Filtration, axial

AV reciprocating tachycardia, 5 108 Axial dispersion coefficient, 10 762 Axial dispersion/mixing, 10 762-763 in adsorption columns, 2 604 in bubble tray absorbers, 2 88-89 chromatographic adsorption, 2 610 in packed column absorbers, 2 61-65 Axial dissolved oxygen profiles, 25 707-708 Axial filtration, 22 385-386 Axial-flow angular-momentum flowmeter, 22 672-673... 

Axial filtration. In most of our bench-scale filtrations cross flow was effected by use of axial filters (10). In this configuration (Figure 4), a membrane is wrapped around a rotor, which is spun in a chamber, into which feed is introduced under pressure. The rotor is perforated, and passages are provided for filtrate (e.g., by an intervening screen) from the membrane to these holes. Filtrate exits through the axis. Rotation speeds providing velocities of up to about 15 ft/sec at the membrane-feed interface can be attained in available equipment. 

Figure 7. Plugging of 1.2-ixm Millipore filter by polymer solutions under 15 psig pressure before and after axial filtration through 1.2-nm Acropor filter (20 psig, 2000 rpm or 11 ft/s). Viscosities of all solutions are 20 to 25 cps.

Fluxes in axial filtration of diluted glucan fermenter broth through 5-iim Nuclepore filters (2000 rpm 11 ft/s)... 

Figures 11 and 12 compare the viscosities of the axial filtration effluents and feeds for the screened and unscreened broths. Here there is a clear advantage of the screening step. There is little decrease of viscosity between feed and filtrate for the screened material, but about a 25% decrease for the unscreened. There are sometimes decreases in viscosity Incurred from diatomaceous earth filtration.

The results in Figures 8-12 are somewhat more favorable than results in reference 6. One might infer that the reason is the use here of Nuclepore filters, which have cylindrical pores normal to the surface in a narrow size range, rather than the Acropor membranes in reference 6, whose pores are more tortuous. We do not have sufficient information for a definitive conclusion on this point. However, a later comparison of fluxes in axial filtration by the two types (Figure 13) does not indicate any great difference. Comparisons of viscosity of feeds and filtrate and of plugging rates of filtrates also did not support a significant difference between Nuclepore and Acropor. 

Figure 9. Unscreened feed for axial filtration plugging of 1.2-iim Nuclepore filters by filtrate from 5-jim Nuclepore filter mounted on axial filter (plugging test pressure...

Figure 11. Unscreened feed comparison of viscosities of feed and filtrates from axial filtration (25 °C axial filtration through S-ftm Nuclepore filter)...

Figure 13. Axial filtration at a rotational speed of 2,000 rpm f-screened fermenter broth at 7.5 1 dilution and 7 psi...

Further optimization of the polishing step is clearly necessary. The mediocre fluxes in cross-flow and axial filtration suggest that systems allowing backwash at frequent intervals may be necessary. Many commercial ultrafiltration systems, including the one used here, have this capability. 

The so-called axial filter , developed in the Oak Ridge National Laboratory, is remarkably similar to Morton s and Kaspar s dynamic filter in that the filter leaf is in the tubular form and the outer shell is also cylindrical. An ultra-filtration module based on this principle has also been described more recently. Unlike the European dynamic filters referred to in the previous paragraph, however, this filter is not suitable for scale-up because it poorly utilizes the available space. The Escher-Wyss pressure filter described (identical copy of this paper also appeared subsequently in Filtration and Separation ) takes the idea of axial filtration a step further in... 

---