IEEE Transactions on Nuclear Science最新文献

{"title":"Development of a Position-Sensitive Monolithic Planar Scintillator for Gamma-Ray Imaging","authors":"Chengshuai Tian;Jian Yang;Guoqiang Zeng;Lie Ni;Xinyu Yang;Haowen Deng;Chuanhao Hu","doi":"10.1109/TNS.2026.3724193","DOIUrl":"https://doi.org/10.1109/TNS.2026.3724193","url":null,"abstract":"In the development of Compton cameras, traditional pixelated scintillator designs face escalating complexity and cost as detector size increases. To address this, a side-readout monolithic planar scintillator scheme is proposed, where silicon photomultiplier (SiPM) arrays are coupled to the four edges. This configuration maximizes the sensitive detection volume while significantly reducing the photodetector requirement from <inline-formula> <tex-math>$N^{2}$ </tex-math></inline-formula> to <inline-formula> <tex-math>$4N$ </tex-math></inline-formula>. A prototype utilizing a <inline-formula> <tex-math>$96 times 96 times 6 ~text {mm}^{3}$ </tex-math></inline-formula> CsI(Tl) monolithic scintillator was constructed to achieve 2-D position sensing using only four electronic channels. To optimize performance, a geometric mean method was employed to correct energy peak drift caused by position-dependent signal variations. Furthermore, a neural network model was introduced to further improve position resolution. Experimental results with 662 keV gamma rays demonstrate a 2-D position resolution of approximately 4.5 mm and an average energy resolution of 7.1%. These findings confirm that the proposed design significantly reduces readout complexity without compromising the position and energy resolutions required for Compton imaging, offering a cost-effective solution for large-area gamma-ray imaging and radiation monitoring.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 10","pages":"4070-4081"},"PeriodicalIF":2.4,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149219330","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":"Single-Event Burnout and Leakage in SiC MOSFETs Under Front-Side TPA Femtosecond Laser Irradiation","authors":"Manyi Ji;Lingxu Kong;Zhengjia Chen;Yanjun Li;Hongyi Xu;Kuang Sheng;Na Ren","doi":"10.1109/TNS.2026.3730831","DOIUrl":"https://doi.org/10.1109/TNS.2026.3730831","url":null,"abstract":"This work investigated single-event burnout (SEB) and single-event leakage current from drain to source (SELC-II) in silicon carbide (SiC) MOSFETs under femtosecond laser irradiation based on two-photon absorption (TPA), with the aim of investigating how laser-induced carrier generation and distribution affect single-event effect (SEE) responses. The devices were irradiated from the front side under controlled laser irradiation conditions, including different laser pulse energies and relative displacements (RDs) of the focal plane, to vary the amount and spatial location of laser-induced carriers. No irradiation-relevant gate damage was observed, indicating that the TPA laser irradiation induced specific SEE responses without introducing observable gate degradation. Results showed that the SEB threshold voltage depended on laser-induced carrier generation conditions and device voltage ratings (1.2, 1.7, and 3.3 kV). The SEB threshold voltage of 1.2 kV SiC MOSFETs exhibited a non-monotonic dependence on RD, with the highest sensitivity observed at RD <inline-formula> <tex-math>$= 3~mu $ </tex-math></inline-formula>m, decreasing from 550 V (RD <inline-formula> <tex-math>$= 0~mu $ </tex-math></inline-formula>m) to 450 V (RD <inline-formula> <tex-math>$= 3~mu $ </tex-math></inline-formula>m) and then increasing to 1050 V (RD <inline-formula> <tex-math>$= 6~mu $ </tex-math></inline-formula>m). These behaviors were interpreted in terms of the effective carrier population within the depletion region, which was influenced by the amount and spatial distribution of laser-generated carriers under different laser energies and focal-plane positions. The observed trends were consistent with reported single-event behavior under low-linear energy transfer (LET) heavy-ion irradiation, highlighting the role of carrier density in understanding the relationship between laser-induced and low-LET heavy-ion-induced SEE responses.