 × √19R23.4°-13CO → (2 × 2)-3CO (panel B) as a function of the applied potential in CO saturated 0.1 M H2SO4, following application of a potential step. Panel A: Eini = 0.1 V (2 s), Efin = 0.93 (a), 0.92 (b), 0.91 (c), 0.90 (d), 0.89 (e), 0.88 (f), and 0.87 V (g) (5 s). Panel B: Eini = 0.92 (5 s), Efin = 0.1 (h), 0.3 (i) V, 0.38 (j), and 0.4 V (k) (2 s). Each of the curves represents the average of ca. 100 Acq. Photon counter: SR400, T = 50 ms/counts, 500 pts/record (A) and SR430, bin width = 163.84 μs, 1000 pts/record (B).

to that found for experiments in which τh was too short for the CO adlayer to undergo significant oxidation. It may be thus be surmised that the overall character of the CO adlayer is preserved following partial oxidation. Further support for this view was obtained by Weaver et al., who concluded based on their in situ infrared reflection absorption spectroscopy (IRAS) studies that extensive island formation survives to small θCO on Pt(111) formed by electrooxidative stripping of fully assembled saturated CO adlayers.23 It may be argued that the COads layer may have acquired high mobility during electrooxidation and reconstitute into fairly immobile structures once the potential was stepped back to 0.1 V. One possible scenario would involve a rapid coalescence of COads into (2 × 2)-3CO islands. Such conditions, however, would lead to an increase in peripheral (island edge) sites and thus in the rate of oxidation of the entire adsorbed CO layer. The virtual overlap between the transients observed for t > ca. 160 ms (see Figure 3) reinforces the view that partial oxidation of the (2 × 2)-3CO phase simply reduces θCO without changing the local molecular disposition (vide inf ra). Dynamics of the (2 × 2)-3CO ↔ √19 × √19R23.4°-13CO Phase Transition. The potentials required for CO oxidation are more positive than those associated with the (2 × 2)-3CO ↔ √19 × √19R23.4°-13CO phase transition,23 a factor that may complicate the analysis of the results. Shown in panel A of Figure 4 are plots of Ipp(2ω) as a function of time for a CO adlayer formed at Eads = 0.1 V, i.e., (2 × 2)-3CO phase, and then stepped to various potential values at which the √19 × √19R23.4°-13CO phase would be present. As indicated, the curves display an initial fast drop followed by a monotonic increase, as would be expected for the (2 × 2)-3CO to √19 × √19R23.4°-13CO transition. It is interesting to note that the magnitude of the dip is of the same order as that which should be expected based on an extrapolation of the linear Ipp(2ω) vs E behavior found earlier in slower scans in both perchloric41 and sulfuric acid.39 Although no significant differences could be discerned for Efin in the range 0.89−0.93 V, much slower rates could be obtained as Efin was decreased. Most importantly, the

phase transition, as was pointed out in an earlier communication,44 occurs over periods of time on the order of seconds. This is in marked contrast with the reverse transition (see panel B in this figure for a series of Efin), which for Efin in the range 0.1−0.38 V occurs over times 3 orders of magnitude shorter. These results are in agreement with those reported earlier using SHG for massive Pt(111) single crystal specimens in similar electrolytes.44 Similar qualitative conclusions were drawn from very recent in situ sum frequency generation experiments reported by Lagutchev et al.24 Theoretical Model. Attention will be focused in what follows on the quantitative analysis of experiments involving a well-defined initial state, i.e., a fully saturated (2 × 2)-3CO (θCO = 0.75), and an equally well-defined fully saturated final state, i.e., a √3 × √7 adsorbed bisulfate layer (θHSO4− = 0.2). The latter would be expected to form at potentials at which the CO adlayer undergoes full oxidation based on the adsorption isotherm for bisulfate on Pt(111) derived from classical interfacial thermodynamics measurements by Feliu, Lipkowski, and co-workers under equilibrium conditions.45 The primitive model to be herein considered assumes that at all times during the oxidation of the CO adlayer all areas of the Pt(111) facet are covered either by a (2 × 2)-3CO or by a √3 × √7 bisulfate (see Scheme 1). These conditions are equivalent to state that the rates of relaxation of the CO adlayer are negligible, and those of bisulfate adsorption (via diffusion from solution) and subsequent surface reorganization are much faster than those associated with COads oxidation. Whether these stringent requirements are met under some of the experimental conditions employed in this study will be determined from a quantitative analysis of the Ipp(2ω) and ΔR/R data in the left panels in Figure 1. Before proceeding further, it would be useful to review key aspects associated with translating optical signals into actual coverages both for CO and bisulfate. Optical Aspects. As recently reported, the analysis of normalized differential reflectance data collected in situ on 27905

