This directory contains the necessary files to generate a pure ionic liquid 
trajectory file with LAMMPS and perform Widom insertions of 
difluoromethane (R-32) and pentafluoroethane (R-125) on the trajectory 
frames.  The chosen ionic liquid for this example is 
1-n-butyl-3-methylimidazolium tetrafluoroborate (BMIM-BF4).

First, run LAMMPS using lmp.inp as the input file.
For the corresponding article, the trajectory was output in the custom 
dump style, but this input file will result in the trajectory being written 
in the portable binary XTC format, which uses less storage space and doesn't 
need to be converted before use, unlike custom dump files. If you want to use 
custom style dump files instead, the required input file lines are provided 
but commented out. Originally, the snapshot at production timestep 0 was 
included in the dump file but manually removed later.  The delay option in 
dump_modify is now used to omit that snapshot so the snapshots written are the
same as those used for the Widom insertion simulations.



Prior to running the Widom insertion simulation in Cassandra, fragment libraries
must be generated using the library_setup.py script included with Cassandra:

library_setup.py [path to Cassandra] cas_widom.inp cassandra_molecules/BMIM.pdb cassandra_molecules/BF4.pdb cassandra_molecules/r32_36.pdb cassandra_molecules/r125_100_new.pdb


Replace [path to Cassandra] with the path to your Cassandra executable.  Simply having 
the Cassandra executable on your shell search path and including only the file name of your
Cassandra executable will NOT work, at least with the current version at time of writing. 
If library_setup.py is not on your path, write the path to it.

To run the Widom insertion simulation with Umax = 708 and estimate appropriate smaller Umax values 
with the same parameters and settings as for the article, execute the following command:

cassandra.exe cas_widom.inp

Note that your Cassandra executable may not be named cassandra.exe, and in fact probably is not, 
so replace cassandra.exe with the name of the Cassandra executable you want to use 
if it is on your path, or replace it with the path your Cassandra executable if it is not 
on your path. Frame <w> values are written to cas_widom.out.spec3.wprp for R-32 and 
cas_widom.out.spec4.wprp for R-125.  Note that w is referred to as widom_var in the output files. 
The W_ij arrays for R-32 and R-125 discussed in the article are written to cas_widom.out.spec3.wprp2 
and cas_widom.out.spec4.wprp2, respectively.  Shifted chemical potential results are written near the 
end of the log file, along with overlap information and Umax recommendations. Extra results that could 
be used to manually estimate recommended Umax with a <w> relative tolerance other than the 
default (10^-10) are written to cas_widom.out.spec3.emax and cas_widom.out.spec4.emax as described 
in the article.

Despite using the same random number seeds as the corresponding simulation performed for the article,
this input file will not give exactly identical results because the simulation was performed with 
64 threads.  Even if you were to run this with the same number of threads and same version of Cassanda, 
the results would not be identical because there is non-deterministic parallel task scheduling.
Repeat simulations will only be deterministically identical if only a single thread is used.

To compute Henry's law constants, you need the <w> of R-32 and R-125 in an empty box with known volume, 
as discussed in the article.  To compute these values, run the following:

cassandra.exe empty_ideal_gas_widom.inp

Note that for the empty box simulation, species 1 is R-32 and species 2 is R-125, although they still use the 
fragment libraries in directories species3/ and species4/ because they were species 3 and species 4 in the 
cas_widom simulation.

The Python script used to generate the GCMC P(µ') data required to make the P(x) isotherm is provided in 
the gcmc/ directory, with file paths changed for the directory arrangement here.


Python scripts and notebooks for analysis are included in the analysis/ directory.  They are only partially 
cleaned and are not necessarily ready-to-use, considering parts of them are specific to the files and 
directory arrangement used and some of the analysis requires other simulations besides the example. 
However, by following the details in the article and using the documentation and example provided
for reference, you should be able to replicate the LAMMPS and Cassandra simulations described in the 
article.