Nanocrystalline materials with grain size less than 100 nm have been receiving much attention because of their unparallel properties compared with their microcrystalline counterparts. Because of its high hardness, nanocrystalline nickel has been used for MEMS. Long term thermomechnical properties and deformation mechanism at both ambient and elevated temperatures need to be evaluated which is vital for reliability of its applications as structural material. In this thesis, nanoindentation creep of nanocrystalline nickel with an as-deposited grain size of 14 nm was characterized at elevated temperatures. The nanoindentation creep rate was observed to scale with temperature and applied load (or stress), and could be expressed by an empirical power-law equation for describing conventional crystalline solids. Creep activation energy was found to be close to that for grain boundary self-diffusion in nickel. The activation volume was also evaluated using a stress relaxation technique. The creep results were compared with those for fine-grained nickel in the literature. Possible mechanisms were discussed in light of the creep rate and temperature ranges. To provide a direct comparison, uniaxial creep tests were conducted on nanocrystalline nickel with an as-deposited grain size of 14 nm at 398 K. It was found that stress exponents under the two test conditions are almost the same, indicating a similar creep mechanism. However, the strain rate measured by nanoindentation creep was about 100 times faster than that by uniaxial creep. The rate difference was discussed in terms of stress states and the appropriate selection of Tabor factor. To further explore the time-dependent plastic behavior, multiple unload-reload tests were conducted on electrodeposited nanocrystalline nickel in both compression and tension. A hysteresis was observed during each unload-reload cycle, indicating irreversible energy dissipation. The dissipated energy was evaluated and the energy dissipation rate was found to increase with the flow stress to the third power and sensitive to the stress state (tension or compression). A mechanistic model based on grain boundary sliding was proposed to describe the unload-reload behavior. Experimental results were found to be in good agreement with the model predictions, suggesting the observed hysteresis was indeed caused by grain boundary sliding.

With the ability to generate fine displacements with a resolution down to sub-nanometers, piezoelectric actuators (PEAs) have found wide applications in various nano-positioning systems. However, existence of various effects in PEAs, such as hysteresis and creep, as well as dynamics can seriously degrade the PEA performance or even lead to instability. This raises a great need to model and control PEAs for improved performance, which have drawn remarkable attention in the literature. Sliding mode control (SMC) shows its potential to the control of PEA, by which the hysteresis and other nonlinear effects can be regard as disturbance to the dynamic model and thus rejected or compensated by its switching control. To implement SMC in digital computers, this research is aimed at developing novel discrete models and discrete SMC (DSMC)-based control schemes for PEAs, along with their experimental validation. The first part of this thesis concerns with the modeling and control of one-degree of freedom (DOF) PEA, which can be treated as a single-input-single-output (SISO) system. Specifically, a novel discrete model based on the concept of auto-regressive moving average (ARMA) was developed for the PEA hysteresis; and to compensate for the PEA hysteresis and improve its dynamics, an output tracking integrated discrete proportional-integral-derivative-based SMC (PID-SMC) was developed. On this basis, by making use of the availability of PEA hysteresis models, two control schemes, named “the discrete inversion feedforward based PID-SMC” and “the discrete disturbance observer (DOB)-based PID-SMC”, were further developed. To illustrate the effectiveness of the developed models and control schemes, experiments were designed and conducted on a commercially available one-DOF PEA, as compared with the existing ones. The second part of the thesis presents the extension of the developed modeling and control methods to multi-DOF PEAs. Given the fact that details with regard to the PEA internal configurations is not typically provided by the manufacturer, a state space model based on the black box system identification was developed for the three-DOF PEA. The developed model was then integrated in the output tracking based discrete PID-SMC, with its effectiveness verified through the experiments on a commercially available three-DOF PEA. The superiority of the proposed control method over the conventional PID controller was also experimentally investigated and demonstrated. Finally, by integrating with a DOB in the discrete PID-based SMC, a novel control scheme is resulted to compensate for the nonlinearities of the three-DOF PEA. To verify its effectiveness, the discrete DOB based PID-SMC was applied in the control experiments and compared with the existing SMC. The significance of this research lies in the development of the discrete models and PID-based SMC for PEAs, which is of great help to improve their performance. The successful application of the proposed method in the control of multi-DOF PEA allows the application of SMC to the control of complicated multi-inputs-multi-outputs (MIMO) systems without details regarding the internal configuration. Also, integration of the inversion based feedforward control and the DOB in the SMC design has been proven effective for the tracking control of PEAs.

