Determining Kinetic Parameters for Isothermal Crystallization of Glasses C. S. Ray w Marshall Space Flight Center, NASA, Huntsville, Alabama 35812 T. Zhang, S. T. Reis, and R. K. Brow Materials Science & Engineering Department, University of Missouri—Rolla, Rolla, Missouri 65409 Non-isothermal crystallization techniques are frequently used to determine the kinetic parameters for crystallization in glasses. The techniques are experimentally simple and quick, compared with isothermal techniques. However, the analytical models used for non-isothermal data analysis, that were derived from models originally developed for describing isothermal transformation kinetics, are fundamentally flawed. The present paper describes a technique for determining the kinetic parameters for isother- mal crystallization in glasses, which eliminates most of the com- mon problems that generally make the study of isothermal crystallization laborious and time consuming. In this technique, the volume fraction of a glass that is crystallized as a function of time during an isothermal hold was determined in a separate experiment using differential thermal analysis. The activation energy (345710 kJ/mole) and Avrami parameter (0.8970.09) for crystallization of Li 2 O . 2SiO 2 glass determined by the pres- ent technique are consistent with the similar values reported in the literature. I. Introduction T HE process of overall crystallization in glasses, which occurs due to the combined effects of nucleation and crystal growth, is best described by the Johnson–Mehl–Avrami–Ko- lmogorov (JMAK) model. 1–5 The JMAK model assumes a transformation occurring strictly under conditions of isother- mal heat treatment, and yields the kinetic parameters for crys- tallization that describe the complete transformation process. Most notable among these parameters are the activation energy for crystallization, E, and the dimensionality of crystal growth or the Avrami parameter, n. E and n are determined from ex- periments in which the volume fraction (x) that is crystallized in a melt is measured as a function of time (t) when the melt is held isothermally at different temperatures, T. Although, several techniques such as X-ray diffraction (XRD) analysis and optic- al microscopy are available 6–9 to measure the extent of x, the thermal analysis techniques, 10–13 including differential thermal analysis (DTA) and differential scanning calorimetry (DSC), are considered the most suitable and convenient means. For measuring x using DTA or DSC, the temperature of the melt (for a glass composition under investigation) is quickly de- creased to a pre-determined value (say, T) and held there until crystallization is complete. The crystallization event is displayed on a temperature–time thermogram as an exothermic peak such as the one shown schematically in Fig. 1(a). The fraction crys- tallized (X t ) at any time, t, is determined from the ratio of the area at time t (A t ) to the total area (A 0 ) of the exothermic peak, X t 5 A t /A 0 , Fig. 1(a). It is often difficult to experimentally obtain an idealized crys- tallization exotherm as shown in Fig. 1(a). Obtaining such an exotherm depends notably upon two factors: (1) the choice of a suitable isothermal hold temperature that is, generally, un- known for an unknown system, and that depends upon the characteristic temperature-time-transformation (TTT) diagram of the melt, and (2) the efficiency and capability to quickly and precisely attain the proper hold temperature that, in most cases, lies within a very narrow temperature range. As a result, one often ends up obtaining from these experiments one of the two types of curves shown in Figs. 1(b) and (c). Figure 1(b) is typical of thermal profiles where the temperature T does not fall within the temperature range for crystallization of the melt. Upon quenching the melt, the temperature shows several oscil- lations before being stable at T. Figure 1(c) typically depicts the simultaneous occurrence of crystallization and temperature os- cillations, making it difficult to identify the crystallization event and, hence, to accurately determine the time taken for a particular volume fraction to crystallize. Determining the kinetic parameters for crystallization by con- ventional isothermal heating techniques is tedious, time-con- Fig. 1. Schematic of an expected temperature-time thermal profile on cooling and isothermally heating (at T) a hypothetical melt with melting temperature T m ; (a) ideal/desired thermal profile for crystallization kinetic studies, (b) and (c) thermal profiles that are commonly observed. E. Zanotto—contributing editor w Author to whom correspondence should be addressed. e-mail: Chandra.S.Ray@nasa. gov Manuscript No. 22172. Received August 25, 2006; approved November 4, 2006. Presented at the 8th International Symposium on Crystallization in Glasses and Liquids, Jackson Hole, USA. J ournal J. Am. Ceram. Soc., 90 [3] 769–773 (2007) DOI: 10.1111/j.1551-2916.2006.01478.x r 2007 The American Ceramic Society 769