J. Phys. Chem. B 2006, 110, 14355-14362
14355
High-Throughput Synthesis and Screening of Ternary Metal Alloys for Electrocatalysis Samuel Guerin, Brian E. Hayden,* Christopher E. Lee, Claire Mormiche, and Andrea E. Russell School of Chemistry, UniVersity of Southampton, Highfield, Southampton SO17 1BJ, United Kingdom ReceiVed: March 30, 2006; In Final Form: May 9, 2006
We report the application of a new method for the high-throughput synthesis and screening of thin film materials and its application to the discovery of electrocatalysts. Results are presented for the PtPdAu ternary alloy system with respect to activity for oxygen reduction. The results reveal an enhancement in activity for a range of PtPd alloy compositions over either of the pure elements. An optimum composition range of ternary alloys with significant activity was also identified. A correlation was also investigated between the surface reduction potential and the activity for oxygen reduction in both binary and ternary alloys. The results demonstrate the potential of the methodology for the discovery and optimization of electrocatalysts for a wide range of applications.
1. Introduction The search for metal alloy electrocatalysts which exhibit an enhancement in activity over the pure components has been the focus of considerable effort with respect to a number of important reactions. Among the most important of these reactions has been that of oxygen reduction, which takes place at the cathode of polymer electrolyte membrane fuel cells (PEMFC) on carbon-supported platinum. There have been a number of reports of an enhancement in the activity of platinum for oxygen reduction by the incorporation of a second element,1,2 and recently this possibility has stimulated theoretical studies in the search for the optimum alloy combination.3 Additionally, alloying components can provide a degree of tolerance to poisoning in low-temperature fuel cells, the most notable cases being that for anodes of reformate fuelled PEMFCs and of the direct methanol fuel cell (DMFC).2 Another notable reaction that benefits from alloying metal components is nitrate reduction.4 In the case of PEMFC, the cost and abundance of platinum make platinum-free catalysts highly desirable; however, the most likely route to the development of PEMFC for automotive applications lies with platinum-lean catalysts.5 In addition to the constituent elements it is of interest to optimize the relative amounts of the active components for electrocatalysts alloys and materials. For example, it is well-known that oxidation of methanol in fuel cells is improved significantly by using PtRu bimetallic electrocatalysts with optimized surface compositions. The preparation of electrodes with defined and wide-ranging surface compositions is challenging. Approaches taken include the co-electrodeposition of the elements,6,7 the underpotential deposition of a second component on a surface,4 chemical deposition of a second component on a surface,8 or studies on bulk alloy samples.9 However, these methods suffer from a number of difficulties when applied to the investigation of the electro-catalytic activity of a surface and certainly do not lend themselves to a high-throughput methodology attempting to investigate large compositional ranges. The disadvantages in this respect include difficulties in ensuring mixing and alloy formation, the separation into bulk phases, and surfaces segrega* Corresponding author.
tion of individual components. All these effects make it difficult to correlate electrochemical activity and surface composition. Gel-transfer synthetic methods applied in a combinatorial methodology also suffer from the difficulty in ensuring welldefined ranges of bulk and surface alloy compositions.10 At the other extreme, electrochemical studies on fully characterized surfaces prepared in UHV conditions and studied ex situ have been achieved on binary surface alloys on single crystals,11-14 but the methods do not lend themselves to the screening of electrochemical activity over a wide compositional range. A range of high-throughput screening methodologies have been developed,15-17 but their combination with an effective synthetic methodology is paramount in their success in identifying promising catalytically active materials. We have developed a method to deposit thin film materials of varying composition through simultaneous deposition of the component elements.18 The method allows the preparation of graded compositional thin films of alloy materials through the controlled and simultaneous deposition of the component elements on a substrate at room temperature. A significant advantage of this method is that the alloys formed are not subjected to annealing steps which could alter the relative bulk and surface compositions and cause the formation of thermodynamically stable phases. Rather, we are able to prepare thin films alloys whose composition is determined by the deposition conditions and which are best described as nonequilibrium solid solutions of the alloy components. Thin film synthesis was carried out in an ultrahigh vacuum (UHV) high-throughput physical vapor deposition (HT-PVD) chamber incorporating six sources (three effusion and three e-beam) arranged at 60° from one another. Substrate entry is via a fast load lock and interconnecting buffer line, which also allows UHV transfer between a number of other chambers, including one with surface analytical facilities. Up to six elements can be deposited simultaneously, and the variation of film thickness across a substrate from any unhindered source is typically Pd > Au as expected. The addition of Au to Pt (Figures 6A and 7A) causes an approximately linear decrease in oxygen reduction activity across the composition range. This suggests that Au is an inert diluent with respect to the active platinum component. Note that the effect of introducing Au into Pt caused a monotonic decrease in the surface reduction potential corresponding to a slight increase in the surface oxygen bond strength (Figure 5A). The addition of Au to Pd (Figures 6C and 7C) shows a sharp
14360 J. Phys. Chem. B, Vol. 110, No. 29, 2006
Guerin et al.
