Leak-free thermal batteries from shell waste: Chitin aerogel locks in heat
Peer-Reviewed Publication
Updates every hour. Last Updated: 3-Apr-2026 06:15 ET (3-Apr-2026 10:15 GMT/UTC)
Researchers from City University of Hong Kong, the Chinese Academy of Sciences, and the Massachusetts Institute of Technology have developed an artificial intelligence-driven workflow called AAPSI (AI-Accelerated PhotoSensitizer Innovation) that integrates expert knowledge, scaffold-based molecule generation, and Bayesian optimization to accelerate the discovery of novel photosensitizers for photodynamic therapy (PDT). Through this workflow, the team generated 6,148 candidate molecules and experimentally validated a hypocrellin-based compound, HB4Ph, which achieves a singlet oxygen quantum yield (ϕΔ) of 0.85 and absorption maxima (λmax) of 645 nm — outperforming all clinical and trial-stage photosensitizers. The work is published in AI for Science .