A convenient method for the investigation of gas transport in alkali halides is offered by the possibility to load single crystals homogeneously with radioactive rare gas ; (1) in the case of potassium chloride with Ar-39 through the reaction K(n, p) Ar. The radiation with neutrons will, however, also produce defects of various kinds in the lattice. The experimental results readily show, that the gas atoms interact with radiation-defects, (2), (3) in that the coefficient of diffusion D can be put on the form D ~ 1/CT, CT being the concentration defects. This, the so called trapping-diffusion-model successfully could explain as a trapping of the gas atoms a the defects (4). Some further work, undertaken to account for certain deviations from expected isothermal degassing curves, did not succeed in bringing these anomalities in consistency with the predictions based on the original assumption mentioned above. The contradictions led to the suggestion that the defects are not stable, but anneal inhomogeneously during the experiment, which in effect means that a time-and-position-depending coefficient of diffusion will have to be used in the formal description. Due to the fact that the gas produced in potassium chloride crystals is almost completely bound at the defects, there is a method at hand for attaining information on the behaviour of the defects, if the gasprofile, or some function of it, can be determined at low enough temperatures. Thus if the irradiated specimen is heated at a constant temperature until a certain amount of gas has left the crystal, the resulting gasprofile will hold information on the defect profile. Any difference from a gas-profile produced in a system with a constant or only time-depending coefficient of diffusion, will mean that the defect concentration in this experiment has changed time-and-position dependently, which then also pertains the coefficient. For the purpose of measuring gas profiles, a small precision-grinding machine was constructed, which permits the specimen to be grinded along one axis in an evacuated vessel. The radioactive gas, which is set free quantitatively during the grinding process, is measured with a proportional counter at appropriate intervals of grinding. Subsequently the change in length of the specimen along the grinding-axes is determined. The experimental result will thus be available in the form of an integrated activity versus position curve. It should be noted, that the profiles and consequently the coefficient of diffusion, can be calculated explicitly only when diffusion profiles in the specimen of interest have an unique position space dependence, e. g. in a spherical but not a cubic specimen. In the case of a cube, a qualitative evaluation is attainable if the experimental curve is compared to the one calculated from the solutions to Fick's-second-law for a specimen of the same form and the same degree of degassing, i. e. the curve that would result if the defects were perfectly stable or would anneal homogeneously. From experiment performed on cubic specimens of potassium chloride and evaluated as outlined above, it seems fair to conclude, that at least a part of the defects anneal at the surface.
Read full abstract