Dual-layer hybrid solid electrolyte for improved interfacial stability in solid-state lithium batteries.

Two-dimensional (2D) nanocomposite membranes have gained significant research interest owing to their high designability, excellent strength, and optimal balance between ion selectivity and flux. However, their characteristically large interlayer spacing introduces a fundamental trade-off, typically at the expense of ion selectivity. Bioinspired asymmetry, combined with accelerated ion transport dynamics, presents a promising avenue for advancing 2D nanocomposite membranes. Herein, we report an asymmetric heterogeneous 2D composite membrane for efficient osmotic energy conversion, that integrates a gradient architecture with an SA-enabled transport-promoting microenvironment, consistent with hopping-assisted Na+ transport. The heterogeneous membrane features a dual-layer architecture: a substrate of sulfonated large-sized graphene oxide (GO) nanosheets and bacterial cellulose (BC) for high ion selectivity, and a functional layer of sulfonated small-sized GO nanosheets, BC, and sodium alginate (SA) for enhanced ion flux. This rationally designed structure delivers a power density of approximately 11 W m-2 under a river water/seawater mixture, comparing favorably with representative GO-based membranes under matched artificial-salinity conditions. Temperature-dependent transport measurements, together with continuum and molecular dynamics simulations, support the beneficial roles of the asymmetric structure and SA-containing functional layer, including a lower apparent transport barrier and reduced concentration-polarization-related losses relative to the corresponding controls. This work establishes an asymmetric membrane design strategy for improving the balance between ion selectivity and ion flux in osmotic energy conversion.

Lens fogging and biological contamination are persistent unmet needs in endoscopic surgery that degrade visualization, disrupt procedural continuity, and compromise patient safety. To address these limitations, we have developed the superlubricious high-clarity interface for endoscopic lens defense cap (SHIELD-C), a single-use snap-on endoscope cap designed to maintain stable endoscopic visualization by forming a durable liquid-liquid interface on the cap surface. SHIELD-C employs a dual-layer architecture consisting of a covalently grafted polydimethylsiloxane (PDMS) brush base layer and an infused silicone oil top layer, forming a stable and optically transparent liquid-liquid interface. This configuration maintains high optical clarity (optical transmittance of >90% across the visible range) while effectively suppressing adhesion of blood and mucus (sliding angle of <10°). Importantly, the liquid interface enables rapid intraoperative recovery of surface functionality through simple lubricant replenishment, allowing immediate restoration of visual clarity without interrupting the surgical workflow. The antifouling durability as well as the optical stability of SHIELD-C were systematically validated through in vitro tests under chemically and mechanically challenging conditions with additional confirmation in clinically relevant ex vivo surgical models. These results identify SHIELD-C as a stable liquid-liquid interface cap that enhances the reliability and continuity of endoscopic visualization under lens-fouling conditions.

The spatial confinement of electromagnetic hotspots (<15 nanometers) in plasmonic nanostructures fundamentally restricts their utility for probing large, heterogeneous targets across diverse material and biological systems. We introduce a cascaded photonic-plasmonic strategy that bridges far-field illumination and near-field enhancement by integrating dielectric silicon dioxide microspheres that form subdiffraction nanojets on a plasmonic, gold-coated silicon dioxide nanoarray. This dual-layer architecture generates spatially extended electromagnetic “hotspaces” exceeding 110 nanometers in lateral extent and sustaining analytical enhancement factors > 106, a regime inaccessible to conventional surface-enhanced Raman scattering (SERS) platforms. In silico simulations and experiments reveal ~20-fold enhancements in signal intensity and spatial reach compared to conventional nanoarrays. As a proof of concept, we demonstrate ultrasensitive, label-free classification of extracellular vesicles, 80 to 200 nm in diameter, derived from patients with colorectal cancer with 99.8% accuracy, surpassing traditional SERS (<87.5%). More broadly, this cascaded excitation strategy shifts the emphasis from nanogap optimization to the engineering of spatially extended fields through hybrid light-focusing architectures, enabling advances in spectroscopy, biosensing, nanophotonics, and diagnostics.

