ww.sciencedirect.com

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7

Available online at w

ScienceDirect

journal homepage: www.elsevier .com/locate /watres

Protein and polysaccharide content of tightly andloosely bound extracellular polymeric substancesand the development of a granular activated sludgefloc

Mahendran Basuvaraj a, Jared Fein b, Steven N. Liss a,*

a School of Environmental Studies and Department of Chemical Engineering, Queen's University, Kingston, ON, K7L

3N6, Canadab Wastewater Department, Rothsay, Rothsay, A Division of Darling International Canada Inc., Dundas, ON, L9H5G1,

Canada

a r t i c l e i n f o

Article history:

Received 2 December 2014

Received in revised form

1 May 2015

Accepted 3 May 2015

Available online 12 May 2015

Keywords:

Sludge characterization

Flocs and granules

Settleability

Dewaterability

Extracellular polymeric substances

Polysaccharide

Protein

Rendering plant wastewater treat-

ment system

Abbreviations: CER, cation exchange resin;flow stirred-tank reactor; DAF, dissolved aisuspended solids; F/M, food to microorganispolymeric substances; LS, lab-scale; MLSS,saccharide ratio; SBRs, sequencing batch revolume index; TB-EPS, tightly bound-extrace* Corresponding author. Tel.: þ1 613 533 693E-mail address: steven.liss@queensu.ca (

http://dx.doi.org/10.1016/j.watres.2015.05.0140043-1354/© 2015 Elsevier Ltd. All rights rese

a b s t r a c t

A full-scale (FS) activated sludge system treating wastewater from a meat rendering plant

with a long history of sludge management problems (pin-point flocs; >80% of floc <50 mm

diameter; poor settling) was the focus of a study that entailed characterization of floc

properties. This was coupled with parallel well-controlled lab-scale (LS) sequencing batch

reactors (SBRs) treating the same wastewater and operated continuously over 1.5 years.

Distinct differences in the proportion of proteins and polysaccharides associated with

extracellular polymeric substances (EPS) were observed when comparing the properties of

flocs from the FS and the LB systems. Further differences in the proportion of tightly bound

(TB) and loosely bound (LB) fractions of EPS were also observed for flocs derived from

conditions where differences in settling and dewatering properties of flocs occurred (i.e. FS

and LS systems). FS flocs contained higher levels of EPS along with a higher proportion of

LB than TB EPS, and possessing characteristics associated with non-filamentous bulking

(SVI >150 mL/g). Floc formed in the LS system, following inoculation from sludge taken

from the FS system, was markedly larger in size (>70% of floc >300 mm diameter), spherical

in shape, compact and firm, and appeared to be granular in form. Flocs formed in the LS

system, when an anoxic phase was introduced into the react stage of the SBR cycle, were

found to be more hydrophobic and contained more TB and less loosely bound (LB) EPS

when compared to the FS floc. TB-EPS contained a greater amount of protein, whereas the

polysaccharide content of LB-EPS was larger. Protein was predominantly localized in the

core region of granular flocs where cells were compactly packed. When assessing the

operating conditions of the FS and LS systems parameters that appear to impact the floc

CLSM, confocal laser scanning microscopy; CST, capillary suction time; CSTR, continuous-r flotation; DO, dissolved oxygen; EPS, extracellular polymeric substances; ESS, effluentm ratio; FS, full-scale; HRT, hydraulic retention time; LB-EPS, loosely bound-extracellularmixed liquor suspended solids; PN, protein; PS, polysaccharide; PN/PS, protein to poly-actors; sCOD, soluble chemical oxygen demand; SRT, sludge retention time; SVI, sludgellular polymeric substances; WWTP, wastewater treatment plant.3; fax: þ1 613 533 6934.S.N. Liss).

rved.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7 105

properties and the transition to a granular form include dissolved oxygen (DO) concen-

tration and food to microorganism (F/M) ratio.

© 2015 Elsevier Ltd. All rights reserved.

