,; , r~

L ..... , ,. ,

~) ,. | •
DESALINATION

ELSEVIER Desalination 119 (1998) 315-322

Ultrafiltration as an advanced tertiary treatment process

for municipal wastewater
George Tchobanoglous a, Jeannie Darby a, Keith Bourgeous a, John McArdle b,
Paul Genest b, Michael Tylla c*
aDepartment of Civil and Environmental Engineering, University of California, Davis, CA 95616, USA,
Tel.." +916-752-5670, Fax: +916-752-7872; bKoch Membrane Systems Inc., 850 Main Street, Wilmington,
MA 01887, USA, Tel.: +978-694-7173, Fax: +978-694-7021, E-mail. [email protected] and
[email protected]; and CKoch International GmbH, Membrane Systems Division, Neusserstrasse 33,
40219 D~isseldorf, Germany, Tel.: +49-211-90195-24, Fax: +49-211-90195-95, E-mail. [email protected]

Received 29 June 1998; accepted 8 July 1998

Abstract

Treatment of secondary and tertiary municipal wastewater by membrane filtration has been investigated at the
University of California, Davis, California, USA, as described in the following paper. Secondary and tertiary
wastewater used during this investigation were produced from activated sludge and media filtration processes,
respectively. Ultrafiltration, hollow-fiber, polysulfone membranes were used in this investigation for the
treatment of secondary and tertiary municipal wastewater. Membrane operating parameters investigated were
cross-flow velocity, with circulation and single pass flow, trans-membrane pressure and backflushing methods.
Total solids and particle size distribution of secondary and tertiary wastewater were analyzed and correlated with
observed membrane performance. Reduction of total solids, biological oxygen demand (BOD), and chemical
oxygen demand (COD) of secondary wastewater was also determined after treatment with membrane and media
filtration processes.

Keywords: Wastewater reuse; Ultrafiltration membrane; Sand filtration; Particle size distribution

1. B a c k g r o u n d using either microfiltration or ultrafiltration

hollow-fiber membranes followed by reverse
Recent d e v e l o p m e n t s in m e m b r a n e osmosis spiral membranes is recognized as
filtration are providing new technically the low cost alternative for municipal
feasible and e c o n o m i c a l m e t h o d s for wastewater reuse plants [1].
treatment of municipal wastewater. For It is expected that the use of membrane
example, a combined membrane process f i l t r a t i o n for t r e a t m e n t of m u n i c i p a l

*Corresponding author
Presented at the Conferenceon Membranesin Drinkingand IndustrialWaterProduction,Amsterdam, September21-24, 1998,
International Water Services Association,European DesalinationSocietyand AmericanWater Works Association
0011-9164/98/$ - See front m~ter © 1998 ElsevierScienceB.V. All rights reserved.
P// S0011-9164(98)00175-1
316 G. Tchobanoglous et al. /Desalination 119 (1998) 315-322

wastewater will steadily increase with more polymers with the optimal combinations of
stringent discharge regulations and fresh properties for this application is perhaps the
water supply limitations. In response, major goals of membrane suppliers.
ultrafiltration and microfiltration membrane
suppliers have developed a number of 1.3. Membrane cartridges
different membrane structures, membrane
p o l y m e r s , m e m b r a n e c a r t r i d g e s and Membrane cartridge design should allow
operating procedures for treatment of for simple installation, fiber integrity
wastewater treatment, described below. checking and maintenance. Special tools and
Suppliers of reverse osmosis membranes training should not be needed for these
have been involved in some of these activities; procedures. Air leak and pressure hold tests
however, the majority of work seems focused are used to check integrity of hollow-fiber
on the use of hollow-fiber microfiltration and membranes. Clear cartridge housings allow
ultrafiltration membranes as pretreatment for visual identification of fiber leaks. Individual
spiral reverse osmosis membranes. fibers are repaired in each cartridge either on
the system itself or using a separate repair
1.1. Membrane structure stand. Cartridges should be installed on the
membrane system in a manner that provides
Ultrafiltration and microfiltration for their c o n v e n i e n t removal and
membranes are used for the treatment of replacement.
municipal wastewater. Ultrafiltration Membrane cartridges are designed for
membranes have tighter pore structures than either outside-in or inside-out flow. Outside-
microfiltration membranes, and remove in flow cartridges may use smaller diameter
larger amounts of solids. Reverse osmosis hollow-fibers because only clean, filtered
membranes have been shown to perform at water moves into the narrow fiber lumen, with
lower operating pressure and longer time solids remaining on the outside of the fiber.
between cleanings after ultrafiltration Inside-out flow cartridges maintain crossflow
compared to after microfiltration [2]. feed distribution while avoiding accumulation
Ultrafiltration membranes have asymmetrical of solids on the outside of the fiber bundle.
"skinned" surface structure. Depth fouling This provides for efficient membrane use and
does not occur with this type of membrane, high productivity. Note that the inside
resulting in high and consistent membrane diameter of a hollow-fiber membrane with
productivity. inside-out flow must be carefully selected to
avoid fiber plugging and loss of membrane
1.2. Membrane polymer productivity.

