IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-05-01 Epub Date: 2026-01-27 DOI: 10.1016/j.ijleo.2026.172682

Bodem Indraja, Arijit Datta, Sreenivasulu Tupakula, Swagata Samanta

Contemporary advances in the spatial-temporal modulation of electromagnetic wavefields have fundamentally reshaped optical beam-shaping paradigms, positioning structured light as a pivotal enabler of next-generation photonic architectures and catalysing transformative developments in nanoscale optics, ultra-sensitive sensing, computational imaging, and high-bandwidth optical communications. Within this evolving landscape, structured-light engineering has unlocked unprecedented control over modal excitation and propagation dynamics in fiber-optic platforms. Nevertheless, despite these advances, most fiber-optic sensing schemes still rely on conventional Gaussian-beam illumination, which is inherently limited in its capacity to preferentially couple energy into higher-order guided modes, resulting in reduced evanescent-field extension and weaker interaction with the surrounding medium. Motivated by this fundamental shortcoming, the present work unveils what we believe to be a novel waveguide-based refractive index sensing system by launching a Laguerre higher-order cosh-Gaussian beam (LHOChGB), whose non-trivial spatial structure enables preferential coupling of higher-order guided modes within a uncladded multimode fiber, thereby enhancing the evanescent field localization and strengthening the guided-mode interaction with the external medium. Furthermore, our study was corroborated by using the beam propagation simulations based on the finite-difference method in OptiBPM to characterize the propagation dynamics of the LHOChGB within the sensor structure. Such rigorous analysis unequivocally demonstrates an impressive peak absolute sensitivity of 33.72 dB/RIU, representing an approximate 10.6-fold enhancement over the conventional Gaussian-excited sensor, which typically yields a sensitivity of 3.18 dB/RIU. Therefore, this incipient genre of structured optical field propels innovative trajectories in the field of optical sensing and exhibits substantial potential for multifaceted physico-chemical and biosensing applications.

电磁波时空调制的当代进展从根本上重塑了光束成形范式，将结构光定位为下一代光子架构的关键推动者，并催化纳米光学、超灵敏传感、计算成像和高带宽光通信的变革发展。在这个不断发展的领域中，结构光工程已经解锁了对光纤平台中模态激励和传播动力学的前所未有的控制。然而，尽管取得了这些进展，大多数光纤传感方案仍然依赖于传统的高斯光束照明，这种照明在优先将能量耦合到高阶制导模式的能力上存在固有的局限性，从而减少了倏逝场扩展和与周围介质的相互作用。基于这一基本缺陷，本研究揭示了一种新的基于波导的折射率传感系统，该系统通过发射拉盖尔高阶cosh-高斯光束（LHOChGB），其非平凡的空间结构使高阶导模在无包层的多模光纤中优先耦合，从而增强了倏逝场定位并加强了导模与外部介质的相互作用。此外，基于OptiBPM的有限差分方法的光束传播模拟验证了我们的研究，以表征LHOChGB在传感器结构中的传播动态。这种严格的分析明确地表明，峰值绝对灵敏度为33.72 dB/RIU，比传统的高斯激发传感器提高了大约10.6倍，传统的高斯激发传感器通常产生3.18 dB/RIU的灵敏度。因此，这种结构光场的早期类型推动了光学传感领域的创新轨迹，并在多方面的物理化学和生物传感应用中显示出巨大的潜力。

{"title":"Structured light-induced enhancement of fiber-optic refractive index sensing via the excitation of Laguerre higher-order cosh-Gaussian mode","authors":"Bodem Indraja, Arijit Datta, Sreenivasulu Tupakula, Swagata Samanta","doi":"10.1016/j.ijleo.2026.172682","DOIUrl":"10.1016/j.ijleo.2026.172682","url":null,"abstract":"<div><div>Contemporary advances in the spatial-temporal modulation of electromagnetic wavefields have fundamentally reshaped optical beam-shaping paradigms, positioning structured light as a pivotal enabler of next-generation photonic architectures and catalysing transformative developments in nanoscale optics, ultra-sensitive sensing, computational imaging, and high-bandwidth optical communications. Within this evolving landscape, structured-light engineering has unlocked unprecedented control over modal excitation and propagation dynamics in fiber-optic platforms. Nevertheless, despite these advances, most fiber-optic sensing schemes still rely on conventional Gaussian-beam illumination, which is inherently limited in its capacity to preferentially couple energy into higher-order guided modes, resulting in reduced evanescent-field extension and weaker interaction with the surrounding medium. Motivated by this fundamental shortcoming, the present work unveils what we believe to be a novel waveguide-based refractive index sensing system by launching a Laguerre higher-order cosh-Gaussian beam (LHOChGB), whose non-trivial spatial structure enables preferential coupling of higher-order guided modes within a uncladded multimode fiber, thereby enhancing the evanescent field localization and strengthening the guided-mode interaction with the external medium. Furthermore, our study was corroborated by using the beam propagation simulations based on the finite-difference method in OptiBPM to characterize the propagation dynamics of the LHOChGB within the sensor structure. Such rigorous analysis unequivocally demonstrates an impressive peak absolute sensitivity of 33.72 dB/RIU, representing an approximate 10.6-fold enhancement over the conventional Gaussian-excited sensor, which typically yields a sensitivity of 3.18 dB/RIU. Therefore, this incipient genre of structured optical field propels innovative trajectories in the field of optical sensing and exhibits substantial potential for multifaceted physico-chemical and biosensing applications.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"348 ","pages":"Article 172682"},"PeriodicalIF":3.1,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146090342","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Orbital angular momentum of vector light fields at the tight focus 紧聚焦处矢量光场的轨道角动量

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-05-01 Epub Date: 2026-01-24 DOI: 10.1016/j.ijleo.2026.172680

