influence of codn ratio on sludge properties and their role in membrane fouling of a submerged mbr

8/17/2019 Influence of CODN Ratio on Sludge Properties and Their Role in Membrane Fouling of a Submerged Mbr

1/10

Inuence of COD:N ratio on sludge properties and their role in

membrane fouling of a submerged membrane bioreactor

L. Hao a , S.N. Liss b, B.Q. Liao a , *

a Department of Chemical Engineering, Lakehead University, 955 Oliver Road, Thunder Bay, ON P7B 5E1, Canadab School of Environmental Studies and Department of Chemical Engineering, Queen's University, Kingston, ON K7L 3N6, Canada

a r t i c l e i n f o

Article history:

Received 21 August 2015

Received in revised form

30 October 2015

Accepted 22 November 2015

Available online 2 December 2015

Keywords:

COD:N

Membrane bioreactor

Membrane fouling

Industrial wastewater

Extracellular polymeric substances

Sludge properties

a b s t r a c t

The effect of COD:N ratio on sludge properties and their role in membrane fouling were examined using a

well-controlled aerobic membrane bioreactor receiving a synthetic high strength wastewater containing

glucose. Membrane performance was improved with an increase in the COD/N ratio (100:5e100:1.8) (i.e.

reduced N dosage). Surface analysis of sludge by X-ray photoelectron spectroscopy (XPS) indicates sig-

nicant differences in surface concentrations of elements C, O and N that were observed under different

COD/N ratios, implying changes in the composition of extracellular polymeric substances (EPS). Fourier

transform-infrared spectroscopy (FTIR) revealed a unique characteristic peak (C]O bonds) at 1735 cm1

under nitrogen limitation conditions. Total EPS decreased with an increase in COD/N ratio, corresponding

to a decrease in the proteins (PN) to carbohydrates (CH) ratio in EPS. There were no signicant differ-

ences in the total soluble microbial products (SMPs) but the ratio of PN/CH in SMPs decreased with an

increase in COD/N ratios. The results suggest that EPS and SMP composition and the presence of a small

quantity of lamentous microorganisms played an important role in controlling membrane fouling.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Integration of membrane technologies with conventional

biologically-based wastewater treatment has been widely applied

to the management of municipal and industrial wastewaters

providing a direct solideliquid separation by membrane ltration.

Compared to the conventional activated sludge process, membrane

bioreactors (MBR) offer several advantages including superior

ef uent quality as well as a smaller footprint (Brindle and

Stephenson, 1996; Visvanathan et al., 2000; Rosenberger and

Kraume, 2003; Marrot et al., 2004). However, membrane fouling

which leads to poor membrane performance and high operationalcosts (Kraume and Drews, 2010; Judd, 2011), continues to be a

major challenge that limits the further development and wide-

spread application of MBRs.

In general, membrane fouling is caused by many factors,

including inuent characteristics, solid retention time (SRT), hy-

draulic retention time (HRT), organic loading rate (OLR), and dis-

solved oxygen levels (Kraume and Drews, 2010). Among these

parameters, COD:N:P ratios remain one of the most important

factors. This is largely because COD:N:P ratio can inuence the

physiological properties of microorganisms and chemical compo-

sitions of biomass in MBRs. The COD:N:P ratio also can inuence

the amount of extracellular polymeric substances (EPS) and the

composition of proteins (PN) and carbohydrates (CH) in EPS which

can affect membrane performance.

Considerable attention has been given to the effect of COD:N

ratio in conventional biological processes for municipal wastewater

treatment on sludge properties including settleability (Durmaz and

Sanin, 2003), lterability (Wu et al., 1982), and occulation and

dewatering (Sanin et al., 2006). Other studies have focused on the

effect of COD:N ratios on nitri

cation and denitri

cation processesas well as nutrient removal ef ciency (McAdam and Judd, 2007;

Meng et al., 2008; Hwang et al., 2009). Studies found that EPS

production increased with an increase in the COD:N ratio in

municipal wastewater treatment (Durmaz and Sanin, 2001;

Miqueleto et al., 2010; Feng et al., 2012).

To date studies have largely focused on the effect of COD:N ratio

in conventional activated sludge processes, and mainly focused on

N removal in municipal wastewater treatment. There have only

been a few reports on the effect of COD:N:P ratio on biomass

properties and membrane fouling in submerged membrane bio-

reactors for municipal wastewater treatment with excess amount* Corresponding author.E-mail address: [email protected] (B.Q. Liao).

Contents lists available at ScienceDirect

Water Research

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c om / l o c a t e / w a t r e s

http://dx.doi.org/10.1016/j.watres.2015.11.052

0043-1354/©

2015 Elsevier Ltd. All rights reserved.

Water Research 89 (2016) 132e141


2/10

of nutrients. On the other hand, the role of nutrients and COD:N

ratios is most important for industrial wastewaters where nutrients

(N and P) are added and required to support microbial growth and

biodegradation (Tchobanoglous and Burton, 1991; Henze et al.,

1997; Ammary, 2003). Optimizing nutrient addition not only im-

proves process ef ciency but also saves chemical costs and reduces

secondary pollution of treated ef uent with added nutrients. The

paper describes the effectof three different COD:N:P ratios (COD:N)

(100:5:1100:2.5:1, and 100:1.8:1) on sludge properties and their

role in membrane fouling of an aerobic MBR for high strength in-

dustrial wastewater, like food industry wastewater, treatment. The

study included examination of membrane ltration resistance,

surface concentrations of elements (C, N, and O) on sludge surfaces,

extracellular polymeric substances (EPS), soluble microbial prod-

ucts (SMP) formed, and the abundance of lamentous microor-

ganisms under different COD:N ratios.

2. Material and methods

2.1. Experimental set-up and operating conditions

The laboratory-scale submerged MBR system, as shown in a

previous publication (Hao and Liao, 2015), had a 6.0 L workingvolume and containedat-sheet microltration membranes (SINAP

Membrane Science & Technology Co. Ltd., Shanghai, China). The

at-sheet membranes were made of polyvinylidene uoride

(PVDF), and had a pore size of 0.3 mm and molecular weight cut-off

(MWCO) of 70,000 Da. Immersed coarse air bubble diffusers (3.8 L/

min (LPM)) were installed under the membrane module for aera-

tion as well as air scouring to limit membrane fouling. Finer air

aeration (2.6 LPM) was also used to provide additional oxygen to

maintain satisfying dissolved oxygen (DO) level large than 2 mg/L

during the aerobic period. Mechanical mixing was achieved using a

magnetic stirrer (Thermolyne Cimarec, Model S47030) located at

the bottom of the reactor. The temperature of the bioreactor was

maintained constant at 35 ± 1 C by means of water jacket. The pH

was monitored by a pH electrode (Thermo Scientic, Beverly, MA),and automatically adjusted to 7.0 ± 0.2 by a pH regulation pump

using 0.25 M NaOH.

