Setting up Qeq for our pair style

Ray,

Thank you for your help. If anyone else has any suggestions I would be grateful.

thanks much
matthew

I imagine in relation to Q3, i.e. reciprocal space contributions, Ray is referring to efficient numerical implementation of the method into lammps, right? I have previously coded SQE models with PBCs on and as far as I remember there are no theoretical limitations. Ray is also correct in stressing that such contributions are generally small with the proper choice of the cutoff, however, in highly symmetric systems without defects reciprocal space terms tend to carry more weight than in your average everyday system.
Carlos

This is a followup on the previous question. What I decided to try to do is create a new pair syle based on coul/cut, which I am calling coul/ring. It basically only counts coulombic contributions between two radial distances, the larger of which is the global cut off from before. I then used pair style hybrid/overlay to combine the two pair styles so that VMM would contribute to the energy within some cutoff and coulombic terms outside it. The input is as follows:

primitive vmm test

units real

atom_style charge

neighbor 0.3 bin

read_data sio.data

fix 1 all qeq/point 1 10 1.0e-6 200 param.qeq1

pair_style hybrid/overlay coul/ring 3.0 10.0 vmm

pair_coeff * * vmm sio.vmm Si O

fix 2 all nve

thermo 0

run 0

I get the following error when I try to run:

ERROR: Pair hybrid sub-style is not used (…/pair_hybrid.cpp:421)

What am I missing?

Matthew

This is a followup on the previous question. What I decided to try to do
is create a new pair syle based on coul/cut, which I am calling coul/ring.
It basically only counts coulombic contributions between two radial
distances, the larger of which is the global cut off from before. I then
used pair style hybrid/overlay to combine the two pair styles so that VMM
would contribute to the energy within some cutoff and coulombic terms
outside it. The input is as follows:

# primitive vmm test

units real

atom_style charge

neighbor 0.3 bin

read_data sio.data

fix 1 all qeq/point 1 10 1.0e-6 200 param.qeq1

pair_style hybrid/overlay coul/ring 3.0 10.0 vmm

pair_coeff * * vmm sio.vmm Si O

fix 2 all nve

thermo 0

run 0

I get the following error when I try to run:

ERROR: Pair hybrid sub-style is not used (../pair_hybrid.cpp:421)

What am I missing?

​there is no pair_coeff statement for coul/ring.
without it hybrid/overlay doesn't know to which pairs of atoms to apply the
style to.

axel.​

Hi Matthew/Ray:
We’ve actually been working on something quite similar (as seen by my earlier email).

Re: long range contributions to electrostatics from qeq

I might be missing something here, but isn’t this term already handled by the kspace solver? Since the qeq coulomb term reduces to the regular 1/r form at long distances, the only correction necessary would be for the real space part of the ewald [i.e. qi*qi*Jab] in the case where Jab is the overlap of slater orbitals, no?

–Mike

Ray,

Thank you for your help. If anyone else has any suggestions I would be grateful.

thanks much
matthew