This research paper presents a comprehensive study of a hybrid book recommendation system that integrates contentbased filtering with popularity metrics to address the critical challenge of recommendation diversity in online book discovery platforms. The system implements a dual-layer architecture combining TF-IDF semantic vectorization (70/30 weighting for research-grade precision) and an optimized production model (60/40 weighting) to eliminate popularity bias while maintaining user satisfaction. The proposed approach achieves a 15% improvement in recommendation diversity compared to traditional popularity-driven systems, while preserving content relevance. This paper documents the complete development lifecycle, technical implementation, performance evaluation, and real-world deployment considerations for an intelligent book discovery application serving diverse user preferences.

Toward a relational paradigm for AI safety: relational synchrony in the heart-inspired dual-layer architecture (HIDLA)

Widely targeted metabolomics reveals conserved and differential metabolite patterns related to nutritional and bioactive compounds in stems and leaves of three Portulaca oleracea cultivars.

Abstract In response to the urgent need for high-precision, large-range planar near-field testing of large antennas, this paper designs and implements a large-stroke dual-tower structure scanning frame, with emphasis on its control system architecture and core planning algorithms. At the control system level, a dual-layer architecture integrating motion control and health monitoring is constructed. The motion control system, based on a servo computer and multi-axis controller, achieves high-precision coordinated motion of X, Y, Z, and polarization (P) axes, real-time position pulse feedback, and power-off position memory with resume testing capabilities. The health monitoring system, through integration of vibration sensors, temperature and humidity sensors, and signal acquisition instruments, realizes online monitoring and intelligent early warning of equipment mechanical status and testing environment. At the control algorithm level, the system supports 11 scanning modes, including continuous and step-wise scanning. Addressing the challenges of large-scale, high-density test point planning and controller hardware limitations, an intelligent trajectory planning algorithm based on dynamic switching of high-speed comparison axes is proposed. This algorithm generates test queues based on test parameters and, through the introduction of coordinate accuracy compensation, dynamic flatness compensation, and a partitioned queue management strategy with comparison axis switching points as boundaries, effectively ensures motion accuracy and execution efficiency for large-scale testing tasks.

Engineerable stabilized bonding hydrogel coatings with dual-layer architecture for scalable drag reduction

Direct recycling lithium-ion battery cathodes, a process that retains the engineered oxide structures from end-of-life materials, presents a cost-effective and energy-efficient alternative to other battery recycling methods. However, while direct-recycled cathodes have demonstrated performance comparable to that of pristine materials at low cycling rates, their high-rate performance remains uncertain. Morphology changes in cathode particles, a main mode of degradation, directly impact rate performance by limiting surface kinetics and solid-phase diffusion. If direct recycling processes do not sufficiently restore pristine-like morphologies, the recycled materials may retain structural defects that hinder high-rate performance. The present work uses a physics-based pseudo-2D model to simulate hybrid electrodes with pristine and artificially “aged/recycled” NMC materials to investigate potential impacts of incorporating performance-limited aged cathode materials into cells. The study highlights how differences in transport and kinetic properties can influence rate capabilities in mixed electrodes— particularly in high-loading cells in high-demand applications. However, model results also reveal a possible mitigation strategy via dual-layer electrode architectures with lower-performing materials positioned near the current collector. Simulations of 4.0 mAh cm −2 cells cycled at 4C using a dual-layer architecture provided approximately 5%–30% more capacity in constant-current protocols compared to homogeneously blended electrode architectures with the same loadings and mixed-material compositions. These findings highlight the importance of strategic electrode design in minimizing potential performance losses and facilitating the integration of recycled materials into high-performance batteries, advancing sustainable and cost-effective battery manufacturing. • A P2D model is developed for electrodes with blended or layered active materials. • The model is calibrated and validated using mixed-NMC experimental data. • Hybrid pristine-recycled electrodes benefit from layering under high-rate cycling. • Direct-recycled materials are viable without fully restoring pristine performance.

