Thermal Fundamental Gap Predictions in DFT via Optimally Tuned Hybrids

Abstract

Predicting electronic fundamental gaps at finite temperature has remained conceptually and practically challenging. We address this in three connected steps. First, we extend generalized Kohn--Sham hybrid density functional theory to thermal ensembles, deriving a Mermin generalized Kohn--Sham framework from a thermal one-particle auxiliary system and an exact density-functional remainder. Second, via an extension of Janak's theorem that holds rigorously in this framework, we recast Hirata's thermal-quasiparticle picture as a thermal orbital gap estimator and derive a closed low-temperature form, the error of which is controlled by the derivative discontinuity. Third, because optimal tuning eliminates this error, the auxiliary orbital gap matches the interacting gap at low temperature, upgrading optimal tuning from a ground-state strategy to the governing principle -- mandatory, not optional -- for accurate finite-temperature gap predictions obtained from gaps of orbital eigenvalues within a hybrid functional framework. We present applications that validate the theory and demonstrate its consequences.