Membrane polymers used for treatment of 1.4. Operating procedures

municipal wastewater include polysulfone,
polypropylene, polyacrylonitrile (PAN), Backflushing with permeate and oxidant
polyvinylidene flouride (PVDF), cellulose chemicals, including sodium hypochlorite
acetate, and blends of these materials. Each and hydrogen peroxide, or compressed air, is
of these polymers has a set of properties used to c o n t r o l m e m b r a n e fouling.
desired for treatment of municipal wastewater Ammonia in municipal wastewater can reduce
including productivity (flux), chemical the effectiveness of chemical backflushing by
resistance, fouling resistance and mechanical converting hypochlorite into less active
strength. However, it is important to chloramine species. Reverse osmosis
recognize that polymers having an especially permeate can be used as backflush solution,
useful property in one area may have a instead of ultrafiltration permeate, to avoid
critical deficit in another area. Identifying reaction of hypochlorite with ammonia
 G. Tchobanoglous et al. /Desalination 119 (1998) 315-322 317

contained in ultrafiltration or microfiltration testing in this report. Koch Membrane

permeate. However, using reverse osmosis Systems manufactures these membranes
permeate for backflushing is not desired specifically for the treatment of municipal
because of the resulting loss of product water. wastewater (Table 1).
Some polymers, notably polypropylene,
are extremely sensitive to attack by oxidant Table 1
chemicals. Removal of residual chlorine by Ultrafiltration membrane description
the addition of reducing chemicals, such as
sodium metabisulfite or thiosulfite, is Manufacturer Koch Membrane Systems, Inc.
generally required with this type of Wilmington, MA, USA
Designation HF-479-35-PM100
membrane. Membrane material Polysulfone
Process flow for membrane systems Cartridge material Polysulfone
include dead-end, single-pass and Cartridge diameter 20 cm (8 inch)
recirculation. Municipal wastewater with Cartridge length 180 cm (72 inch)
higher solids; for example, secondary Fiber diameter(ID) 0.9 mm (35 mil)
Membranearea 45 m2 (479 ft2)
wastewater, may require recirculation flow for MWCO 100,000 daltons (nominal)
optimum membrane performance.
Municipal wastewater with lower solids; for
example, tertiary wastewater, may operate
with dead-end or single-pass flow, with lower
equipment and electricity costs. 3.1. University of California, Davis, CA

Secondary and tertiary wastewater used for

2. Municipal wastewater analysis pilot plant testing at this site was obtained
from an activated sludge process and a
It would be very useful for membrane Parkson Dynasand continuous upflow filter,
suppliers to have an analytical measurement respectively [3].
available to predict membrane performance Process flow used for pilot plant testing at
for the treatment of a specific municipal this site was single - pass with 10% bleed for
wastewater. Availability of this type of secondary and tertiary wastewater. Total water
analytical method could eliminate the need recovery was about 80%, including backflush
for expensive and time consuming pilot plant losses. Backflush started after each twenty
testing. minute production period, with 100-150 mg/1
With this in mind, secondary and tertiary hypochlorite in the backflush solution.
municipal wastewater have been sampled Ultrafiltration permeate was used for the
from several municipal treatment plants. backflush solution.
Total solids and particle size distribution was Membrane productivity was consistent at
determined for each of these samples for use 135-170 lmh (80-100 gfd) for secondary
as a predictor of membrane performance. and tertiary wastewater, and increased as
This method has shown promising results, expected at higher trans-membrane pressure
however, additional work is needed in this (Fig. 1).
area to further develop and refine this Particle sizes for secondary wastewater at
approach. this site are distributed preferentially to the
larger particle range (52.7%). Particle size
distribution shifts from larger to smaller size
3. Membrane testing particles after tertiary filtration, as expected.
Total solids in the secondary and tertiary
Ultrafiltration, hollow-fiber, polysulfone wastewater are 4.61 mg/1 and 1.87 mg/1,
membranes were used for all pilot plant respectively. Tertiary filtration removes about
318 G. Tchobanoglous et al. / Desalination 119 (1998) 315-322

