{"title":"A Knowledge-Enhanced Dual-Timescale Framework for Reliability-Aware Soft Sensing of Crystal Quality in Czochralski Silicon Growth Process","authors":"Jun-Chao Ren;Ding Liu;Yin Wan;Zi-Xing Huang","doi":"10.1109/TSM.2026.3717554","DOIUrl":"https://doi.org/10.1109/TSM.2026.3717554","url":null,"abstract":"In Czochralski silicon crystal growth, the key quality-related indicator <inline-formula> <tex-math>$v/G$ </tex-math></inline-formula> is governed by strong multivariable coupling, nonstationary disturbances, and inherent dual-timescale dynamics, where the pulling rate <inline-formula> <tex-math>$v$ </tex-math></inline-formula> changes within seconds while the axial temperature gradient <inline-formula> <tex-math>$G$ </tex-math></inline-formula> evolves over minutes. These characteristics limit the applicability of purely mechanistic models under process drift and also challenge generic deep temporal models in terms of both prediction accuracy and reliability. To address this issue, this paper proposes a knowledge-enhanced data-driven hybrid framework, termed KE-MoSG-Hybrid, for reliability-aware soft sensing of <inline-formula> <tex-math>$v/G$ </tex-math></inline-formula>. First, physics-informed proxy features are constructed from interface energy balance and multi-window process statistics, and are further used as gating priors to guide feature adaptation. Then, to suppress nonstationary interference, a context-gated multi-kernel Savitzky-Golay module is developed to adaptively decompose the input sequence into trend and residual components. On this basis, a dual-branch fast/slow architecture is designed to model the decoupled dynamics collaboratively, where a global filterbank state space model captures long-term memory and slow drift, while a local Transformer focuses on short-term disturbances and transient nonlinear coupling. In addition, physical consistency constraints is incorporated into the training objective to improve predictive robustness against temporal drift within the tested batch. Finally, uncertainty quantification is achieved through Monte Carlo dropout and consistency inference, enabling calibrated prediction intervals for reliability evaluation. Experiments and ablation studies on a chronological industrial batch dataset demonstrate that the proposed method improves within-batch future-segment prediction accuracy and interval reliability compared with competitive baselines, thereby showing its potential for online monitoring, early warning, and process intervention.","PeriodicalId":451,"journal":{"name":"IEEE Transactions on Semiconductor Manufacturing","volume":"39 4","pages":"832-845"},"PeriodicalIF":2.5,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149399539","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"DEA-GIN: Distance-Enhanced Attentional Graph Isomorphism Network for Efficient Lithography Hotspot Detection","authors":"Junyi Li;Guohao Zhang;Wujian Ye;Tangchen Zhong;Lishun Lu;Chaofeng Zhang","doi":"10.1109/TSM.2026.3722747","DOIUrl":"https://doi.org/10.1109/TSM.2026.3722747","url":null,"abstract":"Lithographic hotspot detection is essential for improving the yield of integrated circuit manufacturing. Although convolutional neural networks have achieved promising performance in this task, their reliance on high-resolution layout images often leads to substantial computational and storage overhead as feature sizes continue to shrink. This paper proposes a compact graph-based hotspot detection method, named Distance-Enhanced Attentional Graph Isomorphism Network (DEA-GIN). In the proposed framework, layout clips are first converted into graph-structured representations, where geometric primitives are modeled as nodes and their spatial relationships are encoded through graph edges. This representation preserves key layout information while significantly reducing the input data volume compared with pixel-based images. To further capture spatial proximity among layout components, an edge distance attention mechanism is designed to modulate message passing according to the Euclidean distances between connected nodes, enabling the model to focus on critical local interactions that are more likely to induce lithographic hotspots. The proposed attention mechanism is integrated into a lightweight Graph Isomorphism Network to improve hotspot classification with limited model complexity. Experimental results on benchmark datasets show that DEA-GIN achieves detection