Hi all
I've came cross many MD works where system charge was balanced (neutralized) through adding some arbitrary counter-ions.
Why did they do that? What if simulating a charged system? Is it necessary to neutralized any charged system before MD simulation?
Bests
If using periodic boundary conditions (PBC) and a grid-based electrostatics evaluation procedure (e.g. PME) you *need* a neutral box otherwise the electrostics will essentially add to infinity from the interaction of your box with the presence of your infinite periodic images.
By having a total charge = zero, your electrostatics will sum to a finite, correct value.
If using periodic boundary conditions (PBC) and a grid-based electrostatics evaluation procedure (e.g. PME) you *need* a neutral box otherwise the electrostics will essentially add to infinity from the interaction of your box with the presence of your infinite periodic images.
By having a total charge = zero, your electrostatics will sum to a finite, correct value.
To expand a bit on Christian's answer above, the Ewald method used to compute electric fields only formally converges if the net electric charge is zero. As a practical matter, you will rarely be able to discern the difference. I have miscounted on occasion and left a stray charge or two floating around without detecting any particular changes in results when I went back and reran the simulation with a neutral box.
You do need to be careful if you are doing something like trying to conduct free energy perturbation calculations where the charge changes. For example, if you are deprotonating a carboxyl residue in the active site to assess likely protonation states, you really should simultaneously mutate a water molecule into an hydronium ion somewhere to keep the overall charge neutrality. (The proton really must go somewhere.)
- Badges
- Science topic