Density-functional theory vs density-functional fits

Area of Science:

• Computational Chemistry
• Quantum Mechanics
• Materials Science

Background:

• Kohn-Sham density-functional theory (DFT) is a cornerstone of modern electronic structure calculations.
• Early DFT models relied on simple, parameter-free approximations.
• Recent advances involve extensive empirical fitting, leading to numerous parameters and potential loss of physical insight.

Purpose of the Study:

• To develop a new density functional that prioritizes theoretical modeling over extensive empirical fitting.
• To evaluate the performance of this theoretically grounded functional against highly parameterized methods.
• To assess the efficacy of physically motivated parameters in DFT approximations.

Main Methods:

• Development of a novel density functional based on theoretical principles for exchange-correlation modeling.
• Performance evaluation using the GMTKN55 chemical reference dataset.
• Comparison against established, heavily parameterized Kohn-Sham functionals.

Main Results:

• The proposed density functional, with few physically motivated parameters, demonstrates competitive accuracy.
• Its performance rivals that of extensively fitted functionals on the GMTKN55 benchmark.
• This suggests that theoretical rigor can yield highly accurate DFT approximations.

Conclusions:

• A theoretically derived DFT functional can achieve high accuracy comparable to empirical methods.
• Minimizing empirical parameters while adhering to physical principles is a viable strategy for developing robust DFT functionals.
• This work advocates for a return to theory-driven development in DFT approximations.