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 10","pages":"3946-3954"},"PeriodicalIF":2.4,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149219315","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":"CPRE64-CZT: A Low-Noise 64-Channel Front-End ASIC With Energy and Timing Readout for CdZnTe Detectors","authors":"Haolong Liao;Ke Wang;Na Wang;Jiale Cai;Xiaoxuan Li;Xiaopan Jiang;Cunfeng Wei;Zhiming Zhang;Lei Shuai;Yao Yang;Yi Liu;Long Wei;Xiaohui Li","doi":"10.1109/TNS.2026.3733362","DOIUrl":"https://doi.org/10.1109/TNS.2026.3733362","url":null,"abstract":"This article presents CPRE64-CZT, a low-noise 64-channel front-end readout application-specific integrated circuit (ASIC) for cadmium zinc telluride (CdZnTe or CZT) detector systems that require configurable-polarity cathode and anode readout. Each channel integrates a low-noise charge-sensitive amplifier (CSA), a high-order shaper, a peak-detect-and-hold (PDH) circuit for energy measurement, and a sample-and-hold circuit for timing measurement, with power dissipation below 5 mW per channel. The three-side pad arrangement supports compact detector coupling and enables expansion to 128 channels with limited area overhead. With a biased CZT detector connected, the measured equivalent noise charge (ENC) is below 150 e<inline-formula> <tex-math>${}^{-}$ </tex-math></inline-formula> rms for the anode channels and below 250 e<inline-formula> <tex-math>${}^{-}$ </tex-math></inline-formula> rms for the cathode channel. The readout chain achieves an integral nonlinearity (INL) within 1.5% over a maximum input charge range of 350 fC. In ¹³⁷Cs measurements with a pixelated CZT detector, depth-dependent energy correction using the cathode-to-anode (C/A) ratio improved the summed 662-keV energy resolution to 1.00% FWHM, with the best-performing pixel reaching 0.78% FWHM. These results demonstrate low-noise spectroscopic readout with energy and timing observables suitable for subsequent detector-level depth correction and position-sensitive CZT analysis.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 10","pages":"4059-4069"},"PeriodicalIF":2.4,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149219384","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":"Single-Event Effects Characterization of Integer-N and Fractional-N PLLs With Unified Architecture for Space-Based Microprocessor in Sub-20-nm FinFET Technology","authors":"Zhenhua Zhang;Hanhan Sun;Zirui Wu;Shuqiang Zhao;He Yin;Bin Liang;Xiuhai Cui;Xingji Li","doi":"10.1109/TNS.2026.3735820","DOIUrl":"https://doi.org/10.1109/TNS.2026.3735820","url":null,"abstract":"This study presents a comparative characterization of single-event effects (SEEs) in integer-N and fractional-N phase-locked loops (PLLs) implemented with the same radiation-tolerant proportional-integral current-path architecture in a sub-20-nm FinFET technology for space-based microprocessor applications. The integer-N PLL and fractional-N PLL have frequency tuning ranges of 0.8–3.2 and 1.6–6.4 GHz, respectively. Preirradiation electrical measurements were performed to establish the baseline clock stability for SEE characterization. Additional same-chip phase-noise measurements show that, at the tested fractional operating point of N = 80.5 and <inline-formula> <tex-math>$alpha = 0.5$ </tex-math></inline-formula>, the integrated rms jitter increases from 9.59 ps in integer mode to 13.68 ps in fractional mode. Separately, the measured period-jitter values of the monitored postdivided baseline clocks are 3.679 and 3.979 ps for the integer-N and fractional-N PLL implementations, respectively, providing baseline clock-stability information for the subsequent SEE characterization. The PLL chip underwent heavy-ion irradiation testing with LET values ranging from 13.4 to 93 MeV <inline-formula> <tex-math>$cdot $ </tex-math></inline-formula> cm²/mg. At LET values of 93 and 83 MeV <inline-formula> <tex-math>$cdot $ </tex-math></inline-formula> cm²/mg for the integer-N and fractional-N PLLs, respectively, no irrecoverable loss-of-lock events were observed. The measured SEE timing-event cross sections, determined using the predefined period-deviation criterion, were <inline-formula> <tex-math>$3times 10^{-7}$ </tex-math></inline-formula> cm² and <inline-formula> <tex-math>$5times 10^{-7}$ </tex-math></inline-formula> cm², respectively. In addition, pulsed-laser testing was performed to investigate the SEE sensitivity of both the complete PLL architecture and individual modules. The laser energy threshold for SEE-induced loss of lock reached 1.6 nJ. The results demonstrate that the proposed unified architecture provides comparable SEE tolerance under the tested integer-N and fractional-N configurations, and that the fractional-N extension, including the delta-sigma modulator (DSM) module, does not exhibit a dominant SEE-sensitive region under the tested conditions.