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

2)-3CO patches (see data and axis in green in Figure 5). This assumption will be tested by determining θCO based strictly on the Ipp(2ω) data as shown next. It has been well established that adsorption of bisulfate suppresses generation of second harmonic from the otherwise bare Pt surface.47 Hence, all contributions to Ipp(2ω) can be regarded as originating exclusively from the presence of COads. From a general perspective, the nonlinear susceptibility of a bare metal surface originates by and large from the polarizability of free electrons associated with surface atoms. Hence, the binding of species to sites on the surface would be expected to decrease the nonlinear susceptibility and thus the magnitude of I(2ω). In fact, experiments performed in ultrahigh vacuum have shown that in the case of atomic oxygen adsorption, Oads, on Ag(110), I(2ω) is proportional to θO2, the square of the coverage of Oads.48 On this basis, it is reasonable to assume that for the system under consideration in this work θCO would be proportional to [Ipp(2ω)]1/2 provided the local geometric disposition of all adsorbate species remains the same for all coverages. The latter is equivalent to assume that the (2 × 2)3CO phase undergoes no relaxation during oxidation of the CO adlayer, as the primitive model proposed above demands. Since the initial and final values of θCO are known with certainty, it becomes possible to convert quantitatively transient Ipp(2ω) into corresponding θCO values. The internal consistency of this model can be tested from the two sets of optical data collected simultaneously. Specifically, unique values of θCO could be predicted from the ΔR/R data, which should then match those derived from the Ipp(2ω) data. In order to perform this comparison, it becomes necessary to correlate the coverages of the two species. Since all areas of the surface are covered at all times either by a (2 × 2)-3CO or by a √3 × √7 bisulfate layer,

Scheme 1

quasi-perfect single crystal Pt(111) facets over a range of bisulfate concentrations in aqueous perchloric acid solutions indicates that ΔR/R varies linearly with the coverage of bisulfate, θHSO4−, i.e., ΔR/R = β + γ θHSO4−·46 It thus seems reasonable to assume that such a linear relationship would also be obeyed for fully covered surfaces involving any two adsorbates, provided the nature of the adsorption sites does not change as a function of their coverage, such as the specific model herein considered, where at all times the (2 × 2)-3CO and √3 × √7 bisulfate patches cover the entire Pt(111) surface. Since θHSO4− varies from 0 at beginning of the experiment to 0.2 following full oxidation of the CO adlayer, the relationship above makes it possible to convert ΔR/R data into θHSO4− along the entire transient (see data and axis in blue in Figure 5). As prescribed by our proposed model, all other areas of the surface, i.e., 1 − θHSO4−, would be covered by (2 ×

Figure 5. Plots of averaged (10 points adjacent smoothing) θHSO4− vs t (inner right ordinate) based on the ΔR/R vs t data and the θCO predicted by the model (outer right ordinate, see text) for Eox = 0.98 V (panel A), Eox = 0.99 V (panel B), Eox = 1.00 V (panel C), and Eox = 1.02 V (panel D). Shown in magenta and black respectively are the values of θCO2 obtained by squaring θCO predicted by the model and those of the normalized Ipp(2ω) (see left ordinate). 27906

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

Figure 6. Best fits to the experimental 0.75 − θCO transient data (scattered points, right axes) to the integral of the Avrami function, i.e., 0.75 − θCO = ∫ t0σTπkNkG2t2 exp(−πkNkG2t3/3) dt (solid blue lines). The solid jagged curves are the same as those in black in Figure 5 in unnormalized form.

