[gmx-users] Cut-off coulombtype heats my system in SPC water

[Michael Shirts]

> Dear All,
>
>I solvate my molecules in SPC water.I do not use PME  but Cut-off  as
>coulombtype in my .mdp file in order to save some simulation time.
>However, I found the temperature after equilibration is 12 degree bigger
>than 300K in my system.In this simulation I only care the properties such
>as rdf, msd and gyration. Do I need to do a new simulation by using PME
>because of the higher temperature?

Cut-off methods do not conserve energy, and therefore add energy to the
system, which results in higher temperature.  I would suggest using Switch for
Van der Waals and PME for electrostatics.  You may have to fiddle with the
parameters to get it working well, but I have had good success with this
combination.  If it is CLOSE to conserving energy, then using Nose-Hoover
(Berensdsen does not quite give the right ensemble), can help keep the
temperature about the same, and give very close to the correct ensemble.  But
that's just essentially allowing the heat that is added in to leak back out.
Better to not gain the extra heat to begin with.

The temperature can definitely affect the rdf and msd, so I would suggest
running it again.

Cheers,
Michael Shirts
Pande Group
Stanford University