Online FirstList of IssuesDocument Types

Volume34,Issue21,2026
Issue CoverIssue Contents

    Modern Applied Optics

  • Ultrafast terahertz modulation based on bound states in the continuum metasurfaces AI Introduction

    HU Siyang, HU Yuze, CHENG Xiang'ai, JIANG Tian
    Vol. 34, Issue 21, Pages: 3309-3317(2026) DOI: 10.3724/OPE.20263421.3309
    Abstract:To address the limitations of conventional active terahertz devices, including constrained response speed, low quality factors (Q-factors), and weak light-matter interactions, an ultrafast all-optical terahertz modulation scheme based on symmetry-protected quasi-bound states in the continuum (quasi-BIC) metasurfaces is proposed. A composite metasurface consisting of split-ring resonators (SRRs) and infinite metal bars is designed. By tailoring the height parameter of the metal bars to break the in-plane spatial symmetry, controllable conversion from non-radiative symmetry-protected BICs to high-Q quasi-BICs is achieved, along with continuous tuning of the resonant frequency and Q-factor. Furthermore, photosensitive amorphous germanium (a-Ge) is incorporated into the SRRs to conduct all-optical modulation experiments of the quasi-BIC mode governed by ultrafast photocarrier recombination dynamics. Experimental and numerical results demonstrate that the proposed metasurface effectively sustains high-Q resonance while maintaining significant field enhancement, achieving an ultrafast terahertz transmission modulation and recovery process within 4.5 ps under femtosecond optical pumping. This work highlights the unique advantages of quasi-BIC metasurfaces in facilitating strong light-matter interactions, providing an effective pathway for the development of active terahertz modulators and ultrafast all-optical switches.  
    Keywords:bound states in the continuum;metasurfaces;terahertz;ultrafast modulation  
    76
    |
    30
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • High-sensitivity atmospheric methane detection based on calibration-free WMS and FFT-PSO-KF algorithm AI Introduction

    GU Minghui, HU Xuanjie, WU Bingrui, HE Tianbo, ZHANG Minghui, LI Jinsong
    Vol. 34, Issue 21, Pages: 3318-3328(2026) DOI: 10.3724/OPE.20263421.3318
    Abstract:To address the key issues that the tunable diode laser absorption spectroscopy (TDLAS) is susceptible to optical interference fringe noise and light intensity fluctuation in the in-situ monitoring of atmospheric trace methane (CH4), a high-sensitivity detection method integrating calibration-free wavelength modulation spectroscopy (CF-WMS) and Fast Fourier Transform-Particle Swarm Optimization-Kalman Filtering (FFT-PSO-KF) is proposed. An experimental system based on a near-infrared distributed feedback (DFB) diode laser and a multi-pass long-range absorption cell was built. The second harmonic and first harmonic normalization (2f/1f) was adopted to eliminate light intensity disturbances, achieving calibration-free concentration inversion without standard gases. In terms of signal noise reduction, FFT was used to accurately extract the frequency domain characteristics of interference noise, and the covariance matrix parameters of the adaptive optimization Kalman filter based on PSO were optimized to construct a dynamic Kalman filtering mechanism, achieving efficient decoupling of absorption signals and optical interference noise under low signal-to-noise ratio. Experimental results show that the FFT-PSO-KF algorithm significantly improves 2f signals; Allan variance analysis shows that detection sensitivity for CH4 reaches 2.04×10-9 within 1 s integration time, and the system long-term drift is effectively suppressed. This method has both high anti-interference ability and long-term stability, providing a reliable online monitoring solution for on-site trace detection of atmospheric greenhouse gases.  
    Keywords:tunable diode laser absorption spectroscopy (TDLAS);calibration-free wavelength modulation spectroscopy (CF-WMS);atmospheric methane detection;Kalman filter;particle swarm optimization (PSO);optical interference noise  
    53
    |
    26
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • Measurement-driven behavioral modeling and validation for silicon photonic delay-line chip AI Introduction