it follows that θCO = 3/4 − 15/4θHSO4−· Shown in magenta and black in Figure 5 respectively are values of θCO derived from the ΔR/R data based on this relationship and those based on the Ipp(2ω) results for Efin = 0.98 (panel A) and Efin = 0.99 (B), 1.00 (C), and 1.02 V (D) vs RHE. As evidenced from the results obtained, the rather good agreement between the two sets of values in panel A affords strong support for the proposed model. This is not the case, however, for the data in panels B, C, and D, where the values predicted from the ΔR/R data are higher than those expected based on the Ipp(2ω) data. The most likely explanation for this behavior may be found in the finite rates of bisulfate diffusion toward the surface, which, for short times, would leave areas of the surface not covered by either CO or bisulfate leading to an overestimate of θCO. Also to be considered is the possibility that the (2 × 2)-3CO undergoes a phase transition to √19 × √19R23.4°-13CO; however, the time constant for this process based on the data in Figure 4 is much longer than that associated with the oxidation of COads and therefore should not proceed to any large extent during the entire oxidation of the CO adlayer. Furthermore, the phase transition should lead to an increase in Ipp(2ω), which is not supported by the data. A quantitative analysis of the Ipp(2ω) transients was performed assuming the (2 × 2)-3CO adlayers retain their structural character during the time required for their full oxidation, as the data presented earlier in this work strongly

suggests. The transient Ipp(2ω) data in panel B of Figure 6 can be converted into θCO vs t using the expression θCO = 3 /4{Ipp(2ω)/I°pp(2ω)}1/2, where I°pp(2ω) is the intensity of the SHG at t = 0. The Avrami model as applied by Love and Lipkowski in their electrochemical studies28 provides values for the transient response of the current, namely, i = σTπkNkG2t2 exp(−πkNkG2t3/3), where σT is the charge contained under the full transient and kN and kG are the rates of nucleation and growth, respectively, and thus directly proportional to dθCO/dt (or, equivalently, d(0.75 − θCO)/dt), the rate at which nonCO-covered sites are being generated. Best fits to the experimental data were obtained using a feature available in Origin version 8.6, which allows the integral of an analytic expression as a user-defined fit function, i.e., 0.75 − θCO = ∫ t0σTπkNkG2t2 exp(−πkNkG2t3/3) dt in this case. It should be emphasized that this model only predicts the time evolution of the system from the time the reaction starts, and as such, it does not account for the induction time, τind. On this basis, it becomes then necessary to exclude from the fit, data collected before the onset of the actual reaction. To this end, the points preceding the onset were eliminated by fitting the data therein to a straight line and looking for changes in the slope from their close to zero values as the data set was extended into the onset region. Enough points were excluded to reach optimum values of R2, i.e., close to unity. The results obtained, shown in Figure 6 (see also best fit parameter values 27907

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

characteristic dimensions determined by the resolution of array) believed to be derived from interactions between the adsorbed species, i.e CO, water, bisulfate, hydroxyl, and/or surface oxides.

in the inset table in Figure 7), indicate the model (solid smooth lines) accurately fits the experimental data (scattered points)

■

CONCLUDING REMARKS The oxidation dynamics of the well-defined (2 × 2)-3CO (θCO = 0.75) and √19 × √19R23.4°-13CO (θCO = 0.68) CO adlayer phases on quasi-perfect Pt(111) facets were investigated in CO-saturated 0.1 M H2SO4 aqueous solutions via a combination of simultaneous, in situ, time-resolved, second harmonic generation (SHG, λinc = 590 nm), normalized differential reflectance spectroscopy (ΔR/R, λinc = 633 nm), and potential step techniques. The main conclusions emerging from this study can be summarized as follows: (i) The onset of oxidation of COads for both of these phases was found to be delayed with respect to the time at which the potential was stepped to values positive enough, Eox, to promote oxidation of the entire CO adlayers. This induction time, τind, was longer for the (2 × 2)-3CO (few tens of milliseconds) compared to the √19 × √19R23.4°-13CO phase (few milliseconds) and decreased monotonically for both of the phases as Eox was increased. (ii) For a fixed Eox, the oxidation rates, as measured by the optical techniques, were higher for the √19 × √19R23.4°13CO compared to the (2 × 2)-3CO phase. (iii) Excellent statistical fits to the θCO transients for the (2 × 2)-3CO phase derived from the SHG data could be obtained using Avrami’s nucleation and growth model and thereby consistent with the oxidation of the fully assembled (2 × 2)3CO phase starting at the edge of the domain and propagating through the entire island until the entire process is completed. (iv) Similar experiments involving CO adlayers on Pt(111) facets below saturation coverages (i.e., θCO < 0.75) formed by oxidative stripping of the fully assembled (2 × 2)-3CO phase yielded much longer τind values than surfaces with similar θCO prepared by dosing the otherwise bare Pt(111) surfaces.