1. Introduction

The link between the physicochemical and structural prop-

erties of microbial flocs and functionality in engineered sys-

tems has been a focus of numerous studies (Li and Yuan, 2002;

Liao et al., 2006; Wil�en et al., 2008; Van Dierdonck et al., 2013).

Extracellular polymeric substances (EPS) represent an impor-

tant constituent of the floc matrix contributing to the surface

properties of floc, floc strength, formation of floc, as well as

representing the major organic fraction in activated sludge

(Mikkelsen and Keiding, 2002; Sponza, 2003;Wil�en et al., 2003).

EPS has been represented as an internal gel-like network but

can also be characterized with respect to the association with

the cell surface as either tightly bound (TB) and capsular

(Seviour et al., 2012) or extending outward from the cell sur-

face towards the surrounding environment as a loosely bound

(LB) and amorphous structure (McSwain et al., 2005). Infor-

mation on these two types of EPS and its constituents, and

their role on bioflocculation, surface properties and sludge

quality is more limited (Ahmed et al., 2007; Li and Yang, 2007;

Yang and Li, 2009; Ye et al., 2011).

Previous studies indicate the role of EPS to be complex and

the particular relationship between the composition, or

changes in composition of EPS, and floc morphology and

structure, or behaviour, and other structures including gran-

ules is unclear. There have been differences of opinion on the

relationship between EPS content and bioflocculation (Chao

and Keinath, 1979; Jin et al., 2003; Li et al., 2006; Liao et al.,

2001). Other studies have suggested that EPS composition

(Goodwin and Forster, 1985) and properties (Liao et al., 2001),

rather than the quantity, are more important in sludge floc-

culation. More specifically, extracellular proteins have been

associated with improvements in flocculation properties

(Flemming and Wingender, 2001). Information about the role

of EPS, and associated constituents, on granule formation, and

its solid liquid separation efficiency is very limited (de Kreuk

et al., 2007). The gel-like properties of granular sludge have

been attributed to exopolysaccharides (Lin et al., 2010; Seviour

et al., 2012), to the extent that the morphological features of

granules resemble biofilmsmore so than flocs. For flocs a high

or excessive polysaccharide content is viewed as being detri-

mental to settling and dewatering properties owing to a higher

water content associated with polysaccharide rich EPS

(Flemming, 1996); Thompson and Forster, 2003).

Proteins and polysaccharides, are major constituents

which can account for 75e90% of the floc EPS mass (Tsuneda

et al., 2003), with lesser amounts of humic substances, uronic

acids, and nucleic acidsmaking up the rest of thematerial (Liu

and Fang, 2002; Sponza, 2002). EPS can be attached to the cell

surface as peripheral capsules that is tightly bound (TB- EPS),

or shed into the surrounding environment as a less organized

(amorphous) slime, or loosely bound EPS (LB-EPS) (Comte

et al., 2006). Common methods for EPS extraction usually

consist of a thorough washing that strips away LB-EPS. This is

typically followed by harsh extraction (Comte et al., 2006),

resulting in a determination of EPS content and composition

that reflects the TB-EPS content.

The present study entailed analyses of floc characteristics

of a full-scale (FS) activated sludge system treating waste-

water from a large meat rendering plant (Rothsay, Dundas,

Ontario, Canada) experiencing varying degrees of settling and

dewatering challenges. This was coupled with parallel lab-

scale (LS) sequencing batch reactors (SBRs) treating the same

wastewater and operated continuously over 1.5 years Flocs

were generally pin flocs (<50 mmdia), and settled poorly. These

structures were found to contain elevated levels of EPS, which

were associated with non-filamentous bulking, and often

causing the deterioration of sludge handling. Laboratory scale

SBRs were operated on site in parallel to the FS system to

investigate operational conditions, in well-controlled experi-

ments, and to assist the plant in addressing its sludge man-

agement challenges. This paper reports on the content and

composition of TB-EPS and LB-EPS fractions of flocs derived

from different sludge samples over a range of settling and

dewatering properties observed. Correlative microscopy and

physicochemical analyses of flocs were carried out and

revealed differences between TB- and LB-EPS reflected in their

relative PN-PS ratios. A transition of floc and to granular forms

in the sludge of the laboratory scale reactors was observed,

and this paper reports on reactor conditions and the changes

in the properties of these structures with respect to EPS.

