Micromachines

Research

Jump to: Review

28 pages, 35935 KB

Open AccessArticle

Efficient Automatic Design of a 2D TMD FET via Machine Learning-Assisted TCAD Simulation

by Na Shi, Zi-Jun Wei and Tong Wu

Micromachines 2026, 17(8), 987; https://doi.org/10.3390/mi17080987 - 21 Aug 2026

Viewed by 72

Abstract

As the scaling of silicon-based devices approaches physical limits, two-dimensional transition-metal dichalcogenide field-effect transistors (2D TMD FETs) have emerged as promising candidates for logic devices in the post-Moore era. However, their design optimization relies heavily on computationally intensive TCAD simulations, thereby limiting efficient [...] Read more.

As the scaling of silicon-based devices approaches physical limits, two-dimensional transition-metal dichalcogenide field-effect transistors (2D TMD FETs) have emerged as promising candidates for logic devices in the post-Moore era. However, their design optimization relies heavily on computationally intensive TCAD simulations, thereby limiting efficient exploration of multidimensional parameter spaces. This paper proposes an efficient automated design framework for 2D TMD FETs under small-sample conditions and validates it using a monolayer MoS₂ FET as a case study. The framework integrates device design, physics-based simulation, performance prediction, and inverse design, establishing a bidirectional mapping between device parameters and electrical performance. Target-driven closed-loop optimization is achieved through TCAD-based feedback validation. Results demonstrate that, using a dataset comprising 300 TCAD samples, the forward model achieves an average coefficient of determination (

R^{2}

) of 0.9503. TCAD revalidation of the inverse-designed devices yields an average mean absolute error (MAE) of 0.0464 and an average mean absolute percentage error (MAPE) of 5.46% for performance metrics. Regarding computational efficiency, while a single TCAD simulation takes approximately 25 to 50 min, the trained model performs inference in under 50 ms, achieving a speedup of at least

3 \times 10^{4}

during the inference phase. Accounting for the generation of the 300 TCAD samples and the training of both forward and inverse models, the framework’s one-time computational cost ranges from 160.27 to 285.27 h. Once the cumulative number of design tasks exceeds approximately 342 to 385, the total computational cost falls below that of direct TCAD simulation, with the computational advantage becoming increasingly significant as the number of tasks grows. Consequently, this method is highly suitable for large-scale parameter sweeps, device screening, and multi-objective, high-frequency design iterations. It drastically reduces repetitive TCAD calls, offering a scalable solution for the efficient, automated design of 2D TMD FETs. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

24 pages, 25465 KB

Open AccessArticle

A Defect Detection Method for Functional Membranes in Flexible Sensors for Vibration Monitoring During Glass Substrate Transfer

by Zhuohao Shi, Han Wang, Yibin Chen, Shuai Chen, Daohua Zhan and Weicheng Ou

Micromachines 2026, 17(8), 933; https://doi.org/10.3390/mi17080933 - 5 Aug 2026

Viewed by 265

Abstract

Vibration monitoring of glass substrate transfer systems is crucial for ensuring the stable operation of Flat Panel Display (FPD) manufacturing equipment. Fabrication defects in the functional nanofiber membrane of flexible vibration sensors can significantly degrade sensing performance and lead to inaccurate monitoring results. [...] Read more.

Vibration monitoring of glass substrate transfer systems is crucial for ensuring the stable operation of Flat Panel Display (FPD) manufacturing equipment. Fabrication defects in the functional nanofiber membrane of flexible vibration sensors can significantly degrade sensing performance and lead to inaccurate monitoring results. To address the challenge of achieving an effective balance between detection accuracy and inference efficiency in such defect-dense scenarios characterized by large variations in defect scale, this paper proposes a novel defect detection model, termed MA-YOLO. The proposed model incorporates four key architectural enhancements: the Multi-level Bidirectional Feature Aggregation Network (MLBAN), the Multi-Receptive Field Adaptive Fusion Module (MRAF), the Morphology-Adaptive Feature Extraction Module (MA-C2f), and the Interactive Dynamic Decoupling Head (IDDH). These components collaboratively improve defect feature extraction, multi-scale feature fusion, and localization performance while maintaining a lightweight architecture and high inference speed. Experimental results on a self-constructed defect dataset demonstrate that MA-YOLO achieves a mean Average Precision (mAP@0.5) of 91.9%, which is a 3.1 percentage point improvement over the baseline model. Moreover, with only 9.15 million parameters and an inference speed of 119.05 FPS, the proposed model exhibits superior overall performance compared with several mainstream and state-of-the-art object detection methods. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

17 pages, 12564 KB

Open AccessArticle

Single and Repetitive Surge Reliability of 1200 V Asymmetric Trench SiC MOSFETs Under Various Gate Biases

by Menglin Yan, Zhizhe Wang, Dazheng Chen, Yuncong Li, Yongle Zhong, Yuansheng Li, Jun Luo and Hao Xia

Micromachines 2026, 17(7), 823; https://doi.org/10.3390/mi17070823 - 10 Jul 2026

Viewed by 468

Abstract

The parameter degradation and failure mechanisms of 1200 V asymmetric trench-type (AT) silicon carbide (SiC) metal oxide semiconductor field-effect transistors (MOSFETs) under various single and repetitive surge currents, with various gate bias voltages (V_GS) of 0 V, −5 V, and −10 [...] Read more.

