[Wien] 3R crystal structure WS2 SCF problem

[Tran, Fabien]

Hi,


I am not sure, but this kind of crash can occur when the basis set is too large and leads to linear dependency. What is the smallest RMT in your cell and what RMT*KMAX did you choose in case.in1? ?


Comment: The position of atoms should not matter at all. In addition, you can not change position of atoms since the symmetry operations will no be the correct ones.


________________________________
From: Wien <wien-bounces at zeus.theochem.tuwien.ac.at> on behalf of Luigi Maduro - TNW <L.A.Maduro at tudelft.nl>
Sent: Wednesday, April 1, 2020 5:10 PM
To: wien at zeus.theochem.tuwien.ac.at
Subject: [Wien] 3R crystal structure WS2 SCF problem


 Dear WIEN2k users,

I am running WIEN2k version 19.1 with the 2017 version 4 intel fortran compiler. I am trying to run an SCF calculation of 3R WS2 unit cell.
The .struct file was created with the crystallographic information found in the american mineralogist crystal structure database:

Spacegroup 160
a = b = 3.158 Angstrom, c = 18.49 Angstrom, alpha = beta = 90, gamma = 120
W = 0 0 0
S =  0 0 0.2497
S = 0 0 0.4917

I used hex2rhomb to write down the atomic coordinates into rhombohedral coordinates, as is needed by WIEN2k, giving the atomic coordinates:

W = 0 0 0
S =  0.2497 0.2497 0.2497
S = 0.4190 0.4190 0.4190

However, when I try to run a regular SCF cycle the calculation crashes at the first SCF cycle when it reaches the lapw1 script.
The following error is found in lapw1.error:

Cholesky INFO =      274
'SECLR4' - POTRF (Scalapack/LAPACK) failed

In the past I noticed that when there are atoms at the borders of the unit cell the SCF calculation crashes, so I thought that maybe that this can be fixed by shifting the atoms a bit in the unit cell. The new atomic coordinates are, after using hex2rhomb:

W = 0.15 0.15 0.15
S =  0.3997 0.3997 0.3997
S = 0.5690 0.5690 0.5690

However, this gives the same error as before.
Any suggestions on how to go forward is highly appreciated.


Kind regards,
Luigi Maduro


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://zeus.theochem.tuwien.ac.at/pipermail/wien/attachments/20200401/5c4ac4cf/attachment.html>