P. F. Gostin et al., In Situ Synchrotron X‐Ray Diffraction Characterization of Corrosion Products of a Ti‐Based Metallic Glass for Implant Applications, Advanced Healthcare Materials, 2018, 7, 1800338 (https://doi.org/10.1002/adhm.201800338)
Free pre-publication full-text version: doc/manuscript_Gostin_2018.pdf
This project automatically:
- downloads raw data from a data repository on Zenodo
- processes the data (calibration, azimuthal integration etc.)
- plots figures and table
To repeat the analysis
git clone https://github.com/flaviu-gostin/xrd_analysis_workflow.gitcd xrd_analysis_workflow/sudo apt install virtualenvpython3 -m venv ./venvYou need Python3.8 for this projectsource venv/bin/activatepython3 -m pip install -r requirements.txtmake allDownloads 14.6 GB (but reserve 30 GB) and performs all the data processing and plotting.
You need to reserve more storage space as downloaded archives and split files will automatically be decompressed/concatenated. To delete the archives and split files:
make clean-data
You should find the results in results/
Please feel free to create an issue or a pull request on GitHub or email me.
- download raw 2D XRD image data from a Zenodo repository
- perform azimuthal integration on those images resulting in 1D XRD patterns using pyFAI
- determine peak position, calculate lattice spacing and write values in a table
- calculate diffraction patterns from structures using pymatgen
- plot measured and calculated diffraction patterns using matplotlib
Calibration uses the Python module pyFAI. This is listed in requirements.txt. When installing this module, a set of command line interfaces (CLIs) are also installed in bin/, e.g. pyFAI-calib. The recommended way to do the calibration is to use the pyFAI-calib CLI (there is also a GUI available for that). With those interfaces the user must click on several diffraction rings to help the program locate the rings. Another reason for using the CLI or the GUI is to be able to check visually that the refined rings match well enough the real rings of the calibrant. However, I intentionally avoided using either interface because clicks cannot be reproduced easily. Instead, I wrote a script which uses the refinement function in pyFAI and takes as arguments a long list of manually selected points. Thus, the calibration is guaranteed to be reproducible. Moreover, one can check that the manually selected points are located on diffraction rings.
The refined (calibrated) geometry is saved in a .poni file. This .poni file is the basis for the azimuthal integrator object used to integrate all the diffraction images. Thus, if the calibration is changed, most of the results might change. This will most likely be a minor change.