The parameter degradation and failure mechanisms of 1200 V asymmetric trench-type (AT) silicon carbide (SiC) metal oxide semiconductor field-effect transistors (MOSFETs) under various single and repetitive surge currents, with various gate bias voltages (V_GS) of 0 V, −5 V, and −10 V, are systematically investigated in this work. It is indicated that V_GS has no impact on the single surge reliability, with the same maximum single surge current (SSC_max) under different V_GS. However, during repetitive surge stress (90% and 60% SSC_max), the maximum surge cycles have increased as V_GS increases from −10 V to 0 V. It may be caused by the enhancement of channel-assisted leakage conduction, allowing more surge current to flow through the channel. It is concluded from gate capacitance (Cg-Vg) and low-frequency noise (LFN) characterizations that lower V_GS increases SiC/SiO₂ interface defect density, accelerating parameter degradation during single and repetitive surge stress. Both chip and package failures are observed for single and repetitive surge stress. For single surge stress, the device failure has resulted from the melted source Al as the metal erodes and penetrates through the interlayer dielectric and the ohmic contact layer between the source metal and the SiC-doped region, respectively, leading to a three-terminal short circuit. For repetitive surge stress, the device failure has been caused by the penetration of Al metal into the interlayer dielectric, leading to a gate-source short circuit. This comprehensive research provides valuable guidance for enhancing the surge reliability of SiC MOSFETs. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

24 pages, 20333 KB

Open AccessArticle

A Novel Fault-Identification Method for Micro Coils of EMECs Based on a Composite Analytical Model Combining a 2D Thermal Model and a 1D-CNN

by Aobo Wang, Jiaxin You, Xu Tan, Yutong Xue and Xinyu Jin

Micromachines 2026, 17(7), 777; https://doi.org/10.3390/mi17070777 - 26 Jun 2026

Viewed by 343

Abstract

This paper proposes a novel fault-identification method for micro-coils in relays with forcibly guided contacts, a type of electromechanical elementary component (EMEC), combining a composite analytical model, a 2D thermal model, and a 1D-CNN. A low-order thermal circuit with one central node and [...] Read more.

This paper proposes a novel fault-identification method for micro-coils in relays with forcibly guided contacts, a type of electromechanical elementary component (EMEC), combining a composite analytical model, a 2D thermal model, and a 1D-CNN. A low-order thermal circuit with one central node and four boundary nodes is established, while a two-dimensional anisotropic Poisson equation is used as a high-order calibration model. The two models are coupled through iterative correction of reusable thermal resistances. For thermal aging, enamel-film delamination, and inter-turn short-circuit faults, thermal-conductivity attenuation, asymmetric branch-resistance perturbation, and localized abnormal heat-source injection are introduced to generate physically constrained temperature sequences. Orthogonal centerline temperature distributions are extracted as one-dimensional feature vectors for 1D-CNN classification. Simulation results show that the hybrid model has an error of approximately 1.7% compared with finite-element results, and the trained 1D-CNN achieves 98.13% accuracy on 160 test samples. Experimental reconstruction and deep-feature visualization further verify its ability to distinguish normal, aging, delamination, and local short-circuit states. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

20 pages, 5179 KB

Open AccessArticle

High-Precision LCCD-Based Focus Metrology for I-Line Lithography: Multi-Sample Repeatability and Adaptability Evaluation

by Hengrui Guan, Xinxin Zhao, Yuheng Chu, Wuhao Liu, Yongxing Yang, Dapeng Kuang, Maoxin Song, Mingchun Ling and Jin Hong

Micromachines 2026, 17(6), 714; https://doi.org/10.3390/mi17060714 - 11 Jun 2026

Viewed by 380

Abstract

Achieving stable local focus-height measurement across different material surfaces is important for I-line-lithography-related inspection, where sub-micrometer height deviations can affect imaging quality, exposure uniformity, and subsequent autofocus performance. This study evaluates the local focus-height repeatability of a linear charge-coupled device (LCCD)-based focus metrology [...] Read more.

Achieving stable local focus-height measurement across different material surfaces is important for I-line-lithography-related inspection, where sub-micrometer height deviations can affect imaging quality, exposure uniformity, and subsequent autofocus performance. This study evaluates the local focus-height repeatability of a linear charge-coupled device (LCCD)-based focus metrology system under several I-line-lithography-related material-surface conditions. The prototype integrates fiber-coupled LED illumination, telecentric projection and imaging optics, reference marks, and a two-step localization procedure based on template matching and centroid estimation; the dual-wavelength source is treated as part of the fixed optical configuration. Tests were performed on silicon wafers, GaAs bright substrates, sapphire, infrared transmissive material, and SiC, covering different reflectivity levels and surface structures. The measured peak-to-valley repeatability was 35–37 nm for highly reflective samples and 40–54 nm for intermediate- or low-reflectivity and microstructured samples, all below the selected 70 nm conservative engineering criterion derived from the depth-of-focus estimate. These results indicate that the integrated LCCD measurement chain maintained stable local repeatability within the tested material-surface range, providing experimental support for further development of local focus metrology and precision optical inspection. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

24 pages, 25588 KB

Open AccessArticle

Development of a Bionic Bistable Compliant Mechanism for the LDI Machine

by Ruizhou Wang, Junhong Li and Hua Wang

Micromachines 2026, 17(6), 640; https://doi.org/10.3390/mi17060640 - 22 May 2026

Viewed by 1203

Abstract

Rigid mechanisms (RMs) are widely adopted in the vision-based measurement (VBM) system of laser direct imaging (LDI) machines. Constant-stiffness compliant mechanisms (CMs) improve the performance of traditional RMs. Unfortunately, constant-stiffness CMs still exhibit high energy consumption and limited adaptability during fast focusing. Inspired [...] Read more.

