Dear LAMMPS-users,
I am using fix indent plane to separate solid and liquid phase. Solid is held fixed, and liquid phase is lifted in +z-direction using the fix indent plane. I am invoking following fix command:
variable zi equal ramp(-21,-8)
fix indent1 lPhase indent 120 plane z v_zi lo units box
Above the liquid phase, I have space (vacuum phase) too. Hence, the system is non-periodic in z-direction. The main aim is to sample the forces on indent plane at the different separations (-21, -20, …-8 Angstroms). I face no problem in running the simulation. I lift the liquid phase quite slowly, i.e., 1 Angstroms in 1 ns. When I run long simulation at any desired separation, I notice some oscillation (up-down) in the entire liquid-phase along with the random motion of atoms at the fix indent. This would be problematic in convergence of forces.
I thought that this could be problem of force constant of fix indent, so I tested a range of force constants (e.g. 1, 10, 100, 1000, 10000 in real units). But the oscillation appear in every cases to some extent. At this point, I am not able understand why the oscillation occurs, and how I can control this.
Can you comment on this why it is happening? How this can be resolved in LAMMPs?
with regards
Dear LAMMPS-users,
I am using fix indent plane to separate solid and liquid phase. Solid is held fixed, and liquid phase is lifted in +z-direction using the fix indent plane. I am invoking following fix command:
variable zi equal ramp(-21,-8)
fix indent1 lPhase indent 120 plane z v_zi lo units box
Above the liquid phase, I have space (vacuum phase) too. Hence, the system is non-periodic in z-direction. The main aim is to sample the forces on indent plane at the different separations (-21, -20, ...-8 Angstroms). I face no problem in running the simulation. I lift the liquid phase quite slowly, i.e., 1 Angstroms in 1 ns. When I run long simulation at any desired separation, I notice some oscillation (up-down) in the entire liquid-phase along with the random motion of atoms at the fix indent. This would be problematic in convergence of forces.
I thought that this could be problem of force constant of fix indent, so I tested a range of force constants (e.g. 1, 10, 100, 1000, 10000 in real units). But the oscillation appear in every cases to some extent. At this point, I am not able understand why the oscillation occurs, and how I can control this.
Can you comment on this why it is happening? How this can be resolved in LAMMPs?
there is very little tangible information here, so one can only
speculate. what may be the issue is, that you make your solid atoms
immobile. that way, the liquid atoms have to "squeeze" through the
available openings. if those are rather small, they would pick up
significant potential energy while being pushed to the free volume,
which could cause the situation you are describing. if my guess is
valid, then it might help (and be more realistic in the first place),
to use fix spring/self to restrain the solid atoms instead of making
them completely immobile. i would also look at the total force on the
indenter divided by its area and see if this is meaningful and without
major oscillations.
axel.
Dear Dr. Axel,
Thank you so much for your comments. Sorry for not giving the full detail of the system. Surface is made of simple polymer chains. The density is sufficiently high, so chains are almost crystalline in nature. As you pointed out, I am already using the fix spring/self to restrain the position of chains. I did another test simulation by putting another fix indent plane at top of the water layer. The second indent is placed in such a way that water between two indents has bulk density. Now, I don’t see any oscillation in the water layer, but of course it is not the real situation of the system. I suspect that the oscillation is because of finite system size. I have only 3 nm water layer over surface ( i.e., set up is like [surface | water layer (3nm) | vacuum | wall]). At any separation within surface-water interaction range when water molecule try to reach surface, it is pushed back by the indent plane, thus it hits the water layer. May be, these hitting are not surviving because of small thickness of water layer. I will check this by taking more thick water-layer, but I am not sure whether this assumption (understanding) is correct or not. A further comment on this, would be helpful.
Thanks once again
regards
Dear Dr. Axel,
Thank you so much for your comments. Sorry for not giving the full detail of the system. Surface is made of simple polymer chains. The density is sufficiently high, so chains are almost crystalline in nature. As you pointed out, I am already using the fix spring/self to restrain the position of chains. I did another test simulation by putting another fix indent plane at top of the water layer. The second indent is placed in such a way that water between two indents has bulk density. Now, I don't see any oscillation in the water layer, but of course it is not the real situation of the system. I suspect that the oscillation is because of finite system size. I have only 3 nm water layer over surface ( i.e., set up is like [surface | water layer (3nm) | vacuum | wall]). At any separation within surface-water interaction range when water molecule try to reach surface, it is pushed back by the indent plane, thus it hits the water layer. May be, these hitting are not surviving because of small thickness of water layer. I will check this by taking more thick water-layer, but I am not sure whether this assumption (understanding) is correct or not. A further comment on this, would be helpful.
sorry, but i don't really understand what the problem is, that you are
trying to solve here. also your description remains very vague.
remember, that i am not your adviser and thus no nothing about your
specific research. this really doesn't sound like a LAMMPS problem,
but like a problem of the design of your simulation. this is not
something i have the time to comment on, since i don't know your
research and its goals. i also don't consider this a good topic for
this mailing list, but rather something you should discuss with your
adviser/supervisor/colleagues.
axel.