Membrane

LAMMPS LAMMPS Mailing List Mirror
1

Dear lammps users and developers
I have a membrane and a sphere and i want to cross the sphere through the membrane , i am using fix rigid for sphere and fix npt iso for membrane but here is the problem during relaxation , as long as the sphere is small and so the simulation box is small there is no problem but when the sphere is big and the simulation box is big , the simulation box shrinks so does the membrane . Because i want to see the crossing when i use big sphere i determine the simulation box along z axis even bigger to see it after the crossing .
I appreciate any suggestions.

2

there is not much useful specific information here to go on. your description hints at the possibility of some serious finite size issues with the process you describe, but there is not enough information and i am not enough of an expert in membrane simulations myself (i know know about it from current and former colleagues).

however choosing fix npt iso is a very odd choice for your kind of system. why not keep the z axis unchanged and only run npt in x and y? it seems rather undesired to have changes of the pressure in z direction affect the size of the box x and y.

axel.

3

Thanks for your reply , i believe my problem was minimization , i did not minimize it enough so i changed the minimization parameters and i think my problem is solved.

4

Thanks for your reply , i believe my problem was minimization , i did not minimize it enough so i changed the minimization parameters and i think my problem is solved.

you may have removed the immediate symptom, but the underlying, more conceptual issues are not addressed by changing minimization parameters.

axel.

5

Thanks for your reply , i believe my problem was minimization , i did not minimize it enough so i changed the minimization parameters and i think my problem is solved.

you may have removed the immediate symptom, but the underlying, more conceptual issues are not addressed by changing minimization parameters.

axel.

I think Axel’s point about using isotropic pressure deserves consideration. You should allow your simulation to change size in the X and Y directions, but not in the Z direction (which is currently what is occurring in your simulation).

If it helps I am using this protocol for my (coarse-grained, zero-tension) lipid membrane simulations.

fix 1 all langevin 1.1 1.1 50.0 48279
fix 2 all nph x 0 0 1000 y 0 0 1000 couple xy

Here I am using a combination of “fix nph” and “fix langevin”, however you can also use “fix npt” instead (with the same settings “x 0 0 1000 y 0 0 1000 couple xy”.) For details how to run these kinds of (anisotropic) pressure simulations see:
https://lammps.sandia.gov/doc/fix_nh.html

(The complete example is here).

Cheers

Andrew

6

Thank you all , but why not control pressure in z direction? I mean couple x and y together and barostat z seperately?

7

Thank you , not only did i change the minimization but also i changed the Pdamp

8

Thank you all , but why not control pressure in z direction? I mean couple x and y together and barostat z seperately?

I have not tried this. (For the biological systems I simulate, both the tension in the membrane and the pressure in the surrounding solvent [if explicitly represented] are close to zero.) But if you want to have a separate pressure in the Z direction, I think LAMMPS will let you do that. (:heart:) I am guessing that you could achieve that using a command like this:

fix 2 all nph x 0 0 1000 y 0 0 1000 z 10 10 1000 couple xy

(also works with “nvt”. “couple xy” is optional)

But I could be wrong. See the documentation at:
https://lammps.sandia.gov/doc/fix_nh.html

Good luck.

Andrew

9

Thank you so much , don’t you think because in a great part of simulation box In the z direction i have nothing so barostating in z is not necessary and it will be zero or if i barostat it, it will keep shrinking in z diection ?

10

Thank you so much , don’t you think because in a great part of simulation box In the z direction i have nothing

Does this you mean that your system only contains a membrane? (There is no surrounding solvent?)

(If so, that’s okay. Some decent coarse grained models of membranes use implicit solvent to reduce computation time. The example I showed earlier does not have any explicit water.)

so barostating in z is not necessary and it will be zero or if i barostat it, it will keep shrinking in z diection ?

I would imagine that if you do not specify a pressure in the Z direction, that the size of the simulation box in the Z direction will remain constant.

Andrew

P.S. This does not necessarily mean the pressure in the Z direction is zero. However if I understood you correctly, your simulation contains only a flat membrane present in the XY plane, and no water, solvent, or other molecules to fill the space above and below the membrane. In that case I suspect the pressure in the Z direction will indeed be close to zero
(…provided that the simulation box in the Z direction is significantly wider than the membrane, and ignoring any long range electrostatic effects that could be effecting your system. However I doubt you would using long range electrostatics in an implicit-solvent coarse-grained membrane simulation) …Either way, I don’t think the simulation box would shrink in the Z direction.