Hydrogen terminated graphene in water using Airebo potential

Hi All

I have encountered a very strange phenomenon when I tried to simulate the hydrogen terminated graphene in water. The bond between carbon atoms and hydrogen atoms can not be formed and the temperature of water is hard to control. However, when I use the same coordinates of carbon and hydrogen without water molecules, I can get a perfect hydrogen terminated graphene sheet and a good control of temperature.
The in.data file for the hydrogen terminated graphene in water is in the follows,
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

#hydrogen terminated graphene with water
units metal
dimension 3
boundary p p p
atom_style full
read_data data.water
neighbor 2.0 bin
neigh_modify delay 5
kspace_style pppm 0.0001

pair_style hybrid airebo 3.0 1 1 lj/cut/coul/long 10.0
pair_coeff * * airebo /Potentials/CH.airebo C H NULL NULL
pair_coeff * 2 lj/cut/coul/long 0.0000 0.0000
pair_coeff * 3 lj/cut/coul/long 0.0000 0.0000
pair_coeff 2 * lj/cut/coul/long 0.0000 0.0000
pair_coeff 3 * lj/cut/coul/long 0.0000 0.0000
pair_coeff 1 4 lj/cut/coul/long 0.0087 3.1900
pair_coeff 4 4 lj/cut/coul/long 0.0067 3.1660

Your model seems odd. Pair hybrid only allows one pair

style per atom type pair. You’ve turned off the 1/2 airebo

with your second pair_coeff command and replaced

it with lj/cut/coul/long. Putting charges in an AIREBO

model and expecting both AIREBO and Coulombics to

give consistent answers together also seems non-standard.

Steve

pair_style hybrid airebo 3.0 1 1 lj/cut/coul/long 10.0
​​
pair_coeff * * airebo /Potentials/CH.airebo C H NULL NULL
pair_coeff * 2 lj/cut/coul/long 0.0000 0.0000
pair_coeff * 3 lj/cut/coul/long 0.0000 0.0000
pair_coeff 2 * lj/cut/coul/long 0.0000 0.0000
pair_coeff 3 * lj/cut/coul/long 0.0000 0.0000
pair_coeff 1 4 lj/cut/coul/long 0.0087 3.1900
pair_coeff 4 4 lj/cut/coul/long 0.0067 3.1660

Your model seems odd. Pair hybrid only allows one pair
style per atom type pair. You've turned off the 1/2 airebo
with your second pair_coeff command and replaced
it with lj/cut/coul/long. Putting charges in an AIREBO
model and expecting both AIREBO and Coulombics to
give consistent answers together also seems non-standard.

​"non-standard" in this case is a friendly euphemism for telling you that
MD follows the "GI-GO"-principle (GI-GO = garbage in, garbage out).
e.g. overwriting the 1-2 interactions with parameters that essentially wipe
them out, fully explains why your hydrogens are not bound to carbons.

what you need are the following:

# this covers 1-1, 1-2, and 2-2

pair_coeff * * airebo /Potentials/CH.airebo C H NULL NULL
# these are still missing
pair_coeff 1 3 ...
pair_coeff 2 3 ...
pair_coeff 3 3 ...
pair_coeff 1 4 ...
pair_coeff 2 4 ...
pair_coeff 3 4 ...
pair_coeff 4 4 ...

in case you want to have the water hydrogens (type 3?) have no interactions
with anything - as is implied by your input, you can simplify this to:

pair_coeff * 3 0.0 0.0
pair_coeff 1 4 ...
pair_coeff 2 4 ...
pair_coeff 4 4 ...

​axel.​