Bioresource Technology 185 (2015) 125133

Contents lists available at ScienceDirect

Bioresource Technology
journal homepage: www.elsevier.com/locate/biortech

Effects of ZnO nanoparticles on wastewater treatment and their removal

behavior in a membrane bioreactor
Magdalene Tan, Guanglei Qiu, Yen-Peng Ting
Department of Chemical and Biomolecular Engineering, National University of Singapore, 4 Engineering Drive 4, Singapore 117585, Singapore

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 Long-term exposure of ZnO NPs

caused moderate deterioration in

pollutants removal.
 ZnO NPs altered activated sludge
properties and caused higher
membrane fouling.
 Shifts in bacteria community
structure were noted but with
unaffected diversity.
 ZnO NPs were removed effectively in
MBR mainly via sludge sorption.
 Membrane ltration contributed
signicantly to ZnO NPs removal.

a r t i c l e

i n f o

Article history:
Received 18 December 2014
Received in revised form 18 February 2015
Accepted 24 February 2015
Available online 28 February 2015
Keywords:
Membrane bioreactor
ZnO nanoparticles (ZnO NPs)
Extracellular polymeric substances (EPS)
Membrane fouling
Bacteria community dynamics

Long term loading of

ZnO NPs

COD, nitrogen removal

Activated sludge properties

Eco-toxicity on
microorganisms

Emerging contaminants
ZnO NPs

Membrane bioreactor

Fate and removal

behavior of ZnO NPs

a b s t r a c t
Long-term effects of ZnO nanoparticles on the system performance of an MBR were investigated together
with their removal behavior in the system. Continuous operation over 242 days showed that ZnO NPs at
both 1.0 and 10.0 mg/L caused moderate deterioration in the removal of COD, nitrogen and phosphorus.
Denitrication was affected upon the exposure but recovered subsequently. Although no signicant acute
effect on ammonia-oxidization was observed, permanent inhibition occurred after long-term exposure.
Nitrite-oxidization was not affected even with 10.0 mg/L ZnO NPs. Signicant changes were observed
in activated sludge properties which resulted in severe membrane fouling. Although ZnO NPs caused
changes in the bacteria community structure, the diversity however remain unchanged. ZnO NPs was
removed effectively in the MBR (>98%) with biosorption being a major removal mechanism.
Membrane ltration also played an important role (20% of the total removal) especially at high ZnO
NPs concentrations (around 10.0 mg/L).
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Over the past decades, membrane bioreactors (MBR) have
emerged as an effective solution to the treatment of various wastewaters into high quality efuent. MBRs have several advantages
such as high efuent quality, high treatment capacity, low sludge
production and small footprint, and are increasingly applied in
municipal and industrial wastewater treatment (Rajesh Banu
Corresponding author. Tel.: +65 65162190; fax: +65 67791936.
E-mail addresses: cheqg@nus.edu.sg (G. Qiu), chetyp@nus.edu.sg (Y.-P. Ting).
http://dx.doi.org/10.1016/j.biortech.2015.02.094
0960-8524/ 2015 Elsevier Ltd. All rights reserved.

Bacterial community
dynamics

et al., 2009, 2011; Qiu et al., 2013a). Currently, there are at least
50 individual MBR membrane suppliers and hundreds of large-scale MBR plants (with treatment capacity >10,000 m3/d) in operation worldwide (Judd, 2011). The capacity and the broadening
application of MBR are expected to increase due to more stringent
regulations and increased water reuse initiatives.
Many consumer products, such as sunscreen, cosmetics, paints
and coatings, etc., commonly incorporate zinc oxide nanoparticles
(ZnO NPs) due to its excellent UV absorption and reective properties. In 20032004, global production of ZnO NPs and TiO2 NPs in
sunscreen products was estimated at approximately 1000 tons

126

M. Tan et al. / Bioresource Technology 185 (2015) 125133

(Borm et al., 2006). By 2012, the annual production of ZnO NPs

increased to 1600 tons in European Union countries alone (Sun
et al., 2014). Rapid increase in the use of ZnO NPs would ultimately
lead to their growing release into municipal wastewater. Modeling
studies suggested a ZnO NPs concentration of 0.341.42 lg/L in
treated wastewater in Europe in 2008 (Gottschalk et al., 2009),
which was increased to 1.721 lg/L in 2012 (Sun et al., 2014).
Taking into account the removal efciency (around 93%) used in
the modeling, an environmentally relevant concentration of ZnO
NPs in wastewater would be around 24300 lg/L (Sun et al.,
2014). Considering the rapid and continuous increase in the production and use of ZnO NPs, its environmental concentration is
expected to inevitably increase. In sludge treated soil, for instance,
the estimated rate of increase is 1.63.3 lg/kg/yr (Gottschalk et al.,
2009). In municipal wastewater, its concentration is likely to be in
the mg/L level in the next few years.
ZnO NPs have been described as some of the most toxic nanomaterials. Its antibacterial properties have been reported (Zhang
et al., 2007), with the half maximal effective concentration
(EC50) ranging from tens of lg/L to several mg/L (Farr et al.,
2009). As these NPs enter wastewater streams and end up at treatment plants, they may inhibit bacterial activities in the activated
sludge and affect the treatment efciency. Studies have been conducted to examine the potential impacts of ZnO NPs on wastewater
treatment in conventional activated sludge (CAS) processes (Zheng
et al., 2011; Puay et al., 2015). Hou et al. (2013) showed that ZnO
NPs at 5.0 mg/L signicantly inhibit nitrifying activity and result
in reduced NH+4-N removal in a sequencing batch reactor (SBR).
Liu et al. (2011) reported a half maximal inhibitory concentration
(IC50) of ZnO NPs of 13.1 mg/L on ammonia-oxidizing bacteria.
The NPs also affected the settleability of the activated sludge, with
subsequent decreases in nitrogen and phosphorus removal in an
SBR study (Puay et al., 2015). Compared to gravitational settling
process in CAS, membrane separation in the MBR effectively
retains both sludge ocs and dispersed bacteria cells (Qiu et al.,
2013b), which may benet the enrichment of high-NPs-resistant
bacteria and thus lead to a better performance against the impacts
of NPs.
As the last barrier before the discharge of treated efuent into
the water bodies, wastewater treatment processes play a crucial
role in controlling the entrance of these emerging pollutants into
the environment. Membrane separation in the MBR process may
result in enhanced removal of ZnO NPs although the potential
effect of the accumulation of ZnO NPs remains to be understood.
Additionally, membrane fouling as a major challenge in MBR is
widely recognized to be governed by soluble microbial products
(SMP) and extracellular polymeric substances (EPS) (Lin et al.,
2014; Wang et al., 2014a). Studies have suggested that shortterm ZnO NPs exposure may result in signicant increase in
the sludge EPS and SMP production (Mei et al., 2014).
Signicant increase in EPS concentration was also observed in
long-term exposure of other NPs (e.g. Ag NPs) (Zhang et al.,
2014). Thus, there is added impetus to understand the long-term
effects of ZnO NPs on activated sludge properties and membrane
fouling, as well as their removal behaviors in the MBR
process.
In this work, a lab-scale MBR for the treatment of municipal
wastewater was set up and operated over 242 days. The longterm effects of 1.0 mg/L and 10.0 mg/L of ZnO NPs on the
system performance in terms of pollutant removal efciencies,
activated sludge properties and membrane fouling were investigated. The fate and removal behaviors of ZnO NPs in the MBR
were also examined. In addition, the effects of ZnO NPs on
the bacterial community dynamics in the system were also
investigated.

