Fresnel zone plate (FZP) stacking is an approach to enhance diffraction efficiency in x-ray focusing and imaging by physically aligning 2 or more FZPs to increase the effective thickness and optical efficiency. Existing methods face limitations in x-ray-source availability, controlled degrees of freedom, alignment accuracy, and contact-induced friction. To overcome these limitations, we develop a double-illumination all-optical nanoalignment (DIANA) system. Designed to leverage the x-ray FZP chip’s nanopatterned characteristics, DIANA uses coherent illumination in reflection mode to implement laser interferometry for accurate tilt and yaw adjustment. It also employs incoherent illumination in transmission mode for high-contrast nanoalignment using Vernier/main scales and moiré fringes. Using DIANA, we demonstrate the stacking of 2 x-ray FZPs with under 30-nm alignment accuracy.

Across Asia's vast drylands, a new study reveals a critical imbalance between degradation and recovery.

Scientists from Harbin Institute of Technology, together with collaborators from The University of Hong Kong and Southern University of Science and Technology, has developed a decoupled titanium dioxide–viologen hybrid material that integrates energy storage and color-changing display functions within a single system. This breakthrough enables high-capacity, high-efficiency electrochromic energy storage and marks an important step toward the next generation of smart, interactive energy devices.

The research conducted by Professor Zou Zui and his team focuses on the immunoregulatory mechanisms in sepsis. As a life-threatening systemic inflammatory response syndrome triggered by infection, sepsis often leads to multiple organ failure, with the NF-κB signaling pathway playing a pivotal role in this pathophysiological process.

Nowadays, compute-intensive programs, like those for training artificial intelligence and machine learning models, are used extensively. Modern compilers use vectorization techniques to exploit parallel processing capabilities to improve the performance of such programs. A group of scientists from the University of Southern California, Cisco AI Research, and Intel Labs designed a data-driven, graph-based learning framework for automatic vectorization called autograph, which utilizes deep reinforcement learning to have an intelligent agent learn an optimal policy. Autograph greatly outperformed other approaches across different datasets. The work was published on June 2, 2025, in an article titled “A Graph-Based Learning Framework for Compiler Loop Auto-Vectorization” in Intelligent Computing, a Science Partner Journal.