AIMD; Periodic Boundary Condition and Minimum Image Convention

liu dongfei

unread,

Aug 30, 2021, 3:07:30 AM8/30/21

to cp2k

Dear cp2k users and experts,

I want to know how to set up some parameters so that the cell meets the Minimum Image Convention in the AIMD simulation.

I know the PBC (Periodic Boundary Condition) setting is in this directory: FORCE_EVAL / SUBSYS / CELL. To meet Minimum Image Convention, there are usually some parameters to set, such as cut-off distance, which is usually less than half the size of the box. However, I don't find any related instructions or parameters. Could anybody give me some suggestions?

Thanks in advance

Dongfei

Marcella Iannuzzi

unread,

Aug 30, 2021, 3:37:42 AM8/30/21

to cp2k

Dear Dongfei,

For QS there is no such cut-off distance, the range of the interaction fields is controlled by the EPS_* parameters.

Regards

Marcella

liu dongfei

unread,

Aug 30, 2021, 4:53:31 AM8/30/21

to cp2k

Dear Marcella,

Thanks a lot for your reply !

If there is no Minimum Image Convention, does that mean there is no limit to the size of the box? And I think when using PBC, there should be some restrictions on the box size. I don't know what the relevant settings are in CP2K.

In the CP2K reference manual, I found this keywords R_CUTOFF just now in this directory: CP2K_INPUT / FORCE_EVAL / DFT / XC / VDW_POTENTIAL / PAIR_POTENTIAL . I think this keyword is also linked to the interaction. So is there a connection between the R_CUTOFF and the size of the box? I'm looking forward to your answer.

Regards

Dongfei

Marcella Iannuzzi

unread,

Aug 30, 2021, 6:42:18 AM8/30/21

to cp2k

Dear Dongfei

You misunderstood my reply. I did not say that there is no limit to the interaction range.

The different terms of the Hamiltonian are computed by considering the interactions within ranges that are controlled by the EPS_* parameters.

If your cell is small and the EPS_* tight, multiple cells are used. With a large cell, most probably the interaction terms are going to be calculated for distances smaller than the cell dimension.

For the specific case of the pair potential used as vdw correction there is indeed a cutoff, but no connection to the box size.

Regards

Marcella

liu dongfei

unread,

Aug 30, 2021, 11:53:55 AM8/30/21

to cp2k

Dear Marcella,

Thanks a lot for your patient explanation. Following your suggestion, I found these EPS_* parameters in the CP2K reference manual. However, just through the instructions on the manual, I still don't know how to set up and use them reasonably. I also tried to find the literature about CP2K (such as doi: 10.1063/5.0007045; I was suprised to find that you are one of the authors, and I am very honored to be able to discuss problems with you and get your reply.) and look for relevant explanation but that didn't have much effect. Maybe I didn't find the right materials. So I really hope you can recommend some relevant materials for me.