CO₂ on Graphene: Benchmarking Computational Approaches to Noncovalent Interactions

Area of Science:

• Computational materials science
• Surface science
• Physical chemistry

Background:

• Designing graphene-based gas sensors requires accurate computational modeling of gas molecule adsorption.
• Selecting cost-effective yet precise methods for geometry optimization and energy calculations is crucial for *in silico* design.

Purpose of the Study:

• To assess various computational methods for modeling CO2 adsorption on graphene.
• To determine the most accurate and efficient theoretical approaches for predicting adsorption energies and geometries.

Main Methods:

• Utilized density functionals (DFs), coupled cluster theory (CCSD(T)), and symmetry-adapted perturbation theory (SAPT).
• Employed finite and periodic surface models of bare and supported graphene.
• Developed a scheme to extrapolate interaction energies from finite clusters to infinite surfaces.

Main Results:

• Extrapolation from finite clusters accurately reproduced periodic results, removing size dependence.
• Inexpensive DFs (e.g., PBE-D3) showed good agreement with CCSD(T) for small models.
• SAPT extrapolation and nonlocal van der Waals functionals yielded interaction energies closest to experimental values for CO2 on graphene.

Conclusions:

• A robust method for *in silico* design of graphene gas sensors was established by extrapolating finite model data.
• SAPT and nonlocal van der Waals functionals are recommended for accurate CO2 adsorption energy calculations on graphene.
• Current methods fail to reproduce the experimentally observed tilted CO2 adsorption geometry on Pt(111), indicating a need for advanced theoretical models or experimental revision.