Crystal Nucleation in Submicron Droplets of Pure Elements
- PDF / 1,794,467 Bytes
- 6 Pages / 414.72 x 648 pts Page_size
- 59 Downloads / 214 Views
LOUIS M. HOLZMAN*, THOMAS F. KELLY*,**, AND W.N.G. HITCHON*,***
*Materials Science Program, **Materials Science and Engr., ***Electrical and Computer Engr., University of Wisconsin, Madison, WI 53706
ABSTRACT
Liquid-to-crystal nucleation has been studied extensively through droplet experiments to locate examples of homogeneous nucleation. However, prior to this work very few examples have been found, which implies that the experiments have not been able to isolate heterogeneous In this research, nucleants in a small percentage of the droplets as is required. electrohydrodynamic atomization (EHD) is used to produce sub-micron droplets of pure elements that are largely free of heterogeneous nucleants. Diffraction patterns of individual EHD-produced droplets are viewed to determine the fraction of crystalline droplets produced as a function of droplet radius. These results are compared to theories for surface and volume heterogeneous nucleation and for homophase nucleation. It is found that Si and Ge nucleate through either homogeneous nucleation or nucleation by homophase impurities. Nucleation results for vanadium and iron were not conclusive.
INTRODUCTION
Previous experimental research on nucleation has shown that it is very difficult to produce homogeneous nucleation. Turnbull [1, 2] established the droplet experiment as a fundamental method of studying homogeneous nucleation. The material to be studied would be broken into a large number of small droplets in order to isolate the heterogeneous nucleants that were present in a small fraction of the droplets. Only homogeneous nucleation would remain as a possible mechanism to cause the crystallization of the droplets. Turnbull studied the crystallization of micron-sized droplets suspended in a variety of solutions. He found that those in mercury benzoate or mercury laurate solutions exhibited nucleation dependent on the droplet volume [2], which is consistent with homogeneous nucleation. Miyazawa and Pound found a volume dependence for nucleation of gallium from the melt [3]. However, experiments by Perepezko, et al. [4] demonstrated that greater supercoolings than those obtained in the research above were possible, showing that conditions allowing homogeneous nucleation had probably not been met. In the experiments discussed above, it is not possible to differentiate qualitatively between volume heterogeneous and homogeneous nucleation. Homogeneous nucleation depends on the droplet cooling rate as well as the volume of the droplet, but the cooling rate of droplets of all sizes was the same. However, if the cooling rate is dependent upon droplet radius, homogeneous nucleation will result in an R 3*Rn dependence, where n represents the dependence of cooling time on radius. In such cases, homogeneous nucleation and volume 233 Mat. Res. Soc. Symp. Proc. Vol. 321. • 1994 Materials Research Society
heterogeneous nucleation can be qualitatively differentiated. Drehman and Turnbull [5] studied a case in which the cooling time was proportional to R 1.6, so homogeneou
Data Loading...
