Related Experiment Video
Updated: Aug 7, 2026
10:10
Neutron Crystallography Data Collection and Processing for Modelling Hydrogen Atoms in Protein Structures
Published on: December 1, 2020
NeuDiff Agent: a governed AI workflow for single-crystal neutron crystallography
Zhongcan Xiao1, Leyi Zhang1,2, Guannan Zhang3
1Neutron Scattering Division Oak Ridge National Laboratory Oak Ridge TN37831 USA.
Journal of Applied Crystallography
|August 6, 2026
Summary
NeuDiff Agent, an AI workflow for TOPAZ at the Spallation Neutron Source, significantly accelerates structural crystallography analysis. This AI system reduces analysis time by over 4.6x, delivering validated crystal structures ready for publication.
Area of Science:
- Crystallography and Materials Science
- Artificial Intelligence in Scientific Research
- High-Throughput Scientific Analysis
Background:
- Large-scale scientific facilities face analysis and reporting latency, hindering throughput for structural studies.
- Iterative reduction, integration, refinement, and validation steps are time-consuming in traditional crystallography workflows.
- Existing methods struggle to keep pace with the data generation capabilities of modern scientific instruments.
Purpose of the Study:
- To introduce NeuDiff Agent, a governed, tool-using AI workflow designed to accelerate crystallographic analysis at the Spallation Neutron Source.
- To automate the process from raw instrument data to a validated crystal structure and publication-ready CIF file.
- To benchmark the performance and efficiency of the NeuDiff Agent workflow compared to manual analysis.
Main Methods:
- NeuDiff Agent was developed as an AI workflow for TOPAZ, coordinating established crystallographic tools.
- The workflow incorporates explicit governance, including allowlisted tools and fail-closed verification gates.
- Performance was assessed via end-to-end runs using two large language model backends, measuring time, intervention burden, and recovery under gating.
Main Results:
- NeuDiff Agent reduced wall time for structural crystallography analysis from 435 minutes (manual) to 86.5–94.4 minutes (4.6–5.0x faster).
- The AI workflow successfully produced validated CIF files with no checkCIF level A or B alerts.
- The system captured complete provenance for inspection, auditing, and controlled replay.
Conclusions:
- NeuDiff Agent offers a practical and efficient route for deploying agentic AI in facility crystallography.
- The AI workflow significantly improves time-to-result and analysis efficiency without compromising validation requirements.
- This approach enhances scientific throughput by automating complex crystallographic analysis pipelines.
