In 1951, Loebenstein and Deitz described an innovative gas adsorption technique that did not require the use of a vacuum. They adsorbed nitrogen out of a mixture of nitrogen and helium that was passed back and forth over the sample between two burettes by raising and lowering attached mercury columns. Equilibrium was established by noting no further change in pressure with additional cycles. The quantity adsorbed was determined by the pressure decrease at constant volume. Successive data points were acquired by adding more nitrogen at the system. The results obtained by Loebenstein and Deitz agreed with vacuum volumetric measurements on a large variety of samples with a wide range of surface areas. They were also able to establish that the quantities of nitrogen adsorbed were independent of the presence of helium.

Nelson and Eggertsen, in 1958, extended the Loebenstein and Dietz technique by continuously flowing a mixture of helium and nitrogen through the powder bed. They used a hot wire thermal conductivity detector to sense the change in effluent gas composition during adsorption and desorption, when the sample cell was immersed into and removed from the bath, respectively. Fig. 1 illustrates a simplified continuous flow apparatus. Fig. 2 is a schematic of the flow path arrangement using a four-filament thermal conductivity bridge.

Figure 1: simplified continuous flow apparatus

Figure 2: flow path arrangement using a four-filament thermal conductivity bridge (D_A is formed by filaments 2 and 4, D_B = 1 and 3. This type of circuit is known as a Wheatstone bridge)

In Fig. 1, a mixture of adsorptive and carrier gas of known concentration is admitted into the apparatus at 'a'. Valve V1 is used to control the flow rate. The analytical pressure is the partial pressure of the adsorptive component of the mixture. When the system has been purged, the detectors are zeroed by balancing the bridge (see Fig. 3). When the sample cell 'b' is immersed in the coolant, adsorption commences and detector D_B senses the decreased nitrogen concentration. Upon completion of adsorption, D_B again detects the same concentration as D_A and the signal returns to zero. When the coolant is removed, desorption occurs as the sample warms and detector D_B senses the increased nitrogen concentration. Upon completion of desorption, the detectors again sense the same concentration and the signal returns to its initial zero value. Wide tubes 'c' act as ballasts to (i) decrease the linear flow velocity of the gas ensuring its return to ambient temperature prior to entering D_B and (ii) to prevent air being drawn over D_B when the cell is cooled and the gas contracts.

Figure 3: Thermal conductivity bridge electronic circuit

Fig. 4 illustrates the detector signals due to adsorption and subsequent desorption. Figs. 5 and 6 illustrate a parallel flow arrangement which has the advantage of requiring shorter purge times when changing gas composition but is somewhat more wasteful of the mixed gases. The symbols shown in Fig. 5 have the same meaning as those used in Fig. 1.

Figure 4: adsorption and desorption peaks in a continuous flow apparatus

Figure 5: parallel flow circuit

Figure 6: Parallel flow path using a four-filament bridge (D_A is formed by filaments 2 and 4 are, 1 and 3 comprise D_B. Dashed lines are gas flow paths.)