NOBUGS 2026

Operando X-ray absorption spectroscopy (XAS) at synchrotron beamlines requires careful monitoring of both spectral data and sample-environment conditions throughout an experiment. During catalysis studies detecting anomalous process states as they develop is critical to ensuring meaningful results. An intelligent, automated analysis pipeline that processes data in near real-time can support researchers by enabling timely intervention rather than post-hoc discovery of problems.

The analysis pipeline presented here forms part of the ROCK-IT [1] project, and leverages a near real-time processing infrastructure to perform automated calibration, normalization, linear combination analysis utilizing the Larch [2] library; alongside continuous monitoring of sample-environment parameters such as temperature profiles and gas composition. This pipeline can issue alerts to beamline scientists and provide direct feedback to the experiment control software to adjust input parameters autonomously. The foundation of this system is Asapo [3], a data streaming framework developed at DESY, and the same infrastructure is designed to scale to experiments with significantly higher data rates at modern and upcoming beamline configurations.

A key use case is the detection of unwanted water formation during operando catalysis studies, an artifact that can arise when specific combinations of temperature and gas flows create conditions for unintended side reactions. The pipeline identifies water formation and triggers corrective actions, either through operator notification or automated feedback to the sample-environment control system, which in this case is Bluesky [4]. This demonstrates how intelligent data pipelines can guarantee sensitive measurements and ensure data of high quality in complex experiments.
References
[1] https://www.rock-it-project.de/
[2] https://github.com/xraypy/xraylarch
[3] https://asapo.pages.desy.de/asapo/
[4] https://blueskyproject.io/