Wastewater Management Fact Sheet

Membrane Bioreactors

INTRODUCTION
The technologies most commonly used for performing secondary treatment of municipal
wastewater rely on microorganisms suspended in
the wastewater to treat it. Although these technologies work well in many situations, they have
several drawbacks, including the difficulty of
growing the right types of microorganisms and
the physical requirement of a large site. The use
of microfiltration membrane bioreactors
(MBRs), a technology that has become increasingly used in the past 10 years, overcomes many
of the limitations of conventional systems. These
systems have the advantage of combining a suspended growth biological reactor with solids
removal via filtration. The membranes can be
designed for and operated in small spaces and
with high removal efficiency of contaminants
such as nitrogen, phosphorus, bacteria, biochemical oxygen demand, and total suspended
solids. The membrane filtration system in effect
can replace the secondary clarifier and sand filters in a typical activated sludge treatment
system. Membrane filtration allows a higher
biomass concentration to be maintained, thereby
allowing smaller bioreactors to be used.

APPLICABILITY
For new installations, the use of MBR systems
allows for higher wastewater flow or improved
treatment performance in a smaller space than a
conventional design, i.e., a facility using secondary clarifiers and sand filters. Historically,
membranes have been used for smaller-flow systems due to the high capital cost of the
equipment and high operation and maintenance
(O&M) costs. Today however, they are receiving
increased use in larger systems. MBR systems
are also well suited for some industrial and
commercial applications. The high-quality effluent produced by MBRs makes them particularly
applicable to reuse applications and for surface

water discharge applications requiring extensive

nutrient (nitrogen and phosphorus) removal.

ADVANTAGES AND DISADVANTAGES

The advantages of MBR systems over conventional biological systems include better effluent
quality, smaller space requirements, and ease of
automation. Specifically, MBRs operate at
higher volumetric loading rates which result in
lower hydraulic retention times. The low retention times mean that less space is required
compared to a conventional system. MBRs have
often been operated with longer solids residence
times (SRTs), which results in lower sludge production; but this is not a requirement, and more
conventional SRTs have been used (Crawford et
al. 2000). The effluent from MBRs contains low
concentrations of bacteria, total suspended solids
(TSS), biochemical oxygen demand (BOD), and
phosphorus. This facilitates high-level disinfection. Effluents are readily discharged to surface
streams or can be sold for reuse, such as irrigtion.
The primary disadvantage of MBR systems is
the typically higher capital and operating costs
than conventional systems for the same throughput. O&M costs include membrane cleaning and
fouling control, and eventual membrane replacement. Energy costs are also higher because
of the need for air scouring to control bacterial
growth on the membranes. In addition, the waste
sludge from such a system might have a low
settling rate, resulting in the need for chemicals
to produce biosolids acceptable for disposal
(Hermanowicz et al. 2006). Fleischer et al. 2005
have demonstrated that waste sludges from
MBRs can be processed using standard technologies used for activated sludge processes.

MEMBRANE FILTRATION
Membrane filtration involves the flow of watercontaining pollutants across a membrane. Water
permeates through the membrane into a separate
channel for recovery (Figure 1). Because of the
cross-flow movement of water and the waste
constituents, materials left behind do not accumulate at the membrane surface but are carried
out of the system for later recovery or disposal.
The water passing through the membrane is
called the permeate, while the water with the
more-concentrated materials is called the concentrate or retentate.

ment is that the membranes prevent passage of

particles the size of microorganisms, or about 1
micron (0.001 millimeters), so that they remain
in the system. This means that MBR systems are
good for removing solid material, but the removal of dissolved wastewater components must
be facilitated by using additional treatment steps.
Membranes can be configured in a number of
ways. For MBR applications, the two configurations most often used are hollow fibers grouped
in bundles, as shown in Figure 2, or as flat
plates. The hollow fiber bundles are connected by
manifolds in units that are designed for easy
changing and servicing.

