56016 research outputs found
Sort by
Liquid-Phase Synthesis of Highly Deformable and Air-Stable Sn-Substituted Li3PS4 for All-Solid-State Batteries Fabricated and Operated under Low Pressures
No full textThe liquid-phase synthesis (LS) of sulfide solid electrolytes (SEs) has promising potential for mass production of practical all-solid-state Li batteries (ASLBs). However, their accessible SE compositions are mostly metal-free. Moreover, liquid-phase-synthesized-SEs (LS-SEs) suffer from high electronic conductivity due to carbon impurities, resulting in below-par electrochemical performance of ASLBs. Here, the LS of highly deformable and air-stable Li3+xP1-xSnxS4 (0.19 mS cm(-1)) using 1,2-ethylene diamine-1,2-ethanedithiol with tetrahydrofuran is reported. A low heat-treatment temperature (260 degrees C) prevents the carbonization of organic residues. Importantly, a remarkable enhancement in the deformability of LS-SEs compared to that of conventional solid-state-synthesized SEs (SS-SEs) is identified for the first time. LiNi0.7Co0.15Mn0.15O2 electrodes employing LS-SEs in ASLBs significantly outperform those using SS-SEs, notably when assembled under a low fabricating pressure (148 vs 370 MPa, e.g., capacity loss: 2 vs 41 mA h g(-1)) or tested under a low operating pressure (12 or 3 MPa), which is attributed to reduced electrochemo-mechanical effects. Finally, when employing SEs that are exposed to air (dew point of -20 degrees C), LiNi0.7Co0.15Mn0.15O2 electrodes employing SEs with Sn-substituted composition or prepared by LS exhibit significantly better capacity retention than conventional SEs with Sn-free composition or prepared by SS (e.g., 92.2% for LS-Li3.2P0.8Sn0.2S4 vs 32.5% for SS-Li3PS4)
Viscosity approximation of the solution to Burgers' equations with shock layers
No full textViscous Burgers' equations with a small viscosity are considered and convergence of vanishing viscosity limit problem is investigated. We examine interior layers of a solution to viscous Burgers' equations, u(epsilon), as a viscosity parameter epsilon tends to zero. The inviscid model, i.e. when epsilon = 0, possesses the structure of scalar hyperbolic conservation laws, hence our studies deliver an important idea that arises in the field of shock discontinuities of nonlinear hyperbolic waves. The heart of the paper is to establish asymptotic expansions and utilize inner solutions of sharp transition, which are called a corrector function. With aid of corrector functions and energy estimates, we improve the convergence rate of ue to u(0) as O(epsilon(1/2)) in L-2(R) (O(epsilon) in L-loc(1)(R)) in the regions including shocks under an entropy condition
????????? ????????? ???????????? ?????? ????????? ????????? ????????? ??????
No full textDepartment of Industrial Engineeringope
Analytical Models to Design and Analyze for Surface-Mounted Permanent Magnet Motor
No full textDepartment of Mechanical Engineeringclos
EBSD ????????? ????????? ????????? ????????? ???????????? ????????? ?????? ??????
No full textDepartment of Materials Science and EngineeringIn general, the microstructure characterization of various materials, especially of metals and ceramics, is an essential step for studying the process-microstructure-property relationships. For polycrystalline microstructures of such materials systems, the typical aim for microstructure characterization is to find the quantitative measures of the microstructural parameters such as the grain morphology, grain size, grain texture, grain boundary texture and even compositional distribution. To obtain the relevant data, various experimental analysis methods such as scanning electron microscopy (SEM), x-ray diffraction (XRD), electron backscatter diffraction (EBSD), etc. are usually performed. Among them, the EBSD measurements have been extensively used because, it is possible to obtain from them local crystal orientation in a relatively fine scale. Using the orientation map, one can easily extract grain morphology, grain size, crystallographic texture, grain boundary characters in the sample measured. In this study, the polycrystalline microstructures of the copper (Cu, metal) thin films and the bulk yttrium iron garnet (YIG, ceramic) are studied through the EBSD measurement. The main goal of this study is to elucidate the microstructural evolution during the deposition of the Cu films and the annealing of the bulk YIG polycrystal.
