Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10026908
S. Arima, Yu Sasaki, Sho Morie, Yuto Kataoka, Chending Mao, Jia Lin
This study introduced the application of VAR-LiNGAM, and Backpropagation Neural Network with node2vec for feasible data-driven modeling of dynamics of semiconductor production system in which the scale and complexity increase more and more. Open testbed SMT2020 is used evaluations.
{"title":"Data-driven Modeling for Production Dynamics","authors":"S. Arima, Yu Sasaki, Sho Morie, Yuto Kataoka, Chending Mao, Jia Lin","doi":"10.1109/ISSM55802.2022.10026908","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10026908","url":null,"abstract":"This study introduced the application of VAR-LiNGAM, and Backpropagation Neural Network with node2vec for feasible data-driven modeling of dynamics of semiconductor production system in which the scale and complexity increase more and more. Open testbed SMT2020 is used evaluations.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"2 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121267754","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10026914
Sergio Garnica, R. Wieland
This paper reports on the implementation of low cost Internet-of-Things enabled power sockets, deployed on the Fraunhofer institute for Microsystems and Solid State Technologies clean room on one of the inspection microscopes used on the CMOS compatible line. The devices were flashed with open source software to ensure local, secure and reliable control without the necesity of an external cloud provider. The architecture of the physical deployment is shown and experimental data is analysed in order to obtain insight into the usage and statistics behind a previously unknown station in the clean room.
{"title":"Secure and Reliable Power Monitoring for Low Consumption Factory Equipment via Programmable IoT Devices","authors":"Sergio Garnica, R. Wieland","doi":"10.1109/ISSM55802.2022.10026914","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10026914","url":null,"abstract":"This paper reports on the implementation of low cost Internet-of-Things enabled power sockets, deployed on the Fraunhofer institute for Microsystems and Solid State Technologies clean room on one of the inspection microscopes used on the CMOS compatible line. The devices were flashed with open source software to ensure local, secure and reliable control without the necesity of an external cloud provider. The architecture of the physical deployment is shown and experimental data is analysed in order to obtain insight into the usage and statistics behind a previously unknown station in the clean room.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"92 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131718805","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10027071
Wang Xu, Ran Jing Yang, Yang Lai, Yang Fa Shun, M. Kui
As a member of ultra-wide band gap semiconductor materials, $beta-text{Ga}_{2}mathrm{O}_{3}$ materials have attracted wide attention from researchers in the semiconductor field in recent years. Etching process is crucial to realize semiconductor devices and integrated circuits based on $beta-text{Ga}_{2}mathrm{O}_{3}$ materials. Based on the reaction ion etching process commonly used in silicon-based semiconductor technology, the etching experiment research of $beta-text{Ga}_{2}mathrm{O}_{3}$ thin film is carried out. The $beta-text{Ga}_{2}mathrm{O}_{3}text{film}$ is etched with SF6, based on the induction coupled reaction ion etching. The effect of RIE etching power, excitation power and bias power, on etching rate of $beta-text{Ga}_{2}mathrm{O}_{3}$ thin film has been studied. SEM characterization results show that the etching rate is the highest at 600W excitation power. The etching rate increases with the increase of bias power. The etching rate at 200W bias power is slightly higher than that at 150W bias power. However, the photoresist used as the etch mask will be damaged at 200W bias power.
