Page 1 of 1

newcomer: DIMER METHOD WITH VTST

Posted: Wed Oct 01, 2014 10:13 am
by perez
hello

I have been looking at the dimer-method example provided with the VTST distribution
and, as a newcomer, I would like to understand one point that is not clear

please let me know why when you grep the total energy inside OUTCAR, you can see that
it going down, and the final energy is the lowest of all

In my undestanding, with the dimer method we are computing a maximum in total energy,
therefore, the final (dimer-converged) energy should be the highest of all (less negative)

thank you for your understanding

regards

Manuel

Re: newcomer: DIMER METHOD WITH VTST

Posted: Wed Oct 01, 2014 1:01 pm
by graeme
The total energy will not be a local maximum or minimum with the dimer method since it is converging to a saddle point. Instead, it is the force that should systematically drop as the dimer converges to the saddle.