172 Q IWA Publishing 2010 Water Science & Technology—WST | 62.

1 | 2010

Improving the performance of membrane bioreactors

by powdered activated carbon dosing with cost
considerations
W. Yang, M. Paetkau and N. Cicek

ABSTRACT

Effects of powdered activated carbon (PAC) dosing on the overall performance of membrane W. Yang
M. Paetkau
bioreactors (MBR) were investigated in two bench-scale submerged MBRs. Positive impacts of N. Cicek (corresponding author)
Department of Biosystems Engineering,
PAC dosing on membrane fouling and the removal of 17b-estradiol (E2) and 17a-ethyinylestradiol
University of Manitoba,
(EE2) were demonstrated over a six-month stable operational period. PAC dosing in the MBR Winnipeg, MB R3T 5V6,
Canada
increased the removal rates of E2 and EE2 by 3.4% and 15.8%, respectively. The average soluble E-mail: [email protected];
[email protected];
extracellular polymeric substances (EPS) and colloidal total organic carbon (TOC) concentrations [email protected]

in the PAC-MBR sludge was 60.1% and 61.8% lower than the control MBR sludge, respectively.
Lower soluble EPS and colloidal TOC concentrations in the PAC-MBR sludge resulted in a slower
rate of trans-membrane pressure (TMP) increase during MBRs operation, which could prolong
the lifespan of membranes. Cost assessment showed that PAC dosing could reduce the
operating cost for membrane cleaning and/or membrane replacement by about 25%.
The operating cost for PAC dosing could be offset by the benefit from its reducing the cost for
membrane maintenance.
Key words | cost analysis, endocrine disrupting compounds, estrogen, membrane bioreactor,
powdered activated carbon, sludge characteristics

INTRODUCTION

Membrane bioreactor (MBR) technology has been The application of PAC dosing in MBRs dated back
advancing rapidly around the world both in research and to 1998 (Ohnishi et al. 1998). In this system, PAC was
commercial applications in the last decades (Yang et al. added primarily to reduce the color of treated wastewater
2006). Membrane fouling, which has been widely recog- for reuse. It was also found that the activated sludge
nized to be one of the major limitations to faster commer- dewaterability could be improved by adding PAC in the
cialization of MBRs, has been investigated from various bioreactor. The first application of PAC addition for flux
perspectives in many studies; these perspectives include the enhancement in a membrane anaerobic bioreactor was
causes, characteristics, mechanisms of fouling, and methods published in 1999 (Park et al. 1999). In this study, a flux
to prevent or reduce membrane fouling. Flux improvement improvement with PAC addition was achieved, especially
can be achieved with powdered activated carbon (PAC) with a higher PAC dosage of up to 5 g/L. In recent years, a
addition because PAC plays an important role in reducing number of studies were conducted, aiming to use PAC
the biomass cake resistance and changing the filtration dosing as a membrane fouling control method (Ng et al.
characteristics of the sludge in bioreactors. In addition, PAC 2006; Hu & Stuckey 2007; Remy et al. 2009).
might have a scouring effect for removing the cake layer The occurrence and behaviour of endocrine disrupting
from membrane surfaces. compounds (EDCs), including estrogens in wastewater
doi: 10.2166/wst.2010.276

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018
 173 W. Yang et al. | Improving the performance of membrane bioreactors Water Science & Technology—WST | 62.1 | 2010

