• Journal of Surface Science and Engineering
• /
• v.46 no.4
• /
• pp.145-152
• /
• 2013

The growth behavior of nanocrystalline diamond (NCD) film has been studied for three different substrates, i.e. bare Si wafer, 1 ${\mu}m$ thick W and Ti films deposited on Si wafer by DC sputter. The surface roughness values of the substrates measured by AFM were Si < W < Ti. After ultrasonic seeding treatment using nanometer sized diamond powder, surface roughness remained as Si < W < Ti. The contact angles of the substrates were Si ($56^{\circ}$) > W ($31^{\circ}$) > Ti ($0^{\circ}$). During deposition in the microwave plasma CVD system, NCD particles were formed and evolved to film. For the first 0.5h, the values of NCD particle density were measured as Si < W < Ti. Since the energy barrier for heterogeneous nucleation is proportional to the contact angle of the substrate, the initial nucleus or particle densities are believed to be Si < W < Ti. Meanwhile, the NCD growth rate up to 2 h was W > Si > Ti. In the case of W substrate, NCD particles were coalesced and evolved to the film in the short time of 0.5 h, which could be attributed to the fact that the diffusion of carbon species on W substrate was fast. The slower diffusion of carbon on Si substrate is believed to be the reason for slower film growth than on W substrate. The surface of Ti substrate was observed as a vertically aligned needle shape. The NCD particle formed on the top of a Ti needle should be coalesced with the particle on the nearby needle by carbon diffusion. In this case, the diffusion length is longer than that of Si or W substrate which shows a relatively flat surface. This results in a slow growth rate of NCD on Ti substrate. As deposition time is prolonged, NCD particles grow with carbon species attached from the plasma and coalesce with nearby particles, leaving many voids in NCD/Ti interface. The low adhesion of NCD films on Ti substrate is related to the void structure of NCD/Ti interface.

• Journal of the Korean Society of Radiology
• /
• v.14 no.7
• /
• pp.965-973
• /
• 2020

The purpose of this study was to quantitatively correlate the change of flow velocity and signal voiding in TOF-MRA. We made our phantom to control the flow velocity, and changed the flow velocity in 16 steps from 8.0 to 127.3 mc/s. The TOF-MRA test was performed using a 3.0T MRI system and the signal intensity was measured by classifying the signal voiding length and image into the In flow, Mid flow, and Out flow. The length of signal voiding was the longest when the flow velocity was 127.3 cm/s and the signal intensity decreased with increasing flow velocity(p<0.05). In flow(-.547) and Mid flow(-.643) were negatively correlated with flow velocitys(p<0.05). In conclusion, it was confirmed that the increase in flow velocity was a major factor causing signal voiding in TOF-MRA. In the future, this study will provide basic data when studying sequences and parameters to reduce signal voiding in models with a high flow velocity.

• Journal of the Korean Crystal Growth and Crystal Technology
• /
• v.24 no.3
• /
• pp.89-93
• /
• 2014

In this study, effects of the V/III ratio on a-plane GaN epitaxial on r-plane grown by HVPE have been investigated. According to increasing of V/III ratio, the value of FWHM of a-plane (11-20) GaN and the value of surface roughness (Ra) were decreased. Growth rate of a-plane GaN epitaxial layer were increased until V/III ratio = 7 as the increasing of V/III ratio, but it was reduced at V/III ratio = 10. At V/III ratio = 10, the FWHM of a-plane (11-20) GaN RC and the surface roughness (Ra) were 829 arcsec and 1.58 nm, respectively, as the lowest value in this study. Also for V/III ratio = 10, cracks under surface or voids were observed the lowest values in images of optical microscope. An M-shaped azimuthal dependence over $360^{\circ}$ angle range was observed for all samples. At V/III ratio = 10, the difference of FWHM of a-plane GaN between $0^{\circ}$ and $90^{\circ}$ was 439 arcsec revealed as the lowest value in the 4 samples.

• Journal of the Korean GEO-environmental Society
• /
• v.9 no.6
• /
• pp.31-41
• /
• 2008

Consolidation characteristics of reclamated ground with dredged soil and methods of evaluating them are investigated in this paper. For a dredged and reclamated ground with a very high water content, self-weight consolidation being progressed, its consolidation characteristics are difficult to find since it is almost impossible to have a undisturbed sample. In order to overcome such a problem, methods of laboratory tests with disturbed sample were studied to obtain consolidation parameters required to analyze consolidation settlement in practices, using the conventional infinitesimal consolidation theory, were evaluated by carrying out various laboratory tests with disturbed soils such as oedometer test, constant rate of deformation test, Rowe-cell tests with ring diameters of 60 mm, 100 mm and 150 mm and the centrifuge model tests with 40 g-levels. Constitutive relations of void ratio - effective vertical stress - permeability were evaluated by using the inverse technique implemented with the finite strain consolidation theory and results of centrifuge model tests. Design soil parameters related to consolidation such as compression index, swelling index, coefficient of volume change and vertical and horizontal consolidation coefficients were proposed properly by analyzing the various test results comprehensively.

