On Tue, Dec 18, 2012, Shenglong Wang wrote:
>
> Thanks a lot for the idea. I have another question, if I just simply
> substrate Coulomb interactions, how to deal with the cutoff issue? For
> Ewald sum or PME, there is cutoff for the short range interactions
> in the real space, how to match this cutoff with the one for Coulomb
> interactions between these two groups of atoms?
I don't understand the question, but will give a couple of notes that might be
relevant:
1. The Coulomb "cutoff" in PME can be thought of as a (hidden) internal detail
of the implementation: you can change the cutoff with effectively no change in
the interactions. There's nothing physical there.
2. The only consistent way I can think of (offhand!) to modify the Coulomb
interaction is to define the modified interaction to exist among a fixed set
of atoms, with no cutoffs based on distance. Then you can subtract a fixed
number of Coulomb terms, and add back in the same number of modified
interaction terms.
3. You probably should study the code to see how 1-2, 1-3 and 1-4 electrostatic
interactions are handled, since this sounds just like what you propose. That
is: 1-2 interactions have a "modified Coulomb interaction" (which happens to
be zero), and 1-4 interactions have a scaled Coulomb interaction. In each
case, the code subtracts out the unmodified Coulomb interaction and adds back
in the modified one.
....dac
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed Dec 19 2012 - 05:00:04 PST
