The development of drug delivery systems (DDSs) is moving from empirical, trial-and-error formulation toward data-driven, artificial intelligence (AI)-guided, and automated workflows. Microfluidics provides precise control of microscale fluids, enabling high-throughput production of relatively homogeneous drug carriers such as liposomes, lipid nanoparticles (LNPs), and polymeric micelles. However, the high-dimensional parameter space of microfluidic reactors often exceeds the capacity of manual optimization. This review examines the integration of AI, particularly machine learning (ML), deep learning (DL), and Bayesian optimization, into microfluidic platforms for accelerating DDS design, optimizing critical quality attributes (CQAs), and supporting self-driving laboratory workflows. We discuss the technical foundations of AI-enabled microfluidics, applications in nanocarrier synthesis and phenotypic screening, and the evolving regulatory landscape for AI-assisted pharmaceutical development.
The pre-silicon reliability analysis of digital integrated circuits relies on fault-injection campaigns to characterize how single-event upsets propagate into distinct system-level outcomes. Exhaustive gate-level injection is expensive, whereas most learning-based accelerators collapse failure manifestations into a single binary label, thereby providing limited support for reliability diagnosis and follow-up analysis. We present a reliability-oriented reduced-campaign framework that predicts the monitor-defined outcome type of flip-flop (FF)-level cases omitted from a fixed circuit/workload campaign. For each circuit, Cadence Xcelium first performs SA0/SA1 screening and retains an FF when at least one stuck-at polarity produces a monitored failure. Within the resulting transient campaign, FeatureCoverage selects FF–time cases using structural attributes and fault-free activity statistics without reading their transient-fault outcomes. HSTGNN then combines a 60-cycle FF logic-value window from the same fault-free waveform, netlist-derived FF–FF topology and gate-path attributes, and module hierarchy to predict five outcomes: C0 No Error, C1 Result Error, C2 Exception Error, C3 Timeout Error, and C4 Safety Error. On the I2C, SPI, FIFO, and RISC-V benchmarks, HSTGNN achieves an 81.2–99.0% macro F1 on more than 38,000 held-out FF–time cases under a 70% labeled fault-injection budget comprising 60% training and 10% validation cases. FeatureCoverage also yields the highest macro F1 across all four label-free split strategies in every evaluated circuit. The proposed framework predicts the remaining 30% of cases while preserving reliability-relevant outcome semantics that binary acceleration discards. The evidence is limited to within-campaign prediction and supports, rather than replaces, quantitative reliability and diagnostic-coverage analysis.
In recent years, digital manufacturing, high-energy-beam processing, and nano- and micro-scale fabrication have developed rapidly, extending the capabilities of modern engineering and broadening their practical application [...]
Compared with bare copper wire, palladium-coated copper (PCC) wire is widely used in microelectronic packaging due to its improved oxidation resistance and enhanced reliability. However, the formation mechanism of free air balls (FABs) and the redistribution behavior of Pd during the electronic flame-off (EFO) process, particularly under different Pd coating thicknesses and processing conditions, have not yet been fully understood. In this work, four types of 1 mil PCC wires with Pd coating thicknesses of 60, 80, 100, and 120 nm were systematically investigated to study the influence of EFO parameters on FAB morphology and Pd redistribution behavior. SEM, FIB, and EDS analyses were employed to provide experimental insights into the coupled relationship between transient thermal input, internal pore distribution and elemental segregation evolution, and Pd redistribution behavior. The results show that the preferred FAB morphology is obtained at 54 mA and 580 μs, with a diameter-to-wire ratio of approximately 2. With increasing Pd coating thickness, the exposed copper area on the FAB surface decreases from 13% to 6%, while the Pd-deficient region gradually shifts toward the bottom of the FAB. This study provides experimental insights into the Pd redistribution behavior during FAB formation under different EFO conditions, which may contribute to the optimization of Pd-coated Cu bonding wires.
