Reasearch project in Bari, spring 2013

During March-May of 2013 I'm working on a research project in Bari, Italy, creating a Monte-Carlo simulation of electron beams in a collisional medium. This is the code.

The code is developed in Fortran, licensed under a modified version of the MIT license - see LICENSE.md.

Contents of this guide

• Obtaining the source code
   

• Building and running the program  

• Program input and output

  • Program input
    • The structure of input.in
    • The structure of the data files
  • Program output

Obtaining the source code

The source code for this project is available on github.com, but chances are you're already well aware of that, since that is also the best place to be reading this readme. Don't forget to read the license information ;)

Building and running the program

The Github repository has three branches: master, linux-build and windows-build. master has everything you absolutely need to build the program - the sources and the makefile - and all development of the project happens here. The *-build branches also include some additional tools that make building and/or using the program a little easier, and when something that changes the program behavior is commited to master, changes are merged into the *-build branches as well.

• linux-build includes a bunch of scripts for bash, awk and gnuplot that help with running and post-processing the program. The preferred way of building and running the program in a Linux-based system is

    $ make
    $ ./simulate.sh
  

  To plot the results, one might run ./scripts/plotall.sh which creates a number of plots in the output directory of the last run.

• windows-build does not include the scripts for linux, but instead has a couple of .bat scripts. The preferred way of building and running the simulation on Windows is to run build.bat followed by simulate.bat - this can be done by double-clicking them from Windows Explorer. (If you get a warning message about executing a potentially dangerous executable from an unknown source, just click "Run anyway".)

Prerequisites

In order to build the program, you need to have both Make and gfortran installed on your computer. These are free tools, available both for linux and Windows.

Program input and output

Program input

The program reads a number of parameters from stdin. In the default setup, this is redirected to the file input.in, but it's absolutely possible to run enter the information manually as long as the given information conforms with the format.
The program then also reads data for collisional cross-sections from a number of other files.

The structure of input.in

1. First, a few simple parameters for the simulation:

2. The number of simulated particles at t=0 (e.g. 1e6)

3. The time, in seconds, for which to run the simulation (e.g. 10e-9 or 1e-8 for 10 ns)

4. The time step, in seconds. EEDF data is output for each time step (real, e.g. 1e-9 for data every nanosecond), and the rate coefficients are evaluated at each time step; thus this is the resolution of the rate coefficients used for calculating populations densities of various excitaitons.

5. The initial energy of the primaries, in electronvolts (1e3 for 1 keV, or 200 for 200 eV)

6. À mathematical expression describing the electron density as a function of time. The function supports syntax as described at the bottom of this page, with the only available variable being t for the time in seconds.

7. Next, information about the collisional processes which should be considered.
   Note: The first of these lines determines how the rest of the file is interpreted by the program, but no verification is done at execution time. If the input file is not consistent, the program might throw an error or return incorrect results - the behavior is undefined.

8. The number of collisional processes to be considered, and the number of species for which population densities should be calculated (e.g. 5 2)

9. The parameter A for each species for which populations should be calculated, in s^-1 (e.g. 2.74e7 1.53e7)

10. The parameter Q for each species for which populations should be calculated, in cm³s^-1 (e.g. 3.67e-11 8.84e-10)

11. Finally, a list of the data files from which data for the differential cross-sections is read, one file per line with paths relative to the executable, enclosed in quotes.
    Note: The ordering is important! The files should be listed in an order such that the first n files correspond to the processes for which populations should be calculated, given in the same order as the parameters A and Q above, and other processes follow.

12. The file must end with (at least) one empty line

Putting it together, this is an example of a valid input file which considers three collisional processes and calculates populations of the species here labelled N2B and N2C, using a unity electron density:

1e6
10e-9
2e-9
1e3
1.0
3 2
2.74e7 1.53e7
3.67e-11 8.84e-10
"data/N2C.dat"
"data/N2B.dat"
"data/N2.dat" 

The structure of the data files

The first line of the data file should contain three whitespace-separated values. The first two the lower and upper boundaries of the energy range for which this collisional process should be considered - the cross-section will be considered 0 outside this range. The third value is a 1 if the process spawns a secondary electron, and 0 otherwise.

The rest of the data files should contain two values per line; the first value is the energy in electronvolts, and the second is the differential cross-section in cm².

The first few lines of an example datafile looks like this:

 15.581    1e3 1
 15.58     0.000E-16
 16.00     0.013E-16
 16.50     0.030E-16

For this file, the cross-section will be interpolated between the given data points between 15.581 and 1000 eV. Each time a collision of this type occurs, a secondary is produced (signified by the 1 at the end of the first line) which will share the available energy with the primary.

Program output

The data is output to three different files:

• eedf.dat contains data for the electron energy distribution function. It has three columns, listing the time, energy and value of the eedf. The time is given in seconds, the energy in eV, and the eedf in eV^-1.

  The following Matlab code plots the eedf at the time 10 ns:

    data = load('eedf.dat');
    rows = data(:,1)==10e-9; % a logical matrix which selects the correct rows
    eedf = data(rows, 2:3);
    plot(eedf(:,1), eedf(:,2))
  

• rate.dat contains data for the rate coefficients for each species in the simulation, with the time in seconds the first column and the rate coefficients in cm³s^-1 in the following columns, in the same order as in the input file.

  The following Matlab script plots the ratio of rate coefficients for species N2B/N2C, given the input file above:

    data = load('rate.dat');
    times = data(:,1);
    ratios = data(:,3)./data(:,2) % second column is N2C, since that was first in input.in
    plot(times, ratios);
  

• pops.dat contains the data for the populations for each species, with the time in seconds in the first column and the population density in cm^-3 in the following columns, in the same order as in the input file.

  The following Matlab script plots the population of the N2B species versus time:

    data = load('pops.dat')
    plot(data(:,1), data(:,3))