Influence of shear on the production of extracellular polymeric substances in membrane bioreactors

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 5

Avai lab le a t www.sc iencedi rec t .com

journa l homepage : www.e lsev ie r . com/ loca te /wat res

Influence of shear on the production of extracellularpolymeric substances in membrane bioreactors

Adrienne Mennitia, Seoktae Kangb, Menachem Elimelechb, Eberhard Morgenrotha,c,*aDepartment of Civil and Environmental Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USAbDepartment of Chemical Engineering, Environmental Engineering Program, Yale University, New Haven, CT 06520, USAcDepartment of Animal Sciences, University of Illinois at Urbana-Champaign, Urbana, IL 61801, USA

a r t i c l e i n f o

Article history:

Received 16 November 2008

Received in revised form

8 June 2009

Accepted 11 June 2009

Published online 1 July 2009

Keywords:

Membrane bioreactor

Extracellular polymeric substances

Soluble microbial products

EPS

SMP

Shear

* Corresponding author. Present address: SwDubendorf, Switzerland.

E-mail address: eberhard.morgenroth@ea0043-1354/$ – see front matter ª 2009 Elsevidoi:10.1016/j.watres.2009.06.052

a b s t r a c t

Shear is used to control fouling in membrane bioreactor (MBR) systems. However, shear also

influences the physicochemical and biological properties of MBR biomass. The current study

examines the relationship between the level of shear and extracellular polymeric substance

(EPS) production in MBRs. Two identical MBRs were operated in parallel where the biomass in

one reactor was exposed to seven times greater shear forces. The concentrations of floc-

associated and soluble EPS were monitored for the duration of the experiment. The stickiness

of extracted floc-associated EPS from each reactor was also characterized using atomic force

microscopy. A mathematical model of floc-associated and soluble EPS production was

applied to quantitatively describe changes in EPS production with shear. Biomass grown in

a high shear environment has lower floc-associated EPS production compared to biomass

grown in a lower shear environment. This decrease in floc-associated EPS production also

corresponds to a decrease in soluble EPS production, which can be explained by both the

lower concentration of floc-associated EPS and the production of stickier floc-associated EPS

that is more erosion resistant in the high shear reactor. This research suggests that

mechanical stresses can have a significant impact on the production rates of floc-associated

and soluble EPSdkey parameters governing membrane fouling in MBRs.

ª 2009 Elsevier Ltd. All rights reserved.

1. Introduction membrane surface to control this fouling. The level of shear is

Water stress, both in the Unites States and worldwide, creates

the need to generate new water resources through water

reuse. Membrane bioreactors (MBR) combine biological

wastewater treatment with membranes for solid–liquid

separation, providing an ideal technology for water reclama-

tion. However, membrane fouling still represents a significant

cost and energy burden for the system. Biomass is deposited

on the membrane during process operation forming a cake

layer and resulting in flux decline. Shear is applied along the

iss Federal Institute of Aq

wag.ch (E. Morgenroth).er Ltd. All rights reserved

a key process parameter in MBR systems as the flux of product

water through the membrane is directly related to this

parameter (Kim and DiGiano, 2006). However, MBRs are

dynamic biological systems. Shear also affects the physico-

chemical and biological properties of MBR biomass.

The microorganisms in activated sludge flocs reside in

a complex matrix of proteins, polysaccharides, lipids and

nucleic acids, which is referred to as floc-associated EPS (Liu

and Fang, 2003). EPS plays a key role in the formation of acti-

vated sludge flocs (Liao et al., 2001; Wilen et al., 2003) and

uatic Science and Technology (Eawag), Uberlandstrasse 133, 8600

.

Table 1 – Experimental conditions.

Parameter Short-termshear

Long-term shear

Experiment1

Experiment2

Shear rate (s�1) 1840 160 (L.S.)

1124 (H.S.)

160 (L.S.)

1124 (H.S.)

Duration 6 h 134 days 56 days

SRT (d) n/a 28 14

HRT (h) n/a 18 16.5

Influent COD

(mg/L)

n/a 400 450

Influent NH4þ

(mg N/L)

n/a 150 45

C/N ratio

(g COD/g N)

n/a 2.67 10

n/a¼ not applicable; L.S. = low shear; H.S. = high shear.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 54306

provides stability to flocs in high shear environments (Mik-

kelsen et al., 2002). Extracellular enzymes responsible for

protein and lipid hydrolysis are also found in the EPS matrix,

allowing the organisms to take up macromolecular and

particulate substrates that are too large to diffuse into the cell

(Gessesse et al., 2003).

Namkung and Rittmann (1986) defined soluble microbial

products (SMP) as compounds of microbial origin that are

released into solution during substrate utilization and

biomass decay. The majority of effluent organic matter in

biological wastewater treatment consists of these microbially

derived soluble compounds rather than original substrate

(Namkung and Rittmann, 1986). Alternately, Nielsen et al.

(1997) called the soluble compounds of microbial origin soluble

EPS and defined them as floc-associated EPS compounds that

have been release into solution either by bacterial self-

hydrolysis or by shear-induced erosion. Nielsen et al. (1997)

also included cell lysis products and hydrolysis products from

influent substrate degradation in their definition of soluble

EPS. Laspidou and Rittmann (2002a) provide a convincing

argument that SMP and soluble EPS refer to the same set of

microbially derived soluble polymers. In MBRs, these soluble

compounds interact with the membrane surface and clog the

membrane pores (Ho and Zydney, 2000; Ognier et al., 2002).

Soluble EPS is widely considered the most important foulant

in MBR systems (Chang et al., 2002; Le-Clech et al., 2006).

Short term increases in shear have been shown to increase

the release of soluble EPS by eroding floc-associated EPS into

solution (Kim et al., 2001; Wisniewski and Grasmick, 1998).

The production of floc-associated EPS may also be linked to

shear intensity. Increased shear has been shown to cause the

overproduction of polysaccharides in the EPS of biofilm

systems (Ramasamy and Zhang, 2005). This increase in poly-

saccharide EPS production is thought to aid in bacterial

adhesion (Liu and Tay, 2002). However, increased poly-

saccharide production in response to elevated shear may be

a relatively short-term phenomenon as Ramasamy and Zhang

(2005) showed that polysaccharide production decreased after

15 to 20 days.

Soluble carbohydrates are often found to be the most domi-

nant EPS compound responsible for membrane fouling in MBRs

(Fan et al., 2006; Rosenberger et al., 2006; Zhang et al., 2006).

Thus, increased carbohydrate production as a response to high

shear in MBR systems coupled with increased erosion rates of

floc-associated EPS could result in increased membrane fouling.

However, few studies have directly examined the effects of

shear on EPS production in flocculated biological systems.

Similar to the biofilm studies discussed previously, Liu et al.

(2005) showed an increase in the floc-associated polysaccharide

concentration with increased shear in short-term batch tests

with activated sludge, but the authors did not study long-term

changes in EPS production. Ji and Zhou (2006) and Meng et al.