Microfluidic Hall-effect biosensors detect superparamagnetic bead labels as they flow past a thin-film Hall element in a microchannel. Designing one couples bead magnetization, stray-field distribution, Hall transport, and channel flow across 14 parameters that finite-element solvers explore only at minutes to hours per configuration. We present a coupled analytical–numerical framework for this signal chain: Clausius–Mossotti bead magnetization with a volume fraction correction, a point-dipole stray field, a volume-averaged Hall voltage, Poiseuille transport, and a Johnson–Nyquist and Hooge 1/f noise model, evaluated across 12 sensor presets compiled from the literature, spanning graphene, III–V semiconductors, Si CMOS, bismuth, and topological insulators; any other platform can be defined from user-supplied transport parameters. Benchmarked against a companion COMSOL Multiphysics 6.0 study, the framework reproduces the Hall voltage to within 4.8% at a favorable bead-to-sensor area ratio and deviates by 22% and 15% at off-optimum geometries, consistent with the point-dipole near-field limit at h/rb=1. Three design rules follow: a signal-to-noise ridge at sensor widths comparable to the bead diameter (w*≈db; area ratios 0.4–1.0 at constant voltage, 0.5–2.6 at constant current), matching reported single-bead geometries; a material choice that must be made under an explicit electrical drive constraint; and a sampling-limited flow-velocity window. Predicted signals agree at the order-of-magnitude level with published InAs and Si CMOS experiments. We release the model as a freely accessible, no-install browser implementation with a built-in 2D axisymmetric magnetostatic finite-element (FEM) solver that maps where the dipole approximation degrades.
Point-of-care (POC) diagnostic testing based on microfluidic technology plays an important role in coagulation management. Its precision is, however, limited by the inefficient mixing between blood and solid activators in laminar microfluidic flow. In this study, a novel micromixer incorporating periodic oscillatory flow was developed to enhance blood-activator mixing. A computational fluid dynamic (CFD) model was established to investigate the microscale hydrodynamic characteristics and solid dispersion behavior in the oscillatory multiphase flow system. The numerical predictions were validated against tracer mixing experiments. Sample calculations for the air-liquid-solid flow system demonstrated that the initial loading position of activator particles significantly affected the dispersion efficiency due to the spatial variation in the radial velocity field. Compared with the center-initialized case, the solid dispersion level in the corner-initialized case decreased by approximately 42% after two oscillation cycles. Furthermore, the influence mechanism of oscillation period on solid dispersion was clarified through multi-physics coupling analysis. This study provided new insights into solid–liquid mixing in oscillatory microfluidic systems and established an effective CFD-based framework for optimizing microdevice design and operating conditions.
Lamb wave devices have the advantage of monolithic multi-frequency integration and show great application potential in mobile communications. However, their performance is often limited by spurious modes. In this work, an arc-shaped apodization electrode structure is proposed to suppress spurious modes in Sc0.2Al0.8N Lamb wave resonators (LWRs). By shortening the overlap length of the interdigital transducer electrodes at the edges, the energy coupling in the spurious-mode regions is reduced while the main-mode region is largely preserved. Finite element method (FEM) simulations confirm the suppression of spurious responses near the anti-resonant frequency. Based on the optimized design, typical and arc-shaped apodization LWRs and filters were fabricated. The fabricated arc-shaped apodization LWR exhibits an effective electromechanical coupling coefficient (keff2) of 3.9%. The corresponding filter operates at a center frequency of 1.774 GHz, with a −3 dB bandwidth of 20 MHz and a minimum insertion loss of 4.50 dB. Compared with the typical filter fabricated on the same wafer, the insertion loss is reduced by 1.22 dB. These results demonstrate the effectiveness of the proposed electrode apodization scheme for suppressing spurious modes in Sc0.2Al0.8N Lamb wave devices.
