Consistent GMTKN55 and molecular-crystal accuracy using minimally empirical DFT with XDM(Z) dispersion

Area of Science:

• Computational chemistry
• Quantum chemistry
• Materials science

Background:

• Density-functional theory (DFT) is crucial for computational chemistry.
• Dispersion corrections like the exchange-hole dipole moment (XDM) model are vital for accurate modeling.
• Existing XDM implementations use a two-parameter Becke-Johnson (BJ) damping function.

Purpose of the Study:

• Introduce and implement a new XDM variant with a one-parameter Z damping function.
• Benchmark Z-damped XDM against BJ-damped XDM using the GMTKN55 database.
• Evaluate transferability of Z damping to solid-state systems.

Main Methods:

• Implemented a new one-parameter Z damping function for XDM.
• Benchmarked XDM variants on the GMTKN55 database with various DFT functionals.
• Performed outlier analysis using the WTMAD-4 metric.
• Tested Z damping on molecular crystal benchmarks.

Main Results:

• The Z-damped XDM variant demonstrated excellent performance across molecular and solid-state benchmarks.
• revPBE0 and B86bPBE0 hybrid functionals paired with Z-damped XDM showed superior results.
• First-time testing of XDM and many-body dispersion (MBD) on the GMTKN55 dataset.

Conclusions:

• The Z-damped XDM variant offers a promising advancement for high-accuracy DFT calculations.
• This method shows excellent performance and consistency for both molecular and solid-state systems.
• The new damping function improves the modeling of large-scale systems in computational chemistry.