Materials Science · Materials Chemistry
A study finds that large language models can recognize patterns in material categories but consistently struggle to use coordinate information, even as the models and datasets grow. For properties shaped mainly by crystal structure, specialized models that represent geometry performed better.
Nanosecond-scale simulations of the solid acids CsH2PO4 and CsHSO4 identify a “proton slingshot” mechanism for fast proton conduction. A rotating phosphate or sulfate group first carries a proton, while reorientation of the proton’s bond to oxygen extends its movement and enables longer jumps.
Researchers developed an AI-powered pipeline that extracts information about concrete materials from scientific papers and organizes it into a database. Applied to a corpus screened from more than 27,000 publications, the system produced nearly 9,000 records for an open database on blended cement concrete.
Researchers have introduced a curated database containing more than 30,000 entries on solid-state hydrogen-storage materials, drawn from over 4,000 experimental literature sources. The database includes storage capacity and key heat- and disorder-related properties, giving researchers a way to compare materials under different operating conditions.
Wild relatives contain adaptive genetic diversity that was reduced during crop domestication, but conventional prediction methods often miss complex DNA changes. A review in BMC Genomics outlines how artificial intelligence and broader genetic maps could help identify these variants for climate-resilient breeding.
