Kinetics and mechanism of the oxygen electroreduction reaction on faceted platinum electrodes in trifluoromethanesulfonic acid solutions

The kinetics of the oxygen electroreduction reaction (OERR) were investigated on (1 1 1)-type and (1 0 0)-type faceted, and polycrystalline platinum electrodes in aqueous (0.05–1.0)m trifluoromethanesulfonic acid (TFMSA) using the rotating disc and ring-disc electrode techniques at 25°C. Reaction orders with respect to oxygen close to either 1/2 or 1 were found, depending on the TFMSA concentration and platinum surface morphology. At all TFMSA concentrations the formation of H2O2 was enhanced at (1 0 0)-type platinum surfaces. The difference in the electrocatalytic activity of platinum surfaces can be explained through data derived from the OERR formalism proposed by Damjanovicet al. The rate of the direct O2 to H2O electroreduction reaction increased steadily with the cathodic overvoltage irrespective of the platinum surface morphology, whereas a maximum H2O2 formation rate was found at about 0.5 V, depending on the TFMSA concentration. The H2O2 decomposition rate on (1 0 0)-type platinum electrode yielding H2O approached zero within a certain potential range.


Introduction
The performance of Pt/phosphoric acid fuel cells is principally limited by the high overvoltage of the oxygen electroreduction reaction (OERR). The rate of this reaction can be increased by using alternative electrolytes, particularly aqueous fluoroalkane sulfonic acids. These acids offer good thermal and electrochemical stability above 80 °C [1][2][3]. Furthermore, at a constant temperature, the OERR rate on platinum in aqueous trifluoromethanesulfonic acid (TFMSA) is almost two orders of magnitude greater than that in concentrated H3PO 4 and similar to that in aqueous HzSO 4 [1]. This behaviour has been attributed to the greater solubility of oxygen in aqueous TFMSA and to the poor adsorption of CF3SO2 O-anion on platinum [2,3]. The kinetics of the OERR on platinum in aqueous TFMSA above 80 °C has shown a better mass transfer coefficient for oxygen and a greater exchange current density as compared to other acid electrolytes, such as aqueous HC104 and H3PO 4 [4].
Two Tafel regions have been confirmed for the OERR on polycrystalline platinum (pc-Pt) in aqueous TFMSA, namely, -0.060V (decade) -1 in the low cathodic overvoltage range (1.o.r.) and -0.120 V (decade) -1 in the high cathodic overvoltage range (h.o.r.), at 25 °C [1,5]. Despite the validity of these results for the OERR on pc-Pt in aqueous TFMSA [6,7], the mechanism of the reaction in these solutions is not yet fully understood as, for instance, the influence of the electrode surface morphology on the OERR kinetics at noble metal electrocatalysts is not firmly established. Research has recently been 740 carried out for several electrochemical reactions on well-defined crystallographic electrode surfaces with different orientations [8][9][10]. However, the scarce knowledge about the stability of low Miller index single crystal platinum surfaces in the aqueous environments and potential range related to the OERR, has led to inconclusive results.
A possible way to overcome this drawback is by using electrofaceted platinum surfaces with different preferential orientations. These surfaces are easily prepared in a reproducible way and are sufficiently stable over a wide range of potential for a long term [11][12][13][14]. The surface characteristics of electrofaceted platinum have been determined by cyclic voltammetry, scanning electron microscopy (SEM) and scanning tunneling microscopy (STM) [14][15][16]. Accordingly, (1 00)-preferentially oriented platinum (hereafter denoted as (1 00)-type Pt), and (1 1 1)-preferentially oriented platinum (hereafter denoted as (1 1 1)-type Pt), can be described as high order Miller index stepped surfaces with a predominance of either (100) or (1 1 1) oriented facets, respectively.
It has been found recently that the OERR on both preferentially oriented and well-defined single crystal platinum electrodes in aqueous HC104, H3PO4, and H2SO 4 [10,14] and alkaline solutions [17], behaves as a structure-sensitive reaction, probably induced, at least partially, by anion adsorption and/or a high surface coverage by OERR intermediates [14,17]. The aim of this work is to gain further insight into the induced-anion structure effect on the OERR kinetics on pc and electrofaceted platinum electrodes in aqueous TFMSA. 0021

Experimental details
Mirror polished pc-platinum disc working electrodes were prepared using successive alumina pastes of different grades, subsequently immersed in a 1 : 1 pure nitric-sulfuric acid mixture for 1 h and, finally, rinsed with Millipore-MilliQ* water.