treatment plants (WWTPs), has gained considerable same feed at similar experimental conditions (Figure 1).
attention worldwide. Some scientists concluded that dis- A hollow fiber membrane module with a surface area of
charge from WWTPs is one of the largest sources of EDCs in 0.08 m2 from Korea Membrane Separation Ltd. was
the Canadian aquatic environment (Servos et al. 2005). submerged in each bioreactor. The pore size of the
Steroid estrogens, especially 17b-estradiol (E2) and membrane fibers was 0.4 mm. The working volume of each
17a-ethyinylestradiol (EE2) are major EDCs of particular bioreactor was 3.0 L. Coarse bubble air-diffusers at the
concern for the aquatic environment. MBRs have been bottom of membrane module were used to provide
proven quite effective in removing both organic and dissolved oxygen (DO) for biomass growth and to introduce
inorganic pollutants as well as biological entities from water vigorous shear force on the membrane surface to reduce
or wastewater streams. Due to their advantages such as high membrane fouling.
sludge age, high biomass concentration and complete
particle retention, MBRs are expected to achieve enhanced Operation of MBRs
performance in removing EDCs. This hypothesis has been
The two investigated MBR systems were operated at
supported by a number of previous studies on the removal of
constant flux mode by withdrawing the permeate through
EDCs in MBR process ( Joss et al. 2004; Chang et al. 2006).
peristaltic pumps from the outside-to-inside hollow mem-
By dosing PAC in the bioreactor, the characteristics of
brane fibers. The two MBR systems were fed with identical
the sludge (PAC-sludge) can be expected to vary. PAC-
synthetic wastewater by gravity. The composition of
sludge is hypothesized to possess higher adsorption
synthetic wastewater has been described in Yang et al.
capacity, which would increase its capacity to adsorb
(2009). The water levels of the bioreactor were controlled by
substances including estrogens to its surface. A couple of
floating valves. To control membrane fouling, a cyclic
studies have been conducted to test the efficiency of PAC in
pumping mode with 10 minutes ON and 2 minutes OFF
further eliminating estrogens in MBR. Korner et al. (2001)
was used. Solids retention time (SRT) of both MBR and
reported that additional activated carbon filtration was very
PAC-MBR remained 25 days by wasting portion of mix
efficient in further eliminating estrogenic activity from
liquor from bioreactors daily. Hydraulic retention time
WWTPs effluents. On the contrary, other investigators
(HRT) of both MBRs was 5.8– 6.0 hour, depending on
have shown that the estrogenic activity in aqueous phase
membrane flux. During the experimental period, trans-
leaving the MBR system was not further reduced by
membrane pressures (TMP) of the two MBR systems were
polishing with either granular activated carbon (GAC) or
monitored every day. Temperature, DO, and pH in the
PAC (Holbrook et al. 2002).
bioreactors were measured on a regular base. The values of
With the expansion of MBR installations for wastewater
DO in the two bioreactors were not less than 4 mg/L during
treatment and water reuse, there needs to be an assessment
the experimental period.
as to whether PAC dosing is able to improve the overall
In the PAC-MBR, PAC with particle size range from 50
MBR performance including the removal of estrogens.
to 150 mm was added into the bio-reactor at a dosage of
Meanwhile, the cost of PAC dosing in MBRs needs be
2.0 g/L. The PAC dosage remained at the same level by
assessed. The objective of this study was to investigate the
effects of PAC dosing on the overall performance of MBRs
Synthetic Pressure meter
and assess the cost of PAC dosing in MBRs. wastewater

Feed tank
Peristaltic
MBR PAC-MBR pump

MATERIAL AND METHODS

Effluent
Experimental set-up Air compressor
Stirrer
Two bench-scale MBRs, one without PAC dosing and the
other with PAC dosing, were operated in parallel with the Figure 1 | Schematic of the bench-scale MBR and PAC-MBR.

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018
 174 W. Yang et al. | Improving the performance of membrane bioreactors Water Science & Technology—WST | 62.1 | 2010