Figure 7. Derivatives of the fits based on Avrami’s model shown in solid lines in Figure 6. The actual best fit parameters to the model are given in the table.

lending support to the assumptions upon which the model is based. The normalized changes in the CO coverage, θCO, predicted from this model, which within a constant should accurately reflect the transient currents due strictly to COads oxidation, can be obtained by taking the derivative of the fits. The results obtained, shown in Figure 7, yielded bell-shaped curves which strongly resemble the chronoamperometric data reported by other workers.28,29,33,34 The major advantage of this analysis is that no assumptions are required regarding the contributions to the total current arising from other sources. This is especially critical for COads oxidation where the capacitive currents due to the adsorption of bisulfate must be included in order to extract actual θCO values. The validity of this strategy when analyzing transient data, however, has not as yet been verified. Furthermore, it is important to note that previous analyses of transient data for the oxidation of COads on Pt(111)33 were based on experimental data collected in the absence of CO in the solution for which θCO is considerably smaller than that associated with the (2 × 2)-3CO adlayer,37 leaving a large number of non CO-covered sites where the oxidation of CO can be easily triggered. Extraordinary insights into the dynamics of electrooxidation of COads on Pt(poly) in CO-saturated aqueous sulfuric acid solutions have been gained from attenuated total reflection FTIR using a liquid N2-cooled MCT focal plane array detector consisting of 64 × 64 IR-sensitive elements.49 In particular, steady state data collected in experiments performed galvanostatically under forced convection afforded unambiguous evidence for the presence of spontaneously formed large domains (tenths to several millimeters in characteristics dimensions) of COads at high coverages surrounded by areas devoid of any detectable COads. Furthermore, areas in which the θCO was not as high, as judged by the intensity of the on-top band CO stretching band displayed over a wide range of coverages a CO stretching frequency virtually identical to that found for the large domains consistent with the COads in the form of small compact microislands (less than 40 μm in

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (D.S.). Author Contributions

D.S. planned and supervised the project. I.F. performed all the experimental measurements and together with H.Z. and Z.F. helped equally in the analysis of the data. Notes

The authors declare no competing financial interest.

■ ■

ACKNOWLEDGMENTS This work was supported by the National Science Foundation (NSF) CHE-1412060. REFERENCES

(1) Iwasita, T.; Nart, F. C. In Situ Infrared Spectroscopy at Electrochemical Interfaces. Prog. Surf. Sci. 1997, 55, 271−340. (2) Korzeniewski, C.; Climent, V.; Feliu, J. M. Electrochemistry at Platinum Single Crystal Electrodes. In Electroanalytical Chemistry. A Series of Advances; Bard, A. J., Zoski, C. G., Eds.; CRC Press: Boca Raton, FL, 2012; Vol. 24. (3) Markovic, N. M.; Grgur, B. N.; Lucas, C. A.; Ross, P. N. Electrooxidation of CO and H-2/CO Mixtures on Pt(111) in Acid Solutions. J. Phys. Chem. B 1999, 103, 487−495.