2. Materials and methods

2.1. Operational condition of full scale and lab scalereactors

2.1.1. Full scale systemThe operational conditions of a full scale (FS) activated sludge

treatment system receiving wastewater from a meat

rendering plant is presented in Table 1. The rendering plant

processes mostly beef byproducts. The primary treatment

consists of a two-step solids removal process, mechanical

screening of course solids followed by finer screenings; as well

as flow equalization and subsequent removal of floating solids

(i.e. fats, oils, and grease) by dissolved air flotation (DAF) (Kerri

et al., 2003). Following primary treatment, the remaining sol-

uble and colloidal organic constituents and nutrients in the

primary effluent are removed by secondary treatment which

consists of conventional activated sludge treatment (Kerri

et al., 2003). Aeration was by turbulent air flow provided by

floating mechanical aerators. The primary and secondary

solids separated from the wastewater are recycled back to the

Table 1 e Operating parameters of laboratory scale andfull scale treatment system.

Parameter Lab scale system Full scale system

Reactor type Sequencing batchc Continuousc

Flow ratea 0.6e1.4 553 ± 53

Dissolved oxygen

(aerobic) (mg/L)

2.0e3.0 2.7e4.5

Dissolved oxygen

(anoxic) (mg/L)

0.3e0.4 0e0.5

pH 6.5e7.4 6.5e7.4

Temperature (�C) 22 ± 2 6.8e28.7

Hydraulic retention

time (days)

2e3 3.5

Organic loading rate

(kg sCOD/m3$d)

0.39 ± 0.1 0.49 ± 0.21

Mean cell residence

time (days)

10.6 ± 3.1 9.4 ± 2.4

Food to microorganism

(F/M) ratiob0.05e0.3 <0.15

Data were acquired at the active operation condition of the reactors

over one year period.a Flow rate was expressed as ml/min for the LS-SBRs (intermittent

feed) and as gallon/min for the FS system (continuous feed).b F/M was calculated based on soluble chemical oxygen demand

(sCOD).c The FS system was not fed over the weekends whereas the LS

reactors were operated continuously.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7106

production plant for conversion to renderedmaterials. WWTP

plant is operated 24 h/day over 5 days in a week, with no

production or wastewater feeding to their WWTP on

weekends.

2.1.2. Lab scale sequencing batch reactors (LS-SBRs)The operational conditions of lab scale (LS) sequencing batch

reactors SBRs were well controlled and operated in parallel to

the FS system. Four SBRs with an effective working volume of

2 L each were operated over 1.5 years. These SBRs are similar

to those described in previous studies showing the effect of

feed and operating parameters on floc structure and phys-

icalechemical properties of sludge (Liao et al., 2001). The

SBRs were fed with fresh primary DAF effluent from the FS

system which was collected daily from 24 h refrigerated

composite samples. The effluent was kept at 4 �C until fed

into the SBRs. Operating parameters of the SBRs are provided

in Table 2 shows the composition of the primary effluent

Table 2 e Characteristic feature of feed and the performance o

Parameter (mg/L) aPrimary effluent Sec

Lab scal

TSS 843 ± 425 155 ± 53

sCOD 1666 ± 726 63 ± 17

cBOD 1893 ± 914 e

TeNH4eN 151 ± 59 1.3 ± 0.9

TP 21 ± 8 11.9 ± 4.5

TSS-total suspended solids; sCOD - soluble Chemical oxygen demand

ammonia nitrogen; TP- Total phosphorus; NA- Not applicable.a Composition of the feed of laboratory scale reactors as similar to the ful

the reactors over one year period.

(feed) to both full and lab scale reactors over the study period.

