博士 / 國立交通大學 / 電子物理系所 / 98 / This dissertation is devoted to study the localized surface plasmons (LSPs) in spatially controlled array structures. In the first part, three periodic Au coated Si nano rod (SiNR) arrays on Si substrate with different distances between the SiNRs were fabricated by electron beam lithography (EBL). The LSPs is induced when these spatially controlled array structures were irradiated by an un-polarized white light from a tungsten halogen lamp, we found that the trends of reflectance spectra indicate that the LSP frequency can be spatially controlled by manipulating the distances between the SiNRs of the arrays. In addition, the experimental results were compared with 2D numerical simulations based on the finite element method. The simulation result of each array in the near field regions can reveal subtle characteristic of the intensity distributions including the grating effects (the higher order diffractions) and more LSP modes than in the far field reflection observation. After the investigation of the spatially controlled LSPs, we started the second part of the research, fluorescence signals of quantum dots (QDs) influenced by the three array structures of the Au coated SiNRs. The QDs were spin-coated on a cover glass. And then, the three array structures were put upside down on the cover glass in sequence. We further developed a new technique which can obtain the optical image of the array structures without losing information of the QD locations at the time of QDs fluorescence measurement. We removed the filter and focused the excitation laser on the substrate to obtain the optical image of the array structures (the substrate is therefore bright, and the top of the SiNR is dark). And then, we put back the filter and focused the excitation laser on the QDs to obtain the fluorescence image of the QDs. After superposing the two images, the relative locations between the QDs and the Au coated SiNRs were defined. The same 50 QDs were successively observed on the three array structures (one by one). On the densest gold coated SiNRs array structure, the highest QDs fluorescence quenching rates are observed. And on the sparsest array structure which provides the LSP frequency closer to the emission band of QDs, the highest QDs fluorescence enhancement rates are observed. The QDs fluorescence enhancement effects are observed on the locations proximity enough (but not touched by the Au) to the Au coated SiNRs, and on the locations near the mid point of two Au coated SiNRs (where the constructive interference is formed). And we found again that when the QDs contact with Au coated SiNRs, instead of enhancement, a non-radiatve process may occur, leading to QDs fluorescence quenching. Finally, both in the near field region, 2D numerical simulation results are consistent with the experiment results.

碩士 / 國立交通大學 / 電機資訊國際學位學程 / 98 / To achieve the goal of higher transmission rate and seamless internet access everywhere, femtocells are foreseen to be a key fixed-mobile convergence technology. As femtocells will be user-installed, self-organization techniques are required to automatically integrate themselves into the network and avoid co-femtocell interference, especially in urban dense deployment area. To cope with these challenges, a two-step femtocell-cooperation algorithm has been designed in this thesis. The first step is the clustering phase, during which femtocells will form groups with other femtocells according either to femtocell Base Station or to femtocell user interference measurements, thus setting two types of methods. Then, a frequency allocation scheme is proposed based on the grouping distribution; it corresponds to the tuning phase. The interference based clustering is the key part of this work; it has therefore been studied and evaluated according to several sets of parameters like power threshold or limit member per cluster. The results compare the performance of the two clustering methods with their parameters and show that the femtocell user SINR and the outage probability can be highly improved but at a capacity cost.

博士 / 國立交通大學 / 機械工程系所 / 98 / Hyperchaos of chaotic systems, Yin-Yang chaos, new fuzzy model, new fuzzy logic controllers, generalized chaos synchronization via GYC partial region stability theory and pragmatical asymptotically stability theorem are studied in this thesis. The main points in the researches are shown as follow: 1.Analyzing Yin chaos of the classical Lorenz system and comparing it with Yang chaos. 2.Hyperchaos in a new Mathieu-van der Pol system is identified by phase portraits, power spectrum, Lyapunov exponents and 2-D and 3-D parameters diagrams. Three positive Lyapunov exponents are found for system with four states. 3.Chaotic control and synchronization for a system by GYC partial region theory. 4.New fuzzy model is proposed to simulate the complicated chaotic behaviors via only two linear subsystems and used to carry out synchronization of complicated chaotic systems and different chaotic systems. 5.Simplified fuzzy logic constant controller (FLCC) is presented to achieve generalized synchronization.

