An initial study of the structural phase transition of SrTiO3

Abstract

SrTiO3 (STO) is a complex oxide perovskite of great technological interest for its superconductivity, blue-light emission and photovoltaic effect. In normal conditions, SrTiO3 crystallizes in the cubic Perovskite structure and undergoes a second-order phase transition to a tetragonal structure known as the antiferrodistortive (AFD) phase of STO at the critical temperature Tc = 105 K. The AFD phase of STO can appear near the interfaces at much higher temperatures if STO is used as a substrate for the growth of thin films or superlattices with other perovskites. In the last decades, both phases of STO have been extensively studied with different schemes of ab initio calculations, but none of the previously published work has been able to give, at the same time, an accurate estimate of the structural and electronic properties of the cubic and AFD phases of STO. In this work, we use Gaussian 09 to fully explain the reason behind this failure using a large spectrum of functionals ranging from pure DFT functionals like LDA and GGA to more modern and complex hybrid functional like HISS and HSE06. We also show how the quality of the basis set compete with the functional effect in predicting the properties of STO, the strongest competition being observed for the AFD phase. In fact, basis sets of low quality tend to seriously inhibit the tetragonality of the AFD phase and sometimes even suppress it. On the other hand, pure DFT functionals tend to overestimate the tetragonality of the AFD phase in agreement with previously reported results in the literature using basis sets of comparable quality. Hybrid functionals predict the structural properties of the cubic and AFD phase in very good agreement with experimental results, especially if used with high quality basis sets. Thus, we present the most reliable combination of functional and Gaussian basis set for STO currently computationally tractable. This combination gave the best agreement with the experimental structural and electronic properties for the cubic and the AFD phases of STO. It is accurate enough to enable us to understand the changes in the band structure during the cubic to AFD phase transition, predict the carrier densities, find the activation barriers for the formation and mobility of defects and the magnetic ordering.