Sampling different configuration from NVT ensemble

User discussions
1

Dear All,
I am trying to create an air-water interface. For this, I am following this paper by Liu et. al “doi.org/10.1021/jp046807l”. Simulation involves multiple steps:

  1. NPT Equilibration for 200 ps
  2. NVT Equilibration for 500 ps
    I am able to perform steps 1 and 2. Next step involves sampling nine independent configurations from the NVT ensemble run (step 2). This will be followed by NVE equilibriation runs.

Can someone please guide me on how to do this sampling using Gromacs.

2

Hi @shivangi

I took a look at that paper and I don’t see where the authors define “independent”. I think they just mean “different”. In that case you could just use trjconv to dump out 9 equally spaced frames from your NVT run. However you’d probably be much better off doing 9 separate NVT runs (with different seeds of course) and using the last conformation of each one. That way all of your structures will be equally well equilibrated.

BTW The reference you cite seems to be mostly about using a polarizable forcefield to model the air:water interface better. What kind of forcefield are you going to use? This is potentially deep water (!) so you may want to get some expert help before you go much further.

Good Luck!

Guy Vigers

3

Thank you gvigers for the explanation and suggestion.
I am going to use TIP4P/FQ Water model for the simulation. If you have any suggestion regarding air water modeling please tell. I am new to Gromacs and md simulation so I really need some expert help.

4

Hi @shivangi

Sadly I’ve never used a polarizable forcefield or worked on air/water interfaces. There are many more expert users in this forum, but you probably need to explain what you are trying to achieve and why.

Sorry that I can’t be more helpful myself!