Rigid mechanisms (RMs) are widely adopted in the vision-based measurement (VBM) system of laser direct imaging (LDI) machines. Constant-stiffness compliant mechanisms (CMs) improve the performance of traditional RMs. Unfortunately, constant-stiffness CMs still exhibit high energy consumption and limited adaptability during fast focusing. Inspired by the hierarchical structure and mechanical behavior of ligaments and tendons, this paper proposes a bionic bistable compliant mechanism (BBCM) to replace constant-stiffness CMs. The BBCM exhibits dynamic stiffness characteristics throughout the focusing stroke, with low stiffness in the transition phase to reduce energy consumption during rapid focusing and high local stiffness near the stable positions to maintain focusing stability. A numerical model is established to analyze the variable-stiffness and bistable characteristics of the proposed BBCM. Prototype tests demonstrate the bistable response, dynamic feasibility, and energy-saving potential of the mechanism. Under the tested camera-loaded flying-shot condition, compared with the constant-stiffness CM, the BBCM reduces electrical and mechanical energy consumption by 12.37% and 9.74%, respectively. The target recognition results indicate that the BBCM-based system maintains comparable visual measurement performance. These results demonstrate that the proposed BBCM provides a feasible mechanism-level solution for energy-efficient dual-position focusing in LDI machines. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

20 pages, 4281 KB

Open AccessArticle

High-Precision Localization Algorithm for Target Symmetry Center in Image-Based Overlay Metrology

by Wuhao Liu, Maoxin Song, Shuming Shi, Mingchun Ling, Hengwei Qin, Hengrui Guan, Jun Wang and Jin Hong

Micromachines 2026, 17(5), 626; https://doi.org/10.3390/mi17050626 - 20 May 2026

Viewed by 595

Abstract

Achieving high-precision overlay target center localization is critical for image-based overlay (IBO) metrology in advanced semiconductor manufacturing. This paper proposes a novel IBO target localization algorithm based on symmetry center matching. Leveraging the symmetry design of the IBO optical system as a physical [...] Read more.

Achieving high-precision overlay target center localization is critical for image-based overlay (IBO) metrology in advanced semiconductor manufacturing. This paper proposes a novel IBO target localization algorithm based on symmetry center matching. Leveraging the symmetry design of the IBO optical system as a physical prior, the algorithm reformulates center localization as a global correlation optimization problem. The grayscale projection profile of a single-sided edge is extracted, spatially mirrored, and used as a reference template for sliding correlation matching against the opposite edge. The symmetry center is then determined from the peak of the Pearson correlation coefficient curve. Simulation results demonstrate a center localization accuracy better than 0.00013 pixels (3

σ

), with repeatability precision remaining within 0.012 pixels (3

σ

) under stringent noise and blur conditions. Experimental validation yields object-space repeatability precision of 0.129 nm (3

σ

) and 0.144 nm (3

σ

) in the X and Y directions, respectively, surpassing the 0.32 nm measurement uncertainty requirement for advanced process nodes. The average single-frame processing time is approximately 0.07 s, demonstrating that the proposed algorithm simultaneously satisfies the demands of high precision and high throughput. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

23 pages, 8071 KB

Open AccessArticle

Intelligent Optimization of Gas-Assisted Electrospinning via LLM-Guided Bayesian Inference

by Jun Zeng, Rongguang Zhang, Weicheng Ou, Xuanzhi Zhang, Shize Huang, Xun Chen and Guojie Xu

Micromachines 2026, 17(5), 619; https://doi.org/10.3390/mi17050619 - 18 May 2026

Viewed by 437

Abstract

Nanofiber-based structures have shown considerable potential in semiconductor-related applications, including ultra-thin dielectric layers and flexible electronic devices, owing to their tunable micro-/nanoscale morphology. However, the manufacturing of these structures is often hindered by the complex multiparameter coupling and poor reproducibility inherent in conventional [...] Read more.

Nanofiber-based structures have shown considerable potential in semiconductor-related applications, including ultra-thin dielectric layers and flexible electronic devices, owing to their tunable micro-/nanoscale morphology. However, the manufacturing of these structures is often hindered by the complex multiparameter coupling and poor reproducibility inherent in conventional electrospinning processes. To address these challenges, this study develops an intelligent optimization framework for gas-assisted electrospinning by integrating Large Language Models (LLMs) with Bayesian Optimization (BO). A Gaussian Process Regression (GPR) surrogate model was established to navigate the high-dimensional parameter space efficiently. Comparative studies demonstrate that the proposed BO+LLM strategy not only outperforms pure data-driven BO and pure knowledge-driven LLM approaches but also surpasses the conventional Response Surface Methodology (RSM) baseline, successfully locating a verified minimum fiber diameter of 239 nm. Furthermore, through response-surface analysis, this work identifies a specific multiphysics collaborative window where electrostatic stretching and aerodynamic assistance are balanced. These findings provide a robust pathway for the reproducible fabrication of nanofiber-based electronic devices. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

19 pages, 6565 KB

Open AccessArticle

Simulation and Response Surface Methodology for Predicting Mass Transfer in Coaxial Electrospun Core-Shell Fibers

by Xun Chen, Weiming Shu, Rongguang Zhang, Shize Huang and Xuanzhi Zhang

Micromachines 2026, 17(5), 606; https://doi.org/10.3390/mi17050606 - 15 May 2026

Viewed by 694

Abstract

Coaxial electrospinning technology enables the fabrication of nanofibers with a core-shell structure, thereby facilitating the encapsulation of functional materials. Its efficacy lies in the precise regulation of mass transfer behavior at the sensing interface. However, achieving the controllable preparation of core-shell fiber structures [...] Read more.

