• Journal of the Semiconductor & Display Technology
• /
• v.18 no.1
• /
• pp.1-4
• /
• 2019

Reactive sputtering is widely used because of its high deposition rate and high step coverage. The deposition layer is often affected by target poisoning because the target conditions are changed, as well, by reactive gases during the initial stage of sputtering process. The reactive gas affects the deposition rate and process stability (target poisoning), and it also leads unintended oxide interlayer formation. Although the target poisoning mechanism has been well known, little attention has been paid on understanding the interlayer formation during the reactive sputtering. In this research, we studied the interlayer formation during the reactive sputtering. A DC magnetron sputtering process is carried out to deposit an aluminum oxide film on a silicon wafer. From the real-time process monitoring and material analysis, the target poisoning phenomena changes the reactive gas balance at the initial stage, and affects the interlayer formation during the reactive sputtering process.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.582-584
• /
• 2023

연합학습(Federated Learning)에서는 여러 참가자가 서로 간의 데이터를 공유하지 않고 협력하여 하나의 모델을 학습할 수 있다. 그 중 수직 연합학습(Vertical Federated Learning)은 참가자 간에 동일한 샘플에 대해 서로 다른 특성(Feature)를 가지고 학습한다. 또한 서로 다른 특성(Feature)에는 입력의 라벨(Label)도 포함하기 때문에 라벨을 소유한 참가자 외에는 라벨 정보 또한 접근할 수 없다. 이처럼 다양한 참가자가 학습에 참여하는 경우 악의적인 참가자에 의해 모델이 포이즈닝 될 여지가 존재함에도 불구하고 수직 연합학습에서는 관련 연구가 부족하다. 포이즈닝 공격 중 백도어 공격은 학습 과정에 관여하여 특정 입력 패턴에 대해서 모델이 공격자가 원하는 타겟 라벨로 예측하도록 오염시키는 공격이다. 수직 연합학습에서는 참가자가 학습과 추론 모든 과정에서 관여하기 때문에 백도어 공격에 취약할 수 있다. 본 논문에서는 수직 연합학습에서의 최신 백도어 공격과 한계점에 대해 분석한다.

• Journal of Wetlands Research
• /
• v.25 no.4
• /
• pp.306-314
• /
• 2023

Stream sediments are an important component of water quality management because they are receptors of various pollutants such as heavy metals and organic matters emitted from upland sources and can be secondary pollution sources, adversely affecting water environment. To effectively manage the stream sediments, identification of primary sources of sediment contamination and source-associated control strategies will be required. We evaluated the performance of machine learning models in identifying primary sources of sediment contamination based on the physico-chemical properties of stream sediments. A total of 356 stream sediment data sets of 18 quality parameters including 10 heavy metal species(Cd, Cu, Pb, Ni, As, Zn, Cr, Hg, Li, and Al), 3 soil parameters(clay, silt, and sand fractions), and 5 water quality parameters(water content, loss on ignition, total organic carbon, total nitrogen, and total phosphorous) were collected near abandoned metal mines and industrial complexes across the four major river basins in Korea. Two machine learning algorithms, linear discriminant analysis (LDA) and support vector machine (SVM) classifiers were used to classify the sediments into four cases of different combinations of the sampling period and locations (i.e., mine in dry season, mine in wet season, industrial complex in dry season, and　industrial complex in wet season). Both models showed good performance in the classification, with SVM outperformed LDA; the accuracy values of LDA and SVM were 79.5% and 88.1%, respectively. An SVM ensemble model was used for multi-label classification of the multiple contamination sources inlcuding landuses in the upland areas within 1 km radius from the sampling sites. The results showed that the multi-label classifier was comparable performance with sinlgle-label SVM in classifying mines and industrial complexes, but was less accurate in classifying dominant land uses (50~60%). The poor performance of the multi-label SVM is likely due to the overfitting caused by small data sets compared to the complexity of the model. A larger data set might increase the performance of the machine learning models in identifying contamination sources.

