Lammps fix cha/reg with LJ units give unexpected number of ions

Hi!
I am running charge regulation with LJ units (unit length = 1nm, unit energy = 1kT, unit mass = 23g/mol, and resulted unit time ~ 0.1ns). I have box of 200 x 200 x 200 with ~14,000 groups that can be negatively charged and set the pIp = 1. When the number of charged groups converges, the system has ~50,000 positive ions (counter ions + extra ions) and ~40,000 negative ions.
I calculate the concentration of the ions to be ~0.008M instead of 0.1M. Even though the number of ions are still increasing but rather slowly.
Is there anything I am missing? The initial configuration, input and output log are attached below.

Best,
Yang

slurm-8510289.out (11.3 KB)
in.0.1M_chareg (6.7 KB)
2fiber.data (608.8 KB)

Since you do have the equivalence of reduced unit settings, do you get the same behavior when converting to “units real”?

Thank you for your reply!
I am a bit confused. Do you mean setting up a system with real units (2000 A x 2000 A x 2000 A box etc) and run the simulation for much longer (since in real units the time is in fs), then check the results? Or do you mean I just keep the configuration the same but just change the units from LJ to real?

Best,
Yang

You could set up the same system, just using real units instead of lj. There is no need to run “longer”, since you would set the time step to the exact same time you would use currently. Units are just units and you can convert between them without changing the actual interactions. 200 nm are the same distance as 2000 A.

That would be a different simulation.