Re: [ESPResSo-users] virtual sites questions

[Rudolf Weeber]

Hello,
On Mon, Mar 21, 2011 at 11:33:02PM +0100, Jiajia Zhou wrote:
> I am trying to use the raspberry model to simulate colloids. I found
> the VIRTUAL_SITES_RELATIVE (in developer's code) is very handy, ie, I
> create a single particle at the center of a sphere, and then decorate
> the surface with virtual sites. But there are still some question I'm
> wondering if some one can help me.
> 
> 1)  I'm using the Langevin thermostat. I read in the ug file that
> there is an additional feature called VIRTUAL_SITES_THERMOSTAT. Do we
> need to turn on this feature in order to have the frictional and
> random force on the virtual sites? I tried to uncomment the line in
> the configuration but doesn't seem to do anything.
You have two options regarding the thermostat,
Either, you thermostat the rotation and translation of the centrl particle. 
This is the standard behaviour.
Or you do not thermostat the central particle and instead thermostat all the 
virtual particles. This is currently only implemented for the case that there 
are no other real particles to thermostat. You can achieve this behaviour by 
setting
THERMOSTAT_VIRTUAL_SITES and THERMOSTAT_IGNORE_NON_VIRTUAL.


The difference between those two options lies in a different effective gamma, 
i.e. relaxation time for the rotational motion of your raspberry. If you are 
not interested in dynamical properties, you can therefore safely chose the 
first option.



Please note that the thermostatting switches for virtual sites have not yet 
been implemented for Lattice-Boltzmann and dpd. If you want to couple the 
raspberry to an explicit fluid simulation, this coupling should definitely act 
on the virtual particles.

Hope that helps!
Regards, Rudolf