[Pw_forum] finite temperature DFT

戴佳钰 daijiayu at nudt.edu.cn
Thu Mar 19 13:51:27 CET 2009

Dear Stefano,

>The Mermin functional is obtained if you fill the electronic states  
>according to the Fermi-Dirac distribution. use occupations='smearing'  
>and smearing='fermi-dirac'. Temperature must be converted in rydberg  
>and passed in the smearing_width variable.
>When using this at high temeperature check carefully that the number  
>of bands used is sufficient to include all partially occupied  
>levels... FD distribution decays rather slowly w.r.t its argument.

Thanks for your answer, and it is very helpful. 

So, you mean the temperature is set to be smearing_width here, all right? And if
want to perform molecule dynamics, i have to use the tempw and the method of
thermostat at the same time? Then, the tempw and the smearing_width must be
And do we need to consider the chemical potential of the ensemble?

Thanks a lot.

Jiayu Dai
National University of Defense Technology, P R China

More information about the users mailing list