Spin and electronic property prediction of IV–V binary monolayers using deep knowledge transfer method

Abstract

The spin–orbit coupling (SOC) correction associated to the electronic and spin properties of IV–V binary monolayers using deep transfer learning model is estimated. The accuracy of more than 99% is obtained for the prediction of electronic properties. In addition, the correlation between the spin-splittings at the K-point for both conduction and valence bands of different compounds is studied by this model. Their corresponding spin properties demonstrate inter-dependent behavior in the conduction band while less correlation is obtained in the valence band. Furthermore, the four most effective band splitting SOC corrections at the valence band are also found using the square of atomic numbers with an error of less than 10−3.