Re: RDF of hydroxide ion in a water box

From: RonitS Chem (
Date: Sun Sep 01 2019 - 14:10:43 CDT

Hello all,

After reading through some literature, I came across that TIP3P water model
does not show long range behavior properly. Besides TIP4P, TIP5P model is
the only one which replicates the second and third solvation shell
properly, as mentioned in this paper: , and its reference
on TIP3P water and a modified model called TIP3P-Ewald.

The only modification that the paper: suggested was in the LJ
potential terms of Oxygen and Hydrogen. After making the same
modifications, I see the second and third solvation shells in the rdf but
there was a significant overbinding, and the density is messed up. Since
the work I am doing strictly focusses on TIP3P water model, I am benching
the idea of getting proper long range behavior, but I do want to understand

I guess my question is why in some models, there are no Hydrogen LJ
parameters, and how to make sense of changes in LJ parameters physically? I
mean without running an actual simulation, how to predict what changes
might occur in a system on tweaking the epsilon and Rmin/2 a little.


On Thu, Aug 22, 2019 at 7:17 PM RonitS Chem <> wrote:

> Hello everyone,
> I am pretty new to NAMD, and MD in general so please forgive me if this
> question seems too naive/simplistic, but I have searched on this mailing
> list a lot and cannot seem to find any solution for the question I have.
> So, while learning how to run NAMD, I simulated just a water box and
> calculated the RDF via VMD and the first solvation shell was fine but I had
> a hard time getting the second solvation shell. I attributed this problem
> to the rigid bonds, but still not sure since it is advised that for
> 2fs/step rigid bonds should be on.
> Now, I am trying to simulate a single hydroxide ion in a water box using
> the parameters given in the toppar_water_ion.str file. The RDF was
> calculated between hydroxide oxygen and water oxygen, and it shows there
> are 9 atoms in the first solvation shell, which does not seem correct.
> I tried testing for ion O and water H, but that also shows 9 in the first
> solvation shell with a distinct second shell, which is even weirder. After
> looking through the literature, the RDF is definitely messed up, but the
> reasoning is unclear.
> I should mention that after parametrizing OH radical and running the same
> input file but with a radical, in place of the ion, the RDFs match at least
> one paper that I have been able to find.
> Should I play with the values of timestep, rigidbonds or is the problem
> elsewhere?
> Thank you for taking the time to read and any input will be appreciated.

This archive was generated by hypermail 2.1.6 : Sat Dec 07 2019 - 23:20:49 CST