2. Methods
2.1. Experimental set-up
A schematic diagram of the MBR set-up is shown in Fig. 1. The
MBR system comprised a rectangular tank with an operating volume of 5.034 L with a at-sheet PVDF membrane module (with a
mean pore size of 0.1 lm and the effective area of 0.018 m2)
(Newton and Stokes, Singapore) submerged in the tank. The bioreactor was aerated at 15.0 L/min with an air diffuser xed directly
below the membrane module and was inoculated with activated
sludge from a local MBR plant (Ulu Pandan, Singapore). The HRT
and SRT of the MBR system were maintained at 12 h and 30 days
respectively. The transmembrane pressure (TMP) was monitored
using a pressure gage (Cole Palmer, USA). The system was operated
for 68 days to achieve steady state before ZnO NPs were added into
the inuent wastewater to achieve a nal concentration of 1.0 mg/
L (Zn) from Day 69. The ZnO NPs concentration was further
increased to 10.0 mg/L (Zn) from Day 161. New membranes were
used at the beginning of the experiment and were replaced on
Day 69 when ZnO NPs were added. Ex-situ physical cleaning of
the membranes was carried out on Day 189 when the TMP exceeded 20 kPa using sponge sweeping before the membranes were
ushed with tap water.
2.2. Wastewater and ZnO NPs
Synthetic municipal wastewater was prepared daily as inuent,
using tap water supplied by the Public Utilities Board, Singapore.
Glucose, NH4Cl and KH2PO4 were added to give the feed COD,
NH+4-N and PO3
4 -P concentrations of 400.0 mg/L, 40.0 mg/L, and
8.0 mg/L respectively. In addition, the synthetic wastewater also
contained 1.0 g/L of NaHCO3, 19.3 mg/L CaCl22H2O, 71.0 mg/L
MgSO47H2O, 17.4 mg/L FeSO47H2O, 0.07 mg/L CuCl22H2O,
0.13 mg/L MnCl24H2O, 0.13 mg/L ZnSO47H2O, 0.03 mg/L
Na2MoO42H2O, 0.025 mg/L H3BO3 and 0.033 mg/L KI.
Commercially-produced ZnO NPs as a suspension of 50 wt% ZnO
NPs in water were purchased from SigmaAldrich, USA. Analysis of
particle size and zeta potential were performed using a Zetasizer
(Malvern Instruments, USA). ZnO NPs suspension was rst diluted
to 100 mg/L using Milli-Q water, followed by sonication for 1 h
using Elmasonic S30H ultrasonicator (Elma GmbH & Co,
Germany). The particle size was determined to be
65.70 38.46 nm with a corresponding zeta potential of 30.5 mV.
2.3. Analyses of wastewater quality and sludge properties
Fresh wastewater sample was collected daily from the inuent
and the efuent. The NH+4-N, PO3
4 -P and COD concentrations were

Pump

Membrane
module

Feed tank

Effluent

Pump
Air Pump
Fig. 1. Schematic diagram of the MBR system.

M. Tan et al. / Bioresource Technology 185 (2015) 125133

analyzed according to Standard methods (APHA, 1999). Total nitrogen (TN) concentration was calculated as the sum of NH+4-N, NO
2 -N
and NO
3 -N concentration. All chemical tests were done at least in
duplicates, with the exception of the measurement of settled
sludge volume used in the calculation of the Sludge Volume
Index (SVI). Analysis of variance (ANOVA) test was performed
using SPSS 13.0 (SPSS Inc., USA). A p-value of less than 0.05 was
taken to be statistically signicant.
As ZnO NPs are slightly soluble, both Zn2+ and total Zn concentration in the inuent, supernatant and efuent wastewater were
analyzed. 150 ml of mixed liquor was collected daily and left to
stand in a measuring cylinder. The supernatant (10 ml) was collected about 1 cm beneath the water surface after 30 min of settling. Analysis of the soluble Zn2+ was conducted using an
Inductively Coupled Plasma-Mass Spectrometer (ICP-MS) (Agilent
Technologies, USA). Acid digestion of the sample wastewater was
conducted in a process similar to 3030E of the Standard Methods
(APHA, 1999). 5 ml of the sample was acidied with 1 ml of trace
metal grade nitric acid and reuxed at 105 C for 2 h. The resultant
solution was ltered through a 0.45 lm lter membrane before
measurement. The released Zn2+ due to the dissolution of ZnO
NPs was determined according to literature (Zheng et al., 2011).
In addition to the water samples, Zn content in the activated
sludge was also analyzed after acid digestion. 10 ml of mixed
liquor was rst centrifuged at 5000 rpm for 5 min and the supernatant removed before the sample was washed with Milli-Q water.
5 ml of nitric acid was then added to the residue and reuxed at
105 C for 2 h, followed by ltration through a 0.45 lm lter membrane. The resultant solution was diluted to a nal volume of 10 ml
using Milli-Q water.
The amount of SMP and EPS produced by the sludge bacteria
was analyzed as protein and polysaccharide concentrations. SMP
was obtained by centrifuging 10 ml of mixed liquor at
12,000 rpm for 10 min and ltering the supernatant through a
0.45 lm membrane lter. EPS was then extracted via thermal heating method. Milli-Q water was added to the sludge residue before
the sample was heated in a water bath at 80 C for 30 min. The
resultant solution was then centrifuged again at 12,000 rpm for
10 min and the supernatant ltered through a 0.45 lm membrane
lter to obtain the EPS.
Protein concentration was determined using a modied Lowry
method, with Bovine Serum Albumin as standard. Polysaccharide
content was determined using the phenolsulfuric acid method,
with glucose as standard (Qiu and Ting, 2014b).
2.4. Sludge sampling and DNA extraction
10 ml of mixed liquor was collected and immediately frozen at
20 C. DNA was extracted from the samples with a QIAamp DNA
Mini Kit (Qiagen, Valencia, CA).
2.5. Polymerase chain reaction (PCR)
In order to increase the yield of PCR products and to facilitate
the denaturing gradient gel electrophoresis (DGGE) analyses, a
nested PCR technique was applied (Qiu and Ting, 2013). For the
total bacterial community, the 16S rRNA genes were amplied
from the DNA extracts using universal primers 27F and 1492R, following a temperature cycling conditions: pre-incubation at 95 C
for 2 min, followed by 25 cycles of 95 C for 1 min, 62 C for
1.0 min, and 72 C for 1.5 min; and a nal elongation at 72 C for
10 min. A nested PCR was then performed on the PCR products
obtained from previously described primers with a second primer
pair 357F-Clamp and 518R, following a cycling program: Pre-incubation at 95 C for 2 min, followed by 30 cycles of 95 C for 1 min,
60 C for 1.0 min, and 72 C for 45 s; and a nal elongation at 72 C

