Double Hybrids and Noncovalent Interactions: How Far Can We Go?

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Molecular Interactions

Background:

• Accurate evaluation of weak noncovalent interactions in large molecular systems (up to 1000 atoms) is computationally challenging.
• Approximations are often necessary to make these calculations feasible.

Purpose of the Study:

• To assess the PBE-QIDH/DH-SVPD protocol's accuracy and robustness compared to other Density Functional Approximations.
• To investigate the impact of computational parameters on the protocol's performance for large systems.

Main Methods:

• The PBE-QIDH/DH-SVPD protocol, combining a nonempirical double hybrid functional with a small basis set, was employed.
• Three datasets (S66, L7, CiM13) with increasing molecular sizes were used for evaluation.
• The performance was analyzed concerning computational parameters and the use of the DLPNO approximation.

Main Results:

• The PBE-QIDH/DH-SVPD protocol demonstrates high accuracy for large molecular systems, including those in the CiM13 set (over 1000 atoms).
• The DLPNO approximation significantly accelerates the calculation of the perturbative correlation term for large systems.
• The study highlights limitations related to assessing reference data quality and sample size for very large molecular complexes.

Conclusions:

• The PBE-QIDH/DH-SVPD protocol is a reliable and accurate method for studying weak noncovalent interactions in large molecular systems.
• Computational efficiency can be enhanced using approximations like DLPNO for large-scale calculations.
• Further validation is needed, considering the challenges in assessing reference data quality for extremely large systems.