The system was continuously fed with a synthetic high strength

wastewater containing glucose, simulating the food industry

wastewaters, stored at 4 C. Synthetic wastewater containing

glucose has been widely as a receipt of laboratory-scale biological

wastewater treatment to develop fundamental understanding of

microbial kinetics, microbialoc structure, biomass separation, and

membrane fouling in MBRs (Kovarova-Kovar and Egli, 1998; Hong

et al., 2002; Liao et al., 2011). The feed was pumped automatically

by a peristaltic pump (Masterex Model 7520-50, Barnant Co.,

USA), which was controlled by a liquid level sensor (Madison Co.,

USA) as well as a controller (Flowline, USA). The ef uent was ob-

tained by means of a suction pump connected to the membranemodule. The operational cycles were applied with 4 min suction

followed by 1 min relaxation. The trans-membrane pressure (TMP)

was monitored by a pressure gauge connected between the

membrane module and the suction pump. An instant membrane

ux of 10 L/m2 h was applied in this study.

The synthetic wastewater included glucose, nitrogen (NH4Cl)

and phosphorus (KH2PO4) in a proportion of chemical oxygen de-

mand (COD): N: P ¼ 100:5:1, 100:2.5:1 and 100:1.8:1 and trace

metals as summarized in Table 1. All chemicals were from Sigma-

eAldrich at analytical grade. The inuent COD was approximately

2500 mg/L. The mixed liquor suspended solids (MLSS) concentra-

tion under steady-state operation was maintained at 8450 ± 325,

7268 ± 289 and 6363 ± 181 mg/L under a COD:N:P ratio of 100:5:1,

100:2.5:1, 100:1.8:1, respectively (Hao and Liao, 2015). During the

operation of the bioreactor, a volume of 400 mL of sludge was

discharged daily from the MBR for bulk sludge characterization and

to maintain the SRT at 15 days. The reactor was operated at an HRT

of approximately one day and an OLR of 2.5 g COD L 1 day1. The

operation of the reactor system included three phases: Phase 1

(0e107 day), the MBR was fed at a COD:N:P ratio of 100:5:1; in

Phase 2 (108e240 day), the reactor was fed with COD:N:P ratio of

100:2.5:1; in Phase 3 (240e380 day) the MBR was operated at

nutrients ratio (COD:N:P) of 100:1.8:1. Stable operation was ach-

ieved after three SRTs operation following moving to each nutrientregime.

2.2. Analytical methods

2.2.1. Water quality measurement

The inuent synthetic wastewater, permeate and mixed liquor

were sampled periodically from the system. The COD and MLSS

were analyzed according to Standard Methods (APHA, 2005). Su-

pernatant COD was determined after centrifuging the mixed liquor

for 20 min at 18,700 g .

2.2.2. Bound extracellular polymeric substances (EPS) extraction

and measurement

The bound EPS from sludge suspensions samples was extracted

according to cations exchange resin (CER) (Dowex Marathon C, Naþ

form, SigmaeAldrich, Bellefonte, PA) method (Frølund et al., 1996).

100 mL sample of the sludge suspension was taken and centrifuged

(IEC MultiRF, Thermo IEC, Needham Heights, MA, USA) at 18,700 g

for 20 min at 4 C. The sludge pellets were re-suspended to their

original volume using a buffer consisting of 2 mM Na3PO4, 4 mM

NaH2PO4, 9 mM NaCl and 1 mM KCl at pH 7. The sludge was then

transferred to an extraction beaker lled with buffer and the CER

(80 g/g-MLSS) and mixed for 2 h at 4 C. The bound EPS was

determined as the sum of bound PN and CH and was measured

colorimetrically by the methods of Lowery et al. (Lowery et al.,

1951) and DuBois et al. (DuBois et al., 1956), respectively. The to-

tal bound EPS was represented by adding the concentrations of

bound CH and PN.

2.2.3. Soluble microbial products (SMP) measurement

SMP means the soluble EPS. The extracted supernatant recov-

ered from mixed liquor following centrifugation as described above

was ltered through 0.45mm membranelters (Millipore) and used

to estimate SMP. Measurements for PN and CH of SMP were as

described above. The total SMP was the sum of the concentrations

of soluble CH and PN.

2.2.4. Surface composition of sludge by X-ray photoelectron

spectroscopy (XPS)

The surface concentrations of elements, including C, O, and N

were examined by XPS. The wet sludge samples were freeze-dried

Table 1

Composition and concentration of micronutrients in the feed.

Components Concentration (mg/L)

MgSO4 ∙ 7H2O 5.07

FeSO4 ∙ 7H2O 2.49

Na2MoO4 ∙ 2H2O 1.26

MnSO4 ∙ 4H2O 0.31

CuSO4 0.25

ZnSO4 ∙ 7H2O 0.44NaCl(mg/L) 0.25

CaSO4 ∙ 2H2O 0.43

CoCl2 ∙ 6H2O 0.41

L. Hao et al. / Water Research 89 (2016) 132e141 133


3/10

at 35 C for one week. The freeze-dried sludge samples were

ground to a powder before being analyzed by a ThermoFisher Sci-

entic K-Alpha XPS Spectrometer equipped with monochromatic

AlK a X-ray source with a spot source of 400 mm (ThermoFisher, E.

Grinstead, UK). Charge compensation was also provided. The po-

sition of the energyscale was adjusted to place the main C1s feature

(CeC) at 285.0 eV except for those samples where the CeO peak

wasmore dominant. A surveyspectrum was taken at lowresolution

(PE 150 eV). High resolution spectra were taken of C1s regions

(PE25 eV). All data processing was performed using the software

(Advantage) provided with the instrument. Three sludge samples

were collected at three different days, under each singly COD:N

ratio condition, during the period close to the end of stable oper-

ation of each COD:N ratio.

2.2.5. Molecular composition of sludge by Fourier transform-

infrared spectroscopy (FTIR)

A Bruker Ten 37 FTIR Spectrometer (Bruker Optics Inc., Billerica,

MA, USA) was employed to determine the major organic functional

groups of freeze-dried sludge and to predict the chemical func-

tional groups of the bulk sludge. The FTIR analysis was performed

using absorbance mode. The sludge samples were freeze-dried

at 35 C for one week prior to analysis. The samples for FTIR

were from the same batches of samples for XPS.

