Development of electrical conductivity estimation method based on tight-binding quantum chemical molecular dynamics simulation

In the development of materials for nano-scale devices, particularly those including impurities, defects, surfaces, and hetero-interfaces, estimation of their electrical conductivity based on quantum chemistry provides important information. Existing quantum mechanics-based methods for theoretically estimating the electrical conductivity require huge computational cost, which prohibit their applications to complex systems. We have already succeeded in the development of our original tight-binding quantum chemical molecular dynamics program "Colors", which realizes a calculation speed over 5000 times faster than the conventional first principles molecular dynamics methods. Here we present a novel electrical conductivity estimation method based on "Colors" combined with Monte Carlo algorithm for estimating the value corresponding to the carrier mobility. We have applied the developed method to systems of C, Si, Ge, and Sn. Calculated results quantitatively agree with experimental data, which indicates the validity of the developed method. We have also applied the developed method to the study of Si surface.