Problem with MeOH using OPLS-AA

Dear LAMMPS users,

I have a problem simulating a box of methanol with the OPLS-AA force field and cannot solve it, so I’m asking you now.
The MeOH molecules behave weird in my simulation, resulting in a crash. I therefore simplified the system and tried to run a simulation with methane, which works.
So I think I have a mistake in my data file. Unfortunately I cannot find it and therefore I wanted to ask if someone can provide me his/her data file from a simulation of a simple alcohol (meoh/etoh/isopropanol) with the OPLS force field (with inserted parameters), so I can compare it with my own.

Paul

PS:

I created the data file with a tcl-script, similar to the ones of A. Kohlmeyer’s topotools tutorial. The parameters were taken from

http://dasher.wustl.edu/ffe/distribution/params/oplsaa.prm

The input file was also created according to the examples in A. Kohlmeyer’s tutorial.

Paul,

Check the link below:
ftp://ftp.ill.fr/pub/cs/workshops/mdanse/tutorials/md_tutorials/

Sec 4.3 of the tutorial covers the ethanol case and in the ethanol dir there is a full subdir for lammps.

Carlos

PS: Found this by browsing the web thus I do not assume responsibility for the accuracy of the data/scripts. Yet, looks pretty well put together…

LAMMPS-1Feb14

Hi Carlos,

thanks for the link. The examples helped me to make it work!
Although, I do not understand why. I still think, though, that the problem lies within the data file.

Let me explain:
The only difference between the example in the link and my own file is the way the files are set up.
I set up 2 cases (all files in the appendix)

  • One case that follows the example from your link, i.e. I have 4 files with data, coeffs, initialization and simulation (the hybrid keyword does not change the calculation, I checked)
  • The other case is my own ‘standard’ procedure, where I have one input file with the initialization and simulation parameters and the other being the data file with the coeffs and everything else.

In both cases the coefficients and simulation parameters are the same, but the example-case works and my own procedure does not (the calculation crashes after a bit).
I did the same with methanol and have the same problem.
So basically I have no idea what is going on. My ‘standard’ way of setting up the simulation does not work and I have to write the coefficients in a separate file.
Do you or someone else have a clue about why that is the case?

Paul

standard-case.tar (10 KB)

example-case.tar (20 KB)

LAMMPS-1Feb14

Hi Carlos,

thanks for the link. The examples helped me to make it work!
Although, I do not understand why. I still think, though, that the problem
lies within the data file.

​yes, it does and it is in the Pair Coeffs​ section.
you "standard" file uses the same syntax as in the pair_coeff command, but
that is not correct.
Pair Coeffs only accepts a single atom type designator. if you want to
specify i,j pairs you have to use PairIJ Coeffs (but then also provide the
mixed terms).

this can be easily validated by turning off the useless hybrid stuff and
then compare the output from write_data between both input decks and the
input with the written file.

axel.