The reaction of CO₂ and H₂ with ZnO/SiO₂ catalyst at 295 K gave predominantly hydrogencarbonate on zinc oxide and a small quantity of formate was evolved after heating at 393 K. Elevation of the reaction temperature to 503 K enhanced the rate of formation of zinc formate species. Significantly these formate species decomposed at 573 K almost entirely to CO₂ and H₂. Even after exposure of CO₂–H₂ or CO–CO₂–H₂ mixtures to highly defected ZnO/SiO₂ catalyst, the formate species produced still decomposed to give CO₂ and H₂. It was concluded that carboxylate species which were formed at oxygen anion vacancies on polar Zn planes were not significantly hydrogenated to formate. Consequently it was proposed that the non-polar planes on zinc oxide contained sites which were specific for the synthesis of methanol.

The interaction of CO₂ and H₂ with reduced Cu/ZnO/SiO₂ catalyst at 393 K gave copper formate species in addition to substantial quantities of formate created at interfacial sites between copper and zinc oxide. It was deduced that interfacial formate species were produced from the hydrogenation of interfacial bidentate carbonate structures. The relevance of interfacial formate species in the methanol synthesis reaction is discussed.

Experiments concerning the reaction of CO₂–H₂ with physical mixtures of Cu/SiO₂ and ZnO/SiO₂ gave results which were simply characteristic of the individual components. By careful consideration of previous data a detailed proposal regarding the role of spillover hydrogen is outlined.

Admission of CO to a gaseous CO₂–H₂ feedstock resulted in a considerably diminished amount of formate species on copper. This was ascribed to a combination of over-reduction of the surface and site-blockage.