2.2.6. Membrane resistance

The extent of fouling rate can be evaluated by the derivative of

the ltration resistance:

Rt ¼ DP

J hT ¼ Rm þ R f þ Rc (1)

hT ¼ h20 C$e0:0239 T 20ð Þ (2)

where R t is the total ltration resistance (m1), J represents the

permeate ux (m3

/m2

h), DP is the trans-membrane pressure dif-ference (Pa), and hT is the permeate dynamic viscosity (Pa s). T is

the permeate temperature in C. R t was calculated with the tem-

perature corrected to 20 C to compensate for the dependence of

viscosity on temperature. The experimental procedure to deter-

mine each resistance value is as follows: R t is total membrane

resistance (m1) and calculated from the ltration data at the end

of operation. R m is intrinsic membrane resistance and evaluated by

the waterux of tap water. Rc is fouling layer resistance (m1). R f is

fouling resistance due to irreversible adsorption and pore blocking

(m1). When fouling occurred, membrane surfaces were wiped

with a sponge and tap water, and then the membrane was sub-

merged in tap water for ux and TMP measurement. New mem-

branes were used in each membrane operation cycle, in order to

maintain the same membrane conditions for comparison and

reproducibility of results in repeated membrane operation cycles at

each nutrient (COD:N) ratio.

2.2.7. Morphology of sludge and textual of cross-section of fouled

membranes

Sludge samples were routinely taken out from the MBR and

viewed for morphology under a conventional optical microscope

(COM) (Olympus IX 51 Inverted Microscope, Olympus America Inc.,

Melville, NY) at a magnication of 40. A cross section of the cake

layer was observed by a COM to investigate the physical structure of

sludge cake layers formed on membrane surface. In order to pre-

vent the structure and thickness of cake layer from change, the cake

layer was saturated with 0.85% NaCl aqueous solution and then

frozen at 22

C before being

xed on to a sample stage using

optimal cutting temperature (O.T.C) compound (Sakura Finetech-

nical Co. Ltd., Tokyo, 103, Japan).

2.3. Statistical analysis

Statistical analysis was conducted using the Statistical Package

for the Social Science (SPSS) 16.0. An analysis of variance (ANOVA)

was employed to identify whether there is signicant difference

between treatment means when evaluating membrane fouling and

EPS concentration under different COD:N ratios. The difference was

considered statistically signicant at a 95% condence interval

(p < 0.05). The student t-test also was applied to analyze the con-

tent of surface chemical composition of bulk sludge. The paired p

values were calculated for the differences between COD:N:P ratios

of 100:5:1 and 100:2.5:1; 100:2.5:1 and 100:1.8:1; and 100:5:1 and

100:1.8:1. Data sets were considered statistically different at a 95%

condence interval (p < 0.05).

3. Results

The MBR was operated at three different COD:N:P ratios,

100:5:1100:2.5:1, and 100:1.8:1 for over one year. COD removal

ef ciency of the MBR was only slightly decreased when N wasreduced (COD removal ranged from 98.4% to 99.5%) (Hao and Liao,

2015). Sludgeyield was reduced and ef uent quality with respect to

nutrient residuals improved indicating the feasibility of operating

an aerobic MBR for high strength wastewater treatment with less

nutrient addition while achieving good biological treatment (Hao

and Liao, 2015). Assessing the impact of nutrient conditions (i.e.

COD:N) when added to achieve biological treatment, on sludge

properties and membrane fouling is equally important and was the

subject of this study.

3.1. Floc size and morphology

The morphology of sludge ocs was observed throughout the

study using a COM. Fig. 1 illustrates the typical morphology of sludge ocs under different COD:N ratios. Sludge ocs at a COD:N

ratio of 100:5 had a lower level (0-1) of lamentous microorgan-

isms and were smaller in sizes than that at a COD:N ratio of 100:2.5

and 100:1.8 (moderate level 3-4), according to the classication of

Jenkins et al. (2003).

3.2. Surface characterization of sludge by XPS

The surface chemical composition of bulk sludge at different

COD:N:P ratios were analyzed by XPS. XPS has been used to study

the surface functional groups of materials, including bacteria, and

each peak corresponds to electrons with a characteristic binding

energy from a particular element (Dengis and Rouxhet, 1996;

Dufrene et al., 1997; Badireddy et al., 2008). As shown in Fig. 2,the C peaks (C1s, C1sA, C1sB, C1sC) were decomposed into four

different bonds: (1) C bound only to C and H (C(C,H) bonds) from

lipids and amino acids side chains at a binding energy of 284.8 eV

(C1s); (2) C singly bound to O or N, C(O, N), including ether,

alcohol, amine, and amide, at a binding energy of 286.3 eV (C1Sa);

(3) C bound to O making two single bonds or one double bond, C]

O or OeCeO, including amide, carbonyl, carboxylate, ester, acetal,

and hemiacetal, at a binding energy of 288.0 eV (C1sB); and (4) a

weak peak at 289.0 eV (C1sC) arises from O]CeOH and O]C-OR,

commonly found in carboxyl or ester groups (Badireddy et al.,

2008). The O peaks (O1s, O1sA, and O1sB) could be attributed to

three bonds: OeC bond, including hydroxide (CeOH), acetal, and

hemiacetal (CeOeC), at a binding energy of 532.7 eV, and O]C in

carboxylic acid, carboxylate, ester, carbonyl and amide at a binding

L. Hao et al. / Water Research 89 (2016) 132e141134


4/10

energy of 531.4 eV. The N peaks (N1s and N1sA) were attributed to

the two different bonds: NeC bond in amide or amine at a binding

energy of 400.12 Ev (N1s) and NeH bonds in ammonia or proton-

ated amine at a binding energy of 402.10 eV (N1sA).

Surface concentrations of element C, N and O on sludge surfaces

under different nutrients conditions (COD:N) were summarized in

Table 2. Although there was no signicant difference in total C

(Student t-test, P > 0.05), signicant differences in the quantity of

C-(C, H), C-(O, N) and C]O were observed between COD:N:P of

100:5:1 and 100:2.5:1, 100:5:1 and 100:1.8:1, 100:2.5:1 and100:1.8:1 (Student t-test, p < 0.05). Furthermore, the total O on

sludge surface increased with an increase in the COD:N ratio

(reduced N dosage) (Student t-test, P < 0.05). Moreover, the general

tendency of total N on sludge surface decreased with an increase in

the COD:N ratio, although no statistically signicance was observed

(ANOVA, P > 0.05). But student t-test suggested signicant differ-

ence in total N on sludge surface existed between the COD:N ratio

of 100:5 and 100:1.8 (student t-test, p < 0.05).

