MMA Package for ellipsometry data import
Source code and installer located at https://github.com/lllamnyp/BIGParserAndMerger/
Simply extract this repository to a separate folder and run install.bat. Correct operation requires that the working directory matches the location of the script. If for some reason this is not the case on your machine, open a terminal (Win + R, cmd), navigate to the install script (cd /path/to/install/bat/) and run the script from the command line.
Alternatively, you may download and run only install_from_hub.bat which automatically (at the time of writing, 07.08.2018) downloads the latest files from GitHub.
Both scripts will overwrite any existing installation!
OPTIONAL: The directory parser functionality (which is written in C and compiled for Win7x64) will not work. The source code for scandir.exe is available in the package and should not be difficult to modify. The Run commands inside the package file (BIGParserAndMerger.m) are likewise written with Windows filesystems in mind. The rest of the package is, to the best of my knowledge, platform-independent.
On Windows the necessary package files (init.m, BIGParserAndMerger.m, and scandir.exe) are placed into %appdata%\Mathematica\Applications\BIGParserAndMerge\. For a different OS consult the Mathematica documentation for the appropriate place to place user packages.
Open Mathematica (v10 required) and run
<< BIGParserAndMerger`The package will take some time to load as many functions are being compiled to C. If a C compiler is not present on the system, warning messages will be issued and the kernel will revert to using the Wolfram Virtual Machine. This causes a slight hit to the performance, but nothing more serious.
It is sufficient to download just 'BIGParserAndMerger.m' and execute
Get["/path/to/BIGParserAndMerger.m"]The scandir functionality, if it has also been placed in the same location, may or may not suffer (untested).
Get["https://raw.githubusercontent.com/lllamnyp/BIGParserAndMerger/master/BIGParserAndMerger.m"]Should also work. scandir functionality will certainly not work.
After loading the package into memory by one of the methods from the previous section, run
MergingInterface[]The following window shall appear
The program was designed with the two types of experimental setups in mind: the infrared Fourier-transform ellipsometers and the near-IR to deep-UV Woolam ellipsometer. Each setup can operate in several spectral ranges. Here you must select the appropriate number of ranges (sets of measurements) to use.
For example, you have a measurement in the far-IR range, a measurement in the mid-IR range with a compensator that works well in the lower end of the mid-IR spectrum, a measurement in the mid-IR range without a compensator that works well in the higher end of the spectrum, and, finally, a measurement in the near-IR range. All are done on the Fourier-transform ellipsometers. That gives 4 FTIR .big file ranges.
On top of that you may have also performed measurements on the Woolam setup, say one with the 0.5 -- 4.0 eV optical fiber, and another with the 0.75 -- 6.5 eV optical fiber. This gives 2 Woolam ranges.
After setting the sliders to the appropriate positions, you may hit "Next" to specify the data files.
Hit the purple "Browse" buttons to bring up the following dialog:
and select one or more .big-files (or one Woollam data file). If multiple .big-files are being selected, it is expected, that they correspond to the same measurement, but for different polarizer azimuths.
Alternatively, you may hit the white "Browse better" button to bring up the improved file selection dialog which uses the scandir.exe functionality. scandir.exe quickly parses all of the .big files in a given folder using much faster C-code to bring up metadata for each .big-file:
Navigate to the desired folder using the left bar and click on the folder name (not the arrow). This will make scandir.exe run, check if the directory has been updated since the last access, cache its contents if that is the case, then load that from the cache and display the contents. Click on the column headings to sort by that column. Click on the filenames to add them to the list of files to load. Click on the appearing filenames in the rightmost bar to remove them from the selection.
After selecting files for each spectral range you may proceed, or you may also configure the compensator. The radio buttons' functionality is not implemented. Leave all buttons at "None/No Ref.". The compensator may be simulated as a constant offset for Psi and Delta across any one given spectral range. In this case, do not select any files for the compensator. Alternatively, a compensator's spectrum may be loaded from an .epd-file, which is basically a three-column data file (wavenumber, psi, delta), or from one or more .big-files generated by measuring the compensator in the straight-through configuration. Use the rightmost "Browse" buttons to select files. Afterwards hit "Next >>".
You will be taken to this dialog:
Here the errors in the polarizer and analyzer zeroes can be inputted manually for each FTIR spectral range. Additionally, the "Find P0/A0" buttons can automatically find the best parameters.
The depolarization degree field accepts either a real value in [0, 1) or an anonymous function such as
0.01 * (700 - #) / 700 &In this case, the depolarization is calculated as a function of the wavenumber in cm-1. The above example would correspond to linearly increasing depolarization from 0 at 700 cm-1 to 0.01 at 0 cm-1. # is standard Mathematica syntax to represent the argument of the function (i.e., the wavenumber).
