Python code for performing Bragg Interferometry, a data analysis technique for 4D-STEM of moiré materials as described in Refs [1,2]. Implemented for and used in Ref [2]. Some further considerations and a more detailed description of the method are at [3].

Please see the following manuscripts for a description of the method.

1: Nathanael P. Kazmierczak* & Madeline Van Winkle*, et al. Nature Materials 20.7 (2021): 956-963.

2: Madeline Van Winkle* & Isaac M. Craig*, et al. Nat Commun 14, 2989 (2023). (In particular, the SI of this work described many details of this codebase)

3: Isaac M. Craig, et al. J. Appl. Phys. 136, 074301 (2024). (Has some additional considerations and numerical details)

The code can also provide more qualitative color-coded virtual dark fields as seen in Nature Nanotechnology (2024): 1-7 and Nature Materials 23.3 (2024): 323-330. Although precise virtual aperture definition for materials with multiple moiré patterns is not yet automated.

Please contact me with any questions, issues, or requests. Isaac M. Craig - craigi[at]berkeley[dot]edu

format your 4d-stem data (gatan dm4 and/or py4dstem formatted h5 accepted) in accordance with the example input directory. Then place a Diffraction_SI.dm4 or dp.h5 in each folder. You can then run the code from the driver script "main.py" which will guide you through the analysis workflow, asking for user input when needed. All processed data and images will be saved in the corresponding data/Material_Name/dsN directories (where N is the dataset number).

An example of a processed dataset is given in data/MoS2_P/ds3. A raw Diffraction_SI.dm4 and processed diskset.pkl files used in Ref [2] are available on https://zenodo.org/record/7865970

If you see a runtime error saying main thred is not in the main loop, its from matplotlib's backend not being thread safe. Please adjust the backend in the import statements. I had included matplotlib.use('Qt5Agg') when running in serial and showing plots with user input and matplotlib.use('Agg') when running in parallel and saving plots. Can be architecture specfic.