3.3. Bound EPS production and components

Bound EPS is dened as the biopolymers originally from cellsattached on the surface of ocs or cells. In order to understand the

role of bound EPS in membrane fouling under reduced N dosage in

industrial wastewater treatment, the content of PN and CH of

bound EPS in sludge were measured. The bound EPS contents at

each COD:N:P ratio are shown in Fig. 3. Statistical analysis using

ANOVA also conrmed that differences in total EPS, PN, CH were all

statistically signicant (ANOVA, p < 0.05) at different COD:N:P ra-

tios studied. The total EPS were 35.5(±3.08),19.12 (±2.04) and 15.21

(±1.22) mg/g MLSS for COD:N:P ratios of 100:5:1, 100:2.5:1 and

100:1.8:1, respectively. An increase in COD:N ratio led to a decrease

in total bound EPS production but more carbohydrates in bound

EPS. This corresponds to the change in PN/CH ratio in bound EPS as

the COD:N ratio changed. The PN/CH values (ANOVA, p < 0.05)

decreased with an increase in the COD:N ratios (Fig. 3(b)).

3.4. SMP

The soluble microbial products (SMPs) are dened as bio-

polymers released from bacteria into solution, as compared to

bound EPS, which are biopolymers attached on bacteria surfaces.

The total SMP and soluble PN and CH concentrations at three

COD:N:P ratios are shown in Fig. 4. The amount of SMP (ANOVA,

p < 0.05) reached to a peak of 32.6 (±3.10) mg/L at the COD:N:P

ratio of 100:5:1. After decreasing nitrogen content in the feed, the

SMP production presented a decrease trend for other two condi-tions. Obviously, the PN content decreased with a decrease in the

nitrogen content in feed, while the biomass produced similar CH

level in SMP at COD:N of 100:5 and 100:2.5. At the COD:N ratio of

100:1.8, the CH concentration in SMPs was almost twice of CH

concentrations produced at the other two COD:N ratios. The PN/CH

values (ANOVA, p < 0.05) showed the similar trend as shown in

bound EPS in that PN/CH ratio decreased with an increase in the

COD:N ratio. CH were found as the predominant fraction and

accounted for more than half of the SMP concentration at COD:N:P

of 100:1.8:1.

3.5. FTIR analysis

The FT-IR spectra under different nutrients (COD:N) conditionsare shown in Fig. 5. All three sludge samples shown peaks at

3300 cm1 (overlapping of bands from the stretching vibrations of

NeH and OeH) (Kumar et al., 2006). Peaks, at wave numbers of

2926, 2847 and 1445 cm1, are reecting CeH bonds in the alkenes

class (Kim and Jang, 2006). The doublets at 2370 and 2343 cm 1

attribute to carbon dioxide from atmosphere adsorbed on the

sample surface. The asymmetrical stretching peak (C]O stretch)

observed at 1735 cm1 was associated with O-acetyl ester bonds

(Badireddy et al., 2008), under a nitrogen shortage of

COD:N:P ¼ 100:2.5:1 and COD:N:P ¼ 100:1.8:1.

3.6. Membrane performance

Fig. 6 shows the changes of

ltration resistance under different

Fig. 1. Morphology of sludge ocs observed by COM under different COD:N ratios (a) 100:5; (b) 100:2.5 and (c) 100:1.8.



5/10


6/10

membrane fouling. Recovery cleaning was performed by physical

cleaning using water with a wet sponge over the surface of the

fouled membrane to reinstall membrane performance.

3.7. Fouling resistance

Membrane ltration resistances evaluated at the end of an

operation cycle, obtained for each of the COD:N regimes studied,

are shown in Table 3. A comparison of the ltration resistances at

the end of an operation cycle among COD:N ratios studied is not

reasonable, due to the fact that different ending TMP was used. The

purpose of measuring the ltration resistance was to identify the

dominant fouling mechanism. The value of Rc/(Rc þ Rf) (100%)

indicates the formation of a cake layer was the dominant mecha-

nism of membrane fouling in the MBRs. Fig. 7 illustrates the

thickness of wet cake layers formed on membrane surfaces There

was variability with respect to the thickness of the cake layer over

the surface of the membrane. A thinner cake layer was usually

observed at the center of membrane module, where there was

likely better air scouring condition. The cake layer thickness ranged

from between 100 and 300 mm under the different nutrients

(COD:N) conditions studied. The cake layer thickness at COD:N

100:5 was usually thinner than that at COD:N 100:5 and 100:1.8 at

the end of an operation cycle.

4. Discussion

As shown in Fig. 1, sludge ocs had irregular shapes containing

different level of laments and changed in oc sizes. The changes in

the abundances of

lamentous microorganisms and

oc sizes

under different COD:N ratios could be explained by our general

knowledge that nutrients (N, P) deciency in inuent would pro-

mote the growth of lamentous microorganisms in activated

sludge ( Jenkins et al., 2003). Thus, a moderate higher level (3-4) of

lamentous microorganisms was observed under N deciency

Fig. 3. Comparison of (a) bound EPS of bulk sludge and (b) PN/CH in EPS under

different COD:N:P ratios (ANOVA, p < 0.05, number of measurements: n ¼ 5 for each

condition).Fig. 4. (a) SMP concentrations and (b) PN/CH in SMPs under different COD:N:P ratios

(ANOVA, p < 0.05, number of measurements: n ¼ 5 for each condition).

Fig. 5. FTIR spectrum of bulk sludge under different COD:N:P ratios (sample number

n ¼ 3 under each COD:N ratio; sample dates: Day 74, 87 and 90 for COD:N 100:5; Day

218, 224 and 229 for COD:N 100:2.5; and Day 339, 347 and 352 for COD:N 100:1.8).



7/10

conditions. Furthermore, it is well-known that lamentous micro-

organisms could serve as a backbone for sludge oc formation and

thus increase oc sizes ( Jenkins et al., 2003). Moreover, under the

conditions of reduced growth due to N limitations, a signicant

Fig. 6. Membrane ltration resistance over time under different COD:N:P ratios (the recorded TMP pattern between days 115 and 129 is due to high membrane fouling rates

observed in this time period).

Table 3

A series of resistances at the end of the study point of membrane cyclic ltration.