    JIN YiFan, LIU Xianming, FENG Jing, CAO Guowei, YU Jiyao, FENG Junbo
    Vol. 34, Issue 21, Pages: 3329-3340(2026) DOI: 10.3724/OPE.20263421.3329
    Abstract:Establishing a reliable transfer path from fabricated device characteristics to system-level responses is critical for scalable silicon photonic microwave-photonic integrated systems. This paper proposes a measurement-driven system-level behavioral modeling and experimental validation method for a 6-bit silicon photonic true-time-delay chip. The proposed workflow starts from wafer-level amplitude and phase characterization of fabricated photonic building blocks, constructs SPICE-compatible behavioral macromodels using complex vector fitting, and deploys the models in Lumerical Interconnect for circuit-level assembly and prediction. By extracting group delays from phase-frequency responses, 64 delay states are quantitatively compared among design targets, measured results, and model predictions. The proposed model achieves a mean prediction error of 1.245 ps, an RMS error of 1.577 ps, a mean relative error of 0.741%, and an average incremental delay error of 0.035 ps. The results indicate that the measurement-driven modeling strategy can preserve fabrication-induced deviations and device nonidealities in system-level simulation, thereby supporting post-fabrication design closure, model reuse, and scalable verification of complex silicon photonic microwave-photonic circuits.  
    Keywords:Silicon photonics;Optical delay line (ODL);microwave photonics;Optical Wafer Acceptance Test (OWAT);Behavioral modeling;Measurement-driven;Complex vector fitting (CVF)  
    80
    |
    62
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • Research progress of topological optimization for large-aperture all-aluminum mirrors AI Introduction

    WU Guanghan, WANG Xu, CUI Yatao, ZHANG Zhiyu, ZHANG Weihong
    Vol. 34, Issue 21, Pages: 3341-3363(2026) DOI: 10.3724/OPE.20263421.3341
    Abstract:Driven by the urgent demands of space optical remote sensing and optomechanical systems for large aperture, lightweighting, and wide-temperature-range adaptability, this paper systematically reviews the research progress of topology optimization for large-aperture optical all-aluminum mirrors, aiming to provide a comprehensive reference for the design and fabrication of high-performance space optomechanical structures. Focusing on the full-chain opto-mechano-thermal integrated development, this review elaborates on two core dimensions: mirror blank topology optimization and support structure design. Furthermore, the critical performance evaluation metrics, standard design workflows for the mirror body and mounting structures, and mainstream manufacturing and optical fabrication routes are systematically summarized. Compared with conventional lightweight design approaches, topology optimization mathematically optimizes the spatial material distribution with high design freedom and efficiency. It substantially improves the lightweight ratio and root-mean-square (RMS) surface figure accuracy while leveraging the inherent advantages of aluminum alloys in rapid fabrication and athermalization design. Current technical bottlenecks regarding multi-constraint coupled optimization, manufacturing constraint integration, and thermal stability are identified, and future research trends toward multi-physics collaborative topology optimization and design for manufacturing (DFM) are prospected.  
    Keywords:all-aluminum mirrors;topology optimization;large aperture mirror;light weight design;athermalization design;surface figure accuracy;support structure  
    68
    |
    35
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18