V.V. Kotlyar , A.A. Kovalev

We introduce here new physical quantities. By analogy with the conventional orbital angular momentum (OAM), they are called polarization OAM and hybrid OAM. It is known that the conventional OAM indicates the presence of an optical vortex in the light field and is equal to the topological charge. The same way, the polarization OAM indicates the presence of a "vortex of the linear polarization direction" in the vector field and is equal to the singularity index of the vector field. Thus, it shows how many full turnovers makes the linear polarization vector along a closed contour around the optical axis. The hybrid OAM demonstrates that along a closed contour around the optical axis, the vector field changes not only by the direction of major axis of the polarization ellipse, but also by the ratio between the major and minor ellipse axes. The hybrid OAM is equal to the number of full turnovers of the polarization ellipse. These new OAMs are derived for several typical initial vector fields to demonstrate their functionality. The initial electric field has only two transverse components, which determine the longitudinal component of the OAM vector. However, at the tight focus, all three components of the E-vector arise, and therefore, longitudinal component of the OAM vector is defined by all three electric field components. Thus, if one of these three OAMs is zero in the initial field, then at the focus, this OAM can become nonzero due to appearing longitudinal field component, contributing to this OAM. We believe that these newly introduced OAMs can be applicable to other vector fields, allowing more complete description of their properties.

我们在这里引入新的物理量。与传统的轨道角动量（OAM）类似，它们被称为极化角动量和混合角动量。已知常规OAM表示光场中存在光涡旋，且等于拓扑电荷。同样，偏振OAM表示矢量场中存在“线性偏振方向的涡旋”，等于矢量场的奇异指数。因此，它显示了多少完全翻转使线性偏振矢量沿着光轴周围的封闭轮廓。混合OAM实验表明，在围绕光轴的闭合轮廓线上，矢量场不仅随偏振椭圆长轴方向的变化而变化，而且随椭圆长轴与椭圆短轴的比值而变化。杂化OAM等于极化椭圆的全匝数。这些新的oam是为几个典型的初始向量场导出的，以演示它们的功能。初始电场只有两个横向分量，这两个分量决定了OAM矢量的纵向分量。然而，在紧焦点处，e矢量的三个分量都出现，因此，OAM矢量的纵向分量由三个电场分量定义。因此，如果这三个OAM中有一个在初始场中为零，那么在焦点处，由于出现纵向场分量，该OAM可能会变为非零，从而导致该OAM。我们相信这些新引入的oam可以应用于其他向量场，从而可以更完整地描述它们的性质。

{"title":"Orbital angular momentum of vector light fields at the tight focus","authors":"V.V. Kotlyar , A.A. Kovalev","doi":"10.1016/j.ijleo.2026.172680","DOIUrl":"10.1016/j.ijleo.2026.172680","url":null,"abstract":"<div><div>We introduce here new physical quantities. By analogy with the conventional orbital angular momentum (OAM), they are called polarization OAM and hybrid OAM. It is known that the conventional OAM indicates the presence of an optical vortex in the light field and is equal to the topological charge. The same way, the polarization OAM indicates the presence of a \"vortex of the linear polarization direction\" in the vector field and is equal to the singularity index of the vector field. Thus, it shows how many full turnovers makes the linear polarization vector along a closed contour around the optical axis. The hybrid OAM demonstrates that along a closed contour around the optical axis, the vector field changes not only by the direction of major axis of the polarization ellipse, but also by the ratio between the major and minor ellipse axes. The hybrid OAM is equal to the number of full turnovers of the polarization ellipse. These new OAMs are derived for several typical initial vector fields to demonstrate their functionality. The initial electric field has only two transverse components, which determine the longitudinal component of the OAM vector. However, at the tight focus, all three components of the E-vector arise, and therefore, longitudinal component of the OAM vector is defined by all three electric field components. Thus, if one of these three OAMs is zero in the initial field, then at the focus, this OAM can become nonzero due to appearing longitudinal field component, contributing to this OAM. We believe that these newly introduced OAMs can be applicable to other vector fields, allowing more complete description of their properties.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"348 ","pages":"Article 172680"},"PeriodicalIF":3.1,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146090341","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

ML-driven gain equalization and enhancement in O+E+S+C multiband optical fiber amplifier ml驱动的O+E+S+C多波段光纤放大器增益均衡与增强

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-05-01 Epub Date: 2026-01-21 DOI: 10.1016/j.ijleo.2026.172675

Krishna Sarma, Divyendu Vats, Mohd Mansoor Khan

This work introduces the design and performance analysis of a hybrid optical amplifier (HOA) operating across the O+E+S+C-bands, achieved by integrating Praseodymium-doped fiber amplifiers (PDFAs), Thulium-doped fiber amplifiers (TDFAs), and Erbium-doped fiber amplifiers (EDFAs) along with a Raman amplifier (RA). The PDFA, TDFA, and EDFA modules are configured in parallel to amplify optical signals spanning from 1270–1560 nm. However, this configuration exhibits reduced gain performance in the 1360–1410 nm and 1500–1520 nm regions, yielding gains of

\sim

18 dB and

\sim

23 dB, respectively, due to the limited spectral coverage of the individual amplifier modules. To overcome these limitations, a Raman amplifier is cascaded with the PDFA–TDFA–EDFA setup to enhance gain performance in the affected regions with an average gain of 30 dB. To optimize the amplifier’s performance across the O+E+S+C-bands, a machine learning (ML) model, P2RAMnet (Pump Parameter Optimization of Raman Amplifier for Multiband Amplification), is developed to predict optimal Raman pump parameters. The model demonstrates high accuracy, with a mean squared error (MSE) of

\sim

0.5 dB² and a mean absolute error (MAE) of

<

0.2 dB. Among all input features, the pump wavelength emerges as the most influential factor, as identified by SHapley Additive exPlanations (SHAP) analysis. Moreover, PDF analysis shows that the majority of prediction errors fall within

\pm

3 dB, highlighting the model’s accuracy and reliability. The pump parameters predicted by P2RAMnet suggest a pump wavelength of 1321 nm at 1500 mW for the 1360–1410 nm region, and a pump wavelength of 1400 nm at 1300 mW for the 1500–1520 nm region. The predicted parameters are validated through OptiSystem

^{®}

simulations, achieving gain flatness of 4 dB in the 1360–1410 nm range and 2.4 dB in the 1500–1520 nm range, while maintaining a gain of

>

26 dB across the 1280–1550 nm spectrum.

