- Email: [email protected]
On the nature of platinum oxides on carbon-supported catalysts
Journal of Electroanalytical Chemistry 728 (2014) 112–117
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
On the nature of platinum oxides on carbon-supported catalysts Yan Huang a,b, Frederick T. Wagner b, Junliang Zhang b,⇑, Jacob Jorné a,⇑ a b
Materials Science Program, University of Rochester, Rochester, NY 14627, USA Electrochemical Energy Research Laboratory, GM Global R&D, Honeoye Falls, NY 14472, USA
a r t i c l e
i n f o
Article history: Received 20 October 2013 Received in revised form 24 June 2014 Accepted 30 June 2014 Available online 8 July 2014 Keywords: Charge number Pt oxide Electrocapillary theory Coverage
a b s t r a c t In fuel cells, the formation of platinum oxides on the Pt surface interferes with the Oxygen Reduction Reaction (ORR). The chemical composition of Pt oxides formed on carbon-supported Pt nanoparticles has been postulated for the first time to be hydroxylated platinum PtOH and hydroxylated platinous PtOH+ species. The thermodynamics of electrocapillary was applied to determine the charge number for Pt oxidation in various pH under a constant ionic strength perchlorate solutions. The charge number of the Pt oxide increased with the holding potential from 0.9 up to 1.5 electrons/H+ at 0.765 and 0.89 V vs. SHE, respectively. This indicates the coexistence of multi oxides, where PtOH is initially formed and is further converted at higher potentials to PtOH+. The charge number for H adsorption was measured as expected to be +1, increasing the confidence in the technique. The charge number allows the calculation of individual oxide coverages, explaining the previously intriguing reports that the fraction coverage was larger than 1 under the simplified assumption of one-electron transfer per Pt atom and without the need to assume a place-exchange mechanism. It is concluded that under increasing holding potential the coverage of PtOH+ increases while that of PtOH decreases. The formation of both oxides can be periodically reversed by a short switch to lower potentials, E < 0.6 V vs. SHE. The reduction of the oxides and consequently the recovery of Pt activity are found to be relatively fast. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Oxygen Reduction Reaction (ORR) is one of the most important electrochemical reactions in nature and technology. It is a highly sluggish reaction in need of catalysis and therefore represents the rate determining step in H2–O2 fuel cells [1–3] and metal–air batteries [4,5]. It is a well studied reaction [6–11], yet its slow mechanism [12] is far from being well understood due to its complexity and the multiple electron transfer involved [13–15]. Platinum and Pt-based catalysts remain the best catalysts for use in H2–O2 proton exchange membrane fuel cells (PEMFC), however the high cost of platinum requires a better understanding in order to minimize the amount of catalyst. Furthermore, the catalytic activity of Pt is complicated by the formation of oxides at high potentials [16]. These oxides progressively cover the Pt surface at high potentials (>0.74 V vs. SHE) thus reducing the fraction of the Pt surface available for ORR. The oxide coverage of Pt at high ⇑ Corresponding authors. Current address: Institute of Fuel Cells, MOE Key Laboratory of Power & Machinery Engineering, Shanghai Jiao Tong University, Shanghai 200240, China. Tel.: +86 21 3420 7439; fax: +86 21 3420 6249 (J. Zhang). Tel.: +1 585 275 4584; fax: +1 585 273 1348 (J. Jorné). E-mail addresses: [email protected] (J. Zhang), jacob.jorne@rochester. edu (J. Jorné). http://dx.doi.org/10.1016/j.jelechem.2014.06.040 1572-6657/Ó 2014 Elsevier B.V. All rights reserved.
potentials have been determined and well discussed, however due to experimental difficulties the exact nature and the chemical composition of these oxides have not been determined. Furthermore, it has been observed that the amount and nature of the oxides are determined by the history of the Pt and the time the Pt electrode has been held at a positive potential. The mechanism of Pt oxides formation and growth has been studied extensively [4–10]. It has been shown that Pt oxides are formed at positive potentials upon exposure to water, while surprisingly dissolved molecular oxygen has little effect [17,18]. Reaction schemes have been suggested by different studies [19–21]. The following one electron reaction has been proposed
Pt þ H2 O ! PtOH þ Hþ þ e
ð1Þ
to be followed by the possible reaction:
PtOH ! PtO þ Hþ þ e
ð2Þ
This mechanism is characterized by a charge number of 1 electron per H+. The purpose of the present work is to challenge this mechanism by experimentally determining the charge number and identifying the oxide species. The tendency at positive potentials toward the formation of Pt oxides results in a temporal decay of ORR activity [16]. Proper
113
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
Nomenclature
c q E
COH R T
surface energy (J/m2Pt) charge density (lC/cm2Pt) electric potential (V vs. SHE) Gibbs excess of adsorbed oxide species (mol/m2Pt) ideal gas constant (8.31 J/mol/K) temperature (298.15 K)
modeling of the reaction requires the knowledge of the actual charge transfer associated with oxide formation or the nature of the oxides and their coverage [22–27]. Substantial effort has been devoted to understanding the mechanism of Pt oxide formation and growth [19], and to identifying oxide species and their coverage [18], but these have been partially frustrated by the difficulty of detecting such species in situ. A logarithmic time growth of oxide [20,28,29], multiple Pt oxides [30,31] and their coverage have been reported [32,33], however, these results were limited to continuous Pt films [18,28,32], transient-state measurements [31,33] and under an unsupported assumption of one-electron transfer per Pt [16,17,24]. These experiments would not necessarily reflect the practical case of carbon-supported Pt nano-particles under high potentials, as in PEM fuel cells [1]. It is commonly assumed that one electron transfer occurs per Pt atom [16,17,33]. This assumption might be valid only for a shorttime oxide growth and is obviously inappropriate for a long-time oxide growth, where oxide fractional coverage was calculated to be significantly greater than 1 [16]. By applying electrocapillary to a solid Pt(1 1 1) electrode, Mostany et al. [34] obtained a charge number of about 1 per oxide which suggests that the oxide species is indeed PtOH. However, the measurements were restricted to a single crystal Pt(1 1 1) and to a potential window over which the cyclic voltammogram shows a reversible behavior of adsorbed oxygen species. It is expected however that the charge number would be different for carbon-supported Pt nanoparticles at large potentials due to different surface states and the polycrystalline structure of the C-supported Pt nanoparticles. In this paper, we determined for the first time the equilibrium charge number of Pt oxides formed during a long time potential hold of C-supported Pt nanoparticles. Electrocapillary theory is used to measure the charge density on Pt particles in various pH solutions (0.1 x) M HClO4 + x M NaClO4 under a constant ionic strength. The identity of the oxides (PtOH and PtOH+) and their coverage are determined as a function of the held potential. The analysis was validated for the reversible H adsorption where the charge number as expected was determined to be 1.
