Energy Gap from Step Structure of the Analytically Inverted Non-Additive Kinetic Potential

Abstract

The bandgap constitutes a challenging problem in density functional theory (DFT) methodologies. It is known that the energy gap values calculated by common DFT approaches are underestimated. The bandgap was also found to be related to the derivative discontinuity (DD) of the exchange-correlation potential in the Kohn-Sham formulation of DFT. Several reports have shown that DD appears as a step on the potential curve. The step structure is a mandatory structure for aligning the KS energy levels in the ionization potentials in a dissociated molecule in both fragments and is a function of electron localisation. Reproducing the step in the DFT framework gives the charge transfer process and the correct energy gap and describes the source of dissociation. This step phenomenon has not yet been studied in the non-additive kinetic potential $v^{\text{NAD}}[\rho_A,\rho_B](\textbf{r})$, a key quantity used in embedding theories. While $v^{\text{NAD}}[\rho_A,\rho_B](\textbf{r})$ is known to be difficult to approximate, in this work, we explain how an accurate energy gap can be produced from the analytically inverted $v^{\text{NAD}}[\rho_A,\rho_B](\textbf{r})$, even if we use the input densities calculated by the local and semi-local functionals. We used the precisely calculated $v^{\text{NAD}}[\rho_A,\rho_B](\textbf{r})$ reported in our previous publication [Phys. Rev. A 106, 042812 (2022)] to produce the energy gap for some model systems and report in this work the promising accuracy of our results through the comparison with the results obtained from one of the most accurate calculations, OEP theory with the KLI local approximation.