This paper presents BIOSYNC 1.0, an innovative attendance automation and visualization system tailored for educational institutions. Traditional biometric solutions, such as those from ESSL, ZKTeco, and other vendors, primarily focus on recording attendance logs and exporting them in raw Excel formats, leaving administrators to manually interpret, compile, and distribute the data. Such systems lack integration with modern communication tools and fail to provide institution-wide analytics. BIOSYNC 1.0 addresses these shortcomings by introducing a dual-layer architecture that integrates local automation with a cloud-based analytics dashboard. At the local automation layer, BIOSYNC 1.0 monitors a designated folder for biometric exports in Excel or CSV formats. Upon detecting a file, the system automatically processes it into structured daily reports. These reports highlight total attendance, absentees, and offline department flags. The pipeline then generates WhatsApp-ready summaries, ensuring that department heads, staff, and administrators receive real-time notifications without manual intervention. Simultaneously, reports are archived into a month-wise backup system, creating a tamper-proof and organized record of attendance over time. The cloud-based visualization layer extends these capabilities. Processed reports are uploaded to a centralized dashboard powered by Supabase (backend) and React (frontend). Once uploaded, the reports are parsed into normalized datasets with key attributes such as year, department, section, and date. This enables stakeholders to visualize attendance patterns using bar charts, line graphs, and trend analyses. The system supports role-based authentication, where administrators manage uploads and users, while faculty access filtered dashboards for their specific classes or departments By combining real-time reporting, secure archival, and advanced visualization, BIOSYNC 1.0 transforms attendance management into an intelligent analytics system. It ensures seamless communication, reduces manual workload, and provides both micro-level (class-specific) and macro-level (institution-wide) insights. This positions BIOSYNC 1.0 as a significant advancement over existing biometric solutions that remain limited to data collection and static reporting.

Surgical site infections (SSIs) remain a primary postoperative complication that conventional passive sutures fail to address. We present a conductive thermoresponsive silk suture (CTS) with a dual-layer architecture, featuring an inner reduced graphene oxide (rGO) layer that enables real-time strain sensing, an outer thermoresponsive hydrogel layer designed for on-demand drug delivery. The integrated design provides mechanical robustness, stable electromechanical and intelligent release, significantly accelerating elution at fever temperatures (40 and 42 °C) compared to the basal release at normal body temperature (37 °C). In vivo studies confirmed CTS significantly modulated inflammatory response by reducing tumor necrosis factor-α (TNF-α) and CD68 levels at day 7, and accelerated tissue integration at days 14 and 28 by promoting angiogenesis (CD31) and collagen deposition. This sense-and-treat suture can track suture tension and wound status, demonstrating clinical potential for precise postoperative management in diverse surgical settings.

Composite polymer electrolytes (CPEs) offer an attractive route to safer lithium metal batteries, yet their practical deployment is often limited by a persistent trade-off between ionic conductivity and mechanical integrity, together with unstable electrode/electrolyte interfaces. In this work, a dual-layer PVDF-based composite electrolyte (PVLC CPE) is constructed, consisting of a lithium aluminum titanium phosphate (LATP)-rich layer to facilitate Li+ transport and a cerium oxide (CeO2) nanofiber-reinforced layer to strengthen the mechanical robustness of membrane and stabilize Li deposition. The optimized PVLC CPE exhibits a high ionic conductivity of 1.02 × 10–4 S cm–1 at 30 °C with a low activation energy of 0.127 eV and robust thermal stability up to 120 °C. A Li/PVLC/Li symmetric cell delivers stable stripping/plating for over 1800 h at 0.1 mA cm–2. Full cell assembled with LiFePO4 cathode, the electrolyte enables 95.7 mA h·g–1 after 200 cycles at 1 C with near-100% Coulombic efficiency. These results demonstrate that a rational dual-layer architecture can simultaneously promote fast Li+ transport and mechanically regulate the interface, offering a practical design principle for durable quasi-solid-state batteries.