Coaxial electrospinning technology enables the fabrication of nanofibers with a core-shell structure, thereby facilitating the encapsulation of functional materials. Its efficacy lies in the precise regulation of mass transfer behavior at the sensing interface. However, achieving the controllable preparation of core-shell fiber structures in complex environments and quantitatively predicting their mass transfer kinetics remain challenging. This study aims to establish a predictive framework combining simulation and experiment. Firstly, finite element simulations using COMSOL clarified that increasing the shell thickness or decreasing its effective diffusion coefficient can significantly delay analyte transport. A model incorporating time-varying parameters further revealed the influence of polymer swelling on the initial release kinetics. Using the diffusion of an aqueous KCl solution as a model system, experiments confirmed that increasing the shell solution concentration is an effective processing strategy for enhancing the mass transfer barrier. Based on the Box-Behnken design and response surface methodology (RSM), a quantitative model linking key process parameters to release kinetic parameters was established. Model diagnostics indicated that the regression equation is significant and reliable. Validation experiments demonstrated that the model possesses good predictive capability for the key release kinetic parameters, with prediction errors within an acceptable range. The framework established in this study indicates that active design of the mass transfer behavior of core-shell fibers can be achieved through process control, providing a quantitative predictive tool and methodological reference for the preparation of controllable mass transfer interfaces for sensing applications. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

19 pages, 12679 KB

Open AccessArticle

Lightweight Semantic-Guided FCOS for In-Line Micro-Defect Inspection in Semiconductor Manufacturing

by Tao Zhang, Shichang Yan and Gaoe Qin

Micromachines 2026, 17(4), 473; https://doi.org/10.3390/mi17040473 - 14 Apr 2026

Cited by 1 | Viewed by 1044

Abstract

The relentless miniaturization of semiconductor components and Printed Circuit Boards (PCBs) has rendered Automated Optical Inspection (AOI) of micro-defects a critical bottleneck in modern manufacturing and metrology. While in-line inspection systems offer economically viable and scalable quality control solutions, they impose stringent constraints [...] Read more.

The relentless miniaturization of semiconductor components and Printed Circuit Boards (PCBs) has rendered Automated Optical Inspection (AOI) of micro-defects a critical bottleneck in modern manufacturing and metrology. While in-line inspection systems offer economically viable and scalable quality control solutions, they impose stringent constraints on both inference latency and detection robustness—particularly for diminutive, sparsely distributed defects (e.g., mouse bites, pinholes) amidst complex, repetitive circuit topologies. To bridge this gap, we present a semantic-enhanced FCOS framework specifically engineered for micro-defect inspection. Our approach introduces two synergistic innovations: (1) a Semantic-Guided Upsampling Unit (SGU) that adaptively reweights channel–spatial features to reconcile the semantic disparity between shallow textural details and deep contextual representations; and (2) a Sparse Center-ness Calibration (SCC) module that enforces high-confidence, spatially sparse supervision to sharpen localization precision and suppress false positives. The SGU is integrated within a Progressive Semantic-Enhanced Feature Pyramid Network (PSE-FPN) that extends multi-scale representations to stride-4 (P2) resolution, while the SCC module is embedded directly into the detection head. Comprehensive evaluations on MS COCO and the real-world DeepPCB dataset validate the efficacy of our design. On COCO, our model achieves 41.8% AP with real-time throughput of 28 FPS on a single NVIDIA 1080Ti GPU. A lightweight variant further attains 41.6% AP at 42 FPS, accommodating high-throughput production environments. For PCB defect detection, the framework delivers 98.7% mAP@0.5, substantially outperforming contemporary detectors. These results demonstrate that semantics-aware, lightweight architectures enable scalable, real-time quality assurance in semiconductor manufacturing. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

15 pages, 3191 KB

Open AccessArticle

High-Uniformity Core-Shell Nanofibers for Semiconductor Packaging: Process Optimization and Performance Study of Airflow-Assisted Coaxial Electrospinning

by Xun Chen, Shize Huang, Rongguang Zhang, Xuanzhi Zhang, Jiecai Long and Guohuai Lin

Micromachines 2026, 17(4), 463; https://doi.org/10.3390/mi17040463 - 10 Apr 2026

Cited by 1 | Viewed by 574

Abstract

Semiconductor miniaturization demands stricter material uniformity. Core-shell nanofibers, promising for semiconductor packaging and flexible circuits, face application limits due to traditional coaxial electrospinning’s electric field instability—causing poor fiber diameter uniformity and challenges with high-viscosity and low-conductivity solutions. To address this, airflow-assisted coaxial electrospinning [...] Read more.

Semiconductor miniaturization demands stricter material uniformity. Core-shell nanofibers, promising for semiconductor packaging and flexible circuits, face application limits due to traditional coaxial electrospinning’s electric field instability—causing poor fiber diameter uniformity and challenges with high-viscosity and low-conductivity solutions. To address this, airflow-assisted coaxial electrospinning leveraged airflow-electric field synergy to enhance fiber stretching. COMSOL Multiphysics 6.4 simulated the influence of different inner diameters of the air flow nozzles on the air flow field, while the response surface method optimized parameters. At 10 kPa air pressure, 16.71 kV voltage, and a gas nozzle inner diameter of 3.42 mm, nanofibers showed regular morphology with a diameter coefficient of variation as low as 9.2%. This study enables stable preparation of highly uniform core-shell nanofibers, providing key process support for their large-scale semiconductor application and advancing flexible electronics and photodetection. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

12 pages, 2082 KB

Open AccessArticle

Design and Experimental Validation of a Dynamic Frequency Sweeping Algorithm for Optimized Impedance Matching in Semiconductor RF Power Systems Under Pulse-Mode Operation

by Zhaolong Fan, Zhifeng Wang, Long Xu, Lili Hou, Long Yao, Siao Zeng and Mingqing Liu

Micromachines 2026, 17(3), 376; https://doi.org/10.3390/mi17030376 - 20 Mar 2026

Viewed by 948

Abstract

The design and implementation of a dynamic frequency sweeping algorithm for a 3 kW RF power source are underpinned by theoretical principles aimed at optimizing impedance matching under pulse-mode operation. The algorithm dynamically adjusts the output frequency within a predefined range to align [...] Read more.

