Abstract
The need for further studies on the Thermal Diffusion effects relative to the separation of Common Gases is demonstrated. Classical theory, as derived by Furry, Onsager and Jones, for Isotopes included a number of assumptions, the validity of which has been examined for binary and ternary mixtures of He, N2, and CO2. The theory has been fully derived to ensure that the significance of each step be considered. Particular emphasis is placed on the investigation of separation and transport phenomena for three-component mixtures. A large thermal diffusion column of plate type was used with a 1" gap. This minimises the edge effects and ensures accurate sampling and temperature measurements. A detailed study of the 'field' of concentration and temperature in the gap, was necessary to check the applicability of the theory. Special sampling apparatus and installation was developed to achieve the accuracy required. The physical parameters, such as viscosity, specific heat, density, heat conductivity, are all temperature dependent and the values of these parameters were determined as functions of temperature within the range of the temperature field. It has been demonstrated that the original equation of the average transport and separation as derived by Furry and Jones and others were not successful in correlating the experimental results either for binary or ternary common gas mixtures. This was established incidental to the main scientific study. The study of 'fields' was especially important as the attempt to correlate the Data by means of classically developed equations were rather unsuccessful. In this respect an initial study was made of the field on the basis of convection and conductance. Nusselt type equations were developed so that the effects of convective disturbances across the gap could be related to concentration and temperature profiles. It was realised from the study as well as was evident from the theories the convective contribution can be preferably considered to be minimal. This is in agreement in fact with the original conception of Furry, Onsager and Jones who neglected them and restricted their development to a study of conductance across the gap. This in no way affects the thermo-gravitational aspect of the mechanism of enhancement of separation. At least three methods were used to determine the contribution of each component in the mixtures to the conductivity and therefore heat flux across the gap. The corrected and measured temperature profiles agreed with all three theoretically predicted profiles of temperatures. In two of these methods, where a restricted gas space (2w) was considered, the agreement was excellent. It was noted, not only were the temperature profiles non-linear, but they gave consistent shapes of curves for many experiments; these shapes are shown in the thesis. There is clear evidence of the effect of the thermo-gravitational effects close to the walls where the curves take on a different shape from the central span. This change in shape is, of course, due mainly to changes in the heat conductivity and also, to a smaller degree, to those of specific heat, density and viscosity. The relationship between temperature and concentration profiles have been derived mathematically. The temperature profiles can be directly derived from the concentration profiles by equations. These derived temperature profiles compared exactly with the measured profiles. Initial work on the reverse procedure has been commenced and the equations, considered as simultaneous, gave correct values of concentration. The problem, however, is not complete yet due to the extensive data and computation required. It is hoped eventually to be able to obtain concentration distribution from boundary conditions only (i.e. temperatures of hot and cold walls). Theory has been developed which examines phehomenologically the separation and transport in a large thermo-gravitational column for the cases of 'total reflux' and for small mass flow of gas through the column. The theories avoid the inaccuracies inherent in too many simplifications and assumptions which in other circumstances such as separation of isotopes are justifiable. The present approach is capable of extensive development for the prediction of separation in multicomponent gas mixtures.