Canonical coupled cluster binding benchmark for nanoscale noncovalent complexes at the hundred-atom scale

Area of Science:

• Computational chemistry
• Quantum chemistry
• Materials science

Background:

• Accurate prediction of noncovalent binding is crucial for understanding molecular interactions.
• Existing benchmarks for large molecular complexes are limited.
• Coupled cluster with single, double, and perturbative triple [CCSD(T)] excitations extrapolated to the complete basis set (CBS) limit is a high-accuracy method.

Purpose of the Study:

• To create and benchmark datasets for nanoscale noncovalent binding.
• To evaluate the accuracy of computational methods for large systems.
• To identify reliable methods for studying noncovalent interactions.

Main Methods:

• Development of two datasets (L14 and vL11) for nanoscale noncovalent binding.
• Benchmarking using canonical and local coupled cluster with single, double, and perturbative triple [CCSD(T)] excitations extrapolated to the complete basis set (CBS) limit.
• Comparison with fixed-node diffusion Monte Carlo (FN-DMC) and evaluation of other electronic structure methods.

Main Results:

• Canonical CCSD(T)/CBS benchmarks extended to 113 atoms, and local CCSD(T)/CBS to 174 atoms.
• Local CCSD(T)/CBS shows good agreement with canonical CCSD(T)/CBS.
• Fixed-node diffusion Monte Carlo (FN-DMC) underestimates binding energies in π-π complexes by over 1 kcal/mol.
• MP2+aiD(CCD), PBE0+D4, and ωB97X-3c are recommended for nanoscale noncovalent interactions.

Conclusions:

• The fixed-node approximation in FN-DMC requires further investigation for binding discrepancies.
• Recommended methods (MP2+aiD(CCD), PBE0+D4, ωB97X-3c) offer a balance of accuracy and computational efficiency for nanoscale systems.
• The developed datasets serve as valuable references for future computational chemistry studies.