(2008) both examined long-term changes in EPS production with

increased shear in MBRs, but the results of these two studies are

contradictory. Although both studies found an initial increase in

soluble EPS release with increased shear, Ji and Zhou (2006)

observed decreased floc-associated EPS production with

increased shear while Meng et al. (2008) found increased floc-

associated EPS production with increased shear. EPS production

in the Ji and Zhou study may have also been influenced by low

dissolved oxygen concentrations (less than 2 mg/L) at lower

shear levels. As a result, the long-term relationship between

shear and EPS production in MBR systems remains unclear.

The specific objective of the current study is to evaluate the

influence of shear on floc-associated EPS production, soluble

EPS release, and membrane fouling potential. The long-term

effects of shear on floc-associated EPS production and soluble

EPS release were characterized for MBR biomass. Atomic force

microscopy was used to characterize the stickiness of floc-

associated EPS produced in differing shear environments. The

fouling potential of different biomass fractions was quantified

using dead-end filtration. A mathematical model of floc-

associated and soluble EPS production is applied to quantita-

tively describe changes in EPS production with shear.

2. Materials and methods

2.1. Experimental systems

2.1.1. Short-term shear experimentsShort-term shear experiments were performed with biomass

from three municipal activated sludge wastewater treatment

plants (WWTP) located in Champaign, Urbana, and Danville,

Illinois. On the day of sampling, the biomass was subjected to

shear in the form of vigorous mixing with a 2-inch (5.08 cm)

Rushton mixing impellor at 1000 revolutions per minute (rpm).

The tank used had the standard reactor geometry defined by

Holland and Chapman (1966). Therefore the average velocity

gradient in the reactor, G, can be based on the mixing intensity.

The duration and shear intensity of the short-term experiments

are provided in Table 1. All short-term experiments were per-

formed in the absence of electron donor to minimize biological

growth.

2.1.2. Long-term shear experimentsTo study the long-term effects of shear, two 9.4 L membrane

bioreactors (Wilson et al., 2002; Zein et al., 2004) were operated

continuously for organic carbon oxidation and nitrification.

A schematic of the MBRs is shown in Fig. 1a and a picture of the

porous plastic membrane is provided in Fig. 1b. The

Mixedliquor

Effluent

NutrientsDilution water

Sludgewaste

4 M NaOH4 M HCl

pHDO

Rigidmembrane

Air

a

b

Fig. 1 – (a) Schematic of MBR with rigid polyethylene

membrane. (b) Picture of the rigid membrane installed

inside the MBRs. Mixed liquor remains inside the

membrane while permeate passes through the porous

walls of the reactor.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 5 4307

experimental conditions are summarized in Table 1. The reac-

tors were operated identically except one reactor experienced

seven times greater shear (Table 1). The MBRs used in the long-

term study were constructed with a rigid membrane directly

incorporated into the walls of the reactor and were designed to

havea standardreactor geometry (Holland and Chapman,1966).

This design allows the direct correlation of impeller rotation

with shear and avoids the complex, poorly characterized

hydrodynamic conditions in standard MBR systems. The

membrane was a 0.48 cm thick porous polyethylene plastic

(Atlas Mineral & Chemical, Mertztown, PA) with a nominal pore

size of about 20 mm. Filtrate passed through the membrane by

gravity and the liquid height in the reactor was maintained

hydraulically. Mixing was achieved with 3-inch (7.62 cm)

diameter Rushton mixing impellors. Biofilms were removed

from the walls of the reactor approximately biweekly during the

long-term shear experiments.

Dissolved oxygen (DO) and pH were measured on-line (WTW

Multi 340i, Weilheim, Germany). The DO and pH data were

collected using a Labview (National Instruments, Austin, TX)

program. The Labview program also controlled the influent

pumps and level sensors. A pH control loop maintained the pH

between 6.5 and 7.5 by adding 4 M HCl or 4 M NaOH as needed.

The solids retention time (SRT) was maintained by wasting

a fixed volume of mixed liquor daily.

The effect of different levels of shear on EPS production was

evaluated in two long-term experiments performed with

differing SRTs and different influent carbon to nitrogen ratios as

shown in Table 1. A sterile, concentrated nutrient feed con-

taining acetate and ammonia was diluted with deionized water

to achieve the desired influent concentrations. Trace mineral

salts were added to the sterile, concentrated nutrient feed while

calcium and iron were added to the dilution water to avoid their

precipitation during autoclaving. Influent concentrations of

mineral salts in the overall synthetic medium were: 3.95 mg/L

MgSO4$7H2O, 14.9 mg/L KH2PO4, 2.2 mg/L MnCl2$4H2O, 4.41 mg/L

CoSO4$7H2O, 7.4 mg/L CuCl2$2H2O, 1.76 mg/L NiSO4$6H2O,

2.64 mg/L Na2MoO4$2H2O, 3.67 mg/L ZnSO4$7H2O, 1.47 mg/L

H3BO3, 4.48 mg/L CaCl2$2H2O, 0.133 mg/L FeCl3.

2.1.3. Reactor performance monitoringChemical oxygen demand (COD), mixed liquor suspended

solids (MLSS) and mixed liquor volatile suspended solids

(MLVSS) were measured according to Standard Methods for

the Examination of Water and Wastewater (APHA et al., 1998).

Ammonia nitrogen was measured either using the Nessler

method (APHA et al., 1998) or using a microtiter based method

(Rhine et al., 1998). Nitrate and nitrite were measured either

using a microtiter based method (Mulvaney, 1994) or using ion

chromatography (APHA et al., 1998). Samples for COD,

ammonia, nitrate and nitrite characterization were filtered

using a 0.45 mm cellulose acetate filter. All samples were

measured on the day of collection.

2.1.4. Reactor operation

2.1.4.1. Experiment 1. The biomass for experiment 1 was first

acclimated to the high influent ammonia concentration,

which could be toxic to organisms unaccustomed to the feed.

Both MBRs were started with biomass from a full-scale

nitrifying CAS process and operated at low shear (G¼ 160 s�1)

throughout the start-up period. The influent ammonia

concentration was slowly increased from 50 mg/L N to

150 mg/L N over an acclimation period of six weeks. The C:N

ratio was held constant at every ammonia increase and 100%

ammonia removal was confirmed before increasing the

ammonia concentration of the influent. To begin the shear

experiment, the contents of the reactors were mixed together,

redistributed and the shear was increased in the high shear

MBR. This day is represented as day zero in the presentation of

the results.

Both MBRs were operated with complete removal of acetate

and ammonia until day 68 when the high shear reactor experi-

enced an upset in nitrification. The nitrification performance in

the high shear reactor remained unstable until day 90 when

complete ammoniaremovalwasagain achieved. However, after

the reactor upset, nitrite accumulation started in the high shear

reactor with nitrite accounting for about 60% of the reduced

nitrogen compounds in the reactor. The low shear reactor also

experienced an upset in nitrification performance on day 89.