127

for 10 min. All the PCR amplications were carried out in a total
volume of 50 ll in 200 ll tubes using a DNA thermocycler (Bio
Rad, USA). The PCR mixture contained 1.25 U of Taq polymerase
(Promega, USA), 1 PCR buffer, 2 mM MgCl2, 0.5 lmol of each primer, each deoxynucleoside triphosphate at a concentration of
200 lM, and 40 ng of template DNA.
2.6. Denaturing gradient gel electrophoresis
DGGE was performed using a D-Code System (Bio Rad, USA)
maintained at a constant temperature of 60 C in 1 TAE buffer.
PCR products were loaded onto 8% (w/v) polyacrylamide gels
(37.5:1, acrylamide/bisacrylamide) using a denaturing gradient
ranging from 35% to 60% denaturant (7 mol urea and 40% formamide in the 1 TAE buffer constituted 100% denaturant) (Qiu
et al., 2013a). Gels were run at 80 V for 12 h and stained with
SYBR Gold nucleic acid stain (Invitrogen, USA). The resultant
image was analyzed using Quantity One 4.6.2 software (Bio-Rad,
USA) to obtain the ngerprint patterns and to calculate the band
similarities. ShannonWiener diversity index (H0 ) was used to
evaluate the variation of structural diversity and species richness
of bacterial communities of the different samples (Qiu et al.,
2013b) as expressed in Eq. (1):

H0 

pi lnpi

where pi is the ratio of i-th group of bacteria to total community.

3. Results and discussion
3.1. Effects of ZnO NPs on system performance
Fig. 2 shows the removal of COD, nitrogen and phosphorus
before and after ZnO NPs addition. Moderate deterioration was
observed in COD removal (Fig. 2a). The average COD removal
was 98.5% prior to ZnO NPs dosage and decreased to 97.5%
(p = 0.000) with continuous exposure to 1.0 mg/L of ZnO NPs for
91 days (from Day 69 to Day 160), which decreased further to
96.3% (p < 0.005) with 10.0 mg/L ZnO NPs (from Day 161 onwards).
The results corroborate studies where both short-term and longterm exposure of 1.0 mg/L of ZnO NPs did not signicantly impact
COD removal (Hou et al., 2013; Puay et al., 2015; Wang et al.,
2014b), suggesting that ZnO NPs at current environmentally relevant concentrations were unlikely to affect COD removal in both
the MBR and CAS processes. However, remarkable impacts were
indeed observed at much higher ZnO NPs concentrations. For
instance, a 240-h study with a Zn loading rate of 0.83 mg/min in
a simulated CAS process showed a mean COD removal efciency
of 71 7%, which was signicantly lower than in the control
(80 5%) (Musee et al., 2014). Huang et al. (2013) also reported
that COD removal was inhibited in batch experiments with MBR
sludge in the presence of 68 mg/L ZnO NPs.
NH+4-N removal in the MBR system was consistently high even
after 91 days of continuous exposure to 1.0 mg/L ZnO NPs
(Fig. 2b). This corroborate a recent study where no major inuence
on nitrication in MBR was observed with long term exposure of
1.0 mg/L ZnO NPs (Wang et al., 2014b). However, a relatively
remarkable decrease (from 99.3% to 97.2%, p < 0.05) was observed
when ZnO NPs concentration was increased to 10.0 mg/L. For TN
removal (Fig. 2c), ANOVA analysis suggested signicant differences
(p < 0.05) among prior ZnO NPs exposure, exposure to 1.0 mg/L
ZnO NPs and 10.0 mg/L ZnO NPs (with average removal of 89.9%,
87.2% and 85.2% respectively). However, the decrease was less
drastic than that observed in a short-term (4.5 h) study in a SBR,
in which the TN removal decreased from 81.5% to 75.6% with
10.0 mg/L ZnO NPs (Zheng et al., 2011), as well as in a long-term

M. Tan et al. / Bioresource Technology 185 (2015) 125133

10 mg/L ZnO NPs

700
Removal efficiency

500

85

400

80

300

75

200

70

100

65

0
20

40

60

1 mg/L ZnO NPs

Influent
Removal efficiency

20

0
0

20

40

60

80 100 120 140 160 180 200 220 240

Operation time, days
1 mg/L ZnO NPs

60
Influent
Effluent
Removal efficiency

90

Nitrite effluent

Nitrate effluent

Removal efficiency

70
60

50

40

30

20

10

20
PO43--P, mg/L

Ammonia effluent

TN removal efficiency, %

80

20

40

60

80

100 120 140 160 180 200 220 240

Operation time, days

40

12

30

20

10
0
0

50

16

10 mg/L ZnO NPs

24

100

40
20

b
10 mg/L ZnO NPs

Effluent
Nitrification rate

80 100 120 140 160 180 200 220 240

9
Effluent concentration, mg/L

60
30

10

Operation time, days

10

80

60
0

100

40

NH4+-N, mg/L

Effluent

10 mg/L ZnO NPs

50
Removal efficiency, %

90
Influent

1 mg/L ZnO NPs

60

95

600
COD, mg/L

100

Removal efficiency, %
Nitrification rate, 0.1mg/L/h

1 mg/L ZnO NPs

800

Removal efficiency, %

128

20

40

60

80

100 120 140 160 180 200 220 240

Operation time, days

Fig. 2. Pollutants ((a) COD, (b) NH+4-N, (c) TN, (d) PO3
4 -P) removal in MBR before and after dosing ZnO NPs.

