Hi lammps user,
I am performing lammps simulation, and I prepared my data file from material studio and then import it to lammps using msi2lmp.exe tool provided with lammps. I am using hybrid potential (Tersoff and PCFF) for my simulation. The simulation runs well without any error but when I implement NPT ensemble, the density is supposed to converge but it’s not converging. Does anyone help me with what wrong is happening with my simulations? The input file and log files are attached with this mail.
min.txt (18 KB)
1.1.in (4.27 KB)
For a sufficiently recent version of LAMMPS the conversion of tersoff potential files from metal to real units is not required, if the file contains the required metadata to indicate the units it was parameterized for (like the files bundled with LAMMPS do). Then LAMMPS can do the conversion automatically when reading the potential parameters from the file.
So what density do you get? what is the reference you are comparing to?
and what do you expect for the combined system instead? and how close do you get to it?
Yes, when I run the simulation individually they attend the desired density. I also check the tersoff potential file which I converted to real units. The results matched with the metal unit results (using the original tersoff file).
It is difficult to debug hybrid systems, thus you have to validate each component individually?
Have you run simulations of each part separately? Have those been able to reach the desired density?
I prepare the model from material studio and then I imported it to lammps. I check the density of the combined system in the material studio and the density of the combined system is 0.996 g/cm3. So, in lammps also I should get the density around this just by using npt ensemble at the same temperature and pressure, but I am not getting. So I used fix deform to compress it a little and then again using npt but the density remains unaffected when I use npt only.
I also check the density of the individual systems on lammps and compared it with the density obtained from material studio software. The individual systems densities obtained from lammps are close to the results obtained from the material studio software.
sorry, but that is not a validation. you only get the density for what you input. my chemical intuition would expect a density significantly lower than what you say.
furthermore I have asked for the equilibrated density of your solvent as determined by a simulation, which you claim you have done. That is most certainly different from that of the combined system.
what if the model as built by materials studio is already flawed? what if the force field is not suitable?
this needs to be validated as well.
there are other flaws in your calculation. the timestep is unsuitable for the elements you have, you might get better performance with a different long-range solver, and running this small a system with as many as 62 MPI ranks cannot be efficient.
we have been over the topic of requiring proper advising and tutoring before. this just manifests this again.
Thank you.
The problem is solved. The only error was: I use iso and you mentioned it aniso system in the other mail. When I use the aniso system the density converges and matches the result obtained from material studio.
Sorry for my mistake, I didn't see the units were real and assumed they were metal. My fault