Tackling real-world crystal structure prediction from powder X-ray diffraction data
Year: 2026
Type: article
Abstract: Crystal structure prediction (CSP) from powder diffraction data is a central challenge in materials chemistry. Machine learning (ML) models show promise, but most are trained on idealized simulated data, limiting reliability on real experiments. Here, we assess real-world behaviour using deCIFer as a representative case. deCIFer is an autoregressive transformer that conditions each step of structure generation on encoded PXRD data, guiding token-wise predictions of space group, lattice parameters, and atomic positions. Using controlled robustness tests, we quantify performance under realistic artefacts (noise, background, peak asymmetry, and Scherrer broadening) and introduce metrics for accuracy and predictive uncertainty. deCIFer adapts smoothly to signal distortions and improves over unconditioned baselines when diffraction features remain informative, while expressing appropriate uncertainty as the PXRD pattern becomes underdetermined. Tests on experimental PXRD show recovery of known structures but also highlight limitations for lower-symmetry and nanocrystalline samples. Overall, ML-based CSP is fundamentally limited by the information content of PXRD, but can accelerate expert workflows by rapidly generating chemically plausible candidates and quantifying uncertainty, making such models valuable human-in-the-loop tools for real-world structure determination.
Source: Digital Discovery
DOI: 10.1039/d6dd00088f
Institution: University of Copenhagen
Language: English
Subfield: Materials Chemistry
Field: Materials Science
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W7196945909