SBR study where 23.7% inhibition in nitrication was observed at

ZnO NPs concentration of 5.0 mg/L (Hou et al., 2013). The reason
could be that the initial long-term exposure to 1.0 mg/L ZnO NPs
allowed the acclimation of sludge bacteria to the toxic effects of
ZnO NPs, thus resulting in a less signicant response when the concentration was increased to 10.0 mg/L. Additionally, the relatively
higher MLSS concentration in the MBR system may bestow higher
resistance to NPs since toxicity effects of compounds are largely
related to the dosage-per-unit of subjects.
A closer examination of the efuent concentrations may provide more details on the effect of ZnO NPs on different bacteria
activities (Fig. 2c). At 1.0 mg/L ZnO NPs, the decrease in TN removal
was due to the signicant increase (p = 0.000) in efuent NO
3 -N,
suggesting the inhibition of the denitrifying activity. However, at
the later part of the 1.0 mg/L phase (from Day 102 onwards), the
efuent NH+4-N started to show an increasing trend, suggesting
the onset of inhibition of the ammonia-oxidizing activity. This
observation corroborates previous studies which showed that
1.0 mg/L of ZnO-NPs did not acutely impact nitrication.
However, inhibitory effect became apparent after long-term exposure (Puay et al., 2015). The efuent NO
3 -N decreased after Day
130, suggesting the acclimation of denitrifying bacteria. When
ZnO NPs was increased to 10.0 mg/L, the decrease in TN removal
was again due to the signicant increase (p = 0.000) in efuent
NH+4-N, yet indicating the inhibition of ammonia-oxidizing activity.
Efuent NO
3 -N level experienced a less evident increase and
peaked at Day 224 before decreasing. The efuent NO
2 -N concentration remained low (<0.5 mg/L) throughout the experiment and
suggested that ZnO NPs apparently did not affect nitrite-oxidizing
activity. These observations on the effects on nitrifying and denitrifying bacteria were dissimilar to the study of Zheng et al. (2011)
where there was no effect on NH+4-N removal but signicant

increase in efuent NO

3 -N (and hence inhibition of denitrication)
under short-term exposure of 10.0 mg/L and 50.0 mg/L ZnO NPs
was reported. However, the present study corroborates a longterm SBR study by Hou et al. (2013). Liu et al., 2011 suggested that
the IC50 of ZnO NPs for ammonia-oxidizing bacteria is 13.1 mg/L.
Long-term exposure may lead to the accumulation of ZnO NPs in
the wastewater treatment system and nally result in the ZnO
NPs concentrations exceeding the tolerance limits of bacteria over
time. These results suggest that the effects of ZnO NPs on bacterial
activities are highly dependent on exposure time. Additionally, different bacteria species (even with the same function, e.g. denitrication) have different tolerance to ZnO NPs (Chen et al., 2014). A
complex bacteria community (as in the activated sludge) may have
a higher ability to acclimatize to ZnO NPs in the long-term through
community dynamics. This could possibly explain the change in
denitrifying activity with exposure to different concentration of
ZnO NPs from Day 130 to Day 222.
The trend in PO3
4 -P removal (Fig. 2d) upon exposure to different ZnO NPs concentrations is particularly noteworthy. Prior to
addition of ZnO NPs, the removal efciency was 47.5%, which
decreased signicantly (p = 0.000) to 34.3% during exposure to
1.0 mg/L ZnO NPs, due to the inhibition effects of ZnO NPs on bacteria growth and phosphorus uptake. The observed biomass yield
(Yobs) in the MBR (Fig. 3a) was an average of 0.13 gMLSS/gCOD during the control period and decreased to 0.06 gMLSS/gCOD after
29 days exposure to 1.0 mg/L ZnO NPs. Yobs however increased
back to an average of 0.12 gMLSS/gCOD by the end of the 1.0 mg/
L ZnO NPs dosing, suggesting an initial inhibition effect of ZnO
NPs on bacteria growth and a subsequent recovery due to the
acclimatization of the bacteria. ZnO NPs has been shown to inhibit
the respiration of aerobic biomass (Hou et al., 2013). High concentrations (10.0 and 50.0 mg/L) of ZnO NPs were also reported to

M. Tan et al. / Bioresource Technology 185 (2015) 125133

1 mg/L ZnO NPs

100
90

12

80
10

70

60

3.2. Effects of ZnO NPs on sludge characteristics and membrane fouling

SVI, ml/g MLSS
MLVSS/MLSS, %

MLSS and MLVSS, g/L

Yobs, 0.1gMLSS/gCOD

the system performance during the operation of the bioreactor

may cause intrinsic uctuation in the removal efciencies.

10 mg/L ZnO NPs

14

50
6

40
30

20
2

10
0

0
0

20

40
MLSS

60

80

100 120 140 160 180 200 220 240

Operation time, days

MLVSS
Yobs
SVI
MLVSS/MLSS

1 mg/L ZnO NPs

90

10 mg/L ZnO NPs

-20

80

-10

70

10

40

20

30

30

20
40

10
0

50
0

20

40

60

80

Operation time, days

Proteins

Polysaccharides

Proteins + Polysaccharides

1 mg/L ZnO NPs

240

10 mg/L ZnO NPs

TMP

-20
-10

200
EPS, mg/g MLVSS

100 120 140 160 180 200 220 240

160

10
120
20
80

TMP, kPa

SMP, mg/g MLVSS

50

TMP, kPa

60

30

40

40

0
0

129

20
Proteins

40

50
100 120 140 160 180 200 220 240 260
Operation time, days
Polysaccharides
Proteins + Polysaccharides
TMP

60

80

Fig. 3. Changes in (a) MLSS, MLVSS, SVI, (b) SMP and (c) EPS during MBR operation.