The growing adoption of flexible and portable electronic devices has created an increasing need for sustainable power sources that can operate without relying on conventional batteries. Triboelectric nanogenerators (TENGs) have attracted significant attention due to their ability to convert ambient mechanical energy into electrical energy. However, practical contact–separation TENGs may require additional spacers, elastic supports, or other structural components to maintain repeated contact and separation. This study presents a spacer-free arch-shaped TENG based on a laminated structure consisting of polytetrafluoroethylene (PTFE), paper, and aluminum. The arch configuration provides an inherent restoring tendency that facilitates contact–separation operation without a discrete spacer or additional elastic support. The fabricated device was experimentally evaluated through open-circuit voltage, short-circuit current, load-dependent electrical characterization, power measurement, capacitor charging, and LED illumination. Under the manual mechanical actuation conditions used in the present study, the device exhibited a maximum open-circuit voltage of 69 V, a reported peak short-circuit current of 68 μA, and a peak power output of 8.576 μW at an optimal load resistance of approximately 55 MΩ. Furthermore, the TENG charged a 4.7 µF capacitor to 7 V within approximately 60 s and directly illuminated 30 commercial light-emitting diodes (LEDs). These results demonstrate the feasibility of combining a simple spacer-free arch-shaped configuration with readily available materials for low-cost triboelectric energy harvesting. Long-term cyclic durability, controlled mechanical excitation, environmental stability, and matched structural control experiments remain important areas for future investigation.
In this work, we present the design and numerical validation of a reconfigurable electro-optic photonic integrated circuit architecture for dynamic 1 × 4 optical routing, based on electro-optic phase shifters and a chain of cascaded multimode interference structures, and implemented in an amorphous silicon platform. By exploiting recent developments in symmetry-engineered silicon photonics to enable phase control in a platform compatible with complementary metal-oxide-semiconductor fabrication requirements, this approach addresses the increasing demand for ultra-fast, compact and energy efficient reconfigurable photonic integrated circuits. The photonic circuit design bases its functionality on self-imaging theory and has been optimized through simulations implementing the beam propagation method, while optical performance and electro-optical behavior have been validated through finite difference time-domain simulations and multiphysics modeling. The numerical results obtained confirm the operational performance of the individual building blocks and demonstrate the proposed architecture as being able to perform as a reconfigurable electro-optic platform and provide the dynamic 1 × 4 optical routing. Hence, this architecture provides a scalable platform for reconfigurable silicon photonics, complementary metal-oxide-semiconductor compatible, and the fabrication-ready layout establishes a practical path to experimental validation and future reconfigurable photonic integrated circuits.
Laser Powder Bed Fusion (LPBF) has revolutionized high-end manufacturing, particularly in aerospace and biomedical fields. However, internal defects such as pores, cracks, and inclusions compromise the structural integrity and service reliability of LPBF components. Laser ultrasonics, a non-contact, broadband non-destructive testing (NDT) method, offers a promising solution for detecting and characterizing these defects. This study systematically investigated laser ultrasonic testing technology for LPBF-fabricated Ti6Al4V using a combined approach of physics-driven simulation modeling and experimental validation. To accurately model material anisotropy, a finite element model was developed that integrated Voronoi algorithm-generated polycrystalline microstructures with orientation-dependent elastic tensors, providing a comprehensive representation of the material’s microstructural heterogeneity. Simulation results revealed that while sub-100-μm defects yield weak ultrasonic scattering signals, the Synthetic Aperture Focusing Technique (SAFT) markedly improves the detection and imaging performance for such small-scale defects. Experimental validation using a laser ultrasonic system identified a 90 μm internal defect in the LPBF Ti6Al4V specimen, though a 75 μm defect was undetectable. This highlights the need for enhanced sensitivity. A signal processing method combining time-truncation principal component analysis (PCA) with targeted noise reduction and SAFT was proposed to reduce high-frequency noise and improve high-resolution imaging, enhancing defect detection accuracy. This study provides theoretical foundations and technical support for high-precision defect detection in metal additive manufacturing components, with significant implications for quality control in high-end equipment manufacturing.