Figure 2. Hollow-fiber membranes (Image

from GE/Zenon)

DESIGN CONSIDERATIONS

Figure 1. Membrane filtration process

(Image from Siemens/U.S. Filter)

Membranes are constructed of cellulose or other

polymer material, with a maximum pore size set
during the manufacturing process. The require2

Designers of MBR systems require only basic

information about the wastewater characteristics,
(e.g., influent characteristics, effluent requirements, flow data) to design an MBR system.
Depending on effluent requirements, certain
supplementary options can be included with the
MBR system. For example, chemical addition (at
various places in the treatment chain, including:
before the primary settling tank; before the secondary settling tank [clarifier]; and before the
MBR or final filters) for phosphorus removal can
be included in an MBR system if needed to
achieve low phosphorus concentrations in the
effluent.
MBR systems historically have been used for
small-scale treatment applications when portions
of the treatment system were shut down and the

wastewater routed around (or bypassed) during

maintenance periods.
However, MBR systems are now often used in
full-treatment applications. In these instances, it
is recommended that the installation include one
additional membrane tank/unit beyond what the
design would nominally call for. This N plus 1
concept is a blend between conventional activated sludge and membrane process design. It is
especially important to consider both operations
and maintenance requirements when selecting
the number of units for MBRs. The inclusion of
an extra unit gives operators flexibility and ensures that sufficient operating capacity will be
available (Wallis-Lage et al. 2006). For example,
bioreactor sizing is often limited by oxygen
transfer, rather than the volume required to
achieve the required SRTa factor that significantly affects bioreactor numbers and sizing
(Crawford et al. 2000).
Although MBR systems provide operational
flexibility with respect to flow rates, as well as
the ability to readily add or subtract units as conditions dictate, that flexibility has limits.
Membranes typically require that the water surface be maintained above a minimum elevation
so that the membranes remain wet during operation. Throughput limitations are dictated by the
physical properties of the membrane, and the
result is that peak design flows should be no

more than 1.5 to 2 times the average design flow.

If peak flows exceed that limit, either additional
membranes are needed simply to process the
peak flow, or equalization should be included in
the overall design. The equalization is done by
including a separate basin (external equalization)
or by maintaining water in the aeration and
membrane tanks at depths higher than those required and then removing that water to
accommodate higher flows when necessary (internal equalization).

DESIGN FEATURES
Pretreatment
To reduce the chances of membrane damage,
wastewater should undergo a high level of debris
removal prior to the MBR. Primary treatment is
often provided in larger installations, although
not in most small to medium sized installations,
and is not a requirement. In addition, all MBR
systems require 1- to 3-mm-cutoff fine screens
immediately before the membranes, depending
on the MBR manufacturer. These screens require
frequent cleaning. Alternatives for reducing the
amount of material reaching the screens include
using two stages of screening and locating the
screens after primary settling.
Membrane Location
MBR systems are configured with the memTurbidimeter

Mixed
Anoxic

Aerobic + ZeeWeed

Pretreated
Wastewater
Feed

Treated
Water

Sludge
Recycle

Clean in
Place
Tank

Blowers

Sludge Wasted
@ 1 - 1.2 wt% TS

Figure 3.

Immersed membrane system configuration (Image from GE/Zenon)

3

Figure 4. External membrane system configuration (Image from Siemens/U.S. Filter)

branes actually immersed in the biological reactor or, as an alternative, in a separate vessel
through which mixed liquor from the biological
reactor is circulated. The former configuration is
shown in Figure 3; the latter, in Figure 4.
Membrane Configuration
MBR manufacturers employ membranes in two
basic configurations: hollow fiber bundles and
plate membranes. Siemens/U.S.Filters Memjet
and Memcor systems, GE/Zenons ZeeWeed and
ZenoGem systems, and GE/Ionics system use
hollow-fiber, tubular membranes configured in
bundles. A number of bundles are connected by
manifolds into units that can be readily changed
for maintenance or replacement. The other configuration, such as those provided by
Kubota/Enviroquip, employ membranes in a flatplate configuration, again with manifolds to allow a number of membranes to be connected in
readily changed units. Screening requirements
for both systems differ: hollow-fiber membranes
typically require 1- to 2-mm screening, while
4

plate membranes require 2- to 3-mm screening

(Wallis-Lage et al. 2006).
System Operation
All MBR systems require some degree of pumping to force the water flowing through the
membrane. While other membrane systems use a
pressurized system to push the water through the
membranes, the major systems used in MBRs
draw a vacuum through the membranes so that
the water outside is at ambient pressure. The
advantage of the vacuum is that it is gentler to
the membranes; the advantage of the pressure is
that throughput can be controlled. All systems
also include techniques for continually cleaning
the system to maintain membrane life and keep
the system operational for as long as possible.
All the principal membrane systems used in
MBRs use an air scour technique to reduce
buildup of material on the membranes. This is
done by blowing air around the membranes out
of the manifolds. The GE/Zenon systems use air
scour, as well as a back-pulsing technique, in
which permeate is occasionally pumped back

