About LDIPOL

Message

nujjj · #1 Post by **nujjj** » Wed Feb 09, 2005 7:51 pm

About the LDIPOL I have another question, so the manual talkes about obtain preconverge result first and then turn on LDIPOL so that it can fastly reach convergence. So does it mean that I have to turn on the writing flag of WAVECAR?

a.soon · #2 Post by **a.soon** » Thu Feb 17, 2005 2:39 am

Hi,
From what I understand, to deal with an asym slab, setting LDIPOL=T and IDIPOL=3 (given that the z-axis sym is broken) should be sufficient. This seemed to give reasonable results.

Regards,
Aloysius

#3 Post by **admin** » Mon Apr 11, 2005 1:41 pm

LDIPOL need not be set for asymmetric slabs, right.
Yes, it improves the convergence speed if the WAVECARS are kept (which is
default) and read for the next step
-- set ISTART = 1, and
-- do not forget to start with CONTCAR of the previous ionic step (copy
it to POSCAR and continue)

konglt · #4 Post by **konglt** » Mon Aug 15, 2005 1:05 pm

"LDIPOL need not be set for asymmetric slabs, right. "
I am confused with this sentense and what has been said in the manual. Should I set LDIPOL=.TRUE. in the INCAR fiel or not on earth?
Thanks.
<span class='smallblacktext'>[ Edited Mon Aug 15 2005, 03:49PM ]</span>