Properties of AgBiI4 using high through-put DFT and machine learning methods
Victor T. Barone, Blair Tuttle, S. V. Khare
Abstract
Silver iodo-bismuthates show promise for optoelectronic and other applications. Within this family of materials, AgBiI4 is a prominent model compound. The complexity of AgBiI4 has prevented a conclusive determination of specific atomic arrangements of metal atoms in the bulk material. Here, we employ high through-put density functional and novel machine learning methods to determine physically relevant unit cell configurations. We also calculate the fundamental properties of the bulk material using newly discovered configurations. Our results for the lattice constant (12.7 Å) and bandgap (1.8 eV) agree with the previous theory and experiment. We report new predictions for the bulk modulus (7.5 GPa) and the temperature-dependent conductivity mass for electrons (m0 at T = 300 K) and holes (7m0 at T = 300 K); these masses will be useful in AgBiI4-based device simulations.