Toward Versatility: A Flexible Generalized Gradient Approximation Exchange Functional

Abstract

In a quest for a versatile generalized gradient approximation (GGA) functional, we have constructed a new flexible exchange (X) functional, named Ghosh-Oshi-Salahub (GOS), that unifies the thermochemistry of a broad range of finite molecules (G2 set), transition metal compounds, weakly bonded complexes and the lattice constant prediction of periodic solids with all types of conductivity, outperforming the PBE and WC X functionals. Rooted in a rational functional framework, GOS smoothly interpolates among slowly, moderately and rapidly varying density regimes using two tunable parameters, offering a bounded, monotonic, damping-controlled and saturating enhancement factor that rigorously satisfies most of the ab initio constraints, while preserving numerical stability and analytical simplicity. GOS is recommended for general use whenever GGA accuracy is deemed to be adequate. Its versatile performance also recommends it for the construction of hybrid functionals and pseudopotentials, enabling state-of-the-art molecular and solid-state applications.