{"title":"Experimentally Study on the Effect of RIE Etching Power on Etching Rate of $beta-text{Ga}_{2}mathrm{O}_{3}$ Thin Film","authors":"Wang Xu, Ran Jing Yang, Yang Lai, Yang Fa Shun, M. Kui","doi":"10.1109/ISSM55802.2022.10027071","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10027071","url":null,"abstract":"As a member of ultra-wide band gap semiconductor materials, $beta-text{Ga}_{2}mathrm{O}_{3}$ materials have attracted wide attention from researchers in the semiconductor field in recent years. Etching process is crucial to realize semiconductor devices and integrated circuits based on $beta-text{Ga}_{2}mathrm{O}_{3}$ materials. Based on the reaction ion etching process commonly used in silicon-based semiconductor technology, the etching experiment research of $beta-text{Ga}_{2}mathrm{O}_{3}$ thin film is carried out. The $beta-text{Ga}_{2}mathrm{O}_{3}text{film}$ is etched with SF6, based on the induction coupled reaction ion etching. The effect of RIE etching power, excitation power and bias power, on etching rate of $beta-text{Ga}_{2}mathrm{O}_{3}$ thin film has been studied. SEM characterization results show that the etching rate is the highest at 600W excitation power. The etching rate increases with the increase of bias power. The etching rate at 200W bias power is slightly higher than that at 150W bias power. However, the photoresist used as the etch mask will be damaged at 200W bias power.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"8 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115449982","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10027157
H. H. Nguyen
Semiconductor manufacturing has data management systems comprising multiple layers, including the cloud layer, the edge layer, and the equipment or device layer, which perform different functions in the system. The equipment layer performs data monitoring and detection of faults-the cloud layer and the edge layer help perform computational tasks. Performance of the computational tasks at the equipment layer is beneficial because they help achieve real-time response to the production and reduce the delays caused by data transfer from the equipment layer to the edge or cloud layer. In semiconductor manufacturing, the host computer located at the edge layer communicates to the equipment through Secs/Gem communication protocol. According to the results from our experiment, it is more efficient and effective to perform data analysis at the equipment level. This paper proposes a new Secs/Gem protocol for performing dynamic AI tasks on the equipment. The protocol allows the host to dynamically assign tasks of analyzing data to the equipment, and the equipment reports the results back to the host.
{"title":"Dynamic AI Computation Tasks with SECS/GEM in Semiconductor Smart Manufacturing","authors":"H. H. Nguyen","doi":"10.1109/ISSM55802.2022.10027157","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10027157","url":null,"abstract":"Semiconductor manufacturing has data management systems comprising multiple layers, including the cloud layer, the edge layer, and the equipment or device layer, which perform different functions in the system. The equipment layer performs data monitoring and detection of faults-the cloud layer and the edge layer help perform computational tasks. Performance of the computational tasks at the equipment layer is beneficial because they help achieve real-time response to the production and reduce the delays caused by data transfer from the equipment layer to the edge or cloud layer. In semiconductor manufacturing, the host computer located at the edge layer communicates to the equipment through Secs/Gem communication protocol. According to the results from our experiment, it is more efficient and effective to perform data analysis at the equipment level. This paper proposes a new Secs/Gem protocol for performing dynamic AI tasks on the equipment. The protocol allows the host to dynamically assign tasks of analyzing data to the equipment, and the equipment reports the results back to the host.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"284 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122401553","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10027076
J. Baderot, Solange Garrais, S. Martínez, J. Foucher, R. Eto, K. Tanida, Takatoshi Yasui, Tomoya Tanaka
Wafer-level direct bonding technology is a key process for the production of backside illuminated (BSI) CMOS image sensor (CIS). Usually, constant-depth mode scanning acoustic microscope (C-SAM) 300mm wafer images are acquired and defect size distribution is provided to monitor defects that degrade bonding yield. Current solutions are not providing information detailed enough to identify the root cause of this degradation. In this paper, we propose a rule-based method for the classification of the defects and automatic segmentation of the defects to extract precise measurements depending on the type of defect. All these information will allow to reduce the time to analyze the images and improve the precision and consistency of the analysis.