re-supplementing the bioreactor with certain amount of efficiencies of E2 from MBR and PAC-MBR sludge were
PAC every other day. The selection of the particle size 82.3% and 71.0%, respectively. The average extraction
of PAC was based on previous studies. Smaller particles efficiencies of EE2 from MBR and PAC-MBR sludge were
provide quicker rates of adsorption, but might tend to 76.7% and 64.2%, respectively.
increase the small particle population of the activated
sludge and shift the flocs size to lower ranges (Ng et al.
2006). However, PAC with an average size of 100 mm was RESULTS AND DISCUSSION
reported to shift the particle size distribution of the mixed
Effects of PAC dosing on MLVSS
liquor broth to a relatively higher range (Park et al. 1999).
Based on these results, the PAC with a relatively higher Two identical MBRs were seeded with activated sludge
particle size range would be suitable for dosing into from a full-scale pollution control center in Winnipeg,
MBR systems. Manitoba, Canada. After eighty days of start-up and
acclimation, stable operation and consistent performance
in terms of stable MLVSS concentrations in the bioreactors
Analytical methods
and satisfactory COD removal was achieved in both MBR
For wastewater and sludge samples, mixed liquor sus- systems. Before PAC was added in one of two bioreactors,
pended solid (MLSS), mixed liquor volatile suspended solid the biomass concentrations were nearly identical at all
(MLVSS), chemical oxygen demand (COD), soluble COD, sampling points. The MLVSS concentrations remained
color, ammonia-nitrogen (NH3-N) and time to filtration stably in the two bioreactors at 5.5 ^ 0.2 g/L for one SRT
(TTF) were measured according to standard methods before PAC dosing.
(APHA-AWWA-WEF 1998). In this study, 50 ml sludge During the first SRT (25 days) period after PAC was
sample was used for each TTF test. Total organic carbon added into PAC-MBR reactor, the MLSS and MLVSS
(TOC) was measured by a TOC analyzer (Phoenix 8000, concentrations in the PAC-MBR increased gradually
USA). The difference of TOC between the filtrate passing whereas the MLSS and MLVSS concentrations in the
through a 1.2 mm filtration paper (GF/C, Whatman, USA) control MBR remained quite stable. The average MLVSS
and the permeate directly collected from the MBRs was concentration of the MBR and the PAC-MBR after 105 days
referred to as colloidal TOC. Soluble extracellular poly- operation was 6.1 ^ 0.4 g/L and 10.1 ^ 0.3 g/L, respect-
meric substances (EPS) were measured as supernatant after ively. As the dosage of PAC in the PAC-MBR remained at
the centrifugation of the sludge samples at 10,000 g for 2.0 g/L and the volatile content at 5508C of PAC is 96.7%,
30 min and calculated by summing the contents of the the MLVSS concentration in the PAC-MBR was expected
carbohydrate and protein fractions. Carbohydrate was to be 1.93 g/L higher than that in the MBR provided the
measured according to Dubois et al. (1956) and protein biomass growth in the control MBR and PAC-MBR were
was measured according to modified Lowry’s method the same. However, the average MLVSS concentration in
(Lowry et al. 1951). The bound EPS was measured as extract the PAC-MBR after 105 days operation was 4.0 g/L higher
after 1 hour heat extraction at 1008C (Chang & Lee 1998). than that in MBR. Some of the higher MLVSS concen-
The particle size of sludge was measured by a laser particle tration of the PAC-MBR sludge can be attributed to the
counter (Spectrex PC-2000, USA). adsorption and accumulation of incoming (or generated)
A yeast estrogen screen (YES) assay was used to soluble organic substances on the PAC surface. Results on
quantify estrogenic activity associated with E2 or EE2 in soluble COD data show a 31 mg/L difference between the
samples. Prior to the YES assay, samples were extracted control MBR and the PAC-MBR. It is also possible that
by using cyclohexane extraction for the enrichment of more PAC was added than extracted from the MBR on
estrogen concentrations. The methodology of YES assay a daily basis due to settling of the PAC in the reactor.
and cyclohexane extraction has been reported in the This could have resulted in a new equilibrium at the MLVSS
previous study (Yang & Cicek 2008). The average extraction levels observed in the PAC-MBR.

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018
 175 W. Yang et al. | Improving the performance of membrane bioreactors Water Science & Technology—WST | 62.1 | 2010

