[gmx-developers] triplet range cut-off

[David van der Spoel]

marcos wrote:
> Hello,
> 
> We are simulating liquid-gas interfaces and we want to include the long
> range forces due to VDW during the course of the simulation. 
> We think to implement a kind of triplet range cut-off. This is to add a
> new (and longer) cut-off to the already existent in the twin-range
> method. This new cut-off will be very long and the forces will be
> updated very infrequently. 
> We plan to apply this only to the VDW interaction, whereas the
> electrostatics will be handled with PME.
> 
> Another possibility is to use a VDW analog of the PME but update the
> forces very rarely.
> 
> 
> Which option do you like more, and where would be the place to start
> modifying the code?
The PME-Van der Waals would be nice to have as an option.

However, for your problem you might consider whether you shouldn't 
rather have an analytical approximation: the gas will feel the liquid a 
lot more than the other way around, but you could have some kind of 
Z-dependent force due to dispersion (if Z is perpendicular to the 
interface). You should in that case also consider the Wall potentials 
that Berk implemented in the CVS branch.

-- 
David van der Spoel, Ph.D.
Molec. Biophys. group, Dept. of Cell & Molec. Biol., Uppsala University.
Box 596, 75124 Uppsala, Sweden. Phone:	+46184714205. Fax: +4618511755.
spoel at xray.bmc.uu.se	spoel at gromacs.org   http://folding.bmc.uu.se