本文介绍了一种工作在O+E+S+ c波段的混合光放大器（HOA）的设计和性能分析，该混合光放大器通过集成掺镨光纤放大器（pdfa）、掺铥光纤放大器（tdfa）和掺铒光纤放大器（edfa）以及拉曼放大器（RA）来实现。PDFA、TDFA和EDFA模块并联配置，可放大1270 - 1560nm的光信号。然而，由于单个放大器模块的频谱覆盖范围有限，这种配置在1360-1410 nm和1500-1520 nm区域的增益性能下降，分别产生~ 18 dB和~ 23 dB的增益。为了克服这些限制，将拉曼放大器与PDFA-TDFA-EDFA设置级联，以提高受影响区域的增益性能，平均增益为30 dB。为了优化放大器在O+E+S+ c波段的性能，开发了一种机器学习（ML）模型P2RAMnet（用于多波段放大的拉曼放大器泵参数优化）来预测最佳拉曼泵参数。该模型具有较高的精度，均方误差（MSE）为0.5 dB2，平均绝对误差（MAE）为0.2 dB。在所有输入特征中，泵浦波长是影响最大的因素，这是SHapley加性解释（SHAP）分析确定的。此外，PDF分析表明，大多数预测误差在±3db以内，突出了模型的准确性和可靠性。P2RAMnet预测的泵浦参数表明，在1360-1410 nm区域，泵浦波长为1321 nm， 1500 mW；在1500 - 1520 nm区域，泵浦波长为1400 nm， 1300 mW。通过OptiSystem®仿真验证了预测参数，在1360-1410 nm范围内实现了4 dB的增益平坦性，在1500-1520 nm范围内实现了2.4 dB的增益平坦性，同时在1280-1550 nm范围内保持了26 dB的增益。

{"title":"ML-driven gain equalization and enhancement in O+E+S+C multiband optical fiber amplifier","authors":"Krishna Sarma, Divyendu Vats, Mohd Mansoor Khan","doi":"10.1016/j.ijleo.2026.172675","DOIUrl":"10.1016/j.ijleo.2026.172675","url":null,"abstract":"<div><div>This work introduces the design and performance analysis of a hybrid optical amplifier (HOA) operating across the O+E+S+C-bands, achieved by integrating Praseodymium-doped fiber amplifiers (PDFAs), Thulium-doped fiber amplifiers (TDFAs), and Erbium-doped fiber amplifiers (EDFAs) along with a Raman amplifier (RA). The PDFA, TDFA, and EDFA modules are configured in parallel to amplify optical signals spanning from 1270–1560 nm. However, this configuration exhibits reduced gain performance in the 1360–1410 nm and 1500–1520 nm regions, yielding gains of <span><math><mo>∼</mo></math></span>18 dB and <span><math><mo>∼</mo></math></span>23 dB, respectively, due to the limited spectral coverage of the individual amplifier modules. To overcome these limitations, a Raman amplifier is cascaded with the PDFA–TDFA–EDFA setup to enhance gain performance in the affected regions with an average gain of 30 dB. To optimize the amplifier’s performance across the O+E+S+C-bands, a machine learning (ML) model, P2RAMnet (Pump Parameter Optimization of Raman Amplifier for Multiband Amplification), is developed to predict optimal Raman pump parameters. The model demonstrates high accuracy, with a mean squared error (MSE) of <span><math><mo>∼</mo></math></span> 0.5 dB<sup>2</sup> and a mean absolute error (MAE) of <span><math><mo><</mo></math></span> 0.2 dB. Among all input features, the pump wavelength emerges as the most influential factor, as identified by SHapley Additive exPlanations (SHAP) analysis. Moreover, PDF analysis shows that the majority of prediction errors fall within <span><math><mo>±</mo></math></span>3 dB, highlighting the model’s accuracy and reliability. The pump parameters predicted by P2RAMnet suggest a pump wavelength of 1321 nm at 1500 mW for the 1360–1410 nm region, and a pump wavelength of 1400 nm at 1300 mW for the 1500–1520 nm region. The predicted parameters are validated through OptiSystem<span><math><msup><mspace></mspace><mi>®</mi></msup></math></span> simulations, achieving gain flatness of 4 dB in the 1360–1410 nm range and 2.4 dB in the 1500–1520 nm range, while maintaining a gain of <span><math><mo>></mo></math></span> 26 dB across the 1280–1550 nm spectrum.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"348 ","pages":"Article 172675"},"PeriodicalIF":3.1,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146080614","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Multi-directional controllable electro-optical switch based on grating coupler 基于光栅耦合器的多向可控电光开关

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-05-01 Epub Date: 2026-02-03 DOI: 10.1016/j.ijleo.2026.172685

Yuling Shang , Qilan Xu , Shujie Deng , Hui Jiang , Chunquan Li , Shaoyun Wu , Guotao Huang , Fei Huang

In this paper, based on a lithium niobate thin film (LNTF) photonic integration platform, addressing the difficulties of high microwave loss and restricted optical signal transmission direction in traditional electro-optic modulators, a multi-directional controllable electro-optical switch structure based on a grating coupler is proposed and designed. By cascading three directional couplers and a grating-assisted contra-directional coupler (GACDC), and combining the low-loss transmission characteristics of tapered waveguide and bent waveguide, a multi-directional controllable optical switch is constructed to support forward and reverse transmission of optical signals. Due to the problem of metal absorption and high microwave loss induced by tiny electrode gaps in LNTF electro-optic modulators, traveling wave electrodes with micro-structures were designed to overcome this limitation. Simulation verification with the input of TE₀ mode light with a wavelength of 1550 nm shows that the coupling efficiencies of the output ports (Port₆, Port₉, Port₁₁, and Port₁₂) reach 0.828, 0.830, 0.748, and 0.750, corresponding to insertion losses of 0.82 dB, 0.81 dB, 1.26 dB, and 1.25 dB, respectively. The insertion loss of the GACDC within the device is only 0.56 dB, and the inverse crosstalk is as low as −21 dB; the transmission response of the TE₀ polarization mode is higher than −1.3 dB at the download port in the wavelength range from 1548 to 1553 nm. The flexible output of optical signals from different ports is realized by dynamically modulating the optical signal paths through electrode groups. This study provides a low-loss, multi-directional controllable solution for high-density photonic integrated circuits.