Copper (Cu) has attracted much attention and has been heavily used in the electronic devices and applications such as ultra largescale integration (ULSI) metallization, flexible and wearable electronics, printed circuit boards, etc. because of its superior electrical and thermal conductivity, high electromigration (EM) resistance and low electronic resistance. These electrical, mechanical and thermal properties of Cu thin films are well known to be influenced by the microstructure parameters and their characteristics. As much as such a microstructure-property relationship is an important research subject, the process-microstructure relationship study is of equivalent interest. Numerous previous studies on the microstructure characterization of the Cu thin films show that the film deposition methods and condition change the final microstructure of the film. Also, the subsequent annealing process after the deposition of the films induce the changes, sometimes drastically, in microstructural characteristics, especially in the selective grain growth and resultant texture. However, there are only a small number of the EBSD studies investigating the microstructural evolutions during the deposition of the Cu films via the physical vapor deposition methods. For this study, we adopt the e-beam evaporation deposition method to prepare the Cu films to study microstructural changes during the deposition. Specifically, we investigate the microstructure characteristics of Cu seed layer deposited on the amorphous SiO2 substrate with various film thickness (100 nm, 300 nm and 500 nm) using e-beam evaporation at room temperature. Such as-deposited microstructures are analyzed using the scanning electron microscopy (SEM), the Ga+ ion beam irradiation, the X-ray diffraction (XRD) and the electron backscatter diffraction (EBSD). From the SEM measurements, it is found that, as the film thickness increases, the surface morphology evolves through the surface diffusion for the morphological characteristics of the film surface is well comparable to those for the zone T in the structure zone model (SZM). As a result of the surface diffusion, the e-beam evaporated Cu thin films coarsen as the film thickness increases, showing a self-similar state in grain morphology. While the surface of the film features a corrugated, ridge-like morphology (i.e., valleys and hills), the Ga+ ion beam irradiation images reveal that the inner microstructure is composed of both the nanocrystalline grains and the abnormally coarsened grains. From the EBSD measurements of the film surface, it is found that the grain size increases with increasing film thickness, following a power law (G = Adn), where G is the average grain size, d is the film thickness, n is the growth exponent and A is a proportionality constant. The growth exponent (n) for the surface grains is found to be 0.23, close to 0.25, theoretical value of the grain coarsening via surface diffusion. However, the texture analysis of the film surface reveals an interesting aspect that both (100) and (111) // ND texture increases as the film is deposited. The results contradict the prediction suggested by Thompson and Carel, where, as the film thickness increases at a certain temperature, (111) // ND texture is initially dominant and decreases, and (100) // ND texture increases while consuming (111) grains. Especially for the current area, the (111) // ND texture is dominant for all samples. Also, it is found that the fraction of the ??3 grain boundary also increases with increasing film thickness. From results of grain texture (PFs, IPFs) and grain boundary texture (MDFs, KAM) analyses, it can be hypothesized that the low-energy (111) surface and/or the low-energy ??3 grain boundary play a crucial role in determining the characteristics of the Cu films deposited at room temperature via the e-beam evaporation method. The results from the Kernel average misorientation analysis suggest that, as the film grows, the strain at the surface reduces, signifying that the coarsening of the surface grains relieve the strain. The reason for this might be either the surface energy reduction during (111) // ND texture development or the strain energy release of the twin growth, or both. This issue is the subject of the future study.
Yttrium iron garnet (YIG, Y3Fe5O12) is a ferromagnetic material widely used for microwave communication devices, and for magneto-optical applications such as circulators, isolators, phase shifters, resonators, etc. In many studies, pulsed laser deposition (PLD), radio frequency (RF) / direct current (DC) sputter, conventional solid-state sintering, or spin coating have been generally used to fabricate the YIG films or bulk samples. Recently, a new method for preparing bulk YIG samples was introduced where a substrate-free polycrystalline YIG was fabricated by sintering the powder obtained from a sol-gel synthesis which is relatively simple and cost-effective, and resulted in fine-grained and relatively homogenous YIG bulk samples. In this study, two YIG samples, compacted from the calcined sol-gel powder, are annealed at 1400 ???C for 350 min and 1750 min, respectively. It is found that, during the annealing, the polycrystalline YIG coarsens via normal grain growth, developing a steady-state in grain size distribution, grain texture, and grain boundary texture. However, the 1750-min-annealed sample bears sharper and narrower x-ray peaks, smaller local misorientation, and more uniform element distribution than the 350-min-annealed one, signifying that the strain decreases and the steady-state in composition is not attained during the annealing at 1400 ???C. Notably, little segregation of metal elements is observed on the grain boundary of the current YIG samples, as opposed to previously reported conventionally sintered YIGs using yttrium-oxide and iron-oxide powders, often featuring either second phases or segregation of Fe on YIG grain boundaries.clos
????????? ????????? ????????? ???????????? ????????? ?????? ?????? ??? ?????? ??????
Get PDFDepartment of Mechanical EngineeringThe potential danger of invisible hazardous substance leakage accident is increasing, such as hazardous chemical leakage accidents in industrial complexes, potential risks of aging nuclear power plants, and international chemical terrorism threats. In particular, hazardous chemical, biological, or radioactive substances leaked into the atmosphere cause irreversible damage to nature, and there is a risk of human damage if prompt action is not taken. Therefore, estimating the emission source and the amount of invisible hazardous substances is required to minimize human casualties and increase public safety. As the risk of hazardous material leakage and potential terrorism increases in random places, it is difficult using traditional systems such as pre-installed ground sensors in a specific area. This thesis proposes autonomous search method for estimating the source of hazardous materials using a mobile sensor attached to an unmanned aerial vehicle (UAV). Since the mobile sensor can be freely deployed to any arbitrary places, it is possible to monitor a wider area with a relatively low cost. Besides, this approach is an unmanned autonomous system, so it has the advantage of minimizing secondary human casualties that may additionally occur during search.
