The molucules stop moving during the umbrella sampling

LAMMPS LAMMPS General Discussion
1

Dear Lammps users,

I meet an issue with the umbrella sampling method computing the PMF: the molecules in the system stop moving after a certain time of simulation.

I want to use the umbrella sampling method to compute the PMF of an cation (Na+) adsorption from the bulk water to the vacuum-water interface with anionic surfactant present at the interface. I use “fix spring tether” command to control the location of the cation. The simulation works well at the beginning, but when the cation approaching to the vacuum-water interface, the (water) molecules slow down while the temperature remain at 300K. The whole system even stop moving at late time. I checked the potential energy of the system, and it remain stable at the beginning. It increases (i.e., more negative) at certain time (certain location of the cation, maybe) and gradually reach a maximum when the all molecules stop moving. Could someone please provide me any suggestion about this issue?

Here are the additional information about my simulation. The water molecules are modeled with SPC/E force field and the surfactants are modeled with the OPLS-AA force field. The simulation is conducted in the NVT ensemble.

Thanks for your help in advance

2

Hello,
Could it be the flying ice cube effect? Are you using a Berendsen thermostat?
Simon

3

Hi Simon,

Thanks for your reply! I actually learned the umbrella sampling from you tutorials (It is really great.). The main changes are that I use the “fix NVT” command to control temperature (I think it should be Nose-Hoover thermostat) and I use another command “fix recenter” to control the center of mass of water located at the center of the simulation box, between two vacuum layers.

I guess the reason should be the “fix recenter” command. However, without this command the water may shift the location when we move the cation and I could not sampling the cation from the bulk water to the vacuum-water interface. Could you please give me some hints about how to address this issue?

Thanks,
Wenqian

4

Hi Wenqian,

Have you considered cancelling a momentum using fix momentum? I would place a fix momentum at least during the equilibration stage of the simulation, and if possible remove it during the production run, or keep it but with a very low frequency (large value of N), so that it impacts your dynamics as little as possible.

Best,
Simon

5

Hi Simon,

Thanks for your help. I will try this method first.

Wenqian