University of California. Davis. CA. USA

4 2/13/98 - 3/17/98 4/30/9S-5/19/98
200
Increase hans-membrane
pressure
180

Secondary Efflue~tt
160

140

120
Secondary and Tertiary Wastewster •--4t--Transmembrane Pressure
Once- Through Flow • Cartridge Pressure Drop
Water Recovery - 90% operating/80% overall 100
A Cartridge Productivity
Backflush Frequency - 20 minutes/130 ppm CIO-
Back/lush Duration - 90 seconds

6O

2O

0 . . . . . . . 0

0 100 200 300 400 500 600 700 800 900 1000
Elapsed Time (hours).
Fig. 1. U n i v e r s i t y o f California, D a v i s , C A .

4 - 100

Gwinnett County. GA. USA 90

9/4/97 - 9/12/97
80

70

• Transmembrane Pressure
• Cartridge Pressure Drop
Second ary W astewater
.t. Cartridge Productivity
Recirculstion Flow
Water Recovery .. 90% operating/84% overall 50 "~
Baekflush Frequency = 20 m m / I 0 0 ppm CIO2"
Backflush Duration - 14,5 sec .4o

,30

,20

0 . . . . . . . . . ' . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0

0 20 40 60 80 100 120 140 160 180

Elapsed Time (Hours)
Fig. 2. G w i n n e t t C o u n t y , G A .
 G. Tchobanoglous et al. /Desalination 119 (1998) 315-322 319

Table 2
Particle size distribution

Site/Stream Particle size distribution (mg/l) Total

0.4-1 1-5 5-12 12-20 >20 (mg/1)
(mg/l) (%) (mg/1) (%) (mg/1) (%) (mg/1) (%) (mg/l) (%)

Davis, CA
Secondary 0.52 11.3 0.76 16.5 0.43 9.3 0.47 10.2 2.43 52.7 4.61
Tertiary 0.48 25.7 0.77 41.2 0.20 10.7 0.17 9.1 0.25 13.4 1.87
Gwinnett, GA
Secondary 0.40 5.9 0.87 12.9 0.56 8.3 0.75 11.1 4.17 61.8 6.75
WF21, CA
Secondary 1.30 38.3 0.93 27.4 0.40 11.8 0.24 7.1 0.52 15.3 3.39

Table 3
Average BOD/COD analysis

Site Stream BOD5 (mg/1) Removal(%) COD (mg/1) Removal(%)

Davis, CA Secondary 8.8 29.3

Tertiary filtered 3.4 61.4 17.2 41.3
Membrane filtered 1.1 87.5 6.2 78.8

60% of total solids from the secondary water recovery was about 85%, including
wastewater (Table 2). backflush losses. Backflush started after each
BOD removal from secondary wastewater twenty minute production period, with 100-
was 61.4% after tertiary filtration and 87.5% 150 mg/1 hypochlorite in the backflush
after tertiary and membrane filtration. COD solution. H y d r o g e n p e r o x i d e was also
removal from secondary wastewater was successfully used in the backflush solution.
41.3% after tertiary filtration and 78.8% after Ultrafiltration permeate was used for the
tertiary and membrane filtration (Table 3). backflush solution.
High and consistent membrane Membrane productivity was consistent at
p r o d u c t i v i t y was o b s e r v e d p r o c e s s i n g 75 lmh (-45 gfd) for secondary wastewater
secondary and tertiary wastewater at this site (Fig. 2), Larger diameter hollow-fibers (35
indicating that removal of large particles mil versus 30 mil) were used to avoid fiber
from this wastewater by tertiary filtration does plugging and to maintain high membrane
not have a significant impact on performance productivity.
of the membrane system. Particle sizes of secondary wastewater at
Gwinnett are distributed preferentially to the
3.2. Gwinnett County, GA larger particle size range (61.8%). Total
solids at this site were higher (6.75 mg/1) than
Secondary wastewater used for pilot plant at UC Davis (Table 2).
testing at this site was obtained from an
activated sludge process. 3.3. Water Factory 21, Orange County, CA
Process flow for pilot plant testing at this
site was recirculation with 10% bleed. Overall Secondary wastewater used for pilot plant
320 G. Tchobanoglous et al. / Desalination 119 (1998) 315-322

