Jiayu Dai wrote: > How could we know the heavily doped semiconductor is a metal or > semiconductor? > Or, the fermi energy calculated is not appropriate in this case?
it depends on how "heavily" the semiconductor is doped. For doping in the order of a few %, you can use a supercell and everything will work. Realistic (i.e. order of magnitudes smaller) doping levels in semiconductors cannot be simulated in this way. You have to resort to phenomenological models in which the doping produces free charges and the Fermi energy determines how many of them. Paolo -- Paolo Giannozzi, Democritos and University of Udine, Italy