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 10","pages":"3906-3917"},"PeriodicalIF":2.4,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149219413","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":"Different Single-Event Burnout Behavior in p-GaN HEMTs Under Zero and Negative Gate Bias","authors":"Xiangjie Cao;Fengkai Liu;Zhongli Liu;Boshang Lu;Zefeng Wang;Lei Wu;Yuanbo Jiao;Hao Jiang;Pengfei Wan;Jianqun Yang;Xingji Li","doi":"10.1109/TNS.2026.3729980","DOIUrl":"https://doi.org/10.1109/TNS.2026.3729980","url":null,"abstract":"High-energy heavy-ion-induced single-event burnout (SEB) in p-GaN high-electron-mobility transistors (HEMTs) under zero and negative gate bias was investigated. Under negative gate bias, simultaneous transient surges occur in both drain and gate currents during irradiation, whereas only a drain current surge manifests under zero gate bias conditions. This distinct behavior indicates bias-dependent SEB mechanisms. Through technology computer-aided design (TCAD) simulations, we reveal that the redistribution of heavy-ion-induced charge critically governs poststrike electric field evolution. Electron cluster accumulation combined with the back-channel effect caused the high electric field near the drain, resulting in the SEB under zero gate bias. Furthermore, analysis of the p-GaN gate energy band structure demonstrates that partial turn-on of the metal/p-GaN Schottky diode caused the high electric field within the p-GaN layer. Negative gate bias intensifies this electric field, thereby promoting breakdown beneath the gate electrode. Simulation results align with the experimental observations. This study provides new physical insights into SEB mechanisms in p-gallium nitride HEMTs (GaN HEMTs).","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 10","pages":"4014-4022"},"PeriodicalIF":2.4,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149219318","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":"Machine Learning-Based COTS Device Lot Identification and Total Ionizing Dose Degradation Prediction via Enhanced Electrical Stress Features","authors":"Jicong Yang;Qi Guo;Xiaolong Li;Jie Feng;Qiwen Zheng;Lin Wen;Yudong Li","doi":"10.1109/TNS.2026.3725537","DOIUrl":"https://doi.org/10.1109/TNS.2026.3725537","url":null,"abstract":"Commercial off-the-shelf (COTS) devices are widely used in aerospace electronic systems, but their process design does not meet the application requirements of the space radiation environment, leading to performance degradation risks caused by total ionizing dose (TID) effects. However, traditional radiation hardness (RH) assessment methods are costly and time-consuming, and fail to effectively address lot-to-lot and within-lot fluctuations in the RH consistency of COTS devices. This article proposes a machine learning model based on physical feature enhancement. A high-quality dataset is constructed via irradiation experiments on devices from multiple manufacturers and lots. By introducing electrical parameter responses under multiple electrical stress conditions as enhanced feature parameters, the model realizes non-destructive identification of device manufacturer and lot identification and prediction of total-dose radiation degradation. Results show that the model achieves identification accuracy of approximately 0.965 for device manufacturers and 0.836 for lots; at four specific dose points, the coefficient of determination <inline-formula> <tex-math>$left ({ mathbf {R}^{mathbf {2}}}right)$ </tex-math></inline-formula> for radiation degradation prediction is above 0.833, outperforming the mean baseline model. Incorporating electrical parameter responses under multiple electrical stress conditions improves the model’s performance in manufacturer and lot identification and TID radiation degradation prediction. This study reveals that differences in the pre-irradiation initial electrical parameters of devices have an implicit correlation with their RH characteristics. Compared with a single test condition, the responses of device parameters to multiple electrical stresses contain richer RH feature information.