Binding energies of molecular solids from fragment and periodic approaches

TL;DR

This study compares fragment-based and periodic boundary condition methods for calculating the binding energies of molecular solids, achieving high precision and analyzing factors affecting accuracy.

Contribution

It provides a detailed comparison of MBE and PBC approaches for molecular solids, highlighting techniques to improve convergence and reduce numerical noise.

Findings

High-precision binding energies within a few tenths of a percent

Real-space Coulomb cut-off improves PBC energy convergence

Modified summation order enhances MBE convergence

Abstract

We calculate binding energies of four molecular solids using the Hartree-Fock (HF) and second-order M{\o}ller-Plesset perturbation theory (MP2). We obtain the energies within many-body expansion (MBE) as well as using periodic boundary conditions (PBC) to compare both approaches. The systems we study are methane, carbon dioxide, ammonia, and methanol. We use tight convergence settings to obtain the binding energies with a high precision, we estimate the uncertainties to be only few tenths of percent. We discuss several issues that affect the quality of the results and which need to be considered to reach high precision for both MBE and within PBC. For example, HF as well as MP2 energies within PBC benefited from the use of real-space Coulomb cut-off technique, the convergence of energies within MBE was improved by modifying the order of summation. Finally, numerical noise made the evaluation of some of the MBE contributions difficult and the effect was reduced by using smaller basis sets for the less critical terms.