• 한국정보시스템학회지:정보시스템연구
• /
• 제19권1호
• /
• pp.19-33
• /
• 2010

생산 수율은 비선형관계를 지닌 여러 요인들에 의해 영향을 받기 때문에 반도체 생산의 경우 예측이 어렵다. 본 논문에서 저자들은 사례기반추론과 자기조직화신경망 기반의 데이터마이닝 기법을 활용하여 수율의 높고 낮음을 밝히는 지능화된 수율예측시스템을 제시한다. 이 시스템은 자기조직회신경망을 사용하여 생산 로트의 공정파라미터 패턴을 파악하고 속성가중치 기반의 사례기반추론을 통해 신규 로트의 수율 수준을 예측한다. 이때 속성가중치는 역전파인공신경망을 통해 계산된다. 웹기반 시스템이 개발되고, 반도체 생산 기업의 실제 자료를 적용하여 본 시스템의 효율을 검증하고 평가한다.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2009년도 제38회 동계학술대회 초록집
• /
• pp.413-413
• /
• 2010

In semiconductor manufacturing field, all equipments have various sensors to diagnosis the situations of processes. For increasing the accuracy of diagnosis, hundreds of sensors are emplyed. As sensors provide millions of data, the process diagnosis from them are unrealistic. Besides, in some cases, the results from some data which have same conditions are different. We want to find some information, such as data and knowledge, from the data. Nowadays, fault detection and classification (FDC) has been concerned to increasing the yield. Certain faults and no-faults can be classified by various FDC tools. The uncertainty in semiconductor manufacturing, no-faulty in faulty and faulty in no-faulty, has been caused the productivity to decreased. From the uncertainty, the rough set theory is a viable approach for extraction of meaningful knowledge and making predictions. Reduction of data sets, finding hidden data patterns, and generation of decision rules contrasts other approaches such as regression analysis and neural networks. In this research, a RGB sensor was used for diagnosis plasma instead of optical emission spectroscopy (OES). RGB data has just three variables (red, green and blue), while OES data has thousands of variables. RGB data, however, is difficult to analyze by human's eyes. Same outputs in a variable show different outcomes. In other words, RGB data includes the uncertainty. In this research, by rough set theory, decision rules were generated. In decision rules, we could find the hidden data patterns from the uncertainty. RGB sensor can diagnosis the change of plasma condition as over 90% accuracy by the rough set theory. Although we only present a preliminary research result, in this paper, we will continuously develop uncertainty problem solving data mining algorithm for the application of semiconductor process diagnosis.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2012년도 제42회 동계 정기 학술대회 초록집
• /
• pp.241-241
• /
• 2012

Semiconductor industry has been taking the advantage of improvements in process technology in order to maintain reduced device geometries and stringent performance specifications. This results in semiconductor manufacturing processes became hundreds in sequence, it is continuously expected to be increased. This may in turn reduce the yield. With a large amount of investment at stake, this motivates tighter process control and fault diagnosis. The continuous improvement in semiconductor industry demands advancements in process control and monitoring to the same degree. Any fault in the process must be detected and classified with a high degree of precision, and it is desired to be diagnosed if possible. The detected abnormality in the system is then classified to locate the source of the variation. The performance of a fault detection system is directly reflected in the yield. Therefore a highly capable fault detection system is always desirable. In this research, time series modeling of the data from an etch equipment has been investigated for the ultimate purpose of fault diagnosis. The tool data consisted of number of different parameters each being recorded at fixed time points. As the data had been collected for a number of runs, it was not synchronized due to variable delays and offsets in data acquisition system and networks. The data was then synchronized using a variant of Dynamic Time Warping (DTW) algorithm. The AutoRegressive Integrated Moving Average (ARIMA) model was then applied on the synchronized data. The ARIMA model combines both the Autoregressive model and the Moving Average model to relate the present value of the time series to its past values. As the new values of parameters are received from the equipment, the model uses them and the previous ones to provide predictions of one step ahead for each parameter. The statistical comparison of these predictions with the actual values, gives us the each parameter's probability of fault, at each time point and (once a run gets finished) for each run. This work will be extended by applying a suitable probability generating function and combining the probabilities of different parameters using Dempster-Shafer Theory (DST). DST provides a way to combine evidence that is available from different sources and gives a joint degree of belief in a hypothesis. This will give us a combined belief of fault in the process with a high precision.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2015년도 제49회 하계 정기학술대회 초록집
• /
• pp.142.2-142.2
• /
• 2015

Plasma etch endpoint detection (EPD) of SiO2 and PR layer is demonstrated by plasma impedance monitoring in this work. Plasma etching process is the core process for making fine pattern devices in semiconductor fabrication, and the etching endpoint detection is one of the essential FDC (Fault Detection and Classification) for yield management and mass production. In general, Optical emission spectrocopy (OES) has been used to detect endpoint because OES can be a simple, non-invasive and real-time plasma monitoring tool. In OES, the trend of a few sensitive wavelengths is traced. However, in case of small-open area etch endpoint detection (ex. contact etch), it is at the boundary of the detection limit because of weak signal intensities of reaction reactants and products. Furthemore, the various materials covering the wafer such as photoresist (PR), dielectric materials, and metals make the analysis of OES signals complicated. In this study, full spectra of optical emission signals were collected and the data were analyzed by a data-mining approach, modified K-means cluster analysis. The K-means cluster analysis is modified suitably to analyze a thousand of wavelength variables from OES. This technique can improve the sensitivity of EPD for small area oxide layer etching processes: about 1.0 % oxide area. This technique is expected to be applied to various plasma monitoring applications including fault detections as well as EPD.

• 산업공학
• /
• 제20권3호
• /
• pp.288-297
• /
• 2007

Success of semiconductor/LCD industry depends on its yield and quality of product. For the purpose, FDC (Fault Detection and Classification) system is used to diagnose fault state in main manufacturing processes by monitoring time series data collected by equipment sensors which represent various conditions of the equipment. The data set is segmented at the start and end of each product lot processing by a trigger event module. However, in practice, segmented sensor data usually have the features of data asynchronization such as different start points, end points, and data lengths. Due to the asynchronization problem, false alarm (type I error) and missed alarm (type II error) occur frequently. In this paper, we propose a robust process fault detection system by integrating a process event detection method and a similarity measuring method based on dynamic time warping algorithm. An experiment shows that the proposed system is able to recognize abnormal condition correctly under the asynchronous data situation.

• 품질경영학회지
• /
• 제47권2호
• /
• pp.381-395
• /
• 2019

Purpose: This study aims to conduct a systematic literature review to suitably identify wide and specific issues and topics on service quality in supply chain. Methods: This study is to investigate service quality in supply chain research using a systematic literature review methodology. In order to extract influential journals and papers, we used the SJR impact factor provided by the SCOPUS database. The collected 169 papers were analyzed using bibliometric analysis, citation analysis as well as keywords network. Results: We conducted a bibliometric analysis to identify top authors contributing to service quality in supply chain and their issues, and further examined important keywords and new emerging keywords. In addition, we extracted five influential papers by PageRank to clarify critical issues and divided into five clusters to identify topics of service quality in supply chain by using network-based approach. In order to examine comprehensive issues and topics of service quality in supply chain, we constructed a keyword network to observe difference in the classification of important keywords across network centrality measures. Conclusion: Our study reviewed literature on service quality in supply chain and explored the future directions and trends of service quality in supply chain.