Figure 5. Peak potentials for the reduction of the surface oxides of (A) Au and Pt in PtAu, (B) Pd and Pt in PdPt, (C) Au and Pd in AuPd binary compositions, and (D) the Pt and Pd components in PtAuPd ternary compositions.
TABLE 1: Comparison of Mean Surface and Bulk Alloy Compositions in the Four Quadrants (Figure 4) of the Thin Film Array segment 1
2
3
4
element:
Pt
Pd
Au
Pt
Pd
Au
Pt
Pd
Au
Pt
Pd
Au
EDS XPS
10 15
82 75
9 10
32 40
40 37
28 23
20 30
75 66
5 4
77 80
10 12
13 8
decrease in oxygen reduction activity over the first 20-30% Au composition range. This corresponds to the region in which the surface oxygen stability on the Pd decreases sharply (Figure 5C). The activities across the PtPd binary composition range are shown in Figures 6B and 7B. There is clearly a maximum in the activity in the composition range of 50-90% Pt that is potential-dependent. At the higher potential (lower overpotential for ORR) where the currents are more sensitive to kinetic control, the maximum in activity is more clearly defined and found in a composition range between 70 and 90% Pt. Over this composition range there is very little downward shift in surface reduction potential during the alloying of Pt by Pd (Figure 5B), and the maximum in oxygen reduction activity corresponds to alloys with a surface reduction potential only very slightly lower than that of pure Pt. The inference is that only a small increase in surface oxygen bond energy over pure Pt (increased overpotential) is required to enhance the oxygen reduction activity. The oxygen reduction activities of the ternary alloys at potentials of 0.7 and 0.8 V are also shown in Figures 6 and 7. It is evident from inspection of these plots that the highest oxygen reduction activity for both potentials is located near the Pt-Pd axis, as one would expect on the basis of the binary results: The effect of Au to both Pd and Pt is to reduce the oxygen reduction activity, and it is clear that addition of Au to the PtPd binary compositions is also to reduce the activity of the alloy. However, the compositions of highest activity for any particular Au composition in the ternary alloy correspond to the alloys containing PtPd ratios similar to those found for the maximum in activity of the PtPd binary. To illustrate this, a
line corresponding to this atomic ratio of PtPd of 65:35 and 80:20 in the ternary phase is shown in Figures 6 and 7, respectively. These are the compositions with maximum activity in the PtPd binary alloys at the two potentials. The regions of maximum activity are centered on these pseudo-binary lines. Note that significant oxygen reduction activities are retained with PdPtAu alloys containing 20-30% Au with these Pt:Pd ratios. Since it is the present price and natural abundance of the Pt component which underlies the search for lean Pt-based fuel cell catalysts, it would not be unreasonable to consider the specific activity normalized to the platinum content to assess the significance of these activity results. For example, for oxygen reduction at 0.8 V, Pt52Pd28Au20 retains a steady-state oxygen reduction activity which is about the same as that for pure Pt. This corresponds to a specific activity related to the atomic platinum component in the ternary alloy of about twice that of pure platinum. There is clearly a significant scope to increase platinum utilization for oxygen reduction with this alloy system. The PtPd binary alloy is likely to be unstable under oxygen reduction conditions, which makes it unsuitable for cathode catalysis in PEMFC. It is not clear at present if the incorporation of Au in the PtPd provides the additional stability which would be required for a realistic oxygen reduction catalyst in PEMFC. It has been suggested that the addition of Au to PdCo improves its stability toward long-term, steady-state oxygen reduction.17 In addition, the cost and abundance of the Au and Pd components are not ideal for the development of a commercially viable Pt lean alloy catalyst. Nevertheless, the methodology presented here shows how the activity of ternary alloy materials can be assessed effectively, and therefore the same methodology
Synthesis and Screening of Ternary Metal Alloys
J. Phys. Chem. B, Vol. 110, No. 29, 2006 14361
Figure 6. Steady-state oxygen reduction activity of (A) PtAu, (B) PdPt, (C) AuPd, and (D) PtAuPd compositions at 0.70 V in 0.5 M HClO4(aq).