Faceted platinum disc electrodes were made by applying the repetitive square wave potential (RSWP) technique in aqueous 0.5M H2SO4, as described elsewhere [13]. The real surface area of the working electrode (AD) was estimated from the Hadatom voltammetric charge density obtained at 0.1Vs -1, after the double layer charge correction [18]. A platinized.Pt plate counter electrode (10 cm 2 in geometric area) facing the disc electrode, and a reversible hydrogen reference electrode (RHE) in the same solution, were used. Potentials in the text are given on the RHE scale. All experiments were performed at 25 °C.
TFMSA (Merck, 98% 'for Synthesis') was purified by double distillation under vacuum at temperatures below 40 °C [19]. Aqueous TFMSA was prepared by dilution with Millipore-MilliQ* water under a purified nitrogen atmosphere. To eliminate electroactive impurities, oxygen-free aqueous TFMSA was preelectrolysed for 48 h using a large platinized-Pt anode. The anode was then removed from the solution, and the platinum working electrode was positioned in the cell. Control cyclic voltammograms were run at sweep rate v = 0.1 V s-I in the 0.05-1.0 V range. Negligible changes were found in those voltammograms run after holding the potential at 0.60V for 10rain ( Fig. 1 Previous to the rotating disc electrode (RDE) and ring-disc electrode (RRDE) measurements, both the disc and ring electrodes were activated by applying a conventional potential routine as described elsewhere [14]. The disc potential (ED) was scanned from 1.0 to 0.05V at 0.01Vs -1, and the ring potential (ER) was held at 1.2 V to electrooxidize H202 to O2 under the limiting current regime. Values of ID and IR, the disc and ring currents, were recorded at different rotation speeds (co) in the 500,.<co~< 3000rpm range. The RRDE collection efficiency was N= 0.22. Potentiostatic polarization curves were obtained in the 0.05V-Erest range, where Erest is the open circuit potential. Steady currents were attained after 4min or thereabouts, irrespective of the potential change direction. Polarization curves run at w = 2000 rpm, after correction for mass transport contribution [14], were plotted as E against log j (Tafel plots).

Cyclic voltammograms for Pt working electrodes in aqueous TFMSA
Cyclic voltammograms for pc and faceted platinum electrodes in aqueous 0.05 M TFMSA at 0.1 V s-1 in the 0.05-1.50 V range (Fig. 2), are qualitatively similar to those found in aqueous H2SO 4 solutions, as they exhibit typical H-atom etectrosorption from 0.05 to 0.45 V, O-atom electrosorption and oxide formarion from ca. 0.7 to 1.5V, and the charging-discharging double layer region in the intermediate potential range. The change in the position and shape of the H and O-atom conjugated electrosorption peaks reflects the influence of platinum surface morphology. Furthermore, as referred to in the literature [3,20], the influence of CF3SO20-anions resembles that ofF-and BF4 anions [21]. In fact, a low and constant surface coverage by CF3SO20-anions has been reported in the 0.55-0.90V range, i.e. I'CF3SO2 o-= 4#Ccm -2 in aqueous 5 x 10-3M TFMSA [20], where F denotes the adsorbed anion excess equivalent concentration.
The H-atom electrosorption voltammogram shows a multiplicity of peaks, which are associated with the different surface morphology dependent H-Pt bonding energies [8]. For (1 0 0)-type Pt in aqueous 0.05 M TFMSA, the contribution of strongly adsorbed Hatoms is enhanced, and the corresponding pair of peaks lies at 0.33 V, a potential which is positively shifted with respect to that found for H-adatoms in aqueous 0.05M H2SO 4. Otherwise, the onset of O-atom-electroadsorption on (100)-type Pt in aqueous 0.05 M TFMSA lies at ca. 0.7 V, this potential shifts positively as the TFMSA concentration is increased.
On the other hand, cyclic voltammograms resulting I I I I Voltammograms resulting from all platinum electrodes show that the H-Pt bond energy in aqueous 0.05M TFMSA is greater as the corresponding H-adatom desorption is completed at potentials more positive than those in aqueous 0.05 M H2SO4. It also appears that in aqueous 0.05 M TFMSA, the O-Pt bond is stronger and the oxygen evolution reaction commences at potentials lower than those in aqueous 0.05 M H2SO 4. The influence of CF3SO20-anions on the H and O-atom electrodesorption is comparable to that reported for aqueous HBF 4 [20], and has been explained by the low adsorbability of BF4 anion on platinum electrodes.