本文基于铌酸锂薄膜（LNTF）光子集成平台，针对传统电光调制器微波损耗高、光信号传输方向受限的难题，提出并设计了一种基于光栅耦合器的多向可控电光开关结构。通过3个定向耦合器和1个光栅辅助定向耦合器（GACDC）级联，结合锥形波导和弯曲波导的低损耗传输特性，构建了支持光信号正向和反向传输的多向可控光开关。针对LNTF电光调制器中微小电极间隙引起的金属吸收和高微波损耗问题，设计了微结构行波电极来克服这一限制。在波长为1550 nm的TE0模式光输入下进行仿真验证，输出端口（Port6、Port9、Port11和Port12）的耦合效率分别达到0.828、0.830、0.748和0.750，对应的插入损耗分别为0.82 dB、0.81 dB、1.26 dB和1.25 dB。器件内GACDC的插入损耗仅为0.56 dB，反串扰低至−21 dB；在1548 ~ 1553 nm波长范围内，TE0偏振模式在下载端口处的传输响应大于−1.3 dB。通过电极组对光信号路径进行动态调制，实现了不同端口光信号的柔性输出。本研究为高密度光子集成电路提供了一种低损耗、多向可控的解决方案。

{"title":"Multi-directional controllable electro-optical switch based on grating coupler","authors":"Yuling Shang , Qilan Xu , Shujie Deng , Hui Jiang , Chunquan Li , Shaoyun Wu , Guotao Huang , Fei Huang","doi":"10.1016/j.ijleo.2026.172685","DOIUrl":"10.1016/j.ijleo.2026.172685","url":null,"abstract":"<div><div>In this paper, based on a lithium niobate thin film (LNTF) photonic integration platform, addressing the difficulties of high microwave loss and restricted optical signal transmission direction in traditional electro-optic modulators, a multi-directional controllable electro-optical switch structure based on a grating coupler is proposed and designed. By cascading three directional couplers and a grating-assisted contra-directional coupler (GACDC), and combining the low-loss transmission characteristics of tapered waveguide and bent waveguide, a multi-directional controllable optical switch is constructed to support forward and reverse transmission of optical signals. Due to the problem of metal absorption and high microwave loss induced by tiny electrode gaps in LNTF electro-optic modulators, traveling wave electrodes with micro-structures were designed to overcome this limitation. Simulation verification with the input of TE<sub>0</sub> mode light with a wavelength of 1550 nm shows that the coupling efficiencies of the output ports (Port<sub>6</sub>, Port<sub>9</sub>, Port<sub>11</sub>, and Port<sub>12</sub>) reach 0.828, 0.830, 0.748, and 0.750, corresponding to insertion losses of 0.82 dB, 0.81 dB, 1.26 dB, and 1.25 dB, respectively. The insertion loss of the GACDC within the device is only 0.56 dB, and the inverse crosstalk is as low as −21 dB; the transmission response of the TE<sub>0</sub> polarization mode is higher than −1.3 dB at the download port in the wavelength range from 1548 to 1553 nm. The flexible output of optical signals from different ports is realized by dynamically modulating the optical signal paths through electrode groups. This study provides a low-loss, multi-directional controllable solution for high-density photonic integrated circuits.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"348 ","pages":"Article 172685"},"PeriodicalIF":3.1,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146190271","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Variable refractive index waveguides for mid-infrared sensing 用于中红外传感的可变折射率波导

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-04-01 Epub Date: 2026-01-06 DOI: 10.1016/j.ijleo.2025.172662

Beibei Kong , Maren Anna Brandsrud , Pranish Karki , Boris Mizaikoff , Achim Kohler

Evanescent fields in attenuated total reflection (ATR) spectroscopy have enabled molecular analysis in the mid-infrared (MIR) region for decades. Recently, thin-film single-mode waveguides have been introduced, significantly improving sample interactions through evanescent fields along their surfaces. However, their implementation demands precise coupling of an infrared laser beam into a sub-wavelength-thick layer, posing design challenges. Here, we introduce gradient refractive index waveguides that achieve robust coupling, substantially enhanced evanescent fields and absorption efficiency, surpassing the performance of conventional single-mode waveguides. Using in-house Finite-Difference Time-Domain (FDTD) simulations validated against commercial software, we demonstrate that these waveguides enhance sample interactions by an order of magnitude. Moreover, their increased robustness simplifies optical system design and broadens their applicability across a wide wavelength range, making them compatible with broadband tunable lasers, LEDs (light-emitting diodes), and they have the potential to accommodate less collimated light sources, such as thermal light sources. This novel waveguide platform promises to advance mid-infrared ATR sensor technologies, particularly in lab-on-a-chip systems and compact devices targeting biomedical and environmental applications.

