[gmx-users] query on table potential

[Mark Abraham]

On 8/06/2012 1:09 PM, Sanku M wrote:
> Hi,
>   I am trying to use a table potential implementing 
> Weeks-Chandler-Anderson (WCA) interaction between a solute and all 
> atoms of a solvent like TIP3P water ( with nonzero interaction in the 
> Hydrogen atoms) . But, I am having a trouble in specifying energy 
> groups required for table potential.
> Here is the problem:
>  I have two atom types in water : OW and HW. and one atom type in 
> solute : POL
>
> Now, I want to use  two Table potentials with WCA interactions between 
> a) POL & HW and b) POL & OW
> So, I created two table potential files: table_POL_HW.xvg and 
> table_POL_OW.xvg
>
> and then
> in the .mdp file, I specified
> energygrps = POL OW HW
> energygrps_table = POL OW POL HW
>
> But, compiling the .mdp file using grompp ( version 4.5.4)  gives me 
> following error
> "Fatal error:
> atoms 33 and 34 in charge group 1 of molecule type 'SOL' are in 
> different energy groups"
>
> Clearly, using OW and HW as separate energy groups is causing the problem.
> So, I am looking for an option on how I can use the table potential 
> between solute and each atoms of solvent.
>

Energy groups can only be unions of complete charge groups. So your 
energy groups require that you split the solvent charge group. The 
solvent charge groups are defined in its [moleculetype]. Splitting will 
mean that you will get integration artefacts at the cut-off unless you 
use a proper long-ranged electrostatic method (i.e. an Ewald family method).

Mark
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20120608/b1023b58/attachment.html>