Geometry Optimizations in a Subsystem Density Functional Theory Formalism: A Benchmark Study.

Area of Science:

• Computational Chemistry
• Quantum Chemistry
• Materials Science

Background:

• Subsystem density functional theory (sDFT) is used for electronic structure calculations.
• Accurate geometry optimization is crucial for predicting molecular properties.

Purpose of the Study:

• To benchmark subsystem density functional theory (sDFT) for equilibrium structure optimization.
• To evaluate new analytical gradient implementations in the SERENITY program.

Main Methods:

• Geometry optimizations were performed on the S22 and A24 test sets using various exchange-correlation (XC) and nonadditive kinetic-energy functionals.
• The study compared Gaussian-type orbital (GTO) and Slater-type orbital (STO) implementations.

Main Results:

• The combination of BP86/LLP91 with D3 dispersion corrections provided the most accurate results for sDFT geometry optimizations.
• Some XC/nonadditive kinetic-energy functional combinations partially mimicked dispersion effects.
• Differences between GTO and STO implementations were generally small.

Conclusions:

• The BP86/LLP91 functional with empirical dispersion corrections is a robust and accurate strategy for sDFT geometry optimizations.
• This approach is preferable to methods relying on error cancellation between XC and nonadditive kinetic-energy functionals.
• The SERENITY program with its new analytical gradient implementation shows promise for sDFT calculations.