Problems of setting collective variables in constrained AIMD simulation

[Bai Li]

Dear CP2K Users,

I am attempting to perform constrained AIMD simulations using CP2K to study the solvation process of a diatomic molecule (X2) transitioning from the gas phase to a solvent model represented by a slab water model. In the constrained AIMD simulation, I define a collective variable (CV) as the distance between the geometry center of X2 and the geometry center of all water molecules in the water slab. I fixed this CV and performed dynamics simulations at 300 K. 

However, in the final simulation results (taking CV = 9.0 Å as an example), some water molecules escape beyond the cell boundaries defined in my input file. To address this, I applied a wrapping operation to the atomic/molecular positions in the output pos.xyz file, bringing the out-of-cell atoms/molecules back into the cell using periodic boundary conditions. Then, I recalculated the CV for each image in the trajectory. Surprisingly, after wrapping all atoms/molecules periodically, the CV is no longer the initial value (9.0 Å), but instead takes on different values. However, if I skip the wrapping operation, the CV remains at the initial value of 9.0 Å. It seems that the CV definition does not account for periodicity during the simulation. 

I would like to ask whether CP2K has this issue, or if there might be other reasons causing this problem in my simulation results.

If someone have time to give me some advice, I would appreciate it.

Best regards