The ammonia concentration was decreased to 100 mg/L N for

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 54308

both reactors on day 96. Following the decrease in influent

ammonia concentration on day 96, the low shear MBRrecovered

and both reactors remained at complete ammonia removal for

the remainder of the experiment. Nitrite accumulation was

never an issue in the low shear reactor.

One consequence of the low shear reactor upset on day 89

wasa shift in the reactor solids concentration. Before the reactor

upset, the average volatile suspended solids concentration from

day zero to day 79 was 1616.5� 62.8 mg/L (average� standard

error) for the low shear reactor and 1768.8� 88.9 mg/L for the

high shear reactor. After the reactor upset and subsequent

decrease in influent concentration, the average volatile sus-

pended solids in the low shear reactor decreased to

803.4� 34.1 mg/L while the high shear reactor only decreased to

1200.3� 43.7 mg/L.

2.1.4.2. Experiment 2. The long-term shear experiment was

repeated with differing C:N ratio and SRT to ensure that the

observed changes were related to the shear conditions being

studied and not other reactor operational parameters. For

experiment 2, the reactors were seeded with sludge from the

same full-scale nitrifying CAS treatment plant used for experi-

ment 1. The reactors were started and operated at low shear

(G¼ 160 s�1) for five days before increasing the shear in the high

shear reactor. Complete nitrification was achieved at the target

influent ammonia concentration after four days. Both reactors

were operated with complete removal of ammonia and acetate

for more than three SRTs. A slight upset in nitrification occurred

in the low shear reactor on day 48 of reactor operation so data

collection for this reactor was stopped after this point. Nitrite

accumulation was never an issue in either reactor for the

duration of the experiment. The average volatile suspended

solids concentration was 795.3� 51.4 mg/L for the low shear

reactor and 805.8� 39.1 mg/L for the high shear reactor.

2.2. Analyses

2.2.1. Extracellular polymeric substance quantificationFloc-associated extracellular polymeric substances were

extracted using a cation exchange resin (CER) following Frolund

et al. (1996). The extraction was performed by mixing CER,

previously washed in phosphate buffer, with solids at a ratio of

70 g of CER per gram of volatile suspended solids. Soluble

compounds were washed from the solids by centrifuging at

12,000 g for 15 min at 4 �C, decanting the supernatant, then

resuspending the remaining solids in phosphate buffer. The

washing procedure was performed twice. The solids and CER

were mixed for four hours at a rate of 175 rpm at 4 �C. The

volume of mixed liquor required for extraction was based on

aMLVSSmeasurement performedtheday before theextraction.

The solids concentration was then verified on the day of

extraction. Extracted EPS was harvested by filtering the CER and

solids mixture with a 1.2 mm Whatman 934-AH glass fiber filter.

Floc-associated EPS was quantified as the total mass of

proteins and carbohydrates extracted per gram of MLVSS.

Soluble EPS was separated from the mixed liquor using a glass

fiber filter (as described below for membrane fouling potential

experiments) and quantified by measuring the concentration

of proteins and carbohydrates. Proteins were measured using

the Lowry method (Gerhardt et al., 1994) with bovine serum

albumin (BSA) as a standard. Carbohydrates were measured

using the Anthrone method (Gerhardt et al., 1994) with

glucose as a standard. A fresh standard curve ranging from

5–100 mg/L BSA or glucose was prepared with each measure-

ment. All EPS samples for protein and carbohydrate analysis

were not further filtered prior to analysis. The samples were

frozen at �20 �C on the day of collection and stored at this

temperature prior to analysis.

The molecular weight distribution of solubleEPS compounds

was quantified using high-pressure size exclusion chromatog-

raphy (SEC) following Zhou et al. (2000). The instrument used

was a Shimadzu VP series high performance liquid chromato-

graph system (Kyoto, Japan) equipped with a photodiode array

detector. The SEC column, a Waters Protein-Pak 125 (Milford,

MA) with a molecular weight range of 2–80 kDa, was calibrated

with sodium polystyrene sulfonate standards (Polysciences,

Inc., Warrington, PA) of sizes 1.8, 4.6, 8, 18, 35, 67 and 780 kDa.

The wavelength range from 200–300 nm was detected contin-

uously in 1 nm increments during measurement. The chro-

matographs presented here were analyzed at a wavelength of

205 nm. This wavelength was chosen because all the samples

measured as well as BSA exhibit a strong peak in absorbance at

this wavelength.Samples for SEC analysiswere from the soluble

fraction (separated as described below for membrane fouling

potential measurements) and further filtered using a 0.22 mm

cellulose acetatemembrane.All SEC sampleswere frozenonthe

day of collection and stored at �20 �C prior to analysis.

2.2.2. Membrane fouling potential measurementFouling potential measurements were performed using an

Amicon stirred cell model 8200 (Millipore, Billerica, MA) mixed

at 175 rpm and pressurized to 1.02 atm (15 pounds per square

inch) of pressure using nitrogen. For short-term shear exper-

iments, 20 kDa polyethersulfone (PES) ultrafiltration (UF)

membranes were used while 30 kDa polyvinylidene fluoride

(PVDF) UF membranes were used for long-term shear experi-

ments. All membranes were purchased from GE Osmonics

(Minnetonka, MN). The clean water flux of the membrane was

characterized for approximately 45 min before measuring the

sample flux and a fresh piece of membrane material was cut

for each experiment.

Flux measurements were performed by weighing permeate

at fixed time intervals on a top loading balance (Model PB3002-S,

Mettler Toledo, Columbus, OH). Flux was calculated assuming

a permeate density of 1 g/mL. Automatic collection of permeate

data was performed using the software WinWedge (TAL Tech-

nologies, Philadelphia, PA). Clean water flux measurements

were performed using NANOpure ultrapure water (Barnstead,

Dubuque, IA).

Following Rosenberger and Kraume (2002), the fouling

potential is defined as the resistance to filtration after filtering

a fixed volume of sample. For short-term shear experiments,

fouling potential is defined as the resistance to filtration after

filtering 40 L of sample per square meter of membrane area

(L/m2). For long-term shear experiments, the fouling potential

is defined as the resistance to filtration after filtering 60 L/m2.

The biomass was separated into two fractions for the

short-term shear tests and for long-term shear experiment

2 in order to quantify the importance of the soluble fraction to

overall membrane fouling. For short-term shear experiments,

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 5 4309

the biomass was centrifuged for 5 min at 1500 g to remove the

particulate matter. The maximum particle size remaining in

the supernatant was calculated to be 6 mm. For long-term

shear experiment 2, the biomass was fractionated using

a Whatman 934-AH glass fiber filter with a nominal pore size

of 1.2 mm.

