[CP2K-user] VIBRATIONAL_ANALYSIS of a single adsorbed molecule on a surface: What is the best computational procedure?

[Stefano Ferrero]

Dear CP2K experts,

I have just started using CP2K and I would be really thankful if anybody 
can help me.
My question is on calculating frequencies of a molecule (as CO) adsorbed on 
a surface (an ice surface in my case).
I will work with energy differences so, there is no need of a vibrational 
analysis of the entire system but just of a fragment of it.
I have to zoom in on the frequencies of the adsorbed molecule and the most 
interacting water which will be the most perturbed modes with respect to 
gas phase frequecies
Looking at other threads,I found that CP2K has the keyword MODE_SELECTIVE 
with which I can zoom in to specific vibrational modes. 

1) How does it do that? Is it displacing and calculating forces just for 
the atom on which I ask to zoom in on?
    Can I use the keyword THERMOCHEMISTRY with the MODE_SELECTIVE one in 
order to obtain the ZPE and thermal correction for just the modes I zoomed 
in on?

2) Would it be better to fix (with CONSTRAINT + FIXED_ATOMS) all atoms for 
which I do not want calculate the normal modes, in order to speed up the 
calculation?
    Are there any problems with this "partial Hessian approach"?
    Would it be possible to use the THERMOCHEMISTRY keyword in this case?

3) Which one will be the best way to calculate the frequencies of just the 
adsorbate plus the most interacting water molecules from the surface, 
keeping in mind that I have to make the thermochemical analysis on that 
little     
   fragment of the entire system?  

Thank you for your help!

Stefano Ferrero
  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20191211/457ca9ce/attachment.htm>