This study presents a comprehensive numerical and experimental investigation on a pneumatically actuated polydimethylsiloxane (PDMS) micropump integrated with passive check valves (PCVs). Advanced three-dimensional (3D) fluid–structure interaction (FSI) simulations were conducted to capture the nonlinear large deformation of the membrane and elucidate the Newtonian flow dynamics. The maximum deformation calculated by the simulations was compared with the theoretical Timoshenko formula, demonstrating excellent agreement across the entire pressure range (3.0–100.0 kPa). The 3D FSI simulations revealed a considerable asymmetry in the flow dynamics between the suction and compression phases. Notably, at applied pressures exceeding 20.0 kPa, the discharge volume substantially outweighed the suction one. To characterize both the ideal and practical volumetric flow rates, curve-fitting analyses revealed that both the simulation and experimental data follow a consistent 1/3-power-law relationship with respect to the applied pressure. Experimentally, the micropump achieved a maximum volumetric flow rate of 2.6 mL/min at 70.0 kPa and 12.0 Hz. Furthermore, the micropump demonstrated a peak pumping efficiency of 56.95% at 35.0 kPa and 10.0 Hz relative to the ideal numerical baseline. By bridging idealized numerical bounds with experimental realities, this validated framework offers a robust predictive tool for optimizing flow asymmetry and pumping efficiency in advanced microfluidic systems.
This study reports the synthesis of a novel methylenebisphosphonic acid calix[4]arene derivative and its application as a receptor layer in a conductometric chemosensor for the highly sensitive detection of L-arginine. The synthesized calixarene contains two methylenebisphosphonic acid fragments at the upper rim of the macrocycle, which serve as recognition sites for L-arginine, and two 3-(methylthio)propoxy groups at the lower rim, which enable immobilization on the gold electrode surface. The sensor based on synthesized calixarene demonstrated a pronounced response to L-arginine, which can be attributed to the cooperative interaction of the protonated amino and guanidinium groups of L-arginine with the methylenebisphosphonic acid groups of the calixarene receptor. Density functional theory (DFT) calculations were employed to elucidate the molecular interactions underlying the recognition of L-arginine by the calixarene receptor. The developed chemosensor demonstrated a low limit of detection for L-arginine (5.6 µM); a wide linear range (up to 1000 μM); a short response time (≤80 s); and a high response reproducibility (RSD = 1.3%). The stability constants determined by HPLC depended on the nature of the amino acid and ranged from logKA = 4.26 for leucine to logKA = 4.52 for L-arginine. Compared with previously reported L-arginine sensors, the developed conductometric chemosensor exhibits a competitive detection limit, a wider linear detection range, and significantly improved response reproducibility. The proposed calixarene-based chemosensor enables simple, highly sensitive, enzyme-free conductometric detection of L-arginine over a broad concentration range in aqueous solutions.
The transition from single-device characterization to array-level simulation remains a critical challenge in the development of three-terminal synaptic transistors for neuromorphic computing, as most existing simulation studies either extract parameters from a single representative device and apply them uniformly, or rely on weight-update strategies originally designed for two-terminal memristors. Here, we establish an experimentally calibrated behavioral simulation framework based on differential conductance-pair mapping (W = G+ − G−, where G+ and G− denote the conductances of the positive and negative devices of each pair), integrating exponential long-term potentiation/depression (LTP/LTD) update rules with a posteriori screening mechanism (isValid) to systematically investigate how update polarity, step size, nonlinearity, and conductance boundaries regulate network computational efficiency. Through comprehensive simulation on the Neural Circuit Policies network, we demonstrate that the update direction must strictly align with the matrix’s role: the G− channel requires unidirectional long-term depression inhibition, while the G+ channel can be frozen or bidirectionally updated. The optimal G−LTD and G+G−LTD strategies achieve accuracies of 0.9208 and 0.9481, respectively. Furthermore, we reveal a unique nonlinear gain effect under long-term depression > 0, where accuracy increases monotonically with nonlinearity level up to 0.9419. Device specification criteria are established: LTP-dominant updates favor large Gmax, while LTD-dominant updates favor high Gmin, with the G−LTD and G+G−LTD strategies showing accuracy fluctuations within ±0.005 across the tested boundary variations. Finally, the array-level implementation is validated through a functional-correctness check and device-parameter ablation experiments on a 23,715-weight array (47,430 differential conductance elements). This work provides an experimentally calibrated behavioral simulation platform and concrete algorithm-hardware co-design guidelines for future neuromorphic hardware, prioritizing synaptic devices with long-term depression > 0, a moderately elevated Gmin, and asymmetric resource allocation toward LTD-side optimization.
