Current version: v4.06.01
Recent updates: Line De-blender and Single Slab Fit functions.
If you are running a previous version, make sure to update to the latest one (see below for instructions)!
Please cite: Jellison et al. 2024 (paper), Johnson et al. 2024 (code), Banzatti et al. 2025 (if you use the "MIRI_general" line list), LMFIT (if you use any of the fitting functions in iSLAT), plus other packages as appropriate.
If you are running iSLAT on an Apple silicon processor, please read "Caveats" below
iSLAT (the interactive Spectral-Line Analysis Tool) is a python package that provides an interactive interface for the visualization, exploration, and analysis of molecular spectra. Synthetic spectra are made using a simple slab model written by Simon Bruderer and originally described in Banzatti et al. 2012; the code uses molecular data from HITRAN (Gordon et al. 2022). iSLAT has been developed and currently tested on spectra at infrared wavelengths as observed at different resolving powers (R = 700-90,000) with: JWST-MIRI, Spitzer-IRS, VLT-CRIRES, IRTF-ISHELL. Examples of these spectra are included for users to practice with the tool functionalities across a range of resolving powers. iSLAT has been built as a flexible tool that should work with one-dimensional molecular spectra observed with other instruments too, provided some requirements are met (see below). iSLAT is presented and described in Jellison et al. 2024. A screenshot of iSLAT's GUI is provided here:
We built iSLAT to make the analysis of infrared molecular spectra an accessible, enjoyable, and rewarding experience for all, including students and junior researchers who are starting in the field today. We believe iSLAT will be useful both to students and to professional researchers, and we will continue to support it and expand it with care for all users. If you find a bug, please be patient and let us know; we want to fix it as much as you do! And if you have ideas to improve the tool, we'd love to hear.
Questions? Feedback? Contributions? Submit an issue, a pull request, or email us at spexodisks@gmail.com
Make sure you have git and pip installed, then run these terminal commands sequentially from within the folder where you wish your local copy of iSLAT to be:
git clone https://github.com/spexod/iSLAT
cd iSLAT
pip install -r requirements.txt
To launch iSLAT, simply type (if something goes wrong, you might just need to update some of your python packages):
cd iSLAT
python iSLAT.py
Remember to update the repository from GitHub fro 8000 m time to time (check at the top of the page to see if you're running an older version); from your local iSLAT folder, type on terminal:
git pull https://github.com/spexod/iSLAT
iSLAT requires a continuum-subtracted spectrum as input data file in csv format, with a wavelength array in μm (assumed to be in a column called "wave") and flux array in Jy (assumed to be called "flux"). The step of estimating and subtracting the continuum can be done e.g. with this tool: ctool. iSLAT's outputs are txt or csv files that save model parameters, spectra, or line lists as defined by users (see more below). These output files are stored in the folders /SAVES, /LINESAVES, and /MODELS. The folder /LINELISTS instead includes some curated line lists provided to users (see Jellison et al. 2024).
At first launch by the user, iSLAT will download from HITRAN the data for a default list of molecules: H2, H2O, CO, CO2, CH4, HCN, NH3, OH, C2H2, C2H4, C2H6, C4H2, HCN, HC3N and some of their isotopologues. These datafiles are stored into the folder "HITRANdata" and are available to load and use in iSLAT (see below). Any additional molecule or isotopologue can be downloaded from HITRAN using the "HITRAN query" function in iSLAT. The current HITRAN data release as of 2024 is described in Gordon et al. 2022. As of version 4.05, we have included two line lists provided in Arabhavi et al. 2024 and the HITEMP line lists for CO and H2O; if you use these, please cite the respective references. A new HITRAN release (2024) is expected in 2025.
(Definitions and instructions are also available by hovering over GUI elements.)
General note for all parameters in the GUI: every time you update any value, hit "return/enter" on your keyboard to apply the new value to the model. This has to be done for each parameter individually, e.g. when submitting a new temperature for a given molecule, or updating the distance for a given target. A confirmation will appear in the message box in iSLAT.
- Molecular model parameters: temperature is in K, radius in au (this is the radius of the equivalent emitting area, not the orbital radius of the emission), column density in cm-2. You can control which molecules are ready for use in the GUI with the buttons: "Default Molecules" and "Add Molecule" (see more below on these functions), and the "Del." button to remove them. You can also select each molecule color by using the "Color" button.
- Other parameters:
- Plot start/range (just for the current plot) and min/max wavelength (the spectral range for computing the slab models) are in μm
- Distance is in pc
- Stellar RV is Heliocentric and in km/s, and will shift the observed spectrum (not the model)
- line FWHM is in km/s, and is used for convolution of the model to match the observed line widths (which can be either set by the instrument resolving power, or by the gas kinematics if the resolving power is high enough to resolve lines)
- line broadening is in km/s and it is the FWHM of the intrinsic line broadening due to thermal motion or turbulence
- line separation is in μm, and is used to identify isolated lines in the model for the molecule selected in the drop-down menu
- FWHM and centroid tol. are in the same units as their parameters, and they control the tolerance range used in the de-blended fit (see below).
(Definitions and instructions are also available by hovering over GUI elements.)
- HITRAN query: opens up a window where you can select and dowload any molecules or isotopologues available in HITRAN; you need to run this first if you wish to use any molecule that is not part of the default list above.