27908

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

(4) Grozovski, V.; Climent, V.; Herrero, E.; Feliu, J. M. The Role of the Surface Structure in the Oxidation Mechanism of Methanol. J. Electroanal. Chem. 2011, 662, 43−51. (5) Lai, S. C. S.; Lebedeva, N. P.; Housmans, T. H. M.; Koper, M. T. M. Mechanisms of Carbon Monoxide and Methanol Oxidation at Single-Crystal Electrodes. Top. Catal. 2007, 46, 320−333. (6) Kitamura, F.; Takahashi, M.; Ito, M. Carbon Monoxide Adsorption on Platinum (111) Single-Crystal Electrode Surface Studied by Infrared Reflection-Absorption Spectroscopy. Surf. Sci. 1989, 223, 493−508. (7) Chang, S. C.; Weaver, M. J. Coverage-Dependent Dipole Coupling for Carbon-Monoxide Adsorbed at Ordered Platinum(111)Aqueous Interfaces - Structural and Electrochemical Implications. J. Chem. Phys. 1990, 92, 4582−4594. (8) Chang, S. C.; Weaver, M. Coverage-Dependent and PotentialDependent Binding Geometries of Carbon-Monoxide at Ordered Low-Index Platinum Aqueous and Rhodium Aqueous Interfaces Comparisons with Adsorption in Corresponding Metal Vacuum Environments. Surf. Sci. 1990, 238, 142−162. (9) Akemann, W.; Friedrich, K. A.; Stimming, U. PotentialDependence of CO Adlayer Structures on Pt(111) Electrodes in Acid Solution: Evidence for a Site Selective Charge Transfer. J. Chem. Phys. 2000, 113, 6864−6874. (10) Lopez-Cudero, A.; Cuesta, A.; Gutierrez, C. Potential Dependence of the Saturation CO Coverage of Pt Electrodes: The Origin of the Pre-Peak in CO-Stripping Voltammograms. Part 1: Pt(111). J. Electroanal. Chem. 2005, 579, 1−12. (11) Lopez-Cudero, A.; Cuesta, A.; Gutierrez, C. Potential Dependence of the Saturation CO Coverage of Pt Electrodes: The Origin of the Pre-Peak in CO-Stripping Voltammograms. Part 2: Pt(100). J. Electroanal. Chem. 2006, 586, 204−216. (12) Oda, I.; Inukai, J.; Ito, M. Compression Structures of Carbon Monoxide on a Pt(111) Electrode Surface Studied by in Situ Scanning Tunneling Microscopy. Chem. Phys. Lett. 1993, 203, 99−103. (13) Villegas, I.; Weaver, M. J. Carbon-Monoxide Adlayer Structures on Platinum(111) Electrodes - A Synergy Between In-Situ ScanningTunneling-Microscopy and Infrared-Spectroscopy. J. Chem. Phys. 1994, 101, 1648−1660. (14) Song, M. B.; Yoshimi, K.; Ito, M. STM Observations of BridgeBonded CO on Pt(111) and Asymmetric On-Top CO on Pt(100). Chem. Phys. Lett. 1996, 263, 585−590. (15) Wakisaka, M.; Asizawa, S.; Yoneyama, T.; Uchida, H.; Watanabe, M. In Situ STM Observation of the CO Adlayer on a Pt(110) Electrode in 0.1 M HClO4 Solution. Langmuir 2010, 26, 9191−9194. (16) Markovic, N. M.; Grgur, B. N.; Lucas, C. A.; Ross, P. N. Surface Electrochemistry of CO on Pt(110)-(1 × 2) and Pt(110)-(1 × 1) Surfaces. Surf. Sci. 1997, 384, L805−L814. (17) Wang, J. X.; Robinson, I. K.; Ocko, B. M.; Adzic, R. R. Adsorbate-Geometry Specific Subsurface Relaxation in the CO/ Pt(111) System. J. Phys. Chem. B 2005, 109, 24−26. (18) Lucas, C. A.; Markovic, N. M.; Ross, P. N. Structural Effects during CO Adsorption on Pt-Bimetallic Surfaces. II. The Pt(111) Electrode. Surf. Sci. 2000, 448, 77−86. (19) Markovic, N. M.; Lucas, C. A.; Rodes, A.; Stamenkovic, V.; Ross, P. N. Surface Electrochemistry of CO on Pt(111): Anion Effects. Surf. Sci. 2002, 499, L149−L158. (20) Tolmachev, Y. V.; Menzel, A.; Tkachuk, A. V.; Chu, Y. S.; You, H. D. In Situ Surface X-Ray Scattering Observation of Long-Range Ordered (√19 x√19) R23.4°-13CO Structure on Pt(111) in Aqueous Electrolytes. Electrochem. Solid-State Lett. 2004, 7, E23−E26. (21) Kitamura, F.; Takahashi, M.; Ito, M. Adsorption Site Interconversion Induced by Electrode Potential of Carbon Monoxide on the Platinum(100) Single-Crystal Electrode. J. Phys. Chem. B 1988, 92, 3320−3323. (22) Leung, L. W. H.; Wieckowski, A.; Weaver, M. J. In Situ Infrared Spectroscopy of Well-Defined Single-Crystal Electrodes: Adsorption and Electrooxidation of Carbon Monoxide on Platinum (111). J. Phys. Chem. 1988, 92, 6985−6990.