The second part of this dialog contains the already-described P0/A0 parameters for the compensator, but also an option to "Invert delta" (i.e., does adding a compensator push Delta to higher or lower values). Finally, Delta may or may not be coerced to [-Pi, 0]. In a standard measurement it is impossible to determine the sign of Delta, but if it is known, that a compensator is used, it may be deduced, that Delta is positive, rather than negative. If the box is left unchecked, such deductions will be made.
After setting everything up, hit next to load all spectra and display them.
The experimental results of an ellipsometric measurement can be characterized with a pair of real numbers at each wavelength. One may choose to show the ellipsometric angles Psi and Delta in degrees, the dielectric function Epsilon (dimensionless real and imaginary parts), the optical conductivity Sigma (real and imaginary inverse Ohm-centimeters), or calculated normal-incidence reflectivity (intensity as a percentage and phase shift in degrees). This is selected by the buttons next to the "Y values" label.
For convenience, the X-axis can be shown in different units, corresponding to energy (cm-1, eV, THz, angular frequency), corresponding to wavelength (micrometers) or to the logarithm of wavenumbers in cm-1. One can also explicitly plot in a log-linear scale, but performance-wise choosing energy as Log(wn) gives essentially the same result as choosing a Log-Linear plot, but is drawn much faster.
Ideally, all measurements in the regions of overlap should match (except for Psi-Delta, since this value is also a function of the angle of incidence), however because of various factors (noise, imprecise calibration, etc) this might not be the case. The multiple sliders at the top of the dialog allow to take various corrections into account.
The sliders P0/A0 allow to account for an inaccurate calibration of the polarizer/analyzer zero azimuths, but in case of a good measurement done at several different polarizer azimuths the correction for the measurement at one polarizer angle will be opposite to the correction of the measurement at the opposite polarizer angle and the net effect will be quite small. The Woollam setup does such calibrations internally, so these sliders are not available for Woollam spectral ranges.
If the angle of incidence isn't determined very accurately, it can cause significant errors. In this case adjust the dTheta slider, aiming to match the adjusted spectrum to one, where you are confident about the correct angle of incidence.
The left/right sliders are used to discard data from edges of the spectral range, where there is more noise.
The Comp-Psi/Comp-Delta sliders are used to simulate a compensator. If no simulation is required, set Comp-Psi to 45 degrees and Comp-Delta to 0. If you have used a compensator that is known to have a roughly constant Psi of 42 degrees and Delta of e.g. 70, but are not using a direct measurement of it, but rather want to simulate its effects, you can set Comp-Psi to 42 and Comp-Delta to one of the following values: -180+70=110, -70, +70, +180-70=+110. The correct choice can be made by knowing approximately the expected properties of the sample and/or having a reference measurement without a compensator.
To focus on certain parts of the plot adjust the numbers in the display limits field.
Once you are satisfied with the final spectrum, hit next to move to the export dialog.
Here the program builds a single continuous function from the multiple datasets gradually transitioning between adjacent datasets in the region of their overlap by a Cos^2 interpolation. The effective angle of incidence is accordingly interpolated. The resulting data is shown in two log-linear plots.
Using the drop-down menu you may select the units of energy to use in the export, the data-columns to include (by the checkmarks), whether or not to include metadata about the exported data (this includes corrections, calibration info, filenames, etc). Select a directory with the "Select directory" button, type in a filename (without the extension) in the filename field and hit one of the export buttons. The extension will be automatically selected according to the export format.
Warning! The program will not prompt you if the file already exists! It will be quietly overwritten!
The program exports in its native format (.bps) which includes the full metadata if chosen and the desired columns. This is a human-readable file and is handy for reuse in programs such as Origin.
It also exports as two two-column data files (energy vs e1/e2), handy for use with WASF, which can only import spectra in this basic format.
It also exports as if this was a measurement done by the Woollam setup. In this case, it does not interpolate the separate spectral ranges, instead it interprets them as measurements done at as many different angles of incidence, as there are spectra. This is convenient if you want to attempt post-processing in WVASE32. This is done by "Export VASE" next to the ".dat" extension.
Finally it can export a ".mat" file, also compatible with WVASE32. In this case you will get an appropriately formatted material datafile with three columns (energy, epsilon1, epsilon2).
Finally, you may save the state of the program by clicking on the "Save Project" button, which is present in the file selection dialog and the adjustment tab. This will save the state of the program exactly as it is where you clicked this button. Helpful to not have to load all the files all over again and then have to input the various corrections.