Re: [ESPResSo-users] Soft sphere interaction

[Axel Arnold]

> On 31.10.2014, at 12:39, Aleksei Kabedev <address@hidden> wrote:
> 
> Dear all,
> 
> I have a question about soft spheres interaction. I tried to use one to get a 
> potential of interactions, corresponding to linear growth of forces while the 
> distance goes to zero (to avoid overlapping on small distances, for LJ I 
> initially had sigma=0.3 and sigma =0.6)
> So I took a formula for potential
> U(r)=169.37*(r-0.65)^2, which corresponds to the 2nd virial coefficient I'm 
> trying to reproduce, and chose the parameters for SS Potential
> a = 169.37, r_off = 0.65 and finally n = -2.
> Probably the last parameter cannot be chosen like that(negative), because I 
> got energy of interaction exactly the same as I expected, but the forces are 
> equal to zero in the whole range of distances. I had a look at the code, but 
> didn't find what the issue can be related with. By the way, there is another 
> problem with potential - it is being cut on 0.9 in case of r_off=0.65. If 
> r_off=0.4, the cut is on 0.65. I don't really understand where this shift of 
> 0.25 comes from.

Well, I guess you gave a cutoff of 0.25  - you didn’t specify the cutoff above, 
but you need to hand it to the soft sphere interaction according to the manual. 
The formula in the user’s guide is wrong, however - the cutoff is at r_cut + 
r_off, that’s why it shifts with the r_off.

When it comes to forces, I can just guess, since you don’t say how you check 
the value of the forces. But the forces are only computed when doing an 
integrate command, so maybe you are lacking an "integrate 0” before evaluating 
the forces.

> Apart from that, could you please tell me if I can use both dpd_interactions 
> between particular types of particles in combination with Lattice-Boltzmann 
> Method?

What do you mean by dpd_interactions? If you refer to the dpd thermostat, then 
using it in conjunction with lattice Boltzmann is not really advisable. You can 
of course from the code point of view, but you will either thermalize some 
degrees of freedom twice if you use the lattice Boltzmann thermostat as well, 
or you will not thermalize all degrees of freedom if you don’t give a 
temperature to the LB fluid. There are still some case when you can do so, but 
you need to be really fit in statistical mechanics to get the fluctuations 
right.

Best,
Axel