inhibit the phosphate release, uptake and net removal (Zheng et al.,
2011). However, in this study, when ZnO NPs was increased to
10.0 mg/L, the average PO3
4 -P removal increased to 47.4%
(p = 0.000), which was similar to pre-ZnO NPs dosage level. This
increase was due to the reaction between ZnO NPs and phosphate
which resulted in the formation of zinc-phosphate and other larger
phosphate complex substances (Qiu and Ting, 2014a). This is plausible, considering the ndings that ZnO NPs in wastewater may
undergo transformation due to complexation/precipitation reactions with phosphate and other ions (Lombi et al., 2012).
Generally, from the results of this work and other studies,
although adverse effects of ZnO NPs were observed on different
functions and bacteria activities in wastewater treatment, it is
unlikely that ZnO NPs at environmentally relevant concentrations
(i.e. below 1.0 mg/L) may cause dramatic effects on the system performance in the MBR. This is especially since potential changes in

3.2.1. MLSS, MLVSS, SVI and sludge morphology

Changes in morphology of the activated sludge were observed
using SEM (Fig. S1). The sludge inoculum showed a relatively compact structure. Similar morphology was observed up to Day 69.
However, upon exposure to ZnO NPs (Day 161 and Day 240), the
structure of the activated sludge became loose with numerous
small aggregates of bacterial cells scattered on the surface which
resulted in dispersed ocs and poorer sludge settleability. SVI values (Fig. 3a) showed an increase from 40 ml/gMLSS (Day 69) to
56 ml/gMLSS (Day 161) after 91 days continuous exposure to
1.0 mg/L ZnO NPs, and further increased to 75 ml/gMLSS (Day
242) after the ZnO NPs concentration was increased to 10.0 mg/L.
This observation corroborates a study which suggested that activated sludge ocs subjected to long-term exposure of 100.0 mg/L
TiO2 NPs resulted in more stable suspension and poorer occulability owing to the predominant repulsive forces between the
NPs (Yang et al., 2013). Additionally, decrease in the MLVSS/MLSS
ratio (by 8%) was observed from the control period up to 10.0 mg/L
ZnO NPs dosage (Fig. 3b), suggesting the deposit of non-volatile
inorganic solids (ZnO NPs and other Zn species) within the activated sludge (Do et al., 2012).
3.2.2. Soluble microbial products (SMP) and extracellular polymeric
substances (EPS)
Signicant effects (p < 0.05) on SMP were observed in the presence of 1.0 mg/L ZnO NPs (Fig. 3b). Both protein and polysaccharides increased immediately in response to 1.0 mg/L ZnO NPs and
peaked at 44.5 mg/g MLVSS and 33.3 mg/g MLVSS (around Day
100) respectively, before decreasing and stabilizing by Day 160.
However, 10.0 mg/L ZnO NPs resulted in another round of
increased in SMP from Day 161 onward. Mei et al. (2014) also
reported that short-term exposure of ZnO NPs (0297.4 mg/L)
resulted in increased SMP production. SMPs can be subdivided into
substrate-utilization associated products (produced during substrate metabolism) and biomass-associated products (formed during hydrolysis of bound EPS). The increase in SMP contents most
probably originated from hydrolysis of bound EPS, as indicated
by the change in the activated sludge morphology.
There was signicant increase in protein content (p = 0.000) in
the EPS (Fig. 3c), from an average of 73.1 mg/gMLVSS (Days 1
68) to 101.1 mg/gMLVSS (Days 69160). This however decreased
to an average of 70.2 mg/gMLVSS during exposure to 10.0 mg/L
ZnO NPs (Days 161242). No signicant change (p > 0.05) was
observed in the average polysaccharides concentration throughout
the entire duration. EPS can function as a physical barrier in protecting the cells by trapping the NPs (e.g. Ag NPs) and reducing
their toxicity (Kiser et al., 2010). The reaction of sludge bacteria
in EPS production in response to ZnO NPs exposure is thus
hypothesized as follow. When the activated sludge was rst
exposed to ZnO NPs, the bacteria responded by increasing EPS production, presumably as a defense mechanism against nanotoxicity.
At a high enough EPS concentration, NPmicrobial cell interaction
was impeded, thereby reducing the need for continued EPS production. Furthermore, it has been shown that organic matter
increases the stabilization of NPs in the aqueous state (Diegoli
et al., 2008). Made up of large amount of soluble organic matter,
SMP may have a similar effect on the stability of NPs. The increase
in the SMP content (Fig. 3b) thus increases the amount of NPs in
the aqueous solution, and reduces their direct contact with the
activated sludge and bacteria (as conrmed in the Zn content analysis, see later discussion in Section 3.4). This could also be a reason

