[Pw_forum] SCF Convergence Problem

[paresh rout]

Dear QE Users and Developers,

I am doing a 100 atomic interface calculation . When I relaxed the 100
atomic bulk with conv_thr=1.D-10 , the  estimated scf accuracy remains
constant like
   estimated scf accuracy   <       0.00000004 Ry   until the maximum
iterations reaches. I reduced my conv_thr value to 1.D-8 and getting the
same problem. Here I am using
      smearing="mv"
      degauss=0.05
      mixing_beta = 0.3 . I had also played with various degauss and beta
values but the result did not improve. By the above parameters and with a
 conv_thr=1.D-7 the scf is converging. The same problem occurs in *slab
interface* *with dipole correction* while relaxing the structure. For slab
calculation I did a single scf without dipole then I started from the
previous potential for optimization but the above problem still persisting .
How to converge the scf  cycle with a lower conv_thr=1.D-9 or -10 ? Any
help would be highly appreciated.

Regards
Paresh Chandra Rout
Research Scholar
Indian Institute of Science Education and Research Bhopal
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20141124/78c2eefb/attachment.html>