Reaction orders with respect to oxygen for Pt electrodes in aqueous TFMSA
The reaction order (p) with respect to oxygen at different electrode potentials for the OERR on platinum [17] was estimated from the log Io against log(1 -ID/ILo) plots at constant w ( Fig. 3), where ILo denotes the limiting current at the disc. For where c~32 and Do2 stand for the concentration and diffusion coefficient of 02 in the bulk acid, respectively, and u is the kinematic viscosity of the acid solution. Values of C°o2, Do2 and u for aqueous 0.1 and 1.0M TFMSA were taken from the literature [4,6,21]. The OERR polarization curves were also obtained at different oxygen partial pressures (0.25atm ~< po2~<l.0atm) in the 0.05--1.0M aqueous TFMSA range. The corresponding Tafel plots for the three types of platinum surface (Fig. 4) were w-independent and the values ofpo2 resulting from these plots agreed with the above reported data. reaction scheme proposed for the OERR [22]. ID against E D and I R against ED plots were obtained within the 500 rpm ,.<co-~< 3000 rpm range in oxygensaturated aqueous TFMSA (Fig. 5) by scanning E D from 1.2 to 0.1V at 0.01Vs -1. An co-independent nonzero ID value appeared at Eo = 0.93V, but mass transport contributions became significant at lower potentials leading to a limiting current for observed. Likewise, as expected for an electrochemical reaction under mixed control, I R also increased with w and attained a maximum value at 0.5V, irrespective of platin~um morphology. IR against ED plots for (1 00)-type Pt comprise higher OERR currents than those for (1 1 1)-type and pc-Pt. These results, which confirm earlier data for aqueous H2SO 4 and KOH [13,15], were attributed to a better ability of (1 0 0)-type Pt to electrogenerate H202. Similar features of the ID against ED and IR against ED plots were found at the highest TFMSA concentration.
The preceding results confirm earlier data for pc-Pt [23][24][25][26][27][28], and further indicate that the overall OERR in the h.o.r, is seemingly controlled by the first electron transfer, the adsorbed intermediates probably obeying a Langmuir isotherm. It should be noted that either a reaction pathway implying a Temkin adsorp-tion for intermediates, or a chemical step following the first electron transfer as r.d.s, have been proposed for the OERR in the 1.o.r. [23].
As previously found for other aqueous electrolytes [14,17], values of bla.o.r, for (1 0 0)-type Pt are greater than those for (1 1 1)-type and pc-Pt. Thus, in aqueous H2SO 4 and KOH, high Tafel slopes have been explained through a blockage of platinum electroactive sites by H20 z intermediates formed during the OERR [14,17]. This explanation also applies to the OERR in aqueous TFMSA and is consistent with the greater contribution of IR for an increasing TFMSA concentration (Fig. 5).
The extrapolation of Tafel lines to the oxygen electrode reversible potential yields Jo, the OERR exchange current density of each Tafel region. Kinetic parameters derived from Tafel plots ( Table  1) indicate an enhancement of the OERR rate on platinum as the TFMSA concentration is increased, although no linear relationship between Jo and TFMSA concentration could be established because of the change in oxygen solubility and diffusivity. Accordingly, Jo values in the 1.o.r. and h.o.r, corrected for the oxygen solubility [4,6] are included in Table 1.

3.3.3.
Analysis of RRDE data. The analysis of RRDE data was based upon the OERR formalism proposed by Damjanovic et al. [29], which consists of a global series versus a parallel reaction pathway. It comprises a direct electroreduction from O2 to H20 (step 1), a parallel electroreduction from O2 to H202 (step 2) and further H202 electroreduction to H20 where superscripts ° and * denote species in the bulk solution, and in the solution adjacent to the electrode surface, respectively, and kD and k~ are the mass transfer coefficients of O2 and H202, respectively. In Scheme I, the catalytic decomposition of peroxo-adsorbates and the adsorption/desorption equi-librium of H202 have been neglected. Scheme I predicts the following relationships between ID, IR and ILO and ca [30]: ro ~R = /1 + S1 co-l/2 Values of ka, k 2 and/~3 were determined from the values of 11, $1 and $2 at different Eo. Finally, from Scheme I, the following expressions for the k's result, RRDE data for the OERR on different types of platinum in aqueous 0.05 M TFMSA (Fig. 6) lead to values of S1 which depend considerably on ED for (1 00)-type Pt, in contrast to those for (1 1 1)-type

I I I
...... from the direct 02 to H20 electroreduction pathway has practically disappeared.

