Machine Learning Accelerates Discovery of Near-Infrared Phosphors

Strategic Insights -

Researchers at the University of Houston have published a groundbreaking study on the use of machine learning to accelerate the discovery of near-infrared (NIR) phosphors. The study, published in the journal Chemistry of Materials, demonstrates a machine-learned regression model that can predict the Dq/B ratio of Cr3+-based NIR phosphors with high accuracy. This breakthrough has significant implications for various applications, including biomedical imaging, night vision, and food quality analysis.

Key Takeaways:

  • The study uses a machine-learned regression model to predict the Dq/B ratio of Cr3+-based NIR phosphors, trained on 193 experimentally determined Dq/B values and their associated compositional and structural features.
  • The model is applied to estimate the Dq/B values of over 6060 known inorganic structures with potential octahedral Cr3+ substitution sites.
  • Eight phosphor hosts are selected for synthesis and characterization, and the measured Dq/B values closely match model predictions.
  • The study demonstrates the utility of this machine-learning framework for accelerating the discovery of application-specific Cr3+-substituted NIR phosphors.
  • Financial supporters for this research include the National Science Foundation (NSF), The Welch Foundation, Harvard University, Engineering & Physical Sciences Research Council (EPSRC), UK Research & Innovation (UKRI), Digital Research Infrastructure programme, and the Institute of Advanced Studies (IAS) at the University of Birmingham.

Statistics:

  • The machine-learned regression model is trained on 193 experimentally determined Dq/B values and their associated compositional and structural features.
  • The model is applied to estimate the Dq/B values of over 6060 known inorganic structures with potential octahedral Cr3+ substitution sites.
  • Eight phosphor hosts are selected for synthesis and characterization, and the measured Dq/B values closely match model predictions.
  • The study is supported by a total of eight financial supporters from the United States, the United Kingdom, and other organizations.

Sources:

  • NewsRx. Studies from University of Houston Have Provided New Data on Machine Learning (Machine-learning-assisted Discovery of Cr 3+ -based Near-infrared Phosphors). Journal of Engineering. October 20, 2025; p 4107.
  • Machine-learning-assisted Discovery of Cr 3+ -based Near-infrared Phosphors. Chemistry of Materials, 2025.
  • Amer Chemical Soc, 1155 16TH St, NW, Washington, DC 20036, USA.