• Nuclear Engineering and Technology
• /
• v.26 no.1
• /
• pp.19-31
• /
• 1994

The objective of this study is to conduct a thermal-hydraulic analysis on the spent fuel pool and to evaluate a parametric effect for the thermal-hydraulic analysis of spent fuel pool. The selected parameters are the Reynolds Number and the gap flow through the oater gap between fuel cell and fuel bundle. The simplified flow network for a path of fuel cells is used to analyze the natural circulation phenomenon. In the flow network analysis, the pressure drop for each assembly from the entrance of the fuel rack to the exit of the fuel assembly is balanced by the driving head due to the density difference between the pool fluid and the average fluid in each spent fuel assembly. The governing equations ore developed using this relation. But, since the parameters(flow rate, pressure loss coefficient, decay heat, density)are coupled each other, iteration method is used to obtain the solution. For the analysis of the YGN 3&4 spent fuel rack, 12 channels are considered and the inputs such as decay heat and pressure loss coefficient are determined conservatively. The results show the thermal-hydraulic characteristics(void fraction, density, boiling height)of the YGN 3&4 spent fuel rack. There occurs small amount of boiling in the cells. Fuel cladding temperature is lower than 343.3$^{\circ}C$. The evaluation of parametric effect indicates that flow resistances by geometric effect are very sensitive to Reynolds number in the transition region and the gap flow is negligible because of the larger flow resistance in the gap flow path than in the fuel bundle.

• The Transactions of The Korean Institute of Electrical Engineers
• /
• v.67 no.1
• /
• pp.114-123
• /
• 2018

This paper concerns a design and learning method of softmax function neural networks based on K-means clustering. The partial discharge data Information is preliminarily processed through simulation using an Epoxy Mica Coupling sensor and an internal Phase Resolved Partial Discharge Analysis algorithm. The obtained information is processed according to the characteristics of the pattern using a Motor Insulation Monitoring System program. At this time, the processed data are total 4 types that void discharge, corona discharge, surface discharge and slot discharge. The partial discharge data with high dimensional input variables are secondarily processed by principal component analysis method and reduced with keeping the characteristics of pattern as low dimensional input variables. And therefore, the pattern classifier processing speed exhibits improved effects. In addition, in the process of extracting the partial discharge data through the MIMS program, the magnitude of amplitude is divided into the maximum value and the average value, and two pattern characteristics are set and compared and analyzed. In the first half of the proposed partial discharge pattern classifier, the input and hidden layers are classified by using the K-means clustering method and the output of the hidden layer is obtained. In the latter part, the cross entropy error function is used for parameter learning between the hidden layer and the output layer. The final output layer is output as a normalized probability value between 0 and 1 using the softmax function. The advantage of using the softmax function is that it allows access and application of multiple class problems and stochastic interpretation. First of all, there is an advantage that one output value affects the remaining output value and its accompanying learning is accelerated. Also, to solve the overfitting problem, L2-normalization is applied. To prove the superiority of the proposed pattern classifier, we compare and analyze the classification rate with conventional radial basis function neural networks.

• Journal of Korean Tunnelling and Underground Space Association
• /
• v.17 no.3
• /
• pp.353-362
• /
• 2015

Platform screen door (PSD) installed at underground subway station has reduced the safety accident, but it may cause poor air ventilation condition due to the isolated exhaust duct in the subway tunnel area. In this study, the additional ventilation system was suggested, which can be installed at a void space (i.e., storage room under stairs) of platform in order to improve efficiency of air ventilation rate. Exhausted air from platform was directed to underneath of platform and joined with existing ventilation duct of train exhaust system (TES). One subway station in Seoul city was selected to predict the effectiveness of the suggested lower exhaust system by using the computational fluid dynamics (CFD) analysis. The predicted mean age of air was decreased by 16.5% which proves the improvement of air ventilation efficiency when the suggested lower exhaust system was applied.