accuracy comparable to representative CNN-based methods, while reducing the model size by one to two orders of magnitude and improving model-inference speed by more than tenfold. Compared with previous GNN-based methods, DEA-GIN provides a more compact model and approximately half the inference time, while maintaining a favorable recall–false-alarm trade-off. These results demonstrate the potential of DEA-GIN for scalable and efficient hotspot detection in advanced VLSI design flows.","PeriodicalId":451,"journal":{"name":"IEEE Transactions on Semiconductor Manufacturing","volume":"39 4","pages":"800-810"},"PeriodicalIF":2.5,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149399540","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Flow Matching-Based Inverse Lithography for Efficient Mask Optimization","authors":"Qiao Xu;Lian Huai;Yuyu Liu;Xingqun Jiang","doi":"10.1109/TSM.2026.3723994","DOIUrl":"https://doi.org/10.1109/TSM.2026.3723994","url":null,"abstract":"With the continuous reduction of feature sizes in integrated circuits, the optical proximity effect has become a critical factor limiting the yield of chip manufacturing. As a vital resolution enhancement technique, inverse lithography technology enables pixel-level optimization of masks. However, traditional methods typically exhibit weak generalization and robustness when encountering new layouts with variations in shape, density, or position; moreover, the iterative refinement processes significantly degrade mask optimization efficiency. In this paper, we develop a Flow Matching-based Inverse Lithography Technique (FM-ILT) framework. A lithography simulation model is embedded into the training framework to constrain the fidelity and complexity of mask optimization via lithography regularization loss terms. The proposed FM-ILT learns probabilistic paths from the data to characterize mask optimization rules, enabling adaptive optimization for unseen layout patterns. The deterministic optimal transport path accelerates model convergence and stabilizes the training process. Experimental results on the <italic>LithoBench</i> benchmark demonstrate that the proposed framework generates masks with excellent imaging fidelity, exhibiting strong generalization to unseen patterns and robustness to process variations. Furthermore, the end-to-end architecture significantly enhances mask optimization efficiency.","PeriodicalId":451,"journal":{"name":"IEEE Transactions on Semiconductor Manufacturing","volume":"39 4","pages":"854-866"},"PeriodicalIF":2.5,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149399542","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Effects of Pulse Repetition Rate on Active Species Production for Nanosecond Spark Discharges in Air","authors":"Iliane Hifi;Igor Timoshkin;Graeme Burt","doi":"10.1109/TPS.2025.3650404","DOIUrl":"https://doi.org/10.1109/TPS.2025.3650404","url":null,"abstract":"Nanosecond pulsed discharges are capable of producing highly transient nonequilibrium plasma in gases which are characterized by high electron temperatures and relatively low rotational and vibrational temperatures. These conditions enable the formation of unstable molecules, namely radicals, which are beneficial to a wide range of biomedical and environmental applications including air and water purification, bio-decontamination, and advanced oxidation processes. This research investigates the effects of repetition rate of ns discharges in static air and the average power delivered into the plasma on the production of the hydroxyl radical (OH<inline-formula> <tex-math>$cdot $ </tex-math></inline-formula>) and the nitric oxide radical (NO<inline-formula> <tex-math>$cdot $ </tex-math></inline-formula>) through optical emission spectroscopy (OES). Plasma pulses were generated with voltage pulses with a constant voltage rate of rise (dV/dt) of 9 kV/<inline-formula> <tex-math>$mu $ </tex-math></inline-formula>s resulting in the formation of a spark discharge with current impulse duration of 16 ns. These pulses were applied at repetition rates from 5 thousand pulses per second (kpps) to 38 kpps. The individual discharge energy was observed to decrease as a result of increasing pulse repetition rate, showing a drop from 1.17 mJ per pulse to 0.28 mJ per pulse at 5 and 38 kpps, respectively. Despite this decrease, the average power delivered into the plasma increased as a result of pulse repetition rate from 5.80 to 11.56 W. The optical