2. Experimental
C H3 Oþ n
CH
hydronium ion concentration (mol/L) charge number at constant chemical potential (electron/ hydronium ion) Gibbs excess of adsorbed H (mol/m2Pt)
2.2. Electrochemical measurements Electrochemical measurements were performed by the rotating-disk electrode (RDE) (Pine Research Instrumentation) technique using a glassy carbon rotating disk electrode in a 3-electrode electrochemical cell setup with a PGZ421 potentiostat (Princeton Applied Research). An Ag/AgCl electrode, connected to the working electrode compartment by a salt bridge, served as a reference. The counter electrode was a large-area Pt foil. Electrolytes were (0.1 x) M HClO4 (VeritasÒ doubly distilled, GFS Chemicals, OH, USA) + x M NaClO4 (97 + %, Sigma–Aldrich, USA) with x = 0, 5 102, 7.5 102, or 9.5 102, all diluted using Millipore ultrapure water. Care was made, by boiling glasswares in de-ionized water and using double-junction electrolyte to prevent chloride contamination from the reference electrode. The potential difference measured between Ag/AgCl and RHE in 0.1 M HClO4 was 290 mV. Potential differences between Ag/AgCl and RHE in electrolyte with other proton concentrations were calculated using the following equation:
! RT 0:1 V DE ¼ 0:29 þ ln F C H3 Oþ
ð3Þ
Electrode potentials in this paper are reported vs. the standard hydrogen electrode (SHE) with unit hydrogen-ion activity. RHE was converted onto SHE by the following formula:
Eðvs: SHEÞ ¼ Eðvs: RHEÞ þ
RT ln C H3 Oþ F
ð4Þ
As electrolyte solutions used in this study are relatively dilute solutions, activity coefficient is approximately 1 and thus, not included in the equations above. CVs were measured in a N2-saturated electrolyte at a sweep rate of 20 mV/s. The charge density was determined by integrating the oxide reduction area. In detail, the oxide reduction area was obtained by holding the electrode at several specific potentials for up to 2 h followed by a potential sweep back to 0.02 V vs. RHE at a rate of 20 mV/s (N2-saturated). The charge density of H was determined by integrating H adsorption area in regular CVs directly. All measurements were carried out at room temperature, and N2 atmosphere was maintained throughout the duration of the experiments.
2.1. Preparation of working electrode 3. Results and discussion We used a commercially available catalyst: 50 wt.% Pt supported on Vulcan carbon black (Tanaka Kikinzoku, Japan). A mixed solution of ultrapure water (Milli-QÒ system, Millipore, MA USA), isopropanol (HPLC grade, Sigma–Aldrich, USA) and a 5.37 wt.% NafionÒ solution (Sigma–Aldrich, USA) with a volume ratio of 200:50:1, mixed with an appropriate amount of catalyst, was prepared and sonicated for 5 min. Ten microliters of this ink was then transferred onto the glassy carbon surface with a geometric area of 0.196 cm2 to achieve Pt loading of 22 lg/cm2. The electrode was dried in air for 5 h before measurement.
3.1. CVs and charge densities Our approach is based on charge densities integrated from CVs in a series of solutions (0.1 x) M HClO4 + x M NaClO4 with a constant ionic strength. Fig. 1a shows CVs in an electrolyte of 0.025 M HClO4 + 0.075 M NaClO4. Before CV scanning, the C-supported Pt catalyst was held at various potentials from 0.74 to 0.89 V vs. SHE for extended periods of time (up to 2 h) in order to approach an equilibrium state for oxide coverage. This is because oxide
114
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
oxide on Pt surface, the electrocapillary equation can be expressed as:
dc ¼ qdEðSHEÞ COH RT d ln C H3 Oþ
ð5Þ
(J/m2Pt),
where c is the surface energy q is the charge density (lC/cm2Pt), and COH is the Gibbs excess of adsorbed oxide species (mol/m2Pt) [34]. A detailed derivation of this equation is shown in the Appendix B1 in the Supporting Material. Thus, c can be determined by integration of qdE at each proton concentration (by courtesy of Professor Jacek Lipkowski who helped these equations derived):
cðEÞ ¼
Z
E
qdE þ cð0:74 VÞ
ð6Þ
0:74 V
COH can be calculated by differentiation of c with respect to RT ln C H3 Oþ : RE @ 0:74 V qdE @ cðEÞ @ cð0:74 VÞ ¼ þ RT @ ln C H3 Oþ RT @ ln C H3 Oþ RT @ ln C H3 Oþ RE @ 0:74 V qdE ¼ þ Cð0:74 VÞ RT @ ln C H3 Oþ
CðEÞ ¼
ð7Þ
or
RE @ 0:74 V qdE ¼ ½CðEÞ Cð0:74 VÞ RT @ ln C H3 Oþ
ð8Þ
Then charge number at constant chemical potential n can be determined by: Fig. 1. (a) Cyclic voltammograms following potential holds (2 h) under each potential from 0.74 to 0.89 V vs. SHE in 0.025 M HClO4 + 0.075 M NaClO4 electrolyte. (b) Charge density as a function of potential in (0.1 x) M HClO4 + x M NaClO4 electrolytes with x = 0, 5 102, 7.5 102, 9.5 102 at a sweep rate of 20 mV/s.
growth on Pt nanoparticles is believed to be much slower [16,24,20,29] than on Pt(1 1 1) at low potentials, where a highly reversible behavior was observed for the single-crystal surface [25,34,35]. It can be seen that the oxide reduction area and the peak potential increased as the potential being held increased. Moreover, part of the oxide reduction charge extended down into the hydrogen underpotential deposition Hupd region, indicating that some of oxides formed on the Pt nanoparticles are highly irreversible. This suggests that oxide can exist in the Hupd region depending on the electrode history (Fig. S1). Irreversible oxides in the Hupd region were also reported at polycrystalline Pt and single crystals [36–38]. Consequently, the area difference between regular CVs and the Hupd area was taken as the integration of oxide reduction current. The total oxide charge densities as functions of hold potential and pH were obtained by integrating the areas under the CV curves and subtracting the hydrogen adsorption/ desorption areas (Fig. 1b). The charging current of the double layer in Fig. 1a serves as the baseline for the integration. Polynomial functions were constructed to best fit the dependence of the charge density on the hold potential and on the pH of the electrolytes.
