Diffusion coefficient of molecules

User discussions
1

Dear GROMACS community,
I tried to calculate the diffusion coefficient D using “gmx msd” for the water molecule(H-O-H), which turns out a following :
5.7147 (+/- 8.8616) (1e-5 cm^2/s)

This value is almost double the reported value from various literature sources. I have tried different settings provided in “gmx msd“, but the diffusion coefficients obtained are quite overestimated with huge errors.

So, I have two very specific questions :

  1. How to accurately and precisely determine the D value for water?
  2. How to obtain accurate D in heterogeneous systems?

Thanks in Advance!

2

You should check the literature for the diffusion constant of your water model. For many water models this is about a factor 2 too high. If you want realistic values, choose a realistic water model.

Heterogeneous systems pose no complication at all. You simply select the molecule type in gmx msd and compute the diffusion constant.

3

Ok. So, it’s all about choosing the correct water model rather than any random water model. Such as TIP3P overestimates the D values for both homogeneous and heterogenous system.