into the membranes to keep the pores cleared

out. Back-pulsing is typically done on a timer,
with the time of pulsing accounting for 1 to 5
percent of the total operating time.
Downstream Treatment
The permeate from an MBR has low levels of
suspended solids, meaning the levels of bacteria,
BOD, nitrogen, and phosphorus are also low.
Disinfection is easy and might not be required,
depending on permit requirements..
The solids retained by the membrane are recycled to the biological reactor and build up in the
system. As in conventional biological systems,
periodic sludge wasting eliminates sludge
buildup and controls the SRT within the MBR
system. The waste sludge from MBRs goes
through standard solids-handling technologies
for thickening, dewatering, and ultimate disposal. Hermanowicz et al. (2006) reported a
decreased ability to settle in waste MBR sludges
due to increased amounts of colloidal-size particles and filamentous bacteria. Chemical addition
increased the ability of the sludges to settle. As
more MBR facilities are built and operated, a
more definitive understanding of the characteristics of the resulting biosolids will be achieved.
However, experience to date indicates that conventional biosolids processing unit operations
are also applicable to the waste sludge from
MBRs.
Membrane Care
The key to the cost-effectiveness of an MBR
system is membrane life. If membrane life is
curtailed such that frequent replacement is required, costs will significantly increase.
Membrane life can be increased in the following
ways:
- Good screening of larger solids before the
membranes to protect the membranes from
physical damage.
- Throughput rates that are not excessive, i.e.,
that do not push the system to the limits of
the design. Such rates reduce the amount of
material that is forced into the membrane and
thereby reduce the amount that has to be re-

moved by cleaners or that will cause eventual

membrane deterioration.
- Regular use of mild cleaners. Cleaning solutions most often used with MBRs include
regular bleach (sodium) and citric acid. The
cleaning should be in accord with manufacturer-recommended maintenance protocols.
Membrane Guarantees
The length of the guarantee provided by the
membrane system provider is also important in
determining the cost-effectiveness of the system.
For municipal wastewater treatment, longer
guarantees might be more readily available compared to those available for industrial systems.
Zenon offers a 10-year guarantee; others range
from 3 to 5 years. Some guarantees include cost
prorating if replacement is needed after a certain
service time. Guarantees are typically negotiated
during the purchasing process. Some manufacturers guarantees are tied directly to screen size:
longer membrane warranties are granted when
smaller screens are used (Wallis-Lage et al.
2006). Appropriate membrane life guarantees
can be secured using appropriate membrane procurement strategies (Crawford et al. 2002).

SYSTEM PERFORMANCE
Siemens/U.S. Filter Systems
Siemens/U.S.Filter offers MBR systems under
the Memcor and Memjet brands. Data provided
by U.S. Filter for its Calls Creek (Georgia) facility are summarized below. The system, as Calls
Creek retrofitted it, is shown in Figure 5. In essence, the membrane filters were used to replace
secondary clarifiers downstream of an Orbal
oxidation ditch. The system includes a fine
screen (2-mm cutoff) for inert solids removal just
before the membranes.
The facility has an average flow of 0.35 million
gallons per day (mgd) and a design flow of 0.67
mgd. The system has 2 modules, each containing
400 units, and each unit consists of a cassette
with manifold-connected membranes. As shown
in Table 1, removal of BOD, TSS, and ammonianitrogen is excellent; BOD and TSS in the effluent are around the detection limit. Phosphorus is
also removed well in the system, and the effluent
5

Figure 5.

Calls Creek flow diagram (courtesy of Siemens/U.S. Filter)

Table 1.
Calls Creek results 2005
Parameter

Influent
Average

Flow (mgd)

Average

0.44

0.26

TSS (mg/L)

248

0.21

0.72

0.10

0.55

0.12

14.8
0.88

Fecal coliforms (#/100 mL)

--

Turbidity (NTU)

--

has very low turbidity. The effluent has consistently met discharge limits.
Zenon Systems
General Electric/Zenon provides systems under
the ZenoGem and ZeeWeed brands. The ZeeWeed brand refers to the membrane, while
ZenoGem is the process that uses ZeeWeed.
Performance data for two installed systems are
shown below.

