Because of its attractive material properties like high hardness, high toughness, and excellent high temperature strength, materials like silicon nitride are becoming more common for use in high performance applications. However, there have been limited studies of the fatigue behavior of small cracks in silicon nitride and other materials toughened by grain bridging mechanisms. This study explores using micro Raman spectroscopy, fatigue crack growth data and results from static fracture experiments to determine a bridging stress profile for silicon nitride doped with MgO and Y₂O₃ as sintering additives. These bridging stress profiles allow for the creation of a geometry specific fatigue threshold R-curve which can be used to develop a fatigue endurance strength prediction tool to aid in the design of products using the material. Cyclical fatigue experiments conducted on bend beams with induced semi-elliptical surface cracks were conducted to verify the prediction tool. The results show that no bend beams with this crack geometry failed below the predicted endurance level. It is expected that this method can be extended to create fatigue endurance strength predictions for other materials similarly toughened by grain bridging and other mechanisms. / Graduation date: 2013

by Poon Ming-Cheong. / Thesis (Ph.D.)--Chinese University of Hong Kong, 1988. / Bibliography: leaves [163-175]

Thesis (M.S.)--University of Texas at El Paso, 2009. / Title from title screen. Vita. CD-ROM. Includes bibliographical references. Also available online.

No description available.

No description available.

We have developed a new coarsening theory that more completely describes grain growth of a system of anisotropic particles such as completely faceted crystals by Ostwald ripening. Our model takes the anisotropy of surface energies on dissimilar facets into account, and the particle sizes are described by a distribution function. The theory is applied to study the coarsening of ,B-silicon nitride in a liquid medium due to the anisotropic growth of grains in different crystallographic directions. A model of the growth of silicon nitride grains is obtained based on the numerical solution of the equations of the new theory. Computer experiments are performed to determine how the distribution function evolves, to investigate the influence of various parameters such as diffusion and interfacial reaction constants on grain growth and to extract grain growth exponents from this model in order to determine the growth mechanisms that are responsible for the anisotropic growth behaviour. Only preliminary numerical results are available thus far due to 1/r instabilities that occur in the theory as r → 0. / Thesis (M.Sc.)-University of Natal, Pietermaritzburg, 2002.

No description available.

The factors influencing the direct nitridation of silicon, including the effects of the native oxide layer covering the surface of silicon, the effects of hydrogen contained in the nitridation gas and the catalytic effects of metals added to the raw material silicon, were investigated, using a tubular flow reactor and a fluidized-bed reactor operated at temperatures ranging from 1150��C to 1390��C in a stream of nitrogen containing 10% hydrogen. The nitridation of silicon is not initiated until the native oxide is removed by an assistance of hydrogen contained in argon during the pretreatment or in the nitridation gas mixture. An induction period is observed before the initiation of the nitridation and depends on the nitridation temperature as well as the pretreatment time, which is associated with the removal of the oxide layer. The presence of hydrogen in the nitridation atmosphere is crucial for the nitridation of silicon. When pretreated silicon grains are exposed to nitrogen without hydrogen for a time period as short as 5 minutes, the subsequent nitridation, even with hydrogen, becomes extremely slow. The concentration of hydrogen as low as 0.3% is effective for sustaining the reactivity of silicon for the nitridation. The results suggest the formation of a protective layer on the surface of silicon when silicon grains are exposed to nitrogen without hydrogen. The protective film is suspected to be silicon oxynitride, or a mixture of silicon oxynitride and silicon dioxide or silicon nitride formed from the reaction of silicon with oxygen and nitrogen, depending on the temperature of its formation. However, the protective film does not form on the native oxide layer, and the reactivity of silicon is resumed upon the removal of the native oxide. An addition of calcium (as low as 0.125%) or yttrium (1.0-2.0%) to silicon suppresses the formation of ��-silicon nitride while iron enhances the formation of silicon nitride. Copper promotes not only the nitridation but also the formation of ��-silicon nitride at 1200��C, but enhances the ��-phase formation at higher temperatures. The role of liquid phases on the formation of ��-/��-silicon nitride was also discussed based on the nitridation of silicon impregnated with copper, calcium, silver, chromium and tungsten. / Graduation date: 2003

2 ��m average sized silicon powder was nitrided with 90% N���/10% H��� in a fluidized-bed reactor, operated at 1200��C, 1250��C and 1300��C. To fluidize silicon powder, alumina particles with an average size of 300 ��m were used as an inert fluidizing conditioner. The feasibility and operating conditions of the fluidization were studied at room temperature. The effects of silicon content and operating temperature on the nitridation of silicon as well as on the formation of ��- and ��-silicon nitride were investigated in batch and semi-continuous operations. The effects of the average residence time of silicon/alumina mixtures in the fluidized-bed reactor on the nitridation process were studied in semi-continuous operations. In batch operations, a maximum mass fraction of 15 wt% silicon powder could be added to alumina particles at temperatures in the range of 1200 to 1300��C without changing the fluidization quality. When the silicon fraction was increased to 20 wt%, fluidization failed immediately. With a mass fraction of 5% silicon powder, almost 100% ��-silicon nitride, which was preferred in applications, was found in the product. ��-silicon nitride was facilitated with an increase in silicon fractions in silicon/alumina mixtures. The nitridation process was strongly affected by the reaction temperature. The overall conversion of silicon increased with an increase in reaction temperature. Higher temperature also promoted the formation of ��-silicon nitride. The overall conversion of silicon into silicon nitride was also enhanced by hydrogen concentrations. An increase in hydrogen concentration facilitated the formation of ��-form silicon nitride. In the semi-continuous operation, the nitridation of 30 wt% silicon/70 wt% alumina mixtures could be achieved without changing the fluidizing quality. Almost 100% ��-silicon nitride was found in the product when a 20 wt% silicon/80 wt% alumina mixture was nitrided at 1250��C for an average residence time of up to 4 hours. However, ��-silicon nitride was formed when the mixture was nitrided at 1300��C for an average residence time of 3 hours. A mathematical model incorporating kinetic data and carryover of silicon powder was developed to described the total conversion of silicon in batch operation. A semi-continuous model was also proposed, which successfully predicted the overall conversion of silicon powder. / Graduation date: 1997

Flash memory is one sort of non-volatile memory, focus on the dates holding and capacity. Conventional non-volatile memory applies poly-crystalline for floating gate material, because the poly-crystalline (like poly-silicon) itself is the semiconductor material, will cause leakage problem, recently, Oxide-nitride-oxide multi-layer structure is under development for the place of conventional floating gate. Because it is the insulator material, can suppress leakage current, and it contains a deeper trapping energy level, and has a partial trapped carriers phenomenon to give a multi-bits memory solution. My effort is to propose a pair of ONO three layers stack, which is located close to the beneath of D/S region and a column like. Such structure can overcome miniaturization limitation of channel length, and a somewhat depth oxide can promise good isolation and separation between the trapping layer and other area, and a reliable distance of the two trapped unit can prevent interference issue. My proposal can suppose a higher devices density and a feasible and flexible solution to develop memory devices, a cost down to be more competitive, certainly bring much favor for the future improvement.