During the feed and aerobic reaction phases, mixing was

accomplished through the aeration system by passing air

through fine-bubble diffusers. This provided gentle mixing

and avoided disruption of floc. The mixing intensity in each

SBR was similarly maintained by setting the air pump at the

bottom of each SBR to the same speed just providing suffi-

cient mixing to suspend sludge floc and maintain the dis-

solved oxygen concentration in each SBR at 2.0e3.0 mg/L.

The reactors were seeded with 250 mL of mixed liquor sus-

pended solids (MLSS, 2400 mg/L) from one of the aeration

basins of the FS system. The SBRs were operated at 22 ± 2 �C.Performance of the SBRs was determined by monitoring the

MLSS, effluent suspended solids (ESS), soluble chemical ox-

ygen demand (sCOD), and total phosphorus content (Table 2).

The reactors were monitored over an extended period char-

acterized by three phases as indicated by changes in sludge

quality, numbers of cycles per day and introduction of an

anoxic phase during the react cycle following a 120 days of

operation (Table 3). The operating cycle consisted for four

stages: 15 min fill with mixing; 300 min reaction; 30 min

settling; and 15 min draw.

2.2. Physical properties of flocs

2.2.1. SamplesMixed liquor samples were collected from both LS and FS re-

actors where differences in activated sludge settling and

dewatering properties were observed, and their physical and

chemical properties were analysed as detailed below.

2.2.2. Floc analysisPhysical properties of mixed liquor flocs were determined by

measuring the floc size, surface charge and relative hydro-

phobicity. Floc size was measured by phase contrast micro-

scopy (Nikon, Eclipse E400) and a particle size analyzer

(Mastersizer 2000, Malvern, UK) permitting the measurement

of particles in the range of 0.1e1000 mm.Mixed liquor samples

were diluted in distilled water to avoid multiple scattering.

The suspension was then continuously recycled through the

sample cell with a peristaltic pump and exposed to a 2 mW

HeeNe laser (wavelength 633 nm). The scattered light was

detected by means of a detector that converts the signal to a

size based on volume (Fan and Zhou, 2007). The average floc

f laboratory scale and full scale treatment system.

ondary effluent % Removal

e Full scale Lab scale Full scale

33 ± 27 NA NA

72 ± 12 86 ± 11 92 ± 6

9 ± 6 e 99 ± 0.3

6.5 ± 10.6 85.4 ± 9.5 94.8 ± 8.8

8.9 ± 8.8 72.7 ±0 .4 63.5 ± 28.4

; cBOD-carbonaceous biological oxygen demand; TeNH4eN -Total

l scale system; Data were acquired at the active operation condition of

Table

3e

Chara

cteristicsofopera

tionalco

nditionsoflabsc

ale

andfu

llsc

ale

treatm

entsy

stem

andsludgequality.

Reactors

Operational

phase

Operation

tim

e(day)

Num

ber

ofreaction

(cycle/day)

Feedflow

(L/cycle)

dAeration

mode

MLSSrange

(mg/L)

SVI(m

L/g)

F/M

ratio

Sludgequality

Sludge

Flocsize

Filam

entous

bacteria

aLS(SBRs)

Phase

-11e60

21.5

Aero

bic

900e1450

440e600

0.5

Filamentous

bulkingsludge

80%

<50mm;W

eek,

openfloc

M.parvicella

andtype1863)

Phase

-261e120

40.75

Aero

bic

1950e2250

120e170

0.11e0.37

Nonbulking

sludge

>80%

of50e150mm;

Compactly

pack

ed

floc(Fig.1b)

Very

few

numbers

of

Noca

rdia

spp

Phase

-3>120

40.5

Aero

bic

3h)/anoxic

(1h)/aero

bic

(2.5

h)

2300e2450

45e70

0.05e0.3

Nonbulking

granular

sludge

Granulatedfloc

(70%

of>300mm;

compactly

pack

ed

floc(Fig.1c)

Very

rare

bFSsy

stem

N/A

Continuous

Continuous

c550

Aero

bic

(2.5

h)/

anoxic(1

h)

1900e3500

260e400

<0.15

Visco

us

bulkingsludge

>80%

<50mm;

Week,openfloc

(Fig.1a)

Very

rare

aLabsc

ale

sequencingbatchbioreactor.

bFullsc

ale

continuoustreatm

entsy

stem.

cgallon/m

.d

Disso

lvedoxygen¼

2e3mg/lduringaero

bic

phase

;N/A

,notapplica

ble.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7 107

size was determined as the mean based on the volume

equivalent diameter.

