Cross-Flow Filtration and Axial Filtration
KURT A. KRAUS, Chemist
Oak Ridge National Laboratory
Oak Ridge, Tennessee 37830
INTRODUCTION
Through legal and other public pressures ever higher demands are being placed on
effluent controls, and conventional and until now adequate treatment methods are
becoming unsatisfactory. Thus, new effluent standards often require decontaminations to
the parts-per-billion range when only in the recent past, decontaminations to the parts-per-
million range were considered good. Perhaps surprisingly, the chemistry of the treatment
processes is often capable of yielding effluents meeting these demanding conditions.
Satisfactory techniques include precipitation of insoluble compounds, adsorption on organic and inorganic materials, and carrying of traces by selected precipitates ("carriers").
Difficulties for reaching high degrees of decontamination seem to come from inefficiencies
of some large-scale systems. Inefficiencies associated with solid-liquid separations seem
largely to arise from use of settlers. In the laboratory, settling in small vessels and with
long residence times yields clear and well decontaminated supernatants; in the field the
large devices with their unavoidable turbulences rarely yield clear products at economical
production rates. Since settlers are convenient and inexpensive, this apparent failure is
disappointing. However, of course, settlers were not designed to achieve extreme separations efficiencies.
The present paper deals with the principal alternative solid-liquid-separations
technique, filtration, in two relatively novel forms: cross-flow filtration and axial filtration.
In particular, it reviews ORNL work largely carried out by people associated with our
Water Research Program. In cross-flow filtration, the feed is pumped past the filtering
surface and in axial filtration the filter, mounted on a rotor, is moved with respect to the
feed. While large-scale application of the axial filter is still in doubt, we believe that it
permits with little expenditure of time and money, duplication of many of the
hydrodynamic aspects of cross-flow filtration and allows convenient resolution of many
questions regarding filtration and other fine-particle handling problems. Because of this,
and to simplify discussion, we shall in this paper frequently use the term "cross-flow"
filtration to include axial filtration.
While I shall restrict discussion here to cross-flow filtration, I do not wish to imply that
other methods, such as sand filtration, multi-media filtration, or precoat filtration are
necessarily inferior. However, cross-flow filtration has advantages over these other
techniques in specific cases, and broader consideration than it is given now is warranted.
Our interest in the method started with work on salt filtration, variously called
hyperfiltration or reverse osmosis. In it, high tangential flow velocities past the membranes
are necessary to avoid excessive buildup of salt at the membrane-solution interface
(concentration polarization). Salt concentration control by hydrodynamic means led
naturally to the use of high tangential flow velocities to remove fouling agents from the
membranes or at least to retard their deposition (1,2). From this it is only a small step to use
high tangential velocities in filtration systems where filter cake buildup is to be avoided or
retarded even if salt filtration is more an objective.
Several years ago we proposed the name "cross-flow filtration" for this technique
where the feed solution is rapidly pumped past the filtering surface at right angles to product
1059