[gmx-users] Re: who can help me with the force-field?

[Mark Abraham]

On 09/21/09, xiao shijun  <xshijun at gmail.com> wrote:
> 
>    Mark wrote:
>        >  xiao shijun wrote:
> 
> >> Hi everyone,
> >>     Recently, I met a problem on the force-field. I am doing something
> >> on the calcium carbonate, and want to simulation it with gromacs.
> >> The shell model for oxygen has been emploied in my system. When I
> >> considered the non-boned potential associated with interactions between
> >> atoms
> >> in the same CO3 group, I only need to take O-O buckingham interaction
> >> into account, but any electric interaction. How could I realize that in
> >> my .top file.
> >>     First, I tried to use [pairs] to eliminate electric interaction, but
> >> I can't add the buckingham interaction of O-O because there is no
> >> buckingham
> >> options for LJC14. I want to know if anyone of you have such the
> >> experience. Thanks!!
>      
>        >Usually you would not define a nonbonded interaction for atoms that are
>        >so close together, since their interaction should be able to be modelled
>        >adequately with the C-O bond and O-C-O angle (and perhaps with CHARMM,
>        >an interaction harmonic the O-O distance as a Urey-Bradley term).
>    
>    Thanks for Mark's suggestion. But I want to reproduce the simulation from the other paper,
>  The paper describe the force-field that O-O nonbonded interaction in the same carbonate should be considered.
>  In addition to that, I can't use buckingham potential in [pairs], even choose funct = 2. I am wondering
>  if I want to describe 1-4 interaction by Buckingham, what could I do? 
This is one application for setting nrexcl for moleculetypes. You don't want exclusions for particles 2 bonds apart, so set this to one. See chapter 5.

If that doesn't work, then if the "bonded" interactions are merely harmonic distance potentials, then GROMACS permits a form of them that doesn't create a bond (bond type 6, IIRC, check chapter 5; or constraint type 2). Thus, the automated exclusion mechanism won't be involved, and you can have Buckingham O-O so long as you have Buckingham for all other LJ interactions. You will need manual exclusions for C-O pairs, I guess.  Thus

[ atomtypes ]
C #nonbonded parameters go here

O #nonbonded parameters go here

[moleculetype]
carbonate

[atoms]
C
O1
O2
O3

[bonds]
C-O1 # type 6
C-O2 # type 6
C-O3 # type 6

[exclusions]
1 2
1 3
1 4

Whether this works may depend on whatever else is in your system, though.

Mark


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20090921/f5172705/attachment.html>