[gmx-users] reflecting boundary conditions wall

[liaoxyi]

Hi dear gmx-users and developers,
 
I'm working on a system involved with solid surface and solution, which using periodic boundary conditions in x and y directions only.
I need to creat a virtual wall at both z=0 and z=z, but it should be a reflecting boundary conditions wall.
It's like this:
pretend that there is a virtual wall at z = z, when a water molecule (for example) is about to go though the wall, the wall plays like a mirror, reflecting the molecule's behavior beyond the wall back to the origin side, with the velocity component at Z direction changed to the opposite side.
Is there a possible way to achieve this, and how?
Should I change the code? and where is it?
 
I tried the wall parameters based on wall_type, wall_atomtype, wall_density etc. with a pure water system, but there comes some problems. When it comes to the table.xvg, I have no exact idea about the parameter setting. The way I do is to copy the gmx_top_path/table6-12.xvg and make the coulomb interation and dispersion interaction zero, leaving the repulsive term's values unchanged. Then rename the table file to table_SOL_wall0.xvg and table_SOL_wall1.xvg separately.
Some settings in .mdp are as below:
nwall = 2
wall_atomtype =SI SI
wall_type = table
wall_r_linpot = 2
wall_density = 2000 2000
wall_ewald_zfac = 3
I make the wall_density especially large, still some water come out of the box in z direction.

I think the parameters setting and table.xvg is a little bit complicative for me.
So, I do hope there is some way to use the reflecting boundary conditions wall, which doesn't need the parameters above.
I need some hints badly and sincerely. Anyone can help me? Thank you!!
 
Sincerely,
Kiara
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20110505/ed59b144/attachment.html>