• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 1998년도 제13차 학술회의논문집
• /
• pp.195-200
• /
• 1998

Since the neural network was introduced, significant progress has been made on data handling and learning algorithms. Currently, the most popular learning algorithm in neural network training is feed forward error back-propagation (FFEBP) algorithm. Aside from the success of the FFEBP algorithm, polynomial neural networks (PNN) learning has been proposed as a new learning method. The PNN learning is a self-organizing process designed to determine an appropriate set of Ivakhnenko polynomials that allow the activation of many neurons to achieve a desired state of activation that mimics a given set of sampled patterns. These neurons are interconnected in such a way that the knowledge is stored in Ivakhnenko coefficients. In this paper, the PNN model has been developed using the plasma enhanced chemical vapor deposition (PECVD) experimental data. To characterize the PECVD process using PNN, SiO$_2$films deposited under varying conditions were analyzed using fractional factorial experimental design with three center points. Parameters varied in these experiments included substrate temperature, pressure, RF power, silane flow rate and nitrous oxide flow rate. Approximately five microns of SiO$_2$were deposited on (100) silicon wafers in a Plasma-Therm 700 series PECVD system at 13.56 MHz.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2005년도 제36회 하계학술대회 논문집 C
• /
• pp.2163-2165
• /
• 2005

A pulsed laser ablation deposition (PLAD) technique has been used for producing fine particle as well as thin film at relatively low substrate temperatures. However, in order to manufacture and evaluate such materials in detail, motions of plume particles generated by laser ablation have to be understood and interactions between the particles by ablation and gas plasma have to be clarified. Therefore, this paper was focused on the understanding of plume motion in laser ablation assisted by Ar plasma at 50(mTorr). Two-dimensional hybrid model consisting of fluid and particle models was developed and three kinds of plume particles which are carbon atom (C), ion $(C^+)$ and electron were considered in the calculation of particle method It was obtained that ablated $C^+$ was electrically captured in Ar plasmas by strong electric field (E). The difference between motions of the ablated electrons and $C^+$ made E strong and the collisional processes active.

• 한국표면공학회지
• /
• 제43권3호
• /
• pp.159-164
• /
• 2010

Surface roughness of deposited or etched film strongly depends on ion bombardment. Relationships between ion bombardment variables and surface roughness are too complicated to model analytically. To overcome this, an empirical neural network model was constructed and applied to a deposition process of silicon nitride (SiN) films. The films were deposited by using a pulsed plasma enhanced chemical vapor deposition system in $SiH_4$-$NH_4$ plasma. Radio frequency source power and duty ratio were varied in the range of 200-800 W and 40-100%. A total of 20 experiments were conducted. A non-invasive ion energy analyzer was used to collect ion energy distribution. The diagnostic variables examined include high (or) low ion energy and high (or low) ion energy flux. Mean surface roughness was measured by using atomic force microscopy. A neural network model relating the diagnostic variables to the surface roughness was constructed and its prediction performance was optimized by using a genetic algorithm. The optimized model yielded an improved performance of about 58% over statistical regression model. The model revealed very interesting features useful for optimization of surface roughness. This includes a reduction in surface roughness either by an increase in ion energy flux at lower ion energy or by an increase in higher ion energy at lower ion energy flux.

• 반도체디스플레이기술학회지
• /
• 제23권2호
• /
• pp.129-134
• /
• 2024

This study aims to talk about the necessity of solving the PFC gas emission problem raised by the recent development of the semiconductor industry and the remote plasma source method monitoring system used in the semiconductor industry. The 'monitoring system' means that the researchers applied machine learning to the existing monitoring technology and modeled it. In the process of this study, Residual Gas Analyzer monitoring technology and linear regression model were used. Through this model, the researchers identified emissions of at least 12700mg CO2 to 75800mg CO2 with values ranging from ion current 0.6A to 1.7A, and expect that the 'monitoring system' will contribute to the effective calculation of greenhouse gas emissions in the semiconductor industry in the future.