衰减全反射（ATR）光谱中的倏逝场已经使中红外（MIR）区域的分子分析成为可能。最近，薄膜单模波导已经被引入，通过沿其表面的倏逝场显著改善了样品的相互作用。然而，它们的实现需要将红外激光束精确耦合到亚波长厚的层中，这给设计带来了挑战。在这里，我们引入了梯度折射率波导，实现了鲁棒耦合，大大增强了倏逝场和吸收效率，超越了传统单模波导的性能。利用内部有限差分时域（FDTD）仿真验证了商业软件，我们证明了这些波导提高了一个数量级的样品相互作用。此外，它们增强的稳健性简化了光学系统的设计，拓宽了它们在宽波长范围内的适用性，使它们与宽带可调谐激光器、led（发光二极管）兼容，并且它们有可能适应较少准直的光源，如热光源。这种新型波导平台有望推进中红外ATR传感器技术，特别是在针对生物医学和环境应用的芯片实验室系统和紧凑型设备中。

{"title":"Variable refractive index waveguides for mid-infrared sensing","authors":"Beibei Kong , Maren Anna Brandsrud , Pranish Karki , Boris Mizaikoff , Achim Kohler","doi":"10.1016/j.ijleo.2025.172662","DOIUrl":"10.1016/j.ijleo.2025.172662","url":null,"abstract":"<div><div>Evanescent fields in attenuated total reflection (ATR) spectroscopy have enabled molecular analysis in the mid-infrared (MIR) region for decades. Recently, thin-film single-mode waveguides have been introduced, significantly improving sample interactions through evanescent fields along their surfaces. However, their implementation demands precise coupling of an infrared laser beam into a sub-wavelength-thick layer, posing design challenges. Here, we introduce gradient refractive index waveguides that achieve robust coupling, substantially enhanced evanescent fields and absorption efficiency, surpassing the performance of conventional single-mode waveguides. Using in-house Finite-Difference Time-Domain (FDTD) simulations validated against commercial software, we demonstrate that these waveguides enhance sample interactions by an order of magnitude. Moreover, their increased robustness simplifies optical system design and broadens their applicability across a wide wavelength range, making them compatible with broadband tunable lasers, LEDs (light-emitting diodes), and they have the potential to accommodate less collimated light sources, such as thermal light sources. This novel waveguide platform promises to advance mid-infrared ATR sensor technologies, particularly in lab-on-a-chip systems and compact devices targeting biomedical and environmental applications.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"347 ","pages":"Article 172662"},"PeriodicalIF":3.1,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145978583","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Fabrication approaches for THz metasurfaces: A comprehensive review 太赫兹元表面的制备方法综述

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-04-01 Epub Date: 2025-12-30 DOI: 10.1016/j.ijleo.2025.172664

Meraline Selvaraj, Sreeja B S

Over the past decade, metamaterials have emerged as a revolutionary class of engineered materials, capturing significant attention across diverse scientific domains. Metasurfaces, composed of two-dimensional arrays of engineered meta-atoms, offer precise control over electromagnetic wave amplitude, phase, and polarization at subwavelength scales. This capability makes them an ideal platform for the design of ultra-compact, high-performance components in the THz regime. A critical challenge in transitioning these innovations to practical applications lies in the scalable fabrication of large-area, uniform, and high-resolution micro/nanostructures compatible with THz operation. This review explores the current state-of-the-art fabrication techniques for THz metasurfaces, including lithography, two-photon polymerization, 3D printing, and other advanced fabrication techniques. Each technique’s applicability, advantages, and limitations in achieving structural precision and optimal performance are highlighted. Additionally, challenges in the THz metasurface fabrication are addressed, along with potential solutions to guide future advancements in high-performance THz devices.

在过去的十年里，超材料已经成为一种革命性的工程材料，在不同的科学领域引起了极大的关注。由工程元原子的二维阵列组成的超表面可以精确控制电磁波的振幅、相位和亚波长尺度的极化。这种能力使它们成为设计超紧凑、高性能太赫兹波段组件的理想平台。将这些创新转化为实际应用的一个关键挑战在于，如何可扩展地制造与太赫兹操作兼容的大面积、均匀和高分辨率微/纳米结构。这篇综述探讨了目前最先进的太赫兹超表面的制造技术，包括光刻、双光子聚合、3D打印和其他先进的制造技术。强调了每种技术在实现结构精度和最佳性能方面的适用性、优点和局限性。此外，还解决了太赫兹超表面制造中的挑战，以及指导高性能太赫兹器件未来发展的潜在解决方案。

{"title":"Fabrication approaches for THz metasurfaces: A comprehensive review","authors":"Meraline Selvaraj, Sreeja B S","doi":"10.1016/j.ijleo.2025.172664","DOIUrl":"10.1016/j.ijleo.2025.172664","url":null,"abstract":"<div><div>Over the past decade, metamaterials have emerged as a revolutionary class of engineered materials, capturing significant attention across diverse scientific domains. Metasurfaces, composed of two-dimensional arrays of engineered meta-atoms, offer precise control over electromagnetic wave amplitude, phase, and polarization at subwavelength scales. This capability makes them an ideal platform for the design of ultra-compact, high-performance components in the THz regime. A critical challenge in transitioning these innovations to practical applications lies in the scalable fabrication of large-area, uniform, and high-resolution micro/nanostructures compatible with THz operation. This review explores the current state-of-the-art fabrication techniques for THz metasurfaces, including lithography, two-photon polymerization, 3D printing, and other advanced fabrication techniques. Each technique’s applicability, advantages, and limitations in achieving structural precision and optimal performance are highlighted. Additionally, challenges in the THz metasurface fabrication are addressed, along with potential solutions to guide future advancements in high-performance THz devices.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"347 ","pages":"Article 172664"},"PeriodicalIF":3.1,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885496","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Image deconvolution using adapted cauchy method 采用自适应柯西方法进行图像反卷积

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-04-01 Epub Date: 2026-01-02 DOI: 10.1016/j.ijleo.2026.172665

