Elsevier

Electrochimica Acta

Volume 47, Issues 22–23, 30 August 2002, Pages 3621-3628
Electrochimica Acta

Quantum-chemical calculations of CO and OH interacting with bimetallic surfaces

https://doi.org/10.1016/S0013-4686(02)00332-8Get rights and content

Abstract

In this work we present results of a periodic density-functional theory study of the adsorption of carbon monoxide (CO) and hydroxyl (OH) on platinum–ruthenium, platinum–molybdenum and platinum–tin alloys as well as the adsorption of CO on a series of transition metals modified with a Pt overlayer. The surfaces are modelled as four-layer slabs (three-layer slab in case of Pt3Sn(111)). The binding energies and geometries of CO and OH are computed. In the case of PtRu, the mixing of Pt by Ru leads to a weaker bond of both CO and OH to the Pt sites, whereas mixing of Ru by Pt causes a stronger bond of CO and OH to the Ru sites. The binding energy trends for CO do not show a clear-cut relationship with its vibrational characteristics. The mixing of Pt by Mo leads to weakly adsorbed CO on both Pt and Mo sites, and OH strongly adsorbed only on Mo sites. This suggests that PtMo could be a better bifunctional catalyst for CO oxidation then PtRu. On Pt3Sn(111) the calculations show that CO binds only to Pt and not to the Sn, whereas OH has an energetic preference for the Sn sites. This also implies that PtSn should be a good CO oxidation catalyst. For Pt–monolayer systems, we demonstrate a relationship between the PtPt distance in the monolayer and the changes in the CO binding energy. The nature of the substrate seems to be of secondary importance.

Introduction

Most currently considered CO tolerant catalysts for low-temperature polymer membrane fuel cell catalysts consist of platinum and a second or even third metal to improve overall performance [1], [2]. The most important issues in contemporary fuel cell electrocatalysis are related to developing a material that either oxidizes CO (as present in fuel reformate or as an intermediate in direct methanol oxidation) at low overpotential, or that adsorbs a limited amount of CO while still capable of oxidizing hydrogen at an acceptable rate [1], [2], [3].

There are many fundamental surface-science studies devoted to the elucidation of the factors involved in CO adsorption and oxidation on well-defined electrode surfaces [3], [4], [5], [5](a), [5](b), [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19]. These studies have mainly focused on bimetallic Pt systems, i.e. Pt either alloyed or modified with other elements such Ru, Sn, Mo, Re, Pd, and Rh. The system that has been studied in the greatest detail is PtRu, as this is still one of the most stable and active CO tolerant catalysts under practical conditions. It is generally believed that the mechanism by which CO is removed from the PtRu surface is the so-called ‘bifunctional mechanism’ [20]. In this mechanism, CO adsorbed on Pt is oxidized preferentially by an oxygen-containing surface species on Ru. Most authors write the oxygen-containing surface species as OH, even if the exact nature of this species is still unknown. In reaction equations, the bifunctional mechanism reads as:H2O+Ru*↔OHRu+H++eCOPt+OHRuCO2+H++e+Ru*+Pt*

However, this mechanism does not take into account a possible change in the CO binding energy on Pt induced by Ru (often referred to as an electronic effect), nor does it directly describe the effect of CO and OH competitive adsorption on the Ru. For instance, recent experiments using well-defined PtSn and PtMo electrodes [7], [8], [21], [22] have shown that both surfaces are better CO oxidation catalysts than PtRu, even though both may not be particularly useful as practical CO tolerant hydrogen or methanol oxidation catalysts. This enhanced performance compared with PtRu was ascribed to a lack of adsorption of CO on either Sn or Mo, leaving more adsorption sites for oxygen-containing species. In a sense, they should correspond more closely to ‘true’ bifunctional catalysts, with one element interacting preferentially with the first reactant, and the other element with the second reactant.

In this paper, we describe the results of quantum-chemical calculations of CO and OH interacting with a variety of bimetallic surfaces, in order to assess from the computational viewpoint the molecular nature of the bifunctional mechanism and the electronic effects involved. We will consider in some detail the results of density-functional calculations of CO and OH adsorption on PtRu, PtMo, PtSn, and a number of Pt-modified transition metals such as Rh and Ir. All surfaces are modeled as semi-infinite slabs in a periodic supercell. The advantage of using quantum-chemical calculations is that they give direct information on the binding energetics of the different species, which are relatively difficult to extract from electrochemical measurements. Moreover, calculations permit to establish the quantum-chemical nature of the surface bond and the different factors involved, and how these might be related to measurable properties of chemisorbates such their vibrational characteristics. In this way, quantum-chemical calculations provide invaluable information to complement and correctly interpret experimental data, even if the calculated systems considered may appear rather idealized compared with the experimental catalysts. This computational approach to understanding catalytic and electrocatalytic systems is obviously not new (for recent reviews, see Refs. [23], [24]), and we will compare our results to other calculations where appropriate. In particular, there have been some pioneering semi-empirical cluster calculations by Anderson related to fuel cell catalysis [25], [26], [27]. More recently, DFT cluster and slab calculations for CO and OH on PtRu have appeared and we will compare our results to those computations [28], [29], [30], [31], [32]. Especially the slab calculations (as considered here) are of interest as these are now well established to give the most reliable results.

