No abstract

Finite-difference methods were used to compute hydrodynamic instability due to natural convection in an enclosed horizontal rectangular region heated from below. Critical Rayleigh numbers were determined for a series of Prandtl numbers and length-to-height ratios. For Prandtl numbers greater than unity excellent agreement was obtained between these calculations and the values predicted by Kurzweg on the basis of a linearized theory. However, for Prandtl numbers less than unity the critical Rayleigh numbers exhibited a dependence on N P~, which was not predicted by the linearized theory. For Rayleigh numbers greater than the critical, complete temperature and velocity fields were determined.The calculations assum,ed that the fluid motion i s two dimensional. Experiments have indicated that the flow may be two or three dimensional depending on minor perturbations in the boundary conditions.Although a number of metastable two-dimensional circulations are possible for symmetrical initial conditions, the calculation always converged to o single, unique solution for any asymmetric initial condition.

shown for absorption conditions using an ethanol-water system that the surface tension will fall from 72 dyn/cm to approximately 49 dyn/cm, a change of approximately 33%.From the photographic study, the droplet falling back toward the interface either (a) was assimilated by the bulk fluid, (b) bounced back into the gas phase one or more times before coalescing, or (c) was assimilated but produced a secondary droplet which reentered the gaseous space. Data concerning this behavior are presented in Table I, and photographs illustrating (b) and (c) are shown in Figures 2, 3, and 4.It is suggested that this anomalous behavior is equivalent to the heat-transfer analog, the Leidenfrost effect. Thus the coalescence time is governed by the action of the vapor being expelled from between the two liquid interfaces. At the same time the droplet approaching the interface deforms the underlying liquid which on restoration imparts an added force causing the droplet to reenter the gas phase.

Methods are developed and illustrated for determining the amount of nonsatisfaction of the overall area (integrated Gibbs‐Duhem equation) test for vapor‐liquid equilibrium data that is to be expected on the basis of random experimental uncertainties. Data which satisfy the overall test within these limits, as well as the local area test, are said to be consistent within the bounds established by their experimental uncertainties, while data which do not are said to show significant systematic error and hence are termed inconsistent.

SynopsisAn atmospheric-pressure, vapor-phase polymerization technique was used to deposit thin (1-10 pm) poly(p-phenyleneterephthalamide) coatings onto polyester film and other sheet substrates. A minimum deposition temperature of 17OOC was found to be critical. When substrate temperatures were below 17OoC, coatings were loosely adhered, powdery, and of low molecular weight; above 17OoC, coherent, well-adhered, high molecular weight coatings were obtained. These vapor-deposited coatings exhibited exceptionally good oxygen barrier properties and were found to consist of fused 0.1-pm particulates. Scanning electron microscopy (SEM) of the top surfaces of these coatings clearly revealed their particulate origin. Similar evidence was also obtiined from SEM examination of fracture surfaces and transmission electron microscopy of microtomed sections. The coalescence of these coatings was demonstrated by SEM of plasma-etched surfaces and oxygen permeability information. The uniplanar orientation of the polymer crystals in these multiparticulate coatings was strong evidence for the epitaxial growth of the crystals.