--

Min Month

145

P (mg/L)

0.35

Max Month

BOD (mg/L)
Ammonia-N (mg/L)

Effluent

0.28
14.2
0.30

20
1.31

0
0.01

Cauley Creek, Georgia. The Cauley Creek facility in Fulton County, Georgia, is a 5-mgd
wastewater reclamation plant. The system
includes biological phosphorus removal, mixed
liquor surface wasting, and sludge thickening
using a ZeeWeed system to minimize the required volume of the aerobic digester, according
to information provided by GE. Ultraviolet disinfection is employed to meet regulatory limits.
Table 2 shows that the removal for all parame-

Table 2.
Cauley Creek, Georgia, system performance
Parameter

Flow (mgd)

Influent

Effluent

Average

Average

Max Month

Min Month

4.27

--

4.66

3.72

2.0

2.0

BOD (mg/L)

182

COD (mg/L)

398

TSS (mg/L)

174

2.0
12

22

3.2

TKN (mg/L)

33.0

1.9

2.9

1.4

Ammonia-N (mg/L)

24.8

0.21

0.29

0.10

5.0

TP (mg/L)

0.1

0.13

0.06

Fecal coliforms (#/100 mL)

--

NO3-N (mg/L)

--

2.8

ters is over 90 percent. The effluent meets all

permit limits, and is reused for irrigation and
lawn watering.
Traverse City, Michigan. The Traverse City
Wastewater Treatment Plant (WWTP) went
through an upgrade to increase plant capacity
and produce a higher-quality effluent, all within
the facilitys existing plant footprint (Crawford
et al. 2005). With the ZeeWeed system, the facility was able to achieve those goals. As of 2006,
the plant is the largest-capacity MBR facility in
North America. It has a design average annual
flow of 7.1 mgd, maximum monthly flow of 8.5
mgd, and peak hourly flow of 17 mgd. The
membrane system consists of a 450,000-gallon
tank with eight compartments of equal size. Secondary sludge is distributed evenly to the
compartments. Blowers for air scouring, as well
as permeate and back-pulse pumps, are housed in
a nearby building.
Table 3 presents a summary of plant results over
a 12-month period. The facility provides excellent removal of BOD, TSS, ammonia-nitrogen,
and phosphorus. Figure 6 shows the influent,
effluent, and flow data for the year.
Operating data for the Traverse City WWTP
were obtained for the same period. The mixed
liquor suspended solids over the period January
to August averaged 6,400 mg/L, while the mixed
liquor volatile suspended solids averaged 4,400
mg/L. The energy use for the air-scouring blow-

ers averaged 1,800 kW-hr/million gallons (MG)

treated.

COSTS
Capital Costs
Capital costs for MBR systems historically have
tended to be higher than those for conventional
systems with comparable throughput because of
the initial costs of the membranes. In certain
situations, however, including retrofits, MBR
systems can have lower or competitive capital
costs compared with alternatives because MBRs
have lower land requirements and use smaller
tanks, which can reduce the costs for concrete.
U.S. Filter/Siemens Memcor package plants
have installed costs of $7$20/gallon treated.
Fleischer et al. (2005) reported on a cost comparison of technologies for a 12-MGD design in
Loudoun County, Virginia. Because of a chemical oxygen demand limit, activated carbon
adsorption was included with the MBR system.
It was found that the capital cost for MBR plus
granular activated carbon at $12/gallon treated
was on the same order of magnitude as alternative processes, including multiple-point alum
addition, high lime treatment, and postsecondary membrane filtration.
Operating Costs
Operating costs for MBR systems are typically
higher than those for comparable conventional
systems. This is because of the higher energy
7

Table 3.
Summary of Traverse City, Michigan, Performance Results
Influent
Average
Flow (mgd)

Effluent
Average

4.3

Max Month

--

Min Month

5.1

3.6

BOD (mg/L)

280

<2

<2

<2

TSS (mg/L)

248

<1

<1

<1

27.9

< 0.08

< 0.23

< 0.03

6.9

0.7

0.95

0.41

17.2

--

Ammonia-N (mg/L)
TP (mg/L)

Influent BOD and TSS (mg/L)

Temperature (deg C)

23.5

11.5

350

35

300

30

250

25

200

20

150

15

100

10

50

0
Aug-05
Inf. BOD

Effl. BOD/TSS (mg/L); Infl. and Effl. NH3-N, PO4-P (mg/L); and
Flow (MGD)

