Story perspectives

Manchester AI Spots Kagome Flat-Band 2D Materials with 98% Accuracy

7/10/2026

Summary

Story summary

The University of Manchester built a physics-informed machine-learning tool to locate flat-band 2D materials.
The tool scores structures by band dispersion and sharp density-of-states peaks, trained on known 2D examples.
After screening over 10,000 unlabelled 2D compounds, it highlighted kagome-like lattices, which quantum calculations confirmed flat-band behavior with 98.2% accuracy.
The study appeared in Science Advances on 8 July 2026 and says the materials could support magnetism, superconductivity and fragile topological phases.

1 of 1

Story summary

The University of Manchester built a physics-informed machine-learning tool to locate flat-band 2D materials.
The tool scores structures by band dispersion and sharp density-of-states peaks, trained on known 2D examples.
After screening over 10,000 unlabelled 2D compounds, it highlighted kagome-like lattices, which quantum calculations confirmed flat-band behavior with 98.2% accuracy.
The study appeared in Science Advances on 8 July 2026 and says the materials could support magnetism, superconductivity and fragile topological phases.