3.2. Charge number at constant chemical potential of Pt oxide Electrocapillary thermodynamics were originally developed for liquid electrodes (e.g., Hg) where the surface energy could be measured directly from the contact angle. They were afterward extended to solid interfaces and in particular to Pt(1 1 1) surfaces [34]. By ignoring surface strain and stress caused by oxide formation, the fundamental Gibbs–Duhem equation could be turned into the well-known electrocapillary equation [39,40]. Considering
1 @q 1 Dq 1 qðE2 Þ qðE1 Þ ¼ F @C C F DC C F CðE2 Þ CðE1 ÞC 1 qðE2 Þ qðE1 Þ ¼ F F½CðE2 Þ Cð0:74 VÞ ½CðE1 Þ Cð0:74 VÞC
n¼
The choice of 0.74 V as the lower integration limit in Eq. (6) is because we are here interested only in the further partial derivative of charge density with respect to Gibbs excess, the charge number at constant chemical potential. Therefore, relative changes of Gibbs excess and charge density (see Eq. (8)) are enough for the calculation of charge number n. So we have greater freedom in the choice of the lower integration limit. This will be demonstrated below for the simpler case of reductive electrosorption of H+. If we were interested in calculating absolute numbers of Gibbs excess COH, the lower limit of this integration would require a knowledge of the potential of zero charge, or at least need to be chosen as a potential at which the surface charge density could be shown to be independent of pH, as was done for Pt(1 1 1) in Ref. [34]. change of the surface energy Dc Fig. 2a shows the relative RE ¼ cðEÞ þ cð0:74 VÞ ¼ 0:74 V qdE as functions of potential at each solution. The calculated Dc increased monotonically with the holding potential at various pHs. Fig. 2b shows that Dc linearly decreased with the logarithm of the proton concentration ln C H3 Oþ at each holding potential. Differentiation of Dc with respect to RT ln C H3 Oþ determines the relative change of Gibbs excess of adsorbed oxide ([C(E) C(0.74 V)]). The oxide charge density r is plotted in Fig. 3 vs. the relative change of Gibbs excess C(E) C(0.74 V), which is obtained from the slopes in Fig. 2b. As the potential increased from 0.765 V to 0.89 V vs. SHE, the magnitude of average Pt oxide charge number determined from Fig. 3 increased up to 1.5 (Fig. 4). As formation of all of the conventionally-proposed surface species (PtOH, PtO, Pt(OH)2, PtO2) and platinum/oxygen place exchange [18,21,41] as well as sub-surface oxides [36–38] involve the transfer of at most one electron per hydrogen ion released to solution (a detailed
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
115
Fig. 4. Average charge number at constant chemical potential of Pt oxide at different potentials.
but this would require PtO to have high basicity allowing it to be protonated in the relatively dilute acid solutions used here. However, another way to think of this would be PtOH losing one electron at high potentials:
PtOH ! PtOHþ þ e
Fig. 2. (a) Surface energy changes as functions of potential (vs. SHE) in various proton concentrations of electrolyte. (b) Surface energy changes as functions of proton concentration under various potentials (vs. SHE).
ð11Þ
Maclagan [42] studied the chemical bond length and bond angle of surface hydroxylated complex PtOH+. In Ref. [34], the potentialindependent charge number for Pt(1 1 1) of 1 could be readily interpreted in terms of reaction (1) with one electron transferred to Pt for each OH ion adsorbed (or, equivalently and perhaps more appropriately for an acid electrolyte, for each H+ ion generated from water). Even for such a simple case, several authors have cautioned about the interpretation of thermodynamic charge transfer numbers in terms of partial or total discharge of an adsorbed ionic species [43]. The charge number of 1.5 that we observed for Pt nanoparticles at the high end of the potential range investigated here suggests an oxide state more complex than has generally been considered. Similar analysis was applied by us to the reversible H adsorption, as shown in Appendix C of the Supporting Material. As expected, a potential-independent charge number of +1 was observed (Fig. 5), in excellent agreement with the well established adsorption of PtH and previous results for H adsorption on Pt(1 1 1) [34], further validating the present application of the electrocapillary to C-supported Pt nano-particles. 3.3. Fractional surface coverage of PtOH and PtOH+
Fig. 3. Charge densities as functions of relative Gibbs excess changes of oxide in various proton concentrations of electrolyte.
derivation is shown in the Appendix B2 in the Supporting Material), the interpretation of a charge number magnitude greater than 1 is a reaction giving 2 e per H+ released:
Pt þ H2 O ! PtOHþ þ Hþ þ 2e
ð10Þ
In combination with a more conventional reaction, such as Eq. (1), it could give an average charge number of 1 < n < 2. One way to think of PtOH+ would be as a protonated PtO species,
Charge number different than 1 indicates the presence of multi oxides. The charge numbers obtained above are in the range between 1 and 2 electrons/hydronium ion, indicating that the oxide species are likely a mixture of PtOH and PtOH+. The possibility of some PtO formation (Eq. (2)) and/or place exchange is ruled out in the present analysis for simplicity, as it appears that PtOH+ is preferably formed in accordance to Eq. (11). In a previous study [17] PtOH was assumed to be the primary surface species. Thus knowing the charge number, the individual oxide fraction coverage hPtOH and hPtOH+ can be calculated for the first time at various potentials. The individual fractional coverage are given by
hPtOH ¼ F CPtOH =qHAD
hPtOHþ ¼ F CPtOHþ =qHAD
ð12Þ
where qHAD is the charge density for H adsorption/desorption. The surface excess CPtOH and CPtOH+ can be calculated by solving the two equations for surface charge and the charge number:
hPtOH þ 2hPtOHþ ¼ Q =Q HAD
ð13Þ
116
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
at potentials lower than 1 V, fractional coverages of PtOH and PtOH+ are calculated from their average charge numbers. It is shown that the growth of the oxides is slow and the Pt electrode must be held in a given potential for a long time (2 h) in order to reach equilibrium coverage. PtOH is the main oxide at potential <0.92 V, however at higher potentials it is being replaced by the PtOH+. The reduction of the oxide is relatively faster than its slow formation. Hence it is suggested that periodic reversal of potential to potentials smaller than 0.6 V vs. SHE can recover the activity of the Pt catalyst. Our study improves the understanding the effect of oxide formation on ORR kinetics on Pt catalysts. This should facilitate a better utilization and activity recovery of Pt catalysts for fuel cells, metal–air batteries and other energy-generation and storage devices. Fig. 5. Charge numbers at constant chemical potential of hydrogen adsorption obtained from charge density integration starting from 0.37 V (black squares) and from 0.2 V (red asterisks). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Conflict of interest There is no conflict of interest. Acknowledgements We thank Professor Jacek Lipkowski, Professor James C.M. Li, Dr. Nuria Garcia-Araez and Dr. Anusorn Kongkanand for thought-provoking discussions. Yan Huang was supported by Ministry of Education of P.R. China and General Motors EERL. Catalyst was supplied by Tanaka Kikinzoku, Japan. Supplementary material Supplementary material on the derivation of electrocapillary equation and the analysis of the charge number of H adsorption can be found in the online version. Supplementary data associated with this article can be found, in the online version, at http:// dx.doi.org/10.1016/j.jelechem.2014.06.040.
Fig. 6. Effect of potential on total and individual oxide coverages of PtOH and PtOH+ in 0.1 M HClO4.