Figure 7. Steady-state oxygen reduction activity of (A) PtAu, (B) PdPt, (C) AuPd, and (D) PtAuPd compositions at 0.80 V in 0.5 M HClO4(aq).
is being used to identify ternary phases of Pt with less expensive alloying components, and indeed Pt-free catalyst materials. It is evident that the extension of ORR activity of PtPd alloy into the Au containing phases (Figures 6 and 7) shows an apparent correlation with the PtPd surface reduction potential
(Figure 5). Ternary alloy compositions with PdPt surface reduction taking place in the region 0.65-0.70 V in general correspond to alloys with maximum oxygen reduction activity. This is the same range of potential observed for the PdPt binary alloy which also exhibits high activity. Ternary alloy composi-
14362 J. Phys. Chem. B, Vol. 110, No. 29, 2006 tions with either higher or lower surface reduction potentials than this range exhibit the lowest activity. These observation are consistent with the results of electronic structure calculations which suggest that a perturbation in O or OH surface bond energy very far from that of platinum will result in a lower oxygen reduction activity.3 The increase in activity with PdPt indicates that there is some scope for optimization electronically through a minor change in surface bond energies (the O and OH binding energies are similar on the two metals).3 There are, however, high gold (>60 at. %) containing regions of alloy composition (Figure 5), which also exhibit surface reduction potentials in the 0.65-0.7 V range, which show no activity for oxygen reduction (Figures 6 and 7). A possible explanation for a breakdown in such a correlation of surface oxygen stability and ORR activity may be the influence of an ensemble requirement for oxygen reduction.
Guerin et al. ternary alloys. Although the optimal composition of oxygen reduction can be assessed using the high-throughput methodology applied here, the scope for the development of such a ternary alloy for PEMFC applications would depend on the ability of Au to stabilize the PtPd system in the cathode environment. Perhaps most importantly these results show that the information obtained allows the exploration of correlations in surface composition and catalytic activity in the search for a better understanding of the composition-activity relationship in multicomponent alloys. Acknowledgment. This work has been supported by the EPSRC of the Combinatorial Centre of Excellence in Southampton through the JIF scheme, EPSRC grant GR/R50639, by Johnson Matthey and by Ilika Technologies Ltd. References and Notes
4. Conclusions A high-throughput method for the synthesis of thin film materials when combined with a high-throughput electrochemical screening system can be successfully used to characterize the surface reduction and oxygen reduction activity of ternary metal alloys.16,18 The alloy components are deposited simultaneously with the substrate at room temperature, and the resulting polycrystalline material consists of randomly mixed component elements (nonequilibrated), with very similar bulk and surface compositions. We suggest that such materials are ideal for the screening of electrocatalytic activity as a function of composition. To achieve the screening of