130

M. Tan et al. / Bioresource Technology 185 (2015) 125133

why EPS concentration did not further increase but instead

decreased at a higher ZnO NPs concentration (10.0 mg/L).
3.2.3. Membrane fouling
Variation in TMP was also compared against SMP and EPS concentrations (Fig. 3b and c). TMP remained generally constant during the initial 40 days before a steady increase occurred until Day
63 followed by a sharp increase until Day 68. This prole follows
the typical three-stages of fouling (i.e. initial passive conditioning;
slow and steady fouling, and a rapid TMP jump) under constant
ux operation (Le-Clech et al., 2006). Although a replacement of
the membrane upon the dosing of 1.0 mg/L ZnO NPs resulted in
the drop in TMP on Day 69, its increase after dosing 1.0 mg/L
ZnO NPs was much faster than that during the control period as
it took only 47 days (i.e. from Day 69 to Day 116) to reach the
pre-ZnO NPs exposure level. During this period, total SMP and
EPS concentration also showed steady increase, suggesting a correlation with the membrane fouling rate. Both EPS and SMP are recognized as major causes of membrane fouling (Wang et al., 2014a).
Their specic effects on the extent of membrane fouling are not
easily distinguished from each other since SMP is generally regarded as a soluble constituent of EPS and both have highly complicated fouling behaviors (Drews et al., 2008). The fouling mechanism of
EPS and SMP are related to surface adhesion, pore clogging, gel/cake formation and induced osmotic effect. Initial colonization of
EPS and SMP on the membrane surface also facilitates subsequent
adhesion of other foulants (Lin et al., 2014).
The addition of 10.0 mg/L ZnO NPs on Day 161 was followed by
a continual sharp increase in TMP, rising to a peak on Day 189
before membrane cleaning was performed. This increase again corresponded with the increase in EPS and SMP. In the nal 43 days,
the increase in TMP was similar to that during the period of
1.0 mg/L ZnO NPs dosage, suggesting a similar membrane fouling
rate although the EPS and SMP production were lower than that
during 1.0 mg/L ZnO NPs dosage. This could be due to changes in
activated sludge property upon prolonged exposure to ZnO NPs,
with decreased oc size and sludge settleability. Smaller ocs possess higher attractive interaction energy per unit mass when in
contact with the membrane surface (Qiu and Ting, 2014b), and
consequently deposit more easily, thus resulting in higher membrane fouling. Additionally, irreversible fouling that was not effectively removed by physical cleaning may have also accelerated
membrane fouling during this period (Wang et al., 2014a).
3.3. Effects of ZnO NPs on bacterial community dynamics
DGGE ngerprint patterns (Fig. 4a) show bacteria species (S7,
S15, S21, S33) which were tolerant at both low (1.0 mg/L) and high
(10.0 mg/L) ZnO NPs concentration. However, some species (e.g.
S11) decreased with the addition of 1.0 mg/L ZnO NPs and completely disappeared with the addition of 10.0 mg/L ZnO NPs, suggesting their elimination due to the toxic effects of ZnO NPs. In
contrast, others (e.g. S33) decreased at 1.0 mg/L ZnO NPs (Day
143) but recovered over time (by Day 161) and were subsequently
present throughout, showing their acclimatization to the ZnO NPs.
At the same time, the development and predominance of new bacteria species (e.g. S18, S31) was also observed, while some (e.g.
S19) loss their domination but still survived until the end of the
experiment. Higher ZnO NPs (10.0 mg/L) resulted in the further
elimination of the low-resistant bacteria species (e.g. S11) and
the development of new high-resistant species (e.g. S3, S8 and
S12). On the whole, the bacterial community showed robust performance over ZnO NPs exposure, since a majority (15 in 21 of
Day 55) of the dominated species was able to survive until the
end of the experiment. Additionally, some of the observed changes
in the bacterial community structure may occur naturally during

the operation of the MBR. Taking into account these natural

dynamics, the impact of ZnO NPs on the bacterial community
structure would be even more insignicant.
A calculation of the ShannonWiener index also indicated no
signicant decrease in the community diversity upon ZnO NPs
exposure (Table 1).
Despite the robust performance of the community, several
shifts in the bacterial community structure were observed in cluster analysis (Fig. 4b). The rst shift was between the inoculum and
the activated sludge before the addition of ZnO NPs on Day 55 followed by the second shift between the sludge on Day 55 and during 1.0 mg/L ZnO NPs exposure on Day 143. The third shift was
between Day 161 and Day 221 following the addition of 10.0 mg/
L ZnO NPs. This suggests that change in bacterial community
occurred after the addition of each concentration of NPs. The community clusters during 1.0 mg/L dosage (i.e. from Day 55 to Day
161) show a lower similarity (68%) compared to that during
10.0 mg/L dosage (i.e. from Day 161 onwards, with a similarity of
78%), suggesting a slightly higher impact on bacterial community
of 1.0 mg/L ZnO NPs than 10.0 mg/L ZnO NPs. This conrms the
above-mentioned speculation that long-term exposure of 1.0 mg/
L ZnO NPs allowed the sludge bacteria to acclimatize to the toxic
effects of ZnO NPs, thus resulting in a less signicant response at
a higher ZnO NPs concentration (10.0 mg/L). The bacterial community on Days 161 and 174 showed about the same similarity to
each other compared to that on Days 190 and 221 (i.e. 82% and
85% respectively). Hence changes in bacterial community during
10.0 mg/L ZnO NPs exposure were less evident over time.
Although a correlation was observed between the overall performance of the MBR and the dynamics in the total bacteria community structures, the effect of ZnO NPs on the specic type of
functional bacteria (e.g. ammonium oxidizing bacteria, nitrite
oxidizing bacteria, and denitrifying bacteria) needs to be further
explored for a better understanding of the impacts of ZnO NPs.
3.4. Fate and removal behavior of ZnO NPs
The percentage removal of Zn was higher than 95% throughout
the entire study, except on a few occasions (Fig. S2). Zn content in
the activated sludge increased substantially from 0.4 mg/gMLSS
(Day 63) before ZnO NPs addition to 12.0 mg/gMLSS on Day 160
(the end of 1.0 mg/L ZnO NPs exposure), and further raised to
75 mg/gMLSS on Day 242 (the end of 10.0 mg/L ZnO NPs exposure)
(Fig. S2). The steady increase in Zn content in the sludge suggested
that biosorption played an important role in ZnO NPs removal from
wastewater. Studies suggested that biomass can remove NPs from
wastewater by trapping them in the EPS or biolm matrixes (Kiser
et al., 2010; Kaegi et al., 2011). Puay et al. (2015) also showed signicant biosorption of ZnO NPs in an SBR system.
The Zn concentration in the inuent, efuent and supernatant
of the mixed liquor are shown in Fig. 5. As the supernatant was collected from the upper layer of the settled sludge, the difference in
Zn content between inuent and supernatant in this case could be
considered to be similar as their removal in CAS system. The
remaining difference between the Zn content in the supernatant
and the efuent Zn could be attributed to membrane rejection.
During the period of 1.0 mg/L ZnO NPs exposure, both supernatant
and efuent Zn content were consistently low. On average, Zn
removal by settling accounted for 80% while overall Zn removal
was 96%. Several studies, including both full scale wastewater
treatment plant and laboratory experiments have reported >90%
removal of NPs (Kaegi et al., 2011; Gottschalk et al., 2009).
However, at the higher ZnO NPs concentration of 10.0 mg/L, Zn
concentration in the supernatant was signicantly higher (up to
5.4 mg/L), compared to 0.2 mg/L in Phase II (i.e. ZnO NPs at
1.0 mg/L), which was possibly due to the stabilization of the NPs

131

M. Tan et al. / Bioresource Technology 185 (2015) 125133

Sludge inoculum
Day 55
Day 143
Day 161
Day 174
Day 190

Day 221
Fig. 4. (a) DGGE ngerprint patterns and (b) cluster analysis of bacterial community dynamics in the MBR.