ðhPtOH þ 2hPtOHþ Þ=ðhPtOH þ hPtOHþ Þ ¼ n
ð14Þ
where Q and QHAD are the oxide charge and the hydrogen adsorption/desorption charge, respectively and n is the experimentally obtained charge number. Fig. 6 shows hPtOH and hPtOH+ as a function of the potential in 0.1 M HClO4. It can be seen that hPtOH+ increases with potential. The total coverage h = hPtOH + hPtOH+ increases with potential and reaches about 0.56 ML at 0.9 V. It is important to notice that under the assumption of one charge transfer per Pt atom, the total coverage of oxide is about 0.74 ML, which agrees well with the result of oxide coverage (under the same assumption) after 15 min of potential hold at 0.9 V vs. RHE [16]. This indicates that oxide growth and species transformation are quite slow at 0.9 V vs. RHE. In contrast, a lower total coverage was calculated at polycrystalline Pt [44]. This indicates that the total oxide coverage at Pt nano-particles is apparently higher than at the extended surface. 4. Conclusions Pt oxides and their average charge number have been studied by electrocapillary thermodynamics, and for the first time their chemical composition and individual charge numbers are determined on carbon-supported Pt nanoparticles, practically used in PEM fuel cells. The charge number of the mixed oxides up to 1.5 electron/hydronium ion, indicates that the oxide is possibly a mixture of PtOH and PtOH+. In the absence of place exchange
References [1] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Appl. Catal. B: Environ. 56 (2005) 9–35. [2] F. Maroun, F. Ozanam, O.M. Magnussen, R.J. Behm, Science 293 (2001) 1811–1814. [3] B.C.H. Steele, A. Heinzel, Nature 414 (2001) 345–352. [4] Y.C. Lu, Z. Xu, H.A. Gasteiger, S. Chen, K. Hamad-Schifferli, Y. Shao-Horn, J. Am. Chem. Soc. 132 (2010) 12170–12171. [5] M. Armand, J.M. Tarascon, Nature 451 (2008) 652–657. [6] H.A. Gasteiger, N.M. Markovic´, Science 324 (2009) 48–49. [7] V.R. Stamenkovic´, B. Fowler, B.S. Mun, G. Wang, P.N. Ross, C.A. Lucas, N.M. Marovic´, Science 315 (2007) 493–497. [8] E. Yeager, Electrochim. Acta 29 (1984) 1527–1537. [9] N.M. Markovic´, T.J. Schmidt, V. Stamenkovic´, P.N. Ross, Fuel Cells (Weinh.) 1 (2001) 105–116. [10] N.M. Markovic´, P.N. Ross, Surf. Sci. Rep. 45 (2002) 117–229. [11] R.R. Adzic, in: J. Lipkowski, P.N. Ross (Eds.), Electrocatalysis, Wiley-VCH, New York, 1998, pp. 197–242. [12] A. Damjanovic, V. Brusic, Electrochim. Acta 12 (1967) 615–628. [13] J. Zhang, K. Sasaki, E. Sutter, R.R. Adzic, Science 315 (2007) 220–222. [14] B. Lim, M. Jiang, P.H.C. Camargo, E.C. Cho, J. Tao, X. Lu, Y. Zhu, Y. Xia, Science 324 (2009) 1302–1305. [15] J. Wu, A. Gross, H. Yang, Nano Lett. 11 (2011) 798–802. [16] Y. Huang, J. Zhang, Y. Liu, N. Subramanian, F. Wagner, J. Jorné, J. Li, ECS Trans. 41 (2011) 1009–1020. [17] Y. Liu, M. Mathias, J. Zhang, Electrochem. Solid-State Lett. 13 (2010) B1–B3. [18] A. Kongkanand, J.M. Ziegelbauer, J. Phys. Chem. C 116 (2012) 3684–3693. [19] D.A. Harrington, J. Electroanal. Chem. 420 (1997) 101–109. [20] B.E. Conway, B. Barnett, H. Angerstein-Kozlowska, B.V. Tilak, J. Chem. Phys. 93 (1990) 8361–8373. [21] H. Angerstein-Kozlowska, B.E. Conway, W.B.A. Sharp, J. Electroanal. Chem. 43 (1973) 9–36. [22] J.X. Wang, N.M. Markovic´, R.R. Adzic, J. Phys. Chem. B 108 (2004) 4127–4133. [23] N.M. Markovic´, H.A. Gasteiger, B.N. Grgur, P.N. Ross, J. Electroanal. Chem. 467 (1999) 157–163. [24] N.P. Subramanian, T. Greszler, J. Zhang, W. Gu, R.R. Makharia, ECS Trans. 41 (2011) 985–1007.
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117 [25] N.M. Markovic´, T.J. Schmidt, B.N. Grgur, H.A. Gasteiger, R.J. Behm, P.N. Ross, J. Phys. Chem. B 103 (1999) 8568–8577. [26] N. Garcia-Araez, V. Climent, J.M. Feliu, J. Electroanal. Chem. 649 (2010) 69–82. [27] J.X. Wang, J. Zhang, R.R. Adzic, J. Phys. Chem. A 111 (2007) 12702–12710. [28] K.J. Vetter, J.W. Schultze, J. Electroanal. Chem. Interfacial Electrochem. 34 (1972) 141–158. [29] D.V. Heyd, D.A. Harrington, J. Electroanal. Chem. 335 (1992) 19–31. [30] T. Nagai, H. Murata, Y. Morimoto, ECS Trans. 33 (2010) 125–130. [31] P. Mathew, J. Meyers, ECS Trans. 41 (2011) 817–825. [32] M. Wakisaka, H. Suzuki, S. Mitsui, H. Uchida, M. Watanabe, J. Phys. Chem. C 112 (2008) 2750–2755. [33] R. Jinnouchi, K. Kodama, T. Hatanaka, Y. Morimoto, Intermediate Coverage in Oxygen Reduction Reaction by First Principles-based Mean Field Model, Fall 2011 ECS Meeting Abstracts, Abstract Number 854, The Electrochemical Society, Boston, 2011. [34] J. Mostany, E. Herrero, J.M. Feliu, J. Lipkowski, J. Electroanal. Chem. 558 (2003) 19–24.
117
[35] R. Gómez, J.M. Orts, B. Álvarez-Ruiz, J.M. Feliu, J. Phys. Chem. B 108 (2004) 228–238. [36] A.I. Danilov, E.B. Molodkina, Y.M. Polukarov, Russ. J. Electrochem. 40 (2004) 585–596. [37] D.V. Savinova, E.B. Molodkina, A.I. Danilov, Y.M. Polukarov, Russ. J. Electrochem. 40 (2004) 683–687. [38] A.M. Gómez-Marín, J.M. Feliu, Electrochim. Acta 82 (2012) 558–569. [39] J. Lipkowski, W. Schmickler, D.M. Kolb, R. Parsons, J. Electroanal. Chem. 452 (1998) 193–197. [40] K.J. Vetter, J.W. Schultze, Ber. Bunsen. Phys. Chem. 76 (1972) 927–933. [41] Z. Nagy, H. You, Electrochim. Acta 47 (2002) 3037–3055. [42] R.G.A.R. Maclagan, J. Mol. Struct. – Theochem. 536 (2001) 117–122. [43] R. de Levie, J. Electroanal. Chem. 562 (2004) 273–276. [44] M. Wakisaka, H. Suzuki, S. Mitsui, H. Uchida, M. Watanabe, Langmuir 25 (2009) 1897–1900.
