Top Page > Browsing

Spi-orbit coupling Date: 2021/01/13 19:40 Name: Artem Pulkin   <gpulkin@gmail.com> What is the difference between setting scf.spinpolarization off scf.spinorbit.coupling off and specifying zeros in scf.so.factor? On the Bi2Se3 example I learned that the first option produces a substantially larger band gap than the second one.

Page: [1]

Re: Spi-orbit coupling ( No.1 ) Date: 2021/08/01 11:48 Name: T. Ozaki Hi, The case: scf.spinpolarization off scf.spinorbit.coupling off corresponds to the collinear spin-unpolarized calculation with scalar relativistic pseudopotentials generated with taking account of the j-degeneracy as V_l = ( (l+1)*V_{l+1/2} + l*V_{l-1/2})/(2*l+1) On the other hand, the case: scf.spinpolarization NC scf.spinorbit.coupling on specifying zeros in scf.so.factor corresponds to the non-collinear calculation with zero spin-orbit coupling, where the pseudpotentials are generated by sum = (V_{l+1/2} + V_{l-1/2})/2 dif = (V_{l+1/2} - V_{l-1/2})/2 V'_{l+1/2} = sum + scf.so.factor * dif V'_{l-1/2} = sum - scf.so.factor * dif Therefore, if a system is non-magnetic, the difference can be attributed to how to generate the scalar relativistic pseudopotentials. These manipulation can be confirmed around lines 902-925 of SetPara_DFT.c. Regards, TO

Page: [1]