The design and implementation of a dynamic frequency sweeping algorithm for a 3 kW RF power source are underpinned by theoretical principles aimed at optimizing impedance matching under pulse-mode operation. The algorithm dynamically adjusts the output frequency within a predefined range to align the source impedance Z_source with the conjugate of the load impedance Z*_load, maximizing the power transfer efficiency and minimizing the reflection coefficient Γ. This is achieved by leveraging the maximum power transfer theorem and adapting to dynamic load variations, such as those induced by the plasma state transitions. The algorithm incorporates adaptive step size adjustments based on the rate of change of Γ, predictive frequency initialization using historical data, and real-time impedance monitoring to ensure efficient convergence within the constrained pulse “ON” time (TON). Integration with pulse mode requires synchronization with the pulse signal, fast convergence, and optimized search strategies. Experimental validation on a 13.56 MHz, 3 kW Automatic Sweep Generator testbed operating at 20 kHz pulse modulation with a 50% duty cycle demonstrates a linear and stable sweep, achieving impedance matching and low reflected power within 5.0172 ms. These findings highlight the algorithm’s potential for high-precision applications, such as RF plasma excitation, and underscore the importance of adaptive techniques in dynamic RF systems. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

25 pages, 2611 KB

Open AccessArticle

Noise-Robust Wafer Map Defect Classification via CNN-ESN Hybrid Architecture

by Hayeon Choi, Dasom Im, Sangeun Oh and Jonghwan Lee

Micromachines 2026, 17(3), 309; https://doi.org/10.3390/mi17030309 - 28 Feb 2026

Cited by 1 | Viewed by 1293

Abstract

Wafer map defect classification plays a critical role in yield monitoring and root-cause analysis in semiconductor manufacturing. Although recent convolutional neural network (CNN)-based approaches have achieved high classification accuracy, most existing models are evaluated primarily on clean datasets and remain vulnerable to unseen [...] Read more.

Wafer map defect classification plays a critical role in yield monitoring and root-cause analysis in semiconductor manufacturing. Although recent convolutional neural network (CNN)-based approaches have achieved high classification accuracy, most existing models are evaluated primarily on clean datasets and remain vulnerable to unseen perturbations and representation-level variability at test time. In this paper, we propose a hybrid CNN–echo state network (ESN) architecture that integrates spatial feature extraction with sequential aggregation to enhance robustness under input perturbations. The CNN backbone extracts two-dimensional feature maps, which are converted into ordered sequences using a multidirectional scanline strategy and processed by an ESN reservoir. The resulting sequential representations are combined with CNN features through a class-specific adaptive fusion mechanism. Using the defect-only eight-class version of the WM-811K dataset, we systematically evaluate robustness under multiple perturbation scenarios, with particular focus on the clean train/noisy test (CT-NT) setting. To ensure a controlled robustness evaluation aligned with the binary nature of wafer map data, we introduce binary-consistent die-flip perturbations and additionally employ additive Gaussian perturbations as a representation-level stress test. Under clean-data conditions, the proposed model showed a 0.61 pp improvement in test accuracy compared to the ResNet34-based CNN, with notably larger gains for rare classes and defect types exhibiting strong structural patterns. In the clean train/noisy test scenario, where the model was trained on clean wafer map data and evaluated under controlled test-time perturbations, the accuracy of the CNN baseline dropped to 77.59% at

σ

= 0.10, whereas the proposed hybrid model maintained an accuracy of 87.30%, resulting in an absolute improvement of 9.71 pp. Per-class analysis reveals that the robustness gain is class-dependent, with pronounced improvements for defect types exhibiting clear and repetitive structural patterns, such as Loc and Edge-Ring. Further mechanistic analysis demonstrates that the robustness improvement arises from enhanced representation stability and bounded reservoir dynamics, rather than from changes in CNN feature extraction or training regularization. These results demonstrate that the proposed CNN-ESN hybrid architecture provides meaningful advantages in terms of robustness under noisy evaluation conditions without requiring noise-aware training or prior knowledge of perturbation characteristics. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

22 pages, 3343 KB

Open AccessFeature PaperEditor’s ChoiceArticle

Post-Release Metallization in MEMS Silicon-to-Silicon Contact Switches for On-Resistance Improvement

by Abdurrashid Hassan Shuaibu, Almur A. S. Rabih, Yves Blaquière and Frederic Nabki

Micromachines 2026, 17(3), 288; https://doi.org/10.3390/mi17030288 - 26 Feb 2026

Cited by 1 | Viewed by 1913

Abstract

This work reports a post-release sputter-metallization process for microelectromechanical systems (MEMS) switches with silicon-to-silicon (Si-to-Si) contacts fabricated by deep reactive ion etching. Platinum (Pt) was selectively deposited on the contacting platforms through a perforated mask. Alternatively, aluminum (Al) was deposited over a thin [...] Read more.

