Structural relaxation with Coulomb interaction cutoff

Running Abinit Ground state
1

Hello everyone,

I was wondering if the Coulomb interaction cutoff (icutcoul) can bu used while doing a relaxation (optcell and ionmov). From the tests I have conducted, it doesn’t seem to be the case. In the example I’m sharing (graphene), the lattice lengths keep increasing until dilatmx is exceeded.

lattice_lengths: [   4.67399,    4.67399,   20.00000, ]
lattice_lengths: [   1.89469,    1.89469,   20.00000, ]
lattice_lengths: [   6.37547,    6.37547,   20.00000, ]
lattice_lengths: [   7.47289,    7.47289,   20.00000, ]
lattice_lengths: [  10.49086,   10.49086,   20.00000, ]

mono_graphene.abi (1011 Bytes)
mono_graphene.abo (41.9 KB)
Does anyone have an idea of what’s going on? Could it be due to the fact that the stresses are inaccurate when using icutcoul?

Thank you in advcance,

Tony Vuu

1 Like
2

Hi Tony,

icutcoul is valid only for energy and forces in the ground-state calculations. The stress correction with respect to the cut-off is not implemented, thus the erratic response.
Bogdan

1 Like
3

In that case, is the best course of action to first relax the system without the cutoff, then run a GS calculation on the relaxed system with the cutoff?