Hi dear LAMMPS users and developers
I am doing a coarse-grained simulation about a toluene molecule using lammps ( LAMMPS 24 Dec 2020). I used the information in this paper and supporting information. (https://doi.org/10.1021/acs.jpcb.6b05925) Based on this paper. Here is the some information from paper:
They employed a rigid model for toluene comprising three SC4 beads (uncharged) held in a rigid isosceles triangle with interbead distances of 0.28,0.28, and 0.23 nm. Lennard-Jones interactions were switched smoothly to zero in the range 0.9-1.4 nm and . Following the standard Martini treatment of electrostatics, Coulomb interactions were shifted to zero in the range 0.0-1.4 nm and they employed a relative dielectric constant of 2.5 corresponding to a low-dielectric apolar toluene. The rigid bonds in the toluene
solvent were fixed to their equilibrium values using the LINCS algorithm. Simulations were conducted in the NPT ensemble using a Nose-́ Hoover thermostat65 and an isotropic Parrinello-
Rahman barostat to maintain pressure at P = 1 bar. The compressibility of the system was set to 9.0 × 10-5 bar-1 in toluene. Newton’s equations of motion were integrated using the leapfrog
algorithm with time step of 20 fs .
I convert the coefficients in the paper to KJ/mol and deleted the coulomb interactions, ( SC4 is uncharged) but the density is about 1.2 gr/cm^3 which is far farm 0.88 gr/cm^3.
Did I miss something in the simulation procedure? Thank you very much in advanced for your kind help.
ps : I attached the moltemplate and lammps script.
Regards
Sepide
system.in.settings (296 Bytes)
system.lt (344 Bytes)
toluene.lt (1.51 KB)
toluene.txt (2.9 KB)
This is near impossible to answer by just looking at the files and requires investing a significant amount of time in running simulations and testing.
There are four possible outcomes (in order of probability):
- you made a mistake translating/converting some parameters or settings or creating the geometries
- the result is correct for the given model
- there is a difference in the potential functions for fixes implemented in LAMMPS versus those in Gromacs
- there is an error in the published parameters
You will have to investigate that yourself. The most obvious (to me) step to do that would be to try an set up the same simulation in gromacs and compare the results.
axel.
I think I solve my problem, so I write back here probably it can help somebody.
In the new version of martini (version 3) benzene (also toluene) is models as 3 beads of TC4. This new
model reduce the density and the error is less than 5 percent.
Thank you very much again Professor Kohlmeyer. I learned a lot from you through the mailing list. 
Regards
Sepide
Dear Professor Kohlmeyer
Thank you very much for your kind help.
Regards
Sepide
This is near impossible to answer by just looking at the files and requires investing a significant amount of time in running simulations and testing.
There are four possible outcomes (in order of probability):
- you made a mistake translating/converting some parameters or settings or creating the geometries
- the result is correct for the given model
- there is a difference in the potential functions for fixes implemented in LAMMPS versus those in Gromacs
- there is an error in the published parameters
You will have to investigate that yourself. The most obvious (to me) step to do that would be to try an set up the same simulation in gromacs and compare the results.
axel.