Piezoelectric actuators are well suited for high precision mechanical and electrical engineering applications. However, its performance in regulator configurations has been limited due to hysteresis. The hysteresis in these actuators means that multiple input states can result in the same output, which introduces a further design variable (initial state) in the regulation problem. It is proposed that certain initial states result in better regulation performance based on the structure of the Preisach model. These initial states are called “neutral states”. In this thesis, hysteresis and piezoelectric actuators are introduced as background information. The Preisach model is used in this work to describe the hysteresis behaviour of a customized shape control unit SS15 due to its convenient general structure and ability to model hysteresis. The representation tests are performed and a Preisach model is subsequently constructed and verified by comparing simulation and experimental results to ensure that the hysteresis inherent in the piezoceramic actuators of the SS15 is suitably described by this model. In order to evaluate the regulation performance for a given desired output, uniformly-distributed noise is injected at the input side of the SS15 in open- and closed-loop tests. It is demonstrated, by both simulation and experimental results, that the system output drifts less when it starts from the neutral state in open-loop tests. A PI regulator is implemented in the closed-loop tests. When the system is driven from the neutral state, both simulation and experimental results demonstrate that the system requires less control effort for closed-loop regulation.

Piezoelectric actuators are well suited for high precision mechanical and electrical engineering applications. However, its performance in regulator configurations has been limited due to hysteresis. The hysteresis in these actuators means that multiple input states can result in the same output, which introduces a further design variable (initial state) in the regulation problem. It is proposed that certain initial states result in better regulation performance based on the structure of the Preisach model. These initial states are called “neutral states”. In this thesis, hysteresis and piezoelectric actuators are introduced as background information. The Preisach model is used in this work to describe the hysteresis behaviour of a customized shape control unit SS15 due to its convenient general structure and ability to model hysteresis. The representation tests are performed and a Preisach model is subsequently constructed and verified by comparing simulation and experimental results to ensure that the hysteresis inherent in the piezoceramic actuators of the SS15 is suitably described by this model. In order to evaluate the regulation performance for a given desired output, uniformly-distributed noise is injected at the input side of the SS15 in open- and closed-loop tests. It is demonstrated, by both simulation and experimental results, that the system output drifts less when it starts from the neutral state in open-loop tests. A PI regulator is implemented in the closed-loop tests. When the system is driven from the neutral state, both simulation and experimental results demonstrate that the system requires less control effort for closed-loop regulation.

The objective of this thesis is to provide a quick process to estimate iron losses of the electric machines with various structures that employ electromagnetic steels in their designs. Due to non-uniform distributed operational magnetic flux densities resulted from the machine structures, the iron losses of machines can not be properly estimated. The aim of the scheme is to modify the deviations among the measurements and the calculations. At first, several standardized steel modules that can be used to assemble the machine structures are established, then a test-bed with closed magnetic path to supply operational magnetic field inside those electric machines is constructed. To calculate the iron loss of each module, the averaged flux densities for each area of the module are measured by needle probe method, and the corresponding magnetic field intensities are calculated from Jiles and Atherton (J-A) model hysteresis model. Finally, the iron loss datasheets for each module are constructed by these measurements, and the iron losses of machines are estimated through assembling the modules. In this thesis, a switched-reluctance machine (SRM) is selected for assessment comparisons, and its iron loss can be calculated through datasheets according to the machine structure. It is believed that the findings of this study can provide a valuable reference and a reliable process in motor designing and manufacturing.

In this study, metal-oxide-semiconductor (MOS) capacitances were prepared with rf magnetron sputtering of Ta2O5 on both n-type GaN and p-type GaN. And metal-insulator-semiconductor (MIS) capacitances were prepared with conjugated rigid-rod polymers PBT on n-type GaN. The processes of fabrication the diodes were shown, and the structures of MOS/MIS diodes were represented. Hysteresis was observed in high-frequency capacitance-voltage (C-V) measurements. And the hysteresis was changed with different scanning delay time on scanning step. They were ascribed to mobile charges and interface charges. The carrier concentration were calculated and compared with the Hall results. The flatband voltage and threshold voltage were calculated and compared with C-V curves which were measured.