(23) Chang, S. C.; Weaver, M. J. In Situ Infrared Spectroscopy of Carbon Monoxide Adsorbed at Ordered Platinum(100)-Aqueous Interfaces: Double-Layer Effects upon the Adsorbate Binding Geometry. J. Phys. Chem. 1990, 94, 5095−5102. (24) Lagutchev, A.; Lu, G. Q.; Takeshita, T.; Dlott, D. D.; Wieckowski, A. Vibrational Sum Frequency Generation Studies of the (2 × 2)→(√19×√19) Phase Transition of CO on Pt(111) Electrodes. J. Chem. Phys. 2006, 125, 154705. (25) Menzel, A.; Chang, K. C.; Komanicky, V.; Tolmachev, Y. V.; Tkachuk, A. V.; Chu, Y. S.; You, H. High-Density Electrosorbed Carbon Monoxide Monolayers on Pt(111) under Atmospheric Pressure. Phys. Rev. B 2007, 75, 035426. (26) Shi, P. F. I.; Scherson, D. A. In Situ, Time-Resolved Reflectance Spectroscopy in the Microsecond Domain: Oxidation of Adsorbed Carbon Monoxide on Polycrystalline Pt Microelectrodes in Aqueous Solutions. Langmuir 2006, 22, 10389−10398. (27) Couto, A.; Rincon, A.; Perez, M. C.; Gutierrez, C. Adsorption and Electrooxidation of Carbon Monoxide on Polycrystalline Platinum at pH 0.3−13. Electrochim. Acta 2001, 46, 1285−1296. (28) Love, B.; Lipkowski, J. Effect of Surface Crystallography on Electrocatalytic Oxidation of Carbon Monoxide on Platinum Electrodes. ACS Symp. Ser. 1988, 378, 484−496. (29) Bergelin, M.; Feliu, J. M.; Wasberg, M. Study of Carbon Monoxide Adsorption and Oxidation on Pt(111) by Using an Electrochemical Impinging Jet Cell. Electrochim. Acta 1998, 44, 1069−1075. (30) Pozniak, B.; Mo, Y.; Scherson, D. A. The Electrochemical Oxidation of Carbon Monoxide Adsorbed on Pt(111) in Aqueous Electrolytes As Monitored by in Situ Potential Step-Second Harmonic Generation. Faraday Discuss. 2002, 121, 313−322. (31) Akemann, W.; Friedrich, K. A.; Linke, U.; Stimming, U. The Catalytic Oxidation of Carbon Monoxide at the Platinum/Electrolyte Interface Investigated by Optical Second Harmonic Generation (SHG): Comparison of Pt(111) and Pt(997) Electrode Surfaces. Surf. Sci. 1998, 402, 571−575. (32) Herrero, E.; Alvarez, B.; Feliu, J. M.; Blais, S.; Radovic-Hrapovic, Z.; Jerkiewicz, G. Temperature Dependence of the Coads Oxidation Process on Pt(111), Pt(100), and Pt(1 1 0) Electrodes. J. Electroanal. Chem. 2004, 567, 139−149. (33) Lebedeva, N. P.; Koper, M. T. M.; Feliu, J. M.; van Santen, R. A. Role of Crystalline Defects in Electrocatalysis: Mechanism and Kinetics of CO Adlayer Oxidation on Stepped Platinum Electrodes. J. Phys. Chem. B 2002, 106, 12938−12947. (34) Lebedeva, N. P.; Koper, M. T. M.; Feliu, J. M.; van Santen, R. A. Mechanism and Kinetics of the Electrochemical CO Adlayer Oxidation on Pt(111). J. Electroanal. Chem. 2002, 524, 242−251. (35) Lebedeva, N. P.; Koper, M. T. M.; Feliu, J. M.; van Santen, R. A. The Effect of the Cooling Atmosphere in the Preparation of FlameAnnealed Pt(111) Electrodes on CO Adlayer Oxidation. Electrochem. Commun. 2000, 2, 487−490. (36) Petukhov, A. V.; Akemann, W.; Friedrich, K. A.; Stimming, U. Kinetics of Electrooxidation of a CO Monolayer at the Platinum/ Electrolyte Interface. Surf. Sci. 1998, 402, 182−186. (37) Cuesta, A. Measurement of the Surface Charge Density of COSaturated Pt(111) Electrodes as a Function of Potential: The Potential of Zero Charge of Pt(111). Surf. Sci. 2004, 572, 11−22. (38) Pozniak, B.; Scherson, D. A. In Situ Dual-Beam Coincidence Second Harmonic Generation as a Probe of Spatially Resolved Dynamics at Electrochemical Interfaces. J. Am. Chem. Soc. 2004, 126, 14696−14697. (39) Fromondi, I.; Scherson, D. A. Mechanical Properties of Bimetallic Crystalline and Nanostructured Nanowires. Faraday Discuss. 2008, 140, 59−68. (40) Clavilier, J. The Role of Anion on the Electrochemical Behaviour of a {111} Platinum Surface; An Unusual Splitting of the Voltammogram in the Hydrogen Region. J. Electroanal. Chem. 1980, 107, 211−216. (41) Fromondi, I.; Scherson, D. A. Oxidation of Adsorbed CO on Pt(111) in CO-Saturated Perchloric Acid Aqueous Solutions: 27909

