Pt-decorated two-dimensional silicene/WSe2 heterostructures for CO, NH3 and NO2 gas adsorption: a density functional theory machine learning combined study

Abstract

This study examines Pt-decorated silicene/WSe2 heterostructures for CO, NH3 and NO2 gas sensing by combining density functional theory and crystal graph convolutional neural networks. The pristine heterostructure exhibits metallic behaviour and robust thermomechanical stability. Gas adsorption induces pronounced modifications in adsorption energetics, charge redistribution, electronic band structure, projected density of states and optical responses. Among the examined species, NO2 shows the strongest interaction, functioning as an electron acceptor and producing substantial spin polarization and electronic reconstruction. Optical characteristics, including the dielectric function, absorption coefficient and joint density of states, are significantly enhanced upon adsorption, with NO2 yielding the most pronounced changes. Transport properties, such as the Seebeck coefficient, electrical conductivity and thermal conductivity, exhibit clear sensitivity to the adsorbed molecule. These findings demonstrate that Pt/silicene/WSe2 heterostructures are promising candidates for future optoelectronic and gas-sensing applications.