Overall performance of the MBR and PAC-MBR efficiencies are consistent with the results from previous
studies in activated sludge process (Cicek et al. 2007).
The comparison of overall performance in the MBR and
The synthetic estrogen EE2 appeared to be removed to a
PAC-MBR at steady state conditions was summarized in
lesser extent than E2, which is reasonable because of its
Table 1. The soluble COD of the mixed liquor in the
more stable chemical structure. PAC dosing increased the
PAC-MBR bioreactor was 47 ^ 3 mg/L, which was con-
removal efficiencies of both E2 and EE2 in the MBR system.
siderably lower than in the control MBR (78 ^ 5 mg/L).
In comparison with the control MBR, the removal rate of
This suggests that PAC is capable of enhancing the
E2 and EE2 in the PAC-MBR was increased by 3.4% and
adsorption of soluble COD to the PAC-sludge. On average,
15.8%, respectively.
31 mg/L more sCOD was adsorbed by the PAC onto
The impact of PAC dosing on EE2 removal was found
PAC-sludge, provided that the biodegradation was the
to be greater than E2 removal. This might be due to their
same in the two reactors. However, there was no significant
difference in hydrophobic properties, which significantly
difference in effluent COD between the control MBR and
affect their adsorption capacity to the sludge and PAC.
PAC-MBR. This indicates that PAC dosing did not remove
Natural estrogen E2 is considered to be weakly hydro-
the refractory COD that could not be removed by biological
phobic. The hydrophobicity of EE2 is approximately 10
degradation or membrane filtration.
times greater than E2. Therefore, removal of EE2 by
For nitrogen removal, both the control MBR and
sorption to sludge could play a more important role in the
PAC-MBR had complete nitrification with ammonia
removal mechanisms of estrogen in biological wastewater
removal rates higher than 99.5%. Ohnishi et al. (1998)
treatment processes. The enhanced adsorption capacity of
found PAC dosed in a wastewater reclamation system at a
PAC-MBR sludge could significantly increase the amount
dosage of 50 mg/L could stably obtain treated water with a
of EE2 being adsorbed onto the sludge.
color of less than 5 true color unit (TCU). This observation
was confirmed in this study. The color of the effluent from
the PAC-MBR was consistently lower than 5 TCU, which
TMP profile in the MBR and PAC-MBR
was lower than the average color of the effluent from the
control MBR of 15 TCU. The enhanced color removal is Trans-membrane pressure (TMP) is an important parameter
beneficial for the application of MBR for direct water reuse. to control the MBR operation as TMP is an indicator of
A natural estrogen E2 and a synthetic estrogen EE2 when membrane modules need be cleaned. To ensure the
were selected as target contaminants to demonstrate the membrane modules in the two MBR systems had identical
removal mechanisms of estrogen in MBR and PAC-MBR, filtration performances, the two investigated MBRs were
because E2 and EE2 have shown to be released from a operated for three membrane-cleaning periods before PAC
wide variety of wastewater treatment plants and possess was added into one of bioreactors. As shown in Figure 2, the
relatively high estrogenic activity (Cicek et al. 2007).
The average removal rate of E2 and EE2 by the control
MBR was 89.0% and 70.9%, respectively. Those removal

Table 1 | Summary of the performance of the MBR and PAC-MBR

MBR PAC-MBR

sCOD (mg/L) 78 ^ 5 47 ^ 3
COD removal rate (%) 95.5 ^ 1.1 95.6 ^ 1.2
NH3-N removal rate (%) .99.5 .99.5
Effluent color (TCU) 15 ^ 5 ,5
E2 removal rate (%) 89.0 ^ 0.7 92.4 ^ 1.4
EE2 removal rate (%) 70.9 ^ 1.2 86.7 ^ 1.1 Figure 2 | Profile of TMPs in the MBR and PAC-MBR. pNo PAC dosing before 80 days.

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018
 176 W. Yang et al. | Improving the performance of membrane bioreactors Water Science & Technology—WST | 62.1 | 2010

daily TMPs in the two MBR systems before PAC dosing Table 2 | Comparison of sludge characteristics between the MBR and PAC-MBR

were nearly the same and the increasing pathway followed

MBR PAC-MBR
the same pattern. The membrane provider recommended a
normal operation pressure range for the membranes of 7.0 Soluble EPS (mg/L) 186.8 ^ 67.6 74.5 ^ 34.1

to 40 kPa. However, during the start-up and acclimation Bound EPS (mg/L) 652.2 ^ 196.3 635.0 ^ 150.4

period, the permeate flux dropped down significantly when Colloidal TOC (mg/L) 17.0 ^ 4.2 6.5 ^ 1.3