Contents lists available at ScienceDirect
Journal of Electroanalytical Chemistry journal homepage: www.elsevier.com/locate/jelechem
On the nature of platinum oxides on carbon-supported catalysts Yan Huang a,b, Frederick T. Wagner b, Junliang Zhang b,⇑, Jacob Jorné a,⇑ a b
Materials Science Program, University of Rochester, Rochester, NY 14627, USA Electrochemical Energy Research Laboratory, GM Global R&D, Honeoye Falls, NY 14472, USA
a r t i c l e
i n f o
Article history: Received 20 October 2013 Received in revised form 24 June 2014 Accepted 30 June 2014 Available online 8 July 2014 Keywords: Charge number Pt oxide Electrocapillary theory Coverage
a b s t r a c t In fuel cells, the formation of platinum oxides on the Pt surface interferes with the Oxygen Reduction Reaction (ORR). The chemical composition of Pt oxides formed on carbon-supported Pt nanoparticles has been postulated for the first time to be hydroxylated platinum PtOH and hydroxylated platinous PtOH+ species. The thermodynamics of electrocapillary was applied to determine the charge number for Pt oxidation in various pH under a constant ionic strength perchlorate solutions. The charge number of the Pt oxide increased with the holding potential from 0.9 up to 1.5 electrons/H+ at 0.765 and 0.89 V vs. SHE, respectively. This indicates the coexistence of multi oxides, where PtOH is initially formed and is further converted at higher potentials to PtOH+. The charge number for H adsorption was measured as expected to be +1, increasing the confidence in the technique. The charge number allows the calculation of individual oxide coverages, explaining the previously intriguing reports that the fraction coverage was larger than 1 under the simplified assumption of one-electron transfer per Pt atom and without the need to assume a place-exchange mechanism. It is concluded that under increasing holding potential the coverage of PtOH+ increases while that of PtOH decreases. The formation of both oxides can be periodically reversed by a short switch to lower potentials, E < 0.6 V vs. SHE. The reduction of the oxides and consequently the recovery of Pt activity are found to be relatively fast. Ó 2014 Elsevier B.V. All rights reserved.
1. Introduction Oxygen Reduction Reaction (ORR) is one of the most important electrochemical reactions in nature and technology. It is a highly sluggish reaction in need of catalysis and therefore represents the rate determining step in H2–O2 fuel cells [1–3] and metal–air batteries [4,5]. It is a well studied reaction [6–11], yet its slow mechanism [12] is far from being well understood due to its complexity and the multiple electron transfer involved [13–15]. Platinum and Pt-based catalysts remain the best catalysts for use in H2–O2 proton exchange membrane fuel cells (PEMFC), however the high cost of platinum requires a better understanding in order to minimize the amount of catalyst. Furthermore, the catalytic activity of Pt is complicated by the formation of oxides at high potentials [16]. These oxides progressively cover the Pt surface at high potentials (>0.74 V vs. SHE) thus reducing the fraction of the Pt surface available for ORR. The oxide coverage of Pt at high ⇑ Corresponding authors. Current address: Institute of Fuel Cells, MOE Key Laboratory of Power & Machinery Engineering, Shanghai Jiao Tong University, Shanghai 200240, China. Tel.: +86 21 3420 7439; fax: +86 21 3420 6249 (J. Zhang). Tel.: +1 585 275 4584; fax: +1 585 273 1348 (J. Jorné). E-mail addresses: [email protected] (J. Zhang), jacob.jorne@rochester. edu (J. Jorné). http://dx.doi.org/10.1016/j.jelechem.2014.06.040 1572-6657/Ó 2014 Elsevier B.V. All rights reserved.
potentials have been determined and well discussed, however due to experimental difficulties the exact nature and the chemical composition of these oxides have not been determined. Furthermore, it has been observed that the amount and nature of the oxides are determined by the history of the Pt and the time the Pt electrode has been held at a positive potential. The mechanism of Pt oxides formation and growth has been studied extensively [4–10]. It has been shown that Pt oxides are formed at positive potentials upon exposure to water, while surprisingly dissolved molecular oxygen has little effect [17,18]. Reaction schemes have been suggested by different studies [19–21]. The following one electron reaction has been proposed
Pt þ H2 O ! PtOH þ Hþ þ e
ð1Þ
to be followed by the possible reaction:
PtOH ! PtO þ Hþ þ e
ð2Þ
This mechanism is characterized by a charge number of 1 electron per H+. The purpose of the present work is to challenge this mechanism by experimentally determining the charge number and identifying the oxide species. The tendency at positive potentials toward the formation of Pt oxides results in a temporal decay of ORR activity [16]. Proper
113
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
Nomenclature
c q E
COH R T
surface energy (J/m2Pt) charge density (lC/cm2Pt) electric potential (V vs. SHE) Gibbs excess of adsorbed oxide species (mol/m2Pt) ideal gas constant (8.31 J/mol/K) temperature (298.15 K)
modeling of the reaction requires the knowledge of the actual charge transfer associated with oxide formation or the nature of the oxides and their coverage [22–27]. Substantial effort has been devoted to understanding the mechanism of Pt oxide formation and growth [19], and to identifying oxide species and their coverage [18], but these have been partially frustrated by the difficulty of detecting such species in situ. A logarithmic time growth of oxide [20,28,29], multiple Pt oxides [30,31] and their coverage have been reported [32,33], however, these results were limited to continuous Pt films [18,28,32], transient-state measurements [31,33] and under an unsupported assumption of one-electron transfer per Pt [16,17,24]. These experiments would not necessarily reflect the practical case of carbon-supported Pt nano-particles under high potentials, as in PEM fuel cells [1]. It is commonly assumed that one electron transfer occurs per Pt atom [16,17,33]. This assumption might be valid only for a shorttime oxide growth and is obviously inappropriate for a long-time oxide growth, where oxide fractional coverage was calculated to be significantly greater than 1 [16]. By applying electrocapillary to a solid Pt(1 1 1) electrode, Mostany et al. [34] obtained a charge number of about 1 per oxide which suggests that the oxide species is indeed PtOH. However, the measurements were restricted to a single crystal Pt(1 1 1) and to a potential window over which the cyclic voltammogram shows a reversible behavior of adsorbed oxygen species. It is expected however that the charge number would be different for carbon-supported Pt nanoparticles at large potentials due to different surface states and the polycrystalline structure of the C-supported Pt nanoparticles. In this paper, we determined for the first time the equilibrium charge number of Pt oxides formed during a long time potential hold of C-supported Pt nanoparticles. Electrocapillary theory is used to measure the charge density on Pt particles in various pH solutions (0.1 x) M HClO4 + x M NaClO4 under a constant ionic strength. The identity of the oxides (PtOH and PtOH+) and their coverage are determined as a function of the held potential. The analysis was validated for the reversible H adsorption where the charge number as expected was determined to be 1.
