WASTEWATER TREATMENT BY MEMBRANES:
WHAT CAN BE DONE ABOUT FOULING?
Georges Belfort, Associate Professor
Benny Marx, Assistant Professor
School of Applied Science and Technology
Hebrew University of Jerusalem
Jerusalem, Israel
INTRODUCTION
The major limiting factor in using pressure-driven membrane processes and in particular hyperfiltration (reverse osmosis) for industrial, agricultural and municipal applications is membrane fouling.
Here we report our continued efforts to understand the mechanism of submicron
colloidal fouling and its effect on performance. We present the results of a new me-
\.h.o<\-fixed and dynamic precoat protective cover method-to reduce the degradation
effect of fouling on permeation flux. In addition, we have modified the standard filtration theory for use in gel-membrane transport.
The results reported here are a direct continuation of our first publication wherein
we used the artifical fouling method to study the influence of membrane fouling and
compaction of hyperfiltration [1]. For additional details on the significance of membrane fouling and its role visa vie compaction and hydrolysis, the reader is referred to
recent reviews by Belfort [2, 31.
METHOD OF ANALYSIS
Membrane-gel transport
For simultaneous comparison of performance of membranes, plotting flux or permeation coefficient (flux per unit pressure) versus time, as is usually done in the reverse
osmosis industry, is very misleading since the accumulated permeation volume, and hence
the extent of fouling, is different for different flux membranes. It appears thus preferable to use a method of analysis where accumulated flux or accumulated permeation
coefficient is plotted versus accumulated permeation volume, V. This normalizes all
membranes and allows their performance to be compared per unit permeated volume.
Three problems exist in applying the standard filtration approach to hyperfiltration [4].
The first has to do with the fact that it does not include the usual cross-flow hydrodynamics of the reverse osmosis system. The second problem is related specifically to our
laminar slit flow experiments where we don't have a constant suspended solids feed
concentration (w ¥= constant) but an exponentially decreasing one, i.e.
w(0) = w(0) exp   {-k0 } (1)
The third problem concerns the pressure differential AP. In standard cake filtration
the AP     across the cloth is assumed to be small with respect to that across the cake
APc. Here we have the reverse, with the added complexity of the effect of the salt on
the driving force i.e. the "reverse" osmotic pressure Air. Thus, AP across the membrane
377