PMF problem

Hello

I utilize colvars module in LAMMPS to calculate PMF. Use Lammps version 2015-07-21. In this process, I use moving restraints: steered molecular dynamics to calculate PMF during a Na+

dragged into the water bulk phase from vacuum. In order to make sure system net charge to be zero, I add a Cl- in vacuum and set fixed.

However, there is a problem: the Na+ PMF alway can’t get a balance platform form vacuum to water bulk phase. In addition, the Na+ PMF can’t get another balance platform in the water bulk phase. I want to know if the presence of Cl- caused the PMF result and how to solve the influence.

Thanks a lot for your suggestions.

Mei

Hello

     I utilize colvars module in LAMMPS to calculate PMF. Use Lammps
version 2015-07-21. In this process, I use moving restraints: steered
molecular dynamics to calculate PMF during a Na+

dragged into the water bulk phase from vacuum. In order to make sure
system net charge to be zero, I add a Cl- in vacuum and set fixed.

However, there is a problem: the Na+ PMF alway can’t get a balance
platform form vacuum to water bulk phase. In addition, the Na+ PMF can’t
get another balance platform in the water bulk phase. I want to know if the
presence of Cl- caused the PMF result and how to solve the influence.

​this is more a question about the science and how to compute valid PMFs
for your system than about ​how to use LAMMPS and the colvars package.

as for the impact of the neutralizing Cl-: what kind of an effect this has
is easy to find out. just set up a calculation without it and compare the
results. without knowing the answer, my sense is that especially placing
the Cl- into the vacuum area where there is no screening of the known-to-be
long ranging coulomb interactions, i would indeed be *very* worried. in
addition, it may be worth considering what the PMF would look like for a
neutral atom, e.g. a united atom methane. that might give additional
insight into the effect of the charge, especially in the vacuum part of
your system.

as for how to properly address this kind of issue: have you searched the
literature and checked what other people have done? i don't think that you
are the first person attempting this kind of calculation. if your total
system is large enough, i don't think that having a total charge of +1 or
-1 will have too much of an impact.

as always, for scientific issue, it is often worth to consult your adviser
for advice. after all, that is the job of an adviser. :wink:

axel.

Thanks for your suggestions very much. According to your suggestion, I removed the Cl- anion from vacuum, and got the balance platform of Na+ cation PMF in vacuum . I searched some papers and found the PMF should be a balance platform in the water bulk phase. But the Na+ cation outputAccumulatedwork W® always increases from interface to the water bulk phase with my system. I want to know the reasons and hope you give me some suggestions again, thanks.

Attach my colvars file:

colvar {

name d

width 0.2

lowerBoundary 0.0

upperBoundary 250.0

outputVelocity on

distanceZ {

componentCoeff 1.0

main {atomNumbers 3601} #Na+

ref {dummyAtom (20,12,0)}

}

}

harmonic {

name distance

colvars d

forceConstant 10.0

centers 90.43994 # steered MD from 90.43994 A…

targetCenters 0.0 # … to 0 A

targetNumSteps 9000000 # with these many time steps

outputCenters on

outputAccumulatedwork on

}

colvarsTrajFrequency 1000

colvarsRestartFrequency 1000

analysis true

Hello Dai, as Axel said, the fact that the results of a simulation (where statistical convergence is an issue) don’t match up with the literature is a strictly scientific issue, and not an issue with the code.

I presume you are aware that simulations in a finite time are subject to non-equilibrium work and insufficient sampling. On top of that, air-water interfaces are not that easy, because models and protocols are usually derived for the bulk phase.

It really sounds like you should plan, discuss and analyze this simulation together with your adviser.

Giacomo