2. Experimental
C H3 Oþ n
CH
hydronium ion concentration (mol/L) charge number at constant chemical potential (electron/ hydronium ion) Gibbs excess of adsorbed H (mol/m2Pt)
2.2. Electrochemical measurements Electrochemical measurements were performed by the rotating-disk electrode (RDE) (Pine Research Instrumentation) technique using a glassy carbon rotating disk electrode in a 3-electrode electrochemical cell setup with a PGZ421 potentiostat (Princeton Applied Research). An Ag/AgCl electrode, connected to the working electrode compartment by a salt bridge, served as a reference. The counter electrode was a large-area Pt foil. Electrolytes were (0.1 x) M HClO4 (VeritasÒ doubly distilled, GFS Chemicals, OH, USA) + x M NaClO4 (97 + %, Sigma–Aldrich, USA) with x = 0, 5 102, 7.5 102, or 9.5 102, all diluted using Millipore ultrapure water. Care was made, by boiling glasswares in de-ionized water and using double-junction electrolyte to prevent chloride contamination from the reference electrode. The potential difference measured between Ag/AgCl and RHE in 0.1 M HClO4 was 290 mV. Potential differences between Ag/AgCl and RHE in electrolyte with other proton concentrations were calculated using the following equation:
! RT 0:1 V DE ¼ 0:29 þ ln F C H3 Oþ
ð3Þ
Electrode potentials in this paper are reported vs. the standard hydrogen electrode (SHE) with unit hydrogen-ion activity. RHE was converted onto SHE by the following formula:
Eðvs: SHEÞ ¼ Eðvs: RHEÞ þ
RT ln C H3 Oþ F
ð4Þ
As electrolyte solutions used in this study are relatively dilute solutions, activity coefficient is approximately 1 and thus, not included in the equations above. CVs were measured in a N2-saturated electrolyte at a sweep rate of 20 mV/s. The charge density was determined by integrating the oxide reduction area. In detail, the oxide reduction area was obtained by holding the electrode at several specific potentials for up to 2 h followed by a potential sweep back to 0.02 V vs. RHE at a rate of 20 mV/s (N2-saturated). The charge density of H was determined by integrating H adsorption area in regular CVs directly. All measurements were carried out at room temperature, and N2 atmosphere was maintained throughout the duration of the experiments.
2.1. Preparation of working electrode 3. Results and discussion We used a commercially available catalyst: 50 wt.% Pt supported on Vulcan carbon black (Tanaka Kikinzoku, Japan). A mixed solution of ultrapure water (Milli-QÒ system, Millipore, MA USA), isopropanol (HPLC grade, Sigma–Aldrich, USA) and a 5.37 wt.% NafionÒ solution (Sigma–Aldrich, USA) with a volume ratio of 200:50:1, mixed with an appropriate amount of catalyst, was prepared and sonicated for 5 min. Ten microliters of this ink was then transferred onto the glassy carbon surface with a geometric area of 0.196 cm2 to achieve Pt loading of 22 lg/cm2. The electrode was dried in air for 5 h before measurement.
3.1. CVs and charge densities Our approach is based on charge densities integrated from CVs in a series of solutions (0.1 x) M HClO4 + x M NaClO4 with a constant ionic strength. Fig. 1a shows CVs in an electrolyte of 0.025 M HClO4 + 0.075 M NaClO4. Before CV scanning, the C-supported Pt catalyst was held at various potentials from 0.74 to 0.89 V vs. SHE for extended periods of time (up to 2 h) in order to approach an equilibrium state for oxide coverage. This is because oxide
114
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
oxide on Pt surface, the electrocapillary equation can be expressed as:
dc ¼ qdEðSHEÞ COH RT d ln C H3 Oþ
ð5Þ
(J/m2Pt),
where c is the surface energy q is the charge density (lC/cm2Pt), and COH is the Gibbs excess of adsorbed oxide species (mol/m2Pt) [34]. A detailed derivation of this equation is shown in the Appendix B1 in the Supporting Material. Thus, c can be determined by integration of qdE at each proton concentration (by courtesy of Professor Jacek Lipkowski who helped these equations derived):
cðEÞ ¼
Z
E
qdE þ cð0:74 VÞ
ð6Þ
0:74 V
COH can be calculated by differentiation of c with respect to RT ln C H3 Oþ : RE @ 0:74 V qdE @ cðEÞ @ cð0:74 VÞ ¼ þ RT @ ln C H3 Oþ RT @ ln C H3 Oþ RT @ ln C H3 Oþ RE @ 0:74 V qdE ¼ þ Cð0:74 VÞ RT @ ln C H3 Oþ
CðEÞ ¼
ð7Þ
or
RE @ 0:74 V qdE ¼ ½CðEÞ Cð0:74 VÞ RT @ ln C H3 Oþ
ð8Þ
Then charge number at constant chemical potential n can be determined by: Fig. 1. (a) Cyclic voltammograms following potential holds (2 h) under each potential from 0.74 to 0.89 V vs. SHE in 0.025 M HClO4 + 0.075 M NaClO4 electrolyte. (b) Charge density as a function of potential in (0.1 x) M HClO4 + x M NaClO4 electrolytes with x = 0, 5 102, 7.5 102, 9.5 102 at a sweep rate of 20 mV/s.
growth on Pt nanoparticles is believed to be much slower [16,24,20,29] than on Pt(1 1 1) at low potentials, where a highly reversible behavior was observed for the single-crystal surface [25,34,35]. It can be seen that the oxide reduction area and the peak potential increased as the potential being held increased. Moreover, part of the oxide reduction charge extended down into the hydrogen underpotential deposition Hupd region, indicating that some of oxides formed on the Pt nanoparticles are highly irreversible. This suggests that oxide can exist in the Hupd region depending on the electrode history (Fig. S1). Irreversible oxides in the Hupd region were also reported at polycrystalline Pt and single crystals [36–38]. Consequently, the area difference between regular CVs and the Hupd area was taken as the integration of oxide reduction current. The total oxide charge densities as functions of hold potential and pH were obtained by integrating the areas under the CV curves and subtracting the hydrogen adsorption/ desorption areas (Fig. 1b). The charging current of the double layer in Fig. 1a serves as the baseline for the integration. Polynomial functions were constructed to best fit the dependence of the charge density on the hold potential and on the pH of the electrolytes.