Biomedical microrobots are often classified by propulsion, materials or fabrication, but clinical translation depends on whether a device can complete a task in a specific biological setting. This review frames biomedical microrobots as task-oriented execution systems. Access and the working environment define the first constraints; actuation and body design convert inputs into useful local work; detection and control close the loop by linking measured states to the next action. Examples from vascular, gastrointestinal, pulmonary, luminal, cellular and biofilm-facing systems show that route, medium, motion, retention, payload function, imaging and post-task fate are coupled rather than separate design labels. Magnetic fields can resist flow or guide motion in confined anatomy, acoustic and optical inputs supply penetrative or local energy, and chemical or biological motors use cues from the surrounding medium. Body design then determines whether movement remains compatible with contact, release, sensing, retrieval and biocompatibility. By focusing on task execution rather than platform identity, this review clarifies what should be preserved in test models, what should be measured, and how fair comparisons can be made before translation.
As the deep integration of the Internet of Things (IoT) and artificial intelligence (AI) technologies has rendered traditional encryption techniques increasingly inadequate in terms of security strength, the implementation of hardware-level security solutions has been identified as a critical issue in the design of information systems. Memristors, as passive circuit components with their intrinsic stochastic switching behavior, multistate storage capability, and low power consumption, have been exploited to provide a viable approach for the construction of hardware security primitives, including true random number generators (TRNGs) and physically unclonable functions (PUFs). In this review, the latest research advances in memristor-based hardware security technologies were systematically surveyed from three perspectives, namely TRNGs, PUFs and other security schemes. With regard to TRNGs and PUFs, respectively, we have systematically summarized their characteristics, including throughput, power consumption and randomness quality, as well as uniqueness, reliability and methods, with particular attention paid to the types of entropy sources and bit-generation strategies. In addition, the potential application of other memristor-based hardware security solutions in information hiding, encrypted transmission and authentication was also analyzed. Subsequently, the principal challenges impeding the transition from laboratory research to mass production were summarized, encompassing synergistic integration with advanced CMOS processes, the balance between reliability and stochasticity, and the absence of design methodologies that integrate hardware and software. Future development trajectories were further delineated, including cross-layer optimization across devices, circuits, and systems, the design of high-security chips resistant to machine learning attacks, CMOS-compatible stable implementation schemes, and the on-chip integration of TRNGs, PUFs, and encryption engines. The present review is intended to serve as a systematic reference for the design of novel hardware security systems, thereby paving the way for the practical deployment of memristor-based security technologies.
Triboelectric nanogenerators (TENGs) offer promising opportunities for distributed wind energy harvesting owing to their low-speed responsiveness, structural flexibility, and adaptability to non-stationary airflow. This review examines wind-driven TENGs from the perspective of realistic wind-field constraints, focusing on three representative scenarios: urban micro-winds, offshore wind–wave environments, and low-altitude complex flows. Scenario-specific advances in device architectures, materials and interfaces, environmental protection, power management, and system integration are systematically reviewed. Representative devices are further quantitatively compared in terms of wind-speed range, activation threshold, electrical output, power density, durability, and system-level energy delivery. Particular attention is given to inconsistent definitions of cut-in wind speed, output normalization, electrical loading, and validation conditions that limit cross-study comparison. Field-validation evidence is assessed from controlled laboratory tests to long-term field operation. Key challenges involving usable regulated energy, environmental reliability, lifetime prediction, array scaling, sustainability, and deployment economics are critically discussed. Finally, five grand challenges with actionable milestones are proposed to facilitate the transition of wind-driven TENGs from laboratory prototypes toward deployable distributed micro-energy systems.