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910

The Journal of Physical Chemistry C

Article

Simultaneous in Situ Time-Resolved Reflectance Spectroscopy and Second Harmonic Generation Studies. J. Phys. Chem. B 2006, 110, 20749−20751. (42) Fromondi, I.; Zhu, H. F.; Scherson, D. A. In Situ Spectroscopy at the Quasi-Perfect Pt(111) Single-Crystal Facet|Aqueous Electrolyte Interface. J. Phys. Chem. C 2012, 116, 19613−19624. (43) Budevski, E. B.; Staikov, G. T.; Lorenz, W. J. Electrochemical Phase Formation and Growth; VCH: Weinheim, 1996. (44) Pozniak, B.; Scherson, D. A. Dynamics of a Surface Phase Transition As Monitored by in Situ Second Harmonic Generation. J. Am. Chem. Soc. 2003, 125, 7488−7489. (45) Herrero, E.; Mostany, J.; Feliu, J. M.; Lipkowski, J. Thermodynamic Studies of Anion Adsorption at the Pt(111) Electrode Surface in Sulfuric Acid Solutions. J. Electroanal. Chem. 2002, 534, 79−89. (46) Fromondi, I.; Scherson, D. (Bi)sulfate Adsorption on Quasiperfect Pt(111) Facets from Acidic Aqueous Electrolytes As Monitored by Optical Techniques. J. Phys. Chem. C 2007, 111, 10154−10157. (47) Lynch, M. L.; Barner, B. J.; Lantz, J. M.; Corn, R. M. In Situ 2nd Harmonic-Generation Studies of Bisulfate Adsorption at Well-Ordered Pt(111) and Pt(100) Electrodes. J. Chim. Phys. 1991, 88, 1271−1289. (48) Heskett, D.; Urbach, L. E.; Song, K. J.; Plummer, E. W.; Dai, H. L. Oxygen and Pyridine on Ag(110) Studied by 2nd HarmonicGeneration - Coexistence of 2 Phases within Monolayer Pyridine Coverage. Surf. Sci. 1988, 197, 225−238. (49) Bauer, P. R.; Bonnefont, A.; Krischer, K. Spatially Resolved ATR-FTIRS Study of the Formation of Macroscopic Domains and Microislands during CO Electrooxidation on Pt. ChemPhysChem 2010, 11, 3002−3010.

27910

dx.doi.org/10.1021/jp5089993 | J. Phys. Chem. C 2014, 118, 27901−27910