Facial nerve injury arising from diseases or clinical interventions can cause significant physiological and psychological harm. Continuous monitoring of facial muscle electrophysiology can mitigate these risks by informing surgical procedures and guiding postoperative rehabilitation. However, conventional electrodes are predominantly invasive and fabricated from mechanically stiff metals that mismatch soft tissues, often leading to secondary injury and discomfort. Here, we develop an ultrasoft eutectogel and employ it as a non-invasive electrode material for a surface electromyography (sEMG) monitoring array. The eutectogel is synthesized by random copolymerization of acrylic acid (AA) and the zwitterionic monomer 3-[N,N-dimethyl-[2-(2-methylprop-2-enoyloxy)ethyl]ammonio]propane-1-sulfonate (SBMA) in a deep eutectic solvent (DES), yielding an electrode that integrates ultra-low modulus, strong adhesion, ionic conductivity, and environmental tolerance. The ultra-low modulus and adhesiveness facilitate conformal, dynamic coupling with epidermal tissues, providing an ideal material platform for non-invasive high-quality acquisition of physiological electrical signals, including facial nerve-related activity. We further validate the feasibility of predicting facial nerve functional integrity using this platform. The eutectogel and the flexible multi-electrode array provide a promising strategy for clinical protection and monitoring of facial nerve function.

A novel class of semiconducting compounds, metal-halide perovskites (MHPs), has emerged as a versatile platform for advanced optoelectronic device architectures, offering a unique combination of exceptional physical properties and facile processing. In this study, we present a monolithic high-speed photodetector capable of directly sensing the time delay between two light pulses with a temporal resolution of at least 170 ps, corresponding to a light propagation distance of ~5 cm—making it well suited for Light Detection and Ranging (LiDAR) applications. This outstanding time resolution is achieved through a signal-balancing detection scheme that effectively overcomes the limitations of conventional photodetectors, whose response speed is inherently limited by charge-carrier lifetime and transit time. The device exhibits an exceptionally low noise spectral density, comparable to that of state-of-the-art silicon photodiodes. The fully symmetric device stack comprises a crystalline CsPbBr₃ absorber layer tens of microns thick, fabricated via a confined melt process. Comprehensive electro-optical characterization reveals charge-carrier lifetimes and mobilities on both microscopic and macroscopic length scales, using transient photoluminescence, time-resolved photocurrent, time of flight, and terahertz pump–probe spectroscopy. The CsPbBr₃ layer exhibits charge-carrier lifetimes exceeding 100 ns, a microscopic electron–hole mobility of 15 ± 1 cm² V⁻¹ s⁻¹, and a macroscopic non-dispersive hole mobility of 8.5 cm² V⁻¹ s⁻¹.

Monitoring sweat loss is an effective method for evaluating dehydration during body thermoregulation. However, current wearable microfluidic sweat sensors often face limitations in terms of breathability and heat dissipation, and textile-based sweat sensors cannot achieve the accurate control and detection of sweat volume. Herein, we report a textile patch with an unprecedented layer-by-layer isovolumetric water transport (IVWT) ability for sweat loss monitoring. The patch features an all-porous laminated fabric with multiple IVWT units and conductive nonwoven fabric sensing units, enabling controllable and quantitative sweat penetration similar to dam-transporting ships. Each layer of the IVWT unit could accurately and stably transport 6.29 ± 0.10 μL of sweat across all tests. This design allows the identification of sweat volume by increasing stepwise jumps in the conductance signals. The dam-inspired textile patch not only provides sweat volume measurements that are highly consistent with those obtained using the absorbent pad method but also offers superior air permeability (~98% of the clothing), excellent heat dissipation (~79% of the uncovered skin), and excellent compatibility, which facilitates seamless integration into various types of wearable garments. A fully integrated wireless transmission device with the textile patch provided a validated predictive model for estimating whole-body sweat loss during dehydration monitoring.

Despite the high theoretical volumetric capacity of aluminum metal anodes (AMAs), their practical use in rechargeable aluminum batteries (RABs) is hindered by low capacity utilization and short-circuit-induced cell failure. Herein, we investigate the aluminum nucleation and growth behavior on a 2D electrode platform to uncover the origins of such failures, integrating experimental analysis with theoretical calculations. We find that the failure capacity is strongly dependent on separator thickness, irrespective of separator type. Short-circuiting arises from unfavorable multi-step reactions, where inefficient Cl⁻ removal promotes vertical Al growth due to localized accumulation of reaction products. Based on these insights, we design a 3D nanostructured graphitic carbon electrode (3D-GCE) to mitigate local AlCl₄⁻ buildup and enhance Al reversibility. Additionally, a Cl-doped polypropylene (Cl-PP) separator is employed to facilitate Cl⁻ transport via the Grotthuss mechanism. This integrated design achieves a record capacity of ~8.2 mAh cm⁻² and stable cycling over 500 cycles with a single thin PP separator.

