A novel methodology based on long short-term memory stacked autoencoders for unsupervised detection of abnormal working conditions in semiconductor manufacturing systems
Restricted accessResearch articleFirst published online October, 2025
A novel methodology based on long short-term memory stacked autoencoders for unsupervised detection of abnormal working conditions in semiconductor manufacturing systems
The detection of the occurrence of working conditions in semiconductor manufacturing systems is fundamental to minimize yield losses and enhance production quality. The main challenges are the lack of data labeled with the machine state (normal/abnormal) and the nonlinear time evolution of the monitored signals. This work develops a novel methodology for the detection of abnormal conditions that, differently from the existing approaches, do not require the availability of labeled data. It consists of: (a) an approach based on k-fold cross-validation for automatically building a training set which does not contain data collected during abnormal conditions; (b) a signal reconstruction model based on stacked autoencoders with Long Short-Term Memory (LSTM) cells for reproducing the machine expected behavior in normal conditions; and (c) a decision module based on an abnormality indicator computed using the Mahalanobis distance. The proposed methodology is shown to outperform other state-of-the-art approaches in two case studies based on data taken from plasma etching machines.
RichardC. (ed.) Semiconductor Manufacturing. In: Understanding semiconductors: a technical guide for non-technical people. Berkeley, CA: Apress, 2022, pp.57–80.
2.
YugmaCBlueJDauzere-PeresS, et al. Integration of scheduling and advanced process control in semiconductor manufacturing: review and outlook. In: 2014 IEEE International Conference on Automation Science and Engineering (CASE), New Taipei, Taiwan, 18–22 August 2014, pp.93–98. New York: IEEE.
3.
NguyenDTDuongQBZamaiE, et al. Fault diagnosis for the complex manufacturing system. Proc IMechE, Part O: J Risk and Reliability2016; 230: 178–194.
4.
KimSHKimCYSeolDH, et al. Machine learning-based process-level fault detection and part-level fault classification in semiconductor etch equipment. IEEE Trans Semicond Manuf2022; 35(2): 174–184.
5.
YehWCChuangMCLeeWC.Uniform parallel machine scheduling with resource consumption constraint. Appl Math Model2015; 39(7): 2131–2138.
6.
LangCISunF-KLawlerB, et al. One class process anomaly detection using kernel density estimation methods. IEEE Trans Semicond Manuf2022; 35: 457–469.
7.
BleakieADjurdjanovicD.Feature extraction, condition monitoring, and fault modeling in semiconductor manufacturing systems. Comput Ind2013; 64(2): 203–213.
8.
KuoWKimT.An overview of manufacturing yield and reliability modeling for semiconductor products. Proc IEEE1999; 87(8): 1329–1342.
9.
ChenSLiuMHouX, et al. Wafer map defect pattern detection method based on improved attention mechanism. Expert Syst Appl2023; 230: 120544.
BatoolUShapiaiMITahirM, et al. A systematic review of deep learning for silicon wafer defect recognition. IEEE Access2021; 9: 116572–116591.
12.
HuangKAgarwalVKThulasiramanK.Diagnosis of clustered faults and wafer testing. IEEE Trans Comput Aided Des Integr Circuits Syst1998; 17(2): 136–148.
13.
ChaJHFinkelsteinM.Burn-in and the performance quality measures in heterogeneous populations. Eur J Oper Res2011; 210(1): 273–280.
14.
ChoiKYiJParkC, et al. Deep learning for anomaly detection in time-series data: review, analysis, and guidelines. IEEE Access2021; 9: 120043–120059.
15.
BaekMKimSB.Failure detection and primary cause identification of multivariate time series data in semiconductor equipment. IEEE Access2023; 11: 54363–54372.
CardinaudCPeignonM-CTessierP-Y.Plasma etching: principles, mechanisms, application to micro- and nano-technologies. Appl Surf Sci2000; 164: 72–83.
18.
HanCKooYKimJ, et al. Wafer type ion energy monitoring sensor for plasma diagnosis. Sensors2023; 23(5): 2410.
19.
ChenSZhangYHouX, et al. Wafer map failure pattern recognition based on deep convolutional neural network. Expert Syst Appl2022; 209: 118254.
20.
KimTBehdinanK.Advances in machine learning and deep learning applications towards wafer map defect recognition and classification: a review. J Intell Manuf2023; 34(8): 3215–3247.
21.
YuHCLinKYChienCF.Hierarchical indices to detect equipment condition changes with high dimensional data for semiconductor manufacturing. J Intell Manuf2014; 25(5): 933–943.
22.
RasayHHadianSMNaderkhaniF, et al. Optimal condition-based maintenance using attribute Bayesian control chart. Proc IMechE, Part O: J Risk and Reliability2024; 238(4): 754–763.
23.
HongSJLimWYCheongT, et al. Fault detection and classification in plasma etch equipment for semiconductor manufacturing e-diagnostics. IEEE Trans Semicond Manuf2012; 25(1): 83–93.
24.
