[Wien] A few basic questions about LAPWSO

[Stefaan Cottenier]

> I want to include spin-orbit interaction in my calcualtion of LiFePO4
> without spin polarization. The non-SO scf converged well. For the SO scf,
> I have a few quick questions.
> 1. How to specify the llmax in case.inso? Does it have a good estimation
> from non-SO files or just taken based on your experience?

This is a default that as far as I know does hardly ever needs to be changed.

> 2. If no magnetic field, shoud I set the direction of magnetization to 000
> or just leave it alone?

Leave it as it is (without runsp there is not such a thing as a  
magnetization anyway).

> 3. How to choose the e-parameter in case.inso? I tried to set it as the
> average of the E-bottom and E-top shown in the case.scf1 with same l. But
> always have ghostband.

It has to be tested, in the same way as RKmax. Start with say 2.5 and  
increase in steps of 0.5 until you have convergence. It depends on the  
properties you are interested in, but very often 3.0 will do the job.

Stefaan


Disclaimer: http://www.kuleuven.be/cwis/email_disclaimer.htm