    Micro\/Nano Technology and Fine Mechanics

  • ZHANG Zeyu, XIAO Weihu, ZHAO Hongda, WU Yulun, YANG Yi, LIU Yizhi
    Vol. 34, Issue 21, Pages: 3364-3372(2026) DOI: 10.3724/OPE.20263421.3364
    Abstract:To address the difficulty of real-time online monitoring of rolling bearing wear under operating conditions and the inaccuracies inherent in static predictive models, this study proposes a method for predicting bearing wear and shafting rotational accuracy by integrating a time-varying Archard wear model with ANSYS transient dynamic simulation. Taking 71816C angular contact ball bearings as the research subject, long-term wear experiments were conducted to dynamically calibrate the temporal evolution of friction and wear coefficients. These time-varying wear parameters were subsequently incorporated into the ANSYS transient dynamics finite element module to numerically simulate the full-cycle contact wear evolution. A geometric averaging equivalent mapping model was then established to quantitatively convert the simulated microscopic wear volume into an engineering-measurable macroscopic shaft runout (radial runout). Finally, the predictive methodology was validated against experimental results obtained from a full-cycle wear test rig. The simulation results matched well with the measured trajectories from two sets of long-term bench tests on 71816C bearings, yielding relative prediction errors for shaft runout of only 6.41% and 1.91%, respectively. The proposed method effectively resolves the predictive drift caused by conventional static wear coefficient assumptions and establishes a cross-scale mapping pathway from microscopic wear volume to macroscopic shafting operational accuracy. It provides a reliable quantitative approach for condition monitoring and accuracy lifetime evaluation of high-precision spindle bearings.  
    Keywords:angular contact ball bearing;Archard model;wear prediction;ANSYS transient dynamics;shaft runout (rotational error);long-term wear test  
    79
    |
    65
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • YU Tianzhu, YIN Xiyao, WANG Zuobin
    Vol. 34, Issue 21, Pages: 3373-3388(2026) DOI: 10.3724/OPE.20263421.3373
    Abstract:Pretreatment conditions were established using 80 μmol/L curcumin for 1 h, followed by the induction of oxidative damage with 0.3% H2O2 for 2 h. Samples were maintained in erythrocyte preservation solution at 4 °C and analyzed within 8 h. Exposure to H2O2 compromised F-actin cytoskeletal continuity, induced nuclear shrinkage, and increased the hemolysis rate to 42.29%. Morphologically, maximum cell height increased from 1.84 μm to 2.22 μm, accompanied by reductions in the major and minor axes from 12.87 μm and 8.29 μm to 12.24 μm and 7.63 μm, respectively. Nanomechanical characterization of the central nuclear region revealed that indentation depth increased from 0.20 μm to 0.26 μm, while the apparent Young’s modulus decreased markedly from 13.37 kPa to 7.74 kPa. Pretreatment with curcumin substantially attenuated these oxidative alterations, reducing the hemolysis rate to 21.57% and moderating cell dimensions (maximum height: 2.03 μm; major axis: 12.75 μm; minor axis: 7.71 μm). Concurrently, indentation depth in the nuclear region decreased to 0.24 μm, and the apparent Young’s modulus recovered to 9.26 kPa. Curcumin substantially mitigates H2O2-induced cytoskeletal impairment and reverses membrane biomechanical stiffness degradation. The established multi-parametric evaluation framework integrating "morphology, ultrastructure, and nanomechanics" provides a sensitive and quantitative approach for elucidating cellular oxidative damage and natural drug protective mechanisms from a micro-biomechanical perspective.  
    Keywords:atomic force microscopy (AFM);chicken erythrocytes;oxidative damage;curcumin;biomechanics;apparent Young's modulus  
    82
    |
    65
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • Material removal mechanism of titanium-containing ULE glass in submerged-jet EEM AI Introduction

    XIE Xianming, YAO Peng, LIU Boxue, ZHANG Sixuan, WANG Weixin, WANG Wei, QU Shuoshuo
    Vol. 34, Issue 21, Pages: 3389-3404(2026) DOI: 10.3724/OPE.20263421.3389
    Abstract:Titanium-doped ultra-low-expansion (ULE) glass, which is widely used in extreme ultraviolet (EUV) lithography and space optical systems, is susceptible to subsurface damage during conventional polishing. Moreover, the complex near-wall flow field in submerged jet elastic emission machining (EEM) makes it difficult to determine appropriate processing parameters. To address these issues, the effects of the near-wall flow field and the material removal mechanism of Ti-doped ULE glass with a TiO₂ mass fraction of approximately 7% were investigated.Computational fluid dynamics (CFD) simulations were combined with fixed-point and scanning polishing experiments to systematically investigate the effects of jet pressure and incidence angle on the near-wall velocity, wall shear stress, and material removal characteristics.The near-wall flow velocity and wall shear stress increased significantly with increasing jet pressure. When the jet pressure increased from 3 MPa to 5 MPa, the maximum material removal depth increased from 46 nm to 100 nm, corresponding to an increase of 117.4%. At a jet pressure of 4 MPa, the maximum removal depth exhibited a non-monotonic dependence on the incidence angle over the range of 45°-90°. Although the maximum removal depth reached 104 nm at an incidence angle of 75°, a central stagnation region and local nonuniformity were observed. In contrast, an incidence angle of 60° produced a moderate removal depth of 83 nm together with a more regular and stable removal profile. Surface scanning polishing was subsequently performed using the optimized parameters of 4 MPa and 60°, achieving continuous and deterministic material removal of approximately 120 nm. The surface roughness Ra was significantly reduced from 0.238 nm to 0.120 nm, while Rq decreased from 0.319 nm to 0.157 nm, with no apparent scratches or edge effects.The results reveal a synergistic material removal mechanism involving normal impact, tangential sweeping, and shear-assisted removal. The study demonstrates that process optimization should consider both material removal capability and removal-function stability. These findings provide key parameter guidelines and theoretical support for achieving atomic-scale, damage-free, ultra-smooth polishing of hard and brittle ultra-low-expansion optical materials.  
    Keywords:ultra-low-expansion (ULE) glass;elastic emission machining (EEM);submerged jet polishing;ultra-precision machining;material removal mechanism;wall shear stress  
    74
    |
    65
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • Two-degree-of-freedom disturbance observer control for moving-base line-of-sight stabilization AI Introduction