2.2.3. Atomic force microscopy (AFM) measurementsAtomic force microscopy was used to characterize the inter-

facial forces between the extracted floc-associated EPS mole-

cules. The force measurements were performed with

a MultiMode AFM connected to a Nanoscope IIIa controller

(Veeco Metrology Group, Santa Barbara, CA). A carboxylate

modified latex (CML) particle (Interfacial Dynamics Corp.,

Portland, OR) was used to make the AFM colloid probe because

proteins and carbohydrates, the largest fraction of polymers

in floc-associated EPS, contain predominantly carboxylic

functional groups. The CML particle (3.0 mm in diameter) was

attached by Norland optical adhesive (Norland Products Inc.,

Cranbury, NJ) to a tipless SiN cantilever having a spring

constant of 0.06 N/m (Veeco Metrology Group, Santa Barbara,

CA) and cured under ultraviolet light for 20 min.

A liquid cell was used to allow force measurements of the

EPS sample solutions, with a polymeric membrane being

located on the bottom of the liquid cell. The liquid cell was

first rinsed with DI water before injecting the EPS test solution

(diluted with phosphate buffer to 110 mg/L COD) to fully

displace the DI water in the cell. The solution was then

equilibrated for 60 min to allow EPS to adsorb onto the CML

probe and the membrane surface (Ang and Elimelech, 2007).

Therefore, the measured interaction force between the EPS-

coated CML probe and the EPS on the membrane surface

provides an estimation of the interaction force between EPS

molecules. The force measurements were conducted at four

different locations on the EPS-coated membrane with 10

force measurements taken at each location. Details on the

procedures for using an AFM colloid probe to determine

intermolecular adhesion forces are given in Lee and Elimelech

(2006).

2.3. Mathematical modeling of EPS production andparameter estimation

A mathematical model of EPS production is used to evaluate

possible mechanisms resulting in experimentally observed

floc-associated and soluble EPS production rates. The math-

ematical formulation of all components of the model is based

on unified theory of extracellular polymeric substances and

soluble microbial products proposed by Laspidou and Ritt-

mann (2002b). The mechanisms and relevant parameters of

floc-associated and soluble EPS production are described

graphically in Fig. 2. The production of both floc-associated

EPS (XEPS) and utilization associated products (SUAP) are

proportional to the substrate utilization rate with stoichio-

metric factors kEPS and k1 for both processes, respectively.

Floc-associated EPS is hydrolyzed producing biomass associ-

ated products (SBAP) with a rate constant khyd. SUAP and SBAP

are two forms of soluble EPS that cannot be chemically

distinguished. A matrix of all processes incorporated into the

parameter estimation and the parameter definitions and units

is provided in the Supplementary information.

Estimation of kEPS and khyd was performed by comparing

measured floc-associated and soluble EPS concentrations

with modeled concentrations of XEPS and the sum of SBAP and

SUAP, respectively. A constant value was assumed for k1 of

0.05 g COD/g COD based on the value provided in Laspidou and

Rittmann (2002b). Approximate confidence regions were

calculated based on

Sc ¼ SR

�1þ p

n� pFp;n�p;a

�(1)

where Sc is the critical sum of squares for a given confidence

level, (1-a)100%. SR is the residual sum of squares, p is the

number of parameters (i.e., p¼ 2), n is the number of observa-

tions, and Fp,n�p,a is the upper a percent value of the F distri-

bution with p and n�p degrees of freedom (Berthouex and

Brown, 2002). All simulations and parameter estimations were

performed using AQUASIM (Reichert, 1994). Approximate

confidence regions were evaluated by performing 10,000

simulations using different randomly generated parameter

combinations of kEPS and khyd and selecting parameter combi-

nations with sum of square errors smaller than Sc in Eq. (1).

State variables for organic matter in the mathematical

model are based on COD units. Measured compounds were

converted to COD units assuming conversion factors of

1.06 g COD/g glucose for carbohydrates, 1.4 g COD/g BSA for

proteins, and 1.42 g COD/g VSS for suspended solids. The

porous plastic membrane used in the long-term MBR experi-

ments was modeled with complete retention of particulate

compounds (Xa, XEPS) but no selective retention of soluble

compounds (S, SBAP, SUAP). Note that the pores in the porous

membrane are much larger than pores in conventional

ultrafiltration or microfiltration membranes.

3. Results and discussion

3.1. Short-term shear

The shear-induced release of soluble EPS was observed during

short-term experiments as an increase of soluble protein

concentration, shown in Fig. 3a. The concentration of soluble

proteins increased after 6 h of shearing by 40%, 38% and 20%

for biomass from Urbana, Champaign, and Danville waste-

water treatment plants, respectively. The soluble carbohy-

drate concentrations before and after shear were all below the

range of the standard curve (less than 5 mg/L). The increase in

proteins with short-term shear agrees with the observations

of Kim et al. (2001) and Wisniewski and Grasmick (1998), who

both observed an increase in soluble COD with shear, and

provides an indication of the relative importance of physical

erosion as a mechanism of soluble EPS production in this

system.

The SEC plots in Fig. 3b show a bimodal molecular weight

distribution of soluble EPS from all three wastewater treat-

ment plants tested before and after shear. The peak of mole-

cules eluting between 6 and 7 min represents molecules larger

than 80 kDa, the maximum separation size of the column.

Substrate (S)Active biomass (Xa)

Floc associated EPS (XEPS)

Biomass associated products (SBAP)

Utilization associated products (SUAP)

Growth

kEPS·rS

k1·rS

Hydrolysiskhyd.XEPS

Solu

ble

EPS

Bio

mas

s

Fig. 2 – Mechanisms for EPS production as the result of two processes: (i) growth and (ii) hydrolysis of floc associated EPS.

State variables in the model are boxed, the two processes are labeled in bold, and process rates are shown in italics (rS is the

rate of substrate utilization). The overall biomass is a combination of Xa and XEPS and the soluble EPS is the sum of SBAP and

SUAP. Model details and parameter values are provided in the Supplemental information.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 54310

Proteins, carbohydrates and organic colloids are expected to

elute in this size range (Rosenberger et al., 2006). Previous

research suggests that molecules in this high molecular

weight peak are largely responsible for membrane fouling in

surface water (Howe and Clark, 2002), wastewater effluent

organic matter (Laabs et al., 2006), and MBR supernatant

(Rosenberger et al., 2006). The second peak of molecules,

0

2

4

6

8

10

12

14

16

18a

b

Urbana, ILWWTP

Champaign,IL WWTP

Danville, ILWWTP

So

lu

ble P

ro

tein

s (m

g/L

)

Befo

re S

hear

After S

hear

Befo

re S

hear

After S

hear

Befo

re S

hear

After S

hear

0

10

20

30

40

50

5 6 7 8 9 10 11 12

UV

A at 205 n

m

Urbana, IL WWTPChampaign, IL WWTPDanville, IL WWTP

Before ShearAfter Shear

Elution Time (min)

Size of Molecule

Moleculeslarger than80 kDa

Fig. 3 – Increase in (a) soluble protein concentrations and

(b) the high molecular weight peak (molecules larger than

80 kDa) in size exclusion chromatography (SEC) as the

result of short-term shearing.

which exits the column between 10 and 11 min just before the

salt boundary, represents small molecules in the size range of

organic acids (Zhou et al., 2000) with molecular weights close

to the lower separation limit of the SEC column (2 kDa).