Neuromorphic computing systems, inspired by the brain's parallel processing capabilities and efficiency, offer promising solutions for artificial intelligence. Spiking neural networks (SNNs), composed of neuron and synapse elements, are a key approach for neuromorphic systems. However, traditional hardware neuron implementations require auxiliary circuits to achieve good training performance of SNNs. Developing appropriate single-device based neural components to enable efficient SNN implementations remains elusive. Here, we introduce a gate tunable MoS₂ memristive neuron. This neuron possesses tunable refractory periods and firing thresholds, emulating key dynamics of neurons without external circuits. Leveraging these adaptable neurons, we develop an early fusion SNN architecture for multimodal information processing based on tunable neuron devices. Through cross-modality weight sharing, proposed neurons can learn common features across modalities and modality-specific features under different gate voltages. This architecture achieves seamless fusion of multisensory data while significantly reducing hardware costs. We demonstrate a 49% reduction in hardware usage along with a major boost in recognition accuracy to 95.45% on an image-audio digit recognition task. Our tunable neuron-enabled SNN provides a pathway for highly efficient neural computing and further integration of neuromorphic intelligence.

Perovskite solar cells (PSCs) have rapidly advanced owing to their excellent optoelectronic properties such as high absorption, long diffusion length, and high carrier mobility, achieving power conversion efficiencies of up to 27%. The ABX₃ crystal structure of perovskites and their various possible material combinations provide broad compositional and dimensional tunability, enabling tailored bandgaps, controlled stability, and targeted optoelectronic features. However, the efficiency of conventional trial-and-error approaches in discovering new materials is limited by the interplay between the compositional, material, and processing variables, highlighting the need for reproducible synthetic protocols and reliable datasets to support the high-throughput exploration of material combinations. Artificial intelligence (AI) technologies, including machine learning and large language models, leverage such datasets to provide predictive and generative capabilities for performance forecasting, compositional and process optimization, inverse design of novel materials, and literature knowledge extraction. Furthermore, the combination of an automated protocol setup, fabrication, high-throughput characterization using AI, and large device-level datasets has paved the way for building autonomous research platforms. Specifically, automation and robotics are integrated with in situ metrology and algorithmic guidance to reduce the build–measure–learn cycle from weeks to hours, thereby accelerating discovery and stability assessment. This review focuses on three central pillars of data-driven and AI research: data platforms, AI methodologies, and self-driving laboratories, which could collectively reshape PSC research into a systematic, autonomous, and scalable framework. By reviewing advances across these domains, we demonstrate how data-driven strategies can transform PSC development from intuition-based exploration to accelerated and reliable innovation, paving the way for practical deployment and commercialization.

In contrast to multi-element ferroelectrics with intricate atomic configurations, single-element ferroelectrics are distinguished by their structural simplicity and low-dimensional nature. In addition, they have been experimentally verified to display robust ferroelectricity at room temperature. These materials are regarded as promising candidates for next-generation flexible optoelectronic devices because of their high electrical conductivity and inherent compatibility with integrated circuits, this review systematically elucidates the underlying mechanisms governing the emergence of ferroelectricity in single-element systems and comprehensively surveys the state-of-the-art fabrication techniques. The fundamental physical properties of several prominent single-element ferroelectrics including tellurium (Te), bismuth (Bi), and black phosphorus (BP) are discussed, highlighting their applications in memristors and optoelectronic devices. Finally, the current research challenges are addressed and future trends in material development, fabrication techniques, and potential applications are presented. The objective of this review is to provide a comprehensive understanding of single-element ferroelectrics, offering valuable insights into their potential for broader applications.

Poor environmental robustness, especially physical damage and chemical corrosion, has remained a primary challenge in practical applications of electromagnetic (EM) wave absorbing materials. The core obstacle involves coupled trade-offs: increasing EM filler loading enhances EM attenuation but restricts polymer healing kinetics needed for rapid self-healing, while also increasing the susceptibility of fillers to oxidation and corrosion. This work proposes a novel photothermally driven MXene/self-healing polyurethane (SPU) elastomer that reconciles these competing requirements. Under xenon lamp (simulated sunlight, 1000 W m⁻²), MXene's photothermal conversion effectively accelerates the reversible bond exchange and chain mobility in the SPU matrix, enabling recovery of both mechanical and EM wave absorbing properties within 30 min after damage (self-healing rate over 70 times faster than under dark condition). Tuning MXene surface terminations further tailors its electrical structure and dielectric response, achieving an effective absorption bandwidth of 5.40 GHz at a thickness of 1.77 mm and a radar cross-section reduction of 23.55 dB·m². Meanwhile, strong hydrogen bonding interaction between SPU and surface termination passivates oxidative sites and forms a protective barrier, effectively suppressing the degradation of MXene and endowing the elastomer with excellent environmental robustness under seawater and harsh acidic/alkaline media. Overall, these findings offer a versatile design paradigm for flexible, durable and self-healing EM wave absorbing materials with potential applications in next-generation wearable electronics, stealth technologies, and marine protection applications.