Surface composition of materials used as catalysts for methanol steam reforming: a theoretical study

PdZn (1:1) alloy is assumed to be the active component of a promising catalyst for methanol steam reforming. Using density functional calculations on periodic supercell slab models, followed by atomistic thermodynamics modeling, we study the chemical composition of the surfaces PdZn(111) and, as a reference, Cu(111) in contact with water and hydrogen at conditions relevant to methanol steam reforming. For the two surfaces, we determine similar maximum adsorption energies for the dissociative adsorption of H(2), O(2), and the molecular adsorption of H(2)O. These reactions are calculated to be exothermic by about -40, -320, and -20 kJ mol(-1), respectively. Using a thermodynamic analysis based on theoretically predicted adsorption energies and vibrational frequencies, we determine the most favorable surface compositions for given pressure windows. However, surface energy plots alone cannot provide quantitative information on individual coverages in a system of coupled adsorption reactions. To overcome this limitation, we employ a kinetic model, from which equilibrium surface coverages of H, O, OH, and H(2)O are derived. We also discuss the sensitivity of our results and the ensuing conclusions with regard to the model surfaces employed and the inaccuracies of our computational method. Our kinetic model predicts surfaces of both materials, PdZn and Cu, to be essentially adsorbate-free already from very low values of the partial pressure of H(2). The model surfaces PdZn(111) and Cu(111) are predicted to be free of water-related adsorbates for a partial H(2) pressure greater than 10(-8) and 10(-5) atm, respectively.