Compared with traditional frame-based machine vision sensors, event cameras are regarded as a promising direction for the next generation of visual systems, thanks to their low latency, minimal data redundancy, and high temporal responsiveness in dynamic vision sensing. However, current event cameras still face significant limitations in dynamic range, contrast sensitivity, and hardware complexity issues that prevent them from fully meeting the requirements of advanced machine vision scenarios like autonomous driving and Industry 4.0. In a recent breakthrough study, Lin et al. proposed an event-driven retinomorphic photodiode (RPD) inspired by the layered structure and signal processing mechanisms of the human retina. This RPD achieves a dynamic range exceeding 200 dB and high-sensitivity adaptive detection, breaking through the performance bottlenecks of existing sensors. This work not only provides a novel paradigm for retinomorphic sensor development but also demonstrates great potential for realizing high-precision, low-power vision perception under complex and dynamic lighting conditions.

The photo-thermal synergistic conversion of CO₂, which harnesses complementary photochemical and thermochemical processes to enhance catalytic efficiency, is gaining increasing attention. Notably, photo-mediated CO₂ cycloaddition exhibits significant promise for practical applications. Catalysts based on metallosalen structures have garnered considerable attention for their role in the formation of cyclic carbonates through CO₂ cycloaddition, owing to their favorable CO₂ and epoxide activation capabilities. In this study, we establish a computational database comprising 300 Salen(Zn)-COFs and screen for optimal photothermal catalysts for CO₂ cycloaddition. Three parameters determine catalyst selection: π-conjugation extent, CO₂ adsorption energy values, and epoxide binding energy values. This approach identifies three pyrene-functionalized COF candidates—Py-EDA-COF, Py-DAC-COF, and Py-OPD-COF—as superior catalytic materials. These three COFs demonstrate outstanding performance in photo-driven CO₂ cycloaddition with various epoxides, particularly phenyl glycidyl ether. Furthermore, we demonstrate that the substituents on the diamine backbone of the salen moiety significantly modulate catalytic activity. Among them, the Py-OPD-COF, which features a benzene unit that enhances π conjugation and exhibits strong electron-donating properties, achieves the highest conversion rate (84%) and a turnover frequency (TOF) of 115.2 h⁻¹, surpassing all previously reported catalysts. This research establishes a viable strategy for the developing of highly efficient catalysts for photo-driven CO₂ conversion.

Metal-organic frameworks (MOFs) are versatile crystalline porous materials with large surface areas, tunable pore architectures, and modular chemical functionalities, enabling diverse applications in materials science and engineering. Hydrogels, with their softness, high water content, biocompatibility, and responsiveness to external stimuli, have been widely explored as smart platforms for flexible electronics and sensing technologies. Integrating MOFs into hydrogel networks synergistically combines the advantages of both material classes, yielding multifunctional composites with enhanced structural and functional properties. MOF–hydrogel composites overcome the limitations of each component, offering improved mechanical robustness, environmental stability, and dynamic responsiveness, making them highly promising for next-generation sensing systems. This review provides a comprehensive overview of recent advances in the synthesis, structural design, and characterization of MOF–hydrogel composites, with a focus on their applications in optical, electrochemical, and electromechanical sensing. Their use across healthcare diagnostics, environmental monitoring, food safety, public health, and flexible electronics is discussed. We highlight how MOF–hydrogel integration influences key sensing metrics such as selectivity, sensitivity, adaptability, detection limits, long-term stability, and dynamic working range. In summary, this review highlights the crucial role of MOF–hydrogel composites in advancing high-performance sensing technologies, outlining key challenges and future directions to inform ongoing research in this field.

Neuromorphic architectures leveraging stochastic device physics present a transformative approach for implementing probabilistic computing paradigms capable of intrinsic uncertainty quantification. In this work, we present a germanium-telluride-based ovonic threshold switch (GeTe_x-OTS) exhibiting inherent stochastic dynamics, integrated into a compact 1-selector-1-transistor-1-resistor (1S1T1R) synaptic unit. The OTS devices are demonstrated within an 8K-array, confirming the future scalability for neuromorphic systems. Experimental validation on the GeTe_x devices shows stable Gaussian-distributed threshold voltage fluctuations (σ = 100 mV), enabling precise control of synaptic activation probability through input pulse modulation. Utilizing this intrinsic stochasticity, we implement mini hardware realization of a Monte Carlo Dropconnect (MC-Dropconnect) neural network, directly demonstrating the feasibility of the system. Applied to COVID-19 diagnosis using chest X-ray images, our system achieves robust uncertainty quantification through predictive entropy, improving classification accuracy to 98.1%, compared to 96.0% with a deterministic baseline. This uncertainty-aware hardware design strategy provides a scalable pathway for implementing energy-efficient neuromorphic systems with native uncertainty estimation capabilities.

The integration of wearable technologies and intelligent algorithms is anticipated to enable synergistic platforms for human motion monitoring and enhanced human-computer interaction. However, the realization of personalized and multi-functional integration of electronic devices through simple fabrication processes remains challenging. Additionally, the current artificial intelligence algorithms lack the universality across different application scenarios. In this work, we present a multifunctional electronic skin (e-skin) system with deep learning-assisted multi-sensing capabilities. The e-skin system integrates a dual-layer 3D-printed e-skin, a flexible multi-channel wireless acquisition circuit, a robust data processing framework, and a multi-action classification neural network (MAC-Net) designed for human activity and sign language recognition. The e-skin can monitor finger bending, joint strain, temperature, humidity, and glucose. Its integrated data processing framework enables an automated workflow from raw signal acquisition to model training and evaluation, enhancing processing efficiency and signal reliability. Experimental results demonstrate that the integration of e-skin with MAC-Net enables accurate daily motion monitoring, sign language recognition, and human-computer interaction, offering a practical and customizable solution for next-generation wearable systems.

Developing two-dimensional (2D) transition metal dichalcogenides (TMDs) that possess uniform crystal orientation and controlled interlayer stacking is essential for next-generation electronic and quantum devices. However, precise control over both lattice alignment and stacking polytype in materials like NbSe₂ remains challenging, primarily due to the small energy difference favoring the hexagonal (H)-stacked phase and the significant lattice mismatch with conventional substrates. Here, we demonstrate the successful growth of mono-oriented rhombohedral-stacked (R-stacked) NbSe₂ ribbons on c-plane sapphire substrates via step-guided epitaxy. This approach utilizes the interaction between atomic steps on the sapphire surface and the NbSe₂ layers: the steps control the orientation and stacking of the NbSe₂ layers. Combined experimental and theoretical analyses reveal that these atomic steps simultaneously guide unidirectional alignment and stabilize the R-stacking configuration. The synthesized NbSe₂ ribbons exhibit promising superconducting properties, with a superconducting transition temperature of 5.6 K and a residual resistance ratio of 4.5. This work paves the way for large-scale integration of single-crystal R-stacked NbSe₂ ribbons, holding immense potential for future applications in superconducting electronics and beyond.