• 한국반도체및디스플레이장비학회:학술대회논문집
• /
• 한국반도체및디스플레이장비학회 2004년도 춘계학술대회 발표 논문집
• /
• pp.154-158
• /
• 2004

Power inductor is usually used in the field of the power circuit of a cellular phone, TFT LCD module etc.. This paper presents the development process of Power Inductor Full Automation Winding/Welding System for TFT LCD. This process, the process algorithm, high precision welding current control, design of welding head, high speed, high precision feeding mechanism, and user interface process control program technologies are included.

• 산업공학
• /
• 제18권1호
• /
• pp.22-34
• /
• 2005

This paper introduces a simulation model which is based on the process analysis of a semiconductor company. The objective of the simulation modelis not only to estimate the overall performancesof the company but also to evaluate the performances of various operation rules for shop floor control. First, in order to develop the simulation model, a time study is performed for each process after analyzing the processes for the company. Second, by using ARENA, a simulation model is constructed based on the process analysis and the time study. After the simulation model is tested and run, its results are discussed.

• 반도체디스플레이기술학회지
• /
• 제19권4호
• /
• pp.121-125
• /
• 2020

With miniaturization of semiconductor, the manufacturing process become more complex, and undetected small changes in the state of the equipment have unexpectedly changed the process results. Fault detection classification (FDC) system that conducts more active data analysis is feasible to achieve more precise manufacturing process control with advanced machine learning method. However, applying machine learning, especially in supervised learning criteria, requires an arduous data labeling process for the construction of machine learning data. In this paper, we propose a semi-supervised learning to minimize the data labeling work for the data preprocessing. We employed equipment status variable identification (SVID) data and optical emission spectroscopy data (OES) in silicon etch with SF6/O2/Ar gas mixture, and the result shows as high as 95.2% of labeling accuracy with the suggested semi-supervised learning algorithm.

• 반도체디스플레이기술학회지
• /
• 제22권4호
• /
• pp.131-135
• /
• 2023

Neural network-based time series models called time series neural networks (TSNNs) are trained by the error backpropagation algorithm and used to predict process shifts of parameters such as gas flow, RF power, and chamber pressure in reactive ion etching (RIE). The training data consists of process conditions, as well as principal components (PCs) of optical emission spectroscopy (OES) data collected in-situ. Data are generated during the etching of benzocyclobutene (BCB) in a SF₆/O₂ plasma. Combinations of baseline and faulty responses for each process parameter are simulated, and a moving average of TSNN predictions successfully identifies process shifts in the recipe parameters for various degrees of faults.

• 국제지역연구
• /
• 제20권2호
• /
• pp.85-111
• /
• 2016

최근, 글로벌 반도체기업들의 중국내 직접투자가 지속적으로 증가하고 있다. 특히, 중국내에서 제조를 위한 공장설립이 늘고 있으며, 이는 저렴한 인건비를 활용하기 위한 차원보다는 늘어나는 중국내 수요에 대응하려 투자하고 있다. 이에 따라 본 연구에서는 한국과 대만 글로벌 반도체기업들에 대한 중국내 직접투자 배경과 과정에 대해서 살펴보았다. 조사결과 삼성반도체는 1996년 Suzhou에 후 공정 공장을 설립하였으나 전 공정의 일괄생산방식으로 투자한 것은 2014년 Xian공장이었다. 그리고 SK하이닉스는 2006년 Wuxi에 처음 공장을 설립하였으며 이후 2009년 중국기업과 합작으로 후 공정 공장인 Hitech Semiconductor를 설립하였고 이어 2015년 Chongqing에 후 공정 공장을 설립하였다. 아울러 TSMC는 2004년 Shanghai에서 처음 공장을 가동하였으며 나아가 2018년에 Nanjing공장이 완공예정이다. 마지막으로 UMC는 2000년대 초반 중국 현지기업인 HJT에 지분을 소유하는 방식으로 중국에 진출하였으며, 이후 직접투자로 빠르면 2016년 말 Xiamen공장이 완공된다. 결과적으로 각사 대부분은 주로 중국시장을 공략하기 위한 목적으로 중국에 공장설립을 한 것으로 나타났으나 이외에도 삼성반도체는 리스크 관리 그리고 SK하이닉스는 상계관세를 회피하기 위한 목적이 있었다. 본 연구결과를 토대로 다른 반도체기업들에 차후 중국내 직접투자에 있어 도움이 되는 전략적 시사점들을 제시하였다.

