espressomd-users
[Top][All Lists]

Re: [ESPResSo] Question about Tabulated Interaction

 From: Jon Halverson Subject: Re: [ESPResSo] Question about Tabulated Interaction Date: Wed, 02 Jul 2008 14:47:36 +0200 User-agent: Thunderbird 2.0.0.12 (X11/20080213)

Hi Lorenzo,


I think we've resolved this. Another thought I had is that even if two particles are closer than the minimum separation (as defined in your tabular potential) at t = 0, since they are moving, at some later time they will be at some separation where your tabular potential is defined. So as long as force capping is turned on during some initial period you should be okay.

Jon

Jon Halverson wrote:

> Yes, you will have to be careful during initialization if you use a tabulated potential. Starting the simulation with the particles arranged in a lattice would solve the problem. An alternative is to use some repulsive force with force capping during initialization.
>
> Jon
>
>
> Lorenzo Isella wrote:
>> Sorry, a moment of confusion; you are 100% right. The potential blows
>> up at d=2R NOT at d=0.
>> You see, initially I am displacing many particles randomly in a box;
>> what if for several couples of particles the two centres are closer
>> than 2R? If the potential is not defined in that region, then nothing
>> will bring them apart. This is what worries me.
>> That is why I was thinking of introducing some constant force there.
>> Or does the minimum tabulated distance automatically take care of that?
>> Cheers
>>
>> Lorenzo
>>
>>> Hi Lorenzo,
>>>

>>> In the paper it says d is the distance between the centers of the particles. >>> With this definition everything looks fine. There are repulsive terms in the >>> potential that are sufficient to keep particles from overlapping. You should >>> be able to use the potential in an analytic or tabular form. Of course, the
>>> former is preferred but if it requires too much work then the table is

>>> probably okay (but people not familiar with ES will wonder why you made that
>>> choice).
>>>
>>> Jon
>>>
>>>
>>> Lorenzo Isella wrote:

>>>> Bad English maybe; by divergence I mean the blowing up of the potential at
>>>> d=0.
>>>> Cheers
>>>>
>>>> Lorenzo
>>>>
>>>>> Hi Lorenzo,
>>>>>

>>>>> It's possible to include a new potential but it requires making changes
>>>>> to

>>>>> nearly twenty files. So it's possible but may or may not be worth it. I'm
>>>>> not up on the details.
>>>>>

>>>>> I still don't understand the trouble even if you use a tabular potential. >>>>> What are you're exponents for the repulsive and attractive terms? They
>>>>> must

>>>>> be less than 12 and 6 - since you're integrated out the atomic degrees of >>>>> freedom - so it should be an easier interaction to handle in comparison
>>>>> to
>>>>> the Lennard-Jones.
>>>>>
>>>>> Jon
>>>>>
>>>>>
>>>>> Lorenzo Isella wrote:
>>>>>> Hello Jon,
>>>>>> Well, the potential I would like to use has an analytical form, but
>>>>>> that is not available among the potentials in Espresso.

>>>>>> I thought it was possible to include a new potential only by tabulating
>>>>>> it.
>>>>>>
>>>>>>> Hi Lorenzo,
>>>>>>>
>>>>>>> Why not just use the analytical expression? Why use an interaction
>>>>>>> table?
>>>>>>>

>>>>>>> Am I right that your potential parameters are \sigma, \epsilon, \rho_i,
>>>>>>> and

>>>>>>> R_i, where \rho_i is the average number density of LJ atoms in sphere i
>>>>>>> and
>>>>>>> R_i is the radius? Can you send the potential?
>>>>>> Not exactly. Some of this parameters are hidden in the Hamaker

>>>>>> constant and the resulting potential is a function of the radia of the >>>>>> interacting nanoparticles and the surface-to-surface separation, both
>>>>>> scaled by the r_min of the Lennard-Jones potential.

>>>>>> I have no problems in sending you the analytical expression (actually, >>>>>> in case you are interested, I could send you the reference), but first
>>>>>> I need to know if the analytical potential can be implemented inside
>>>>>> Espresso.
>>>>>> Cheers
>>>>>>
>>>>>> Lorenzo
>>>>>>
>>>>>>
>>>>>>
>>>>>>> Jon
>>>>>>>
>>>>>>>
>>>>>>> Lorenzo Isella wrote:
>>>>>>>> Dear All,

>>>>>>>> I would like to run some simulations to investigate the dynamics of >>>>>>>> nanoparticles which are supposed to interact via a potential derived
>>>>>>>> from the integration of a Lennard-Jones potential over two spheres
>>>>>>>> (each particle is much larger than a molecule).

>>>>>>>> Now, I have the analytical expression for both the potential and the >>>>>>>> force, so I am able to generate my own tabulated interaction. I simply >>>>>>>> wonder how I should choose the minimum separation between these two
>>>>>>>> particles.

>>>>>>>> Namely, the potential blows up becoming infinitely repulsive when the >>>>>>>> surface-to-surface distance, d, becomes zero and exhibits a minimum
>>>>>>>> above that distance before decaying quickly to zero.

>>>>>>>> So, physically, there is no possibility of an interpenetration of the >>>>>>>> two particles, but I do not know how I should treat it numerically. It >>>>>>>> seems that I have to specify the tabulated potential as a function of
>>>>>>>> the centre-of-mass-to-centre-of-mass separation between the two
>>>>>>>> particles, if I understood correctly.
>>>>>>>> But strictly speaking this is not defined for a

>>>>>>>> centre-of-mass-to-centre-of-mass distance smaller than 2r_p, where r_p
>>>>>>>> is the particle hard-core radius.
>>>>>>>> I think that a similar problem should arise also in the standard

>>>>>>>> Lennard-Jones potential treated for particle separations below r_off.
>>>>>>>> So, how is that dealt with in Espresso?

>>>>>>>> I am concerned that, during the evolution, a particle may end up in a >>>>>>>> region where it cannot possibly have access to and that it would not
>>>>>>>> feel any repulsion at all.
>>>>>>>> What about introducing an artificial repulsive force for small
>>>>>>>> separations?

>>>>>>>> I know this has been done in similar situations, but first I'd like to
>>>>>>>> know what is the standard treatment of such cases in Espresso.
>>>>>>>> Many thanks
>>>>>>>>
>>>>>>>> Lorenzo
>>>>>>>>
>>>>>>>> _______________________________________________
>>>>>>>> ESPResSo mailing list
>>>>>>>> https://fias.uni-frankfurt.de/mailman/listinfo/espresso
>>>>>>>>
>>>>>>>> This email was Anti Virus checked by Astaro Security Gateway.
>>>>>>>> http://www.astaro.com
>>>>>>>>
>>>>>>
>>>>
>>>>
>>
>>
>>
>