The Effect of Pressure Modulation on the Flow of Gas through a Solid Membrane: Permeation and Diffusion of Hydrogen thro
- PDF / 812,380 Bytes
- 10 Pages / 594 x 774 pts Page_size
- 95 Downloads / 161 Views
I.
INTRODUCTION
THE diffusion of hydrogen isotopes
in metals and alloys has been widely studied, but reviews, such as that by Volkl and Alefeld,~ show notable anomalies. Data on some materials, nickel in particular, are consistent; data on other materials, notably iron and its alloys, are dispersed. Measurements involving the macroscopic flow of diffusant across a surface tend to be seen as causing the problem. Thus, Volkl and Alefeld in looking for best values to describe the diffusion of hydrogen in niobium, tantalum, and vanadium accept only data from surface independent techniques: Gorsky effect, Mossbauer effect, QNS, relaxation of resistivity. This approach may give a good estimate of rates of atomic motion, but it is unhelpful for a number of technological problems, such as the design of hydrogen storage cells or of containment vessels for fusion reactor fuels and reaction products, since for these it is the macroscopic flow that is of consequence. There remains, therefore, a need for experimental techniques which will identify the processes which influence diffusant flow. Many possibilities have been suggested to explain the scatter of data on diffusion. 2 They include: trapping at impurities, at point, line, or surface defects; self-trapping in di-interstitials; micropores or macropores; slow surface reaction rates for adsorption, dissociation, and solution on clean surfaces, or impervious surface layers on unclean surfaces. None of these suggestions is open to useful quantitative verification in the course of current macroscopic experiments, even at the general level of distinguishing surface effects from those in the volume of the material. Most macroscopic measurements on hydrogen diffusion come from time-lag experiments. 3,4 In these, two vacuum chambers are separated by an experimental membrane. Diffusant gas, admitted rapidly to one chamber, is detected in the second after a delay which is the subject of experimental measurement: in a complementary form of experiment, gas
D.L. CUMMINGS, Research Fellow, and D.A. BLACKBURN, Director, are with the Oxford Research Unit, The Open University, Oxford, England. R.L. REUBEN, formerly with The Open University, Oxford, United Kingdom, is now Lecturer, School of Mechanical and Offshore Engineering, Robert Gordon's Institute of Technology, Schoolhill, Aberdeen, AB9 1FR, Scotland. Manuscript submitted July 28, 1983. METALLURGICALTRANSACTIONSA
is removed rapidly from one chamber and the delay preceding the consequential drop in the second is measured. Delays are interpreted as dependent on the diffusion coefficient and the membrane thickness alone. Delays may in part be due to trapping or to slow surface reactions, but this is difficult to detect since such processes leave no strong and characteristic signature on the output trace. For this reason a detailed fit is rarely sought. The paper by Shah e t a l . 5 illustrates the use of a generalized time-lag method when phase boundary processes are rate limiting. Against this background an interesting experimenta
Data Loading...