3.2. Charge number at constant chemical potential of Pt oxide Electrocapillary thermodynamics were originally developed for liquid electrodes (e.g., Hg) where the surface energy could be measured directly from the contact angle. They were afterward extended to solid interfaces and in particular to Pt(1 1 1) surfaces [34]. By ignoring surface strain and stress caused by oxide formation, the fundamental Gibbs–Duhem equation could be turned into the well-known electrocapillary equation [39,40]. Considering
1 @q 1 Dq 1 qðE2 Þ qðE1 Þ ¼ F @C C F DC C F CðE2 Þ CðE1 ÞC 1 qðE2 Þ qðE1 Þ ¼ F F½CðE2 Þ Cð0:74 VÞ ½CðE1 Þ Cð0:74 VÞC
n¼
The choice of 0.74 V as the lower integration limit in Eq. (6) is because we are here interested only in the further partial derivative of charge density with respect to Gibbs excess, the charge number at constant chemical potential. Therefore, relative changes of Gibbs excess and charge density (see Eq. (8)) are enough for the calculation of charge number n. So we have greater freedom in the choice of the lower integration limit. This will be demonstrated below for the simpler case of reductive electrosorption of H+. If we were interested in calculating absolute numbers of Gibbs excess COH, the lower limit of this integration would require a knowledge of the potential of zero charge, or at least need to be chosen as a potential at which the surface charge density could be shown to be independent of pH, as was done for Pt(1 1 1) in Ref. [34]. change of the surface energy Dc Fig. 2a shows the relative RE ¼ cðEÞ þ cð0:74 VÞ ¼ 0:74 V qdE as functions of potential at each solution. The calculated Dc increased monotonically with the holding potential at various pHs. Fig. 2b shows that Dc linearly decreased with the logarithm of the proton concentration ln C H3 Oþ at each holding potential. Differentiation of Dc with respect to RT ln C H3 Oþ determines the relative change of Gibbs excess of adsorbed oxide ([C(E) C(0.74 V)]). The oxide charge density r is plotted in Fig. 3 vs. the relative change of Gibbs excess C(E) C(0.74 V), which is obtained from the slopes in Fig. 2b. As the potential increased from 0.765 V to 0.89 V vs. SHE, the magnitude of average Pt oxide charge number determined from Fig. 3 increased up to 1.5 (Fig. 4). As formation of all of the conventionally-proposed surface species (PtOH, PtO, Pt(OH)2, PtO2) and platinum/oxygen place exchange [18,21,41] as well as sub-surface oxides [36–38] involve the transfer of at most one electron per hydrogen ion released to solution (a detailed
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
115
Fig. 4. Average charge number at constant chemical potential of Pt oxide at different potentials.
but this would require PtO to have high basicity allowing it to be protonated in the relatively dilute acid solutions used here. However, another way to think of this would be PtOH losing one electron at high potentials:
PtOH ! PtOHþ þ e
Fig. 2. (a) Surface energy changes as functions of potential (vs. SHE) in various proton concentrations of electrolyte. (b) Surface energy changes as functions of proton concentration under various potentials (vs. SHE).
ð11Þ
Maclagan [42] studied the chemical bond length and bond angle of surface hydroxylated complex PtOH+. In Ref. [34], the potentialindependent charge number for Pt(1 1 1) of 1 could be readily interpreted in terms of reaction (1) with one electron transferred to Pt for each OH ion adsorbed (or, equivalently and perhaps more appropriately for an acid electrolyte, for each H+ ion generated from water). Even for such a simple case, several authors have cautioned about the interpretation of thermodynamic charge transfer numbers in terms of partial or total discharge of an adsorbed ionic species [43]. The charge number of 1.5 that we observed for Pt nanoparticles at the high end of the potential range investigated here suggests an oxide state more complex than has generally been considered. Similar analysis was applied by us to the reversible H adsorption, as shown in Appendix C of the Supporting Material. As expected, a potential-independent charge number of +1 was observed (Fig. 5), in excellent agreement with the well established adsorption of PtH and previous results for H adsorption on Pt(1 1 1) [34], further validating the present application of the electrocapillary to C-supported Pt nano-particles. 3.3. Fractional surface coverage of PtOH and PtOH+
Fig. 3. Charge densities as functions of relative Gibbs excess changes of oxide in various proton concentrations of electrolyte.
derivation is shown in the Appendix B2 in the Supporting Material), the interpretation of a charge number magnitude greater than 1 is a reaction giving 2 e per H+ released:
Pt þ H2 O ! PtOHþ þ Hþ þ 2e
ð10Þ
In combination with a more conventional reaction, such as Eq. (1), it could give an average charge number of 1 < n < 2. One way to think of PtOH+ would be as a protonated PtO species,
Charge number different than 1 indicates the presence of multi oxides. The charge numbers obtained above are in the range between 1 and 2 electrons/hydronium ion, indicating that the oxide species are likely a mixture of PtOH and PtOH+. The possibility of some PtO formation (Eq. (2)) and/or place exchange is ruled out in the present analysis for simplicity, as it appears that PtOH+ is preferably formed in accordance to Eq. (11). In a previous study [17] PtOH was assumed to be the primary surface species. Thus knowing the charge number, the individual oxide fraction coverage hPtOH and hPtOH+ can be calculated for the first time at various potentials. The individual fractional coverage are given by
hPtOH ¼ F CPtOH =qHAD
hPtOHþ ¼ F CPtOHþ =qHAD
ð12Þ
where qHAD is the charge density for H adsorption/desorption. The surface excess CPtOH and CPtOH+ can be calculated by solving the two equations for surface charge and the charge number:
hPtOH þ 2hPtOHþ ¼ Q =Q HAD
ð13Þ
116
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117
at potentials lower than 1 V, fractional coverages of PtOH and PtOH+ are calculated from their average charge numbers. It is shown that the growth of the oxides is slow and the Pt electrode must be held in a given potential for a long time (2 h) in order to reach equilibrium coverage. PtOH is the main oxide at potential <0.92 V, however at higher potentials it is being replaced by the PtOH+. The reduction of the oxide is relatively faster than its slow formation. Hence it is suggested that periodic reversal of potential to potentials smaller than 0.6 V vs. SHE can recover the activity of the Pt catalyst. Our study improves the understanding the effect of oxide formation on ORR kinetics on Pt catalysts. This should facilitate a better utilization and activity recovery of Pt catalysts for fuel cells, metal–air batteries and other energy-generation and storage devices. Fig. 5. Charge numbers at constant chemical potential of hydrogen adsorption obtained from charge density integration starting from 0.37 V (black squares) and from 0.2 V (red asterisks). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Conflict of interest There is no conflict of interest. Acknowledgements We thank Professor Jacek Lipkowski, Professor James C.M. Li, Dr. Nuria Garcia-Araez and Dr. Anusorn Kongkanand for thought-provoking discussions. Yan Huang was supported by Ministry of Education of P.R. China and General Motors EERL. Catalyst was supplied by Tanaka Kikinzoku, Japan. Supplementary material Supplementary material on the derivation of electrocapillary equation and the analysis of the charge number of H adsorption can be found in the online version. Supplementary data associated with this article can be found, in the online version, at http:// dx.doi.org/10.1016/j.jelechem.2014.06.040.
Fig. 6. Effect of potential on total and individual oxide coverages of PtOH and PtOH+ in 0.1 M HClO4.