individual fields of various compositions, a microfabricated chip incorporating 100 individually addressable elements has been employed. Both cyclic voltammetric and potential step measurements have been achieved quickly and consistently. This is demonstrated with the screening of the PtPdAu ternary alloy phase for oxygen reduction activity, and the results of measurements on ca. 800 samples are presented. Comparison of the surface reduction potential of the PtPd component, and the activity of the alloy for oxygen reduction, reveal a correlation of surface oxygen stability and catalytic activity. Optimal activity in the binary PtPd alloy corresponds to a surface reduction potential slightly lower than that on pure platinum, and deviation of the surface reduction potential significantly from this value corresponds to a reduction in catalytic activity in the ternary alloy. The results are broadly in line with the predictions of ab initio calculations, which indicate near optimal oxygen reduction activity for Pt and Pd.3 Nevertheless, our results make it apparent that there is scope for increased utilization of platinum in both the PtPd binary and PtPdAu
(1) Thompsett, D. Pt alloys as oxygen reduction catalysts. In Handbook of Fuel Cells-Fundamentals, Technology and Applications; Vielstich, W., Lamm, A., Gasteiger, H. A., Eds.; John Wiley & Sons: Chichester, 2003; Vol. 3, p 467. (2) Antolini, E. Mater. Chem. Phys. 2003, 78, 563. (3) Norskov, J. K.; Rossmeisl, J.; Logadottir, A.; Lindqvist, L.; Kitchin, J. R.; Bligaard, T.; Jonsson, H. J. Phys. Chem. B 2004, 108, 17886. (4) de Vooys, A. C. A.; van Santen, R. A.; van Veen, J. A. R. J. Mol. Catal. A: Chem. 2000, 154, 203. (5) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal., B 2005, 56, 9. (6) Frelink, T.; Visscher, W.; van Veen, J. A. R. Surf. Sci. 1995, 335, 353. (7) Lukaszewski, M.; Grden, M.; Czerwinski, A. J. Solid State Electrochem. 2005, 9, 1. (8) Lee, C. E.; Bergens, S. H. J. Phys. Chem. 1998, 102, 193. (9) Gasteiger, H. A.; Markovic, N.; Ross, P. N.; Cairns, E. J. J. Phys. Chem. 1993, 97, 12020. (10) Jayaraman, S.; Hillier, A. C. Measure. Sci. Technol. 2005, 16, 5. (11) Davies, J. C.; Hayden, B. E.; Pegg, D. J.; Rendall, M. E. Surf. Sci. 2002, 496, 110. (12) Hayden, B. E.; Rendall, M. E.; South, O. J. Am. Chem. Soc. 2003, 125, 7738. (13) Davies, J. C.; Hayden, B. E.; Pegg, D. J. Surf. Sci. 2000, 467, 118. (14) Hayden, B. E. Single-Crystal Surfaces as Model Platinum Based Hydrogen Fuel Cell Electrocatalysts. In Catalysis and Electrocatalysis at Nanoparticle Surfaces; Wieckowski, A., Ed.; Marcel Dekker: New York, 2002. (15) Reddington, E.; Sapienza, A.; Gurau, B.; Viswanathan, R.; Sarangapani, S.; Smotkin, E. S.; Mallouk, T. E. Science 1998, 280, 1735. (16) Guerin, S.; Hayden, B. E.; Lee, C. E.; Mormiche, C.; Owen, J. R.; Russell, A. E. J. Comb. Chem. 2004, 6, 149. (17) Fernandez, J. L.; Walsh, D. A.; Bard, A. J. J. Am. Chem. Soc. 2005, 127, 357. (18) Guerin, S.; Hayden, B. E. J. Comb. Chem. 2006, 8, 66. (19) Markovic, N.; Gasteiger, H.; Ross, P. N. J. Electrochem. Soc. 1997, 144, 1591. (20) Davey, W. P. Phys. ReV. 1925, 25, 753. (21) Bredig, G.; Allolio, R. Z. Phys. Chem. 1927, 126, 41. (22) Woods, R. Electrochim. Acta 1971, 16, 655.