    TIAN Dapeng, WANG Zhao, WANG Wen
    Vol. 34, Issue 21, Pages: 3405-3414(2026) DOI: 10.3724/OPE.20263421.3405
    Abstract:Aiming at the inherent deficiency of the disturbance observer (DOB), which amplifies the disturbance response in specific frequency bands while suppressing the equivalent disturbance at low frequencies and thereby reduces the stability margin of the outer closed loop, a two-degree-of-freedom disturbance observer (2-DOF DOB) structure and its H∞-based systematic design method are proposed. By introducing a pre-filter, an independently adjustable element is added to the forward path of the DOB; based on an equivalent H∞ control architecture, the low-pass filter (LPF) in the feedback path and the parameters of the pre-filter are solved simultaneously, so that the robust performance and robust stability of the control system against disturbances are jointly guaranteed. Design verification and experimental comparison were carried out on the line-of-sight (LOS) stabilization control of a moving-base electro-optical stabilization platform. The results show that the traditional DOB induces system oscillation when the outer-loop control gain is high, with a maximum oscillation amplitude of 58.22 (°)/s in the LOS stabilization experiment, and the system cannot operate stably; other inner-loop robust control methods must reduce the outer-loop gain to remain stable; whereas the proposed method maintains stability without reducing the outer-loop control gain, and reduces the steady-state error by 32.65%. This study extends the theoretical framework of the DOB, effectively resolves its stability bottleneck under high-gain outer-loop configurations, and provides a reliable robust control solution for high-precision LOS stabilization control of moving-base electro-optical tracking and pointing systems.  
    Keywords:disturbance observer (DOB);two-degree-of-freedom (2-DOF) control;H∞ robust control;line-of-sight (LOS) stabilization;moving base;electro-optical stabilization platform;pre-filter  
    84
    |
    67
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18

    Information Sciences

  • Frequency attention-enhanced Vmamba for multi-label chest X-ray classification AI Introduction

