A Statistical Understanding of Nucleation
A. Izmailov, A.S. Myerson, and S. Arnold
J. Cryst. Growth 196, 234-242 (1999)

In order to study a stochastic phenomenon such as nucleation it is necessary to collect a large enough set of nucleation data to obtain nucleation statistics. This is done by performing nucleation experiments with the same solution under exactly the same conditions many times N (N - 150-300). Such an experiment based on simultaneous levitation of N-150-300 identical microdroplets (1-20Ám in diameter) of supersaturated solutions in a solvent atmosphere, is possible by employing the linear quadrupole electrodynamic levitator trap (LQELT). The LQELT is supplemented with a special optical system which is based on scattering of monochromatic polarized light. This will enable fast observation of nucleation and, thus, induction times in each of the levitated droplets. The N different induction times, counted from the moment to at which supersaturation is established are recorded. This data provides nucleation statistics (induction time statistics). The numerical and analytical studies of nucleation statistics and paraneters provide an insight into statistical properties of the underlying nucleation phenonmenon.