博士 / 國立交通大學 / 光電工程系所 / 98 / Field-sequential-color (FSC) liquid-crystal-display (LCD) has been driven by the demand for energy-efficient Green display in recent years. The driving scheme and slow LC response, however, have led to issues such as color gamut shrinkage, color crosstalk between successive fields, and insufficient backlight intensity, etc., not to mention the visual artifact, color break-up (CBU). This research is to study those colorimetric issues and the feasible solution, based on spatial-temporal modulation, and to develop a methodology to evaluate CBU. A novel two-field driving scheme and the associated system have been proposed to facilitate the FSC LCD without color filters. A spatially-modulated color backlight is implemented to provide degrees of freedom in spatial domain in addition to the two temporal ones for conveying full-color information. This configuration of FSC LCD specifically features the feasibility of commercial LC modes, e.g. Multi-domain Vertical Alignment (MVA), without sacrificing the mentioned colorimetric properties. Additionally, the innate nature of reduced contrast sensitivity between two primary-mixed fields suppresses the CBU to be less sensitive than that appears in the fundamental three-field FSC LCD. Finally, the psychophysical index, relative contrast sensitivity (RCS), has demonstrated the applicability to evaluate CBU among various driving schemes, which can be a useful feedback for system optimization.

博士 / 國立交通大學 / 光電工程系所 / 98 / TFT-LCD Technology has become the mainstream in replacing CRTs in TV and other applications. In order for thin-film-transistor liquid-crystal display (TFT-LCD) technology to claim an even larger share of the TV market, manufacturers must continue to improve image quality. The issue of motion blur effect in TFT-LCD TVs has been widely discussed lately. Among the reported LC modes the Pi-cell, also known as Optically Compensated Bend (OCB) mode, has been found to be a strong candidate to reduce the motion blur effect. Due to the LC-cell structure and the driving scheme, Pi-cell is not only a wide viewing angle display technology, but also the fastest-response LCD mode among the commercialized LCD modes. Because of the fast response of the Pi-cell, the field sequential color LCD (FSC-LCD) which displays R, G and B colors in sequence in a pixel is promising for the high-resolution display. However, Pi-cell possesses intrinsic transition and recovery issues which lead to compromised optical properties, thus, limited in its applications. In this dissertation, we proposed two modified Pi-cells to resolve these issues. 1.Proposed Nanostructure Enhanced Pi-cell (NE-Pi-cell) modified the surface of alignment layer (PI) to create nuclei for speeding up the transition rate. The transition time of a NE-Pi-cell was reduced from 2 minutes to less than 1 sec compared with the conventional one. Moreover, the transition process was uniformly completed without applying high voltage pulse (~18V). 2.Even if the previous topic could speed up the transition rate to almost zero, the critical voltage also needed to be kept over 2V for maintaining the bend state of a Pi-cell. Based on the light dispersion property of liquid crystals, the critical voltage of R, G and B were different; i.e. VR< VG< VB. In order to confirm that the Pi-cell could operate in bend state, a critical voltage larger than VB needed to be chosen, which led to compromised optical qualities. Therefore, we suggested reactive monomer modified Pi-cell (RMM-Pi-cell) to eliminate the splay-to bend state transition. Besides, because of smaller residual retardation at the dark state, the static contrast ratio of a RMM-Pi-cell, compared with conventional Pi-cell, was improved from 26 to 288 (the test samples fabricated in laboratory), up to a factor of 11 without using compensation films. We have demonstrated a novel alignment layer modified method of a Pi-cell for uniform and fast transition without high voltage pulses. Moreover, a transition-free and high optical performance Pi-cell has also proposed in this dissertation. The proposed novel Pi-cells only need one or two simple extra processes in conventional manufacturing of TFT-LCD. Combining the research results with current high image quality LCD technology, the novel high image quality Pi-cell will be realized. The results can not only improve the image qualities for general TFT-LCD applications, but also realize the FSC-LCDs.