{"title":"Automatic Classification of C-SAM Voids for Root Cause Identification of Bonding Yield Degradation","authors":"J. Baderot, Solange Garrais, S. Martínez, J. Foucher, R. Eto, K. Tanida, Takatoshi Yasui, Tomoya Tanaka","doi":"10.1109/ISSM55802.2022.10027076","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10027076","url":null,"abstract":"Wafer-level direct bonding technology is a key process for the production of backside illuminated (BSI) CMOS image sensor (CIS). Usually, constant-depth mode scanning acoustic microscope (C-SAM) 300mm wafer images are acquired and defect size distribution is provided to monitor defects that degrade bonding yield. Current solutions are not providing information detailed enough to identify the root cause of this degradation. In this paper, we propose a rule-based method for the classification of the defects and automatic segmentation of the defects to extract precise measurements depending on the type of defect. All these information will allow to reduce the time to analyze the images and improve the precision and consistency of the analysis.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"289 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132571455","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10027125
Yun-Che Hsieh, Chieh-Yu Chen, Da-Yin Liao, Peter B. Luh, Shi-Chung Chang
Anomaly detection (AD) by exploiting machine learning (ML) of equipment sensory data can make significant contributions to yield improvements. Data cleansing is critical to provide ML-based AD with fixed-length input without distortion of data characteristics. We present a novel data cleansing design. Design innovations are: process step and mode-based input data length determination, importance indicator of sample data based on relative difference, and data cleansing priority by exploiting importance indicator and entropy. Experiment results demonstrate our cleansing design is superior to two frequently used methods in preserving data characteristics for effective AD by using an unsupervised ML approach.
{"title":"Equipment Sensor Data Cleansing Algorithm Design for ML-Based Anomaly Detection","authors":"Yun-Che Hsieh, Chieh-Yu Chen, Da-Yin Liao, Peter B. Luh, Shi-Chung Chang","doi":"10.1109/ISSM55802.2022.10027125","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10027125","url":null,"abstract":"Anomaly detection (AD) by exploiting machine learning (ML) of equipment sensory data can make significant contributions to yield improvements. Data cleansing is critical to provide ML-based AD with fixed-length input without distortion of data characteristics. We present a novel data cleansing design. Design innovations are: process step and mode-based input data length determination, importance indicator of sample data based on relative difference, and data cleansing priority by exploiting importance indicator and entropy. Experiment results demonstrate our cleansing design is superior to two frequently used methods in preserving data characteristics for effective AD by using an unsupervised ML approach.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"210 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132185546","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10026922
H. Tsuchiyama, Holland M. Smith
We have created a line balancing algorithm that uses queueing theory to calculate ideal WIP (Wafer In Process) targets by product and step taking into account the current factory bottlenecks and status, which realize higher equipment utilization, more outs, better WIP bubble/bottleneck management, and reduction of opportunity loss. In this paper, the system architecture and deployment result are described.
{"title":"A Novel Approach to Dynamic Line Balance Control and Scheduling with a Digital Twin Production","authors":"H. Tsuchiyama, Holland M. Smith","doi":"10.1109/ISSM55802.2022.10026922","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10026922","url":null,"abstract":"We have created a line balancing algorithm that uses queueing theory to calculate ideal WIP (Wafer In Process) targets by product and step taking into account the current factory bottlenecks and status, which realize higher equipment utilization, more outs, better WIP bubble/bottleneck management, and reduction of opportunity loss. In this paper, the system architecture and deployment result are described.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"286 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134205415","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-12-12DOI: 10.1109/ISSM55802.2022.10026911
Kohki Murakawa, N. Mayama, T. Mizuno
We experimentally studied the influence of high temperature N2 annealing on the photoluminescence (PL) of SiC and Si quantum-dots (QDs) in SiO2 layer fabricated by hot ion implantation technique. We demonstrated the increase of PL intensity of SiC- and Si-QDs after N2 annealing, compared with that after Ar annealing, which is probably attributable to the reduction of dangling bond density at SiO2/QD interface terminated by N atom trapping.