• 한국근적외분광분석학회:학술대회논문집
• /
• 한국근적외분광분석학회 2001년도 NIR-2001
• /
• pp.1244-1244
• /
• 2001

In linear discriminant analysis there are two important properties concerning the effectiveness of discriminant function modeling. The first is the separability of the discriminant function for different classes. The separability reaches its optimum by maximizing the ratio of between-class to within-class variance. The second is the stability of the discriminant function against noises present in the measurement variables. One can optimize the stability by exploring the discriminant variates in a principal variation subspace, i. e., the directions that account for a majority of the total variation of the data. An unstable discriminant function will exhibit inflated variance in the prediction of future unclassified objects, exposed to a significantly increased risk of erroneous prediction. Therefore, an ideal discriminant function should not only separate different classes with a minimum misclassification rate for the training set, but also possess a good stability such that the prediction variance for unclassified objects can be as small as possible. In other words, an optimal classifier should find a balance between the separability and the stability. This is of special significance for multivariate spectroscopy-based classification where multicollinearity always leads to discriminant directions located in low-spread subspaces. A new regularized discriminant analysis technique, the principal discriminant variate (PDV) method, has been developed for handling effectively multicollinear data commonly encountered in multivariate spectroscopy-based classification. The motivation behind this method is to seek a sequence of discriminant directions that not only optimize the separability between different classes, but also account for a maximized variation present in the data. Three different formulations for the PDV methods are suggested, and an effective computing procedure is proposed for a PDV method. Near-infrared (NIR) spectra of blood plasma samples from mastitic and healthy cows have been used to evaluate the behavior of the PDV method in comparison with principal component analysis (PCA), discriminant partial least squares (DPLS), soft independent modeling of class analogies (SIMCA) and Fisher linear discriminant analysis (FLDA). Results obtained demonstrate that the PDV method exhibits improved stability in prediction without significant loss of separability. The NIR spectra of blood plasma samples from mastitic and healthy cows are clearly discriminated between by the PDV method. Moreover, the proposed method provides superior performance to PCA, DPLS, SIMCA and FLDA, indicating that PDV is a promising tool in discriminant analysis of spectra-characterized samples with only small compositional difference, thereby providing a useful means for spectroscopy-based clinic applications.

• 한국근적외분광분석학회:학술대회논문집
• /
• 한국근적외분광분석학회 2001년도 NIR-2001
• /
• pp.1042-1042
• /
• 2001

In linear discriminant analysis there are two important properties concerning the effectiveness of discriminant function modeling. The first is the separability of the discriminant function for different classes. The separability reaches its optimum by maximizing the ratio of between-class to within-class variance. The second is the stability of the discriminant function against noises present in the measurement variables. One can optimize the stability by exploring the discriminant variates in a principal variation subspace, i. e., the directions that account for a majority of the total variation of the data. An unstable discriminant function will exhibit inflated variance in the prediction of future unclassified objects, exposed to a significantly increased risk of erroneous prediction. Therefore, an ideal discriminant function should not only separate different classes with a minimum misclassification rate for the training set, but also possess a good stability such that the prediction variance for unclassified objects can be as small as possible. In other words, an optimal classifier should find a balance between the separability and the stability. This is of special significance for multivariate spectroscopy-based classification where multicollinearity always leads to discriminant directions located in low-spread subspaces. A new regularized discriminant analysis technique, the principal discriminant variate (PDV) method, has been developed for handling effectively multicollinear data commonly encountered in multivariate spectroscopy-based classification. The motivation behind this method is to seek a sequence of discriminant directions that not only optimize the separability between different classes, but also account for a maximized variation present in the data. Three different formulations for the PDV methods are suggested, and an effective computing procedure is proposed for a PDV method. Near-infrared (NIR) spectra of blood plasma samples from daily monitoring of two Japanese cows have been used to evaluate the behavior of the PDV method in comparison with principal component analysis (PCA), discriminant partial least squares (DPLS), soft independent modeling of class analogies (SIMCA) and Fisher linear discriminant analysis (FLDA). Results obtained demonstrate that the PDV method exhibits improved stability in prediction without significant loss of separability. The NIR spectra of blood plasma samples from two cows are clearly discriminated between by the PDV method. Moreover, the proposed method provides superior performance to PCA, DPLS, SIMCA md FLDA, indicating that PDV is a promising tool in discriminant analysis of spectra-characterized samples with only small compositional difference.

• 조명전기설비학회논문지
• /
• 제19권4호
• /
• pp.1-8
• /
• 2005

본 연구에서는 2차원적 유체 모델을 통하여 보다 실질적인 플라즈마를 이해하고자 하였으며, 기하학적인 방전전극 구조를 반영하도록 전극단에서 챔버 외벽의 거리를 변화시키면서 플라즈마의 특성을 정량적으로 비교 분석하고자 하였다. 방전 챔버의 구조로서, 전극의 반경과 방전 챔버의 높이는 일정하게 유지하면서 방전 챔버의 넓이를 변화시킴에 따라 형성되는 플라즈마의 특성을 분석하였다. 그 결과, 전극단과 챔버 외벽의 거리가 짧을수록 그 영역에서 전계가 강하게 형성되어, 외벽을 향하는 각 입자들의 움직임도 매우 활발하다는 것을 알 수 있었다. 또한, 전각단과 외벽과의 거리가 짧을수록 전극 면상에서 형성되는 입자들의 수밀도와 유속의 변화가 일정하게 형성되는 것을 알 수 있었다. 이러한 결과는 웨이퍼의 대구경화에 따른 플라즈마의 균일성을 고려할 경우에 매우 효과적일 것으로 고려되어 진다.

