I’m trying to repeat the results of other paper for simulation of water in CNT. I know that my question is not related to the software. But I would appreciate if you can help me to figure out the problem of my simulation.
I attached the results for density profile of my simulation and the output parameters during minimization and running the simulation. My results show a big jump at the center of CNT; however, the results of the paper show one smaller jump a little bit far from the center. I am wondering if you can help me to know the reason of high jump exactly at the center in my case?
During the minimization, the temperature reaches to zero and after running the simulation goes to 100 K (The temp of water was set at 298K). The pressure reaches to 35000 atmosphere during minimization and goes to 31000 atmosphere during running the simulation. I also ran another simulation for water without CNT which I believe gave me reasonable results. (I plot them together with results of water in CNT).
I have used moltemplate to build the molecule structures of the system. Andrew mentioned that moltemplate doesn’t connect the carbon atoms together with bonds. Instead, we can hold the structures rigid by not integrating their equation of motions as I applied in the input file. Can this cause big problem for the simulation?
your script is complicated. I suggest you first
simulate flexible water (no SHAKE, no CNT),
Verify you can thermostat it at 300K with
reasonable pressures (correct density, etc).
Then turn on SHAKE and do the same,
Then add a frozen CNT,
…to prevent LAMMPS from attempting to calculate interactions between the immobilized carbon atoms (in the “Cgraphene” group). This will solve the problem you are having with the pressure and energy. (See attached file “run.in.nvt”)
It looks like you based your input script on the “run.in.nvt” file distributed in the moltemplate examples. When I tested that file, I only cared if the atoms were moving in a reasonable/plausible way. It didn’t occur to me to check the pressures or energies. I apologize for that. Thank you for your careful checking.
I will update all of the moltemplate examples with immobile atoms to consistently use “neigh_modify exclude”. (I will also update run.in.npt which is also currently not working.) I have a few moltemplate updates coming soon.
-Andrew
P.S.
If I remember correctly, the Lennard-Jones parameters for the carbon atoms (taken from the Saito,…Dreselhaus Chem Phys Lett 2001 paper) were chosen to to create reasonable water-carbon interactions. (At least that was the goal. However you mentioned you were having a problem with water density inside the tube.) In any case, interactions between the carbon atoms using these forces will definitely be unrealistic (and extremely repulsive, apparently). These large repulsive interactions overwhelm the energy and pressure although they don’t effect the movement of the water unless, because we immobilized the carbon atoms by omitting them from the group of “mobile” atoms submitted to “fix nvt” or “fix nve”, as you are doing. (Incidentally, there is no need need fix setforce to immobilize the carbon atoms.)