Ill
In summary, ID/IR against w -1/2 plots show two distinguishable modes of OERR behaviour in aqueous TFMSA on platinum, i.e., a large and EDdependent value of 11 for (1 1 1)-type and pc-Pt, and I1 ~-10 for (1 0 0)-type Pt, a figure which becomes potential independent for ED~<0.8V. These results resemble those previously found for the OERR in alkaline solutions [17].
The IL,D/(IL,D--ID) against co -1/2 plots in aqueous 0.05M TFMSA (Fig. 7) show that the value of $2 in the 0.45-0.8 V range for (1 0 0)-type Pt are smaller than those obtained for (1 1 1  Values of kl, k2 and /~3 can be determined from Equations 4 to 6, at different ED (Fig. 8), Values of /~1 correspond to the largest rate constant at all platinum surfaces, and diminish steadily with the cathodic overvoltage. However, for ED < 0.4 V,/~1 is nearly constant (0.4cms-]). This means that the OERR in aqueous TFMSA proceeds principally through the direct electroreduction from O2 to H20 at all platinum surfaces at a limiting rate (Fig. 5). On the other hand, a peaked value of/~2 is attained at ca. 0.5 V, and for (1 0 0)-type Pt values of/~2, which are twice or three times greater than those obtained on other platinum surfaces, are found. This indicates a greater contribution of the H202 electroformation reaction for (1 00)-type Pt. The value of /~3 also decreases steadily with the cathodic overvoltage, but /~3 = 0 in the 0.80-0.65V range for (1 00)-type Pt, indicating the presence of H202 in the bulk solution. This has been previously reported for the OERR in aqueous KOH at the same platinum faceted surface [17]. For higher TFMSA concentrations, the value of k3 increases at all platinum surfaces (Fig. 8), however /c3 = 0 for (100)-type Pt at higher ED values, i.e., at ED ~<0.75V in 1.0M TFMSA.

Kinetic andmechanistic aspects of the OERR on Pt in aqueous acid solution
The value bLo.r =-2.303 (RT/F) for the OERR on (1 1 1)-type and pc-Pt in oxygen-saturated (Po2 = 1.0atm) aqueous 1.0M HzSO 4 [14] was associated with either a chemical rate determining step (r.d.s.) following the primary electron transfer or the same first electron transfer itself as r.d.s, under Temkin adsorption conditions for intermediates [26]. On the other hand, the value bla.o.r ' = -2.303 (2RT/F) was interpreted through a mechanism involving the first single electron transfer as r.d.s, under Langmuir adsorption conditions for intermediates [23], although it was found that the presence of strongly adsorbable anions, such as SO~-and HSO~, changes the OERR kinetics on platinum as Ibh.o.r.I > --2.303 (2RT/F). This fact was explained by a competitive adsorption between anions and O-containing species produced in the course of the OERR for platinum electroactive sites.
On the other hand, two Tafel slopes for the OERR on (100)-type Pt in aqueous 1.0M H2SO 4 were found, namely bl.o.r. =-0.06V(decade) -1 and bh.o.r. =-0.165V(decade) -1. In this case, weaker anion adsorption effects were explained as a mismatch between the SO42-and HSO4 anion structtires and the predominant surface atomic arrangement in (1 0 0)-type Pt. Otherwise, at potentials greater than 0.9 V, the anion desorption competes with the initiation of the oxide layer on platinum. Under these conditions a common bl.o.r, value for the OERR is involved, irrespective of the platinum surface nature.