In this thesis, the a-PMT (amorphous lead magnesium titanate) membranes have been prepared by sol-gel technique as H+ ion sensitive layers. The C-V measurements of the a-PMT/SiO2/Si EIS structures prepared by spin-on coating are used for examining the fabrication parameters and sensing properties. There exhibits the quasi-Nerstain response of 55¡V59 mV/pH in the range of pH 2¡V12, fabricated with the Mg-modified content of 4 mole%, the firing temperature of about 400 ¢J and the thickness of about 0.5

In this thesis, the aluminum nitride (AlN) thin film was selected as a sensing membrane for the H+ ion-sensitive field-effect transistor (ISFET). The AlN thin films were prepared by a rf sputtering technology on the reference electrode/electrolyte /AlN /SiO2/p-Si/Al structure. The capacitance-voltage (C-V) measurement was used to detect the H+ ion concentration and the C-V characteristic curves were obtained in the different pH buffer solutions. On the other hand, AlN thin films were also prepared on the double layer structure of AlN/SiO2 gate ISFET devices. After packaging, the current-voltage (I-V) measurement with a PID temperature controller was utilized to measure a series of the I-V characteristic curves. The threshold voltage can be obtained to evaluate the pH sensitivity in the different pH buffer solutions. Additionally, the effects of non-ideal factors, such as temperature effect, drift and hysteresis phenomenon on the characteristics of the ISFET are also measured, analyzed and compared with other sensing materials. According to the experimental results, it can be found that the ISFET based on aluminum nitride thin film has a superior high pH sensitivity of approximately 50~58 mV/pH at 25¢J. The drift and hysteresis are dependent on the H+ ion concentration in pH=1~11, in which the drift rate increases with the pH value increased and the hysteresis magnitude depends on the measuring time and route. It is found that the hysteresis widths measured in pH=7®3®7®11®7 cycle at 960s, 1920s and 3840s loop time are 1.0, 1.5 and 4.5 mV, respectively. When the temperature effect was considered, it was found that the ISFET could be operated at 5~65¢J, in which, the pH sensitivity increased as the ambient temperature increased with the temperature coefficient of sensitivity of about 0.13 mV/pH¢J. In addition, the output voltage of AlN pH-ISFET can be obtained by a constant current constant voltage (CCCV) read out circuit with a fairly linear response, stability and reproducibility in the pH measuring cycle. From the characteristics mentioned above, the AlN thin film can be as a sensing membrane for pH-ISFET applications.

Hydrogen ion-sensitive field effect transistor with sol-gel-derived La-modified lead titanate gate Department of Electrical Engineering, National Sun Yat-Sen University *Jen-Fu Su **Ying-Chung Chen Abstract In this thesis the amorphous lead lanthanum titanate (Pb1-xLaxTi1-x/4O3, PLT) membrane has been prepared by sol-gel method as a novel pH-sensitive layer. The lead lanthanum titanate membrane was directly deposited on the SiO2(1000Å)/p-Si substrate by spin-on coating to form the PLT/SiO2/Si EIS structure. The C-V measurement was used for examining the fabrication parameters and sensing properties. Moreover, the PLT membrane was grown onto the SiO2 gate ISFET as the PLT/SiO2 gate ISFET. The electrical properties with the different parameter conditions can be obtained by the I-V measurement. Experimental results show that the fabrication parameters and characteristics of the PLT membrane are determined at the La-modified content about 3 mol% and the firing temperature of 400¢J via the EIS structure. There exhibits the pH response of about 44-52 mV/pH in the range of pH 2-12. Furthermore, the nonideal factors, such as drift of 0.1-0.3 mV/h, hysteresis of 2-13 mV and lifetime decay of about 72 mV/pH-day, can be also obtained via the I-V properties of the ISFET. Finally, the hardware architecture of pH measuring system has been built up. The system makes use of constant current and voltage bias technique to ensure that the variations of the output voltage can give directly the variations of pH value. For the purpose of achieving the function of data calibration and driving the liquid crystal display (LCD), the 8051 microprocessor is employed. Keywords: ISFET, Sol-gel, Lead lanthanum titanate, Drift, Hysteresis, pH meter * Student ** Advisor

Surface stabilized ferroelectric liquid crystal device (SSFLCD) has fast response time and exhibits excellent bistability. The SSFLCD has wide view angle because it operates in the in plane switching (IPS) mode. However, SSFLCD is a thin device, the uniform of its cell gap is difficult to control. When surface stabilized ferroelectric liquid crystal align, it is easy to bring in zigzag defect. In this study, the use of SU-8 photoresist to make photospacer by using photolithographic technique is studied. We control the thickness of photospacer at 1.08£gm. By using photospacer, we can get the cell with uniform cell gap. The influence of the property of the alignment layer and photospacer on surface stabilized ferroelectric liquid crystal alignment are discussed. We find that when the liquid crystal is injected anti-parallel to the rubbing direction, the alignment of Surface stabilized ferroelectric liquid crystal is easily influenced by photospacers. Wetting characteristics of the substrates is found to strongly affect the alignment of the ferroelectric liquid crystal.