how to confirm a TS after the NEB run has converged ?

kay12 · Post by **kay12** » Fri Nov 03, 2006 2:12 pm

Dear administrators,

I just started using NEB to search for TS for a H2 dissociation onto a graphite sheet. I used a total of 8 images. The results look fine. For exmaple, in neb.dat, I have a clear energy maximum at image 7. Presumably, one would consider that as a TS, am I right? However, the geometry of the image 7 doesn't look like a TS: the H-H length too long. I did frequency calculations on image7 and other nearby images with the VASP4.6 code, and found no imaginary modes in all the images. My question is how to identify a TS, after a seemingly converged NEB run? Many thanks.

Kay
P.S. If the input files are needed for diagnosis, please instruct me how to upload those files.

andri · Post by **andri** » Fri Nov 03, 2006 3:40 pm

How does the mep.eps figure generated by the nebresults.pl script look like? Is the seventh images sitting at the top of the MEP? If so then it should be a saddle point. Are you using the climbing image NEB?

Also how did you calculate the normal modes for that image? Sometimes it can be quite tricky to get concrete values for the frequencies.

Post by **graeme** » Fri Nov 03, 2006 4:10 pm

It would be helpful to see your files. If you have access to a web server, you can post the link to them. I can also post them if you email them to me: henkelman at mail.utexas.edu

andri · Post by **andri** » Fri Nov 03, 2006 5:45 pm

Also note, if using our scripts, that the sign of frequency (or omega^2) is determined by the 0/1 tag. 0 means positive and 1 is negative.

kay12 · Post by **kay12** » Fri Nov 10, 2006 5:05 pm

thanks for the help. I realized I made a mistake for viewing the geometry. In frequency calculation, I ignored the i sign. Sorry for my silly quesiton.