World Water January/February 2010
Membrane Technology 33
and Clinical Professor Mark Clark in Civil and
Environmental Engineering at Northwestern
determined that there was no correlation
between membrane fouling and dissolved
organic carbon (DOC). They did suggest that
fouling resulted from particles in the size range
of 3 to 20 nm (0.003 to 0.020 microns) for LP
membranes. Their experiments determined
that 80 percent of non-scaling fouling resulted
from colloidal matter. They also found that
membrane fouling decreased by not filtering
the coagulated water.
The American Water Works Association
(AWWA) noted in their literature review that the
steady state condition of a cake on LP
membranes improved the permeate flux. Studies
carried out on coated low-fouling membranes
indicated that these membranes did not exhibit
better operating characteristics. A study
undertaken by Professor S. Wright at the
University of Michigan showed (using a loss
ignition test) that 60 to 80 percent of the foulant
is organic in nature. It is because of data like this
that one assumes the fouling is organic and that
biofouling is the culprit. As the AWWA has
pointed out, a significant gap exists in fouling
research in distinguishing between active
biofilms and abiotic fouling associated with
NOM of a microbial origin (e.g.. organic
colloids from cell fragments). The data does not
support the concept that fouling occurs as a
result of biofilm growth.
Fouling is generally a linear relationship. For
example, as water is filtered through a
membrane, the fouling starts immediately and
increases on a linear relationship. This type of
observed data supports fouling from colloidal
particles. Data based on recent studies indicate
pretreatment should be for colloidal particle
removal prior to the membranes.
One can view membranes as sheets of plastic
(synthetic organic polymers -0.2-mm-thick) or
ceramic-based materials with small holes (pores).
Positive Ca ions, which are in the waters to be
filtered, will coat the negatively-charged
membrane surface and will result in the
attachment of negative colloidal particles. The
author's hypothesis is that colloidal particles are
organic or inorganic materials that are typically
angular in nature and tend to concentrate at the
interface of the membrane. They are attracted by
the Ca surfacing on the membrane and become
pegged or trapped in the pores allowing a netting
or web structure to occur at the interface of the
membrane. The pegging and webbing of the
colloids cause fouling. Consequently, the focus
on pretreatment should be the removal of
suspended solids and colloidal solids. If this is
the focus, one has to question as to why
membranes are used at all for suspended solid or
colloidal particle removal. If there is a need to
remove these particles before they are in contact
with the membrane, then why are LP
membranes used to remove these particles?
Results from a November 2007 study
completed at the University of Washington by a
research team (Kim et al) provide valuable data
that helps understand the role of MF and UF
membranes in prefiltration. A co-author of this
study, Dr. Mark Benjamin, a professor for the
Civil and Environmental Engineering
Department, and his team undertook a revealing
review of the literature on organic fouling. It has
always been thought that if the water prior to
MF or UF membrane treatment was pretreated
with powder activated carbon (PAC), or 50 to
200 mg/l of Alum or FeCl3 then the organics
would be removed and fouling would be
reduced significantly; however, there is no
consistency in the studies as to whether organic
removal improves flux management or not.
The Kim et al study also recognized the
inherent problem with metal hydroxides, which
act like a gel on the surface of a membrane. In
the study, iron and aluminum metal hydroxides
were dehydrated using heat to a powder (1.5 to
20-�) with an average particle size of 5 microns
They applied the heated iron oxide particles
(HIOP) or aluminum oxide particles (HAOP)
and the PAC in three different ways. First, they
tried the standard approach by mixing organic
adsorbents into water, allowing particles to
settle, then filtering the effluent before feeding it
through the membrane module. Second, they
tried mixing adsorbents and applying the
mixture to the membrane. Finally, they
precoated the membrane with three different
adsorbents and fed the wastewater to the
precoated membrane.
The research team determined that the best
approach for maintaining the flux on the
membrane was the precoat concept, followed
by the mixed-solution approach. The standard
treatment method produced significantly worse
results. This study also observed the following:
� UF membrane flux reduction increased linearly
versus water treated after filtering Lake
Washington surface water through a 0.45-�
filter. The sieve-type prefiltering to 0.45
microns did not reduce membrane fouling.
� The UF membrane retained a small fraction of
organics (3 percent).
� When the standard pretreatment concept is
used and the solids are filtered with a
0.45-micron prefilter, the organics decrease
significantly by the pretreatment, but
membranes still fouled and retained very little
organics.
� Organic removal was best with PAC, but the
fouling with PAC was worse than the other
two adsorbents using the standard method of
pretreatment.
In subsequent tests performed by Dr.
Benjamin, he found that by precoating the
membranes with an adsorbent, the length of
time to fouling was directly proportional to the
As a result, pretreatment
of colloids is a key
treatment process to
eliminate non-scaling
fouling in any membrane.
David Bromley
1. Clean unused membrane
2. Fouled Membrane � no positive charged precoat
3. Membrane with precoat of PAC and then rinsed
4. Membrane with precoat of HAOP's and then rinsed
Images by: Dr. Mark Benjamin, J. Kim,
Cai Zhenxiao, Seminar presentation on Micro
Granular Adsorptive Membrane Filtration,
University of Illinois, November 2009
1
3
2
4

Page 1Page 2Page 3Page 4Page 5Page 6Page 7Page 8Page 9Page 10Page 11Page 12Page 13Page 14Page 15Page 16Page 17Page 18Page 19Page 20Page 21Page 22Page 23Page 24Page 25Page 26Page 27Page 28Page 29Page 30Page 31Page 32Page 33Page 34Page 35Page 36Page 37Page 38Page 39Page 40Page 41Page 42Page 43Page 44Page 45Page 46Page 47Page 48Page 49Page 50Page 51Page 52 Produced by PageSuite