碩士 / 國立交通大學 / 顯示科技研究所 / 98 / Liquid crystal displays (LCDs) have played an important role in flat panel display (FPD) industry and have already achieved a significant market impact, in recent years. Cholesteric Liquid Crystals (CLCs) have great advantages for reflective mode displays due to their lower power consumption, backlight-free, polarizer-free, color filter-free structure and good readability. This let they are suitable for the flexible display applications such as electronic paper, signboard, bill-board, label, etc. The reflectance of CLC can be changed by applying voltage. The blue shift, however, accompany the process. This phenomenon is a hindrance to display applications. In this thesis, the blue shift were analyzed by color difference (岛E*ab), because the 岛E*ab could get an exact value for color change than the reflection spectrum. To solve the blue shift phenomenon, the various host dielectric anisotropic (岛ε) and birefringence (岛n) were prepared by mixing negative and positive nematic LCs. The experiment results showed the values of 岛E*ab was increased to 14 for the 岛ε of LCs was 7.8, while the applying incremental voltage (岛V) was 7.3V. The same 岛E*ab=14 was obtained when the applied voltage increased to 25V in the 岛ε=1.3 material. The result indicated the low 岛ε CLC mixtures could effectively suppress blue shift phenomenon and produce smooth gray scale. In addition, the lower 岛n CLC mixtures were found to have the smaller 岛E*ab.

碩士 / 國立交通大學 / 外國語文學系外國文學與語言學碩士班 / 98 / This thesis adopts the psychoanalytic perspective to analyze Margaret Atwood’s novel, The Edible Woman. Under the framework of psychoanalysis, we discuss the protagonist’s, Marian’s, awareness of being a woman and also her subjectivity of being a hysteric woman. Besides, this thesis will re-examine the relationship between Marian and other characters via the problems confronted in a woman’s life and offer a new definition of femininity. The thesis is divided into five parts. Chapter One describes the significance of the thesis topic, the literature review of the novel, and the methodology I will apply for in the following chapters. Chapter Two concerns what Marian desires. In addition to explaining why the hysteric’s body is an erotogenic body, I will further elucidate that Marian’s desire is actually an unsatisfied desire. In this chapter, I will also discuss the relationship between Marian’s anorexia nervosa and her mother. Chapter Three discusses Marian’s sexual confusion and what it is to be a woman for her. In Chapter Four, I want to prove that transference between Duncan and Marian is the key point for the restoration of Marian’s health and her “normal” life. Finally, after a series of discussions, we understand what it would be like to be a woman for Marian and redefine femininity in psychoanalytic perspective. Therefore, in my conclusion, I will combine the above theses to unveil the message hidden in the novel’s title, “The Edible Woman.”