    CHEN Hui, HE Yuxuan
    Vol. 34, Issue 21, Pages: 3415-3426(2026) DOI: 10.3724/OPE.20263421.3415
    Abstract:To address the insufficient utilization of frequency-domain texture features of subtle lesions, the difficulty of fine-grained local morphological representation, severe long-tailed class imbalance, and inadequate modeling of spatial co-occurrence relationships among disease labels in multi-label chest X-ray (CXR) classification, this paper proposes FDA-VMamba, a multi-label chest X-ray classification network integrating frequency-domain attention with a visual state space model. Built upon the Visual State Space Model (VMamba), the network adopts a heterogeneous hybrid architecture in which Multi-Feature Fusion State Space modules(MFF-SSM) and lightweight Convolutional Gated Multilayer Perceptrons (ConvGatedMLP) are alternately stacked. A Fourier-Wavelet Attention (FWA) module is designed to achieve the collaborative decoupling and fusion of global frequency domain representations and multi-scale local texture features. Combined with ConvGatedMLP, it precisely captures the boundary contours of subtle lesions. Furthermore, a Multi-Scale Graph Frequency Attention Head (MSGFHead) is introduced to explicitly model high-order semantic co-occurrence topologies among multiple disease categories in anatomical contexts. An asymmetric category-focused composite loss function is also incorporated during training to effectively mitigate gradient domination caused by extreme class imbalance. Extensive multi-metric evaluations on the benchmark ChestX-Ray14 dataset demonstrate that FDA-VMamba achieves a mean AUC of 0.861 across 14 disease categories, significantly outperforming existing mainstream convolutional neural networks and Vision Transformer models. The proposed method exhibits strong discriminative capability and generalization robustness in identifying low-contrast and visually confusable lesions, providing a reliable technical foundation for intelligent screening of chest diseases in clinical computer-aided diagnosis.  
    Keywords:chest X-ray (CXR);multi-label image classification;visual state space model (VMamba);frequency-domain attention;wavelet transform;class imbalance  
    27
    |
    16
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • GUO Yingjie, MING Xing, SHI Haodong, LI Guanlin, YU Jin, LIU Xinyue, ZHANG Weichen
    Vol. 34, Issue 21, Pages: 3427-3444(2026) DOI: 10.3724/OPE.20263421.3427
    Abstract:To address the issues of low spatiotemporal accuracy and limited simulation capability during low-elevation entry and exit phases in unmanned aerial vehicle (UAV)-based satellite pass simulation, an adaptive flight path planning algorithm based on velocity compensation and slant range optimization is proposed, aiming to improve both simulation accuracy and route flyability under complex conditions. First, real-time velocity compensation is performed according to the timing errors of the UAV reaching each waypoint, thereby reducing spatiotemporal errors throughout the mission and achieving high-precision satellite pass simulation. Second, a linear fitting relationship between the range scaling factor and the slant range is established to improve the traditional range scaling method, eliminating the defect of monotonicity reversal and enabling effective simulation during low-elevation entry and exit phases. Then, by adaptively selecting between the fixed-altitude method and the improved range scaling method based on the maximum route velocity, an adaptive trajectory planning model is constructed to ensure simulation accuracy while enhancing mission feasibility under complex pass scenarios. Flight experiments under multiple operating conditions and equivalence evaluations demonstrate that the proposed algorithm significantly improves both the spatiotemporal accuracy of satellite pass simulation and route flyability at low elevation angles. Specifically, under the fixed-altitude mode, the position error is reduced from the meter level to the sub-meter or even centimeter level. The proposed algorithm provides reliable technical support for UAV-based satellite pass simulation under electro-optical tracking and pointing (EOTP) operating conditions as well as subsequent electro-optical countermeasure (EOCM) applications.  
    Keywords:unmanned aerial vehicle (UAV);satellite pass simulation;flight path planning;velocity compensation;slant range optimization;range scaling method  
    115
    |
    86
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
  • Cross-modal structural prior-guided latent diffusion model for side-scan sonar data augmentation AI Introduction

    HUANG Peng, DONG Chao, HE Xi, CUI Xuerong, LIU Tongmu, XU Huan, ZHENG Bing
    Vol. 34, Issue 21, Pages: 3445-3457(2026) DOI: 10.3724/OPE.20263421.3445
    Abstract:Existing side-scan sonar image generation methods provide insufficient constraints on target structure, limiting the reliable use of generated images for target detection, particularly for underrepresented categories. To address these limitations, a sonar image augmentation method based on structural priors and latent diffusion is proposed. A cross-modal structural-prior extraction module is developed to mitigate the appearance differences between optical and side-scan sonar images. Both image types are represented as unified three-channel structural control maps comprising edges, boundaries, and target cores, thereby providing consistent structural guidance for cross-modal generation. Because strictly paired optical–sonar samples are unavailable, a pairing strategy based on category consistency and structural similarity is adopted to construct training pairs for optical structural adaptation. The generated images preserve target contours, main-body positions, and category-specific features. When incorporated into the target-detection training set, the generated samples increase detection precision by 5.8%. The average precision (AP) for the underrepresented person category rises from 41.1% to 87.0%, while overall average precision is 11.3% higher than that achieved by the traditional method. These results demonstrate that the proposed method can effectively augment real side-scan sonar training data and improve detection performance for underrepresented categories, offering a practical approach to sonar image augmentation.  
    Keywords:side-scan sonar (SSS);latent diffusion model (LDM);cross-modal generation;Structural prior;data augmentation;few-shot object detection  
    84
    |
    59
    |
    0
    <HTML>
    <L-PDF> <WORD> <Meta-XML>
    Updated:2026-11-18
Previous IssueNext Issue