Final Report

OAK B135 The formation of metastable crystalline phases in lithium disilicate glass has been a subject of controversy for decades. Here, one aspect of this problem relating to the stability of these non-equilibrium phases when glasses are heated for extended time periods in the nucleation regime is addressed. The results of a systematic experimental investigation on the persistence of metastable phases and the factors that may influence the appearance of such phases, e.g., water content, impurities, glass composition, and glass preparation procedure are presented. Growth rates of lithium disilicate crystals in lithium disilicate glass are measured as a function water concentration in the glass and of temperature in the deeply undercooled regime. The growth rate data obtained in this work are combined with data reported in the literature and used to assess the applicability of standard models of crystal growth for the description of experimental results over a very broad temperature range. The reduced growth rate versus undercooling graph is found to consist of three regimes. For undercoolings less than 140°C, the reduced growth rate curve is suggestive of either 2-D surface nucleation or screw dislocation growth. For undercoolings greater than 400°C, the reduced growth rate plot suggests the operative crystal growth mechanism is 2-D surface nucleation, but detailed calculations cast doubt upon this conclusion. In the intermediate undercooling range, there appears to be some sort of transitional behavior for which none of the standard models appear to be applicable. Further, it is observed that small differences in the viscosity data employed can produce enormous differences in the predicted growth rates at larger undercoolings. Results of the kinetic analyses conducted herein seem to indicate that the nature of the kinetic rate coefficient used in the standard growth models may be incorrect. Nucleation rates of sodium metasilicate crystals in a sodium silicate glass of composition 43Na2O-57SiO2 (mol%) are investigated using the development technique. The results of this study are compared with the nucleation rate results recently obtained for this composition using a novel DTA method. The two techniques are found to agree within experimental error