[CP2K-user] how to constraint the H2O molecule using the G3X3

Travis polla... at gmail.com
Wed May 8 16:35:28 UTC 2019

As I mentioned previously, CP2K stumbles over your close contact. You would 
have to manually edit the generated PSF file to correct the issues. 
Alternatively, you can open your structure in a molecular editor and just 
drag the offending water away from the other water OR recreate your 
structure in Packmol with a cell that's about 1.2 - 1.5x the size you 
expect, then let it relax during MD with your force field. This is a very 
common strategy to cope with Packmol's tendency to generate close contacts 
between peripheral atoms, even with the distance constraints.


On Tuesday, May 7, 2019 at 4:10:05 AM UTC-3, ni... at gmail.com wrote:
> Hi Travis,
> Thanks for your reply. Indeed, I have attempted to generate the 
> connectivity file (***.psf) by the CP2K.  I can successfully run the test 
> input file (see the h2o-32-g3x3-mult.inp in tests). However, it seems to 
> have some bugs (see attachment) in the CP2K when I modified the 
> coordination values according to my case (other parameters is same as those 
> in h2o-32-g3x3-mult.inp). Would you have any experiences about this?
> 在 2019年5月7日星期二 UTC+8上午1:04:59,Travis写道:
>> Append 'H2O' as the 5th column in the XYZ coordinates for waters. You can 
>> provide a connectivity file or allow CP2K to generate one, see 
>> https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/SUBSYS/TOPOLOGY.html#list_CONN_FILE_FORMAT. 
>> You can then modify this file to define 'H2O'. The automatic generation 
>> will probably mess up on your close contact, so be sure to double check 
>> that. The bit below will function to hold the bonds constant but it needs 
>> to know what 'H2O' is in your model. Change the distances to reflect the 
>> water model you're using.
>> <within CONSTRAINTS>
>>     &G3X3
>>       DISTANCES [angstrom] 1.0 [angstrom] 1.0 [angstrom] 1.63298
>>       MOLNAME H2O
>>       ATOMS 1 2 3
>>     &END G3X3
>> On Monday, May 6, 2019 at 12:58:22 PM UTC-3, ni... at gmail.com wrote:
>>> Hi everyone,
>>> I am running the MD using the FIST method for a cell including one CaF 
>>> and 61 H2O. To avoid the decomposition of water molecule, I want to 
>>> constraint the O-H bond of 61 water molecules using G3X3. However, it seems 
>>> that there has some bugs in the input file. I have attached the input file. 
>>> Would you please to have a look? Any points are welcome.
>>> Best,
>>> Ning
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190508/a51b5a6e/attachment.htm>

More information about the CP2K-user mailing list