In experimental investigations of the transient behavior of batch thermogravitational thermal diffusion columns it is expedient to obtain more than one sample from the column during a single run. Such sampling modifies the true batch nature of the column and thereby influences the experimental measurements to some degree. A combined theoretical and experimental investigation was carried out to determine the extent of the influence of sampling rate and thereby provide a means of predetermining a suitable sampling rate.

The theoretical analysis was made by approximating the intermittent sampling as a continuous flow through the column. The transport equation of Furry, Jones, and Onsager was applied in the analysis, and an analytical solution was obtained for small separations of equicomposition binary solutions. The series solution is presented in the form of a graph. Experimental data were obtained in two thermogravitational columns with different plate spacings to test the theory. The effect of sampling rate was investigated, and theory and experiments were found to be in good agreement. It is concluded that the theory is entirely adequate to permit the prediction of sampling rates which will yield the maximum number of samples without disturbing the true batch behavior of the column.