A strong O2-Pt interaction in aqueous solutions is required for the OERR to take place at a reasonable rate. Then, the oxygen-adsorbate stabilization results from a compromise between an increase in the Pt-O and a decrease in the Pt-Pt bond strength. Accordingly, a two-fold Pt-O configuration ('bridge', peroxo-like adsorbate) appears to be harder to break than a one-fold Pt-O configuration ('on-top'). The filling up with electrons of 7r* levels in the oxygenadsorbate increases the strength of the O-O bond leading to a lower interatomic distance [34]. On the other hand, the stretching of the O-O bond leads to a decrease in the ~r* level energy and to a better coupling between the O-O bond and platinum bands. Therefore, not only the O271"* and the or* levels are involved in oxygen-adsorption and dissociation, but also antibonding orbitals on the Pt (1 1 1) surface. This matter has been recently studied for oxygen-adsorption on Pt(1 1 1) and Pt(1 0 0) clusters by applying an 'extended Hfickel molecular orbital' (EHMO) method corrected for repulsion energies according to Calzaferri's [34,39,40]. Accordingly, oxygen molecular adsorption on Pt(1 0 0) and oxygen dissociative adsorption on Pt(1 1 1) yielding O and OH-species are favoured. Once the oxygen-adsorbate is formed, the subsequent step involves at least two competitive charge transfer processes, namely, Oadsorbates yielding H20 on Pt(1 1 1) and oxygenadsorbates leading to H202 and H20 on Pt(1 0 0).
The H20 2 electroformation during the OERR on platinum in aqueous TFMSA is favoured on (1 0 0)type Pt like in aqueous H2SO 4 and KOH [14,17]. This is consistent with the values of /~2 and k3 analysed in Section 3.3.3., after applying the reaction formalism proposed by Damjanovic et al. [29].
Let us first consider the value of p, the reaction order with respect to oxygen, in the absence of anion adsorption interference. This situation is approached to some extent in dilute TFMSA solutions. Then, results on (1 1 1)-type and pc-Pt surfaces suggest that a dissociative oxygen-adsorption mechanism operates, as p = 0.5. Accordingly, the following reaction pathway predominates: 2Pt + OzW 2[PtO]ad (7) k-a where Po2 is the 02 partial pressure. Therefore, the steady rate equation, expressed as current density, Jc, is where K involves conventional potential independent parameters. According to Equation 11, the value p_~ 1/2 should be expected, as found for (111)-type and pc-Pt. This conclusion is no longer valid when the TFMSA concentration exceeds 0.5M, as a change from p ___ 1/2 to nearly 1 is then observed.
On the other hand, the value ofp for the OERR on (100)-type Pt in aqueous TFMSA is close to 1, a figure which suggests a predominant molecular oxygen adsorption for this type of platinum surface. This fact can be explained by the following reaction pathway: This interpretation is consistent with the larger amounts of H202 found from RRDE data for this type of platinum. It is likely that CF3SO20-anion adsorption also becomes negligible on (1 0 0)-type Pt and this makes molecular oxygen adsorption more favourable.

Possible interpretation of the influence of CF3S020-anions on the kinetics and mechanism of the OERR
The rate of the OERR in aqueous TFMSA (Table 1) is similar to that in aqueous H2SO4 and two orders of magnitude greater than that in concentrated aqueous H3PO4. An explanation for the influence of the solution composition on the kinetics of the OERR on platinum may be advanced.
TFMSA is totally ionized in H30 + and CF3SOzOand the adsorption characteristics of this anion on platinum has been studied less than that of HSO~and SO 2-anions [14]. For these anions, depending on the Pt-O interactions, three types of adsorption configurations on platinum are expected [41]. Thus, the adsorption of HSO~-on Pt(1 1 1) has been related to the predominant infrared absorption band appearing in the 1200-1300 crn -1 spectral range, which was assigned to the asymmetric SO3 stretching of trigonally coordinated HSO~-species. A potential independent absorption band at about 1100 cm -~, which was assigned to the SO ]-vibration mode in the solution also appears for Pt(1 1 1) [42].
Otherwise, the main FTIR absorption bands of CF3SO20-on platinum [43], appear only when E > 0.6V. Therefore, this potential value can be considered as a threshold adsorption potential of CF3SO20-anions. The broad band which covers the 1250 to 940cm -1 range for E = 0.8V involves a peak multiplicity due to the S-O bond vibrational resonance. The peak located at 1155cm -1 is related to the S-O asymmetric stretching mode, and peaks at 1080, 1045 and 1005cm -1 correspond to the S-O symmetric stretching frequencies. Absorption bands covering from 1380 to 1290cm -1 are assigned to C-F bond vibrations.
