Clam shells sound warning of Atlantic ‘tipping point’
Peer-Reviewed Publication
Updates every hour. Last Updated: 23-Oct-2025 14:11 ET (23-Oct-2025 18:11 GMT/UTC)
Researchers at Harvard and Northwestern have developed a machine learning method that can design intrinsically disordered proteins with custom properties, addressing nearly 30% of all human proteins that are currently out of reach of AI tools like AlphaFold. The new approach uses automatic differentiation, traditionally a deep learning tool, to optimize protein sequences for desired properties.
Researchers have for the first time measured the true properties of individual MXene flakes — an exciting new nanomaterial with potential for better batteries, flexible electronics, and clean energy devices. By using a novel light-based technique called spectroscopic micro-ellipsometry, they discovered how MXenes behave at the single-flake level, revealing changes in conductivity and optical response that were previously hidden when studying only stacked layers. This breakthrough provides the fundamental knowledge and tools needed to design smarter, more efficient technologies powered by MXenes.
For 200 years, scientists believed heat always spreads the same way—smoothly, like ink dissolving in water. But at the nanoscale, where the world of tomorrow’s chips and energy devices lives, heat behaves very differently. It can ripple like waves, remember where it came from, or even flow like a liquid. Auburn University physicist Prof. Jianjun “JJ” Dong and collaborator Dr. Yi Zeng of DOE’s National Renewable Energy Laboratory have now created the first unified theory that explains all these strange behaviors in one framework. By connecting the atomic motion deep inside materials to the way heat actually propagates, their breakthrough opens the door to designing faster, cooler, and more efficient technologies—from AI hardware to renewable energy systems.