The concentration of molecules along the entire range of

molecular weights increased with shear (Fig. 3b). However, the

most notable feature is the strong increase in the concentra-

tion of high molecular weight molecules. The high molecular

weight peak area increased by 94% for the Urbana biomass

and increased by 79% for the Champaign biomass. There was

variation in the shear response of biomass from different

sources as the peak area for the Danville biomass increased by

only 12%, which was in agreement with the smaller increase

in protein concentration for this sludge.

SEC characterization using a UVdetector ata low wavelength

(205 nm) is well suited to measure changes in the protein

concentrations because proteins exhibit relatively high molar

absorptivity at low wavelengths (Her et al., 2002b). However,

carbohydrates contain primarily carbon-carbon single bonds,

which cannot be detected by UV absorbance measurements

(Her et al., 2002a). Therefore, the SEC measurements presented

here primarily represent the molecular weight distributions of

proteinaceous material in the soluble EPS.

The change in membrane fouling potential with shear and

the contribution of the soluble fraction to overall membrane

fouling are presented in Fig. 4. All three sludge samples exhibit

a strong increase in membrane fouling potential after shearing.

The total membrane fouling potential of biomass from Urbana,

Champaign and Danville increased by 46%, 37%, and 31%,

respectively. Furthermore, the soluble fraction represents more

than 70% of the resistance to filtration for all biomass samples

(both before and after shear) and as much as 97% for the

Champaign biomass after shear. The intense shear conditions

maximize the effects of shear for these short-term shear

experiments. However, since the short-term shear experiments

are performed without substrate, cell lysis could contribute to

the increase in soluble EPS observed after six hours of vigorous

shear. The short duration of these experiments was chosen to

minimize the contribution of cell lysis products to soluble EPS.

Furthermore, the increase in membrane fouling potential due to

the shear-induced release of soluble EPS is consistent with the

0

1

2

3

4

5

6

7

8R

esistan

ce (x10

12 m

-1)

Soluble Particulate

Urbana, ILWWTP

Champaign, ILWWTP

Danville, ILWWTP

Be

fo

re

S

he

ar

Be

fo

re

S

he

ar

Afte

r S

he

ar

Be

fo

re

S

he

ar

Afte

r S

he

ar

Afte

r S

he

ar

Fig. 4 – Increase in membrane fouling potential as the

result of short-term shearing.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 5 4311

observations of others (Park et al., 2005; Rosenberger and

Kraume, 2002; Wisniewski and Grasmick, 1998). Thus, these

results suggest that increased mechanical shear, as used in the

long-term shear experiments, should increase the erosion of

floc-associated EPS.

3.2. Long-term shear

3.2.1. Experiment 1The reactors in experiment 1 were operated for more than

three SRTs before collecting EPS data to ensure sufficient time

to reach steady state at the shear conditions in each reactor.

Steady state floc-associated and soluble EPS concentrations

and membrane fouling potential results for the high and low

shear MBRs are summarized in Table 2. The soluble EPS

concentration was quantified as the area under the high

molecular weight peak resulting from the SEC measurement.

The average floc-associated EPS and soluble EPS concen-

trations in the low shear reactor were significantly greater

than those from the high shear reactor with 95% confidence.

Furthermore, the membrane fouling potential of the low shear

biomass was also significantly greater than the high shear

membrane fouling potential with 95% confidence. The

membrane fouling potential results provide a strong indica-

tion that the lower concentration of EPS compounds in the

high shear reactor corresponds to a lower membrane fouling

potential for the biomass in that reactor. More importantly,

these results demonstrate that the shear-induced removal of

Table 2 – Average results from long-term experiment 1 forfloc-associated EPS concentration, soluble EPSconcentration, and membrane fouling potential(average ± standard error).

Measurement Unit n Low shear High shear

Floc-associated EPS mg/g 6 46.4� 3.3 25.2� 1.2

Soluble EPS UVA$min 6 14.6� 0.53 12.3� 0.69

Membrane fouling potential 1012 m�1 3 5.07� 0.92 3.76� 0.80

EPS sampled 98, 105, 112, 118, 132, and 134 days after increasing

shear. Membrane fouling potential evaluated 104, 111, and 133 days

after increasing shear.

floc-associated EPS in the high shear reactor generates

biomass with both lower floc-associated and soluble EPS

concentrations. This lower EPS production decreased the

membrane fouling potential of the high shear biomass.

3.2.2. Experiment 2In experiment 2, floc-associated and soluble EPS concentra-

tions and membrane fouling potential were followed during the

start-up period of the reactors in order to examine the time

scale of the change in EPS production due to shear. The change

in floc-associated EPS concentration with time is shown in

Fig. 5a. The floc-associated EPS in the seed biomass was

quantified and day zero represents the increase in shear of the

high shear reactor. The floc-associated EPS concentration in

both reactors decreased significantly between day zero and day

10, suggesting an acclimation period to new reactor conditions

for both sludges. By day 10, the concentration of EPS in the low

shear reactor was already greater than the high shear reactor

and it remained higher for the rest of the experiment.

The soluble EPS results are shown in Fig. 5b and Fig. 6. The

soluble EPS results shown in Fig. 5b demonstrate lower

concentrations of soluble EPS in the high shear reactor, which

also had lower floc-associated EPS concentrations. The lower

concentration of soluble EPS in the high shear reactor is sup-

ported by the SEC results shown in Fig. 6. The high shear reactor

alsohad a lowerhighmolecularweight peakon days14and 50of

reactor operation. The results from experiment 2 are consistent

with experiment 1 where high shear conditions decreased both

floc-associated and soluble EPS concentrations.

The fouling potential of the total biomass and of the

soluble fraction for both MBRs is shown in Fig. 5c. The total

fouling potential as well as the fouling potential of the soluble

fraction is lower in the high shear reactor, further demon-

strating the importance of EPS concentration to membrane

fouling potential. However, the relative contribution of the

soluble fraction to overall membrane fouling potential was

greater in the high shear reactor than the low shear reactor

from day 27 to day 48.

The importance of the particulate fraction in the low shear

reactor is likely an indication of the higher floc-associated EPS

concentrations. The higher concentration of floc-associated

EPS in the low shear biomass could produce a cake layer with

a greater specific cake resistance (Cho et al., 2005), increasing

the importance of the particulate fraction in the low shear

reactor even in the presence of increased soluble EPS

concentrations.