Parameter

0
Sep-05
Inf. TSS

Nov-05
Inf. NH3-N

Jan-06
Inf. PO4-P

Feb-06
Eff. BOD

Apr-06
Eff. NH3-N

Jun-06
Eff. TSS

Jul-06
Eff. PO4-P

Flow (MGD)

Figure 6. Performance of the Traverse City plant

costs if air scouring is used to reduce membrane

fouling. The amount of air needed for the scouring has been reported to be twice that needed to
maintain aeration in a conventional activated
sludge system (Scott Blair, personal communication, 2006). These higher operating costs are
often partially offset by the lower costs for
sludge disposal associated with running at longer
sludge residence times and with membrane
thickening/dewatering of wasted sludge.
Fleischer et al. (2005) compared operating costs.
They estimated the operating costs of an MBR
system including activated carbon adsorption at
$1.77 per 1,000 gallons treated. These costs were
8

of the same order of magnitude as those of alternative processes, and they compared favorably to
those of processes that are chemical-intensive,
such as lime treatment.

ACKNOWLEDGMENTS
The authors acknowledge Dr. Venkat Mahendraker, GE/Zenon, Mr. John Irwin, Siemens/
U.S. Filter, and Mr. Scott Blair and Mr. Leroy
Bonkoski of the Traverse City WWTP for
their assistance in obtaining data and system
information. EPA acknowledges external peer

reviewers Pat Brooks, Alan Cooper, and Glenn

Daigger for their contribution.

PRODUCT LITERATURE USED

Enviroquip/Kubota. Sales literature.
Siemens. Product literature.
<http://www.usfilter.com/en/Product+Lines/
Envirex_Products/Envirex_Products/
envirex_mbr_xpress_packaged_plant.htm>.
Zenon. Case studies: Cauley Creek, Georgia.
<http://www.zenon.com/resources/case_studies/
water_reuse/CauleyCreek.shtml>.
Zenon. Case studies: Traverse City, Michigan.
<http://www.zenon.com/resources/case_studi
es/wastewater/TraverseCity.shtml>.

REFERENCES
Crawford, G., G. Daigger, J. Fisher, S. Blair, and
R. Lewis. 2005. Parallel Operation of Large
Membrane Bioreactors at Traverse City. In
Proceedings of the Water Environment Federation 78th Annual Conference &
Exposition, Washington, DC, CD-ROM,
October 29Nov 2, 2005.
Crawford, G., A. Fernandez, A. Shawwa, and G.
Daigger. 2002 Competitive Bidding and
Evaluation of Membrane Bioreactor EquipmentThree Large Plant Case Studies. In
Proceedings of the Water Environment Federation 75th Annual Conference &
Exposition, Chicago, IL, CD-ROM, September 28Oct 2, 2002.

Fleischer, E.J., T.A. Broderick, G.T. Daigger, A.

D. Fonseca, R.D. Holbrook, and S.N. Murthy.
2005. Evaluation of Membrane Bioreactor
Process Capabilities to Meet Stringent Effluent Nutrient Discharge Requirements. Water
Environment Research 77:162178.
Fleischer, E. J., T. A. Broderick, G. T. Daigger,
J. C. Lozier, A. M. Wollmann, and A. D.
Fonseca. 2001. Evaluating the Next Generation of Water Reclamation Processes. In
Proceedings of the Water Environment Federation 74th Annual Conference & Exposition,
Atlanta, GA, CD-ROM, October 1317, 2001.
Hermanowicz, S.W., D. Jenkins, R.P. Merlo, and
R.S. Trussell. 2006. Effects of Biomass Properties on Submerged Membrane Bioreactor
(SMBR) Performance and Solids Processing.
Document no. 01-CTS-19UR. Water Environment Federation.
Metcalf & Eddy. 2003. Wastewater Engineering,
Treatment and Reuse. 4th ed. McGraw-Hill,
New York.
Wallis-Lage, C., B. Hemken, et al. 2006. MBR
Plants: Larger and More Complicated. Presented at the Water Reuse Associations 21st
Annual Water Reuse Symposium, Hollywood, CA, September 2006.

Crawford, G., D. Thompson, J. Lozier, G. Daigger,

and E. Fleischer. 2000. Membrane
BioreactorsA Designers Perspective. In

Proceedings of the Water Environment

Federation 73rd Annual Conference &
Exposition on Water Quality and
Wastewater Treatment, Anaheim, CA,
CD-ROM, October 14-18, 2000.

EPA 832-F-07-015
Office of Water
September 2007