12
$\begingroup$

I am trying to compute the electrostatic potential profile from a distribution of point-charges that were output from an MD simulation. Does anyone know of ready-made Poisson solver packages that are suited for MD? (specifically: can handle an arbitrary distribution of charges, work in 3D, can handle different mesh spacings in different dimensions, and can handle periodic boundary conditions)

I have done a lot of searching, but most of the solvers appear to be applicable to setups with 2D grids, equal mesh spacing in x and y, and non-periodic boundary conditions.

Improve this question
$\endgroup$
5
  • 1
    $\begingroup$ Is your system periodic? $\endgroup$
    – CKl
    Jun 23, 2020 at 3:41
  • $\begingroup$ Yes, the system is periodic. $\endgroup$
    – anneb101
    Jun 23, 2020 at 14:07
  • $\begingroup$ COMSOL Multiphysics can do that, but it's not freeware. $\endgroup$
    – Emil Zak
    Jun 23, 2020 at 19:53
  • $\begingroup$ In general, these kinds of problems are solved with multigrid methods, so it's probably worth searching specifically for multigrid solvers. I've used DL_MG before, I don't know whether it can do everything you want but it can do at least some of it: bitbucket.org/dlmgteam with documentation here: dlmgteam.bitbucket.io/dlmg_docs $\endgroup$
    – Phil Hasnip
    Jun 29, 2020 at 2:08
  • $\begingroup$ This might be useful? github.com/maroba/multipoles $\endgroup$
    – B. Kelly
    Dec 26, 2021 at 1:41

1 Answer 1

Reset to default
6
$\begingroup$

It would help to know what you need the potentials for, since this will affect the techniques that are necessary to evaluate the potential. Point charges are trivial, as the potential generated by each charge is just -Z/r; the trick is mostly how to make the evaluation efficient, this is achieved with methods like fast multipoles or particle mesh Ewald. (Continuous charge distributions are completely different animal since you don't need the solution to Poisson's equation a priori.)

However, evaluation of the Coulomb potential is a task that's handled by MD packages, since determining the non-bonded Coulomb interactions requires knowing the potential in the system. Since you got the structure from an MD simulation, maybe there's a way to extract the potential from the MD software package you used?

Improve this answer
$\endgroup$
1
  • 1
    $\begingroup$ Thank you for your response. I have been looking into the simulation, but it is a bit challenging to implement. That is why I am seeing if there is a quick post-processing tool out there. $\endgroup$
    – anneb101
    Jun 23, 2020 at 14:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .