colloid force calculation

Hi

I am working with a system of small granular particles, but I want to include the van der Waals and electrostatic forces to the system as well. This means using pair/hybrid and adding the “colloid” and “yukawa/colloid” pair styles. However, I don’t agree with the way the forces are calculated in those two pair styles. Can anyone please clarify these pair style forces for me? I am using LAMMPS version dated: 10Mar10.

Firstly, I feel both forces are of the wrong sign (-ve/+ve sign). Secondly, the forces have been divided by the separation distance between particles. This gives the “forces” units of energy/distance^2 and not the correct units of energy/distance. Why was this done? Where these forces made specifically for “LJ” unit systems? I work in the “si” or “cgs” units systems.

Also, I have not seen the “Ur” potential (repulsive component of the colloid-colloid potential in the “colloid” pair style) mentioned anywhere in literature other than the Everaers papers referenced in the documentation. What is the purpose of this component? Why don’t other sources consider it?

Any insight would be appreciated.

Regards
Evan

Hi

I am working with a system of small granular particles, but I want to
include the van der Waals and electrostatic forces to the system as well.
This means using pair/hybrid and adding the "colloid" and "yukawa/colloid"
pair styles. However, I don't agree with the way the forces are calculated
in those two pair styles. Can anyone please clarify these pair style forces
for me? I am using LAMMPS version dated: 10Mar10.

please consider updating to the current version.
you will benefit from bugfixes and performance
improvements.

Firstly, I feel both forces are of the wrong sign (-ve/+ve sign). Secondly,
the forces have been divided by the separation distance between particles.
This gives the "forces" units of energy/distance^2 and not the correct units
of energy/distance. Why was this done? Where these forces made specifically

please have a closer look at the way the force calculation is done
and note that fpair is multiplied by delx/dely/delz to get the force
contributions in x/y/z and thus gets you the right units. don't you
think people would have noticed, if force calculations were so
incorrect in such a bad way?

for "LJ" unit systems? I work in the "si" or "cgs" units systems.

no. force kernels are independent of the choice of units.
the only cases where the units are required are when the
potential parameters are read from a file.

Also, I have not seen the "Ur" potential (repulsive component of the
colloid-colloid potential in the "colloid" pair style) mentioned anywhere in
literature other than the Everaers papers referenced in the documentation.
What is the purpose of this component? Why don't other sources consider it?

sorry, i don't know much about colloid potentials and thus cannot help
you there.

axel.

Evan,

Sorry about the delay in answering your question. Everaers potential
is not the same that you will see that has been included in DLVO type
potentials. Rather the repulsive component, I believe, is simply
there to enforce the pauli principal, i.e. fermions and regular matter
cannot occupy the same physical space. Note, that this will also
prevent granular type collisions, which allow for deformation and a
coefficient of restitution. You will have to modify the lammps code
in order to allow for granular simulations, I suggest looking at the
code in pair_colloid.cpp and seeing what easy changes you can make.
Also make sure to test your modified code on binary collisions, where
you can predict the results analytically.

Evan,

Sorry about the delay in answering your question. Everaers potential
is not the same that you will see that has been included in DLVO type
potentials. Rather the repulsive component, I believe, is simply
there to enforce the pauli principal, i.e. fermions and regular matter
cannot occupy the same physical space. Note, that this will also
prevent granular type collisions, which allow for deformation and a
coefficient of restitution. You will have to modify the lammps code
in order to allow for granular simulations, I suggest looking at the
code in pair_colloid.cpp and seeing what easy changes you can make.
Also make sure to test your modified code on binary collisions, where
you can predict the results analytically.

Our group at Sandia has used these colloidal and yukawa
potentials extensively. I don't believe the issues
you list are bugs, i.e. the sign of force, units of force
are all correct so far as I know and compatible
with the formulas listed on the doc pages.

You should use the most current LAMMPS version
as Axel suggested.

Steve