[Pw_forum] Convergence within electric field

Sergey Lisenkov proffess at yandex.ru
Wed Apr 20 08:34:53 CEST 2005


Dear PWscf authors and users,

I would like to study a system under electric field. First of all, I relaxed my structure without electric field. I got the convergence (with respect to k-points and geometry) using 5 k-points:

1kpt.out:     Final energy            =    -890.3164211588 ryd
2kpt.out:     Final energy            =    -890.2930246276 ryd
3kpt.out:     Final energy            =    -890.2927598985 ryd
4kpt.out:     Final energy            =    -890.2930024805 ryd
5kpt.out:     Final energy            =    -890.2930121231 ryd

After that I started the calculations using relaxed geometry and applied electric field. So, the convergence with respect to both k-points and geometry is not yet achieved:

 6kpt.out:     Final energy            =    -900.8160339804 ryd
 7kpt.out:     Final energy            =    -900.8186989102 ryd
 8kpt.out:     Final energy            =    -900.8184366599 ryd
 9kpt.out:     Final energy            =    -900.8178203905 ryd
10kpt.out:     Final energy            =    -900.8176632906 ryd
11kpt.out:     Final energy            =    -900.8174330457 ryd
12kpt.out:     Final energy            =    -900.8171295549 ryd
13kpt.out:     Final energy            =    -900.8172131375 ryd
14kpt.out:     Final energy            =    -900.8170159823 ryd
15kpt.out:     Final energy            =    -900.8166501310 ryd

For each calculation (I mean increasing k-points) it takes around 30 bfgs steps. Is it normal behavior?

 Thank you very much,
   Sergey






More information about the users mailing list