Errors with core level shift calculation in vasp 5.2.2

[victory2401]

Hi everyone,

I am trying to do some SCLS on a series of core-shell catalyst. I try to reproduce the CLS on Rh(111) as describe in Kholer and Kresse. Phys.Rev B 70, 165405 first. However when I tried to run the Final state approximation then vasp5.2 stop immediately with those report:

running on 10 nodes
distr: one band on 1 nodes, 10 groups
vasp.5.2.2 15Apr09 complex
POSCAR found : 1 types and 24 ions
ERROR: SBESSELITER : nicht konvergent
rm_l_7_13721: (1.910156) net_send: could not write to fd=6, errno = 9
rm_l_7_13721: p4_error: net_send write: -1
p4_error: latest msg from perror: Bad file descriptor
p6_13563: p4_error: interrupt SIGx: 13
p1_14312: p4_error: interrupt SIGx: 13
p2_14392: p4_error: interrupt SIGx: 13
p8_17463: p4_error: net_recv read: probable EOF on socket: 1
rm_l_8_17538: (1.636719) net_send: could not write to fd=5, errno = 32
rm_l_9_17617: (1.382812) net_send: could not write to fd=6, errno = 9
rm_l_9_17617: p4_error: net_send write: -1
p4_error: latest msg from perror: Bad file descriptor
rm_l_4_13478: (2.691406) net_send: could not write to fd=6, errno = 9
rm_l_4_13478: p4_error: net_send write: -1
p4_error: latest msg from perror: Bad file descriptor
p8_17463: (3.640625) net_send: could not write to fd=5, errno = 32
p1_14312: (19.472656) net_send: could not write to fd=5, errno = 32
p6_13563: (18.179688) net_send: could not write to fd=5, errno = 32
p2_14392: (19.222656) net_send: could not write to fd=5, errno = 32

I did it many times but the report is still the same. If I change to use the VASP 4.6 then it can be run, but I found that even I set the value for CLNT is 24 (meaning the Rh atom on the surface), but in the OUTCAR the value for CLNT is always CLNT = 1.

So my questions are:

1) What is wrong with my INPUT files that it can not be run on vasp5.2 but can be run on vasp 4.6

2) The value of CLNT is the consequence of the atom in the POSCAR, or the consequence in the POTCAR?

Thank you very much. I appreciate if anyone can reply to me.

<span class='smallblacktext'>[ Edited ]</span>

[jbray2]

Hi,

I have run into the same problem with vasp 5.2.11. It works fine for ICORELEVEL = 1, but when I try to run the final state calculation with ICORELEVEL = 2, vasp crashes with a segmentation fault. But the same input runs fine with vasp 4.6. I have tried changing a number of other INCAR parameters to see if there were any dependencies or other causes, but nothing has worked so far.

I'm not sure why there is still no support for this feature, after more than 5 years of questions on this forum as well as several promises that it would be fully supported in vasp 5. There is still nothing in the vasp manual on this topic, and apparently it no longer works at all. Does anyone have any information about the source of this error, and if it will be fixed at some point?

Regarding the other issue, which was also asked in an earlier thread (but not responded to), you must treat the atom you want to excite as a different atom type. So the line in your POSCAR with the number of atoms should say "23 1" instead of "24", for example. Then list the atom you want excited last as the "1" atom. You must also change the POTCAR file to match, so in this case you would have Rh twice in the POTCAR, once for the first 23 Rh atoms, and once for the 1 special Rh atom.

Perhaps you have already figured this out, but hopefully this is helpful to others with the same question.

[victory2401]

Hi jbray2,

Thank you so much. I could solve this problems. Seems the CLNT should be corresponding to the consequence in the POTCAR, so if we want to specify the surface Rh atom, we should make the two same POTCARs to distinguish between them.
<span class='smallblacktext'>[ Edited Thu Apr 28 2011, 11:09AM ]</span>