京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第18281号 / 工博第3873号 / 新制||工||1594(附属図書館) / 31139 / 京都大学大学院工学研究科電子工学専攻 / (主査)教授 川上 養一, 教授 藤田 静雄, 准教授 須田 淳 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DFAM

京都大学 / 0048 / 新制・課程博士 / 博士(工学) / 甲第19997号 / 工博第4241号 / 新制||工||1656(附属図書館) / 33093 / 京都大学大学院工学研究科電子工学専攻 / (主査)教授 木本 恒暢, 教授 藤田 静雄, 准教授 船戸 充 / 学位規則第4条第1項該当 / Doctor of Philosophy (Engineering) / Kyoto University / DGAM

京都大学 / 0048 / 新制・課程博士 / 博士(エネルギー科学) / 甲第20476号 / エネ博第345号 / 新制||エネ||69(附属図書館) / 京都大学大学院エネルギー科学研究科エネルギー社会・環境科学専攻 / (主査)教授 石原 慶一, 教授 野平 俊之, 准教授 奥村 英之 / 学位規則第4条第1項該当 / Doctor of Energy Science / Kyoto University / DFAM

No description available.

No description available.

No description available.

No description available.

No description available.

The material properties of low stress silicon nitride make it a possible replacement material for beryllium in X-ray windows. In this study, X-ray windows made of LPCVD deposited low stress silicon nitride are fabricated and characterized. The Young's modulus of the LPCVD low stress silicon nitride are characterized and found to be 226±23 GPa. The residual stress is characterized using two different methods and is found to be 127±25 MPa and 141±0.28 MPa. Two support structure geometries for the low stress silicon nitride X-ray windows are used. X-ray windows with thicknesses of 100 nm and 200 nm are suspended on a silicon rib support structure. A freestanding circular geometry is used for a 600 nm thick X-ray window. The 100 nm and 200 nm thick low stress silicon nitride X-ray windows with a silicon support structure are burst tested, cycling tested and leak rate tested. The average burst pressure for the 100 and 200 nm films on a silicon support structure are 1.4 atm and 2.2 atm respectively. Both 100 nm and 200 nm windows are able to withstand a difference in pressure of 1 atm for over 100 cycles with a leak rate of less than 10-10 mbar-L/s.The low stress silicon nitride with 100 nm and 200 nm thicknesses, the 600 nm freestanding low stress silicon nitride windows and freestanding 8 micron thick beryllium windows are mechanical shock resistance tested. The support structure low stress silicon nitride and beryllium windows are tested with an applied vacuum. The freestanding 600 nm thick low stress silicon nitride windows burst at 0.4 atm and are therefore mechanical shock wave tested without an applied vacuum. The support structure low stress silicon nitride windows fractured when subjected to an acceleration of roughly 5,000 g. The 8 micron thick beryllium windows are subjected to accelerations of over 30,000 g without fracturing. A quasistatic model is used to show that for low stress silicon nitride with a freestanding circular geometry, an acceleration of 106 g is required to have the same order of magnitude of stress caused by a pressure differential of 1 atm. Low stress silicon nitride can act as a replacement for beryllium in X-ray windows, but the support geometry, residual stress, and strength of the material need to be optimized.

The resonance frequency of a silicon nitride (SiN) nano-electromechanical systems (MEMS/NEMS) can be measured precisely due to their large quality factor that is associated to low thermomechanical fluctuations. While these properties enable the fabrication of high performance sensors, their use will eventually raise questions regarding their long-term stability, notably for calibration purposes. The long-term frequency stability and aging of SiN are less studied than the short-term fluctuations such as thermomechanical noise. Long-term aging studies exist for quartz clocks as well as MEMS silicon clocks and accelerometers, but not for SiN resonators with high quality factors. Thus, in this work we conduct the aging study of SiN membranes fabricated by our lab, by constantly tracking changes of the resonance frequency of the device over a long period. The evolution of the frequency drift is tracked, by optical interrogation, continuously for 135 days with a digital phase locked loop (PLL). Our device is placed in a cell under high vacuum to suppress air damping on our resonating membrane. Furthermore, due to its high sensitivity to temperature changes, our silicon nitride resonator and vacuum chamber are placed in an air bath providing a stable temperature (within 0.5 K over 135 days in the present case). To compensate further the frequency drifts induced by temperature changes, a multimeter measures the resistance of a calibrated thermistor placed inside the vacuum environment. The measured frequency drift for the aging periods of 135 days was of 300 parts per million (ppm) and was consistent with previously reported double logarithmic models for quartz oscillators. The initial stage of negative frequency drift, in our aging data, is consistent with the behaviour expected from the desorption of water due to the transition from ambient air environment to high vacuum. We review models explaining how water adsorption/desorption impacts our membrane's frequency by (1) inducing chemical reaction stresses (most important effect), (2) through the contribution of the water surface tension stress (non-negligible effect), and (3) through mass loading from water molecules (weakest effect). After this initial negative trend, the membrane frequency drift inverts and increases almost linearly, in a fashion consistent with loss of mass from desorption of other chemical species. To identify these chemical species, X-ray photoelectron spectroscopy measurements were conducted on a reference membrane stored in an ambient setting and on our membrane placed under vacuum during our aging studies. The aged membrane, compared to its reference counterpart, contained substantially less alkaline ion contaminants (i.e., sodium, calcium and potassium), most likely due to desorption of these species during the aging measurement, and to the increase in adsorption occurring on the reference membrane concurrently. We therefore hypothesize that trapped negative charges, which is a typical phenomenon within dielectric materials such as SiN, might progressively attract positive ion contaminants over time when the device is exposed to ambient air.