Solid State Communications, Vol. 93. No. 8, pp. 713- 717, 1995 Elsevier Science Ltd Printed in Great Britain 0038-1098/95 59.50 + .OO zyxwvut oo?&-10!28(94)00754-3 DISCOVERY OF A POTENTIALLY HOMOGENEOUS-NUCLEATION-BASED CRYSTALLIZATION AROUND THE GLASS TRANSITION TEMPERATURE IN SALOL Takaaki Hikima, Minoru Hanaya and Masaharu Oguni* Department of Chemistry, Faculty of Science, Tokyo Institute of Technology, Ookayama, Meguro-ku, Tokyo 152, Japan zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO (Received 21 July 1994; in revisedform 8 September 1994 by C.N.R. Rao) Heat capacities of fragile liquid salol were measured with an adiabatic calorimeter. Two crystallization processes were observed separately in temperature; one in the range 218-225 K just above a glass transition region and the other above 227K. Two crystals formed at 220 and 297 K gave the same X-ray diffraction patterns indicating that they were in the same crystalline phase. It was concluded that the process proceeding around 220 K was probably the homogeneous-nucleation- based crystallization as found out in o-terphenyl, and that the decrease in the interfacial energy between crystal embryo and liquid must have played an important role for the crystallization to be observed macroscopically. Keywords: B. crystal growth, D. heat capacity, D. phase transition. 1. INTRODUCTION BELOW THE equilibrium fusion point Z-r,,, the Gibbs energy of the liquid is higher than that of the crystal. Supercooled liquid is therefore thermo- dynamically metastable and should und.ergo a transition to crystal. The crystallization, being a first-order phase transition, proceeds through two processes of the nucleation of a crystalline phase and the growth of the nucleus to a large crystal [l, 21. The homogeneous nucleation occurs through local struc- tural fluctuations in the liquid and has an appearance as a structured cluster of molecules, and when having become larger than some critical size the cluster named crystal embryo grows to a crystal nucleus. Development of the structured clusters is generally recognized to be more enhanced as the temperature is decreased down to the glass transition temperature Tg. Below Tg, however, the rearrangemeni. motion of the molecule becomes sluggish so that the forma- tion rate of the clusters should be kinetically reduced. The homogeneous nucleation rate would be thus considered to have a maximum around Tg in many cases. The growth process essentially proceeds in a * To whom correspondence should be addressed. molecule-by-molecule fashion to form a large crystal. The maximum rate has been observed rather in the middle between Tg and Tf,, in many systems. The maximum rates of the homogeneous crystal nuclea- tion and of the growth have been thus expected to appear at considerably different temperatures in general [3]. No direct observation of the nucleation process has, however, been reported so far to our knowledge. Information about the nucleation rate has been obtained in the experiments where the supercooled liquid sample was held for a specified period in the nucleation temperature region and then the nuclei formed were counted at the higher tem- peratures in the crystal growth region [4]. The complicated thermal history thereby brings some difficulty in detailed characterization of the nuclea- tion process, and thus the nucleation process has been hardly understood experimentally nor theoretically. We have recently found out a probably homo- geneous-nucleation-based crystallization in o- terphenyl at very low temperatures just above the glass transition region [5]. This phenomenon was interpreted to be based on the remarkable stabiliza- tion of crystal embryos as fine crystal particles, and 713