I'm relatively new to MD simulation, and am studying a highly viscous system of very concentrated aqueous biphasic systems of two different electrolytes. The higher-concentration systems are notoriously slow to equilibrate, it can take over 500 ns to equilibrate (from literature). When I run npt equilibrations with the barostat set to 1 bar, the average pressure through the simulation increases with the concentration - for example, for 10 molal in both phases, av. pressure is 2 bar, whereas for 20 molal it's 7 bar. However, the density of the systems doesn't change through the sim. Would this suggest that the systems are not equilibrated, or is this just to do with how the pressure is calculated?

I'm aware that the pressure can fluctuate a lot with RMSD in the hundreds. I just thought it weird that the average pressure would increase with increasing concentration?

I'm using a modified AMBER force field for one phase with scaled charges of 0.85 for ions, the Madrid force field for the other, and TIP4P/2005 for the water. Some boxes are 5 x 5 x 10 nm and others 10 x 10 x 30 nm

  • $\begingroup$ Can you share the details how long unit simulation has run and what's the barostat and thermostat parameters the using? $\endgroup$
    – user35952
    Dec 12, 2022 at 2:56
  • $\begingroup$ Sims have been running for over 600 ns, and I'm using Parrinello-Rahman barostat with a time constant of 2.0 and compressibility of 4.5e-5. The thermostat is Nose-Hoover with a time constant of 0.5. Thanks $\endgroup$
    – H W
    Dec 12, 2022 at 11:29
  • $\begingroup$ I think if you can plot running time-average of pressure (let's say over every 100 steps), we might be able to see if there is a systematic increase/decrease with time. If you see any such increase/decrease it might indicate the lack of equilibration. Can you may be add such plots? $\endgroup$
    – user35952
    Dec 12, 2022 at 12:00


Browse other questions tagged .