Thermal Stress Fracture of Optical Glasses

Since glass is brittle and thermally insulating, significant thermal stresses, especially in larger components often lead to catastrophic failure. My work combines experimental measurement of glass thermomechanical failure, as well as computational simulations of transient thermal stresses arising in precision optical components used in fusion research. For the experimental work, I have quenched our samples and correlated the critical temperature difference for thermal shock with surface finish. I have estimated the heat transfer coefficient and the depth of the strength controlling cracks on the edge of disk samples. I have also carried out ring -on-ring biaxial strength tests to measure strength degradation of BK7 disks after thermal shock. For the simulation work, I have calculated thermal stress using both fin approximation and finite element analysis of quenched BK7 disks. The effect of cooling fluid has been examined. To examine the effects of size and glass properties, the radiative cooling processes of BK7 and LHG8 plates are also discussed. I have compared the temperature and thermal stress evolution of BK7 and LHG8 glass plates. Crack size evolution and crack growth rate have been computed. I have used three approaches to describe crack growth: the strength approach, the fracture toughness approach and the slow crack growth approach. The effect of linear variation of surrounding temperature and mitigation strategy are discussed. In this study I also examine in detail the effects of crack geometry near an edge. In addition, I have examined the thermal stresses induced by rinsing. Rinsing is a configuration in-between thermal shock and slow heating/cooling. The effects of rinsing parameters have also been investigated, especially as they affect surface tensile stresses.;The main conclusions are the critical temperature drop for catastrophic failure is in the range 124 - 140 °C for surface prepared by lapping with alumina abrasives ranging in size from 5 mum to 40 mum. Samples with finer surface finish have better thermal shock resistance. There is no obvious strength degradation for BK7 glass disks and the retained strength depends on surface finish. Thermal shock tests in silicone oil show there is significant difference for retained strength after soaking in silicone oil which is evidence of chemical effect. The simulation results of radiative cooling process of BK7 and LHG8 glass show that BK7 plate is stronger than LHG8 plate and the maximum thermal stress occurs at the center of long edge. In addition, linear variation of ambient temperature with time is more detrimental than exponential decay. The mitigation strategy shows two conductive layers work best to reduce the maximum thermal stresses. Slow crack growth approach is used to calculate the crack size evolution. Stress intensity factor of three dimensional corner cracks is simulated and arbitrary crack shape is included. My results show that even though initial crack fronts may have different length scales, the stress intensity factor is largest at the shorter length scale. The implication is that an edge crack eventually will grow to a quarter circular crack shape. Piranha rinsing process is simulated and compared with Piranha rinsing. Thermal stress follows temperature profile with delay. According to my simulation results water rinsing is more detrimental than Piranha rinsing.