FW: Neigh_modify exclude group problem

Hi Axel,

I have tried the pair_style hybrid option which you mentioned.
These are the commands I used:

I got this idea from the manual (the section on pair_style hybrid gives a very similar example to this)
My irradiation atoms are originally assigned as type 4.

pair_style hybrid airebo 0 0 0
pair_coeff * * airebo CH.airebo C C C NULL
pair_coeff 1 4 none

Seems a very clean way of doing it since it doesnt require another potential to be defined.
But in the log file just after the first run, I get this error:

timestep 0.0000625
run 50000
ERROR: All pair coeffs are not set

As I have followed the example in the manual I’m not sure whats wrong.

Daniel

Hi Axel,

I have tried the pair_style hybrid option which you mentioned.
These are the commands I used:

I got this idea from the manual (the section on pair_style hybrid gives a
very similar example to this)
My irradiation atoms are originally assigned as type 4.

pair_style hybrid airebo 0 0 0
pair_coeff * * airebo CH.airebo C C C NULL
pair_coeff 1 4 none

Seems a very clean way of doing it since it doesnt require another potential
to be defined.
But in the log file just after the first run, I get this error:

timestep 0.0000625
run 50000
ERROR: All pair coeffs are not set

As I have followed the example in the manual I'm not sure whats wrong.

you may have followed the example, but you have not properly read the
description.
the issue is what the error says. where are the coefficients for 2-4,
3-4, and 4-4?

axel.

Thanks Axel,

I don't see that I need to define further coefficients.
By saying "pair_coeff * * "
I have defined the coeffs for all type combinations including 2-4,
3-4, and 4-4.

Here is the bit from the manual:

Thanks Axel,

I don't see that I need to define further coefficients.
By saying "pair_coeff * * "
I have defined the coeffs for all type combinations including 2-4,
3-4, and 4-4.

no. not if you skip atom type 4 through using the NULL assignment that
leaves all coefficients involving atom type 4 unset. keep in mind that
manybody potentials set parameters on a per atom-type basis.

the section that you are looking at is describing something different
from what you want to do.

[....]

I think the problem is with the fact that I use NULL for the fourth atom
type:
"pair_coeff * * airebo CH.airebo C C C NULL"

This means that I guess (4,4) has no potential defined and likewise for
(2,4) and (3,4)

yes.

So I have now changed it to:
"pair_coeff * * airebo CH.airebo C C C C"
pair_coeff 1 4 none

Do you agree or am I misreading things?

i disagree. this way you introduce interactions that you don't want, e.g. 4-4

you need to use this instead.

pair_coeff * * airebo CH.airebo C C C NULL
pair_coeff * 4 none

axel.

Very good....I think you are right!!
I don't want (4,4) interactions.
I will see if this works!

Thanks a lot