• 한국진공학회:학술대회논문집
• /
• 한국진공학회 1999년도 제17회 학술발표회 논문개요집
• /
• pp.216-216
• /
• 1999

The usage of a stationary plasma thruster (SPT) ion source, invented previously for space application in Russia, in experiments with surface modifications and film deposition systems is reported here. Plasma in the SPT is formed and accelerated in electric discharge taking place in the crossed axial electric and radial magnetic fields. Brief description of the construction of specific model of SPT used in the experiments is presented. With gas flow rate 39ml/min, ion current distributions at several distances from the source are obtained. These was equal to 1~3 mA/$\textrm{cm}^2$ within an ion beam ejection angle of $\pm$20$^{\circ}$with discharge voltage 160V for Ar as a working gas. Such an extremely high ion current density allows us to obtain the Ti metal films with deposition rate of $\AA$/sec by sputtering of Ti target. It is shown a possibility of using of reactive gases in SPT (O2 and N2) along with high purity inert gases used for cathode to prevent the latter contamination. It is shown the SPT can be operated at the discharge and accelerating boltages up to 600V. The results of presented experiments show high promises of the SPT in sputtering and surface modification systems for deposition of oxide thin films on Si or polymer substrates for semiconductor devices, optical coatings and metal corrosion barrier layers. Also, we have been tried to establish in application of the modeling expertise gained in electric and ionic propulsion to permit numerical simulation of additional processing systems. In this mechanism, it will be compared with conventional DC sputtering for film microstructure, chemical composition and crystallographic considerations.

• 한국광학회지
• /
• 제31권3호
• /
• pp.125-133
• /
• 2020

본 논문에서는 다변량 분석법과 결합된 레이저 유도 플라즈마 분광법을 사용하여 겹친 유류 지문을 분리하는 혁신적인 방법을 연구하였다. LIPS는 겹친 유류 지문의 화학 성분에 대한 데이터뿐 아니라 실시간 분석 및 고속 스캐닝이 가능한 분광법이다. 레이저 유도 플라즈마 분광법을 통해 도출된 스펙트럼은 적절한 다변량 분석이 적용되어 법의학적 분류와 겹친 유류 지문의 재구성에 유용한 화학적 성분을 제공한다. 본 연구에서는 LIPS 스펙트럼에서 4가지의 유류 지문을 분류하기 위하여, 주성분 분석 방식과 부분 최소 제곱 회귀 분석을 사용하였다. 제안된 방법은 SIMCA 및 PLS-DA와 같은 구별 방식을 사용하여 4개의 유류 지문의 분류를 성공적으로 입증하였다. 본 연구의 결과는 대략 85% 이상의 정확도를 가졌으며, external validation 실험에서도 분류의 가능함을 보였다. 최종적으로, 125 ㎛의 공간 간격으로 레이저 스캐닝 분석을 통한 겹친 유류 지문의 2차원 형태의 분리가 가능함을 입증하였다.

• KIEE International Transactions on Electrophysics and Applications
• /
• 제4C권5호
• /
• pp.220-229
• /
• 2004

We have examined a technique of one-dimensional (1D) fluid modeling for radio-frequency Ar capacitively coupled plasmas (CCP) between unequal-sized powered and grounded electrodes. In order to simulate a practical CCP reactor configuration with a grounded side wall by the 1D model, it has been assumed that the discharge space has a conic frustum shape; the grounded electrode is larger than the powered one and the discharge space expands with the distance from the powered electrode. In this paper, we focus on how much a 1D model can approximate a 2D model and evaluate their comparisons. The plasma density calculated by the 1D model has been compared with that by a two-dimensional (2D) fluid model, and a qualitative agreement between them has been obtained. In addition, 1D and 2D calculation results for another reactor configuration with equal-sized electrodes have also been presented together for comparison. In the discussion, four CCP models, which are 1D and 2D models with symmetric and asymmetric geometries, are compared with each other and the DC self-bias voltage has been focused on as a characteristic property that reflects the unequal electrode surface areas. Reactor configuration and experimental parameters, which the self-bias depends on, have been investigated to develop the ID modeling for reactor geometry with unequal-sized electrodes.