2.2.2.1. Surface charge. Floc surface charge was measured by

a zeta-potential analyser (Nano-Z, Malvern, UK). Diluted

samples for floc particle size analysis were used without any

modification for surface charge analysis. Each sample was

transferred to a thoroughly cleaned measurement cell which

was capped tightly to avoid air bubbles inside the cell. The

analysiswas performed immediately using deionizedwater as

reference.

2.2.2.2. Hydrophobicity. For the relative hydrophobicity anal-

ysis of floc, each mixed liquor sample was washed twice by

centrifugation and re-suspended in 50 mL of distilled water.

The mixture was homogenized by sonication at 50 W for 30 s

to disrupt the flocs aggregates. Distilled water was then added

as required to obtain an optical density of 1.5 ± 0.2 at a

wavelength of 400 nm. The initial absorbance (I0) of the sus-

pension was measured at 400 nm (Spectronic 20Dþ, Ther-

moelectron Corporation), then 1mL of hexadecane was added

to a 10 mL of the suspension and vortexed for 2 min. The

suspension was then allowed to settle in a separating funnel

for 10 min and the absorbance of the aqueous phase was

measured. For the relative hydrophobicity (%), the difference

between the initial and final absorbance was normalized over

the initial value.

2.2.3. Extracellular polymeric substances (EPS)A two-step extraction method was used to extract loosely

bound (LB) and tightly bound (TB) fractions of EPS. Themethod

was a modification of the cation exchange resin (CER) EPS

extraction procedure based on previously published work

(Mahendran et al., 2012; Comte et al., 2006; Frølund et al.,

1996). The readily extractable EPS fraction was termed as

loosely bound EPS (LB-EPS), and condensed EPS fraction as

tightly bound EPS (TB-EPS).

2.2.3.1. LB-EPS fraction. Each mixed liquor sample (50 mL)

was first centrifuged (Eppendorf, 5810R) in a 50-mL tube at

4000 g for 5 min. The liquor supernatant was removed and the

pellet in the tube was then resuspended with 15 mL of EPS

extraction buffer, pH 7.2 (2 mM Na3PO4, 4 mM NaH2PO4, 9 mM

NaCl and 1 mM KCl). The mixture was then diluted with the

extraction buffer solution to its original volume of 50 mL. The

suspension was then mixed using a vortex mixer (Analog

vortex mixer, VWR) for 1 min, followed by centrifugation at

4000 g for 10 min. After centrifugation, the supernatant con-

taining the extracted LB-EPS fraction was collected.

2.2.3.2. TB-EPS fraction. The pellet remaining following re-

covery of LB-EPS was resuspended with extraction buffer so-

lution to the original volume of 50 mL. An aliquot of 1 mL

suspension was transferred to Eppendorf tubes, and then

200 mg of pre-washed cation exchange resin (CER) (DowexR,

Naþ form, Sigma Aldrich) was added. The tubes were then

placed on vortex adaptor (VA08G1-24, Mo Bio Labs) and vor-

texed gently for 60 min at 4 �C (Analog vortex mixer, VWR).

The sample was centrifuged at 9000 g for 10 min to pelletize

the CER/biomass. Supernatant was collected and centrifuged

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7108

again at 9000 g for 10 min to remove any suspended biomass

leftover in the supernatant. The supernatant which contained

the extracted TB-EPS fraction was collected.

2.2.3.3. EPS quantification. Both extracted LB-EPS and TB-EPS

fractions were analysed for EPS components. In this study, EPS

components measured were protein (PN) and polysaccharide

(PS)whichaccounted for 75e90%of the EPS confirming that they

were the major EPS components as reported by Tsuneda et al.

