The homogeneous nucleation of nonane

Abstract

INTRODUCTION The first complete theory of homogeneous nucleation, known as the classical theory, was developed by Volmer and Weber [i], Farkas, and Becker and Doering. Classical theory treats a cluster of molecules as a droplet whose properties (e.g., surface tension, density) are the same as those of the liquid. The free energy of the droplet is thus approximated by the free energy of the same number of molecules in the liquid plus the droplet's surface free energy. This assumption has been very controversial, mostly because it is not known to what extent the surface term includes implicit translational and rotational free energy contributions. Several authors have argued that a cluster is similar to a large molecule, so its free energy must also contain contributions from its translation and rotation. Frenkel and Kuhrt considered this effect, and for differing reasons found it to be negligible. In 1962 Lothe and Pound [2] argued that it is not negligible. In fact, nucleation rates for n-nonane at 330 K calculated using their theory are 1020 larger than the rates obtained using classical theory, thus decreasing the critical supersaturation by 30%. Reiss et al.[3] disagreed with the Lothe-Pound arguments and presented a model (referred to as the Reiss-Katz-Cohen theory) which assumes that the free energy of a drop with fixed boundaries is equal to the free energy of the same number of molecules in the liquid phase plus a surface free energy. Statistical mechanics is then used to calculate the difference in free energy of a drop with fixed boundaries and a cluster. This theory results in rates of nucleation which are 103-105 larger than those predicted by classical theory, which decreases the predicted critical supersaturation by 8%-10%. Classical theory has been accepted as correct because of its excellent agreement with the results of critical supersaturation measurements by Volmer and Flood [4] using a piston cloud chamber on water and several organic compounds, and because of more recent measurements by others. However, measurements of critical supersaturations, even if measured over a wide range of temperatures, are not a complete evaluation of nucleation theory they do not test the dependence of the rate of nucleation on supersaturation. Nucleation theory can be evaluated in a very sensitive fashion when the temperature and supersaturation dependences of the rate of nucleation are decoupled, e.g., by measuring the rates as a function of supersaturation on various isotherms. Several groups of investigators have recently developed reliable ways to make such measurements. Schmitt et al.[5] measured nucleation rates for n-nonane using a fast-expansion piston cloud chamber. Their data covered nucleation rates from 102 to 105 nuclei/cm3/sec, and a temperature range of 217 to 266 K. Wagner and Strey [6] measured nucleation rates of n-nonane using a two-piston expansion cloud chamber. Their data 6 covered nucleation rates from 10 to 1010 nuclei/cm3/sec, and a temperature range of 203 to 238 K. We have measured nucleation rates using the upward thermal diffusion cloud chamber [7,8] and here report our results. Our n-nonane data covers nuelea1 3 tion rates from 10 -4 to I0 nuclei/cm /sec, and a temperature range of 233 to 315 K.