3.3. Adhesion force and stickiness of EPS

The measured adhesion forces between the carboxylate

modified latex probe and extracted floc-associated EPS are

presented in Fig. 7. High and low shear floc-associated EPS

samples from days 112 and 130 of long-term shear experiment

1 and days 30 and 50 of experiment 2 were tested. The AFM

results indicate that while the high shear EPS has a lower

concentration, high shear EPS compounds exhibit stronger

adhesion forces. The greater adhesion force in the high shear

EPS suggests that increased mechanical erosion in the high

shear system generated a physiological change in the type of

EPS produced by the microorganisms. The possibility exists

0

20

40

60

80

100F

lo

c E

PS

(m

g/g

)

Low shear reactorHigh shear reactor

Low shear reactorHigh shear reactor

0

10

20

30

40

50

So

lu

ble C

OD

(m

g/L

)

0

2

4

6

8

10

12

-10 0 10 20 30 40 50 60

-10 0 10 20 30 40 50 60

-10 0 10 20 30 40 50 60

Resistan

ce (10

12 m

-1)

Day of Reactor Operation

Low shear reactor total foulingLow shear reactor soluble foulingHigh shear reactor total foulingHigh shear reactor soluble fouling

a

b

c

Fig. 5 – Long-term effect of low or high shear conditions

(Experiment 2): (a) Floc-associated EPS. Error bars indicate

the standard error between duplicate extractions.

(b) Soluble EPS measured using soluble COD. (c) Membrane

fouling potential for the overall and for the soluble fraction

of the mixed liquor.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 54312

that this physiological change – stickier EPS – also decreased

the CER extraction efficiency of the EPS, emphasizing the

observed decrease in EPS production. However, as with any

study on floc-associated EPS production, floc-association EPS

is defined by the characterization method and the character-

ization method applied here demonstrated decreased floc-

associated EPS production. The high shear biomass produced

less EPS but the polymers interact more strongly, providing

the organisms with a greater ability to remain flocculated in

the high shear environment.

The observation of decreased EPS production in the current

study differs from what has been described for biofilm

systems where increased shear resulted in an overproduction

of polysaccharide EPS (Ramasamy and Zhang, 2005). In these

biofilm systems, organisms able to increase EPS production in

response to shear gain a competitive advantage through the

ability to maintain biofilm growth under high shear condi-

tions. Both the biofilm system previously studied and the

porous pot MBR in the current study share the fact that their

bacteria have to aggregate to avoid washout. The 20 mm pore

size of the porous membrane material allows dispersed

organisms and very small flocs to washout with the effluent.

In the current study, MBR bacteria produced less but stickier

EPS while in the study of Ramasamy and Zhang (2005), the

biofilms increased the amount of polysaccharide EPS

produced. Parameter estimation for EPS related processes

(Fig. 2) was performed by assuming a fixed value for k1 of

0.05 g COD/g COD (based on Laspidou and Rittmann (2002b))

and then estimating kEPS and khyd. Best fit values and

approximate 75% confidence regions for kEPS and khyd are

shown in Fig. 8. Elongated confidence regions demonstrate

that the parameters are correlated and cannot be indepen-

dently estimated, but the fact that the confidence regions do

not overlap indicates that parameter values in the high and

low shear reactors are clearly different from each other. The

confidence regions in Fig. 8 show that for a fixed value for kEPS,

the high shear reactor always has a higher value of khyd.

A larger value of khyd with increased shear could be reasonable

as increased shear in short-term experiments resulted in

increased release of floc associated EPS (i.e. an increased value

of khyd). However, the least squares results in Fig. 8 indicate

a constant value of khyd for both reactors and a reduced value

of kEPS (i.e. reduced floc associated EPS production) in the high

shear MBR. Such a constant value of khyd would be consistent

with AFM results that indicated that the high shear floc-

associated EPS is more cohesive than the low shear floc-

associated EPS. This increased stickiness would reduce the

soluble EPS released for a given level of shear. Thus, a possible

interpretation of the experimental data combined with

parameter estimation is that a long-term increase in shear

forces could be balanced by stronger and more shear resistant

floc-associated EPS in the high shear reactor, resulting in very

similar first order hydrolysis rates (khyd) between the two

reactors even though the level of shear was seven times

greater in the high shear reactor.

The EPS model presented by Laspidou and Rittmann

(2002b) and used in this study provides two parallel pathways

for the production of soluble EPS. Soluble EPS is composed of

utilization associated products (SUAP) that are stoichiometri-

cally produced during growth and biomass associated prod-

ucts (SBAP) that result from the hydrolysis of floc-associated

EPS. Using bulk chemical parameters such as soluble COD,

proteins, or carbohydrates, SUAP and SBAP cannot be differen-

tiated. Furthermore, parameter estimation using experi-

mental results for bulk chemical parameters does not allow

the unique identification of individual parameter values, but

mathematical modeling does allow some possible mecha-

nisms to be ruled out. For example, from the results of short-

term experiments, it could be hypothesized that an increase

in shear would result only in an increase of hydrolysis of floc-

associated EPS (khyd) while the physiology of EPS production

remains unaltered (k1, kEPS constant). Based on our long-term

0

10

20

30

40

50

UV

A at 205 n

m

High Shear ReactorLow Shear Reactor

High Shear ReactorLow Shear Reactor

Moleculeslarger than80 kDa

Moleculeslarger than80 kDa

0

10

20

30

40

50

5 6 7 8 9 10 11 12

5 6 7 8 9 10 11 12

UV

A at 205 n

m

Elution Time (min)

Size of Molecule

a

b

Fig. 6 – Long-term effect of low or high shear conditions

(Experiment 2) on the high molecular weight peak

(molecules larger than 80 kDa) in size exclusion

chromatography (SEC) results: (a) day 14 of reactor

operation. (b) day 50 of reactor operation.

0.00

0.20

0.40

0.60

0.80

1.00

0.00 0.02 0.04 0.06 0.08 0.10

kh

yd (d

-1)

kEPS

(mg COD/mg COD)

Low Shear Reactor High Shear Reactor

Fig. 8 – 75% confidence region for estimates of the

parameter combination kEPS and khyd. The value of k1 is

fixed at 0.05 g COD/g COD. The least squares estimate for

each data set is shown as the red triangle. The least

squares estimates in the high shear reactor are

khyd [ 0.28 dL1 and kEPS [ 0.018 mg COD/mg COD and for

the low shear reactor are khyd [ 0.27 dL1 and

kEPS [ 0.039 mg COD/mg COD. (For interpretation of the

references to colour in this figure legend, the reader is

referred to the web version of this article.)

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 5 4313

results, such a simplified explanation can be ruled out as the

model would predict a decrease in the floc-associated EPS

concentration coupled with an increase in the soluble EPS

concentration. Such an increase in the soluble EPS concen-

tration does not match with experimental observations where

both floc-associated and soluble EPS concentrations

decreased with increased shear.

-0.45

-0.4

-0.35

-0.3

-0.25

-0.2

-0.15

-0.1

-0.05

00 50 100 150

F/R

(m

N/m

)

Floc EPS (mg/g)

Low Shear MBRHigh Shear MBR

Fig. 7 – Adhesion forces (normalized by the radius of the

AFM colloid probe) for low and high shear floc-associated

EPS. Results are shown for day 112 and 130 of experiment

1, and day 30 and day 50 of experiment 2. The error bars

indicate the standard error (n [ 4).

