Request for support: Non-periodic boundary condition and dipole correction implementation

[transparent-cicada]

Dear Developers,

Firstly, thank you for developing such a powerful toolkit that combines artificial intelligence with ab initio methods for computational chemistry research. It has been an invaluable resource for our work.

While utilizing the toolkit in my research, I encountered a limitation: the lack of support for non-periodic boundary conditions in the z-direction. My system requires the removal of periodicity along the z-direction and the addition of dipole correction to properly model its physical properties. However, the current setup does not support specifying partial periodicity. For example, when using CP2K, the CELL section is automatically generated, and it is not straightforward to add the PERIODIC XY directive.

Non-periodic systems are generally in computational chemistry. They are particularly critical in studies of electrocatalysis, aqueous batteries, and other systems involving interfaces or electric fields. The absence of this feature limits the toolkit's application to such fields.

To clarify, my specific requirements are:

1. The option to remove periodicity along the z-direction for systems modeled with slab geometries or similar setups.
2. The implementation of dipole correction to account for artifacts arising from the remaining periodic boundaries.

I believe adding this capability would significantly enhance the flexibility and applicability of this toolkit. Many researchers in the field would benefit from these functionalities, particularly those investigating interfacial phenomena and other non-periodic systems.

Please let me know if there are any technical details I can provide to assist in evaluating or implementing this feature. Thank you for considering this enhancement, and I look forward to your feedback.

Best regards

[ChiahsinChu]

Hi,

Could you please specify which part of the codes you are working with? In the CLL MLP training workflow with cp2k labelling, you should be able to have your own setup, i.e., turning on SURFACE_DIPOLE_CORRECTION or modifying the POISSON section.
If it does not work, could you please provide a MINIMAL example to reproduce your issue?

Best