Zohair Al-Ameen

Generating clearer artifact-suppressed images is vital in various real-world scenarios, where hardware and software limitations often introduce blur. Image deconvolution (ID) aims to reverse this degradation, where prevailing methods vary in intricacy and frequently struggle to maintain stability and acuity across iterations. Driven by the need for non-complex and expeditious solutions, this study explores the use of the classical Cauchy method for ID, a method that, to the best of my knowledge, has not been previously investigated. Accordingly, two contributions are proposed based on this principle. First, the classical Cauchy method (CCM) is modified by embedding the blurring kernel into the iterative update structure, allowing CCM to work as an ID method. Second, an adapted Cauchy method (ACM) is introduced based on CCM. ACM further improves stability and suppresses artifacts better by utilizing a gradient reblurring formulation and an optimized step length rule. ACM is implemented and tested with images of real-life scenes, compared with six existing algorithms, and evaluated using four measures. Experimental results demonstrated that ACM showed a strong ability to restore diverse images while suppressing artifacts, surpassing existing methods in this field. This study showed that a non-complex method can achieve competitive and distinguished results, making it a valuable tool for image deconvolution. The related source codes are available at: https://github.com/qi-zohair/ACM.

在各种现实场景中，生成更清晰的伪影抑制图像是至关重要的，硬件和软件的限制通常会导致模糊。图像反卷积（ID）旨在扭转这种退化，其中流行的方法在复杂性上各不相同，并且经常难以在迭代中保持稳定性和敏锐性。由于需要非复杂和快速的解决方案，本研究探索了经典柯西方法在ID中的使用，据我所知，这是一种以前没有研究过的方法。因此，根据这一原则提出了两项建议。首先，对经典柯西方法（CCM）进行改进，将模糊核嵌入到迭代更新结构中，使CCM可以作为ID方法使用。其次，介绍了一种基于CCM的自适应柯西方法（ACM）。ACM进一步提高了稳定性，并通过利用梯度重模糊公式和优化的步长规则更好地抑制伪影。ACM使用真实场景的图像进行了实现和测试，与六种现有算法进行了比较，并使用四种度量进行了评估。实验结果表明，ACM在抑制伪影的同时表现出较强的复原多种图像的能力，超越了该领域现有的方法。该研究表明，一种不复杂的方法可以获得具有竞争性和显著性的结果，使其成为一种有价值的图像反卷积工具。相关源代码可从https://github.com/qi-zohair/ACM获得。

{"title":"Image deconvolution using adapted cauchy method","authors":"Zohair Al-Ameen","doi":"10.1016/j.ijleo.2026.172665","DOIUrl":"10.1016/j.ijleo.2026.172665","url":null,"abstract":"<div><div>Generating clearer artifact-suppressed images is vital in various real-world scenarios, where hardware and software limitations often introduce blur. Image deconvolution (ID) aims to reverse this degradation, where prevailing methods vary in intricacy and frequently struggle to maintain stability and acuity across iterations. Driven by the need for non-complex and expeditious solutions, this study explores the use of the classical Cauchy method for ID, a method that, to the best of my knowledge, has not been previously investigated. Accordingly, two contributions are proposed based on this principle. First, the classical Cauchy method (CCM) is modified by embedding the blurring kernel into the iterative update structure, allowing CCM to work as an ID method. Second, an adapted Cauchy method (ACM) is introduced based on CCM. ACM further improves stability and suppresses artifacts better by utilizing a gradient reblurring formulation and an optimized step length rule. ACM is implemented and tested with images of real-life scenes, compared with six existing algorithms, and evaluated using four measures. Experimental results demonstrated that ACM showed a strong ability to restore diverse images while suppressing artifacts, surpassing existing methods in this field. This study showed that a non-complex method can achieve competitive and distinguished results, making it a valuable tool for image deconvolution. The related source codes are available at: <span><span>https://github.com/qi-zohair/ACM</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"347 ","pages":"Article 172665"},"PeriodicalIF":3.1,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145885499","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Physics-informed neural networks for predicting ultrashort optical pulse dynamics in nonlinear fiber systems 非线性光纤系统超短光脉冲动力学预测的物理信息神经网络

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-04-01 Epub Date: 2026-01-17 DOI: 10.1016/j.ijleo.2026.172673

Régis Donald Hontinfinde , Charbel Z.J. Mamlankou , Senan Ida Valérie Hontinfinde , Marc Amour Ayela , Gaston Edah

We present a deep learning approach for predicting the nonlinear propagation dynamics of ultrashort optical pulses in fiber-optic systems using Physics-Informed Neural Networks (PINNs). The method solves the Hirota equation, which governs femtosecond pulse propagation in long-haul optical communications by incorporating higher-order dispersion and nonlinear delay effects. By embedding physical conservation laws directly into the neural network loss function, our approach achieves high-precision predictions with absolute errors ranging from

10^{- 4}

to

10^{- 3}

when validated against finite-difference reference solutions. Numerical experiments across multiple parameter configurations demonstrate the robustness and computational efficiency of the PINNs framework. The predicted soliton evolution profiles reveal stable pulse propagation characteristics essential for designing next-generation high-capacity optical telecommunication systems. Our results establish PINNs as a powerful computational tool for analyzing complex nonlinear wave phenomena in optical fibers.

我们提出了一种利用物理信息神经网络（pinn）预测光纤系统中超短光脉冲非线性传播动力学的深度学习方法。该方法通过结合高阶色散和非线性延迟效应，解决了远程光通信中控制飞秒脉冲传播的Hirota方程。通过将物理守恒定律直接嵌入到神经网络损失函数中，我们的方法实现了高精度的预测，当对有限差分参考解进行验证时，绝对误差范围为10−4到10−3。跨多参数配置的数值实验证明了该框架的鲁棒性和计算效率。预测的孤子演化曲线揭示了稳定的脉冲传播特性，这对设计下一代高容量光通信系统至关重要。我们的研究结果表明，pinn是分析光纤中复杂非线性波现象的强大计算工具。