RostamiHBlueJYugmaC.Automatic equipment fault fingerprint extraction for the fault diagnostic on the batch process data. Appl Soft Comput2018; 68: 972–989.
25.
YehW-CChuC-L.Feature selection for data classification in the semiconductor industry by a hybrid of simplified swarm optimization. Electronics2024; 13: 2242.
26.
WiseBMGallagherNBButlerSW, et al. A comparison of principal component analysis, multiway principal component analysis, trilinear decomposition and parallel factor analysis for fault detection in a semiconductor etch process. J Chemometrics1999; 13: 379–396.
27.
YuJ.Fault detection using principal components-based Gaussian mixture model for semiconductor manufacturing processes. IEEE Trans Semicond Manuf2011; 24(3): 432–444.
28.
MahadevanSShahSL.Fault detection and diagnosis in process data using one-class support vector machines. J Process Control2009; 19(10): 1627–1639.
29.
KimDKangPChoS, et al. Machine learning-based novelty detection for faulty wafer detection in semiconductor manufacturing. Expert Syst Appl2012; 39(5): 4075–4083.
30.
PugginiLMcLooneS.An enhanced variable selection and Isolation Forest based methodology for anomaly detection with OES data. Eng Appl Artif Intell2018; 67: 126–135.
31.
KimCLeeJKimR, et al. DeepNAP: deep neural anomaly pre-detection in a semiconductor fab. Inf Sci2018; 457–458: 1–11.
32.
AzamfarMLiXLeeJ, et al. Deep learning-based domain adaptation method for fault diagnosis in semiconductor manufacturing. IEEE Trans Semicond Manuf2020; 33(3): 445–453.
33.
GormanMDingXMaguireL, et al. Anomaly detection in batch manufacturing processes using localized reconstruction errors from 1-D convolutional autoencoders. IEEE Trans Semicond Manuf2023; 36(1): 147–150.
34.
KimYLeeHKimCO.A variational autoencoder for a semiconductor fault detection model robust to process drift due to incomplete maintenance. J Intell Manuf2023; 34(5): 529–540.
35.
KimEChoSLeeB, et al. Fault detection and diagnosis using self-attentive convolutional neural networks for variable-length sensor data in semiconductor manufacturing. IEEE Trans Semicond Manuf2019; 32(3): 302–309.
36.
TchatchouaPGratonGOuladsineM, et al. A comparative evaluation of deep learning anomaly detection techniques on semiconductor multivariate time series data. 2021 IEEE 17th international conference on automation science and engineering (CASE), Lyon, France23–27 August 2021, pp.1613–1620. New York: IEEE.
37.
FanSKSHsuCYTsaiDM, et al. Data-driven approach for fault detection and diagnostic in semiconductor manufacturing. IEEE Trans Autom Sci Eng2020; 17(4): 1925–1936.
38.
ChenC-YChangS-CLiaoD-Y.Equipment anomaly detection for semiconductor manufacturing by exploiting unsupervised learning from sensory data. Sensors2020; 20(19): 5650.
39.
MaggipintoMBeghiASustoGA.A deep convolutional autoencoder-based approach for anomaly detection with industrial, non-images, 2-dimensional data: a semiconductor manufacturing case study. IEEE Trans Autom Sci Eng2022; 19(3): 1477–1490.
40.
FanSKSHsuCYJenCH, et al. Defective wafer detection using a denoising autoencoder for semiconductor manufacturing processes. Adv Eng Inform2020; 46: 101166.
41.
AhmedIHosseinpourFBaraldiP, et al. An artificial intelligence-based framework for burn-in reduction in the semiconductor manufacturing industry. In: van DrielWDPresselKSoyturkM (eds) Recent advances in microelectronics reliability: contributions from the European ECSEL JU project iRel40. Cham: Springer International Publishing, 2024, pp.117–133.
42.
ChenYLSacchiSDeyB, et al. Exploring machine learning for semiconductor process optimization: a systematic review. IEEE Trans Artif Intell2024.
43.
NingSSunJLiuC, et al. Applications of deep learning in big data analytics for aircraft complex system anomaly detection. Proc IMechE, Part O: J Risk and Reliability2021; 235(5): 923–940.
44.
HsuCYLiuWC.Multiple time-series convolutional neural network for fault detection and diagnosis and empirical study in semiconductor manufacturing. J Intell Manuf2021; 32(3): 823–836.
45.
WangTQiaoMZhangM, et al. Data-driven prognostic method based on self-supervised learning approaches for fault detection. J Intell Manuf2020; 31(6): 1611–1619.
46.
BaraldiPDi MaioFTuratiP, et al. Robust signal reconstruction for condition monitoring of industrial components via a modified Auto Associative Kernel Regression method. Mech Syst Signal Process2015; 60–61: 29–44.
47.
YuJYooJJangJ, et al. A novel hybrid of auto-associative kernel regression and dynamic independent component analysis for fault detection in nonlinear multimode processes. J Process Control2018; 68: 129–144.
48.
LicenSDi GilioAPalmisaniJ, et al. Pattern recognition and anomaly detection by self-organizing maps in a multi-month E-nose survey at an industrial site. Sensors2020; 20(7): 1887.