Nutrients COD:N:P Resistances Cake resistance ratio (%)

R m R c R p R t R c/(R c þ R p)

100:5:1 (1.82 ± 0.07) 1012 (6.32 ± 0.10) 1012 0 (8.15 ± 0.08) 1012 100

100:2.5:1 (1.80 ± 0.02) 1012 (4.53 ± 0.01) 1012 0 (6.33 ± 0.01) 1012 100

100:1.8:1 (1.81 ± 0.07) 1012 (2.71 ± 0.18) 1012 0 (4.52 ± 0.11) 1012 100

Sample average ± relative error, number of measurements: n ¼ 2 for each COD:N:P condition.

Fig. 7. Cross section of cake layer observed by COM under different COD:N ratios (a) 100:5; (b) 100:2.5 and (c) 100:1.8.



8/10

portion of feed CODwould be converted into storage compounds as

reected by the higher level of CH in bound EPS, which also

resulted in the increase in oc sizes and changes in oc shape by

enhanced bioocculation. Shin et al. (2000) found an increase in

PN/CH in bound EPS correlated to a higher magnitude of negative

surface charge and thus inhibited bioocculation.

Membrane ltration performance of sludge is closely associated

with sludge morphology and chemical/physical characteristics, and

sludge ltration performance directly affects the trans-membrane

pressure (TMP). The larger oc sizes and decreased MLSS concen-

tration at a reduced nitrogen level (COD:N ¼ 100:2.5 and 100:1.8)

might also explain the better membrane performance under these

conditions, as smaller ocs have a higher tendency to attach to the

membrane surface and contributing to membrane fouling at a

COD:N ratio of 100:5. Although the cake layer thickness

(100e150 mm) at COD:N 100:5 was thinner than that

(200e300 mm) at COD:N 100:2.5 and 100:1.8 (Fig. 7), the longer

operation period before TMP jump (less fouling) at COD:N 100:2.5

and 100:1.8 might suggest a looser structure of cake layer, as the

sludge oc size increased, the cake layer porosity increased (Cao

et al., 2015). Thus, less membrane fouling was observed under N

deciency conditions. Some studies found lamentous bulking

sludge caused more serve membrane fouling (Li et al., 2008; Menget al., 2006a,b), but the presence of a small quantity of lamentous

microorganisms did improve membrane ltration (Meng et al.,

2006b). The results from this study is consistent with the nding

of Meng et al. (2006b) in that the presence of a small quantity of

lamentous microorganisms had a positive impact on membrane

permeation (less fouling). Therefore, control the growth of la-

mentous microorganisms at a certain level is important in con-

trolling membrane fouling. A possible explanation is that the

laments will serve as a net and form large oc and thus lead to a

less dense or more porous cake layer (Cao et al., 2015) and reduce

membrane fouling. The results from this study suggest that sludge

oc structure (morphology, oc size, and the abundance of la-

mentous microorganism) is an important factor in controlling

membrane fouling. Another possible explanation would be basedon the potentially signicant difference in microbial communities

under different COD:N ratios, in addition to the differences in the

abundances in lamentous microorganisms. Potential correlations

between microbial communities and EPS/SMP production needs

further studies using advanced molecular tools in the future to

develop new insights in membrane fouling.

The XPS results suggest signicant differences in the surface

concentrations of elements N and O and C bonds ( Table 2). This

could be attribute to the decreasing nitrogen level (increasing

COD:N ratio) in the system. As XPS detects the surface concentra-

tions of elements, like C, N and O, in a few nm thickness, which are

overlap with the bound EPS extracted from oc surfaces. In prin-

ciple, the results of XPS analysis and bound EPS extraction and

chemical analysis are related. The XPS results strongly suggestthere are signicant difference in bound EPS composition and the

properties of the surface under different COD:N ratios. This might

not be surprising that microorganisms responding to the stress of

nutrient (N) decient would produce different biopolymers

(Omoike and Chorover, 2004). It is well known that the bound EPS

consist of proteins, polysaccharides, nucleic acids, lipids, humic

acids, which are located at or outside the cell surface. Proteins in the

activated sludge bound EPS are the primary source for the

elemental nitrogen. Earlier XPS studies (Dengis and Rouxhet, 1996;

Dufrene et al., 1997; Omoike and Chorover, 2004; Badireddy et al.,

2008; Liao et al., 2011) have reported that the C-(C, H) bonds might

originate from lipids or from amino acid side chains. Poly-

saccharides or CH contain hydroxide and acetal or hemiacetal

building blocks. The higher content of Ce

(C, H) and N and a lower

concentration of O in COD:N:P of 100:5:1 might correlate to a

higher surface concentration of lipids and proteins and a lower

surface concentration of CH in the sludge. This is supported by

similar observations from direct chemical analysis of extracted EPS

below.

EPS are generally considered as “extracellular polymeric sub-

stances of biological origin that participate in the formation of

microbial aggregates” (Geesey, 1982). In recent years, a more

broaden denition of EPS was introduced to include both bound

EPS and soluble EPS (Laspidou and Rittmann, 2002). Bound EPS is

dened as the biopolymers attached on the surface of ocs or cells.

It is well-known that bound EPS play an important role in mem-

brane fouling (Wang et al., 2013; Lin et al., 2014). A change in bound

EPS composition and concentration would modify the interactions

between oc surface and membrane surface or between oc sur-

face and fouling layer surface and thus affect cake layer formation

rate and structure. At the lowest COD:N:P ratio of 100:5:1, the

microorganisms are more likely to produce more EPS with the

highest PN/CH ratio. This observation is consistent with previous

ndings (Durmaz and Sanin, 2001; Liu and Fang, 2003). After

reducing the nitrogen content in the feed (increased COD:N ratio),

apparently, the PN contents in bound EPS presented a rapid

decrease. This is mainly because the nitrogen in the system wasutilized for synthesis of PN and insuf cient nitrogen was supplied

for PN synthesis when COD:N ratio increased. At the lowest COD:N

ratio of 100:5, the majority of carbon source was used in biomass

synthesis instead of CH production. However, with an increase in

the COD:N ratio, the CH concentration in the bound EPS increased.

This is due to the fact that a large portion of feed COD was con-

verted into storage compounds like CH in bound EPS under N

deciency conditions. These results are parallel to the results by

Durmaz and Sanin (2001).

