We present results of a combined electrochemical and in situ spectroscopy study on kinetic and mechanistic aspects of the reduction of CO2 on Pt model electrodes and compare these with the performance of a Pt/C membrane electrode assembly (MEA) in a polymer electrolyte fuel cell (PEFC) for using pure H-2 or H-2/CO2(25%) mixtures (synthetic reformate). Based on highly sensitive surface enhanced IR absorption spectrocopy (SEIRAS) measurements on a thin film Pt electrode and on-line differential electrochemical mass spectrometry (DEMS) results obtained on a carbon supported Pt catalyst electrode we conclude that (i) linearly and multiply bonded COad is the only adsorbed reaction product, that GO COad formation by CO2 reduction saturates after about 30 min at coverages of about 0.45 monolayers (ML) at potentials between 0.06 and 0.2 V-RHE, and drops down to zero at 0.35 V-RHE, that (iii) in acidic solution COad formation is independent of the presence of H-2 in the electrolyte, and that (iv) even at saturation COad coverage (0.45 ML), under steady state conditions, hydrogen oxidation is hardly affected by the presence of CO2 in H-2/CO2(20-25%) mixtures. The results indicate that CO2 reduction proceeds by reaction with coadsorbed Had rather than by reaction with dissolved H-2 and is therefore favored in the H-upd region, i.e., at fuel cell relevant anode potentials. CO2 reduction is kinetically hindered by self-poisoning and blocking of Pt ensembles at COad coverages, which are still low enough for H-2 oxidation. These limitations in CO2 reduction are responsible for the relatively small performance losses for PEFCs operating on H-2/CO2 mixtures, compatible with a picture where H-2 Oxidation proceeds in holes of the CO adlayer, which are too small for CO2 reduction. (C) 2005 Elsevier Ltd. All rights reserved.