
Latest Science News
Large AI Models Accelerate Catalyst Discovery
Tohoku University researchers show large AI models, including MLIPs and LLMs, predict catalyst performance before synthesis, revolutionizing clean energy tech. This integrates databases with AI for faster innovation in fuel cells and hydrogen production.
The approach promises closed-loop platforms reducing lab time and resources.
AI Reveals Liquid-Like Ion Flow in Solid-State Batteries
Machine learning predicts Raman spectra, identifying low-frequency signals of rapid ion motion in crystals like Na3SbS4. This speeds discovery of superionic materials for safer, denser batteries.
The workflow simulates spectra at low computational cost with high accuracy.
Record-Breaking Photodetector Captures Light in 125 Picoseconds
Duke University's ultrathin photodetector senses light across the electromagnetic spectrum in record 125 picoseconds. It's the fastest pyroelectric detector, enabling ultrafast applications.
The device powers advanced imaging and sensing tech.