• Journal of the Microelectronics and Packaging Society
• /
• v.18 no.4
• /
• pp.49-53
• /
• 2011

High-speed Cu filling into a through-silicon-via (TSV) and simplification of bumping process by electroplating for three dimensional stacking of Si dice were investigated. The TSV was prepared on a Si wafer by deep reactive ion etching, and $SiO_2$, Ti and Au layers were coated as functional layers on the via wall. In order to increase the filling rate of Cu into the via, a periodic-pulse-reverse wave current was applied to the Si chip during electroplating. In the bumping process, Sn-3.5Ag bumping was performed on the Cu plugs without lithography process. After electroplating, the cross sections of the vias and appearance of the bumps were observed by using a field emission scanning electron microscope. As a result, voids in the Cu-plugs were produced by via blocking around via opening and at the middle of the via when the vias were plated for 60 min at -9.66 $mA/cm^2$ and -7.71 $mA/cm^2$, respectively. The Cu plug with a void or a defect led to the production of imperfect Sn-Ag bump which was formed on the Cu-plug.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.10A
• /
• pp.912-921
• /
• 2005

Optical burst switching(OBS) is a promising solution for building terabit optical routers and realizing If over WDM. Channel scheduling Algorithm for reduce contention is one of the major challenges in OBS. We address the issue of how to provide basic burst channel scheduling in optical burst switched WDM networks with fiber delay lines(FDLs). In OBS networks the control and payload components or a burst are sent separately with a time gap. If CHP arrives to burst switch node, because using scheduling algorithm for data burst, reservation resources such as wavelength and transmit data burst without O/E/O conversion, because contention and void between burst are happened at channel scheduling process for data burst that happened the link utilization and bust drop probability Existent proposed methods are become much research to solve these problems. Propose channel scheduling algorithm that use Release Time of bust to emphasize clearance between data and data dissipation that is happened in data assignment in this treatise and maximize bust drop probability and the resources use rate (RTUC : Release Time Unscheduled Channel). As simulation results, Confirmed that is more superior in terms of data drop and link utilization than scheduling algorithm that is proposed existing. As simulation results, confirmed improved performance than scheduling algorithm that is proposed existing in terms of survival of burst, efficiency resource and delay. However, In case load were less, degradation confirmed than existent scheduling algorithm relatively, and confirmed that is superior in data drop aspect in case of load increased.

• Proceedings of the Korean Vacuum Society Conference
• /
• 1998.02a
• /
• pp.77-77
• /
• 1998

Present silicon dioxide (SiOz) 떠m as intennetal dielectridIMD) layers will result in high parasitic c capacitance and crosstalk interference in 비gh density devices. Low dielectric materials such as f f1uorina뼈 silicon oxide(SiOF) and f1uoropolymer IMD layers have been tried to s이ve this problem. I In the SiOF ftlm, as fluorine concentration increases the dielectric constant of t뼈 film decreases but i it becomes unstable and wa않r absorptivity increases. The dielectric constant above 3.0 is obtain어 i in these ftlms. Fluoropolymers such as polyte$\sigma$따luoroethylene(PTFE) are known as low dielectric c constant (>2.0) materials. However, their $\alpha$)Or thermal stability and low adhesive fa$\pi$e have h hindered 야1리ru뚱 as IMD ma따"ials. 1 The concept of a plasma processing a찌Jaratus with 비gh density plasma at low pressure has r received much attention for deposition because films made in these plasma reactors have many a advantages such as go여 film quality and gap filling profile. High ion flux with low ion energy in m the high density plasma make the low contamination and go어 $\sigma$'Oss피lked ftlm. Especially the h helicon plasma reactor have attractive features for ftlm deposition 야~au똥 of i앙 high density plasma p production compared with other conventional type plasma soun:es. I In this pa야Jr, we present the results on the low dielectric constant fluorocarbonated-SiOF film d밑JOsited on p-Si(loo) 5 inch silicon substrates with 00% of 0dFTES gas mixture and 20% of Ar g gas in a helicon plasma reactor. High density 띠asma is generated in the conventional helicon p plasma soun:e with Nagoya type ill antenna, 5-15 MHz and 1 kW RF power, 700 Gauss of m magnetic field, and 1.5 mTorr of pressure. The electron density and temperature of the 0dFTES d discharge are measUI벼 by Langmuir probe. The relative density of radicals are measured by optic허 e emission spe따'Oscopy(OES). Chemical bonding structure 3I피 atomic concentration 따'C characterized u using fourier transform infrared(FTIR) s야3띠"Oscopy and X -ray photonelectron spl:’따'Oscopy (XPS). D Dielectric constant is measured using a metal insulator semiconductor (MIS;AVO.4 $\mu$ m thick f fIlmlp-SD s$\sigma$ucture. A chemical stoichiome$\sigma$y of 야Ie fluorocarbina$textsc{k}$영-SiOF film 따~si야영 at room temperature, which t the flow rate of Oz and FTES gas is Isccm and 6sccm, res야~tvely, is form려 야Ie SiouFo.36Co.14. A d dielec$\sigma$ic constant of this fIlm is 2.8, but the s$\alpha$'!Cimen at annealed 5OOt: is obtain려 3.24, and the s stepcoverage in the 0.4 $\mu$ m and 0.5 $\mu$ m pattern 킹'C above 92% and 91% without void, res야~tively. res야~tively.