Problems with localized f-electrons

Message

marsoner · #1 Post by **marsoner** » Fri May 30, 2014 3:18 am

Hello all,

I'm experinecing the following problem. I'm trying to calculate a Sm compound and want to treat the f-electrons explicitly.

I have preoptimized the structure using the Sm_3 POTCAR and had no problems but upon switching to the Sm one I only get:

Error EDDDAV: Call to ZHEGV failed. ...

Now I have seen all the suggestions on this forum on how to resolve this type of problem but non seem to apply in my case...

I am using DFT+U as a first step wih U,J being taken from the literature and want to switch ro HSE06 later on.

Does anyone have any suggestions on how to resolve this?

Thanks and have a nice day,
Lukas

#2 Post by **admin** » Mon Jun 02, 2014 10:38 am

This error appears due to wrong geometry. Atomic dimensions and bonding properties of Sm_3 and Sm are different. Try to use another starting geometry, not that preoptimized with Sm_3.

marsoner · #3 Post by **marsoner** » Tue Jun 03, 2014 4:29 am

Dear head Admin,

thank you for the response.

I have tested different settings but was unsuccessful. I used the theoretical values from the literature (http://journals.aps.org/prb/pdf/10.1103 ... .75.045114) which was very close to the Sm_3 POTCAR value to begin with, but had the same problems i.e. crash after one electronic step. I tried different initial values for the lattice constant as well but that did not help.

Calculations with hybrid DFT and standard PBE gave the same problem.

rabi · #4 Post by **rabi** » Thu Sep 25, 2014 11:41 pm

I got the same error when I switch from Sm_3 POTCAR file to Sm POTCAR, no matter whether I used optimized (with Sm_3 potential) or with experimental structure. Could you please let me know if you were able to fix this issue? or This is bug with the potential supplied with the vasp itself.
Thank you!
Rabi

#5 Post by **admin** » Tue Dec 02, 2014 4:36 pm

Dear Lukas, please attach your input files.

My Community