The bound EPS extraction and chemical analysis results are

consistent with the results of XPS as discussed above in that sig-

nicant difference in bound EPS composition and concentration

existed under different COD:N ratios. In this study, a decrease in the

PN/CH ratio in bound EPS correlated to a decrease in membranefouling rate. Cetin and Erdincler (2004) found that sludge lter-

ability and compactibility were improved considerably with a

decrease in the PN/CH ratio in bound EPS. Lee et al. (2001) sug-

gested that PN are more hydrophobic and stickier and have more

tendencies to adhere on membrane surface and inducing mem-

brane fouling. Thus, a higher PN/CH ratio in bound EPS at COD:N

100:5 would enhance cake layer formation thus corresponding

with greater membrane fouling. This observation is in consistent

with previous studies (Lin et al., 2011). It is also believed that total

amount of bound EPS was correlated to the membrane fouling as

the increase in the bound EPS content could deteriorate membrane

fouling (Al-Halbounia et al., 2008; Wang et al., 2009), while Geng

and Hall (2007) found that the content of bound EPS was not

directly associated with membrane fouling. Considering the argu-ment in the literature on the role of bound EPS content in mem-

brane fouling, it is very likely the PN/CH ratio (or the bound EPS

composition) played a more signicant role than the quantity of

bound EPS in controlling membrane fouling. It is the specic

composition (like PN and CH) of bound EPS, as reected by the

change in PN/CH ratio, that determines the specic interactions

(like hydrophobic interaction, van der Waals interaction, cation

bridging, electrostatic interaction) between sludge ocs and

membrane surfaces and thus affect cake formation.

A unied theory was developed by Laspidou and Rittmann

(2002) to elaborate the interrelations between SMP and bound

EPS. Accordingly, SMP are equal to soluble EPS (Laspidou and

Rittmann, 2002). SMP include biomass-associated products (BAP)

and substrate-utilization-associated products (UAP). SMP might



9/10

have different role in membrane fouling, as compared to bound

EPS. Li et al. (2013) found that bound EPS played a signicant role in

fouling development at stage 1, while SMP were the major con-

tributors to the self-accelerating fouling at stage 2. It is more likely

SMP would cause pore blocking or gel-layer formation, while

bound EPS plays a more important role in cake layer formation

(Meng et al., 2009; Wu and Fane, 2012). The higher CH content of

SMPat a COD:Nof 100:1.8 might suggest that more soluble CH were

released from sludge ocs under the extreme N limitation condi-

tion, considering the fact that glucose is a easily biodegradable

substrate. Morel (1983) indicates that if nutrients (N and P) are

present in very low concentrations, SMP may be produced to

scavenge the required nutrient. Ultraltration experiments (results

not shown) with a membrane having a molecular weight cut-off

(MWCO) of 1000 Da indicates that oligosaccharides were present

in SMP, as carbohydrates with a MWCO less than 1000 Da were

detected. The total SMPcontent could notexplain the improvement

of membrane performance at an increase in the COD:N ratio but an

increase in the PN/CH ratio is positively correlated to the membrane

fouling rate (e.g. a decrease in the length of membrane operation

before TMP jump) (Figs. 4 and 6). This might suggest that it is the

composition and PN/CH ratio but not the total SMP content that

controls membrane fouling rate. SMP at a COD:N 100:5 would havea higher af nity to membrane surface, due to its higher PN content

in SMP. Table3 shows that cake layer formationwas the mechanism

of membrane fouling with no pore clogging or adsorption. Thus, it

could be concluded that pore clogging or adsorption caused by SMP

did not occur, and gel layer formation by SMP was likely the

mechanism of SMP involved in membrane fouling. Furthermore,

the higher PN/CH ratio in SMP at COD/N 100:5 could accelerate the

self-accelerating fouling phenomena (increase cake layer formation

rate) in stage 2, as observed by Li et al. (2013). It is hypothesized

that the high molecular weight SMP would be accumulated on the

surface of membrane during ltration and thus would modify the

surface properties of membrane and outer cake layer to enhance

cake layer formation. The higher PN/CH ratio of SMPcorrelatedwell

with the higher PN/CH ratio in bound EPS. Thus, sludge with ahigher PN/CH ratio in bound EPS at COD:N of 100:5 would have a

higher af nity to the modied membrane surfaces with a higher

PN/CH ratio in accumulated SMP and thus increase cake layer for-

mation rate.

The unique FTIR peak at 1735 cm1 (due to C]O stretch) under

N deciency (COD:N of 100:2.5 and 100:1.8) suggest a unique

compound in sludge produced. This result indicated that the large

amount of fat or saturated aliphatic aldehyde could be produced

when biomass grown under the stressof nutrients limitation (Hung

et al., 1996). Additionally, the relative intensity of the peak at

1735 cm1 was stronger under COD:N 100:1.8 than that under

COD:N 100:2.5. A stronger intensity of the unique characteristic

peak at 1735 cm1 correlated to a bettermembrane performance (e.

g. a lower membrane fouling rate or a longer operation time boreTMP jump). The peaks located at 1660 (stretching vibration of C]O

and CeN amide I) and 1540 (NeH deformation and C]N stretching

amide II) and 1245 cm1 are important as they indicate the pres-

ence of proteins (Croue et al., 2003). The peaks of 1384 and

1244 cm1 suggest the presence of amide III ( Jun et al., 2007). A

peak near 1100 cm1 (due to CeO stretching) was due to the

functional group of carbohydrate (Croue et al., 2003). According to

the FTIR spectra, it can be concluded that the sludge under the

nitrogen limitation conditions can generate a great amount of fat or

saturated aliphatic aldehyde. The characteristic peak at 1735 cm1

under nitrogen limitations (COD:N of 100:2.5 and 100:1.8) may be

used as an indicator of nutrients conditions in feed. FTIR may be

used as a tool to indirectly monitor the nutrients condition in MBRs.

Increasing

ltration resistance in the operation of MBRs is an

important indicator of membrane fouling since it directly reects

deterioration in membrane permeability. Importantly, cake layer

resistance (Rc) under each COD:N condition was equivalent to 100%

of the total foulingresistance (Rt) (Table3), suggesting there wasno

irreversible membrane fouling by adsorption and/or pore blocking.

This might be due to the use of new membranes in each TMP cycle,

in order to check the reproducibility of membrane performance

under each tested COD:N condition. In such a short-term operation,

no irreversible membrane fouling likely developed and thus the

cake layer was the only fouling mechanism. Dominant factors that

determine cake layer formation would likely be the change in

bound EPS and SMP composition (such as PN/CH ratio) and the

presence a small quantity of laments that led to increasedoc size

and looser cake layer structure.

5. Conclusions

* An increase in COD:N ratio from 100:5 to 100:1.8 led to an

improved membrane performance and a longer operation

period before membrane cleaning.

* The XPS results demonstrated signicant differences in surface

concentrations of elements O and N and C bonds under different

COD:N ratios, which suggest signicant difference in bound EPS

composition.

* Bound EPS/SMP composition, like PN/CH ratio, played a more

important role in controlling membrane fouling than the

quantity of bound EPS and SMP.