{"title":"Physics-informed neural networks for predicting ultrashort optical pulse dynamics in nonlinear fiber systems","authors":"Régis Donald Hontinfinde , Charbel Z.J. Mamlankou , Senan Ida Valérie Hontinfinde , Marc Amour Ayela , Gaston Edah","doi":"10.1016/j.ijleo.2026.172673","DOIUrl":"10.1016/j.ijleo.2026.172673","url":null,"abstract":"<div><div>We present a deep learning approach for predicting the nonlinear propagation dynamics of ultrashort optical pulses in fiber-optic systems using Physics-Informed Neural Networks (PINNs). The method solves the Hirota equation, which governs femtosecond pulse propagation in long-haul optical communications by incorporating higher-order dispersion and nonlinear delay effects. By embedding physical conservation laws directly into the neural network loss function, our approach achieves high-precision predictions with absolute errors ranging from <span><math><msup><mn>10</mn><mrow><mo>−</mo><mn>4</mn></mrow></msup></math></span> to <span><math><msup><mn>10</mn><mrow><mo>−</mo><mn>3</mn></mrow></msup></math></span> when validated against finite-difference reference solutions. Numerical experiments across multiple parameter configurations demonstrate the robustness and computational efficiency of the PINNs framework. The predicted soliton evolution profiles reveal stable pulse propagation characteristics essential for designing next-generation high-capacity optical telecommunication systems. Our results establish PINNs as a powerful computational tool for analyzing complex nonlinear wave phenomena in optical fibers.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"347 ","pages":"Article 172673"},"PeriodicalIF":3.1,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146023148","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Dual-resonance multilayer Tamm plasmon polariton biosensor for cancer cell detection 用于癌细胞检测的双共振多层Tamm等离子激元生物传感器

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-04-01 Epub Date: 2025-12-18 DOI: 10.1016/j.ijleo.2025.172642

Ahmed Sameer Hatem Alani, Aqilah Baseri Huddin, Norhana Arsad

This work presents a highly efficient multilayer biosensor based on dual‑resonance Tamm plasmon polaritons (TPPs), designed for ultra‑sensitive detection of lung cancer cells. The sensor employs a thin metal-dielectric interface integrated atop an all-dielectric photonic crystal, enabling two ultra-narrowband resonances under normal incidence without the use of coupling prisms or gratings. Numerical modeling using the transfer matrix method (TMM) confirms strong field confinement at the metal-sample interface, resulting in high spectral selectivity and dual resonance modes with full width at half maximum (FWHM) values of 0.133 nm and 0.143 nm. The corresponding quality factors (Q) reach 4029 and 3778, and figures of merit (FoM) are 90 RIU⁻¹ and 106 RIU⁻¹, respectively. The device exhibits sensitivities of 12 nm/RIU and 15.1 nm/RIU for a refractive index variation of Δn_S = 0.01. For biological testing with A549 lung cells, resonance shifts of 0.20 nm and 0.23 nm are observed between healthy (n = 1.3662) and cancerous (n = 1.3568) samples. Considering a spectrometer resolution of 0.1 nm, the minimum detectable refractive index changes are 8.3 × 10⁻³ RIU and 6.6 × 10⁻³ RIU for the two resonances. Due to its ultra‑narrow linewidth, high FoM, label‑free operation, and compatibility with standard thin‑film fabrication, the proposed TPP‑based biosensor offers a compact and accurate platform for real‑time optical cancer diagnostics.

本工作提出了一种基于双共振Tamm等离子激元极化（TPPs）的高效多层生物传感器，设计用于超灵敏检测肺癌细胞。该传感器采用集成在全介电光子晶体顶部的薄金属-介电界面，在正常入射下实现两个超窄带共振，而无需使用耦合棱镜或光栅。利用传递矩阵法（TMM）的数值模拟证实了金属-样品界面处的强场约束，从而产生了高光谱选择性和半最大值全宽度（FWHM）值分别为0.133 nm和0.143 nm的双共振模式。相应的质量因子(Q)达到4029和3778，品质因数（FoM）分别为90 RIU−1和106 RIU−1。灵敏度分别为12 nm/RIU和15.1 nm/RIU，折射率变化为ΔnS = 0.01。在A549肺细胞的生物学检测中，健康样本（n = 1.3662）和癌变样本（n = 1.3568）的共振位移分别为0.20 nm和0.23 nm。考虑到0.1 nm的光谱仪分辨率，两个共振的最小可探测折射率变化分别为8.3 × 10−3 RIU和6.6 × 10−3 RIU。由于其超窄线宽、高FoM、无标签操作以及与标准薄膜制造的兼容性，所提出的基于TPP的生物传感器为实时光学癌症诊断提供了一个紧凑而准确的平台。

{"title":"Dual-resonance multilayer Tamm plasmon polariton biosensor for cancer cell detection","authors":"Ahmed Sameer Hatem Alani, Aqilah Baseri Huddin, Norhana Arsad","doi":"10.1016/j.ijleo.2025.172642","DOIUrl":"10.1016/j.ijleo.2025.172642","url":null,"abstract":"<div><div>This work presents a highly efficient multilayer biosensor based on dual‑resonance Tamm plasmon polaritons (TPPs), designed for ultra‑sensitive detection of lung cancer cells. The sensor employs a thin metal-dielectric interface integrated atop an all-dielectric photonic crystal, enabling two ultra-narrowband resonances under normal incidence without the use of coupling prisms or gratings. Numerical modeling using the transfer matrix method (TMM) confirms strong field confinement at the metal-sample interface, resulting in high spectral selectivity and dual resonance modes with full width at half maximum (FWHM) values of 0.133 nm and 0.143 nm. The corresponding quality factors (<em>Q</em>) reach 4029 and 3778, and figures of merit (FoM) are 90 RIU<sup>−1</sup> and 106 RIU<sup>−1</sup>, respectively. The device exhibits sensitivities of 12 nm/RIU and 15.1 nm/RIU for a refractive index variation of Δ<em>n</em><sub><em>S</em></sub> = 0.01. For biological testing with A549 lung cells, resonance shifts of 0.20 nm and 0.23 nm are observed between healthy (<em>n</em> = 1.3662) and cancerous (<em>n</em> = 1.3568) samples. Considering a spectrometer resolution of 0.1 nm, the minimum detectable refractive index changes are 8.3 × 10<sup>−3</sup> RIU and 6.6 × 10<sup>−3</sup> RIU for the two resonances. Due to its ultra‑narrow linewidth, high FoM, label‑free operation, and compatibility with standard thin‑film fabrication, the proposed TPP‑based biosensor offers a compact and accurate platform for real‑time optical cancer diagnostics.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"347 ","pages":"Article 172642"},"PeriodicalIF":3.1,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145842307","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An automated measurement method for dendritic structures in laser cladding coatings using deep learning 基于深度学习的激光熔覆层枝晶结构自动测量方法

