Assessing in Silico Models and Methods for Non-Covalent Interactions between Graphene and Small Molecules

Designing and optimising graphene-based gas sensors, which involve physisorption of analytes on the sensor surface, requires theoretical insights into the strength and nature of such non-covalent interactions. This modelling entails constructing appropriate atomistic representations for an infinite graphene sheet and its complex with the analyte, then selecting accurate yet affordable methods for geometry optimisations and energy computations. In this work, density functionals from the 2^nd to 5^th rungs of Jacob’s ladder, coupled cluster theory, and symmetry-adapted perturbation theory in conjunction with a range of surface models, from benzene to the periodic system, were tested for their ability to reproduce experimental adsorption energies of CO₂ on graphene in a low-coverage regime. The best agreement with the reference computations was found for global and double hybrid density functionals, while experimental adsorption energies were reproduced within chemical accuracy by extrapolating the SAPT0//DSD-BLYP-D3 interaction energies from finite clusters to infinity. This simple yet powerful extrapolation scheme effectively removes size dependence from the data obtained using finite cluster models, and the latter can be treated at more sophisticated levels of theory relative to periodic systems.