Why I get a 5000 cm-1 vibration frequency for single H2 molecule?

[Florian Schiffmann]

The idea is, that there is translational and rotational invariance 
(frequency of 0 cm-1). That means you only have 3N-6 degrees of freedom. In 
a periodic material, that is not true, as a rotation of your unit cell 
clearly has an effect on the total energy (kink in pbc). However, with the 
defaults cp2k performs the same removal in a periodic system, but here the 
removed modes will not correspond to a CO2 rotation but your whole crystal, 
leaving the CO2 rotations mainly unaffected.
As always with cp2k, it doesn't remove these modes automatically but a 
keyword can activate/ deactivate this option. 
The keyword in question is fully_periodic, which is not obvious in the 
first place but in combination with the points above makes perfectly sense. 
The default here is false, which means these modes are removed by default. 
If you are interested in these modes for any reason or (as the keyword 
states), run a periodic calculation you should set it to true.

Flo

Am Donnerstag, 4. Dezember 2014 23:06:06 UTC+1 schrieb Jingyun Ye:
>
> Hi, Flo
>
> Thanks very much. I checked the data in cccbdb. It really depends on the 
> basis set. I have a another question and hope you have some idea. Why did 
> only get 3 vibration modes for CO2 ? The degenerate modes around 600 cm-1 
> should have two but only shows one in my output. And I still checked the 
> cccbdb data base. There are 3 frequency modes for CO2 molecule shown in the 
> table. And in the full list, there indeed has 4 modes. I also got 3 modes 
> for CO2 before when I used VASP.  
>
> Another question:  In the .mol  file of single CO2 frequency calculation, 
> I only got 3 modes which no rotation or translation modes. But when I put 
> the CO2 in porous material to study the adsorption property, the .mol file 
> has the rotation and translation modes for CO2. I just want to know how 
> does this happen? I remember Alex mentioned in previous posts that the CP2K 
> output will automatically purify the R and T modes. If this is true, how 
> does the R and T modes shows in the outputs for CO2 in porous material but 
> don't show for single CO2 molecule.  (Note: all the input files are same 
> with my post one.)
>
> Thanks very much for your help!
>
>
>
> Thanks again!
>
>
> On Thursday, December 4, 2014 5:31:37 AM UTC-5, Florian Schiffmann wrote:
>>
>> Hi,
>>
>> looking at cccbdb, you can see that using a bad basis can have a 
>> miserable effect. In your case a DZVP-SR basis has the same effect, remove 
>> if you use the normal DZVP optimize and run the analysis you will get a 
>> pretty nice frequency
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20141205/9fd472f6/attachment.htm>