How to set a component of the force which is not parallel to any axis?

Dear Lammps Developers and Users,

I’m doing a MD simulation about nanowire (see the attachment).

Periodic boundary condition is used along the axis of nanowire. There’s a slip plane which is inclined to the axis.

I want the nanowire to be relaxed just along the direction parallel to the slip plane. That’s to say, the vertical component of the atomic force should be zero (f⊥=0), and only parallel component f∥ exists in the force field.

How can I set the fix command?

Thank you.

H. Wei

NW.jpg

-- I'm not experienced in this kind of simulation, but I would look
at these commands:
http://lammps.sandia.gov/doc/fix_addforce.html
http://lammps.sandia.gov/doc/group.html
Divide the system into two groups, and apply a force in opposite
directions f∥ and -f∥ to each group. (Cool use of unicode, by the
way.)

-- You must insure that the slip-plane is parallel to one of the
boundary planes. In your example, your slip plane should be vertical
or horizontal. See attached figure.

By default you will have 2 slip planes (the boundary between the
groups moving f∥ and -f∥ directions, above and below). I don't think
this is a problem.

-- If you only want a single slip plane, you must disable boundary
conditions in the direction perpendicular to your slip plane
http://lammps.sandia.gov/doc/boundary.html

Cheers
Andrew

bumpy_slip_plane_pbc_problem.png

Dear Lammps Developers and Users,

I'm doing a MD simulation about nanowire (see the attachment).

Periodic boundary condition is used along the axis of nanowire. There's a
slip plane which is inclined to the axis.

The addforce command allows you to choose any direction for the force.
http://lammps.sandia.gov/doc/fix_addforce.html
http://lammps.sandia.gov/doc/group.html

-- You must insure that the slip-plane is parallel to one of the
boundary planes. In your example, your slip plane should be vertical
or horizontal. See attached figure.

I should have been more careful in my reply.
I said this because I assumed you wanted periodic boundary conditions
in all directions. If not, you can use the boundary command to turn
off periodicity in the x and y directions using:
boundary f f p
http://lammps.sandia.gov/doc/boundary.html

Cheers
Andrew

I should have been more careful in my reply.
I said this because I assumed you wanted periodic boundary conditions
in all directions. If not, you can use the boundary command to turn
off periodicity in the x and y directions using:
boundary f f p
http://lammps.sandia.gov/doc/boundary.html

I think I was wrong again. Perhaps you should use:
boundary f f f

(Sorry for all of the emails. I should think more before posting.)
Andrew

Dear Lammps Developers and Users,

I'm doing a MD simulation about nanowire (see the attachment).

Periodic boundary condition is used along the axis of nanowire. There's a
slip plane which is inclined to the axis.

The addforce command allows you to choose any direction for the force.
http://lammps.sandia.gov/doc/fix_addforce.html
http://lammps.sandia.gov/doc/group.html

One last detail:
If you want to generate a stress-strain curve (force-distance curve),
then perhaps it is better to use "fix move" instead of "fix addforce"?

http://lammps.sandia.gov/doc/fix_move.html

You can use this to move a group of atoms in any direction (f∥ or -f∥,
at a constant velocity using the "linear" keyword). If you use "fix
move", you would not want to apply this fix to all of the atoms in
each half of the wire, only the atoms far away from the slip plane
(near the ends of the wire). (You'd have to define separate groups
for these boundary atoms, as well as the mobile atoms closer to the
slip-plane.) Lastly, make sure there is enough extra space in all
directions so that all atoms remain in the boundary box during the
simulation.

Unfortunately I don't know how to insure that the fracture occurs near
the center of the wire.
Also: if you pull the two ends too quickly, then you have to worry
about non-equilibrium effects (Jarzynski/Crook). I'm sure there are
other issues I haven't thought of.
Perhaps somebody more knowledgeable should take over this thread...

Good luck.

Andrew

In addition to the comments Andrew made,
you can look at the fix lineforce and fix planeforce commands,
which allow you to discard/keep components of force in a specified
direction. But you need to know the direction in advance, which
isn't clear to me from your description of a slip plane which might
occur dynamically.

Steve