Date of Award

12-2014

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Legacy Department

Mechanical Engineering

Advisor

Joseph, Paul F

Committee Member

Biggers, Sherril

Committee Member

Thompson, Lonny

Committee Member

Blouin, Vincent

Abstract

The Precision Glass Molding (PGM) process was established as an economical and sustainable option for the production of aspherical glass optics to satisfy the increased industrial demand. Applications of precision molded aspherical lenses range from consumer electronics products such as cell phone cameras to defense and medical systems. An aspherical lens can eliminate the spherical and optical aberrations as compared to a spherical lens thus making the lens system more compact and lighter. In spite of being a clean and environmentally friendly process, the lens molding operation suffers from a few drawbacks such as lens profile deviation, stress birefringence/refractive index drop and lens cracking. Prior research has identified a lack in accurate and reliable thermo-mechanical characterization of optical glasses as an obstacle to the application of computational mechanics to resolve these issues. The work presented in this dissertation addresses the importance of a precise determination of the thermo-mechanical material property inputs of optical glass for an accurate prediction of the state of stress during the complex thermo-mechanical loading of a glass preform during the Precision Glass Molding (PGM) process. In addition to an accurate prediction of the residual stress state in a lens, birefringence and fracture were also considered as these are direct consequences of stress. Due to the complexity of glass behavior in the relatively large temperature range where the material behavior transitions from that of an elastic solid to a viscous fluid, it is essential to characterize accurately the time and temperature dependence of the stress relaxation behavior. After understanding the weaknesses in existing stress relaxation characterizations, a set of careful experiments was designed that utilized Parallel Plate Viscometer (PPV) to perform the cylinder compression test on a glass sample. It was determined that the uniaxial compression of a cylindrical sample at an uniform temperature, with a known friction condition at the interface, yields a high quality creep data that was used to determine accurate viscosity and viscoelastic constants of two moldable glasses - L-BAL35 and NBK-7 glass at the given temperature. Comparison of the computational solutions with closed form approximations used in an ASTM standard, revealed deficiencies at viscosity near and above 108 Pa*s due to specimen bulging and interface slip, and led to the development of an approximate expression for a reasonable estimate of viscosity above 108 Pa*s for the full range of interface friction behavior. This study highlighted the importance of an accurate characterization of the stress relaxation function of a moldable glass which enabled the numerical examination of the effect of different levels of modeling detail of the relaxation function on the lens molding simulations. The choice of the material model and the level of detail required in performing the creep and relaxation experiments, is dependent on the problem being solved. The use of simplest viscoelastic stress relaxation function with a single exponential relaxation time that lacks much of the transient effects present in a full viscoelastic relaxation, showed minimal effect on the profile deviation of a lens but leads to an over-estimate of residual stress for the two lens shapes studied. A similar effect was observed on the stress birefringence of a lens after molding. Using numerical experiments, residual stresses were shown to be sensitive to the lower temperature limit of the viscoelastic assumption (TL). The fracture assessment inside a molded lens was made for both radial and circumferential crack configurations. The stress state for the two configurations revealed that the radial crack orientation was more prone to failure among the two. The full viscoelastic relaxation assumption also led to higher crack tip opening displacement (CTOD) values than the simplified relaxation assumption

Share

COinS