In this study, we compare four electrode models, i.e., solid model, added mass model, thin-film model, and a combining model of thin-film with added mass, for a 76.8 MHz AT-cut quartz resonator. Relative to the high-fidelity solid electrode benchmark, the combining model accurately captures the thickness-shear mode but overestimates displacement confinement due to mass loading and boundary effects. The thin-film model fails to replicate practical device states by ignoring mechanical mass and stiffness. In contrast, the added mass model optimally balances fidelity and efficiency, matching the benchmark’s frequency and resistance predictions with a 43% reduction in computation time. Parametric studies further indicate that cut angle deviations (33°~37°) alter equivalent shear stiffness, while electrode mass variations (mass coefficient 0.99~1.02) modify vibrational inertia. Both of them are critical for frequency tuning. Our findings clarify the scope and limitations of these strategies, guiding high-precision simulations and designs for high-frequency resonators.
Zn–Sn–O (ZTO) thin-film transistors (TFTs) are promising indium-free oxide semiconductor devices, but their electrical stability is limited by oxygen-related defect states. In this study, Ga and Si incorporated ZTO TFTs were systematically compared using an identical bottom-gate top-contact device architecture to investigate dopant-dependent defect modulation and bias-temperature stability. Both Ga and Si incorporation induced a positive threshold-voltage shift and reduced the relative contribution of oxygen-deficient bonding components, suggesting modification of oxygen-related defect environments in the ZTO channel. Optical analysis further showed reduced Urbach energies after dopant incorporation, suggesting a decrease in localized band tail states and reduced structural disorder. Under negative bias temperature stress (NBTS), SZTO exhibited the smallest threshold-voltage shift, demonstrating the most effective stability enhancement. These results indicate that Ga incorporation preserves high field-effect mobility while improving stability, whereas Si incorporation more effectively reduces oxygen-related defect features and provides enhanced NBTS stability. This study provides insight into the dopant-dependent defect engineering for the improved reliability of indium free oxide TFTS.
In this paper, we propose a layered structure composed of Weyl semimetal (WSM) and one-dimensional photonic crystal (1D-PhC) to achieve low-threshold and tunable optical bistability (OB). By exploiting the bandgap characteristics of the photonic crystal and the strong third-order nonlinearity of WSM in the terahertz regime, reflective OB is realized within the photonic bandgap. It is shown that at certain frequencies inside the bandgap, the reflectance exhibits a jump with an increasing incident electric field, manifesting a clear hysteresis loop. The bistable threshold and hysteresis width can be effectively tailored by adjusting the Fermi energy and the thickness of the WSM layer. In addition, the influence of the incident angle and the parameters of the spacer layer on the OB threshold has also been elucidated. After parameter optimization, the incident electric field threshold is reduced to the order of 106 V/m. We believe that this structure can provide a reference for constructing optical bistable schemes with simple structures and low thresholds.
Photonic crystals possessing mirror symmetry have been widely investigated to obtain perfect transmission properties. However, complex asymmetric nanostructures can achieve perfect transmission via phase-matching, while one-dimensional photonic crystals exhibit pronounced angle- and polarization-dependent transmission characteristics under oblique incidence. In this study, we systematically investigated the optical properties of one-dimensional asymmetric photonic crystals containing graded-index films, which yielded a broad perfect transparent photonic band that differed from those occurring in one-dimensional asymmetric photonic crystals comprising solid-index films. The transparent band redshifted, and its bandwidth decreased with increasing amplitude of the graded index. Moreover, the broad perfect transparent band obtained over a wide range of incident angles exhibited greater insensitivity to the incident angle for transverse magnetic polarization than for transverse electric polarization. These results provide a theoretical foundation for designing perfect broadband optical devices using photonic crystals with graded-index films.
