Phase transitions like condensation, vaporization, melting and freezing are ubiquitous and important in a host of applications across science and technology. A phase transition is initiated by nucleation, the process where fragments of a new phase begin to form in a supersaturated mother phase. This is followed by the growth of these fragments after they reach a critical size, and, finally, the new stable phase is formed by ageing. Understanding phase transitions, especially the nucleation step, has been an active area of research for over a century with approaches ranging from the development of theory, the refinement of experiments and, more recently, direct in silico simulations. The lack of a robust unified theory that can quantify the process of nucleation for all substances, across a broad range of pressures and temperatures, underscores the challenges involved.
Experimental studies on vapor-liquid nucleation have primarily focused on substances like water, straight chain alcohols and long chain alkanes. Given the complexity of these substances, the uncertainties in key physical properties and a lack of knowledge regarding the intermolecular potentials, meaningful comparisons between theory, experiments and simulations is challenging. A primary goal of this research is, therefore, to experimentally investigate the nucleation behavior of “simple” molecules including argon and nitrogen. Following on the work of Sinha (Dissertation, Ohio State University, 2008), the current work extends the range of experimental argon condensation data in a cryogenic supersonic nozzle (SSN) to lower temperatures and higher supersaturations. Based on the experimental measurements and estimated nucleation rates of 1017±1 cm-3s-1 in our SSN, Classical Nucleation Theory (CNT) predicts nucleation rates that are lower by 11-13 orders of magnitude. In contrast, rates predicted by Mean Field Kinetic Nucleation Theory (MKNT), a recent theory developed within the construct of statistical mechanics, are within 1-2 orders of magnitude of the experimental estimates. The experimental approach used for argon is successfully extended to study the condensation of nitrogen, a challenge given that expansions start at ~85 K, about 15 - 20 K lower than the argon experiments. For nitrogen, MKNT rate predictions again agree better with experiments, by ~11 orders of magnitude, than the predictions of CNT. The possible reasons for the success of MKNT are explored by combining the high rate data measured here with the lower rate data measured by Iland et al. (J. Chem. Phys. 127, 154506, 2007; J. Chem Phys., 130, 114508, 2009) in order to estimate the critical cluster properties, including the cluster size, excess internal energy, and excess entropy, for both argon and nitrogen. Although both CNT and MKNT over-predict the critical cluster size, the predictions of the excess internal energy and entropy predicted by MKNT are in significantly better agreement with experimental estimates of these quantities.
In this work, the freezing of supercooled heavy water (D2O) droplets in a supersonic nozzle is studied by applying three in situ position resolved experimental techniques including static pressure measurements, small angle X-ray scattering and Fourier transform infra-red spectroscopy. Combining the information from these three techniques yields the size, phase, composition and number density of the droplets, as well as the flow variables including temperature, density and velocity. Freezing of supercooled D2O droplets in the size range of 3 to 9 nm occurs at droplet temperatures ranging from 222 K to 226 K while the corresponding ice nucleation rates are ~1023 cm-3s-1 assuming the phase transition can occur throughout the volume of the droplet and ~1016 cm-3s-1 assuming nucleation is initiated at the surface. The current D2O ice nucleation data are more consistent with nucleation initiated at the surface of the droplet, but uncertainty in the experiments makes it difficult to definitively state which process dominates. Some of the difficultly may simply be that for the smallest drops the outer 1 nm of the droplet constitutes ~70% of the volume. Using current estimates of the thermophysical properties of the condensed phases of D2O, the theoretical ice nucleation rates reproduce the experimental data trends qualitatively but not quantitatively.
Finally, given the importance of understanding the kinetics of formation of clathrate hydrates, a “proof of concept” study shows that these complex structures form on a microsecond time scale in a supersonic nozzle, when tetrahydrofuran (THF) is the guest molecule.