Table 1
ShannonWiener diversity index of the bacterial community in the MBR.
Sample

Diversity index (H0 )

Sludge inoculum
Day 55
Day 143
Day 161
Day 174
Day 190
Day 221

2.836
2.895
2.992
2.885
2.873
2.984
3.012

by SMP as mentioned previously. Zn concentration in the efuent

remained low, as the overall Zn removal remained at an average

of 98% when the Zn removal by settling decreased to <50%, which

suggested that membrane rejection played an important role in Zn
removal at higher ZnO NPs concentration.
3.5. Mass balance of Zn
A mass balance on the total Zn loading in the inuent, efuent,
supernatant and in the reactor was performed (Fig. 6). The cumulative Zn that exited and remained in the system was close to 93.5%
of the cumulative Zn loaded into the system before Day 160.
However, this decreased to about 72.3% of the cumulative Zn at
the end of the operation. Taking into account the amount of Zn
deposited in the membrane foulant (which amounted to 8.03 mg

132

M. Tan et al. / Bioresource Technology 185 (2015) 125133

1 mg/L ZnO NPs

Zn concentration in the influent, mg/L

12

Mixed
liquorof
supernatant
Supernatent
MLSS
Effluent
Influent

10

0
63

66

70

73

83

Zn concentration in the supernatant and the

effluent, mg/L

10 mg/L ZnO NPs

14

86 128 140 156 172 186 200 216 230 242

Operation time, days

Fig. 5. ZnO NPs removal during MBR operation.

Acknowledgements
This research was funded by the Singapore National Research
Foundation under its Competitive Research Program for the project
entitled Advanced FO Membranes and Membrane Systems for
Wastewater Treatment, Water Reuse and Seawater Desalination
(Grant number: R-279-000-338-281).
Appendix A. Supplementary data

10.0 mg/L ZnO NPs

1.0 mg/L ZnO NPs

nitrifying activity was manifested only after long-term exposure.

ZnO NPs caused noteworthy changes in activated sludge properties
(i.e. reduced sludge oc size, poorer settleability and increased EPS
and SMP production), thus resulting in increased membrane fouling. Shifts in the bacterial community structures were revealed
by PCR-DGGE. ZnO NPs were removed effectively in MBR (>98%)
mainly via biosorption. Membrane ltration also contributed to
removing ZnO NPs, especially at high concentrations.

10000
Cumulative mass of Zn, mg

Influent

Supplementary data associated with this article can be found, in

the online version, at http://dx.doi.org/10.1016/j.biortech.2015.02.
094.

Effluent

8000

Sludge in reactor
Sludge in reactor + Desludge

6000

References

4000

2000

0
60

80

100

120

140
160
180
Operation time, days

200

220

240

Fig. 6. Total Zn mass balance in the MBR.

and 7.26 mg on Day 189 and Day 242, respectively), the deviation
is still signicant. This deviation may be due to the dosing of ZnO
NPs which was apportioned in two concentrations, i.e. 1.0 mg/L
and 10.0 mg/L. Another possible source of error could be due to
sampling, especially of the sludge, which tends to settle or to accumulate on reactor surfaces.
By Day 242, the distribution of Zn content was about 65.9%
remaining in waste sludge, 33.8% in the reactor, and 0.3% in the
efuent. This supports the conclusion that a major removal pathway of ZnO NPs from wastewater is through sorption by the activated sludge (i.e. sludge in the reactor and waste sludge) which
accounted for 71.9% of the added Zn. The dissolution prole of
ZnO NPs shows that Zn2+ concentration remained at average values
of 0.13 mg/L and 0.33 mg/L in Phase II (i.e. 1.0 mg/L ZnO NPs) and
Phase III (i.e. 10.0 mg/L ZnO NPs) of the experiment, showing that
the majority of Zn remained in the solid form. At a concentration
(10.0 mg/L) higher than the present environmentally-relevant concentration, results here show that the risk that ZnO NPs may further pose to the water environment was signicant reduced after
wastewater treatment by MBR (with high and constant removal
of 98%). MBR could be an effective process in the control of the
release of ZnO NPs to the environment. However, since a high portion of Zn remained in the biosolids, it presents a problem in the
downstream treatment of the ZnO NPs rich waste sludge.
4. Conclusion
Long-term exposure of ZnO NPs caused moderate decrease in
COD, nitrogen and phosphorus removal in MBR. Denitrication
was affected upon exposure but recovered over time. Effect on