ðhPtOH þ 2hPtOHþ Þ=ðhPtOH þ hPtOHþ Þ ¼ n
ð14Þ
where Q and QHAD are the oxide charge and the hydrogen adsorption/desorption charge, respectively and n is the experimentally obtained charge number. Fig. 6 shows hPtOH and hPtOH+ as a function of the potential in 0.1 M HClO4. It can be seen that hPtOH+ increases with potential. The total coverage h = hPtOH + hPtOH+ increases with potential and reaches about 0.56 ML at 0.9 V. It is important to notice that under the assumption of one charge transfer per Pt atom, the total coverage of oxide is about 0.74 ML, which agrees well with the result of oxide coverage (under the same assumption) after 15 min of potential hold at 0.9 V vs. RHE [16]. This indicates that oxide growth and species transformation are quite slow at 0.9 V vs. RHE. In contrast, a lower total coverage was calculated at polycrystalline Pt [44]. This indicates that the total oxide coverage at Pt nano-particles is apparently higher than at the extended surface. 4. Conclusions Pt oxides and their average charge number have been studied by electrocapillary thermodynamics, and for the first time their chemical composition and individual charge numbers are determined on carbon-supported Pt nanoparticles, practically used in PEM fuel cells. The charge number of the mixed oxides up to 1.5 electron/hydronium ion, indicates that the oxide is possibly a mixture of PtOH and PtOH+. In the absence of place exchange
References [1] H.A. Gasteiger, S.S. Kocha, B. Sompalli, F.T. Wagner, Appl. Catal. B: Environ. 56 (2005) 9–35. [2] F. Maroun, F. Ozanam, O.M. Magnussen, R.J. Behm, Science 293 (2001) 1811–1814. [3] B.C.H. Steele, A. Heinzel, Nature 414 (2001) 345–352. [4] Y.C. Lu, Z. Xu, H.A. Gasteiger, S. Chen, K. Hamad-Schifferli, Y. Shao-Horn, J. Am. Chem. Soc. 132 (2010) 12170–12171. [5] M. Armand, J.M. Tarascon, Nature 451 (2008) 652–657. [6] H.A. Gasteiger, N.M. Markovic´, Science 324 (2009) 48–49. [7] V.R. Stamenkovic´, B. Fowler, B.S. Mun, G. Wang, P.N. Ross, C.A. Lucas, N.M. Marovic´, Science 315 (2007) 493–497. [8] E. Yeager, Electrochim. Acta 29 (1984) 1527–1537. [9] N.M. Markovic´, T.J. Schmidt, V. Stamenkovic´, P.N. Ross, Fuel Cells (Weinh.) 1 (2001) 105–116. [10] N.M. Markovic´, P.N. Ross, Surf. Sci. Rep. 45 (2002) 117–229. [11] R.R. Adzic, in: J. Lipkowski, P.N. Ross (Eds.), Electrocatalysis, Wiley-VCH, New York, 1998, pp. 197–242. [12] A. Damjanovic, V. Brusic, Electrochim. Acta 12 (1967) 615–628. [13] J. Zhang, K. Sasaki, E. Sutter, R.R. Adzic, Science 315 (2007) 220–222. [14] B. Lim, M. Jiang, P.H.C. Camargo, E.C. Cho, J. Tao, X. Lu, Y. Zhu, Y. Xia, Science 324 (2009) 1302–1305. [15] J. Wu, A. Gross, H. Yang, Nano Lett. 11 (2011) 798–802. [16] Y. Huang, J. Zhang, Y. Liu, N. Subramanian, F. Wagner, J. Jorné, J. Li, ECS Trans. 41 (2011) 1009–1020. [17] Y. Liu, M. Mathias, J. Zhang, Electrochem. Solid-State Lett. 13 (2010) B1–B3. [18] A. Kongkanand, J.M. Ziegelbauer, J. Phys. Chem. C 116 (2012) 3684–3693. [19] D.A. Harrington, J. Electroanal. Chem. 420 (1997) 101–109. [20] B.E. Conway, B. Barnett, H. Angerstein-Kozlowska, B.V. Tilak, J. Chem. Phys. 93 (1990) 8361–8373. [21] H. Angerstein-Kozlowska, B.E. Conway, W.B.A. Sharp, J. Electroanal. Chem. 43 (1973) 9–36. [22] J.X. Wang, N.M. Markovic´, R.R. Adzic, J. Phys. Chem. B 108 (2004) 4127–4133. [23] N.M. Markovic´, H.A. Gasteiger, B.N. Grgur, P.N. Ross, J. Electroanal. Chem. 467 (1999) 157–163. [24] N.P. Subramanian, T. Greszler, J. Zhang, W. Gu, R.R. Makharia, ECS Trans. 41 (2011) 985–1007.
Y. Huang et al. / Journal of Electroanalytical Chemistry 728 (2014) 112–117 [25] N.M. Markovic´, T.J. Schmidt, B.N. Grgur, H.A. Gasteiger, R.J. Behm, P.N. Ross, J. Phys. Chem. B 103 (1999) 8568–8577. [26] N. Garcia-Araez, V. Climent, J.M. Feliu, J. Electroanal. Chem. 649 (2010) 69–82. [27] J.X. Wang, J. Zhang, R.R. Adzic, J. Phys. Chem. A 111 (2007) 12702–12710. [28] K.J. Vetter, J.W. Schultze, J. Electroanal. Chem. Interfacial Electrochem. 34 (1972) 141–158. [29] D.V. Heyd, D.A. Harrington, J. Electroanal. Chem. 335 (1992) 19–31. [30] T. Nagai, H. Murata, Y. Morimoto, ECS Trans. 33 (2010) 125–130. [31] P. Mathew, J. Meyers, ECS Trans. 41 (2011) 817–825. [32] M. Wakisaka, H. Suzuki, S. Mitsui, H. Uchida, M. Watanabe, J. Phys. Chem. C 112 (2008) 2750–2755. [33] R. Jinnouchi, K. Kodama, T. Hatanaka, Y. Morimoto, Intermediate Coverage in Oxygen Reduction Reaction by First Principles-based Mean Field Model, Fall 2011 ECS Meeting Abstracts, Abstract Number 854, The Electrochemical Society, Boston, 2011. [34] J. Mostany, E. Herrero, J.M. Feliu, J. Lipkowski, J. Electroanal. Chem. 558 (2003) 19–24.
117
[35] R. Gómez, J.M. Orts, B. Álvarez-Ruiz, J.M. Feliu, J. Phys. Chem. B 108 (2004) 228–238. [36] A.I. Danilov, E.B. Molodkina, Y.M. Polukarov, Russ. J. Electrochem. 40 (2004) 585–596. [37] D.V. Savinova, E.B. Molodkina, A.I. Danilov, Y.M. Polukarov, Russ. J. Electrochem. 40 (2004) 683–687. [38] A.M. Gómez-Marín, J.M. Feliu, Electrochim. Acta 82 (2012) 558–569. [39] J. Lipkowski, W. Schmickler, D.M. Kolb, R. Parsons, J. Electroanal. Chem. 452 (1998) 193–197. [40] K.J. Vetter, J.W. Schultze, Ber. Bunsen. Phys. Chem. 76 (1972) 927–933. [41] Z. Nagy, H. You, Electrochim. Acta 47 (2002) 3037–3055. [42] R.G.A.R. Maclagan, J. Mol. Struct. – Theochem. 536 (2001) 117–122. [43] R. de Levie, J. Electroanal. Chem. 562 (2004) 273–276. [44] M. Wakisaka, H. Suzuki, S. Mitsui, H. Uchida, M. Watanabe, Langmuir 25 (2009) 1897–1900.