Refinement of a real-time self-diffusivity measurement methodology

Abstract

We have developed and extensively tested a technique for the measurement of mass diffusivities in liquids over wide temperature ranges. Using this real-time methodology diffnsivities at three different temperatures are obtained per sample. In addition, tens to hundreds of individual measurements are made at each temperature. Thus, this methodology is very amenable for low-gravity experiments where experiment time can be limited. For these diffusion studies, a radiotracer, initially located at one end of the cylindrical diffusion sample, is used as the diffusant. The sample is positioned in a concentric isothermal radiation shield with collimation bores located at defined positions along its axis. The intensity of the radiation emitted through the collimators is measured vs. time with solid state detectors and associated energy discrimination electronics. Diffusivities are calculated from the signal difference between pairs of collimation bores. Because of the mathematical algorithm used in this technique the diffusivity at several temperatures can be measured utilizing a single sample. Selfdiffusivities obtained with “4mIn /In in space and on Earth illustrate the high precision obtainable with this technique. The “4”‘In /In space data were close to the ground results, however, the data scatter was much less. By employing a tracer that emits photons of different energy, and thus, different self-absorption, transport in the bulk of the sample can be distinguished from that in the proximity of the wall. No measurable differences in the diffusivity values were found. In support of this work 2-D numerical modeling of the * Member, t Corresponding author. Copyright Q 2000 by R. Michael Banish Published by the American Institute of Aeronautics and Astronautics, Inc. with permission effect of various blockages on the concentration profile and the resulting apparent diffusivity were conducted. For the methodology that we use very little effect was seen except in the case of extreme blockages. These results have been experimentally verified (using the above measurement location methodology) with “4”‘In /In diffusion studies. The resulting diffnsivity of samples run with blockages of greater than 50% between the radiotracer and host sections and “voids” of greater than 15% were essentially the same as normally run samples. Introduction Diffusivities obtained in liquids at normal gravity are prone to be contaminated by uncontrollable convection. As emphasized for liquid diffusivity measurements by Verhoeven’, any horizontal component of a density gradient results in convection without a threshold. Simple scaling arguments illustrate the difficulty of obtaining purely diffusive transport in liquids. In a system of diffusivity lo‘cm’/sec and a typical diffusion distance of 1 cm, the characteristic diffusion velocity is of order 10” cm/s. Hence, if true diffusion is to be observed, convective flow velocities normal to the concentration gradient must be of order lo-’ cm/s or less. Thus, in liquids, the attainment of diffusion-dominated transport over macroscopic distances at normal gravity is obviously not a simple task. Numerical modeling efforts in our group and others has shown that in liquid metals, with their typical viscosities of lo-’ poise, temperature nonuniformities of a few hundredths of a degree are sufficient to generate convective contributions equal to the diffusive flux”. Thus, even in systems with