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 10","pages":"3991-4002"},"PeriodicalIF":2.4,"publicationDate":"2026-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149219324","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":"A Rapid Reconstruction Method of Gamma Radiation Field Based on Normalized Proper Orthogonal Decomposition","authors":"Kai Tan;Hojoon Son;Fan Zhang","doi":"10.1109/TNS.2026.3718773","DOIUrl":"https://doi.org/10.1109/TNS.2026.3718773","url":null,"abstract":"In nuclear facilities, rapid and reliable reconstruction of gamma radiation fields is crucial for post-incident safety assessment and repair planning, yet direct human intervention is often impossible because of high dose rates. This task remains challenging because of unknown source distributions, complex shielding effects, and operational constraints on sensor deployment. To address these challenges, this work proposes a robot-assisted radiation field-mapping framework that equips a mobile robot with gamma detectors and couples normalized proper orthogonal decomposition (NPOD) with a minimax adaptive selection of measurement locations. By integrating the model-reduction capability of NPOD with intelligent measurement-location optimization, the proposed approach reconstructs spatially complex gamma radiation fields with high fidelity and low computational cost from sparse measurements. In a representative configuration with 160 robot stops selected from 9948 candidate spatial locations, the framework achieves a mean absolute relative reconstruction error of 0.28%. A controlled 2-D experiment further confirms the feasibility of the method and demonstrates that the uncertainty-aware weighted least squares (WLSs) formulation reduces the weighted root mean square error (WRMSE) compared with a naive variant that ignores measurement uncertainty. Compared to the genetic algorithm (GA) baseline, the proposed adaptive measurement-location selection approach requires fewer objective evaluations and achieves a substantially lower worst case reconstruction error without exhaustive search. Owing to its lightweight online reconstruction cost, the framework provides a practical basis for rapid radiation-field mapping and guided inspection in hazardous nuclear environments.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 9","pages":"3579-3590"},"PeriodicalIF":2.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148974738","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":"Design and Validation of a High-Power Radiation-Hardened DC/DC Converter Based on Bidirectional Magnetic Isolation","authors":"Xiaojin Lu;Gang Dong;Qifeng He;Ziyu Zhou;Xinfang Liao","doi":"10.1109/TNS.2026.3723944","DOIUrl":"https://doi.org/10.1109/TNS.2026.3723944","url":null,"abstract":"This article proposes a bidirectional magnetically isolated feedback technique for isolated regulation. By using an isolation transformer to transfer feedback/control information bidirectionally via magnetic coupling, it enables a fully isolated closed-loop regulation and, at the mechanism level, avoids regulation drift and failure risks caused by optocoupler current transfer ratio (CTR) degradation under ionizing radiation. Based on this technique, a 150-W isolated DC/DC converter is designed and implemented for cost-sensitive low-Earth-orbit (LEO) platforms. The converter adopts thick-film hybrid integrated circuit (HIC) technology and a single-ended forward power stage, with the magnetic feedback link driven by radiation-tolerant analog components. The prototype delivers a regulated 28-V output over a 16–50 V input range, achieves a power density of 100 W/in³ in a <inline-formula> <tex-math>$63.5times 38.3times 10.16$ </tex-math></inline-formula> mm³ hermetic package, and maintains <inline-formula> <tex-math>$ge 85$ </tex-math></inline-formula>% full-load efficiency at room temperature. To validate the reliability of this technical route, single-event effect (SEE) and total ionizing dose (TID) tests were conducted: the initial SEE campaign at the Heavy Ion Research Facility in Lanzhou (HIRFL) identified single-event-induced overstress risks in the auxiliary supply and secondary feedback circuitry, and targeted hardening measures were implemented accordingly; after hardening, no catastrophic failures were observed. TID testing up to 100 krad(Si) shows strong dose-rate dependence, with more pronounced degradation at 0.01 rad(Si)/s, consistent with enhanced low-dose-rate sensitivity (ELDRS).","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 9","pages":"3618-3628"},"PeriodicalIF":2.