pair mixing soft-core potentials into fep

LAMMPS LAMMPS Mailing List Mirror
1

Dear lammps users,

I wanted to ask, mostly to the developer of the soft-core version of the potentials. On the documentation is said that soft potentials supports mixing with pair modify, however then is not possible to use the fix/adapt fep or compute fep regarding the mixed interaction right? The interaction coefficient must be explicetely imposed in the script?

for instance

pair_style hybrid lj/cut/coul/long/soft 1 0.5 10 15.0 10.0 etc

pair_coeff 1 1 lj/cut/coul/long/soft 0.5 3 1.0
pair_coeff 2 2 lj/cut/coul/long/soft 0.6 3.1 1.0

pair_modify pair lj/cut/coul/long/soft mix geometric
…

fix adapt all adapt fep {Nadapt} pair lj/cut/coul/long/soft lambda 1 2 v_lambda compute FEP all fep {Temp} pair lj/cut/coul/long/soft lambda 1 2 v_lambda

It will not act since pair_coeff 1 2 is not explicitely reported? If someone knows this for sure i would like to have confirmation

Kind regards,
Lorenzo Chiavarini

2

Hi,
In order to adapt the parameters for the unlike interactions you need to supply all pairs explicitly and not use pair_modify mix.

Cheers,
Agilio

3

Yeah, my problem was that I have a complex organic molecule which interacts with water, therefore I generate the coeff from Opslaa and i have 820 pair coeffs, so i can’t write them explicitly.
I also tried to use hybrid/overlay and define more pair styles like:

pair_style hybrid/overlay lj/cut …… lj/cut/coul/long …… lj/cut/coul/long/soft

pair_coeff 1 1 lj/cut 0.0 1.0

…

Pair_coeff 820 820 lj/cut 0.0 1.0

And then define the one I’m interested by hand while using:

pair_modify lj/cut mix arithmetic

but it says that not all pair coeffs are set. I do not understand why it says so since the mixing should work only on that pair_style also no?

Kind regards,

Lorenzo

4

Here is what i wanted to do but seems not to be allowed by pair_modify…

in_singlelayer_mixing2.gro2lam (35 KB)

5

Yeah, my problem was that I have a complex organic molecule which interacts with water, therefore I generate the coeff from Opslaa and i have 820 pair coeffs, so i can’t write them explicitly.

Why not? A script can go through the like pairs and generate the unlike parameters. This can be put in its own file and then read into your input file using the include command.

I find impressive that you have 820 atom types in terms of the pair potential. Partial charges do vary a lot for the same element in different molecules/chemical environments but the LJ parameters vary much less.

I also tried to use hybrid/overlay and define more pair styles like:
pair_style hybrid/overlay lj/cut ….. lj/cut/coul/long ….. lj/cut/coul/long/soft

pair_coeff 1 1 lj/cut 0.0 1.0
..
Pair_coeff 820 820 lj/cut 0.0 1.0

And then define the one I’m interested by hand while using:
pair_modify lj/cut mix arithmetic

Mixing mix :slight_smile: with overlay has its peculiarities. See the doc page of pair_style hybrid/overlay.

Agilio