emission spectra from both the OH<inline-formula> <tex-math>$cdot $ </tex-math></inline-formula> and NO<inline-formula> <tex-math>$cdot $ </tex-math></inline-formula> radicals, and atomic oxygen showed a strong dependence on pulse repetition rate for low repetition rates, followed by a clear correlation with power delivered into the plasma for repetition rates over 15 kpps. The plasma rotational and vibrational temperatures were obtained using 3 methods, with a high resolution N<sub>2</sub> second positive system (SPS) measurement used to evaluate the baseline temperatures. This was compared with a low resolution N<sub>2</sub> SPS measurement using a local thermal equilibrium (LTE) assumption, which showed an error of over 1000 K between the LTE <inline-formula> <tex-math>$T {_{text {vib}}}$ </tex-math></inline-formula> and the actual <inline-formula> <tex-math>$T {_{text {vib}}}$ </tex-math></inline-formula>, proving the limitations of the LTE assumption with relation to transient spark discharges. The rotational temperature of OH<inline-formula> <tex-math>$cdot $ </tex-math></inline-formula> was evaluated and found to follow a non-Boltzmann behavior, resulting in an error in <inline-formula> <tex-math>$T {_{text {rot}}}$ </tex-math></inline-formula> of up to 189% as compared to the N<sub>2</sub> SPS <inline-formula> <tex-math>$T {_{text {rot}}}$ </tex-math></inline-formula>. This study provides clear insights into the o","PeriodicalId":450,"journal":{"name":"IEEE Transactions on Plasma Science","volume":"54 11","pages":"5630-5640"},"PeriodicalIF":1.6,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149383756","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}
Ann Maria George;S. Shilpa;P. S. Abishek;Manesh Michael
{"title":"Impact of Collisions and Exchange-Correlation Effects on the Dynamics of Nonlinear Waves in Dense Laser-Produced Plasma","authors":"Ann Maria George;S. Shilpa;P. S. Abishek;Manesh Michael","doi":"10.1109/TPS.2026.3736230","DOIUrl":"https://doi.org/10.1109/TPS.2026.3736230","url":null,"abstract":"The propagation of 1-D nonlinear waves, specifically solitons and shocks, for a three-component highly intense laser-produced plasma is studied. The three components comprise stationary background ions and two groups of electrons (hot and cold electrons). The hot and cold electrons are incorporated with collisional and exchange-correlation effects, respectively. The influence of quantum forces, including the electrostatic force, pressure gradient, Bohm potential, and viscous effects, on wave propagation is systematically analyzed. Quantum hydrodynamic (QHD) modeling, along with the reductive perturbation method, is employed to derive the damped Korteweg–de Vries (KdVs) and Korteweg–de Vries–Burgers (KdVBs) equations governing the nonlinear evolution of the plasma waves such as solitons and shocks, respectively. Results demonstrate that lower hot electron collision frequency and higher electron degeneracy enhance soliton amplitude, while strong shock localization is favored by high cold electron viscosity and low electron degeneracy. Exchange-correlation effects, governed by cold electron parameters, significantly affect both soliton and shock dynamics. These results are thus crucial for interpreting plasma behavior in extreme conditions, predicting wave-mediated processes, and designing laboratory experiments.","PeriodicalId":450,"journal":{"name":"IEEE Transactions on Plasma Science","volume":"54 11","pages":"5850-5860"},"PeriodicalIF":1.6,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149383886","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}
{"title":"Parametric Coil Design Methodology and Innovative Pulse Optimization Technique for Enhanced Force and Velocity Profiles in Coilguns","authors":"Vasimmahamad H. Ghanchi;Ketan P. Badgujar","doi":"10.1109/TPS.2026.3735371","DOIUrl":"https://doi.org/10.1109/TPS.2026.3735371","url":null,"abstract":"Electromagnetic coilguns, comprising reluctance-based and induction-based configurations, offer a promising alternative to conventional projectile launchers for defense, aerospace, and industrial applications. The literature lacks a unified methodology for optimizing key coil parameters—winding layers, turns per layer, and wire gauge—while conventional capacitor-based pulsed supplies often suffer uncontrolled discharge, armature capture, and reduced efficiency. This study addresses both challenges through a structured parametric