Credit: © 2010 ACS

Hydrogen feedstocks are generally produced through the reforming of hydrocarbons but they contain significant amounts of carbon monoxide. This creates a problem for proton-exchange membrane fuel cells (PEMFC), which use hydrogen to generate electricity, because their anodes are poisoned by CO. A solution to this problem is to lower the levels of CO in the hydrogen feedstock by selectively oxidizing it without similarly oxidizing hydrogen. Now, Bryan Eichhorn from the University of Maryland, Manos Mavrikakis from the University of Wisconsin–Madison and colleagues have used1 a combined theoretical and experimental approach to find, and understand the mechanism of, improved catalysts for the selective oxidation of CO.

A major problem with the platinum-based catalysts used to oxidize CO is that they are also poisoned by the reactants. To address this, Eichhorn, Mavrikakis and colleagues used density functional theory (DFT) to guide catalyst design. They calculated the energetics of CO adsorption to platinum surfaces supported on various transition metals (Pt/M, where M = Ru, Rh, Ir, Pd, Au) at different levels of CO coverage and how their reactivity was affected. They found that too much CO coverage inhibited the adsorption and activation of other reactive species. Therefore the bimetallic catalyst surface with the weakest CO binding (Pt/Ru) should be the most active and the surface with the strongest binding (Pt/Au) should be the least active.

The team proved this experimentally using 'M@Pt' core–shell nanoparticles analogous to the surfaces studied by DFT. After the nanoparticles were thoroughly characterized, their reactivity was studied and it showed the same order of catalytic activity predicted theoretically. All the catalysts, apart from Au@Pt, selectively oxidized CO rather than hydrogen.