C60

LAMMPS LAMMPS Mailing List Mirror
1

Dear all
I need a little help about applying pressure on C60.
I use displece_atoms command to impose pressure on each atom of C60 fullerene,
and in order to integrate equations of motion I use fix nvt command, as below:

…

label looplable

displace_atoms all move v_press_x v_press_y v_press_z
#fix forcefinal all addforce v_press_x v_press_y v_press_z

fix fix1 all nvt temp 300 300 0.01
thermo 5000
thermo_style custom step temp pe v_sigma
run 5000

next Nstep
jump in.CNT looplable

My problems:
due to using fix nvt, C60 start to rotate because there are not any constraints on C60 atoms.
How can I do to prevent C60 from rotating? as you know, fix rigid … torque … command is not proper for this code.
and, why “fix addforce” command can not work for this study? (through following dump file I see that force on each atom is increased but VMD doesn’t show any signs of pressure.
Thanks.

2

Hi

Something under the effect of pressure will shrink in a direction and not move! For example, if we increase the outside pressure on a balloon, the atoms on its right side will move toward the left while the atoms on the left side will move to the right. Therefore, “displacing” all the atoms in a direction is not beneficial at all.

Similarly, adding “the same force” to “all” the atoms in a direction neither increase the pressure nor prevents a group from rotating.

I believe it is very difficult to restrain the rotation (or translation) of a group of atoms without influencing the physics of your problem. You can use “set command” to fix the position of only some atoms (at least 3 atoms) while leaving the other atoms freely vibrate.

Another solution is attaching C60 in some points (at least 3) to new very heavy atoms. As momentum is the product of mass*velocity, the heavier the atoms the lower their velocity will be. Thus, those heavy atoms almost remain stationary. This makes them good options to be used as supports (the same as supports in statics) which prevent C60 from both rotating and translating.

You could also attack some strong springs to C60 (a pair of springs in any direction which are attached at different points!).

As the spring constants are not infinity, they could not prevent the rotation completely and a “little” (depending on the value of spring constant) vibration around all the axis will remain. the same is happening in the case of heavy atoms.

Albeit, the producers will have better solutions.

Regards

MN

Sent with Mailtrack

3

Dear all
I need a little help about applying pressure on C60.
I use displece_atoms command to impose pressure on each atom of C60
fullerene,

i disagree with your wording. what you describe would be called
"compressing a C60 molecule" or "applying strain to a C60 molecule".
pressure is a property of a system and more specifically the force per
area in a defined volume against its boundaries.

second, i don't see much of a meaning in your setup. what is the point
of applying a thermostat to a single molecule? what is the physical
meaning of your calculation?

and in order to integrate equations of motion I use fix nvt command, as
below:

...
label looplable

displace_atoms all move v_press_x v_press_y v_press_z
#fix forcefinal all addforce v_press_x v_press_y
v_press_z
fix fix1 all nvt temp 300 300 0.01
thermo 5000
thermo_style custom step temp pe v_sigma
run 5000

next Nstep
jump in.CNT looplable

My problems:
due to using fix nvt, C60 start to rotate because there are not any
constraints on C60 atoms.

well, to some degree, it is intrinsic in your model. if your
displacement variables result in only the smallest difference (and
that is likely due to using floating point math, you'll have a net
angular momentum. using a thermostat will usually emphasize it.
also, if you want your C60 to be compressed isotropically *and* have
kinetic energy, where should the kinetic energy go? bond/angle
vibrations will break symmetry, you don't want rotation and if all the
kinetic energy would go into translation, then you can just do a
calculation at 0K (see the recurring discussions of the "flying ice
cube syndrome" on this list).

How can I do to prevent C60 from rotating? as you know, fix rigid ... torque
... command is not proper for this code.

before worrying about this, it is more important to discuss, what you
want to learn from this. i don't see much use in the model you
describe (incompletely), since i cannot imagine any setup, where a
perfectly isotropic strain is imposed on a C60 molecule. anything else
*will* result in (some) rotation. and that kind of rotation will
dissipate only on interactions with other molecules (e.g. a solvent or
other C60s),

and, why "fix addforce" command can not work for this study? (through
following dump file I see that force on each atom is increased but VMD
doesn't show any signs of pressure.

from what you describe fix addforce *does* work, but you may be
underestimating the force on each atom required to (isotropically) and
visibly compress a C60 molecule. this assumes that there are no
mistakes in your input.

please also keep in mind, that a stress/strain curve for a perfectly
isotropic compression of a C60 molecule requires running at 0K and
then - thanks to the high symmetry of C60 - it should be
straightforward to infer it from the force field parameters you are
using.

axel.