Page 1 of 1

NEB result analysis problems

Posted: Mon Jun 12, 2006 11:51 pm
by hwang4
Hi, recently I calculated NEB, and tried to analyze the results by using the scripts downloaded from Dr Henkelman's website: http://theory.cm.utexas.edu/vtsttools/scripts/

I used nebef.pl
the output is like this:
0 0.00000000 -271.22179600 0.00000000
1 0.00000000 -271.19307400 0.02872200
2 0.00000000 -271.04762400 0.17417200
3 0.00000000 -270.56514300 0.65665300
4 0.00000000 -270.80316400 0.41863200
This output seems OK.
But when I use nebbarrier.pl, the content of the output file neb.dat is:
0 0.000000 0.000000 0.000000 0
1 0.000000 0.028722 0.000000 1
2 0.000000 0.174172 0.000000 2
3 0.000000 0.656653 0.000000 3
4 0.000000 0.418632 0.000000 4

The second column is supposed to give me the reaction coordinate distance of each image, and the fourth column is supposed to give me the forces.
With the above neb.dat, I can not use nebspline.pl to get the final result spline.dat. The error is:
Illegal division by zero at ~/bin/nebspline.pl line 64.

I manually checked my OUTCAR of each image, there is some distance in the OUTCAR, e.g.
" left and right image 0.350349 0.350349 A"
I could manually finish the analysis task by checking this distance of each OUTCAR, but I would like to take advantage of the script, which I think should be correct, right? So anyone can give a hint? What's possibly wrong? Anyone got similar sitiuations?

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

NEB result analysis problems

Posted: Tue Jun 13, 2006 12:03 pm
by konglt
I wonder if you allowed some degree of freedom, probably you fixed all atoms and no relaxation is actually performed.

NEB result analysis problems

Posted: Tue Jun 13, 2006 1:27 pm
by hwang4
Hi, Konglt, thanks for reply. the ions were allowed to relaxed, actually I can get the distances by checking OUTCAR manually.

NEB result analysis problems

Posted: Wed Jun 14, 2006 12:17 pm
by konglt
Then the problem might be related with the scripts, please go to
http://theory.cm.utexas.edu/forum/viewf ... 77f587b31c
to ask for help, or you can check the script yourself and fix the problem.