Geometry optimisations with a nonlocal density-functional theory method based on a double Hirshfeld partitioning.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Theoretical Chemistry

Background:

• Nonlocal density-functional theory (NLDFT) is a powerful method for electronic structure calculations.
• Accurate energy gradients are crucial for geometry optimization and understanding molecular properties.
• Previous NLDFT implementations lacked efficient gradient calculations.

Purpose of the Study:

• To derive and validate energy gradients for the nonlocal density-functional theory (NLDFT) method.
• To assess the accuracy of the derived NLDFT gradient for molecular geometry optimization.
• To compare NLDFT results with established high-level computational methods.

Main Methods:

• Analytical derivation of energy gradients for the NLDFT functional.
• Utilized Hirshfeld weights expressed via analytic atomic densities from Slater's rules.
• Performed geometry optimizations on 76 tripeptide molecules and small noncovalently bonded dimers.

Main Results:

• Successfully derived analytical energy gradients for NLDFT.
• Optimized structures using NLDFT gradients showed good agreement with coupled cluster and Møller-Plesset methods.
• Conformer and intermolecular interaction energies were in fair agreement with dispersion-corrected DFT methods.

Conclusions:

• The derived NLDFT energy gradients provide an accurate and efficient tool for computational chemistry.
• NLDFT with analytical gradients is suitable for geometry optimizations of peptides and noncovalently bonded systems.
• This advancement facilitates more reliable predictions of molecular structures and energies.