Page 1 of 1

Cl-NEB problem

Posted: Thu Jul 19, 2012 10:04 pm
by panli0202
I ran a Cl-NEB to find a TS. It converged but there is no imaginary vibrational mode for my "TS". How to solve this problem?

One more thing,shouldnt climbing NEB guarantee that the TS is a true tranistion state since it is pushing the highest energy image UPHILL until the forces are zero?
Thanks.

Re: Cl-NEB problem

Posted: Fri Jul 20, 2012 9:31 am
by graeme
Yes, the climbing image should converge to a first order saddle.

There are some pathological cases (related to high symmetry critical points) where the NEB can pass through a higher order saddle. But these are rare cases and not to be generally concerned about.

Most likely, the problem is related to having good enough convergence of the saddle (to < 0.01 eV/Ang) or even more likely, that you have a converged dynamical matrix calculation. Make sure to have accurate forces, using ediff=1e-8 and check for convergence with displacements in the range of 0.001 to 0.01 Ang.