[gmx-developers] different nb parameters for inter and intramolecular interactions

[Berk Hess]

Berk Hess wrote:

> David wrote:
>
>> On Wed, 2005-07-27 at 15:36 +0200, Lubos Vrbka wrote:
>>  
>>
>>> hi guys,
>>>
>>> sorry to post it here, but i already tried gmx-users with no luck. 
>>> maybe someone who knows the code better might now...
>>>
>>> is it possible to have in gromacs 2 sets of nonbonded parameters for 
>>> the
>>> same atom types?
>>> our potential for water consists of following terms:
>>> coulombic interaction between all particles (including intramolecular)
>>> buckingham potential for O-H intramolecular
>>> buckingham potential for O-H intermolecular
>>>   
>>
>> Can't you use a Morse potential (bonded) instead?
>>
>>  
>>
>>> therefore, i've two atom types (O and H) that have different nb 
>>> parameters for intramolecular and intermolecular interactions. can i 
>>> somehow define and use this setup in gromacs? i don't see any 
>>> solution...
>>>   
>>
>> No you can not.
>>  
>>
> You could use tabulated LJ potentials.
> If you really only have two types of interactions
> you can fill the dispersion table with the one and the repulsion table 
> with
> the other Buckingham potential and then set the c6 and c12 parameters
> appropriately.

I forgot one essential part.
You should then set exclusions for in the intramolecular interactions
and add pair interactions of O-H. This allows you to use different
parameters for the intra and inter molecular interactions.

But now I think of it you probably can't use this as you probably
also have O-O intermolecular interactions.

In that case the CVS code might help, which allows different tablulated
for the pair interactions. With the previously mentioned trick you then
still have two different potential tables available for the intermolecular
interactions.

Berk.