[gmx-developers] Re: TIP3P and TIP4P parameters

[Erik Lindahl]

Yep.

In theory we could of course use the exact C6/C12 values in the 
forcefields that specify the parameters that way, but I'm not too 
enthusiastic about the idea of getting slightly different TIP3P 
parameters depending on whether you selected Gromos96/OPLS/Amber for a 
forcefield... asking them seems like a good idea - if nothing else to 
make sure we use the same values as the model authors.


Cheers,

Erik

On Wednesday, June 11, 2003, at 09:03  PM, Michael Shirts wrote:

>
> OK, let me write this all out. The problem comes from the fact that the
> original defintions of TIP3P and TIP4P in the 1983 paper in the 
> previous
> e-mail come from the equation:
>
> U = A/r^12 - C/r^6
>
> For TIP3P, the constants are A = 5.820 x 10^5, C = 5.950 x 10^2
> For TIP4P, the constants are A = 6.000 x 10^5, C = 6.100 x 10^2
>
> Whereas later papers use the form:
>
> U = 4*eps*[(sigma/r)^12 + (sigma/r)^6]
>
> The original paper's values for A and C yield:
>
> TIP3P:
> eps = 3.1505651 A     sig = 0.1520725 kcal/mol
> eps = 0.31505651 nm   sig = 0.6362713 kj/mol
>
> TIP4P:
> eps = 3.1535779 A     sig = 0.1550416 kcal/mol
> eps = 0.31535779 nm   sig = 0.6489640 kj/mol
>
> In at least one later paper (Mahoney and Jorgensen, jcp, vol 122 no. 
> 20, 22
> May 2000, p. 8910), these parameters are listed as:
>
> TIP3P:
> sig = 3.15061 A       eps = 0.1521 kcal/mol
>
> TIP4P:
> sig = 3.15365 A       eps = 0.1550 kcal/mol
>
> Clearly there has been rounding, in both the sigma and the epsilon.  
> Though
> the sigma results are a litle strange even then . . .  I've checked 
> these
> several times, but if somebody else would once more just to be sure, 
> it would
> be nice.
>
> I'll do some test to try to figure out how much difference these 
> changes could
> make. Perhaps I'll send an e-mail to Jorgensen and/or Madura about it 
> once I
> run those tests . . .
>
> Cheers,
> Michael
>
>
> _______________________________________________
> gmx-developers mailing list
> gmx-developers at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-developers
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-developers-request at gromacs.org.