This work reports a post-release sputter-metallization process for microelectromechanical systems (MEMS) switches with silicon-to-silicon (Si-to-Si) contacts fabricated by deep reactive ion etching. Platinum (Pt) was selectively deposited on the contacting platforms through a perforated mask. Alternatively, aluminum (Al) was deposited over a thin chromium (Cr) adhesion layer. Electrical measurements showed that Pt enabled a contact resistance on the order of 406

Ω

at a 1

m

A

test current, whereas the resistance of Al/Cr coatings decreased from

7.94

k

Ω

at 1

m

A

to 270

Ω

at 25

m

A

, a change that was potentially linked to oxidation of the Al. These results demonstrated successful coating, with uniform top-surface and edge coverage as revealed by energy-dispersive X-ray spectroscopy imaging. Overall, the results indicate that post-release metallization has the potential to improve the operational repeatability of Si-to-Si contact MEMS switches in static and dynamic tests; the findings also point to process refinements to further optimize contact resistance. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

18 pages, 2983 KB

Open AccessArticle

A Physics-Informed Hybrid Neural Network for High-Precision Temperature Prediction in Semiconductor Process Equipment

by Jiefeng Peng, Liang Hu, Rui Su, Yingnan Shen, Jing Wang, Xin Fu and Xiaodong Ruan

Micromachines 2026, 17(3), 287; https://doi.org/10.3390/mi17030287 - 25 Feb 2026

Cited by 1 | Viewed by 1478

Abstract

High-precision thermal regulation in semiconductor process equipment is critical for product quality, yet it is challenged by actuator transport delays, limited actuator bandwidth due to hardware dynamics, and broadband inlet disturbances in temperature-controlled process fluids. This paper presents a systematic solution integrating architecture [...] Read more.

High-precision thermal regulation in semiconductor process equipment is critical for product quality, yet it is challenged by actuator transport delays, limited actuator bandwidth due to hardware dynamics, and broadband inlet disturbances in temperature-controlled process fluids. This paper presents a systematic solution integrating architecture optimization with a physics-informed hybrid prediction model to enable effective feedforward compensation. Frequency-domain analysis justifies placing the temperature fluctuation attenuator (TFA) upstream of the heater to filter mid-to-high-frequency disturbances without compromising feedback stability. To address actuation delays, a Physics-CNN-LSTM predictor is developed using a residual learning strategy. This framework employs a mechanism model for baseline estimation and a deep learning network to correct persistent low-frequency residuals caused by unmodeled dynamics. Comparative experiments on industrial data demonstrate that the model achieves a Root Mean Square Error (RMSE) of

3.56 \times 10^{- 5}

K under low-to-mid-frequency inlet disturbances, reducing error by approximately 51.8% compared to a standard LSTM. The model also exhibits strong robustness against disturbance frequency shifts (

R^{2} > 0.996

on unseen data). Furthermore, closed-loop simulations confirm that the proposed feedforward compensation enhances temperature stability in high-precision thermal control. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

20 pages, 3262 KB

Open AccessArticle

Glass Fall-Offs Detection for Glass Insulated Terminals via a Coarse-to-Fine Machine-Learning Framework

by Weibo Li, Bingxun Zeng, Weibin Li, Nian Cai, Yinghong Zhou, Shuai Zhou and Hao Xia

Micromachines 2026, 17(1), 128; https://doi.org/10.3390/mi17010128 - 19 Jan 2026

Viewed by 1211

Abstract

Glass-insulated terminals (GITs) are widely used in high-reliability microelectronic systems, where glass fall-offs in the sealing region may seriously degrade the reliability of the microelectronic component and further degrade the device reliability. Automatic inspection of such defects is challenging due to strong light [...] Read more.

Glass-insulated terminals (GITs) are widely used in high-reliability microelectronic systems, where glass fall-offs in the sealing region may seriously degrade the reliability of the microelectronic component and further degrade the device reliability. Automatic inspection of such defects is challenging due to strong light reflection, irregular defect appearances, and limited defective samples. To address these issues, a coarse-to-fine machine-learning framework is proposed for glass fall-off detection in GIT images. By exploiting the circular-ring geometric prior of GITs, an adaptive sector partition scheme is introduced to divide the region of interest into sectors. Four categories of sector features, including color statistics, gray-level variations, reflective properties, and gradient distributions, are designed for coarse classification using a gradient boosting decision tree (GBDT). Furthermore, a sector neighbor (SN) feature vector is constructed from adjacent sectors to enhance fine classification. Experiments on real industrial GIT images show that the proposed method outperforms several representative inspection approaches, achieving an average IoU of 96.85%, an F1-score of 0.984, a pixel-level false alarm rate of 0.55%, and a pixel-level missed alarm rate of 35.62% at a practical inspection speed of 32.18 s per image. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

22 pages, 5744 KB

Open AccessArticle

MCHB-DETR: An Efficient and Lightweight Inspection Framework for Ink Jet Printing Defects in Semiconductor Packaging

by Yibin Chen, Jiayi He, Zhuohao Shi, Yisong Pan and Weicheng Ou

Micromachines 2026, 17(1), 109; https://doi.org/10.3390/mi17010109 - 14 Jan 2026

Cited by 1 | Viewed by 716

Abstract

In semiconductor packaging and microelectronic manufacturing, inkjet printing technology is widely employed in critical processes such as conductive line fabrication and encapsulant dot deposition. However, dynamic printing defects, such as missing droplets and splashing can severely compromise circuit continuity and device reliability. Traditional [...] Read more.