博士 / 國立交通大學 / 材料科學與工程學系 / 98 / Low temperature polycrystalline silicon (LTPS) Ni-metal induced lateral crystallization (NILC) thin-film transistors (TFTs) have been investigated in this thesis. Ni impurities trapped inside the NILC poly-Si films is an issue. Therefore the Ni-gettering method is proposed to effectively reduce Ni residues within the NILC poly-Si films. It involves using gettering layers/etching stop layers as the Ni-gettering structure. The 100-nm-thick top α-Si and phosphorous-doped α-Si layers serve as the gettering layers, while the middle ~5-nm-thick chem-SiO2 layer is used as an etching stop layer. Moreover the proposed gettering method is utilized in the fabrication of LTPS NILC TFTs and Si nanowire (NW) channel transistors to investigate the effect of Ni-metal inside poly-Si on the device performance. First, the α-Si film is employed to getter Ni-silicides within NILC poly-Si film and Ni reduction is demonstrated by SEM. After the gettering process, fewer and smaller silicide-etching holes are found at the NILC/NILC boundaries. It is found that top α-Si films, Ni-gettering layers, transfer into NILC poly-Si verified by SEM. This means that during the gettering process, Ni atoms diffused from the NILC poly-Si film through chem-SiO2 into the Ni-gettering layer due to the concentration gradient. Compared with the previous α-Si/PECVD-SiNx study, the thermal budget is greatly reduced. It’s because of low Ni diffusivity in SiNx films resulting a long annealing time as 90 h at 550oC in N2 ambient. In order to improve the Ni-gettering efficiency, phosphorous-doped α-Si films are further used by ion implantation. After a gettering process, there are almost no silicide-etching holes observed at the NILC/NILC boundaries. These results indicate that phosphorous dopants could improve the gettering efficiency of α-Si due to the solubility enhancement of Ni impurities. But the gettering efficiency do not obviously improves until doping phosphorus ions reach a dose of 1×1016 cm-2. The concentration distribution of Ni is similar to that of phosphorous atoms since the projection range of phosphorous ions is set at the middle of the α-Si film. This result also indicates that phosphorus could trap Ni atoms. The proposed gettering method is further utilized in the fabrication of LTPS NILC TFTs. As NILC TFTs are treated with a Ni-gettering process, they reveal lower leakage current, higher on/off current ratio, higher mobility, and better uniformity. These improvements are all attributed to the reduction of Ni impurities in gettered poly-Si films. Finally, a simple method and low-cost process is used to manufacture the NW channels. The feature of process is the method of forming sidewall spacer of MOSFET. The poly-Si sidewall spacer NW channels self-alignment form in the process of defining source/drain (S/D). Both the vertical width (WNW) and the horizontal sidewall thickness (TNW) of poly-Si NWs are about 70 nm. The cross-section of fabricated poly-Si NWs is similar to triangular shape by an anisotropic etching. Compared with the traditional TFTs, the side-gated NWs TFTs have higher channel controllability. Moreover, the performance of NILC NWs TFTs is improved after a Ni-gettering process. This is because of the reduction of Ni and NiSi2 precipitates randomly trapped at poly-Si/gate oxide interfaces and poly-Si grain boundaries.

博士 / 國立交通大學 / 資訊科學與工程研究所 / 98 / Optical packet switching (OPS) has been considered to be a promising paradigm to support a wide range of applications with different time-varying and high bandwidth demands for future optical metropolitan area networks (MANs). This thesis presents the design of an experimental optical-header processing and access control system (OPACS) for an OPS WDM metro slotted-ring network. OPMACS is endowed with two distinctive features. First, OPMACS has been designed for a dual unidirectional slotted ring network using in-band signaling control. Each control header is in-band time-division-multiplexed with its corresponding payload within a slot. OPACS enables the optical headers across all parallel wavelengths to be efficiently received, modified, and re-transmitted by means of a wavelength-time conversion technique. Moreover, OPACS embodies a versatile medium access control (MAC) scheme, referred to as the distributed multi-granularity and multi-window reservation (DMGWR) mechanism, which is particularly advantageous for traffic of high and varying loads and burstiness. Basically, DMGWR requires each node to make reservation requests prior to transmissions while maintaining a distributed queue for ensuring fair access of bandwidth. By “multi-granularity”, each node can make a reservation of multiple slots at a time. By “multi-window”, each node is allowed to have multiple outstanding reservations within the window size. Simulation pit the OPACS network against two other existing networks, simulation results show that the OPACS network outperforms these networks with respect to throughput, access delay, and fairness under various traffic patterns. Experimental results demonstrate that all optical headers are removed and combined with the data in a fully synchronous manner, justifying the viability of the system.

碩士 / 國立交通大學 / 資訊學院資訊科技產業專班 / 98 / Real-Time Global illumination in computer graphics is a very important topic. In the past, a popular approach was ambient occlusion. Then CryTek Company improves this algorithm. They propose an approach to approximate ambient occlusion in the screen space. In this thesis, we introduce a further enhancement to approximate global illumination by combining screen space ambient occlusion and reflective shadow map in a deferred shading context. We include local direct illumination and global indirect illumination by extending screen space ambient occlusion. On the other hand, we reduce GPU overhead by using screen-space interpolation. Since the algorithm works in screen space, it does not depend on the geometric complexity of 3D scenes. We demonstrate the result which can be displayed for large and fully dynamic scenes at real-time frame rates.