HOMOGENEOUS NUCLEATION OF SUPERCOOLED WATER DROPS

Abstract

The experimentally observed crystallization of supercooled water near ?40C is examined in terms of the theory of homogeneous nucleation. The thermodynamic and molecular-kinetic nature, of the nucleation process is outlined to show why supercooling in natural clouds can occur so frequently. Past efforts to explain the ?40C transition are examined critically, and are found to contain a number of significant errors. Because the theoretical nucleation rates are extremely sensitive to the numerical value of the specific surface free energy of a water-ice interface, particular attention is devoted to the refinement of previous estimates of this parameter. It is shown that both Krastanow's and Mason's estimates were inaccurate, and that in the latter's approach, neglect of the distortion energy of the surface layer of ice led to a marked under-estimate of the nucleation efficiency which was concealed by the effects of several counteracting errors. Difficulties lying in the way of a direct calculation of the distortion energy for ice are examined and found to be very serious. A crude correction for distortion effects leads to a theoretically predicted temperature of ?26C for the threshold of spontaneous nucleation of drops of cloud-particle size. It is concluded that although this result lies well above the experimentally observed range of transition temperatures, it is close enough to that range (considering the inherent difficulty of assessing the effect of distortion), to strengthen the belief that the ?40C transition is due to homogeneous nucleation.