IF 3.1 3区物理与天体物理 Q2 Engineering

Optik

Pub Date : 2026-04-01 Epub Date: 2025-12-19 DOI: 10.1016/j.ijleo.2025.172643

Fang Guo , Longmei Luo , Xueming Zhang , Mengsi Zhang , Youji Zhan , Guofu Lian

To enhance the efficiency of intelligent recognition of the microstructure of laser cladding coatings and optimize the evaluation methods for coating performance and quality, we introduce an image segmentation model called DendriticNet and develops an automated measurement method for the dendritic structure images of laser cladding coatings. The DendriticNet model enables precise pixel-level segmentation and recognition of three types of dendritic structures—dendritic, bulk, and equiaxed crystals—effectively overcoming the limitations of traditional metallographic recognition methods. This provides efficient and accurate technical support for the microstructural analysis of laser cladding coatings. The DendriticNet model combines transfer learning with the Res_CSP attention mechanism module, enabling accurate recognition in complex scenarios involving varying scales and overlapping dendritic structures. Additionally, it achieves qualitative characterization of dendritic microstructures at the micron scale. Experimental results demonstrate that the DendriticNet model achieves an mIoU of 80.31 %, an MPA of 88.43 %, and an accuracy of 93.01 %. By integrating the predicted images produced by the DendriticNet model with the automatic measurement method for dendritic microstructure images of laser cladding coatings, quantitative analysis is conducted to automatically calculate the density ratios of various dendritic structures. Experimental results demonstrate that the measurement accuracy for the total dendritic density ratio exceeds 96.6 %, confirming the feasibility and effectiveness of this automatic measurement method. This approach not only enables the quantitative characterization of dendritic structures at the micron scale but also offers an innovative solution for exploring the relationship between the microstructure and macroscopic performance of coatings.

为了提高激光熔覆涂层微观结构的智能识别效率，优化涂层性能和质量的评价方法，我们引入了一种名为DendriticNet的图像分割模型，并开发了一种用于激光熔覆涂层树枝状结构图像的自动测量方法。DendriticNet模型能够精确的像素级分割和识别三种类型的枝晶结构-枝晶，块状和等轴晶体-有效地克服了传统金相识别方法的局限性。这为激光熔覆涂层的显微组织分析提供了高效、准确的技术支持。DendriticNet模型将迁移学习与Res_CSP注意机制模块相结合，能够在涉及不同规模和重叠树突结构的复杂场景中进行准确识别。此外，它还实现了微米尺度上枝晶微观结构的定性表征。实验结果表明，该模型的mIoU为80.31 %，MPA为88.43 %，准确率为93.01 %。将DendriticNet模型生成的预测图像与激光熔覆涂层枝晶微观结构图像自动测量方法相结合，进行定量分析，自动计算出各种枝晶结构的密度比。实验结果表明，总枝晶密度比的测量精度超过96.6% %，验证了该自动测量方法的可行性和有效性。这种方法不仅能够在微米尺度上定量表征树枝状结构，而且为探索涂层微观结构与宏观性能之间的关系提供了一种创新的解决方案。

{"title":"An automated measurement method for dendritic structures in laser cladding coatings using deep learning","authors":"Fang Guo , Longmei Luo , Xueming Zhang , Mengsi Zhang , Youji Zhan , Guofu Lian","doi":"10.1016/j.ijleo.2025.172643","DOIUrl":"10.1016/j.ijleo.2025.172643","url":null,"abstract":"<div><div>To enhance the efficiency of intelligent recognition of the microstructure of laser cladding coatings and optimize the evaluation methods for coating performance and quality, we introduce an image segmentation model called DendriticNet and develops an automated measurement method for the dendritic structure images of laser cladding coatings. The DendriticNet model enables precise pixel-level segmentation and recognition of three types of dendritic structures—dendritic, bulk, and equiaxed crystals—effectively overcoming the limitations of traditional metallographic recognition methods. This provides efficient and accurate technical support for the microstructural analysis of laser cladding coatings. The DendriticNet model combines transfer learning with the Res_CSP attention mechanism module, enabling accurate recognition in complex scenarios involving varying scales and overlapping dendritic structures. Additionally, it achieves qualitative characterization of dendritic microstructures at the micron scale. Experimental results demonstrate that the DendriticNet model achieves an mIoU of 80.31 %, an MPA of 88.43 %, and an accuracy of 93.01 %. By integrating the predicted images produced by the DendriticNet model with the automatic measurement method for dendritic microstructure images of laser cladding coatings, quantitative analysis is conducted to automatically calculate the density ratios of various dendritic structures. Experimental results demonstrate that the measurement accuracy for the total dendritic density ratio exceeds 96.6 %, confirming the feasibility and effectiveness of this automatic measurement method. This approach not only enables the quantitative characterization of dendritic structures at the micron scale but also offers an innovative solution for exploring the relationship between the microstructure and macroscopic performance of coatings.</div></div>","PeriodicalId":19513,"journal":{"name":"Optik","volume":"347 ","pages":"Article 172643"},"PeriodicalIF":3.1,"publicationDate":"2026-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145814298","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optik最新文献