* The presence of a small quantity of lamentous microorganisms

under N deciency led to an increased oc sizes and improved

membrane performance.

* A unique characteristic peak corresponding to unique fat or

saturated aliphatic aldehyde production at 1735 cm1 was

observed by FTIR in sludge under N limitations. FTIR can be used

as a tool to indirectly monitor nutrients (N) conditions in MBRs.

Acknowledgments

We thank the anonymous reviewers and editors for their valu-

able comments and suggestions on revising and improving the

work. This study was nancially supported by Natural Sciences and

Engineering Council of Canada (NSERC) (RGPIN-2014-03727). Spe-

cial thanks go to Dr. Rana Sodhi at Surface Interface Ontario at the

University of Toronto, Toronto, ON, for his help in XPS analysis.

References

Al-Halbounia, D., Traberb, J., Lykoc, S., Wintgensc, T., Melinc, T., Tacked, D., Janote, A.,Dotta, W., Hollenderb, J., 2008. Correlation of EPS content in activated sludge atdifferent sludge retention times with membrane fouling phenomena. WaterRes. 42, 1475e1488.

Ammary, B.Y., 20 03. Nutrients requirements in biological industrial wastewater

treatment. Afr. J. Biotechnol. 3, 236e238.APHA, 2005. Standard Methods for the Examination of Water and Wastewater, 21st

ed. American Public Health Association (APHA)/American Water Works Asso-ciation (AWWA)/Water Environment Federation (WEF), Washington, DC.

Badireddy, A.R., Korpol, B.R., Chellam, S., Gassman, P.L., Engelhard, M.H., Lea, A.S.,Rosso, K.M., 2008. Spectroscopic characterization of extracellular polymericsubstances from Escherichia coli and Serratia marcescens: suppression usingsub-inhibitory concentrations of bismuth thiols. Biomacromolecules 9,3079e3089.

Brindle, K., Stephenson, T., 1996. The application of membrane biological reactorsfor the treatment of wastewaters. Biotechnol. Bioeng. 49, 601e610.

Cao, T.A., Van De Staey, G., Smets, I.Y., 2015. Integrating activated sludge oc sizeinformation in MBR fouling modeling. Water Sci. Technol. 71 (7), 1073e1080.

Cetin, S., Erdincler, A., 2004. The role of carbohydrate and protein parts of extra-cellular polymeric substances on the dewaterability of biological sludge. WaterSci. Technol. 50 (9), 49e56.

Croue, J.P., Benedetti, M.F., Violleau, D., Leenheer, J.A., 2003. Characterization andcopper binding of humic and nonhumic organic matter isolated from the South

Platte River: evidence for the presence of nitrogenous binding site. Environ. Sci.



10/10

Technol. 37, 328e336.Dengis, P.B., Rouxhet, P.G., 1996. Preparation of yeast cells for surface analysis by

XPS. J. Microbiol. Methods 26, 171e183.DuBois, M., Gilles, K.A., Hamilton, J.K., Rebers, P.A., Smith, F., 1956. Colorimetric

method for determination of sugars and related substances. Anal. Chem. 28,350e356.

Dufrene, Y.F., VanderWal, A., Norde, W., Rouxhet, P.G., 1997. X-ray photoelectronspectroscopy analysis of whole cells and isolated cell walls of gram-positivebacteria: comparison with biochemical analysis. J. Bacteriol. 179, 1023e1028.

Durmaz, B., Sanin, F.D., 2001. Effect of carbon to nitrogen ratio on the composition

of microbial extracellular polymers in activated sludge. Water Sci. Technol. 44,221e229.

Durmaz, B., Sanin, F.D., 2003. Effect of carbon to nitrogen ratio on the physical andchemical properties of activated sludge. Environ. Technol. 24, 1331e1340.

Feng, S.P., Zhang, N.N., Liu, H.C., Du, X.L., Liu, Y.L., Lin, H., 2012. The effect of COD/Nratio on process performance and membrane fouling in a submerged bioreactor.Desalination 285, 232e238.

Frølund, B., Palmgren, R., Keiding, K., Nielsen, P.H., 1996. Extraction of extracellularpolymers from activated sludge using a cation exchange resin. Water Res. 30,1749e1758.

Geesey, G.G., 1982. Microbial exopolymers: ecological and economic considerations.ASM News 48, 9e14.

Geng, Z., Hall, E.R., 2007. A comparative study of fouling-related properties of sludgefrom conventional and membrane enhanced biological phosphorus removalprocesses. Water Res. 41 (19), 4329e4338.

Hao, L., Liao, B.Q., 2015. Effect of organic matter to nitrogen ratio on membranebioreactor performance. Environ. Technol. 36 (20), 2674e2680.

Henze, M., Harremoes, P., Jansen, J., Arvin, E., 1997. Wastewater Treatment, seconded. Springer.

Hong, S.P., Bae, T.H., Tak, T.M., Hong, S., Randall, A., 2002. Fouling control in acti-vated sludge submerged hollow ber membrane bioreactors. Desalination 143,219e228.

Hung, W.T., Chang, I.L., Hong, S.G., Young, J.D., Chen, G.W., Lee, D.J., 1996. Flocmigration and chemical composition change in a freezing chamber. J. Environ.Sci. Health Part A Environ. Sci. Eng. Toxic Hazard. Subst. Control 31, 1053e1066.

Hwang, J.H., Cicek, N., Oleszkiewicz, J., 2009. Effect of loading rate and oxygensupply on nitrication in a non-porous membrane biolm reactor. Water Res.43, 3301e3307.

Jenkins, D., Richard, M.G., Daigger, G.T., 2003. Manual on the Causes and Control of Activated Sludge Bulking, Foaming, and Other Solids Separation Problems, thirded. Lewis Publishers, CRC Press, NY, USA.

Judd, S., 2011. In: Judd, S., Judd, C. (Eds.), The MBR BOOK: Principles and Applica-tions of Membrane Bioreactors for Water and Wastewater Treatment, seconded. Elsevier Ltd., Oxford, UK.

Jun, Z., Yang, F.L., Meng, F.G., Peng, A., Di, W., 2007. Comparison of membranefouling during short-term ltration of aerobic granular sludge and activatedsludge. J. Environ. Sci. China 19, 1281e1286.

Kim, I.S., Jang, N., 2006. The effect of calcium on the membrane biofouling in themembrane bioreactor (MBR). Water Res. 40, 2756e2764.Kovarova-Kovar, K., Egli, T., 1998. Growth kinetics of suspended microbial cells:

from single-Substrate-controlled growth to mixed-substrate kinetics. Microbiol.Mol. Biol. Rev. 62 (3), 646e666.