TMPs reached 15 – 20 kPa. The membranes were cleaned by TTF (s) 86 ^ 18 54 ^ 8

soaking in 0.3% NaOCl for 3 hours when permeate flux Medium particle size (mm) 49.4 60.3

decreased significantly, even though the TMPs were still

relatively low. It is difficult to explain why the permeate flux are EPS excreted from cells (Chang & Lee 1998). As shown
decreased at low TMPs because the conditions of biomass in Table 2, the average soluble EPS of the PAC-MBR sludge
and membrane fibers at the start-up and acclimation period was 60.1% lower than the control MBR sludge. In contrast,
were changing and different from the stable operation. there was no significant difference in bound EPS between
After PAC was added in one of the two MBRs, the the PAC-MBR sludge and MBR sludge. This suggests that
profile of TMPs in PAC-MBR was quite different from the soluble EPS would have greater impacts on membrane
control MBR. Generally speaking, PAC dosing resulted in a fouling than would bound EPS. This might be due to the fact
slower rate of TMP increase in the PAC-MBR. Based on the that the bound EPS are mainly present with in the MLSS
results during four membrane cleaning cycles, the average flocs while the soluble EPS are the main components of
daily TMP increasing rates in the control MBR and colloidal substances in the sludge. Colloidal particles have
PAC-MBR were 0.467 kPa/d and 0.269 kPa/d, respectively. been widely recognized as being more responsible for
In practice, a slower TMP increase rate means a longer membrane fouling than MLSS flocs. It should be pointed
operating period between membrane cleaning events. out that bound EPS mentioned in this study referred to
Consequently, the maintenance cost for membrane cleaning the portion of carbohydrate and protein in sludge that could
could be reduced. be extracted by the method described in the previous
In addition, PAC dosing was found to be beneficial to section. It is possible that more EPS substances were
recover TMPs to a lower level after chemical cleanings accumulated in the PAC-MBR sludge but could not be
(Yang 2009). This could result in a longer lifespan of extracted by current extraction method due to the high
membrane fibers because membrane can be cleaned adsorption ability of PAC.
more often without compromising their performance. The results of colloidal TOC measurements showed
Consequently, the operating cost for membrane replace- that the average of colloidal TOC of the PAC-MBR sludge
ment could also be reduced. and MBR sludge was 6.5 mg/L and 17.0 mg/L, respectively.
At each sample point, the colloidal TOC of the PAC-MBR
sludge was consistently lower than the MBR sludge.
Effects of PAC dosing on sludge filterability
The lower colloidal TOC in PAC-MBR sludge might
characteristics
be associated with the slower rate of TMP increase in
To investigate the effects of PAC dosing on sludge the PAC-MBR.
characteristics, those characteristics associated with sludge PAC dosing may affect the floc size of the activated
filterability such as EPS, colloidal TOC, TTF and particle sludge in the bioreactor. In comparison with the MBR
size of the sludge in the MBR and PAC-MBR were analyzed sludge, the particle size distribution of the PAC-MBR sludge
and shown in Table 2. The results show that there was was shifted to a relatively higher range. The mean floc size
significant difference in the filterability characteristics of the PAC sludge was 60.3 mm, higher than that of the
between the MBR sludge and the PAC-MBR sludge. control MBR sludge 49.4 mm. This could be explained by
There are a number of studies which have shown the the fact that the dosed PAC could adsorb and coagulate
main foulants with regard to membrane fouling in MBRs dissolved organics, colloidal particles and free bacteria

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018
 177 W. Yang et al. | Improving the performance of membrane bioreactors Water Science & Technology—WST | 62.1 | 2010