The tetrahedral arrangement of the F3C-group in the CF3SO20-anion would tend to favour the anion adsorption on Pt(1 1 1) sites. However, F atoms in the anion should produce a rather strong repulsive interaction with platinum atoms, leading to weak anion adsorption on those sites. Otherwise, for Pt(1 00) sites [14], steric effects hinder CF3SOzO-anion adsorption through the F3C-group. These structural features account for the fact that the degree of surface coverage by CF3SO20-anions on pc-platinum in aqueous 0.02M TFMSA is lower than 0.1 [44].
The above description of the electrochemical interface can be extended by considering the possibility of OH-anion and H20 coadsorption since the Pt-Pt distance at Pt(1 1 1) planes (dpt-Pt = 0.277nm) is sufficiently large and makes possible the simultaneous location of both CF3SO20-and OH-anions, either as bare, or partially hydrated, species. This possibility can be sustained by the fact that for the same TFMSA solution, the onset of O-atom electrosorption on (1 1 1)-type Pt (Fig. 2) is positively shifted with respect to that on (1 0 0)-type Pt. This potential shift may be due to the partial shielding of Pt(1 1 1) sites by tetrahedral CF3-groups.
On the other hand, TFMSA electrodecomposes on platinum at E > 1.2V yielding (CH3)2SO 4 and SO 2anions [43], a fact which offers a possible explanation for the change in the value of p for the OERR on (1 1 1)-type Pt when the TFMSA concentration exceeds 0.5M. In this case, the CF38020-anion decomposition leads to SO 2-anions which can be specifically adsorbed on platinum [45]. This situation arises from a competitive adsorption of SO 2-anions and 02 for Pt(1 1 1) sites, which may assist the change of the reaction mechanism expressed by Reactions 7 to 9 to that expressed by Reactions 12 to 14. This change implies that the oxygen-adsorbate configuration on (1 1 1)-type Pt, also changes from the 'bridge side-on' peroxo-type structure to the linear 'end-on' structure.
The adsorption of oxygen on (100)-type Pt probably comprises 'bridge side-on' adsorbed species due to the lower O27c* Mulliken population [31] as compared to that on (1 1 1)-type Pt. Furthermore, as the CF3SO20-anion adsorption at Pt(1 0 0) sites is always low the bridge side-on adsorbate configuration would remain, irrespective of the TFMSA concentration. Then, at increasing cathodic overvoltage, two possible reactions are either the direct H202 desorption from the 'bridge side-on' peroxo-adsorbate or the H20 desorption after the O-O bond breaking step. It was experimentally found that both possibilities occur although their relative contribution depends mainly on the cathodic overvoltage and TFMSA concentration, as already discussed. As the surface coverage by CF3SO20-anions on platinum through CF3-groups should be negligible, only a rather mobile adsorption can be assigned to these anions. Accordingly, the oxygen adsorption would be favoured because platinum sites exclusively interact with H20 and its decomposition products. However, the steric arrangement of F3C-groups may cause partial shielding and fluctuation of the electric field at the interface, which should be greater for (111)-type Pt than for (100)-type Pt. Presumably, the orientation of the CF3SO2 O-anion on platinum is determined by CF3-groups facing the platinum surface displaying the sulfonic acid moiety towards the solution [41]. This type of interaction involves the greatest overlapping between the highest occupied molecular orbital (HOMO) of the anion and the available empty Pt state. The HOMO orbitals are responsible for the forward cr bonding with empty 5d Pt sites.
Calculations for the most likely adsorbed configuration of the CF3SO20-anion on Pt(1 1 1) and Pt(100) clusters carried out by the modified 'extended Hiickel molecular orbital' method [46,47], allowed us to infer a relatively greater stability of the CF3SO20anion on Pt(1 1 1) than on Pt(1 0 0) clusters for the 'on-top' adsorbed anion configuration through the CF3-group. The calculated distances for the SO3-group at a frozen Pt-Pt bond distance leads to 0.0625nm 2 cross section. Since the estimated cross section of the Pt(1 1 1) sites is 0.1089nm 2, a partial shielding of these sites by CF3SO20-can be structurally justified.
