Re: Excluding intra-molecular electrostatic interactions

[Rudolf Weeber]

Hi Margaret,
On Thu, Sep 17, 2020 at 05:06:32PM +0200, Margaret Rosenberg wrote:
> I'm attempting to use pypresso to simulate a system of anisotropic charged
> colloids. Each of these colloids is represented using the raspberry model.
> Following the documentation, I've excluded steric interactions between
> virtual sites on the same molecule. However, I also need to add a charge to
> each virtual particle: how can I exclude Coulomb interactions between
> virtual sites in the same colloid?
There is this bonded interaction:
http://espressomd.org/html/doc/espressomd.html#espressomd.interactions.BondedCoulombSRBond
and also a variant for pure 1/r.

Depending on what you intend to do, excluding the electrostatic ia between the 
particles making up a rigid body may not be necessary.
In theory, pair-wise interactions within the virtual sites of a rigid body add 
up to zero due to Newton's thhird law.
For steep potentials like Lennard-Jones, some of the contributions to the sum 
might be huge though, if your virtual sites are "overlapping" based on the 
Lennard-Jones sigma.
So due to numerical errors, the total force might end up being non-zero.
For 1/r, it might well be less of an issue. 
To check, integrate a single free raspberry for a few time steps and check that 
the force and torque on the central, non-virtual particle remain zero.

I mostly just use a specific particle type for the beads in a raspberry and 
only define interactions with respect to other particle types. I.e., there is 
no LJ interaction within the raspberry in the first place.

Regards, Rudolf