49.
ZioE.A support vector machine integrated system for the classification of operation anomalies in nuclear components and systems. Reliab Eng Syst Saf2007; 92(5): 593–600.
50.
Arcos JiménezAGómez MuñozCQGarcía MárquezFP, et al. Dirt and mud detection and diagnosis on a wind turbine blade employing guided waves and supervised learning classifiers. Reliab Eng Syst Saf2019; 184: 2–12.
51.
KontopoulouVIPanagopoulosADKakkosI, et al. A review of ARIMA vs. Machine learning approaches for time series forecasting in data driven networks. Future Internet2023; 15(8): 255.
52.
SalehinejadHSankarSBarfettJ, et al. Recent advances in recurrent neural networks. arXiv preprint arXiv:1801.01078v3, 2018.
53.
BruneoDDe VitaF.Detecting faults at the edge via sensor data fusion echo state networks. Sensors2022; 22(8): 2858.
54.
MaassWNatschlägerTMarkramH.Real-time computing without stable states: A new framework for neural computation based on perturbations. Neural Comput2002; 14(11): 2531–2560.
55.
SteilJJ.Backpropagation-decorrelation: online recurrent learning with O(N) complexity. Proc IEEE Int Joint Conf Neural Netw2004; 3: 843–848.
56.
ZhangYPengPLiuC, et al. A sequential resampling approach for imbalanced batch process fault detection in semiconductor manufacturing. J Intell Manuf2022; 33: 1057–1072.
57.
Van HoudtGMosqueraCNápolesG. A review on the long short-term memory model. Artif Intell Rev2020; 53: 5929–5955.
58.
OmenzetterPBrownjohnJMWMoyoP.Identification of unusual events in multi-channel bridge monitoring data. Mech Syst Signal Process2004; 18: 409–430.
59.
NguyenHDTranKPThomasseyS, et al. Forecasting and anomaly detection approaches using LSTM and LSTM autoencoder techniques with applications in supply chain management. Int J Inf Manage2021; 57: 102282.
60.
YangZBaraldiPZioE, et al. A method for fault detection in multi-component systems based on sparse autoencoder-based deep neural networks. Reliab Eng Syst Saf2022; 220: 108278.
61.
AdemKKılıçarslanS.Performance analysis of optimization algorithms on stacked autoencoder. IEEE Int Conf Big Data2019; 1234–1243.
62.
KingmaDPBaJ.Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980, 2015.
63.
MalekiSMalekiSJenningsNR.Unsupervised anomaly detection with LSTM autoencoders using statistical data-filtering. Appl Soft Comput2021; 108: 107443.
64.
YangZXuBLuoW, et al. Autoencoder-based representation learning and its application in intelligent fault diagnosis: a review. Measurement2022; 189: 110460.
65.
González-MuñizADíazICuadradoAA, et al. Health indicator for machine condition monitoring built in the latent space of a deep autoencoder. Reliab Eng Syst Saf2022; 224: 108482.
66.
ZhangCHuDYangT.Anomaly detection and diagnosis for wind turbines using long short-term memory-based stacked denoising autoencoders and XGBoost. Reliab Eng Syst Saf2022; 222: 108445.
67.
JiangGXiePHeH, et al. Wind turbine fault detection using a denoising autoencoder with temporal information. IEEE/ASME Trans Mechatron2018; 23(1): 89–100.
68.
YoudenWJ.Index for rating diagnostic tests. Cancer1950; 3(1): 32–35.
69.
PowersDMW. Evaluation: from precision, recall, and F-measure to ROC, informedness, markedness & correlation. J Mach Learn Technol2011; 2: 37–63.
70.
YinCWangYHeY, et al. Early fault diagnosis of ball screws based on 1-D convolution neural network and orthogonal design. Proc IMechE, Part O: J Risk and Reliability2021; 235(5): 783–797.
71.
Dabo-NiangSFerratyFHyndmanR, et al. Bagplots, boxplots and outlier detection for functional data. In: FerratyFDabo-NiangS (eds) Functional and operatorial statistics. Lyon, France: Springer, 2008, pp.201–207.
72.
HosseinpourFAhmedIBaraldiP, et al. An unsupervised method for anomaly detection in multi-stage production systems based on LSTM autoencoders. In: Proceedings of the 32nd European safety and reliability conference (ESREL 2022). Singapore: Research Publishing, 2022, pp.1346–1352.
73.
XiongMWangHCheC, et al. Toward safer aviation: Application of GA-XGBoost-SHAP for incident cognition and model explainability. Proc IMechE, Part O: J Risk and Reliability2023; 238(6): 1195–1208.
74.
HalbouniAGunawanTSHabaebiMH, et al. CNN-LSTM: hybrid deep neural network for network intrusion detection system. IEEE Access2022; 10: 99837–99849.
75.
YuNXuQWangH.Wafer defect pattern recognition and analysis based on convolutional neural network. IEEE Trans Semicond Manuf2019; 32(4): 566–573.