• Proceedings of the Korean Vacuum Society Conference
• /
• 2012.08a
• /
• pp.314-315
• /
• 2012

최근들어 반도체 및 디스플레이 소자의 구조가 복잡해짐에 따라 다층 박막 증착에 대한 중요성이 날로 증가하고 있다. 본 연구에서는 다층 박막을 효율적으로 증착하기 위해 회전이 가능한 육각건을 개발하였고, 이를 이용하여 에너지 절약형 단열 유리 증착 공정을 구현 하였다. 개발된 회전형 육각건은 기존 플래너형 스퍼터링 건의 확장형으로서 최대 6개의 물질을 하나의 챔버에서 증착이 가능하도록 구성되었다. 기존 공정의 경우 서로 다른 물질 증착을 위해서는 각각의 챔버가 필요한 반면, 회전형 육각건을 이용할 경우 하나의 챔버에서 공정을 진행할 수 있어 원가 절감이 가능하다. Fig. 1은 개발된 회전형 육각건의 모식도로서, 스퍼터링 타겟이 장착 가능한 건과, 회전부로 구성되어 있다. 이를 이용하여 투명전극-금속-투명전극-금속-절연체로 구성되어 있는 에너지 절약형 단열 유리용 다층 박막 증착 공정을 개발하였다. 이때 알루미늄이 도핑된 ZnO (AZO)는 RF 마그네트론 스퍼터로, 금속 박막은 DC 스퍼터, $SiO_2$ 및 SiN과 같은 절연 박막은 $O_2$와 $N_2$ 분위기에서 반응성 RF 스퍼터로 각각 증착하였다. Base pressure는 $10^{-7}$ torr였으며, 증착 시 공정 압력은 1~3 mTorr로 조정하였다. 증착 균일도 향상을 위해 20 rpm의 속도로 기판을 회전시켰다. Fig. 2(a)는 ZnO-Ag-ZnO 구조로 이루어진 다층 박막의 단면을 관찰한 투과전자 현미경 사진으로 각 층간의 계면이 뚜렷하게 나타남을 확인할 수 있으며, 각 층간의 intermixing 현상이 발생하지 않음을 확인 가능하다. 이를 보완하기 위해 Fig. 2(b)에서 보는 바와 같이 XPS를 이용하여 depth profile을 측정하였다. 각 층에서 서로 다른 물질이 발견되는 현상, 즉 교차 오염이 발생함에 따라 나타나는 intermixing 없이 거의 순수한 형태의 ZnO, Ag 박막 성분이 검출되었다. 이는 6개의 서로 다른 물질이 장착된 회전형 육각건을 이용하여 고 품질의 다층 박막 증착이 가능함을 제시하는 결과이다. 증착된 다층 박막의 균일도는 3.8%, 가시광선 영역에서 80% 이상의 투과도, 면저항 값은 3 ${\Omega}/{\Box}$ 이하를 보임으로서 에너지 절약형 단열 유리로서의 사양을 만족시키는 결과를 제시하였다.

• Journal of Korean Society of Environmental Engineers
• /
• v.31 no.1
• /
• pp.51-57
• /
• 2009

