Balancing physical constraints and empiricism

Here is our newest contribution to the zoo of density functional approximations.
J. Chem. Phys.
154, 061101 (2021)
It’s a collaboration with John P. Perdew and coworkers from University of Bonn, Tulane University, Temple University as well as Dassaults Systèmes. We balance physical constraints with minor empiricism. This way, we construct the dispersion corrected meta-generalized gradient approximation r²SCAN-D4. It’s unique features are:

  • Satisfying 16 exact physical constraints
  • Numerically robust
  • Computationally affordable with cubic scaling cost
  • Highly accurate molecular geometries
  • Excellent performance on main group and metal-organic thermochemistry, kinetics, and non covalent interactions
Balancing physical constraints and empiricism