The reduction of both floc-associated and soluble EPS can

only be explained as a physiological response to shear

affecting k1 and/or kEPS. For example, a reduction of kEPS with

increasing shear can explain experimental results. These

qualitative results have significant implications for the oper-

ation of MBR systems. If mechanical stresses in MBRs have

a significant influence on the production rates of soluble EPS –

a key MBR foulant– then optimization of MBR operation could

aim at reducing the microbial production of the foulants.

Further research on the specific mechanisms of soluble EPS

production in pilot scale MBRs is necessary to quantify the

practical significance of these shear-related differences in EPS

production.

3.4. Selection mechanisms in MBRs

The lower floc-associated EPS concentrations with increased

shear found in this study agree with the observations of Ji and

Zhou (2006), while Meng et al. (2008) found increased floc-

associated EPS concentrations with increased shear. Unlike

both Ji and Zhou (2006) and Meng et al. (2008), who showed

increased soluble EPS concentrations with increased shear,

increased shear in the current study resulted in lower soluble

EPS production. Both previous studies examined MBR systems

with polymeric membranes that selectively retained not only

flocs but also the large molecular weight fraction of soluble EPS

within the reactor. The porous plastic MBR used in the current

study does not retain macromolecules, which is advantageous

to study mechanisms of EPS production as it reduces the influ-

ence of soluble EPS degradation. The soluble EPS produced by

erosion or hydrolysis of floc-associated EPS compounds is

assumed to be very slowly degradable based on the kinetic

parameters provided in Laspidou and Rittmann (2002b).

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 54314

An initial accumulation of soluble EPS with shear in both

previous MBR studies (Ji and Zhou, 2006; Meng etal., 2008) agrees

with this assessment. In microfiltration and ultrafiltration

membrane-based MBR systems, the accumulation of a slowly

degradable fraction makes it difficult to evaluate soluble EPS

production, retention and degradation independently.

In the MBR studied here, un-degraded soluble EPS leaves

the system with the effluent due to the large 20 mm pore size.

This pore size allows the production of soluble EPS to be

evaluated independently from its degradation since the time-

scale of degradation of soluble EPS produced by the erosion

mechanism is slower than the hydraulic retention time of the

reactor based on the parameters provided in Laspidou and

Rittmann (Laspidou and Rittmann, 2002b). However, this

larger pore size also allows washout of dispersed organisms

and small flocs, which means that the selection pressures in

the current study are somewhat different from MBRs with

microfiltration or ultrafiltration membranes. In traditional

MBR systems, all small flocs and dispersed organisms are

retained so the selection pressures for flocculated organisms

are lower. Although organisms in MBRs remain flocculated to

great extent, the concentration of floc-associated EPS in MBRs

is observed to be lower than in conventional activated sludge

(Merlo et al., 2007) and the concentration of dispersed organ-

isms is higher (Cicek et al., 1999). EPS is a major foulant in

MBRs and results from this study indicate that changes in

reactor operation (e.g., mechanical shear) can influence the

microbial physiology in terms of amount and type of EPS

production. As stated above, further research linking the

mechanisms of soluble EPS production to the process condi-

tions in MBR systems is necessary to gain an understanding of

the practical importance of the relationship. The selection

mechanisms in MBR systems (i.e., complete retention and

high shear) may provide valuable insights into the physiology

of soluble EPS production in MBRs.

4. Conclusions

� Biomass grown long-term in a high shear environment has

lower floc-associated EPS production compared to biomass

grown in a lower shear environment. This decrease in floc-

associated EPS production also corresponds to a decrease in

soluble EPS production. Decreased soluble EPS production

under high shear conditions can be explained by two factors:

(1) the lower concentration of floc-associated EPS and (2) the

production of stickier floc-associated EPS that is more erosion

resistant in the high shear reactor.

� Short-term increases in shear increase the release of soluble

EPS through the erosion of floc-associated EPS while long-

term exposure to high shear decreases soluble EPS produc-

tion. In both short and long-term experiments, larger

concentrations of soluble EPS resulted in increased

membrane fouling potential.

Acknowledgements

This work was partially supported by The WaterCAMPWS,

a Science and Technology Center of Advanced Materials for

the Purification of Water with Systems under the National

Science Foundation agreement number CTS-0120978.

Appendix.Supplementary information

Supplementary information associated with this article can be

found in the online version, at doi:10.1016/j.watres.2009.06.052.

r e f e r e n c e s

Ang, W.S., Elimelech, M., 2007. Protein (BSA) fouling of reverseosmosis membranes: implications for wastewaterreclamation. Journal of Membrane Science 296 (1–2), 83–92.

APHA, WEF, AWWA, 1998. Standard Methods for the Examinationof Water and Wastewater. American Public HealthAssociation, Washington, DC.

Berthouex, P.M., Brown, L.C., 2002. Statistics for EnvironmentalEngineers. CRC Press, Lewis Publishers, Boca Raton.

Chang, I.S., Le Clech, P., Jefferson, B., Judd, S., 2002. Membranefouling in membrane bioreactors for wastewater treatment.Journal of Environmental Engineering - ASCE 128 (11), 1018–1029.

Cho, J., Song, K.G., Yun, H., Ahn, K.H., Kim, J.Y., Chung, T.H., 2005.Quantitative analysis of biological effect on membrane foulingin submerged membrane bioreactor. Water Science andTechnology 51 (6–7), 9–18.

Cicek, N., Franco, J.P., Suidan, M.T., Urbain, V., Manem, J., 1999.Characterization and comparison of a membrane bioreactorand a conventional activated-sludge system in the treatmentof wastewater containing high-molecular-weight compounds.Water Environment Research 71 (1), 64–70.

Fan, F., Zhou, H., Husain, H., 2006. Identification of wastewatersludge characteristics to predict critical flux for membranebioreactor processes. Water Research 40 (2), 205–212.

Frolund, B., Palmgren, R., Keiding, K., Nielsen, P.H., 1996.Extraction of extracellular polymers from activated sludgeusing a cation exchange resin. Water Research 30 (8),1749–1758.

Gerhardt, P., Murray, R.G.E., Wood, W.A., Krieg, N.R., 1994.Methods for General and Molecular Bacteriology. AmericanSociety for Microbiology, Washington, DC.

Gessesse, A., Dueh‘olm, T., Petersen, S.B., Nielsen, P.H., 2003.Lipase and protease extraction from activated sludge. WaterResearch 37 (15), 3652–3657.

Her, N., Amy, G., Foss, D., Cho, J., Yoon, Y., Kosenka, P., 2002a.Optimization of method for detecting and characterizing NOMby HPLC-size exclusion chromatography with UV and on-lineDOC detection. Environmental Science & Technology 36 (5),1069–1076.

Her, N., Amy, G., Foss, D., Cho, J.W., 2002b. Variations ofmolecular weight estimation by HP-size exclusionchromatography with UVA versus online DOC detection.Environmental Science & Technology 36 (15), 3393–3399.

Ho, C.C., Zydney, A.L., 2000. A combined pore blockage and cakefiltration model for protein fouling during microfiltration.Journal of Colloid and Interface Science 232 (2), 389–399.

