[ase-users] Thermochemistry

[Eric Hermes]

On Fri, 2017-12-08 at 14:53 +0300, Alexander Romanov via ase-users
wrote:
> Hello,
> 
> I investigate diffusion of atoms on metal surfaces and want to use
> thermochemistry module for energy estimation of transition states.
> What can I do with imaginary frequencies?  If I understand properly,
> Gaussian simply subtract them from calculation, but ase module can't
> do this automatically.  Can I do the same manually? 

Which frequencies you pass to the Thermo classes is up to you. You are
free to exclude any imaginary frequencies returned by the Vibrations
class from the list you pass to the Thermo class if you want.

However, be aware of what you are doing by excluding certain
frequencies from the Thermo classes. You need to ensure that you are
considering all 3N degrees of freedom of your system.

For an ideal gas, 5 or 6 of those are translations/rotations, so 5 or 6
of your vibrational frequencies should be small and potentially
imaginary (since you will capture the effects of those modes with a
rigid rotor/particle in a box model separately).

In the context of transition state theory, you actually are removing a
degree of freedom from your system, so the imaginary mode of a
transition state corresponding to the reaction coordinate should be
removed from the thermodynamic analysis.

In most other cases, simply removing a vibrational frequency will
result in significant errors in the system entropy, which is very
important if you are trying to compare the free energy of two system
configurations.

Eric

> 
> Best regards,
> Alex
> _______________________________________________
> ase-users mailing list
> ase-users at listserv.fysik.dtu.dk
> https://listserv.fysik.dtu.dk/mailman/listinfo/ase-users