Theoretical and experimental study of the CO2 methanation reaction on Ni-Co catalysts.

Abstract

The CO2(g) methanation reaction on Ni, Co and NiCo catalysts was studied experimentally and theoretically. Supported Ni, Co and NiCo catalyst with narrow particle size distribution were synthetized for kinetic, isotopic and spectroscopic experiments. Monometallic and intermetallic bulk, nanoparticle, (111) and (100) surface models were constructed for DFT simulations.

A single crystalline phase was observed in all the prepared catalysts and near-surface studies for the bimetallic catalysts (EDX and CO-induced segregation) suggest that both, Ni and Co atoms, are present in the catalyst surface. DFT support the stable formation of a Ni-Co phase where Ni draws electronic density from Co, affecting their charges, magnetism and electronic structures. Co/SiO2 has the highest CO2(g) activity but lower CH4(g) selectivity than Ni/SiO2, while the NiCo/SiO2 catalysts showed lower CH4(g) formation rates and selectivity compared to both, Ni and Co catalysts. CO(g) formation rates scale linearly with the Co content of the catalysts. Kinetic and isotopic results suggest that CH4(g) is formed through an H-assisted pathway but H is not involved in the CO(g) formation. Direct CO2(g) dissociation to *CO was experimentally observed, Co and NiCo show moderate *CO2 activation barriers (~ 30 kJ/mol) via DFT and weaker *CO binding compared to Ni, consistent with their higher CO(g) selectivity. FTIR and DFT results suggest that *CO is the most abundant surface intermediate, binding in linear and/or multi-bond modes according to the metal composition. Stronger *C binding on Ni surfaces is consistent with their usual C poisoning while the more oxophilic Co surfaces may also deactivate by oxidation. The binding strength of *C and *O act as descriptors for the adsorption trends of most surface species. These descriptors also reflect highlight their structural sensitivity, different oxophilicities and suggests that bimetallic catalysts behave more similar to the Co catalysts. (111) free energy profiles show a Co<NiCo<Ni trend and lower activation barriers for the assisted HCO route with *CO+*H and HCO dissociation steps as the largest (~148 kJ/mol) and highest (200-240 kJ/mol) barriers, respectively. These barriers are ~50 kJ/mol higher than barriers in the (100) profiles. *CH2/*CH3 hydrogenations are the highest reaction barriers (120-160 kJ/mol) for Ni(100) and Co(100), followed by the *CO direct dissociation and COH barriers, respectively. Both of these activation paths are the highest barriers of the NiCo(111) profile (~160 kJ/mol). These paths produce stable *C and *OH species that may contribute to deactivate (100) surfaces. Both surface geometries show unfavorable CO2(g) adsorptions and activation barriers (>150 kJ/mol) higher than apparent experimental values (60-100 kJ/mol) and the CO(g) desorption energies (~120 kJ/mol), suggesting that other less coordinated surfaces may have larger contributions to the methanation rates and trends experimentally observed.

This work focuses on trends for the geometry and surface composition aiming to provide insight on the fundamental interactions and processes involved on the CO2(g) methanation activity, selectivity and stability of Ni, Co and NiCo catalysts while developing a solid background, workflow and tools for future theoretical studies in heterogeneous catalysts.