onto the PAC-MBR sludge surface, thus the overall particle Cost analysis
distribution with PAC dosing was changed to a greater
Table 4 demonstrates operational parameters of a typical
particle size range. The larger particle sizes of sludge would
MBR with a capacity of 1,000 m3/d and an operating
be beneficial to alleviate membrane fouling during MBR
cost analysis of PAC dosing in such a MBR system at a
operation because larger particles typically generate a more
dosage of 2.0 g/L. The majority of the cost for PAC dosing
permeable cake layer and larger particles are also easier to
is the cost for PAC replenishment. The labour cost for
remove from the membrane surface. This was supported by
PAC dosing is negligible because the PAC replenishment
the fact that the average TTF of the PAC-MBR sludge was
can be done within half an hour every other day, even every
37.2% lower than the control MBR sludge.
five days. Previous studies (Park et al. 1999; Ng et al. 2006)
have shown PAC dosing could achieve significant flux
enhancement at a dosage as low as 1.0 g/L. If the dosage of
Effects of PAC dosing on the frequency of 1.0 g/L were applied for PAC dosing in MBRs, the cost
membrane cleaning for PAC replenishment would drop by 50% down to
9,235 US$/year.
The TMP profile in Figure 2 showed a slower rate of TMP
The cost for membrane maintenance primarily consists
increase in the PAC-MBR than the control MBR. This
of costs for membrane cleaning and membrane replace-
means PAC dosing could prolong the operating period
ment. When membrane flux decreases dramatically below
between membrane cleaning events. The membrane man-
the design value, membrane cleaning is required to recover
ufacturer recommends a membrane cleaning event by
the membrane permeability. Chemical cleaning has been
soaking membrane modules in 0.3% NaOCl solution for
extensively used for recovering membrane permeability.
3 – 5 hours when TMP reaches 60 kPa. Table 3 summarize
The choice of chemical agents for recovering membrane
the time for TMP reaching to 60 kPa (T) in the MBR and
flux depends on the nature of the foulants at different MBR
PAC-MBR in the four membrane cleaning cycles. The data
systems. There are no generic guidelines for this technique.
of T were obtained by exponential regression using the daily
Operators tend to rely on their experience and technical
TMPs in the MBR and PAC-MBR. The R-squared value
support from MBR providers or manufacturers. Acids,
(R 2) of each regression was shown in Table 3 as well.
bases, oxidants and surfactants are typically used agents for
As shown in Table 3, the average time for TMP reaching
membrane cleaning in MBR systems. Sodium hypochlorite
60 kPa in a membrane cleaning cycle in the PAC-MBR
was averagely 32.3% longer than the control MBR. In
Table 4 | Cost analysis of PAC dosing in MBRs
practice, membrane modules in MBRs need be cleaned
every 2 – 6 months (Yang et al. 2006). The prolonged Parameter Value Comments

membrane cleaning cycle would eventually prolong the MBR Capacity (m3/day) 1,000 Municipal wastewater
membrane lifespan and result in a lower cost for membrane treatment
replacement. HRT (hour) 6
SRT (day) 25
PAC dosage (g/L) 2.0 Same with the
bench-scale
Table 3 | Comparison of T between the MBR and PAC-MBR
experiment
Period between MBR PAC-MBR Cost for initial PAC 1,265 2.53 US$/kgp
membrane cleanings T (d) R2 T (d) R2 dosing (US$)

I 90 0.9663 122 0.9691 PAC replenishment 20

(kg/day)
II 66 0.9547 84 0.9011
Cost for PAC replenishment 18,469 2.53 US$/kgp
III 67 0.9662 94 0.958 (US$/year)
IV 61 0.9105 77 0.9453 p
CMR Online, http://www.the-innovation-group.com/chemprofile.htm, (March 9, 2009).

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018
 178 W. Yang et al. | Improving the performance of membrane bioreactors Water Science & Technology—WST | 62.1 | 2010

