Minimum Image Convention for COLVAR?

[Daniel Depew]

Hello CP2K Community,

I'm doing some exploratory investigation of collective variables from an unbiased MD trajectory. Specifically when testing the conditioned distance variable, I was seeing larger distances than I expected.

I've scripted up my own implementations of the collective variables since I have an existing postprocessing codebase, and I noticed that the conditioned distance variable doesn't seem to apply minimum image convention to the atom positions. However, I match the CP2K output if I wrap the atoms back into the simulation box.

On the one hand, it makes sense to use atom positions inside the periodic domain in general, but for distance-based collective variables, I expected the minimum image to be used since that governs the strongest interaction.

For what it's worth, I noticed that some other distance functions have a PBC option, so I imagine the default is simply true for CONDITIONED_DISTANCE. Is there any way to observe the minimum image convention, or will I have to resort to PLUMED if I need this capability?

Thanks,

Daniel

[Jürg Hutter]

Hi

I had a quick look at the code that should be used for those colvars and I find that all
distances are indeed calculated using MIC. As I'm not an expert in this part of the code
I might be wrong. Can you post the exact colvar section that you are using?

regards
JH

________________________________________
From: cp...@googlegroups.com <cp...@googlegroups.com> on behalf of Daniel Depew <depew....@gmail.com>
Sent: Wednesday, February 7, 2024 7:29 AM
To: cp2k
Subject: [CP2K:19884] Minimum Image Convention for COLVAR?

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com<mailto:cp2k+uns...@googlegroups.com>.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/3b45651e-c85a-4bf5-8d38-078efe22c7b4n%40googlegroups.com<https://groups.google.com/d/msgid/cp2k/3b45651e-c85a-4bf5-8d38-078efe22c7b4n%40googlegroups.com?utm_medium=email&utm_source=footer>.

[Marcella Iannuzzi]

Hi 

I agree with Juerg, 

I also checked the code and PBC are always applied when distances are calculated.

Regards

Marcella