The source term estimation (STE) using mobile sensors is considered to be a challenging problem because the sensor measurements from atmospheric gas dispersion are sparse, intermittent, and time-varying due to the turbulence and the sensor noise. Thus, Bayesian inference-based estimation technique, sequential Monte Carlo method (i.e., particle filter), is used to estimate the source by using the inaccurate measurements which is easily influenced by air turbulence and sensor noise in this thesis. The autonomous search algorithms using information theory are also proposed. In the proposed algorithms, the information entropy (i.e., uncertainty of estimation) is calculated by using information theory and the agent choose the action to move to the next sensing location that can minimize the expected uncertainty. In other words, the proposed information-theoretic search algorithm is reward-based decision making approaches that use information entropy as a reward. The receding horizon and Gaussian mixture model clustering approaches are adopted to improve the search performance in various environment. Since the time required to compute all of the respective rewards increases as the number of action candidates increases, the policy-based autonomous source term search and estimation algorithm is proposed using deep neural network and reinforcement learning approach to determine efficient search path considering continuous action space. Furthermore, this thesis proposes a cooperative search method for multiple unmanned mobile vehicles based on game theory. The inaccuracy of sensor measurement values can be reduced by using multiple mobile sensors with the fusion approach, so the source of hazardous substances can be quickly estimated. The negotiation based on the game theory can improve the group search performance for source term estimation and search. Finally, to verify the performance of the proposed algorithm, numerical simulation and flight test results using an actual gas measurement sensor and multicopter drone are presented.ope
????????? ??? ???????????? ????????? ??????
Get PDFDepartment of Mathematical SciencesFor an irrational number x in (0,1), we consider its approximation by continued fraction expansion and by decimal expansion. Comparing the two approximations, kn(x) is defined by the number of digits of continued fraction expansion of x that are determined by those of the n-th decimal approximation of x, for any n ??? 1. G. Lochs proved in 1964 that the ratio kn(x) n converges to a constant 6log2log10 ?? 2 as n ??? ??? (see [5]). The observable point is that we can prove the theorem without using the decimal expansion. Focusing on this fact, we drop the base notation and take ???Nx??? Nx as an approximation of x for any integer N. This approximation converges to x as N increases like the n-th decimal expansion approximation. Therefore we define a quantity KN(x) in a similar sense, comparing the continued fraction approximation with ???Nx??? Nx . In this thesis, we restate the theorem of Lochs in terms of KN(x) and logN and prove the renewed statement. Furthermore we restate and reprove several results about the distribution of KN(x) logN given by C. Faivre when N = 10n.ope
?????????????????????/????????? ??????????????? ?????? ???????????? ?????? ???????????? ???/?????? ??????????????? ?????? ??????????????????
Get PDFDepartment of Materials Science and EngineeringAs the performance of next-generation organic devices is approaching that of inorganic devices, the importance of the stability of organic devices is being increasingly emphasized. Organic materials are vulnerable to high temperature, and solvent exposure, and can degrade when exposed to light for a long timethey are particularly vulnerable to moisture. Unencapsulated organic devices are virtually impossible to use in real life. Therefore, it is essential to improve the encapsulation technologies for the commercialization of organic devices. As electronic products based on inorganic materials are not affected by these external factors, there is a general lack of studies on encapsulation technology for stability. In this study, I fabricated a crack-free organic/inorganic hybrid thin film that suppresses the generation of cracks, which act as a moisture penetration paths, by introducing a CTE (Coefficient of thermal expansion) relaxation layer and adding SiO2 nanoparticles to the organic layer. The film structure was optimized by adjusting the purge time in the atomic layer deposition process, to obtain a dense 30 nm thick Al2O3 film with a refractive index of 1.6 or more. The CTE mismatch between the Al2O3 and PDMS films was mitigated by introducing Parylene C to suppress the cracking of the film. The water-blocking effect of the PDMS-SiO2 nanocomposite film was also investigated at various SiO2 nanoparticle concentrations. Perovskite/Si solar cells encapsulated in the optimized crack-free multilayer thin film maintained 90% of the initial performance even after 90 h of operation at 55 ??C and 85% relative humidity.ope
????????? ???????????? ????????????: ????????? ????????? ????????? ??????????????? ????????? ?????? ?????? ?????????
No full textDepartment of Creative Design Engineeringclos
DoubleStep : Efficient Gradient Compression in Distributed DNN Training System
No full textDepartment of Computer Science and Engineeringclos