Spurious frequencies with ZORA/ADF solved

Just a quick note for those using ADF and ZORA, especially for optimizations
and transition state searches. There is small mismatch between the potential
and energy expressions, and always has been with ZORA.

In previous versions, this could lead to differences of 0.01 Å for bonds;
this is now improved in the 2010.01 version, where it is stated that it is
reduced to 0.0001 Å. In itself for chemical purposes this is irrelevant,
BUT it can become highly relevant when trying to calculate frequencies,
where because of this mismatch spurious imaginary frequencies may
occur, which were sometimes impossible to get rid of.

In the 2010.01 version, this seems to be solved; the problem is in the optimization,
not in the frequencies. So, simply redoing the frequency calculation does not help,
you should redo the optimization. I have noticed from a project where we had
one such annoying system with spurious imaginary frequency (of around -30 cm-1).
Using the 2010.01 version, and reoptimizing did the trick.


Popular Posts