hybrid/overlay versus single combined pair potential

Hello LAMMPS users:

I am trying to define pair style and parameters for my polymer model. One way I did is

pair_style lj/class2/coul/long 12.0 12.0
pair_coeff 1 1 0.0640000000 4.0100000000
pair_coeff 2 2 0.0540000000 4.0100000000
pair_coeff 3 3 0.0540000000 4.0100000000
pair_coeff 4 4 0.0200000000 2.9950000000
pair_coeff 5 5 0.2400000000 3.5350000000
pair_coeff 6 6 0.0540000000 4.0100000000
pair_coeff 7 7 0.0540000000 4.0100000000
pair_coeff 8 8 0.2400000000 3.5350000000
pair_coeff 9 9 0.0130000000 1.0980000000

Using pair_style hybrid/overlay, I thought the following way would be equivalent:

pair_style hybrid/overlay lj/class2 12.0 coul/long 12.0
pair_coeff 1 1 lj/class2 0.0640000000 4.0100000000
pair_coeff 2 2 lj/class2 0.0540000000 4.0100000000
pair_coeff 3 3 lj/class2 0.0540000000 4.0100000000
pair_coeff 4 4 lj/class2 0.0200000000 2.9950000000
pair_coeff 5 5 lj/class2 0.2400000000 3.5350000000
pair_coeff 6 6 lj/class2 0.0540000000 4.0100000000
pair_coeff 7 7 lj/class2 0.0540000000 4.0100000000
pair_coeff 8 8 lj/class2 0.2400000000 3.5350000000
pair_coeff 9 9 lj/class2 0.0130000000 1.0980000000
pair_coeff * * coul/long

Both shared the same kspace style (ewald\n) and mixing rule (6th power). The two method gave quite different coul energy and vdwl energy. Can someone kindly point out what mistake I made? Thanks very much.

Actually I was going to use coul/wolf and the only way I see to use it together with lj/class2 is through pair_style hybrid/overlay. But pair_style hybrid/overlay lj/class2…coul/wolf gave wrong energy. Thus I want to test pair_style hybrid/overlay lj/class2…coul/long first.

Shaorui

Hello LAMMPS users:

I am trying to define pair style and parameters for my polymer model. One
way I did is

pair_style lj/class2/coul/long 12.0 12.0
pair_coeff 1 1 0.0640000000 4.0100000000
pair_coeff 2 2 0.0540000000 4.0100000000
pair_coeff 3 3 0.0540000000 4.0100000000
pair_coeff 4 4 0.0200000000 2.9950000000
pair_coeff 5 5 0.2400000000 3.5350000000
pair_coeff 6 6 0.0540000000 4.0100000000
pair_coeff 7 7 0.0540000000 4.0100000000
pair_coeff 8 8 0.2400000000 3.5350000000
pair_coeff 9 9 0.0130000000 1.0980000000

Using pair_style hybrid/overlay, I thought the following way would be
equivalent:

pair_style hybrid/overlay lj/class2 12.0 coul/long 12.0
pair_coeff 1 1 lj/class2 0.0640000000 4.0100000000
pair_coeff 2 2 lj/class2 0.0540000000 4.0100000000
pair_coeff 3 3 lj/class2 0.0540000000 4.0100000000
pair_coeff 4 4 lj/class2 0.0200000000 2.9950000000
pair_coeff 5 5 lj/class2 0.2400000000 3.5350000000
pair_coeff 6 6 lj/class2 0.0540000000 4.0100000000
pair_coeff 7 7 lj/class2 0.0540000000 4.0100000000
pair_coeff 8 8 lj/class2 0.2400000000 3.5350000000
pair_coeff 9 9 lj/class2 0.0130000000 1.0980000000
pair_coeff * * coul/long

Both shared the same kspace style (ewald\n) and mixing rule (6th power). The
two method gave quite different coul energy and vdwl energy. Can someone
kindly point out what mistake I made? Thanks very much.

i don't think that mixing would work with hybrid/overlay
in your case. you'll have to specify all mixed pair interactions.

axel.

Hi Axel:

Is it possible to generate a separate file containing all the pair parameters and let my input script read this file? The number is bit too large here…

Thanks

Shaorui

Hi Axel:

Is it possible to generate a separate file containing all the pair parameters and let my input script read this file? The number is bit too large here…

Yes. Check out the ‘include’ command.
We do this routinely for our coarse grain model.

Axel