• Title/Summary/Keyword: Multi-layer Modular

Search Result 15, Processing Time 0.021 seconds

Filling the Submicron Contact Holes with Al Alloys (AI 합금의 Contact Hole Filling 에 관한 연구)

  • 김용길
    • Journal of the Korean Vacuum Society
    • /
    • v.2 no.4
    • /
    • pp.474-479
    • /
    • 1993
  • Submicron contact hole filling with aluminum alloys has been achieved with a multistep metallization method, which utilizes a metal " flow" or self-diffusion process at elevated temperatures after the metal was sputter-deposited. A multi-chamber, modular sputtering system was employed to deposit aluminum alloys and subsequently to anneal the deposited metal films under vacuum at high temperatures. The film were deposited on 200 mm wafers with planar, dc magnetron sputtering sources without anysubstrate bias. The basic process steps studied for the multistep metallization include an initial layer deposition at low temperatures less than $100^{\circ}C$, and an annealin gstep at elevated temperatures, between 450 and $550^{\circ}C$. The degree of planarization or step coverage was dependent strongly upon the temperature and time of the flow step and complete filling of the submicron contacts with aluminum alloys was achieved. Responsible mechanisms for the enhancement in step coverge and factros determining uniform and reproducible flow of aluminum alloys during the high temperauture step are discussed.discussed.

  • PDF

A reconfigurable modular approach for digital neural network (디지털 신경회로망의 하드웨어 구현을 위한 재구성형 모듈러 디자인의 적용)

  • Yun, Seok-Bae;Kim, Young-Joo;Dong, Sung-Soo;Lee, Chong-Ho
    • Proceedings of the KIEE Conference
    • /
    • 2002.07d
    • /
    • pp.2755-2757
    • /
    • 2002
  • In this paper, we propose a now architecture for hardware implementation of digital neural network. By adopting flexible ladder-style bus and internal connection network into traditional SIMD-type digital neural network architecture, the proposed architecture enables fast processing that is based on parallelism, while does not abandon the flexibility and extensibility of the traditional approach. In the proposed architecture, users can change the network topology by setting configuration registers. Such reconfigurability on hardware allows enough usability like software simulation. We implement the proposed design on real FPGA, and configure the chip to multi-layer perceptron with back propagation for alphabet recognition problem. Performance comparison with its software counterpart shows its value in the aspect of performance and flexibility.

  • PDF

Experimental Study on Modular Community Planting for Natural Forest Restoration (자연림 복원을 위한 모듈군락식재 실험연구)

  • Han, Yong-Hee;Park, Seok-Gon
    • Korean Journal of Environment and Ecology
    • /
    • v.36 no.3
    • /
    • pp.338-349
    • /
    • 2022
  • This study aims to investigate whether modular community planting, which entailed planting a variety of species of seedlings at high density, was more effective in restoring natural forests than the existing mature tree planting. We also investigated whether the planting density of the modular community planting facilitates growth or improves the tree layer coverage. We conducted outdoor experiments in which the samples were divided into a mature tree planting plot (control plot), where mature trees were planted at wide intervals, and a modular community planting (MCP) plot (treatment plot), where multiple seedlings were planted in high density. The MCP plot was further divided into the plot in which 3 seedlings were planted per m2 and the plot of 1 seedling per m2. We measured the specimens' survival rate, growth rate (tree height, crown width, and root collar diameter), and cover rate for 26 months from May 2019 and the predicted future tree height growth using the measured tree height. The survival rate and relative growth rate of the MCP were higher than those of the mature tree planting plot. The vertical coverage rate of the tree crown in the MCP exhibited complete coverage of the ground before 23 months, while the coverage rate of the mature tree planting decreased due to transplantation stress. The seedlings in the MCP, which were planted at high density, grew well and were predicted to grow higher than the mature trees in the large tree planting plot within 5 to 6.5 years after planting. It was due to multiple species, seedlings, high-density planting, and planting foundation improvements, such as soil enhancement and mulching. In other words, the seedlings planted in the MCP had a higher survival rate as their environmental adaptation after planting was better, and their early growth was also larger than the trees in the mature planting plot. The high-density mixed planting of various native species not only mitigated the inter-complementary environmental pressures but also facilitated growth by inducing competition between species. Moreover, the planting foundation improvement effectively increased the seedlings' viability and growth rate. A reduction in follow-up management costs is expected as the tree layer coverage sharply increases due to the higher planting density. In the MCP (3 seedlings per m2 and 1 seedling per m2), the tree height growth was promoted with the higher planting density, and the crown width and root collar diameter tended to be larger with the lower planting density, but these differences were not statistically significant.

Partially Connected Multi-Layer Perceptrons and their Combination for Off-line Handwritten Hangul Recognition (오프라인 필기체 전표용 한글 인식을 위한 부분 연결 다층 신경망과 결합)

  • 백영목;임길택;진성일
    • Journal of the Korean Institute of Telematics and Electronics C
    • /
    • v.36C no.4
    • /
    • pp.87-94
    • /
    • 1999
  • This paper presents a study on the off-line handwritten Hangul (Korean) character recognition using the partially connected neural network (PCNN), which is based on partial connections between the input receptive fields and the hidden nodes. The hidden nodes of three PCNNs have ten receptive fields and different input feature sets. And we introduce modular partially connected neural network (MPCNN), The MPCNN combines three PCNNs with a merging network. The learning scheme of the proposed networks is composed of two steps: PCNN learning step and the merging step of combining three PCNN s. In the merging step, another merging PCNN network is introduced and trained by regarding the hidden output of each PCNN as a new input feature vector. The performance of the proposed classifier is verified on the recognition of 18 off-line handwritten Hangul characters widely used in business cards in Korea.

  • PDF

Primer Evaluation for the Detection of Toxigenic Microcystis by PCR (독소 생성 Microcystis 검출을 위한 PCR primer의 평가)

  • 이현경;김준호;유순애;안태석;김치경;이동훈
    • Korean Journal of Microbiology
    • /
    • v.39 no.3
    • /
    • pp.166-174
    • /
    • 2003
  • Microcystin produced by cyanobacteria in surface waters, such as eutrophic lake and river, is a kind of serious environmental problems due to its toxicity to human and wild animals. Microcystin is synthesized nonribosomally by the large modular multi-functional enzyme complex known as microcystin synthetase encoded by the mcy gene cluster. Amplification of mcy genes by PCR from cultures and environmental samples is a simple and efficient method to detect the toxigenic Microcystis. In order to evaluate primers designed to detect toxic microcystin-producing strains, 17 cyanobacterial strains and 20 environmental samples were examined by PCR with 7 pairs of primers. Some microcystin-producing cyanobacteria were not detected with FAA-RAA, TOX4F-TOX4R and FP-RP primers. The fragment of unexpected size was amplified with NSZW2-NSZW1 primers in Microcystis strains isolated from the lakes in Korea. TOX1P-TOX1F primers failed in amplification of toxin-producing strains. Only MSF-MSR and TOX2P- TOX2F primers amplified the fragments of mcy genes from 11 strains of microcystin-producing Microcystis. The water samples taken from 20 lakes in Korea were analyzed by PCR using each of the primers. In all the water samples, cyanobacteria capable of producing microcystin were detected by the PCR with TOX2P-TOX2F primers. These results indicate that TOX2P-TOX2F primers are better than the other primers for detection of microcystin-producing Microcystis strains in Korea. The nucleotide sequences of mcy gene in Microcystis aeruginosa NIER10010 suggest genetic diversity of Korean isolates.