coil-optimization methodology combined with a controlled pulse-optimization strategy for improved force–time alignment, evaluated for reluctance and induction coilguns using iron, copper, and aluminum projectiles at 400–1200 V, using a coupled field–circuit model in ANSYS Maxwell with a linear B–H permeability model, validated through MATLAB and Twin Builder co-simulation. Optimized excitation increased maximum velocity from 59.5 to 75 m/s for the reluctance launcher and 21 to 35 m/s for the induction launcher. The methodology was extended to a two-stage reluctance launcher, where optimizing intercoil spacing, trigger delay, and excitation polarity identified an optimum 40-mm gap, producing exit velocities of about 165 m/s, twofold over the optimized single-stage case. Benchmarking against literature configurations shows the framework enhances performance while mitigating armature capture. Experimental verification was carried out on a laboratory-scale single-stage reluctance coilgun operated at low to moderate voltage and current levels, for both conventional capacitor-based pulsed power supply (CPPS) and optimized multipulse excitation, demonstrating good agreement with simulation trends. Numerical and experimental results confirm clear improvements in electromagnetic force continuity, energy utilization, and projectile velocity.","PeriodicalId":450,"journal":{"name":"IEEE Transactions on Plasma Science","volume":"54 11","pages":"6126-6141"},"PeriodicalIF":1.6,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149383930","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}
{"title":"Active-Learning Spectral Transformer-Driven Design of a Polarization-Insensitive Ultrabroadband Single-Layer Graphene THz Absorber","authors":"Ajay Kumar Dwivedi;Ankita Shukla;Nagesh Kallollu Narayanaswamy;Vivek Singh","doi":"10.1109/TPS.2026.3735713","DOIUrl":"https://doi.org/10.1109/TPS.2026.3735713","url":null,"abstract":"Designing an ultrabroadband terahertz (THz) absorber through repeated full-wave simulations is computationally demanding, particularly when the complete spectral response and robustness requirements must be considered together. This work proposes an ensemble-guided active-learning spectral Transformer, termed ALSpecFormer, for the design of a polarization-insensitive single-layer graphene absorber. The model predicts the complete 181-point absorption spectrum directly from the three geometric parameters of a concentric disk–ring resonator. An ensemble of five independently trained models evaluates the reliability of each predicted spectrum through their level of agreement and identifies promising geometries. In each active-learning round, five selected geometries are verified in HFSS, and the resulting spectra are added to the training set for model refinement. The closed-loop procedure is repeated for five rounds, while the validation and test sets remain unchanged. On the fixed test set, ALSpecFormer achieves an MAE of 0.01334 and <inline-formula> <tex-math>$R^{2}=0.98736$ </tex-math></inline-formula>, outperforming TabPFN and Extra Trees. Full-wave verification confirms absorption above 90% from 2.7 to 5.7 THz, corresponding to a fractional bandwidth of 71.43%. The optimized absorber preserves its principal resonant features under polarization rotation and shows useful angular tolerance, with better stability under transverse-magnetic (TM) incidence than under transverse-electric (TE) incidence.","PeriodicalId":450,"journal":{"name":"IEEE Transactions on Plasma Science","volume":"54 11","pages":"6179-6190"},"PeriodicalIF":1.6,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149383938","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}
{"title":"Study on the Impact of Adding Transverse Slots on Rail-Armature Contact Characteristics and Repeatability of Railgun Launch Experiments","authors":"Dongdong Zhang;Kejiang Zhou;Yuan Zhou;Youjun Kong","doi":"10.1109/TPS.2026.3736572","DOIUrl":"https://doi.org/10.1109/TPS.2026.3736572","url":null,"abstract":"During the development of railgun technology, the repeatability of launch experiments has become a significant issue, partly due to the poor initial contact characteristics between the rail and the armature, which is reflected in an excessively low proportion of effective contact area (PECA) and a low uniformity of contact pressure distribution. In this article, an optimization method is proposed to enhance the repeatability of launch experiments