liquid between sheared walls

Dear lammps users,

I'm doing simulations of some tribological systems. For doing this I need to use systems which consist of liquid confined between two shearing walls. However, I have some questions about this kind of simulations.

1. How do I impose shear on the walls? If I use fix move with fix noforce then I can not update the positions of atoms in the walls which is principal for my problem. If I switch on the forces then I don't have constant shear velocity which I need. The option fix deform changes the whole box while I need just to move the walls. Is it possible to do something like this with sllod?

2. In some rather old mails I found that coulombic interaction is not implemented with fix sllod. Is it still the case or coulombic forces are already included in it?

Great thanks for any suggestions!

Yury

Comments below.

Steve

Dear lammps users,

I'm doing simulations of some tribological systems. For doing this I need to use systems which consist of liquid confined between two shearing walls. However, I have some questions about this kind of simulations.

1. How do I impose shear on the walls? If I use fix move with fix noforce then I can not update the positions of atoms in the walls which is principal for my problem. If I switch on the forces then I don't have constant shear velocity which I need. The option fix deform changes the whole box while I need just to move the walls. Is it possible to do something like this with sllod?

Fix move should work by itself. Or you can assign them a velocity,
use fix setforce 0.0, and
integrate them with fix nve. If you have expliciit atomistic walls,
you don't want fix deform
or sllod.

2. In some rather old mails I found that coulombic interaction is not implemented with fix sllod. Is it still the case or coulombic forces are already included in it?

I think you mean that long-range Coulomics don't work with a triclinic
box, e.g. fix deform,
which is correct.

Dear Steve,

Thank you very much for your reply! Just to make sure that I understand you correctly: fix move displaces a group of particles. At the same time I can integrate them in usual nve or nvt ensemble. Like this I can have an interacting shearing wall. Is it correct?

Thanks once more!

Yury Fomin

fix move takes the place of an integrator for that subsett
of atoms - only use fix nve on other freely-mobile atoms.

Steve

Thank you for your reply, Steve. However, may be I was not very clear with my question. Let me to describe it in a bit more details.

The story is that in some cases friction between the liquid and the substrate induces changes of the substrate structure which alters the friction coefficient. Therefore for correct computation of tribological properties it is necessary to integrate the substrate particles in usual MD and at the same time to push them for creating the shear. And the question is if I can do it simultaneously. Using fix move switches of the nve or nvt integrator, i.e. I have my wall moving but without changes in it's internal structure. Is this kind of "double" move of particles implemented in lammps?

Thanks for your help!

Yury Fomin

Dear lammps users !

Is there an easy way of calculating e.g. minimal interatomic
distance during runs ? I did not find a compute for this.
This could be crucial for e.g. cluster impact simulations,
when it is useful to check whether there are atoms too
close to each other or not.
Manual did not help me.

Best regard, Péter Sule

Hi

caculate the rdf: http://lammps.sandia.gov/doc/compute_rdf.html

Christian

-------- Original-Nachricht --------

You can integrate the boundary atoms with fix nve. You have a lot of
control over
what the force and velocity of those atoms are. You can add various constraint
forces for example, e.g. fix addforce.

Steve

If you mean you want to monitor the closest that any 2 atoms
get, then you can use compute reduce on a local vector
that stores the pairwise distance of all atom pairs via
compute pair/local. This is not a cheap computation as
it has to compute the pairwise distance between all
atoms in the neighbor list.

Steve

It is not possible to impose a velocity to and thermalize the same
atoms, but your can decompose each wall in two layers. The layer in
contact with the liquid can be thermostated, and the outer layer can
be moved with a constant velocity. Then you just have to ensure the
cohesion between the two layers, using either pair interactions or
bonds.

Best,
Laurent

1 Like

Thanks a lot! This is what I needed!

Best regards,
Yury Fomin

01.08.2011, 20:14, "Laurent Joly" <[email protected]...>: