Argonne Lab’s ‘DONUT’ AI Accelerates X-ray Data Analysis Tenfold for Real-Time Materials Discovery

August 17, 2026

Argonne National Laboratory News Room USA

Overview

Argonne National Laboratory has developed DONUT, a physics-aware machine learning tool that enables real-time X-ray data analysis at the Advanced Photon Source (APS). DONUT allows materials scientists to make on-the-fly decisions and adapt experiments, accelerating insights into advanced material structures by over a factor of ten. This capability is crucial for understanding material behavior in technologies like batteries, catalysts, and advanced electronic devices, significantly shortening development cycles.

In Depth

Key Findings

Argonne National Laboratory has unveiled DONUT, a physics-aware machine learning tool designed for real-time X-ray data analysis at the Advanced Photon Source (APS). This innovation is set to dramatically accelerate materials science research, allowing scientists to make instantaneous decisions and adapt experiments dynamically. DONUT unlocks deeper insights into the structure and behavior of advanced materials, potentially speeding up discovery by more than ten times.

Technical / Clinical Details

DONUT’s core capability lies in its ability to process and interpret massive datasets from X-ray diffraction, scattering, and spectroscopy in real time. This allows researchers to rapidly understand how materials behave in critical technologies such as batteries, catalysts, and advanced electronic or magnetic devices. The physics-aware machine learning models within DONUT are specifically designed to extract physically meaningful parameters directly from raw experimental data, identifying material microstructure and dynamic behavior on the fly. This dramatically enhances experimental efficiency, reducing analysis times from days or weeks to mere minutes, thereby enabling more agile and responsive research workflows.

Background & Context

X-ray-based materials characterization is indispensable for developing new materials, but the sheer volume and complexity of generated data often create a significant bottleneck in the research pipeline. Real-time elucidation of material synthesis processes and functional mechanisms has long been a challenge in materials science. Tools like DONUT are critical for addressing this, accelerating the entire process from material design and characterization to optimization. This not only speeds up the market introduction of high-performance new materials but also strengthens the competitive edge of related industries, including those in clean energy and advanced electronics.

Strategic Significance & Outlook

The introduction of DONUT establishes a new paradigm in materials science research. Scientists can now optimize experimental conditions, such as reagent concentrations, temperature, and pressure, based on immediate feedback, leading to more efficient synthesis of materials with desired properties. This translates to a rapid acceleration in the discovery and optimization of new materials, fostering innovation across a wide range of fields, including sustainable energy, high-performance electronics, and biomedical technologies. In the long term, DONUT’s capabilities position it as a foundational technology for the development of fully autonomous materials laboratories, further revolutionizing the scientific discovery process.

Source: https://www.anl.gov/article/a-taste-of-discovery-realtime-xray-data-analysis-with-donut-accelerates-materials-science-at-argonne

Get our weekly technology intelligence — free

Receive an infographic that lets you judge at a glance whether each field’s analysis report is worth reading.

Subscribe Free — Weekly Tech Intelligence

By subscribing, you’ll receive Troy-Technical’s weekly technology intelligence newsletter.