• 대한산업공학회지
• /
• 제38권1호
• /
• pp.46-56
• /
• 2012

The purpose of virtual metrology (VM) in semiconductor manufacturing is to predict every wafer's metrological values based on its process equipment data without an actual metrology. In this paper, we propose novelty detection-based reliability estimation models for VM in order to support flexible utilization of VM results. Because the proposed model can not only estimate the reliability of VM, but also identify suspicious process variables lowering the reliability, quality control actions can be taken selectively based on the reliance level and its causes. Based on the preliminary experimental results with actual semiconductor manufacturing process data, our models can successfully give a high reliance level to the wafers with small prediction errors and a low reliance level to the wafers with large prediction errors. In addition, our proposed model can give more detailed information by identifying the critical process variables and their relative impacts on the low reliability.

• 반도체디스플레이기술학회지
• /
• 제12권4호
• /
• pp.39-44
• /
• 2013

In the semiconductor heat-treatment process, the temperature uniformity determines the film quality of a wafer. This film quality effects on the overall yield rate. The heat transfer of the wafer surface in the heat-treatment process equipment is occurred by convection and radiation complexly. Because of this, there is the nonlinearity between the wafer temperature and reactor. Therefore, the accurate prediction of temperature on the wafer surface is difficult without the direct measurement. The thermal camera and the T/C wafer are general ways to confirm the temperature uniformity on the heat-treatment process. As above ways have limit to measure the temperature in the precise domain under the micro-scale. In this study, we developed the thin film type temperature sensor using the MEMS technology to establish the system which can measure the temperature under the micro-scale. We combined the experiment and numerical analysis to verify and calibrate the system. Finally, we measured the temperature on the wafer surface on the semiconductor process using the developed system, and confirmed the temperature variation by comparison with the commercial T/C wafer.

• 한국기계가공학회지
• /
• 제4권3호
• /
• pp.32-38
• /
• 2005

Electrical discharge machining is the method of using thermal energy by electrical discharge. Generally, if the material of workpiece has conductivity even though very hard materials and complicated shape which are difficult to cut such as quenching steel, cemented carbide, diamond and conductive ceramics, the EDM process is favorable one of possible machining processes. But, the process is necessarily required of finish cut and heat treatment because of slow cutting speed, no mirror surface, brittleness and crack due to the residual stress for manufactured goods. In this experimental thesis, the super EDM drilling was developed for high precision semiconductor die steel and for minimization of leadframe width. It was possible to development of EDM drilling machine for high precision semiconductor die with the electrode guide and its modelling and stress analysis. The development of electrode with the copper pipe type was conducted to drill the hole from the diameter of 0.1mm to 3.0mm with the error of from 0.02mm to 0.12mm. From the SEM and EDX analysis, the entrance of the EDM drill was found the resolidification of not only the component of tungsten but also the component of copper.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 2012년도 제42회 동계 정기 학술대회 초록집
• /
• pp.171-171
• /
• 2012

Neutral beam assisted chemical vapor deposition (NBa-CVD) process has been developed as a nove,l room temperature deposition process for the light-soaking free nano-crystalline silicon (nc-Si) thin films including intrinsic and n-type doped thin film. During formation of nc-Si thin films by the NBa-CVD process with silicon reflector at room temperature, the energetic particles enhance doping efficiency and crystalline phase in nc-Si thin films without additional heating at substrate. The effects of incident NB energy controlled by the reflector bias have been confirmed by Raman spectra analysis. Additionally, TEM images show uniform nc-Si grains which imbedded amorphous phase without incubation layer. The nc-Si films by the NBa-CVD are hardly degenerated by light soaking; the degradations of photoconductivity were just a few percents before and after light irradiation.

• 반도체디스플레이기술학회지
• /
• 제15권2호
• /
• pp.32-37
• /
• 2016

Sidewall angle (SWA) of an absorber stack in extreme ultraviolet lithography mask is considered to be $90^{\circ}$ ideally, however, it is difficult to obtain $90^{\circ}$ SWA because absorber profile is changed by complicated etching process. As the imaging performance of the mask can be varied with this SWA of the absorber stack, more complicated optical proximity correction is required to compensate for the variation of imaging performance. In this study, phase shift mask (PSM) is suggested to reduce the variation of imaging performance due to SWA change by modifying mask material and structure. Variations of imaging performance and lithography process margin depending on SWA were evaluated through aerial image and developed resist simulations to confirm the advantages of PSM over the binary intensity mask (BIM). The results show that the variations of normalized image log slope and critical dimension bias depending on SWA are reduced with PSM compared to BIM. Process margin for exposure dose and focus was also improved with PSM.