by increasing the contact pressure values and uniformity through the addition of transverse slots on the surface of the armature tails. To investigate the impact of adding transverse slots on contact characteristics, a finite element simulation model of the initial interference contact between the rail and the armature was established. A comparative analysis of contact characteristics was conducted before and after the addition of transverse slots; also, the optimization effects and stress distribution of different slot placement, depths, and width-to-spacing ratios on contact characteristics were intensively studied. Based on the simulation parameters in this article, when the slotting depth is 3 mm (the optimal condition), the PECA increases from 23.4% (without slots) to 38.8%, and the coefficient of variation (CV) of contact pressure decreases from 0.71 to 0.49. These results indicate a significant improvement in contact uniformity between the rail and armature.","PeriodicalId":450,"journal":{"name":"IEEE Transactions on Plasma Science","volume":"54 11","pages":"6119-6125"},"PeriodicalIF":1.6,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149384053","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}
Song Jiang;Ziqiang Wang;Yonggang Wang;Mei Ma;Jiang Wu;Zhonghang Wu
{"title":"Research on Water Treatment by Liquid-Phase Discharge Plasma Synergistic With UVC","authors":"Song Jiang;Ziqiang Wang;Yonggang Wang;Mei Ma;Jiang Wu;Zhonghang Wu","doi":"10.1109/TPS.2026.3736639","DOIUrl":"https://doi.org/10.1109/TPS.2026.3736639","url":null,"abstract":"Low-temperature plasma technology suffers from inherent drawbacks in water treatment, including short lifetime of <inline-formula> <tex-math>$cdot $ </tex-math></inline-formula>OH radicals, low utilization efficiency of reactive species and excessive energy consumption. A coupled water treatment system integrating underwater needle-plate discharge plasma and UVC irradiation was established in this study. Indigo carmine (IC) was selected as the simulated organic pollutant. Systematic investigations were carried out to reveal the effects of UVC spatial configuration and pulse frequency on degradation efficiency, energy consumption, and mineralization performance. The results showed that the synergistic system exhibited significant synergistic effects, achieving an IC degradation rate of 82% within 20 min at 14 kV, with a pulsewidth of <inline-formula> <tex-math>$0.8~mu $ </tex-math></inline-formula>s, a frequency of 1 kHz, and the underwater UVC configuration, significantly higher than the 30.62% of the plasma system alone. Using a quasi-first-order kinetic model for fitting, the apparent rate constant of the synergistic system reached 0.07552 min<sup>−1</sup>, which is 4.2 times that of the single-plasma system, and its synergistic factor S reached 3.41. The total organic carbon (TOC) results indicate that the synergistic system exhibits better mineralization ability. After 60 min of reaction, TOC decreased from 25.2 to 18.8 mg/L, with a removal rate of 25.4%, corresponding to a mineralization synergy factor of 1.66. As the frequency increases, the system achieves higher degradation efficiency while also significantly reducing the electrical energy per order (EEO). Combined with experimental results on the concentration variations of H<sub>2</sub>O<sub>2</sub> and <inline-formula> <tex-math>$cdot $ </tex-math></inline-formula>OH, it can be concluded that UVC can photolytically decompose intermediate products such as H<sub>2</sub>O<sub>2</sub> generated during the discharge process, promoting the continuous regeneration of <inline-formula> <tex-math>$cdot $ </tex-math></inline-formula>OH. Meanwhile, the tert-butanol (TBA) quenching experiment further confirmed the dominant role of <inline-formula> <tex-math>$cdot $ </tex-math></inline-formula>OH in the degradation process. Compared to other advanced oxidation technologies, the synergistic system developed in this study not only accelerates the conversion of key intermediates but also offers significant advantages, such as operational simplicity and high energy utilization efficiency.","PeriodicalId":450,"journal":{"name":"IEEE Transactions on Plasma Science","volume":"54 11","pages":"5918-5925"},"PeriodicalIF":1.6,"publicationDate":"2026-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149384144","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"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}