4 • 1~

.90
W',d~ F~torv 2~ CA, USA
60.2/W - 7 / I ~ .80

.70
i
,60
Secondary W a s ~ s v a t ~
~tian Flow ~ TrammembranePressure

p , , , , w .40
kddlmh Duration - 120 se:ondl d:
.30

211

10

' I I I a .0
I00 2O0 3OO 4OO 5OO
~ a p ~ l T ~ (Hours)
Fig. 3. Water Factory 21, CA.

testing at this site was obtained from a blend concentration of solids of secondary
of activated sludge and trickling bed wastewater in the smaller particle range at WF
processes. 21 is higher (1.30 mg/l) than at Gwinnett (0.4
Process flow for pilot plant testing at this mg/1) and UC Davis (0.52 mg/l).
site was recirculation with 10% bleed. Overall Membrane productivity with secondary
water recovery was about 85%, including wastewater at WF 21 and Gwinnett are about
backflush losses. Backflush started after each the same with larger diameter fibers (35
twenty minute production period, with 100- versus 30 rail) used at Gwinnett to avoid fiber
150 mg/1 hypochlorite in the backflush plugging. This result is explained by the
solution. Hydrogen peroxide was also higher total solids concentration of Gwinnett
successfully used in the backflush solution. secondary wastewater, especially in the larger
Ultrafiltration permeate was used for the particle size range.
backflush solution. Membrane productivity with secondary
Membrane productivity was consistent at wastewater at WF 21 using recirculation flow
65 lmh (~40 gfd) for secondary wastewater is lower than at UC Davis using single-pass
(Fig. 3). Cartridge pressure drop was about flow, indicating that secondary wastewater at
the same at WF 21 and Gwinnett, however, WF 21 is more difficult to process than at UC
crossflow velocity at Gwinnett was lower than Davis. Total solids of WF 21 secondary
at WF 21 because of the larger diameter wastewater is lower than at UC Davis;
fibers at Gwinnett. however, the concentration of smaller
Particle sizes of secondary wastewater at particles at WF 21 (1.30 mg/1) is larger than at
WF 21 are distributed preferentially to the UC Davis (0.52 mg/1). This observation
smaller particle size range (38.3%). Total indicates smaller particles have a
solids of secondary wastewater at WF 21 is d i s p r o p o r t i o n a t e impact on m e m b r a n e
lower (3.39 mg/1) than at Gwinnett (6.75 performance than larger particles.
mg/l) and UC Davis (4.61 mg/1). However, WF 21 secondary wastewater has a higher
 G. Tchobanoglous et al. /Desalination 119 (1998) 315-322 321

concentration of smaller particles (1.30 mg/l) Total solids and particle size distribution
than Gwinnett secondary wastewater (0.40 analysis of secondary and tertiary municipal
mg/1); however, larger diameter fibers were w a s t e w a t e r appear to p r o v i d e a useful
needed at Gwinnett to achieve about the same information for predicting the performance
membrane productivity as at W F 21. As of hollow-fiber ultrafiltration membranes.
previously reported, total solids of secondary Additional pilot plant testing is planned to
wastewater are higher at Gwinnett than at WF further investigate this approach.
21. This observation indicates membrane
performance depends on both total solids and
particles size distribution.
References

4. Conclusions [ll G.L. Leslie et al., Performance and economic

evaluation of membrane processes for reuse
applications, Proceedings Water Reuse
Pilot plant testing at several municipal Conference, Lake Buena Vista, FL, USA, 1998.
wastewater treatment plants has demonstrated [2l G.L. Leslie et al., Pilot testing of microfiltration
that optimal use of hollow-fiber ultrafiltration and ultrafiltration upstream of reverse osmosis
and microfiltration membranes to process during reclamation of municipal wastewater,
secondary and tertiary wastewater requires the Proceedings American Desalting Association,
Biennial Conference and Exposition, Monterey,
careful combination of membrane structure, CA, 1996.
membrane polymer, membrane cartridge and [3] G. Tchobanoglous et al., Continuous backwash
operating procedures of the membrane upflow filtration for primary effluent, Water
system. Environmental Research, 66 (1994).