[gmx-users] mixing rules for LJ parameters?

[Erik Lindahl]

Hi,

On Dec 18, 2006, at 10:48 PM, Nicolas Winter wrote:
>
> I'm trying to understand how the LJ parameters are determined in  
> the Gromacs FF, namely the file ffgmxnb.itp. According to the  
> mixing rules given on page 88 of the manual c6 and c12 for  
> interactions between atoms should just be the square root of the  
> product of the individual atom parameters. However when I tried to  
> take the individual c6 and c12 for O and N for example, and mix  
> them according to this rule I come up with the correct number for  
> for the O - N c6 value, but not for c12. So I'm curious how the c12  
> is determined? Here are the numbers I'm working with, the  
> individual atom parameters taken directly from the Gromacs FF, file  
> ffgmxnb.tip. The O-N values are what I get from using the mixing  
> rule combination 1 on p88.
>

Mixing rules only apply when no specific parameters for an  
interaction are provided. "Gromacs FF" is mostly GROMOS87 (i.e.,  
don't use it for new runs), which has a handful of exceptions from  
the standard combination rules.

In Gromacs it would actually work fine to _only_ list the exception  
interactions, and rely on combination rules for the rest, but since  
GROMOS specifies it as a huge matrix of all combinations it is easier  
to just copy it.

Cheers,

Erik

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://maillist.sys.kth.se/pipermail/gromacs.org_gmx-users/attachments/20061219/eb64b8f3/attachment.html>