The increasing reliance on automated recruitment tools has introduced challenges such as inaccurate filtering, loss of qualified candi- dates, and lack of transparency in resume evaluation. This study presents CVision, an AI-powered web-based platform designed to enhance the re- cruitment process through intelligent resume analysis, automated docu- ment generation, and optimized job matching. By integrating traditional rule-based evaluation with advanced natural language processing (NLP) and machine learning models, CVision provides a holistic assessment that goes beyond keyword matching to interpret context, structure, and skill relationships. The platform operates through a dual-layer architecture — a baseline rule-based scoring engine and an AI-driven analytical layer — both connected via a FastAPI backend and React-based frontend inter- face. The system allows candidates to upload resumes, receive personal- ized improvement feedback, and explore job recommendations tailored to their profiles. In controlled evaluations on real-world datasets, CVision demonstrated a 24% improvement in accuracy, reduced false positives by 67%, and achieved a 19% faster processing rate compared to con- ventional Applicant Tracking Systems (ATS). Additionally, the platform maintained 94% compatibility across irregular document formats, high- lighting its adaptability in real-world hiring scenarios. The outcomes of this study showcase CVision’s ability to redefine automated recruitment through contextual understanding and performance-driven intelligence, paving the way for fairer and more efficient hiring ecosystems.

The development of dependable in vitro models that replicate the pulp-dentin complex is important for regenerative endodontics, biomaterials testing, and disease modelling. However, most existing approaches concentrate on isolated techniques, providing limited guidance regarding their comparative performance, practical limitations, or translational applicability. Furthermore, they often present only the final optimized method without elaborating on the decision-making process, including the challenges faced and the rationale for not pursuing alternative approaches. This study aimed to address this gap by evaluating multiple three-dimensional (3D) strategies for the development of pulp-dentin models. Two principal assembly routes were explored: manual and digital. Manual assemblies used natural dentin rings combined with different 3D culture strategies to generate the pulp core. Scaffold-free spheroids were formed using ultra-low attachment plates (ULA) and non-adherent overlays (Agarose and Matrigel), while scaffold-based systems employed Matrigel encapsulation. These were manually integrated within dentin rings to establish pulp-dentin interfaces. Digital assemblies utilized extrusion-based 3D bioprinting to fabricate composite constructs composed of dentin powder-reinforced Gelatin Methacryloyl (GelMA) or alginate outer rings and collagen or alginate-based cellular cores. All constructs were evaluated for structural stability, reproducibility, cellular organization, and interaction with the dentin interface, primarily through morphological and histological analyses. Both manual and digital assembly strategies successfully produced 3D pulp-dentin constructs with distinct characteristics. In the manual assemblies, scaffold-free and scaffold-based approaches enabled spheroid formation and matrix-supported tissue organization, respectively. When integrated with natural dentin rings, these cultures established localized pulp-dentin interfaces with Dentin Sialophosphoprotein (DSPP) positive cells, indicating odontogenic differentiation. However, construct uniformity and stability were influenced by spheroid size, hydrogel degradation, and dentin ring geometry. Digital bioprinting enabled precisefabrication of biphasic constructs comprising a dentin powder-reinforced outer phase and acell-laden inner core. Stable and reproducible dentin-mimetic outer rings were achieved atdentin powder concentrations ≤20% (w/w), with GelMA exhibiting slower degradation andgreater mass retention than alginate. The inner pulp-like core was bioprinted using cell-ladenbioinks containing 5 × 10⁵ cells, with LifeInk 220 collagen providing consistent print fidelity andhomogeneous cell distribution. The resulting dual-layered architecture enhanced the structuraland biological resemblance of the model to native pulp-dentin tissue, with dentin powderincorporation contributing to dentin-like features and supporting odontogenic cell organization. This study establishes a methodological framework for developing in vitro pulp-dentin models by comparing manual and digital assembly strategies. Rather than identifying a single optimal approach, the work highlights how each method contributes unique advantages and challenges. Documenting both successful outcomes and technical limitations provides valuable guidance for optimizing 3D pulp-dentin constructs and advancing their application in regenerative endodontics, biomaterial evaluation, and translational research.