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148974752","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":"Experimental Investigation of Terrestrial-Neutron-Induced Single-Event Burnout on Carrier Stored Trench-Gate Bipolar Transistor (CSTBT)","authors":"Xiao-Feng Huang;Chen-Ming Li;Li-Hua Mo;Zhi-Liang Hu;Hai-Bin Wang","doi":"10.1109/TNS.2026.3716340","DOIUrl":"https://doi.org/10.1109/TNS.2026.3716340","url":null,"abstract":"In this article, the sensitivity of terrestrial-neutron-induced single-event burnout (SEB) in carrier stored trench-gate bipolar transistors (CSTBTs) under different bias conditions is systematically analyzed. Neutron irradiation experimental results show that, when the gate-emitter voltage (<inline-formula> <tex-math>$V_{mathrm {GE}}$ </tex-math></inline-formula>) is set to 0 V, the SEB failure rates of the CSTBTs at each collector-emitter voltage (<inline-formula> <tex-math>$V_{mathrm {CE}}$ </tex-math></inline-formula>) are lower than those of the conventional trench-gate field-stop insulated gate bipolar transistors (Conv. TFS-IGBTs). As <inline-formula> <tex-math>$V_{mathrm {GE}}$ </tex-math></inline-formula> decreases, the SEB failure rates of both the CSTBTs and the Conv. TFS-IGBTs increase, with the SEB sensitivity of the CSTBTs depending more strongly on <inline-formula> <tex-math>$V_{mathrm {GE}}$ </tex-math></inline-formula>. To investigate the differences in the SEB response between the two devices, SEB simulations under various bias conditions were performed using a TCAD simulation tool. The simulation results reveal that the carrier stored layer (CSL) reduces the probability of the latch-up effect in the CSTBTs at <inline-formula> <tex-math>$V_{mathrm {GE}} = 0$ </tex-math></inline-formula> V, thereby decreasing the SEB sensitivity. However, due to the blocking effect of the CSL on holes, the CSTBTs are at a higher risk of gate oxide breakdown under a negative gate voltage, resulting in their SEB sensitivities being highly dependent on <inline-formula> <tex-math>$V_{mathrm {GE}}$ </tex-math></inline-formula>.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 9","pages":"3653-3661"},"PeriodicalIF":2.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148974756","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":"Transferability of Data-Driven Optimization Results Across Multiple Pixelated CdZnTe Spectrometers","authors":"Thomas D. MacDonald;Hannah S. Parrilla;Jayson R. Vavrek","doi":"10.1109/TNS.2026.3716594","DOIUrl":"https://doi.org/10.1109/TNS.2026.3716594","url":null,"abstract":"Recent work by Vavrek et al. (2026) showed that machine learning methods can be used to exploit spatial patterns of performance variations within the highly-segmented H3D M400 gamma spectrometer to improve an overall spectroscopic performance metric. That work also introduced the <monospace>spectre-ml</monospace> software, which tests various greedy, heuristic, random, and machine learning clustering algorithms to find the best-performing mask for excluding detector regions to improve a user-defined performance metric by training on a given dataset. In this work, we build off of Vavrek et al. (2026) and seek to determine to what extent an optimized binary voxel mask trained on a given dataset can generalize to other datasets. In particular, this article evaluates the transferability of masks trained on one M400 dataset to another M400 detector, in order to determine whether the total effort required in designing masks for different detectors and applications can be substantially reduced by using a single common mask. It also examines testing and training on different subsets of the same dataset to determine the natural level of variability in optimization results. In the inter-detector analysis, as expected, the best-performing model on each detector is often one trained on that dataset, with an average performance enhancement of 16% when considering the relative uncertainty in a Doniach fit to the 186 keV peak. In comparison, the best transferred masks, with the best on-average performance metric across all six detectors, show only a slightly smaller improvement of 13%. These results suggest that high-performing, well-transferable masks can be shared among detectors, reducing or even eliminating the laborious processes of collecting a training dataset and performing the optimization for each detector, ultimately improving safeguards efficiency.","PeriodicalId":13406,"journal":{"name":"IEEE Transactions on Nuclear Science","volume":"73 9","pages":"3758-3766"},"PeriodicalIF":2.4,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148974686","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}