Challenges and Opportunities in Modeling Oxides for Energy and Information Devices

The growth of computational resources has enabled investigations of large-scale and highly correlated problems by using first principles computational techniques such as density functional theory (DFT). In context of oxide materials, these problems include oxide surface reconstructions (Diebold et al. 2010), diffusion and reaction barriers in heterogeneous systems (Chizallet and Raybaud 2014; Aksyonov et al. 2018), phase diagrams for transition metal oxides (Park et al. 2014; Leonov 2015), and point defects as well as extended defects (Youssef and Yildiz 2012; Sun et al. 2015). These developments have opened up new opportunities for predicting not only the bulk crystal properties of oxides, but also the effect of complex microstructures such as associated point defects (Hu et al. 2013; Liu et al. 2012; Zhang et al. 2014; T-Thienprasert et al. 2012), grain boundaries (Polfus et al. 2012; McKenna and Shluger 2009; Hojo et al. 2010), dislocations (Sun et al. 2015; Hojo et al. 2011; McKenna 2013), and surfaces (Lee and Morgan 2015; Freysoldt and Neugebauer 2018; Bajdich et al. 2015) under thermodynamic drivers. These developments can ultimately allow for ab initio prediction of realistic device performance. Yet, challenges remain on both the theoretical and algorithmic level to accurately predict oxide materials properties on a complex potential energy surface. Here we summarize several growing fields in addressing these challenges and present our perspectives on future directions that these methods will enable