On the location of the maximum homogeneous crystal nucleation temperature

Detailed considerations are given to the location of the temperature of maximum homogeneous nucleation as predicted by classical nucleation theory. It is shown quite generally that this maximum temperature, T-asterisk, must occur above the Kauzmann temperature and that the T-asterisk is such that T-asterisk is greater than T(m)/3, where T(m) is the melting temperature. Also, it is demonstrated tha T-asterisk may be considered to be approximately dependent upon two parameters: gamma, the ratio of the difference in specific heat between the crystal and liquid divided by the entropy of fusion, and E, a reduced activation energy for viscous flow. The variation of T-asterisk with these parameters is described. The relationship of the relative location of T-asterisk to the glass transition temperature, is discussed too. This discussion is couched within the framework of the strong and fragile liquid notion introduced by Angell (1981) and coworkers. Finally, the question of the ultimate limits to the undercooling of liquid metals is considered and its relationhsip to computations of the maximum nucleation temperature in such systems.