solution has been commonly used in commercial MBR In real applications, the additional costs generated by PAC
systems (Yang et al. 2006). Even with effective membrane dosing, e.g. higher aeration cost due to higher concentration
cleaning, the membrane modules will need to be replaced of MLVSS and cost for sludge management, should be
somewhere between five and ten years with the current considered in order to evaluate the economic feasibility of
technology (Gander et al. 2000; DeCarolis et al. 2007; PAC dosing in MBRs.
Fletcher et al. 2007). While the membrane price has
dramatically decreased over the past years, these modules
can still be classified as expensive (Yoon et al. 2004).
Table 5 summarizes a cost assessment of membrane CONCLUSIONS
maintenance in the MBR and PAC-MBR for municipal Two bench-scale MBRs, one without PAC dosing and the
wastewater treatment with a capacity of 1,000 m3/d. It was other with PAC dosing, were operated in parallel at similar
found that the annual depreciation for membrane replace- experimental conditions to examine the effects of PAC
ment accounts for over 97% of the cost for membrane dosing on the overall performance of MBRs. The following
maintenance in MBRs. PAC dosing could retard TMP conclusions can be drawn from the study:
increase and result in a longer lifespan of membranes,
(1) PAC dosing could improve the color removal in MBR
which can reduce annual cost for membrane maintenance
system. However, PAC dosing could not remove the
by about 25%. The benefit from reducing the operating cost
refractory soluble COD that could not be removed by
for membrane maintenance by PAC dosing in a 1,000 m3/d
biodegradation or membrane filtration. PAC dosing in
MBR is 14,320 US$ per year, which can offset 77.5%
the MBR increased the removal rate of E2 and EE2 by
operating cost for PAC replenishment at a dosage of
3.4% and 15.8%, respectively. The greater impact of
2.0 g/L. If 1.0 g/L PAC were applied in such a MBR system
PAC dosing on EE2 removal was due to its more
and it could achieve similar performance, the benefit of
hydrophobic property.
PAC dosing would exceed the operating cost for PAC
(2) PAC dosing significantly reduced the concentrations
dosing. It should be pointed out that the cost analysis in this
of soluble EPS and colloidal TOC in the PAC-MBR
paper was based on the bench-scale experimental data.
sludge. Also, the overall particle size distribution of the
sludge with PAC dosing was shifted to a greater
Table 5 | Cost analysis of membrane maintenance in the MBR and PAC-MBR particle size range. The average TTF of the PAC-MBR
sludge was 37.2% lower than the control MBR sludge.
Parameter MBR PAC-MBR
Those effects resulted in a slower rate of TMP increase
Treatment capacity (m3/day) 1,000
in the PAC-MBR during the operation of MBRs.
Operating flux (m3/m2 d) 0.25
(3) The cost analysis suggests that PAC dosing could
Surface area of membrane (m2) 4,000
reduce the operating cost for membrane cleaning
p
Capital cost for membrane (US$) 400,000
and/or membrane replacement. The operating cost
Cost for each membrane 369
for PAC dosing can be offset by the benefit from its
cleaning† (US$)
reducing the cost for membrane maintenance.
Membrane lifespan (year) 7 9.3
Cleaning frequency (times/year) 4 3
Cost for membrane cleaning (US$/year) 1,476 1,107
Depreciation for membrane replacement 57,143 43,192
(US$/year) ACKNOWLEDGEMENTS
Total cost for membrane maintenance 58,619 44,299 The authors would like to acknowledge Natural Sciences
(US$/year)
and Engineering Research Council of Canada (NSERC) for
p
Based on 100 US$/m2 (Yoon et al. 2004).
†
funding this research. We also thank Korea Membrane
Based on labour 2 £ 20 US$/h for 8 h, NaOCl (12.5%) price 0.64 US$/gallon, 2 £ 2 £ 3 m
tank for membrane cleaning. Separation Ltd. for providing membrane modules.

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018
 179 W. Yang et al. | Improving the performance of membrane bioreactors Water Science & Technology—WST | 62.1 | 2010

Korner, W., Spengler, P., Bolz, U., Schuller, W., Hanf, V.