APHA, 1999. Standard Methods for the Examination of Water and Wastewater,
19th ed. American Public Health Association Inc., Washington, DC.
Borm, P., Klaessig, F.C., Landry, T.D., Moudgil, B., Pauluhn, J., Thomas, K., Trottier, R.,
Wood, S., 2006. Research strategies for safety evaluation of nanoparticles, part
v: role of dissolution in biological fate and effects of nanoparticles. Toxicol. Sci.
90, 2332.
Chen, J., Tang, Y., Li, Y., Nie, Y., Hou, L., Li, X., Wu, X., 2014. Impacts of different
nanoparticles on functional bacterial community in activated sludge.
Chemosphere 104, 141148.
Diegoli, S., Manciulea, A.L., Begum, S., Jones, I.P., Lead, J.R., Preece, J.A., 2008.
Interaction between manufactured gold nanoparticles and naturally occurring
organic macromolecules. Sci. Total Environ. 402, 5161.
Do, K.U., Rajesh Banu, J., Son, D.H., Yeom, I.T., 2012. Inuence of ferrous sulphate on
thermochemical sludge disintegration and on performance of wastewater
treatment in an anoxic/oxic MBR. Biochem. Eng. J. 66, 2026.
Drews, A., Vocks, M., Bracklow, U., Iversen, V., Kraume, M., 2008. Does fouling in
MBRs depend on SMP? Desalination 231, 141149.
Farr, M., Gajda-Schrantz, K., Kantiani, L., Barcel, D., 2009. Ecotoxicity and analysis
of nanomaterials in the aquatic environment. Anal. Bioanal. Chem. 393, 8195.
Gottschalk, F., Sonderer, T., Scholz, R.W., Nowack, B., 2009. Modeled environmental
concentrations of engineered nanomaterials (TiO2, ZnO, Ag, CNT, Fullerenes) for
different regions. Environ. Sci. Technol. 43, 92169222.
Hou, L., Xia, J., Li, K., Chen, J., Wu, X., Li, X., 2013. Removal of ZnO nanoparticles in
simulated wastewater treatment processes and its effects on COD and NH4-N
reduction. Water Sci. Technol. 67, 254260.
Huang, F., Wang, Z., Mei, X., Wu, Z., 2013. Effects of short-term presence of
nanoparticles on MBR sludge properties and membrane fouling. Technol. Water
Treat. 8, 1426.
Judd, S., 2011. The MBR Book: Principles and Applications of Membrane Bioreactors
for Water and Wastewater Treatment, 2nd ed. Butterworth-Heinemann,
Oxford.
Kaegi, R., Voegelin, A., Sinnet, B., Zuleeg, S., Hagendorfer, H., Burkhardt, M., Siegrist,
H., 2011. Behavior of metallic silver nanoparticles in a pilot wastewater
treatment plant. Environ. Sci. Technol. 45, 39023908.
Kiser, M.A., Ryu, H., Jang, H.Y., Hristovski, K., Westerhoff, P., 2010. Biosorption of
nanoparticles to heterotrophic wastewater biomass. Water Res. 44, 41054114.
Le-Clech, P., Chen, V., Fane, A.G., 2006. Fouling in membrane bioreactors used in
wastewater treatment. J. Membr. Sci. 284, 1753.
Lin, H., Zhang, M., Wang, F., Meng, F., Liao, B., Hong, H., Chen, J., Gao, W., 2014. A
critical review of extracellular polymeric substances (EPSs) in membrane
bioreactors: characteristics, roles in membrane fouling and control strategies. J.
Membr. Sci. 460, 110125.
Liu, G., Wang, D., Wang, J., Cesar, M., 2011. Effect of ZnO particles on activated
sludge: role of particle dissolution. Sci. Total Environ. 409, 28522857.
Lombi, E., Donner, E., Tavakkoli, E., Turney, Terence W., Naidu, R., Miller, Bradley W.,
Scheckel, Kirk G., 2012. Fate of zinc oxide nanoparticles during anaerobic
digestion of wastewater and post-treatment processing of sewage sludge.
Environ. Sci. Technol. 46, 90899096.
Mei, X., Wang, Z., Zheng, X., Huang, F., Ma, J., Tang, J., Wu, Z., 2014. Soluble microbial
products in membrane bioreactors in the presence of ZnO nanoparticles. J.
Membr. Sci. 451, 169176.
Musee, N., Zvimba, J.N., Schaefer, L.M., Nota, N., Sikhwivhilu, Lucky M., Thwala, M.,
2014. Fate and behavior of ZnO- and Ag-engineered nanoparticles and a

M. Tan et al. / Bioresource Technology 185 (2015) 125133

bacterial viability assessment in a simulated wastewater treatment plant. J.
Environ. Sci. Health A 49, 5966.
Puay, N.Q., Qiu, G., Ting, Y.P., 2015. Effect of ZnO nanoparticles on biological
wastewater treatment in a sequencing batch reactor (SBR). J. Cleaner Prod. 88,
139145.
Qiu, G., Ting, Y.P., 2013. Osmotic membrane bioreactor for wastewater treatment
and the effect of salt accumulation on system performance and microbial
community dynamics. Bioresour. Technol. 150, 287297.
Qiu, G., Ting, Y.P., 2014a. Direct phosphorus recovery from municipal wastewater
via osmotic membrane bioreactor (OMBR) for wastewater treatment. Bioresour.
Technol. 170, 221229.
Qiu, G., Ting, Y.P., 2014b. Short-term fouling propensity and ux behavior in an
osmotic membrane bioreactor for wastewater treatment. Desalination 332, 91
99.
Qiu, G., Song, Y., Zeng, P., Duan, L., Xiao, S., 2013a. Combination of upow anaerobic
sludge blanket (UASB) and membrane bioreactor (MBR) for berberine reduction
from wastewater and the effects of berberine on bacterial community
dynamics. J. Hazard. Mater. 246247, 3443.
Qiu, G., Song, Y., Zeng, P., Duan, L., Xiao, S., 2013b. Characterization of bacterial
communities in hybrid upow anaerobic sludge blanket (UASB)membrane
bioreactor (MBR) process for berberine antibiotic wastewater. Bioresour.
Technol. 142, 5262.
Rajesh Banu, J., Do, K.U., Yeom, I.T., 2009. Nutrient removal in an A2/O-MBR reactor
with sludge recycling. Bioresour. Technol. 100, 38203824.

133

Rajesh Banu, J., Do, K.U., Kaliappan, S., Yeom, I.T., 2011. Effect of sludge
pretreatment on the performance of anaerobic/anoxic/oxic membrane
bioreactor treating domestic wastewater. Int. J. Environ. Sci. Technol. 8, 281
290.
Sun, T., Gottschalk, F., Hungerbuhler, K., Nowack, B., 2014. Comprehensive
probabilistic modelling of environmental emissions of engineered
nanomaterials. Environ. Pollut. 185, 6976.
Wang, Z., Ma, J., Tang, C., Kimura, K., Wang, Q., Han, X., 2014a. Membrane cleaning
in membrane bioreactors: a review. J. Membr. Sci. 468, 276307.
Wang, Z., Huang, F., Mei, X., Wang, Q., Song, H., Zhu, C., Wu, Z., 2014b. Long-term
operation of an MBR in the presence of zinc oxide nanoparticles reveals no
signicant adverse effects on its performance. J. Membr. Sci. 471, 258264.
Yang, X., Cui, F., Guo, X., Li, D., 2013. Effects of nanosized titanium dioxide on the
physicochemical stability of activated sludge ocs using the thermodynamic
approach and Kelvin probe force microscopy. Water Res. 47, 39473958.
Zhang, L., Jiang, Y., Ding, Y., Povey, M., York, D., 2007. Investigation into the
antibacterial behavior of suspensions of ZnO nanoparticles (ZnO nanouids). J.
Nanopart. Res. 9, 479489.
Zhang, C., Liang, Z., Hu, Z., 2014. Bacterial response to a continuous long-term
exposure of silver nanoparticles at sub-ppm silver concentrations in a
membrane bioreactor activated sludge system. Water Res. 50, 350358.
Zheng, X., Wu, R., Chen, Y., 2011. Effects of ZnO nanoparticles on wastewater
biological nitrogen and phosphorous removal. Environ. Sci. Technol. 45, 2826
2832.