Two-timescale hierarchical DRL for resource allocation in UAV-assisted 6G edge networks

Most existing approaches to drug combination discovery are disease-centered, aiming to identify drug pairs for specific diseases. Complementarily, a drug-centered strategy starts from known drug combinations and explores new therapeutic indications, facilitating translational applications by leveraging combinations with established safety profiles. Here, we introduce DualKG-DC, a drug centered computational framework that provides a complementary perspective by identifying potential disease indications for a given drug combination. The dual layer knowledge graph architecture, which is pretrained on a foundation biomedical knowledge graph and subsequently refined on a task specific drug combination subgraph, may reduce reliance on large, labeled datasets by leveraging existing knowledge on drug targets, biological pathways, and observed phenotypic effects. In systematic benchmarking against three state-of-the-art models, DualKG-DC outperformed all comparison models, achieving an average Hits@10 of 0.48, MRR of 0.30, AUROC of 0.99, and AUPRC of 0.31. Notably, in cold start scenarios, DualKG-DC outperformed baseline methods in predicting indications for unseen drug combinations, achieving superior results with an average Hits@10 score of 0.32, an MRR of 0.18, an AUROC of 0.98, and an AUPRC of 0.23. These results highlight DualKG-DC as an effective platform for systematically discovering therapeutic opportunities of drug combinations. By leveraging a dual-layer architecture, the model enables effective knowledge transfer, enhancing predictive performance and robustness, particularly for previously unseen drug combinations.Supplementary InformationThe online version contains supplementary material available at 10.1007/s10916-026-02416-y.

A dual-layer architecture for lithium-ion batteries' state of health estimation: Adaptive local convolution and global temporal-frequency retention network

Healthcare systems increasingly rely on digital twin technology to create virtual representations of patients, medical devices, and clinical workflows. However, secure data sharing across medical digital twin edge networks faces critical privacy and security challenges when handling sensitive patient information and treatment protocols. This paper presents a medical digital twin blockchain sharding (MDTBS) framework that leverages blockchain sharding technology to address security and privacy concerns in healthcare data sharing while maintaining the real-time responsiveness required for clinical operations. The framework incorporates a novel dual-layer architecture combining local medical data sharing chains using directed acyclic graph consensus for intra-hospital communications with global medical data sharing chains employing delegated proof of stake consensus for inter-hospital collaboration. Considering the dynamic characteristics of medical environments and mapping errors between physical healthcare systems and their digital twins, we formulate an adaptive resource allocation model that jointly optimizes medical cluster head selection, hospital base station consensus access, and spectrum and computation resource allocation to maximize blockchain sharding transaction throughput. A medical digital twin edge network-empowered two-layer proximal policy optimization algorithm solves the complex optimization problems while adapting to time-varying medical workflows and equipment configurations. Simulation experiments demonstrate that the framework achieves superior performance across all evaluation metrics compared to baseline methods, including 15-25% improvements in transaction throughput with statistical significance (p-value less than 0.001), sub-three-second emergency response times, and 85%+ privacy preservation efficiency scores.

In this work, we present the first fully packaged 16×16 Benes optical switch based on 3D photonic integration operating in the O-band, and demonstrate, for the first time, its system-level evaluation. The 3D switch employs a dual-layer architecture with 56 thermo-optic 2×2 Mach-Zehnder interferometer-based crossbar units and vertical multimode interferometers enabling the transition between the two parallel waveguiding layers and effectively eliminating in-plane waveguide crossings. The static optical performance of the 3D optical switch prototype was evaluated based on insertion and polarization-dependent losses, crosstalk, power consumption, and switching speed. The 3D switch is fabricated on a polymer-based photonic platform, known as PolyBoard, which is suitable for developing 3D structures. In order to perform a system-level demonstration, the switch was integrated into a 25 Gb/s full-duplex optical link using a commercial Network Interface Card and SFP transceivers. Stable throughput was maintained across back-to-back and extended-reach fiber links up to 10 km, demonstrating the viability of this 3D-integrated switch for scalable, energy-efficient intra-data center interconnect.