[gmx-users] Walls and PME for long range electrostatics

[Alexander Miessen]

Dear all,


I am simulating a Au(111) slab with explicit electrons inside (included 
as a "free electron gas"). The gold atoms are frozen, the electrons are 
free but confined inside the slab by a more repulsive pair interaction 
with the Au-surface atoms. This works fine and simulations are stable.

I am using a xy-periodicity. For this I introduced walls on both sides 
of the box. My wall-atoms dont interact with any of my atoms.

For a couple weeks now, I have been trying to reproduce image charge 
effects with this model, freezing a test charge (no LJ-interaction, only 
a charge of +1e assigned to it) at a certain distance from the Au 
surface. In the documentation it is stated that z_box should be 3 times 
as big as the system when using walls with a slab geometry. I did this. 
But the forces I calculate on the test charge neither match for the far 
field, nor for the near field.

Do you have any suggestions on how the walls and therefore PME might 
bring artifacts in? I am running out of ideas on how to fix this problem.


Thanks,

Alex