Argonne, IL


ATREX - integrated open source data analysis software for mineral and environmental sciences

ATREX (Advanced Tools for Research in Extreme Xtallography).

  • Principal investigators:
    • PI: Przemyslaw Dera (University of Hawaii)
    • Co-PI: Bin Chen (University of Hawaii), Lars Ehm (Stony Brook University)
  • GitHub repository: source code and software distributions.

Project summary

Synchrotron radiation user facilities are critical resources, which enable state of the art research and training in mineralogy, mineral physics and environmental science. Access to these facilities is very competitive (science-merit-based peer review process), and time allocated for experiments always very constrained. The most popular and fundamental type of experiments conducted at synchrotrons is X-ray diffraction, which produces information about the details of the crystal structure, symmetry, chemical bonding and density of minerals and other crystalline solids. Depending on the form of the sample, diffraction experiments can be conducted with single crystals, multiple crystals (e.g. very coarse powder), bulk powders, nano-powders and glasses. Modern synchrotron-based diffraction experiments are typically conducted with the use of area detectors, in which case the data is recorded as digital diffraction images. Conventional diffraction experiments (carried out at ambient conditions with good quality samples, typically of only one crystalline phase) usually produce data that can be analyzed with a combination of the currently available software. However, the situation is much more difficult with unconventional data, e.g. produced during experiments at non-ambient conditions, involving samples embedded in thin sections, or exploring the real time progress of chemical or physical processes. The unconventional experiments are extremely valuable in lab-scale modeling of technological (e.g. CO2 sequestration, new hydrogen storage material discovery, gas separation), as well as geological processes (e.g. deformation and phase transitions of mantle minerals), but introduce significant complexities into the data, often beyond the capabilities of available analysis software.

Technology used in synchrotron experiments evolves rapidly increasing the speed of the data collection and producing massive volumes of experimental data, posing new serious challenges for data analysis, particularly for unconventional experiments. The experiments are often decision-driven, and require at least partial real time data interpretation to provide the experimenter with the necessary information on how to proceed (e.g. monitoring a progress of sample transformation, or detecting a subtle discontinuity which may require a more careful investigation). Software offering such real time analysis capabilities is currently not available. Utilizing a combination of existing, well tested and widely used software components developed in the Interactive Data Language by the PI Dera, Python code developed by the co-PI Ehm at the Brookhaven National Lab and Bayesian statistics Matlab code by Project Collaborator Wolf at the University of Michigan we propose to create a new integrated multiplatform, open-source Python software package ATREX (Advanced Tools for Research in Extreme Xtallography), with unique capabilities to process diffraction image data from samples in all forms from glasses and melts, bulk powders, though coarse multi grains, to single crystals, which will support new data types produced by novel ultrafast X-ray imaging detectors, offer extensive automated serial processing capabilities for massive data sets and will allow real time data analysis for time-constrained decision-driven synchrotron experiments. ATREX will include database access capabilities utilizing the free American Mineralogist Crystal Structure Database.

The Intellectual Merits of this project focus on unique combination of novel tools for diffraction-based synchrotron research in Earth and environmental sciences including: (IM1) New capabilities for structure determination of unknown phases and identification of known phases in natural and synthetic samples; (IM2) Opening new possibilities for real time analysis in studies of solid state dynamic processes; (IM3) Robust Bayesian analysis of outliers, uncertainties and missing data; and (IM4) Creation of new free advanced tools for utilizing the available mineralogical database in synchrotron research.

The Broader impacts of the proposed activities will include (BI1) easy accessibility of the new tools, which will be deployed and distributed through the open source code repository site github, as well as through all major synchrotron facilities for Earth science where novel types of experiments are carried out; (BI2) training opportunities, both utilizing online guides and tutorials, as well as workshops; (BI3) open source availability of the software components encouraging collaborations with other developers; (BI4) attractiveness of the ATREX tools outside Earth science community, e.g. in chemistry, materials science and engineering, as well as condensed matter physics.