REFERENCES & Metzger, J. W. 2001 Substances with estrogenic activity in
APHA-AWWA-WEF 1998 Standard Methods for the Examination effluents of sewage treatment plants in southwestern Germany.
of Water and Wastewater, 20th edition, Washington, 2 Biological analysis. Environ. Toxicol. Chem. 20(10),
DC, USA. 2142 –2151.
Chang, S. J., Jang, N. J., Yeo, Y. H. & Kim, I. S. 2006 Fate and Lowry, O. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. 1951
transport of endocrine-disrupting compounds (oestrone and Protein measurement with the folin phenol reagent. J. Biol.
17 beta-oestradiol) in a membrane bio-reactor used for water Chem. 193, 265 –275.
re-use. Water Sci. Technol. 53(9), 123– 130. Ng, C. A., Sun, D. & Fane, A. G. 2006 Operation of membrane
Chang, I. S. & Lee, C. H. 1998 Membrane filtration characteristics bioreactor with powdered activated carbon addition. Sep. Sci.
in membrane coupled activated sludge system-the effect of Technol. 41(7), 1447 –1466.
physiological states of activated sludge on membrane fouling. Ohnishi, M., Okuno, Y. & Ohkuma, N. 1998 Decoloration system
Desalination 120, 221–233. using rotating membrane UF module. Water Sci. Technol.
Cicek, N., Londry, K., Oleszkiewicz, J. A., Wong, D. & Lee, Y. 2007 38(6), 35– 43.
Removal of selected natural and synthetic estrogenic Park, H., Choo, K. H. & Lee, C. H. 1999 Flux enhancement with
compounds in a Canadian full-scale municipal wastewater powdered activated carbon addition in the membrane
treatment plant. Water Environ. Res. 79(7), 795 –800. anaerobic bioreactor. Sep. Sci. Technol. 34(14), 2781 – 2792.
DeCarolis, J., Adham, S., Pearce, W. R., Hirani, Z., Lacy, S. Remy, M., van der Marel, P., Zwijnenburg, A., Rulkens, W. &
& Stephenson, R. 2007 Cost trends of MBR systems for Temmink, H. 2009 Low dose powdered activated carbon
municipal wastewater treatment. Proceedings of the 80th addition at high sludge retention times to reduce fouling in
Annual Water Environment Federation Technical Exhibition membrane bioreactors. Water Res. 43(2), 345 –350.
and Conference (WEFTEC.07). San Diego, USA. Servos, M. R., Bennie, D. T., Burnison, B. K., Jurkovic, A., McInnis,
Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A. & Smith, F. R., Neheli, T., Schnell, A., Seto, P., Smyth, S. A. & Ternes,
1956 Colorimetric method for determination of sugar and T. A. 2005 Distribution of estrogens, 17 beta-estradiol and
related substances. Anal. Chem. 28, 350– 356. estrone, in Canadian municipal wastewater treatment plants.
Fletcher, H., Mackley, T. & Judd, S. 2007 The cost of a Sci. Total Environ. 336(1 –3), 155 –170.
package plant membrane bioreactor. Water Res. 41(12), Yang, W. & Cicek, N. 2008 Treatment of swine wastewater by
2627 –2635. submerged membrane bioreactors with consideration of
Gander, M., Jefferson, B. & Judd, S. 2000 Aerobic MBRs for estrogenic activity removal. Desalination 231, 200 –208.
domestic wastewater treatment: a review with cost Yang, W., Cicek, N. & Ilg, J. 2006 State-of-the-art of membrane
considerations. Sep. Purif. Technol. 18(6), 119– 130. bioreactors: worldwide research and commercial applications
Holbrook, R. D., Novak, J. T., Grizzard, T. J. & Love, N. G. 2002 in North America. J. Membr. Sci. 270(1–2), 201–211.
Estrogen receptor agonist fate during wastewater and biosolids Yang, W., Paetkau, M. & Cicek, N. 2009 Removal mechanisms of
treatment processes: a mass balance analysis. Environ. Sci. estrogens in membrane bioreactors with powdered activated
Technol. 36(21), 4533 – 4539. carbon dosing. Proceedings of the 82nd Annual Water
Hu, A. Y. & Stuckey, D. C. 2007 Activated carbon addition to a Environment Federation Technical Exhibition and Conference
submerged anaerobic membrane bioreactor: effect on (WEFTEC.09), Orlando, USA, October pp. 10 – 14.
performance, transmembrane pressure, and flux. J. Environ. Yang, W. 2009 Investigation of endocrine disputing compounds
Eng.—ASCE 133(1), 73 – 80. in membrane bioreactor and UV processes. Ph.D. Thesis,
Joss, A., Andersen, H., Ternes, T., Richle, P. R. & Siegrist, H. 2004 University of Manitoba.
Removal of estrogens in municipal wastewater treatment Yoon, T. I., Lee, H. S. & Kim, C. G. 2004 Comparison of pilot scale
under aerobic and anaerobic conditions: consequences performances between membrane bioreactor and hybrid
for plant optimization. Environ. Sci. Technol. 38(11), conventional wastewater treatment systems. J. Membr. Sci.
3047 –3055. 242(1– 2), 5–12.

Downloaded from https://iwaponline.com/wst/article-pdf/62/1/172/446145/172.pdf

by guest
on 01 November 2018