Revisit of the band gaps of rutile SnO2 and TiO2: a first-principles study

Abstract

From the recent experimentally observed conduction band offset and previously reported band gaps, one may deduce that the valence band offset between rutile SnO2 and TiO2 is around 1 eV, with TiO2 having a higher valence band maximum. This implication sharply contradicts the fact that the two compounds have the same rutile structure and the Γ3+ VBM state is mostly an oxygen p state with a small amount of cation d character, thus one would expect that SnO2 and TiO2 should have small valence band offset. If the valence band offset between SnO2 and TiO2 is indeed small, one may question the correctness of the previously reported band gaps of SnO2 and TiO2. In this paper, using first-principles calculations with different levels of computational methods and functionals within the density functional theory, we reinvestigate the long-standing band gap problem for SnO2. Our analysis suggests that the fundamental band gap of SnO2 should be similar to that of TiO2, i.e., around 3.0 eV. This value is significantly smaller than the previously reported value of about 3.6 eV, which can be attributed as the optical band gap of this material. Similar to what has been found in In2O3, the discrepancy between the fundamental and optical gaps of SnO2 can be ascribed to the inversion symmetry of its crystal structure and the resultant dipole-forbidden transitions between its band edges. Our results are consistent with most of the optical and electrical measurements of the band gaps and band offset between SnO2 and TiO2, thus provide new understanding of the band structure and optical properties of SnO2. Experimental tests of our predictions are called for.