Scientists from China have developed a highly scalable on-chip photonic neural network that solves key bottlenecks long limiting the progress of optical computing. The team's new architecture, called a partially coherent deep optical neural network (PDONN), achieves unprecedented network depth by using a cascadable nonlinear activation function with positive net gain. This, combined with the innovative use of more accessible, partially coherent light sources (like LEDs) instead of narrow-linewidth lasers , enable s a chip with the largest input size and deepest structure of its kind to date. The chip successfully performed image classification tasks with high accuracy, marking a critical step toward energy-efficient, scalable, and widely accessible optical computing.

Professor Wang's team and their collaborators have creatively combined the three-dimensional (3D) magic cube configuration with the design structure of metamaterials, opening up a channel connecting information science and mathematical physics. A new paradigm of mechanically reconfigurable metamaterials characterized by high information entropy and visual interactivity has been successfully established. Different magic cube architectures and variable meta-elements allow for complicated and precise customization of electromagnetic waves, holding potential applications in electromagnetic shielding, target camouflage, and holographic encryption. The results of this work were recently published in Science Bulletin.

Optimal packing problems have inspired mathematicians for centuries. Biophysicists now add a layer to the question: how do chloroplasts arrange themselves optimally within cells, when the meaning of “optimal” changes over time? Researchers from the University of Amsterdam and Emory University in Atlanta show how certain plants have managed to solve this problem strikingly well.

Supersolids, a state of matter that combines the rigidity of a solid with the frictionless flow of a superfluid, exhibit surprising synchronization when rotated. Innsbruck researchers found that quantum vortices—tiny whirlpools in the quantum fluid— cause the precession and revolution of the superfluid crystal structure to synchronize their motion. This discovery provides a new tool for studying fundamental properties of quantum systems.

Researchers from Nanjing University and Peking University have developed an innovative method for integrating two transverse superconducting nanowire single-photon detectors (SNSPDs) onto a single silicon waveguide, achieving a detection efficiency of over 99%. This breakthrough provides a scalable and materials-compatible solution for on-chip photon detection in quantum computing and communication applications.

Electrochemical synthesis in aqueous solution has emerged as a sustainable strategy to replace traditional fossil-fuel-driven chemical production, but it is often limited by the sluggish oxygen evolution reaction (OER) that wastes energy and yields low-value products. Recent advances are transforming this challenge into an opportunity by replacing OER with faster, value-added oxidation reactions paired with reduction processes beyond hydrogen production. This review highlights cutting-edge catalysts, hybrid electrolyzers, and in situ characterization tools that enable the simultaneous generation of two valuable products in a single electrochemical process. Such integrated systems not only enhance efficiency and economic viability but also pave the way for greener chemical manufacturing aligned with global net-zero goals.

When dust sticks to a surface or a lizard sits on a ceiling, it is due to ‘nature’s invisible glue’. Researchers at Chalmers University of Technology, Sweden, have now discovered a quick and easy way to study the hidden forces that bind the smallest objects in the universe together. Using gold, salt water and light, they have created a platform on which the forces can be seen through colours.