Is it OK to use ewald-geometry = 3d for a crystal slab/water simulation?

GROMACS version: 2020.5
GROMACS modification: Yes/No

Hi All,

I am simulating a crystal slab with water and salt added above the slab. I will be using pbc=xyz. I was wondering if it is OK for me to use ewald-geometry=3d?

Additional details: Using semi-isotropic pressure coupling, PME electrostatics, and position restraints on the atoms in the crystal slab.

Thank you for your help and support.


Tej S. Choksi

Assistant Professor
School of Chemical and Biomedical Engineering
Nanyang Technological University, Singapore
Block N1.2, B1-18, 62 Nanyang Drive
Singapore 637459, Office: +65 6316 8940

Ballenegger, et al (2009). Journal of Chemical Physics, 131(9), 1–10. doi 10.1063/1.3216473
I.-C. Yeh and M. L. Berkowitz, J. Chem. Phys. 111, 3155 (1999),
doi 10.1063/1.479595

in the introduction of Ballenegger there is a good overview for the case of neutral systems. I seem to recall that there used to be some discussions regarding the net dipole correction term in the ewald summation, ew3d vs ew3dc, you better do some literature search there before starting any large scale simulations.

If you have solvent periodically connecting the slabs, 3d is the only option. If there is some vacuum in between and you want to simulate an isolated slab, you should use 3dc.