{"title":"Influence of High Temperature N2 Annealing on Photoluminescence of SiC and Si Quantum Dots in SiO2 Layer","authors":"Kohki Murakawa, N. Mayama, T. Mizuno","doi":"10.1109/ISSM55802.2022.10026911","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10026911","url":null,"abstract":"We experimentally studied the influence of high temperature N2 annealing on the photoluminescence (PL) of SiC and Si quantum-dots (QDs) in SiO2 layer fabricated by hot ion implantation technique. We demonstrated the increase of PL intensity of SiC- and Si-QDs after N2 annealing, compared with that after Ar annealing, which is probably attributable to the reduction of dangling bond density at SiO2/QD interface terminated by N atom trapping.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123704763","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
In the field of semiconductor defect inspection, it has been possible to detect defects with high accuracy thanks to the object detection model (defect detector) composed of the deep learning model. The performance of the deep learning model depends highly on training data; therefore, during the operational phase at the customer site, we need to frequently evaluate the model's performance to deal with shifts of appearance for defects. However, frequently executing general evaluation methods is difficult at the customer site; hence, we need a method to automatically evaluate performance. In this study, for the purpose of automatically evaluating the performance of the defect detector, we propose the Positive/Negative Decision via Outlier Detection (PNDOD). PNDOD decides on positive/negative for detection results based on comparing features corresponding to the detected result with statistics computed from training data. By using this method, we can calculate the estimated precision from the ratio of the estimated number of positive detections to the number of total detections, and we can evaluate the model performance automatically based on this estimated precision. In experiments using SiC wafer images, we confirmed that PNDOD can decide on positive/negative with high accuracy, and we can precisely evaluate the model's performance.
在半导体缺陷检测领域，由深度学习模型组成的物体检测模型(缺陷检测器)使得高精度检测缺陷成为可能。深度学习模型的性能高度依赖于训练数据;因此，在客户站点的操作阶段，我们需要频繁地评估模型的性能，以处理缺陷的外观变化。然而，在客户现场频繁执行一般评估方法是困难的;因此，我们需要一种自动评估性能的方法。在本研究中，为了自动评估缺陷检测器的性能，我们提出了Positive/Negative Decision via Outlier Detection (PNDOD)。PNDOD通过将检测结果对应的特征与从训练数据中计算的统计量进行比较，来决定检测结果的正/负。利用该方法，我们可以通过估计阳性检测次数与总检测次数的比值来计算估计精度，并根据该估计精度自动评价模型的性能。在使用SiC晶圆图像的实验中，我们证实了PNDOD可以高精度地判断正/负，并且我们可以精确地评估模型的性能。
{"title":"Positive/Negative Decision Via Outlier Detection Towards Automatic Performance Evaluation for Defect Detector","authors":"Toshinori Yamauchi, Kentaro Ohira, Takefumi Kakinuma","doi":"10.1109/ISSM55802.2022.10027074","DOIUrl":"https://doi.org/10.1109/ISSM55802.2022.10027074","url":null,"abstract":"In the field of semiconductor defect inspection, it has been possible to detect defects with high accuracy thanks to the object detection model (defect detector) composed of the deep learning model. The performance of the deep learning model depends highly on training data; therefore, during the operational phase at the customer site, we need to frequently evaluate the model's performance to deal with shifts of appearance for defects. However, frequently executing general evaluation methods is difficult at the customer site; hence, we need a method to automatically evaluate performance. In this study, for the purpose of automatically evaluating the performance of the defect detector, we propose the Positive/Negative Decision via Outlier Detection (PNDOD). PNDOD decides on positive/negative for detection results based on comparing features corresponding to the detected result with statistics computed from training data. By using this method, we can calculate the estimated precision from the ratio of the estimated number of positive detections to the number of total detections, and we can evaluate the model performance automatically based on this estimated precision. In experiments using SiC wafer images, we confirmed that PNDOD can decide on positive/negative with high accuracy, and we can precisely evaluate the model's performance.","PeriodicalId":130513,"journal":{"name":"2022 International Symposium on Semiconductor Manufacturing (ISSM)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125831978","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：