Biofouling in seawater reverse osmosis (SWRO) desalination process causes many problems such as flux decline, biodegradation of membrane, increased cleaning time, and increased energy consumption and operational cost. Therefore biofouling is considered as the most critical problem in system operation. To control biofouling in early stage, detection of the most problematic bacteria causing biofouling is required. In this study, six model bacteria were chosen; Bacillus sp., Flavobacterium sp., Mycobacterium sp., Pseudomonas aeruginosa, Pseudomonas fluorescens, and Rhodobacter sp. based on report in the literature and phylogenetic analysis of seawater intake and fouled RO membrane. The adhesion to RO membrane, the high pressure resistance, and the hydrophobicity of the six model bacteria were examined to find out their fouling potential. Rhodobacter sp. and Mycobacterium sp. were found to attach very well to RO membrane surface compared to others used in this study. The test of hydrophobicity revealed that the bacteria which have high hydrophobicity or similar contact angle with RO membrane ($63^{\circ}$ of contact angle) easily attached to RO membrane surface. P. aeruginosa which is highly hydrophilic ($23.07^{\circ}$ of contact angle) showed the least adhesion characteristic among six model bacteria. After applying a pressure of 800 psi to the sample, Rhodobacter sp. was found to show the highest reduction rate; with 59-73% of the cells removed from the membrane under pressure. P. fluorescens on the other hand analyzed as the most pressure resistant bacteria among six model bacteria. The difference between reduction rates using direct counting and plate counting indicates that the viability of each model bacteria was affected significantly from the high pressure. Most cells subjected to high pressure were unable to form colonies even thought they maintained their structural integrity.

• Journal of the Microelectronics and Packaging Society
• /
• v.27 no.4
• /
• pp.83-89
• /
• 2020

For the past few decades, as part of efforts to protect the environment where fossil fuels, which have been a key energy resource for mankind, are becoming increasingly depleted and pollution due to industrial development, ecofriendly secondary batteries, hydrogen generating energy devices, energy storage systems, and many other new energy technologies are being developed. Among them, the lithium-ion battery (LIB) is considered to be a next-generation energy device suitable for application as a large-capacity battery and capable of industrial application due to its high energy density and long lifespan. However, considering the growing battery market such as eco-friendly electric vehicles and drones, it is expected that a large amount of battery waste will spill out from some point due to the end of life. In order to prepare for this situation, development of a process for recovering lithium and various valuable metals from waste batteries is required, and at the same time, a plan to recycle them is socially required. In this study, we introduce a nanoscale pattern transfer printing (NTP) process of Li2CO3, a representative anode material for lithium ion batteries, one of the strategic materials for recycling waste batteries. First, Li2CO3 powder was formed by pressing in a vacuum, and a 3-inch sputter target for very pure Li2CO3 thin film deposition was successfully produced through high-temperature sintering. The target was mounted on a sputtering device, and a well-ordered Li2CO3 line pattern with a width of 250 nm was successfully obtained on the Si substrate using the NTP process. In addition, based on the nTP method, the periodic Li2CO3 line patterns were formed on the surfaces of metal, glass, flexible polymer substrates, and even curved goggles. These results are expected to be applied to the thin films of various functional materials used in battery devices in the future, and is also expected to be particularly helpful in improving the performance of lithium-ion battery devices on various substrates.

• Korean Journal of Remote Sensing
• /
• v.36 no.5_3
• /
• pp.1053-1066
• /
• 2020

Sulfur dioxide (SO₂) is primarily released through industrial, residential, and transportation activities, and creates secondary air pollutants through chemical reactions in the atmosphere. Long-term exposure to SO₂ can result in a negative effect on the human body causing respiratory or cardiovascular disease, which makes the effective and continuous monitoring of SO₂ crucial. In South Korea, SO₂ monitoring at ground stations has been performed, but this does not provide spatially continuous information of SO₂ concentrations. Thus, this research estimated spatially continuous ground-level SO₂ concentrations at 1 km resolution over South Korea through the synergistic use of satellite data and numerical models. A stacking ensemble approach, fusing multiple machine learning algorithms at two levels (i.e., base and meta), was adopted for ground-level SO₂ estimation using data from January 2015 to April 2019. Random forest and extreme gradient boosting were used as based models and multiple linear regression was adopted for the meta-model. The cross-validation results showed that the meta-model produced the improved performance by 25% compared to the base models, resulting in the correlation coefficient of 0.48 and root-mean-square-error of 0.0032 ppm. In addition, the temporal transferability of the approach was evaluated for one-year data which were not used in the model development. The spatial distribution of ground-level SO₂ concentrations based on the proposed model agreed with the general seasonality of SO₂ and the temporal patterns of emission sources.