In semiconductor packaging and microelectronic manufacturing, inkjet printing technology is widely employed in critical processes such as conductive line fabrication and encapsulant dot deposition. However, dynamic printing defects, such as missing droplets and splashing can severely compromise circuit continuity and device reliability. Traditional inspection methods struggle to detect such subtle and low-contrast defects. To address this challenge, we propose MCHB-DETR, a novel lightweight defect detection framework based on RT-DETR, aimed at improving product yield in inkjet printing for semiconductor packaging. MCHB-DETR features a lightweight backbone with enhanced multi-level feature extraction capabilities and a hybrid encoder designed to improve cross-scale and multi-frequency feature fusion. Experimental results on our inkjet dataset show a 29.1% reduction in parameters and a 36.7% reduction in FLOPs, along with improvements of 3.1% in mAP@50 and 3.5% in mAP@50:95. These results demonstrate its superior detection performance while maintaining efficient inference, highlighting its strong potential for enhancing yield in semiconductor packaging. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

18 pages, 12671 KB

Open AccessArticle

Numerical Study on Heat Transfer Performance of Turbulence Enhancement Configurations for Galinstan Based Mini-Channel Cooling

by Fajing Li, Junxi Han, Zhifeng Wang, Yi Dai and Peizhu Chen

Micromachines 2026, 17(1), 83; https://doi.org/10.3390/mi17010083 - 7 Jan 2026

Cited by 3 | Viewed by 704

Abstract

The escalating heat flux density and temperature in highly integrated microelectronic devices adversely affect their reliability and service life, making efficient thermal management crucial for stable operation. This study utilizes Galinstan liquid metal as the coolant to investigate the flow and heat transfer [...] Read more.

The escalating heat flux density and temperature in highly integrated microelectronic devices adversely affect their reliability and service life, making efficient thermal management crucial for stable operation. This study utilizes Galinstan liquid metal as the coolant to investigate the flow and heat transfer performance in microchannel heat sinks incorporating various turbulator configurations. It is revealed that for microchannels featuring expanded regions, turbulators that create highly symmetric flow fields are preferable due to improved flow distribution. The long teardrop-shaped turbulator provides the best heat transfer performance among all the investigated heat transfer enhancement structures. And this turbulator yields a 13.8–25.9% higher enhancement effectiveness compared to other configurations, at the expense of a 28–41% increase in pressure loss. However, the sudden cross-sectional expansion in the expanded region causes a significant reduction in fluid velocity. Consequently, microchannels with expanded regions and turbulators exhibit a higher bottom surface temperature than the original, straight microchannels, leading to an overall deterioration in heat transfer performance. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

13 pages, 3866 KB

Open AccessArticle

Near-Field Electrospray ZnO Thin Film for Ultraviolet Photodetectors

by Liyun Zhuo, Tao Peng, Jiaxin Jiang and Gaofeng Zheng

Micromachines 2026, 17(1), 69; https://doi.org/10.3390/mi17010069 - 31 Dec 2025

Cited by 2 | Viewed by 1107

Abstract

ZnO thin-film ultraviolet photodetectors are widely used in the military, space, environmental protection, medicine, and other fields. Accurate printing of ZnO photoelectric-sensitive films plays a key role in the detection results. Therefore, obtaining printing technology with a simple process and high precision has [...] Read more.

ZnO thin-film ultraviolet photodetectors are widely used in the military, space, environmental protection, medicine, and other fields. Accurate printing of ZnO photoelectric-sensitive films plays a key role in the detection results. Therefore, obtaining printing technology with a simple process and high precision has become a challenge for ZnO photoelectrically sensitive films. By adjusting the distance between the nozzle and the collecting plate, the jet is atomized in a straight line and deposited directly on the collecting plate, which effectively improves the stability and controllability of the jet spraying and deposition processes. ZnO thin films with a uniform distribution of nanoparticles, significantly improved density, and controllable deposition area linewidth were successfully prepared. The effects of different ZnO film structures on the performance of ultraviolet photodetectors were tested. When the ultraviolet light intensity is 500, 1000, and 2500 mW/cm², the

I_{l i g h t}

of the photodetector is 4.62, 9.38, 14.67 mA, The on/off ratio (

I_{l i g h t}

/

I_{d a r k}

) is 20.7, 42.1, 65.8, implying satisfactory photoelectric performance as well as high stability and repeatability, providing an effective technical means for the precise printing application of micro-nano functional devices. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Graphical abstract

14 pages, 2545 KB

Open AccessArticle

Study on the Core-Shell Structure of Gas-Assisted Coaxial Electrospinning Fibers: Implications for Semiconductor Material Design

by Rongguang Zhang, Xuanzhi Zhang, Jianfeng Sun, Shize Huang, Xuan Zhang, Guohuai Lin, Xun Chen, Zhifeng Wang, Jiecai Long and Weiming Shu

Micromachines 2026, 17(1), 20; https://doi.org/10.3390/mi17010020 - 24 Dec 2025

Cited by 2 | Viewed by 802

Abstract

Gas-assisted coaxial electrospinning (GACES), a simple and versatile technique for the large-scale fabrication of coaxial nanofiber membranes, possesses significant industrial potential across advanced manufacturing sectors including semiconductors—particularly for fabricating high-precision dielectric layers, high-uniformity encapsulation materials, and flexible semiconductor substrates requiring tailored core-shell architectures. [...] Read more.

