[gmx-users] Three-site model for acetonitrile

[Christoph Freudenberger]

Hi there,

I got into trouble looking for an apropriate model for acetonitrile
to use in gromacs.
I checked three potentials from the literatur by doing some short
simulations, finding them to give similar results, so i decided to
use the following for my system.

E. Guardia, Mol. Simul., 26, 2001, 287;
The parameters given there are:

		eps (KJ/mol)	sig(A)		 q(e)
Me 		0.7824		3.775		 0.206
C		0.544		3.650		 0.247
N		0.6276		3.200		-0.453

Here is what i converted them to for my .itp:
[ atomtypes ]
;type   mass           charge    ptype c6            c12
MeAN    15.03500       0.000       A   0.90571E-02   0.26212E-04 ;
  CAN    12.01100       0.000       A   0.51454E-02   0.12167E-04 ;
  NAN    14.00670       0.000       A   0.26955E-02   0.28943E-05 ;

[ moleculetype ]
; name  nrexcl
Acetonitril        2

[ atoms ]
;   nr  type    resnr   residu  atom    cgnr    charge  mass
1       MeAN     1       MeCN    C1AN    1        0.206  15.035
2       CAN      1       MeCN    C2AN    1        0.247  12.011
3       NAN      1       MeCN    N3AN    1       -0.453  14.007

The problem is, that the system "freezes" in longterm simulations.
The molecules do not move anymore (except some minor oszillations)
as if the viscosity of the system was extremly high.
I figured that the density of the system (950g/l) might be to high,
so I changed the LJ-Parameters a little bit, to accieve better
agreement with the expt. density. First every thing looked fine
but again in a longterm simulation the system froze to a cystal-like
behavior, although the density remained constant at the new value
(~770g/l).

Can you give me a hint, what is wrong here?
Do you know a set of parameters for MeCN that works with gromacs?

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