Kraume, M., Drews, A., 2010. Membrane bioreactors in waste water treatment estatus and trends. Chem. Eng. Technol. 33, 1251e1259.

Kumar, M., Adham, S.S., Pearce, W.R., 2006. Investigation of seawater reverseosmosis fouling and its relationship to pretreatment type. Environ. Sci. Technol.40, 2037e2044.

Laspidou, C.S., Rittmann, B.E., 2002. A unied theory for extracellular polymericsubstances, soluble microbial products, and active and inert biomass. WaterRes. 36, 2711e2720.

Lee, J., Ahn, W.Y., Lee, C.H., 2001. Comparison of the ltration characteristics be-tween attached and suspended growth microorganisms in submerged mem-brane bioreactor. Water Res. 35, 2435e2445.

Li, J.F., Li, Y.Z., Ohandja, D.-G., Yang, F.L., Wong, F.S., Chua, H.-C., 2008. Impact of lamentous bacteria on properties of activated sludge and membrane-foulingrate in a submerged MBR. Sep. Purif. Technol. 59 (3), 238e243.

Li, J.F., Zhang, X.X., Cheng, F.Q., Liu, Y., 2013. New insights into membrane fouling insubmerged MBR under sub-critical ux condition. Bioresour. Technol. 137,404e408.

Liao, B.Q., Lin, H.J., Langevin, S.P., Gao, W.J., Leppard, G.G., 2011. Effects of temper-ature and dissolved oxygen on sludge properties and their role in bio-occulation and settling. Water Res. 45, 509e520.

Lin, H.J., Liao, B.Q., Chen, J.R., Gao, W.J., Wang, L.M., Wang, F.Y., Lu, X.F., 2011. Newinsights into membrane fouling in a submerged anaerobic membrane biore-actor based on characterization of cake sludge and bulk sludge. Bioresour.Technol. 102, 2373e2379.

Lin, H.J., Zhang, M.J., Wang, F.Y., Meng, F.G., Liao, B.Q., Hong, H.C., Chen, J.R.,Gao, W.J., 2014. A critical review of extracellular polymeric substances (EPSs) inmembrane bioreactors: characteristics, roles in membrane fouling and controlstrategies. J. Membr. Sci. 460, 110e125.

Liu, Y., Fang, H.H.P., 2003. Inuences of extracellular polymeric substances (EPS) onocculation, settling, and dewatering of activated sludge. Crit. Rev. Environ. Sci.Technol. 33, 237e273.

Lowery, O.H., Rosebrough, N., Farr, A.L., Randall, R.J., 1951. Protein measurementwith the folin phenol reagent. J. Biol. Chem. 193, 265e275.

Marrot, B., Barrios-Martinez, A., Moulin, P., Roche, N., 2004. Industrial wastewatertreatment in a membrane bioreactor: a review. Environ. Prog. 23, 59e68.

McAdam, E.J., Judd, S.J., 2007. Denitrication from drinking water using a mem-brane bioreactor: chemical and biochemical feasibility. Water Res. 41,4242e4250.

Meng, F.G., Zhang, H.M., Yang, F.L., Li, Y.S., Xiao, J.N., Zhang, X.W., 2006a. Effect of lamentous bacteria on membrane fouling in submerged membrane bioreactor.

J. Membr. Sci. 272, 161e168.Meng, F.G., Yang, F.L., Xiao, J.N., Zhang, H.M., Gong, Z., 2006b. A new insight into

membrane fouling mechanism during membrane ltration of bulking andnormal sludge suspension. J. Membr. Sci. 285, 159e165.

Meng, F.G., Chae, S.R., Drews, A., Kraume, M., Shin, H.S., Yang, F.L., 2009. Recentadvances in membrane bioreactors (MBRs): membrane fouling and membranematerial. Water Res. 43, 1489e1512.

Meng, Q.J., Yang, F.L., Liu, L.F., Meng, F.G., 2008. Effects of COD/N ratio and DOconcentration on simultaneous nitrication and denitrication in an airlift in-ternal circulation membrane bioreactor. J. Environ. Sci. China 20, 933e939.

Miqueleto, A.P., Dolosic, C.C., Pozzi, E., Foresti, E., Zaiat, M., 2010. Inuence of carbonsources and C/N ratio on EPS production in anaerobic sequencing batch biolmreactors for wastewater treatment. Bioresour. Technol. 101, 1324e1330.

Morel, F.M.M., 1983. Principles of Aquatic Chemistry. John Wiley Interscience, USA,New York.

Omoike, A., Chorover, J., 2004. Spectroscopic study of extracellular polymeric sub-stances from Bacillus subtilis: aqueous chemistry and adsorption effects. Bio-macromolecules 5, 1219e1230.

Rosenberger, S., Kraume, M., 2003. Filterability of activated sludge in membranebioreactors. Desalination 151, 195e200.

Sanin, F.D., Vatansever, A., Turtin, I., Kara, F., Durmaz, B., Sesay, M.L., 2006. Opera-

tional conditions of activated sludge: inuence on occulation and dewater-ability. Dry. Technol. 24, 1297e1306.Shin, H.S., Kang, S.T., Nam, S.Y., 2000. Effect of carbohydrates to protein ratio in EPS

on sludge settling characteristics. Biotechnol. Bioprocess Eng. 5, 460e464.Tchobanoglous, G., Burton, F., 1991. Wastewater Engineering: Treatment, Disposal,

and Reuse, third ed. Metcalf and Eddy Inc., McGraw-Hill, New York.Visvanathan, C., Ben Aim, R., Parameshwaran, K., 2000. Membrane separation

bioreactors for wastewater treatment. Crit. Rev. Environ. Sci. Technol. 30, 1e48.Wang, Z.W., Wu, Z.C., Tang, S.J., 2009. Extracellular polymeric substances (EPS)

properties and their effects on membrane fouling in a submerged membranebioreactor. Water Res. 43, 2504e2512.

Wang, Z.W., Han, X.M., Ma, J.X., Wang, P., 2013. Recent advances in membranefouling caused by extracellular polymeric substances: a mini-review. Desalin.Water Treat. 51 (25e27), 5121e5131.

Wu, B., Fane, A.G., 2012. Microbial relevant fouling in membrane bioreactors:inuencing factors, characterization, and fouling control. Membranes 2,565e584.

Wu, Y.C., Smith, E.D., Novak, R., 1982. Filterability of activated sludge in response togrowth-conditions. J. Water Pollut. Control Fed. 54, 444e456.


influence of codn ratio on sludge properties and their role in membrane fouling of a submerged mbr

Documents