Holland, F.A., Chapman, F.S., 1966. Liquid Mixing and Processing inStirred Tanks. Reinhold Publishing Corporation, New York, NY.

Howe, K.J., Clark, M.M., 2002. Fouling of microfiltration andultrafiltration membranes by natural waters. EnvironmentalScience & Technology 36 (16), 3571–3576.

Ji, L., Zhou, J., 2006. Influence of aeration on microbial polymersand membrane fouling in submerged membrane bioreactors.Journal of Membrane Science 276 (1–2), 168–177.

w a t e r r e s e a r c h 4 3 ( 2 0 0 9 ) 4 3 0 5 – 4 3 1 5 4315

Kim, J.S., Lee, C.H., Chang, I.S., 2001. Effect of pump shear on theperformance of a crossflow membrane bioreactor. WaterResearch 35 (9), 2137–2144.

Kim, J., DiGiano, F.A., 2006. Defining critical flux in submergedmembranes: influence of length-distributed flux. Journal ofMembrane Science 280 (1–2), 752–761.

Laabs, C.N., Amy, G.L., Jekel, M., 2006. Understanding the size andcharacter of fouling-causing substances from effluent organicmatter (EfOM) in low-pressure membrane filtration.Environmental Science & Technology 40 (14), 4495–4499.

Laspidou, C.S., Rittmann, B.E., 2002a. A unified theory forextracellular polymeric substances, soluble microbialproducts, and active and inert biomass. Water Research 36(11), 2711–2720.

Laspidou, C.S., Rittmann, B.E., 2002b. Non-steady state modelingof extracellular polymeric substances, soluble microbialproducts, and active and inert biomass. Water Research 36 (8),1983–1992.

Le-Clech, P., Chen, V., Fane, T.A.G., 2006. Fouling in membranebioreactors used in wastewater treatment. Journal ofMembrane Science 284 (1–2), 17–53.

Lee, S., Elimelech, M., 2006. Relating organic fouling of reverseosmosis membranes to intermolecular adhesion forces.Environmental Science & Technology 40 (3), 980–987.

Liao, B.Q., Allen, D.G., Droppo, I.G., Leppard, G.G., Liss, S.N., 2001.Surface properties of sludge and their role in bioflocculationand settleability. Water Research 35 (2), 339–350.

Liu, Q.S., Liu, Y., Tay, J.H., Show, K.Y., 2005. Responses of sludgeflocs to shear strength. Process Biochemistry 40 (10),3213–3217.

Liu, Y., Fang, H.H.P., 2003. Influences of extracellular polymericsubstances (EPS) on flocculation, settling, and dewatering ofactivated sludge. Critical Reviews in Environmental Scienceand Technology 33 (3), 237–273.

Liu, Y., Tay, J.H., 2002. The essential role of hydrodynamic shearforce in the formation of biofilm and granular sludge. WaterResearch 36 (7), 1653–1665.

Meng, F.G., Yang, F.L., Shi, B.Q., Zhang, H.M., 2008. A comprehensivestudy on membrane fouling in submerged membranebioreactors operated under different aeration intensities.Separation and Purification Technology 59 (1), 91–100.

Merlo, R.P., Trussell, R.S., Hermanowicz, S.W., Jenkins, D., 2007.A comparison of the physical, chemical, and biologicalproperties of sludges from a complete-mix activated sludgereactor and a submerged membrane bioreactor. WaterEnvironment Research 79 (3), 320–328.

Mikkelsen, L.H., Mascarenhas, T., Nielsen, P.H., 2002. Keyparameters in sludge dewatering: testing for the shearsensitivity and EPS content. Water Science and Technology 46(10), 105–114.

Mulvaney, R.L., 1994. Nitrogen-inorganic forms. SSSA Book SeriesNo.5. Methods of Soil Analysis. Part 3. Chemical Methods.SSSA-ASA, Madison.

Namkung, E., Rittmann, B.E., 1986. Soluble microbial products(Smp) formation kinetics by biofilms. Water Research 20 (6),795–806.

Nielsen, P.H., Jahn, A., Palmgren, R., 1997. Conceptual model forproduction and composition of exopolymers in biofilms.Water Science and Technology 36 (1), 11–19.

Ognier, S., Wisniewski, C., Grasmick, A., 2002. Influence ofmacromolecule adsorption during filtration of a membranebioreactor mixed liquor suspension. Journal of MembraneScience 209 (1), 27–37.

Park, J.S., Yeon, K.M., Lee, C.H., 2005. Hydrodynamics andmicrobial physiology affecting performance of a new MBR,membrane-coupled high-performance compact reactor.Desalination 172 (2), 181–188.

Ramasamy, P., Zhang, X., 2005. Effects of shear stress on thesecretion of extracellular polymeric substances in biofilms.Water Science and Technology 52 (7), 217–223.

Reichert, P., 1994. Aquasim - a tool for simulation and data-analysis of aquatic systems. Water Science and Technology 30(2), 21–30.

Rhine, E.D., Sims, G.K., Mulvaney, R.L., Pratt, E.J., 1998. Improvingthe Berthelot reaction for determining ammonium in soilextracts and water. Soil Science Society of America Journal 62(2), 473–480.

Rosenberger, S., Kraume, M., 2002. Filterability of activated sludgein membrane bioreactors. Desalination 146 (1–3), 373–379.

Rosenberger, S., Laabs, C., Lesjean, B., Gnirss, R., Amy, G., Jekel, M.,Schrotter, J.C., 2006. Impact of colloidal and soluble organicmaterial on membrane performance in membrane bioreactorsfor municipal wastewater treatment. Water Research 40 (4),710–720.

Wilen, B.M., Jin, B., Lant, P., 2003. Relationship betweenflocculation of activated sludge and composition ofextracellular polymeric substances. Water Science andTechnology 47 (12), 95–103.

Wilson, G.J., Pruden, A., Suidan, M.T., Venosa, A.D., 2002.Biodegradation kinetics of MTBE in laboratory batch andcontinuous flow reactors. Journal of EnvironmentalEngineering - ASCE 128 (9), 824–829.

Wisniewski, C., Grasmick, A., 1998. Floc size distribution ina membrane bioreactor and consequences for membranefouling. Colloids and Surfaces A: Physicochemical andEngineering Aspects 138 (2–3), 403–411.

Zein, M.M., Suidan, M.T., Venosa, A.D., 2004. MtBE biodegradationin a gravity flow, high-biomass retaining bioreactor.Environmental Science & Technology 38 (12), 3449–3456.

Zhang, J., Chua, H.C., Zhou, J., Fane, A.G., 2006. Factors affectingthe membrane performance in submerged membranebioreactors. Journal of Membrane Science 284 (1–2), 54–66.

Zhou, Q.H., Cabaniss, S.E., Maurice, P.A., 2000. Considerations inthe use of high-pressure size exclusion chromatography(HPSEC) for determining molecular weights of aquatic humicsubstances. Water Research 34 (14), 3505–3514.