In the face of increasingly complex application scenarios, there is an urgent need for (Radio Frequency Identification) RFID systems that are capable of accurately identifying microwave signals of different frequency bands. Based on the acumen detection characteristics of microwave signals by lithium niobate electro-optic modulators, applying large-bandwidth thin-film lithium niobate electro-optic modulation to RFID systems can achieve efficient operation across multiple frequency bands. This study discusses, in detail, the design, simulation, fabrication, and testing process of the electro-optic modulator to obtain a high-performance, large-bandwidth lithium niobate electro-optic modulator. By using multilayer lithography techniques to prepare thick traveling-wave electrodes, the problem of irregular cross-sections during the fabrication of thick electrodes has been successfully reduced, improving the stability and controllability of the device. Test results show that the insertion loss of the electro-optic modulator is about 6 dB, the extinction ratio is 36.5 dB, the optical waveguide mode field is 1 μm, the full-band characteristic impedance is 50 Ω, the test bandwidth is 50 GHz, and the half-wave voltage is 1.8 V. Compared with existing optimization schemes, this design not only achieves a large bandwidth and a small half-wave voltage, but also proposes a new fabrication process scheme, optimizing the process and resulting in samples with stable performance.

To address the challenge of high-dimensional multi-objective optimization in the design of highperformance lithium niobate on insulator (LNOI) acousto-optic modulators (AOMs), this paper integrates digital twins with Bayesian optimization to establish an efficient intelligent design framework. This framework constructs a high-fidelity, parameterized multi-physics coupling digital twin model that accurately replicates the electro-acoustic-optic energy conversion process. An improved multiobjective Bayesian optimization algorithm is employed to efficiently explore a seven-dimensional design space. Within 100 simulation iterations, the framework successfully identifies a Pareto front comprising 17 non-dominated solutions, revealing the inherent trade-offs among key performance metrics. A representative balanced design achieves a diffraction efficiency of 71.5% and a half-wave drive power of 38.7 mW, representing a 16.8% improvement and a 27.9% reduction, respectively, compared to conventional empirical benchmarks. Experimental validation on eight fabricated devices demonstrates good agreement between measured performance and digital twin predictions, with an average absolute relative error of approximately 8.48%. By calibrating the digital twin model with experimental data, the average prediction error is reduced to about 2.0%, establishing a closed-loop workflow of “simulation-optimization-experimentation-calibration”. This work provides a generalizable and efficient paradigm for the inverse design of complex photonic integrated circuits.

Abstract—In this letter, a trilayered heterojunction of tin sulfide (SnS)/hexagonal BN (h-BN)/SnS₂ and corresponding 405-nm NUV photodetectors based on such heterojunctions are proposed. They exhibit a supreme self-powered performance with a remarkable responsivity of up to 1.8 A/W, a large detectivity of 1.51 × 10¹² Jones and a low dark current of 1.2 pA, which surpasses those of NUV photodetectors based on SnS/SnS₂ bilayered heterojunctions and most other 2D materials, and even commercial photodetector (Thorlabs FDS100). This work offers an efficient optimization strategy for SnS₂-based self-powered photodetectors, thereby paving the way for their utilization and application.

Hollow fiber membranes are critical components in modern separation systems, yet conventional integrity inspection methods rely on liquid infiltration and suffer from long testing cycles, complex procedures, and limited defect localization capabilities. These limitations hinder their use in rapid quality control and on-site maintenance. This paper proposes a dry-state non-contact integrity inspection method based on deformation induced by gas permeation and three-dimensional (3D) morphology analysis. Under dry conditions, defective fibers exhibit significantly higher gas permeability. A controlled internal air pressure is applied to the membrane module, inducing localized deformation on a flexible film covering the module end face. High-resolution surface morphology is captured before and after pressurization using line-structured light 3D scanning. Defects are identified via highprecision point cloud registration and differential analysis. A laser stripe center extraction algorithm is developed, termed the Multi-scale Hessian-Gaussian (MHG) algorithm, integrating multi-scale Hessian feature enhancement with direction-adaptive Gaussian fitting to handle highly reflective and densely curved end faces. A Coarse-to-fine Hybrid Point Cloud (CHP) registration strategy is also introduced to improve alignment accuracy and deformation detection. Experiments were conducted on 20 PVDF hollow fiber membrane modules using the GB/T 36137-2018 internal pressure method as a benchmark. The proposed method achieved 100% defect detection consistency under dry conditions, with an average false-positive rate of 0.7% and a detection time of approximately 85 seconds per module. The proposed method offers a rapid, quantitative, and nondestructive solution for membrane module integrity assessment, demonstrating strong potential for applications in manufacturing quality control and operational maintenance.

By leveraging the high sensitivity of acousto-optic modulators to microwave signals, their integration into RFID systems can significantly improve the detection of faint microwave signals. Traditional acousto-optic modulators, which use bidirectional uniform comb transducers (IDTs), suffer from a 50% microwave signal loss due to the diffraction efficiency being less than 50%, complicating the accurate identification of already weak signals. This paper presents a solution to the 50% loss issue by replacing the IDT with a floating electrode unidirectional transducer (FEUDT), a key structure in detecting microwave signals. The resulting samples exhibited an optical waveguide insertion loss of 8.73 dB, a mode field diameter of 6.48 $\mathsf{\mu }$m, an operating frequency of 141.2 MHz, and a diffraction efficiency of 61.72%, and required only 0.57 W of driving power. Comparative empirical tests with traditional transducers under identical conditions demonstrated that the FEUDT-based acousto-optic modulator surpassed the 50% diffraction efficiency limit of conventional modulators, which facilitated the detection of weak microwave signals.