- Default molecules: load the default list of molecules for the GUI; this list includes: H2O, OH, HCN, CO2, C2H2, CO.
- Add molecule: loads a molecule into the list of available molecules at the top left of the GUI; the new molecule must already be downloaded from HITRAN and stored in a .par file in the folder "HITRANdata"; the user can also load multiple times the same molecule, as long as different labels are assigned (for instance to simultaneously plot different temperature components of a same molecule).
- Save parameters: save in an output file the current model parameters (T, R, N) for each molecule, plus the distance, RV, FWHM, and broadening values set by users for any input spectrum; the output file will have the same name as the input observed spectrum plus "-molsave.csv" and will be stored in the folder iSLAT/SAVES.
- Load parameters: loads previously saved model parameters from output file created with "Save Parameters" from the folder iSLAT/SAVES; it will use the input spectrum file name to identify the correct save file, so if you change the input spectrum name make sure you update the save filename too.
- Export models: export specific or all model spectra in an output csv file in the folder iSLAT/MODELS.
- Selecting a line: by dragging a region in the top spectrum plot, the spectral region gets visualized in the bottom left plot with the addition of the individual transitions that dominate the emission; the strongest of these transitions is highlighted in orange and its properties are reported in the text box at the bottom left of the GUI. Starting with version 4.04, other lines can be selected interactively in this plot and the population diagram by clicking on them; this new feature allows users to explore and save different lines within the same selected range.
- Save line: save the currently selected line into an output csv file that will include all its line parameters; this function needs the definition of an output file in the box "Output Line Measurements" that will be saved in the LINESAVES folder.
- Fit line: fit selected line with a Gaussian function using LMFIT; full fit results are reported on terminal, and a selection in the GUI text box.
- Show saved lines: marks in the plot all lines saved into the file selected in "Input Line List" (this list can be one of those provided with iSLAT, or one you make using the "Save Line" function).
- Fit saved lines: this is the same as "Fit line" but it fits in one shot all lines defined in the "Input Line List" and will save their measurements in the LINESAVES folder in the file defined in the "Output Line Measurements" box. The output will also include rotation diagram values for each line, which you can use to make a rotation diagram like the one in the GUI, but with measured lines instead of a model.
- Show atomic lines: marks and labels atomic lines from the list provided in the folder LINELISTS
- Find single lines: identifies and marks isolated lines using the value in "Line separ." as their spectral separation.
- Molecule drop-down menu: select which molecule is considered for the zoomed-in plot, the rotation diagram, the "Find single lines" function, and the "Save line" function.
- Single slab fit: fit single slab model for molecule selected in the drop-down menu; it needs a line flux measurement file as "Input line list", which can be made by running the "Fit saved lines" function (see above).
- Line De-blender: extract individual transitions from blended feature selected in the Line Inspection Plot; it uses LMFIT and gaussian functions, where centroid and FWHM can be given a tolerance range in the GUI (the fit assumes line centroids from HITRAN for the selected molecule, and FWHM as fixed by the user in the FWHM box). The results are saved into /LINESAVES.
iSLAT's release includes some continuum-subtracted spectra of protoplanetary disks that users can use to get familiar with iSLAT and its applications across a range of resolving powers:
- one M-band spectrum from iSHELL with R~60,000 (FZ Tau) from Banzatti et al. 2023a
- two spectra from MIRI with R~2000-3000 (CI Tau and FZ Tau) from Banzatti et al. 2023b and Pontoppidan et al. 2024, respectively
- one spectrum from Spitzer-IRS with R~700 (CI Tau) from Banzatti et al. 2020 , reduced as described in Pontoppidan et al. 2010
The original spectra (not continuum-subtracted) are or will be available on spexodisks.com. Users who wish to use any of these spectra in their research work should use the original data and cite the original publication papers.
The simple slab model used in iSLAT includes line opacity and saturation with a curve-of-growth method, but does not account for mutual saturation of lines that overlap within the thermal/turbulent line broadening (see details in the appendix of Banzatti et al. 2012). This implies that some emission lines in iSLAT will look stronger than they should be, specifically where there are clusters of transitions that are partially or fully optically thick. Users that wish to model those lines correctly can use other available slab model codes: spectools-ir or iris.
If you use iSLAT on an Apple-silicon Mac (M1, M2, etc.), there is a known issue with tkinter where the GUI often seems to freeze and does not take a new input or click; the solution is to click and drag very slightly on any element you're working on (whether you're clicking a button or updating a parameter value), and the GUI will take your input. Hopefully this will be fixed in a new release of tkinter (this is not an issue of iSLAT).
It is hard to remember to acknowledge and cite all the tools we use in research, and often it is impossible to offer more than personal gratitude to the many people who have contributed to the tools and instruments we use. Science is a community endeavor that always builds "on the shoulders of giants", but these giants are by now a multitude and are most often unknown. They simply did good work and shared it, and left a heritage that became useful to others.
iSLAT's story is similar: it is the outcome of experience, passion, and complementary contributions of multiple people spanning more than 14 years, even if we focus on just the four main authors. We offer our gratitude for all the tools we have used to build iSLAT. If you find iSLAT useful for your research, and if you can remember doing it, please add its citation in your publications and oral/poster presentations. It's just a way to say "thank you". The main reference to cite is: Jellison et al. 2024