ABINIT Discussion Forums

Dear all
Can I used LDA pseudopotential file with ixc = 11, 23 of GGA-PBE, GGA-WC respectively

Best.
Abdulmutta

No, I think that just changing the ixc line in the pseudopotential file, or setting ixc differently in the abinit input from the pseudopotential files, are both bad ideas. You really need to have your pseudopotential files consistent with abinit and also internally self-consistent. If you want to use a different exchange-correlation model, the only real alternative is to re-make the pseudopotential files with the new ixc.

hi, all
Dear mutta and Prof: jzwanzig
see this message :
* Report this post
* Reply with quote
psuedopotential and ix
Postby cheti2010 » Sun Apr 11, 2010 8:51 am
Dear all:
is it possible for example to used pseudopotential with LDA approximation (pspnc) and ixc with GGA (ixc=11) ? , on other hand,
cheti2010
Posts: 1
Joined: Sat Apr 10, 2010 9:27 pm
* Private message
Top
* Report this post
* Reply with quote
Re: psuedopotential and ix
Postby gonze » Sun Apr 11, 2010 8:14 pm
Dear Cheti2010,
It is possible to have ABINIT input ixc=11 (GGA-PBE) while the pseudopotential has not been obtained
using this exchange-correlation functional. However this is not consistent.
You should have look at M. Fuchs et al, Phys. Rev. B 57, 2134 (1998)
Best wishes,
X
• the links is: Board index ‹ Running Abinit ‹ Response calculations<psuedopotential and ix
• In Page 3

Yes, I know it's *possible*, my point was that it's a bad idea and shouldn't be done. I think XG is saying roughly the same thing by saying that it's "not consistent".