On the other hand, the most likely oxygen-adsorbate configuration on both Pt(1 1 1) and Pt(1 00) results from a 'bridge side-on' interaction [31]. Then, the activation energy for the oxygen-adsorbate dissociation is influenced by the presence of coadsorbates, including strongly adsorbed anions. However, when the electrochemical interphase involves a (1 1 1)-type Pt surface and a low TFMSA concentration, the adsorption of the oxygen molecule would be independent of the presence of CF3SO20-anions. Under these conditions, steps 13 to 16, involving the above discussed structural aspects, can be represented by the structural scheme shown in Fig. 9. In this scheme, (a) corresponds to the bridge side-on adsorption configuration ofO 2 on Pt(1 1 1) together with a neighbour CF3SOzOanion; (b) describes how this anion assists the OERR process through a local proton transfer leading to the oxygen-adsorbate dissociation as depicted by (c). Accordingly, in aqueous TFMSA, proton transfer is assisted by HzO--OSO2CF3 hydrogen bonding interactions, and the dissociation of the oxygen-adsorbate prior to the first monoelectronic charge transfer yields the O-containing species as depicted by (d). On the other hand, when the TFMSA concentration exceeds 1.0 M, CF3 SO20-anions will be able to shield adjacent platinum sites, and therefore, the oxygenadsorbate configuration will turn into a 'bridge end-on' configuration, as represented by the structural scheme given in Fig. 10.
In this case, (a t) represents a CF3SO20-anion and an 'end-on' oxygen molecule adsorbed on adjacent Pt(111) sites. The initial stages of the OERR would imply a number of successive configurations represented by (b') to (e').
The presence of CF3SO20-anions assists the proton transfer to the oxygen molecule (b t) without blocking platinum sites. However, when the TFMSA solution concentration is sufficiently large, the oxygen-adsorbate configuration could change as shown in (a'). The subsequent stages of the OERR on Pt(111) is a dissociation process after the first electron transfer, implying structures such as (d) and (d). The change in the configuration mode of the adsorbed ensemble (a t ) also leads to an increase in the reaction order with respect to oxygen, that is, from nearly p = 1/2 top = 1.
The influence of aqueous TFMSA and H2SO 4 concentration on the kinetics of the OERR on (1 1 D-type Pt is rather similar, as the anions of both acids are hydrogen bound to water, and this type of specific interaction seems to enhance the rate of the OERR by favouring local proton transfer. Nevertheless, there is an essential difference in the adsorption behaviour of those acids, as HSO4 and SO ]-anions are rather strongly adsorbed on platinum over a wide potential range, in contrast to the poor TFMSA adsorption. This means that in aqueous H2SO4 solution the assisted proton transfer increases the rate of the OERR, but this effect is counterbalanced by the increase in the platinum surface coverage by HSO4 and SO ]-anions. Conversely, for TFMSA at concentrations lower than 1.0M, only the enhancement of the OERR rate is observed. 0.50M, but p ~ 1 as the TFMSA concentration was increased. Conversely, for (100)-type Pt, p = 1 at all TFMSA concentrations. This change in the value of p was assigned to different CF3SO20anion-Pt interactions, which depend on the morphology of the electrode surface. (iii) RRDE data for the OERR showed that a greater amount of H202 was produced on (10 0)-type Pt than on other platinum surfaces. This behaviour has already been found for (100)-type Pt in aqueous H2SO 4 and KOH and it agrees with a 'bridge sideon' peroxo-type adsorbate formed on (100)-type Pt, but 'end-on' 02 adsorption on (1 11)-type and pc-Pt. (iv) On the basis of the Damjanovic OERR formalism, the electrochemical rate constants of the reaction steps were calculated. At all platinum surfaces, the 02 to H20 electroreduction reaction was the most important and its rate increased with cathodic overpotential. On the other hand, the rate constant for the 02 to H202 electroreduction reaction reached a maximum value at ca. 0.5V. The fastest electrodecomposition of H202 to H20 appeared for (11 1)type and pc-Pt, and its rate increased with the cathodic overpotential.
(v) The different electronic behaviour of platinum surfaces towards oxygen adsorption was assigned to the different filling of the O27r * orbitals and electronic energy of platinum surfaces, leading to a dissociative oxygen adsorption on Pt(1 11) and a molecular oxygen adsorption on Pt(1 00). A 'bridge side-on' 02 configuration (peroxo-adsorbate) was also proposed for (100)-type Pt and (111)-type Pt at low TFMSA concentrations, whereas a 'bridge end-on' 02 configuration was suggested for (11 1)-type Pt at high TFMSA concentrations.
The kinetics and mechanism of the OERR in TFMSA aqueous solutions were discussed in terms of a nonlocalized CF38020-anion adsorption on platinum surfaces. The higher OERR rates in these solutions were explained by means of a CF38020-anion assisted proton transfer to oxygen-adsorbates.