Electronic transport of organic-inorganic hybrid perovskites from first-principles and machine learning

Abstract

Using the data-mining machine learning technique and the non-equilibrium Green's function method in combination with density functional theory, we studied the electronic transport properties of the organic-inorganic hybrid perovskite MAPbI3. The band structures of MAPbI3 from first-principles show that the ferroelectric and antiferroelectric dipole configurations have very little influence on the energy bandgap. Furthermore, we investigated the tunnel junctions made of MAPbI3 and 48 different metal electrodes, with the same fixed lattice constant as MAPbI3. With the increase in the number of perovskite unit cells, the electron transmission coefficients are found to decrease exponentially in general. For data mining studies, several different methods are employed to develop models for predicting electron transport properties. In particular, the gradient boosting regression tree model was tested and found to be the most effective tool among all these algorithms for fast prediction of the electron transmission coefficients and performance ranking of all studied metal electrodes.