[gmx-users] Three-site model for acetonitrile

[Christoph Freudenberger]

Hi David,

Thanks for your help so far. Before we continue on that issue I'd like to ask,
if this is getting to specific for the group and if we should
continue our disccusion via email.
AFAIC setting up solvent topolgies and importing parameters from the
literature is of general interest for gromacs users.

David wrote:
> That looks right. To make sure, you can just use any output conformation
> from the simulation and measure the distances in rasmol or something
> like that...
The MD has been running overnight without errors. The results look
quite good. I have 1ns of the converged solvent box in NPT at
300K/1bar:

Energy                      Average       RMSD     Fluct.      Drift  Tot-Drift
-------------------------------------------------------------------------------
Potential                  -30.0018   0.201429          0 0.00278205    2.78206
Density (SI)                731.741    2.57664    2.57525 -0.000293333  -0.293334

The density is too smal, but the potential energy looks good, isn't it?
(You posted DHvap=32.94 kJ/mol.)
Should I do an md in NVT to have better comparable values? DH corresonds to
constant volume and temperature AFAIR my physchem lectures...

I'm just thinking about a different constrained topology to avoid the error from
redistributing the mass of the dummie:

something like:
[ constraints ]
;  ai  aj funct           b0
    1   3       1          0.26200
    1   2       1          0.14600
    2   3       1          0.11600

Would that work also?

regards
-- 
Christoph Freudenberger
Abt. Organische Chemie I AK Siehl - Uni Ulm -Tel: ++49-731-502-2785