Gas-assisted coaxial electrospinning (GACES), a simple and versatile technique for the large-scale fabrication of coaxial nanofiber membranes, possesses significant industrial potential across advanced manufacturing sectors including semiconductors—particularly for fabricating high-precision dielectric layers, high-uniformity encapsulation materials, and flexible semiconductor substrates requiring tailored core-shell architectures. However, there is still a lack of relevant studies on the effective regulation of the core-shell structures of coaxial fibers based on GACES, which greatly limits the batch preparation and wide application of coaxial fibers. Finite element simulation analysis of the flow field and development of the coaxial jet mechanics model with a gas-driven flow field—two key methodologies in this study—successfully uncovered the influence mechanism of gas-assisted flow fields on the core-shell structures of coaxial nanofibers. By adjusting the gas-assisted flow fields parameters, we reduced the total diameter of coaxial fibers by 47.33% (average fiber diameter: 334.12 ± 16.29 nm → 175.98 ± 1.18 nm), decreased the shell thickness by 72.98%, increased the core-shell ratio by 289% (core-shell ratio: 0.49 → 1.91), and improved the uniformity of the total diameter distribution of coaxial fibers by 30.64%. This study delivers a practical conceptual framework and robust experimental underpinnings for the scalable fabrication of coaxial nanofiber membranes with controllable core-shell structures, thereby promoting their practical application in semiconductor devices such as ultra-thin dielectric layers, precisely structured encapsulation materials, and high-uniformity templates for nanoscale circuit patterning. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

12 pages, 5636 KB

Open AccessArticle

Enhancement of Piezoelectric Properties in Electrospun PVDF Nanofiber Membranes via In Situ Doping with ZnO or BaTiO₃

by Zhizhao Ouyang, Jinghua Lin, Renhao Rao, Guoqin Huang, Gaofeng Zheng and Changcai Cui

Micromachines 2026, 17(1), 12; https://doi.org/10.3390/mi17010012 - 23 Dec 2025

Cited by 1 | Viewed by 1306

Abstract

High-performance piezoelectric poly(vinylidene fluoride) (PVDF) has great application potential in the field of microsensors, but achieving efficient polarization remains a challenge. Here, the in situ doping electrospinning technique is employed to enhance the piezoelectric properties by introducing a single dose of zinc oxide [...] Read more.

High-performance piezoelectric poly(vinylidene fluoride) (PVDF) has great application potential in the field of microsensors, but achieving efficient polarization remains a challenge. Here, the in situ doping electrospinning technique is employed to enhance the piezoelectric properties by introducing a single dose of zinc oxide (ZnO) or barium titanate (BaTiO₃,BTO) dopants. The effects of key processing parameters on the morphology of nanofiber membranes were systematically investigated. In addition, the influence of zinc oxide (ZnO) or barium titanate (BTO) dopant concentrations on the piezoelectric properties of PVDF was examined. The microstructure, electrical performance, and β-phase content of the composite membranes were characterized. Results indicate that the composite film with a doping formulation of 16 wt% PVDF and 10 wt% ZnO exhibits optimal overall performance: the β-phase content of PVDF reaches 52.8%, and the output voltage reaches 1.5 V, which is 2.5 times higher than that of the undoped PVDF nanofiber membranes. This study provides an effective doping strategy for the fabrication of high-performance piezoelectric nanofiber membranes. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Figure 1

Review

Jump to: Research

32 pages, 21854 KB

Open AccessReview

Advancements in MV₂O₆-Based Particulate Systems for Solar-Light Water Splitting

by Parnapalle Ravi and Jin-Seo Noh

Micromachines 2026, 17(8), 904; https://doi.org/10.3390/mi17080904 - 29 Jul 2026

Viewed by 309

Abstract

The development of efficient visible-light-driven semiconductor photocatalysts is essential for scalable and sustainable green hydrogen production. Among ternary metal oxides, MV₂O₆ (M = Zn, Ni, Cu, Mn, Co, etc.) metavanadates have attracted considerable interest because of their narrow band gaps [...] Read more.

The development of efficient visible-light-driven semiconductor photocatalysts is essential for scalable and sustainable green hydrogen production. Among ternary metal oxides, MV₂O₆ (M = Zn, Ni, Cu, Mn, Co, etc.) metavanadates have attracted considerable interest because of their narrow band gaps (~1.8–2.5 eV), strong visible-light absorption, and unique edge-sharing VO₆ octahedral framework that promotes charge separation. This review summarizes recent advances in the design, synthesis, and electronic engineering of MV₂O₆-based photocatalysts for solar water splitting. Since direct particulate overall water splitting has only been demonstrated for MnV₂O₆, whereas ZnV₂O₆, NiV₂O₆, and CuV₂O₆ have mainly been investigated as photoelectrodes, both particulate photocatalytic and photoelectrochemical (PEC) systems are critically examined. The review clearly distinguishes these two configurations, highlighting how PEC studies provide valuable insights into charge transport, interfacial processes, and reaction kinetics while recognizing the additional challenges associated with suspension-based photocatalysis. Fundamental crystal structures, electronic band alignments, and charge-transfer characteristics of MV₂O₆ compounds are discussed, followed by recent advances in synthesis strategies, including hydrothermal, sol–gel, and deep eutectic solvent (DES)-assisted methods, together with morphology and defect engineering. Particular attention is given to oxygen-vacancy formation and its influence on visible-light absorption and charge separation. Modification strategies, including elemental doping, cocatalyst loading, and the construction of Z-scheme and step-scheme (S-scheme) heterojunctions, are critically evaluated for improving photocatalytic efficiency. Finally, the review discusses the key challenges that limit practical applications, including unfavorable band-edge positions, rapid carrier recombination, sluggish surface reaction kinetics, photostability, and the need to establish composition–structure–activity relationships. Future perspectives emphasize rational materials design through advanced characterization, theoretical calculations, and scalable synthesis approaches to accelerate the development of efficient MV₂O₆ photocatalysts for solar-driven hydrogen production. Full article

(This article belongs to the Special Issue Emerging Technologies and Applications for Semiconductor Industry)

► Show Figures

Graphical abstract

Journal Menu

Journal Browser

Emerging Technologies and Applications for Semiconductor Industry

Share This Special Issue

Editors

Special Issue Information

Keywords

Benefits of Publishing in a Special Issue

Related Special Issue

Published Papers (22 papers)

Research

Review

Further Information

Guidelines

MDPI Initiatives

Follow MDPI