Section snippets

Computational methods

DFT-GGA periodic slab calculations were performed using the Vienna Ab Initio Simulation Package VASP [33], [33](a), [33](b), [33](c). Unless specified otherwise, the (√3×√3)R30° unit cell was used to model the adsorption of CO and OH, corresponding to a 1/3 ML coverage on the surface. The surfaces were modeled as four-layer slabs, with the atomic distances fixed to the calculated bulk lattice distance, and a vacuum, corresponding to five equivalent layers. The top metal layer was relaxed in all

CO and OH on PtRu

The current state-of-the-art CO tolerant electrocatalyst is a platinumruthenium bimetallic alloy. As mentioned in the introduction, the mechanism for CO tolerance is believed to involve the adsorption of oxygen containing species (OHads) on the ruthenium sites such that CO adsorbed on platinum can participate in a bimolecular reaction with the activated oxygen species.

Recently we reported a detailed study of CO and OH adsorption on PtRu alloy surfaces [32]. The binding energies and

Conclusions

In this paper, we have presented the results of a periodic DFT-GGA study of CO and OH interacting with variety of bimetallic surfaces.

As our most important conclusions, concerning the PtRu alloy, we found that mixing of Pt by Ru weakens the binding of both CO and OH to the Pt surface sites. By contrast, the CO and OH binding gets stronger as Pt is mixed in. The surface with the weakest CO binding energy in our calculation is PtML–Ru(0001). Interestingly, a similar surface was recently found to

Acknowledgements

This work was financially supported by the Energy research Center of the Netherlands (ECN) and the Royal Netherlands Academy of Arts and Sciences (KNAW).

References (56)

  • H.A Gasteiger et al.

    Catal. Lett.

    (1996)
  • N.M Markovic et al.

    Electrochim. Acta

    (2000)
  • R Ianniello et al.

    Electrochim. Acta

    (1994)
  • W Chrzanowski et al.

    Catal. Lett.

    (1998)
  • A Kabbabi et al.

    J. Electroanal. Chem.

    (1998)
  • W.E Lin et al.

    J. Phys. Chem. B

    (1999)
  • H Massong et al.

    Electrochim. Acta

    (2000)
  • S Mukerjee et al.

    Electrochem. Solid State Lett.

    (1999)
  • M.T.M. Koper, R.A. van Santen, M. Neurock, in: E.R. Savinova, C.G. Vayenas, A. Wieckowski, (Eds.), Catalysis and...
  • R.C Binning et al.

    Int. J. Quant. Chem.

    (2000)
  • Q Ge et al.

    J. Phys. Chem. B

    (2001)
  • M.C Payne et al.

    Rev. Mod. Phys.

    (1992)
  • K Bedürftig et al.

    J. Chem. Phys.

    (1999)
  • M.O Pedersen et al.

    Surf. Sci.

    (1999)
  • P.N Ross

    J. Vac. Sci. Technol. A

    (1992)
  • G Hoogers et al.

    Cat. Tech.

    (2000)
  • L Carrette et al.

    Fuel Cells

    (2001)
  • N.M. Markovic, P.N. Ross Jr., Surf. Sci. Reports, 45 (2002)...
  • H.A Gasteiger et al.

    J. Phys. Chem.

    (1994)
  • H.A Gasteiger et al.

    J. Phys. Chem.

    (1995)
    H.A Gasteiger et al.

    J. Phys. Chem.

    (1995)
  • N.M Markovic et al.

    Electrochim. Acta

    (1995)
  • B.N Grgur et al.

    J. Phys. Chem. B

    (1998)
  • B.N Grgur et al.

    Electrochim. Acta

    (1998)
  • W Chrzanowski et al.
  • J.C Davies et al.

    Electrochim. Acta

    (1998)
  • R Liu et al.

    J. Phys. Chem. B

    (2000)
  • I Igarashi et al.

    Phys. Chem. Chem. Phys.

    (2001)
  • H Massong et al.

    Electrochim. Acta

    (1998)
  • Cited by (201)

    • Materials in PEM Fuel Cells

      2021, Encyclopedia of Smart Materials
    • Understanding the interplay of bifunctional and electronic effects: Microkinetic modeling of the CO electro-oxidation reaction

      2020, Journal of Catalysis
      Citation Excerpt :

      Bimetallics in which the secondary site is selective for OH* adsorption – as is usually argued to be the case for Pt-Sn and Pt-Mo systems – contain secondary sites (i.e. Sn, Mo) that have a low affinity for CO* binding, but a high affinity for OH* [45]. This differs from the Pt-Ru system, where the Ru sites bind both CO* and OH* more strongly than Pt, and therefore the ability of Ru to provide OH* to remove CO* from the Pt sites could, under certain conditions, be precluded by CO* poisoning on the Ru sites themselves [28,44]. This distinction highlights the importance of accounting for the coverages of all species on both site types when performing a kinetic analysis of potentially-bifunctional catalysts.

    View all citing articles on Scopus
    View full text