(2003). PN and PS were quantified in a micro-plate assay

format. PS measurement was determined using the phenol-

sulphuric acid method as per Masuko et al. (2005). PN was

measured as described by Fryer et al. (1985). The EPS content

corresponded to thesumof thePNandPSfractions, and total EPS

corresponded to the sum of the LB-EPS and TB-EPS fractions.

2.3. Settleability and dewaterability of activated sludge

Settleability of activated sludge was evaluated by measuring

sludge volume index (SVI) (Bye and Dold, 1998; Liao et al.,

2006). Capillary suction time (CST) (instrument and manu-

facturer instructions: Model 294-50, OFITE, OFI Testing

Equipment Inc., Houston, Texas, USA) was measured to

evaluate the dewaterability of sludge. SVI and CST are widely

accepted approaches to the measure of settleability and

dewaterability, respectively (Lee and Liu, 2001; Buzatu et al.,

2012; Gabarr�on et al., 2013). Measurements of SVI and CST of

sludges from both LS and FS systems were based on a MLSS

concentration of 2250 ± 250 (mg/L).

2.4. Microscopy

2.4.1. Phase contrast microscopeThe morphology and size distribution of the activated sludge

floc were determined by phase contrast microscopy. Wet

mount preparations were used for the analysis. The

morphology of floc, the filamentous index, andhigher life forms

(protozoa, amoeba, ciliates, rotifers and nematodes) were

monitored according to a classification scheme (Jenkins et al.,

1993). Reverse-staining of mixed liquor floc with India ink was

applied to visualize the exopolymeric matrix in intact floc.

2.4.2. Confocal laser scanning microscopy2.4.2.1. Sample preparation. Floc were collected by filtering

5 mL of MLSS sample through a nitrocellulose filter paper

(0.45 mm) (Millipore, Billerica, MA) and were placed onto a

semi-solid state agar surface by inverting the filter paper onto

the agar surface ensuring contact between the filtered floc and

the agarose. The agarose solution (0.75%; low melting point

agarose, electrophoresis grade (SigmaeAldrich, Oakville, ON)

(0.1-mL) was pipetted and spread evenly onto a microscopic

slide well (Fluoro slides with two polished spherical de-

pressions per slide with 18 mm diameters, 0.8 mm de-

pressions) at 35 �C in a water bath (Erie Scientific, Asheville,

NC) to maintain the a semi-solid state. Slides were kept aside

at room temperature (22� ±2 �C) to solidify the floc on the agar

surface. After 10 min, the filter paper was carefully removed

from the agar surface using fine forceps and the slides were

kept at 4 �C until staining to avoid dehydration of the sample.

2.4.2.2. Staining. Acridine Orange (5 mg/L, Invitrogen), a

nucleic acid selective fluorescent dye, which stains each in-

dividual cell, was used to visualize the floc structure. Conca-

navalin A (Con-A) with Alexa Flour 633 conjugate (5 mg/L,

Invitrogen) was used to target PS (a-Mannose, a-Glucose) and

SYPRO orange (5 mg/L, Invitrogen) was used to target PN as

described earlier (Lin et al., 2009; Mahendran et al., 2012). In

brief, the staining probes were dissolved (5 mg/L) in filter-

sterilized (pore size, 0.20 mm) 0.1 M phosphate buffer (pH

7.0). Approximately 0.1 mL of the staining solution was

applied to the floc embedded on slides. The samples were

incubated in the dark at room temperature for 30 min. After

staining, the samples were carefully rinsed twice with 0.1 M

phosphate buffer (pH 7.0) to remove any unbound probes.

2.4.2.3. Microscopy. The stained samples were imaged by

CLSM (Leica DM RE microscope connected to a Leica TCS SP2).

Both 10� and 20� objectives, and a 63� water immersion

objective, were used to examine the samples. Signals were

recorded in the green channel (excitation 488 nm, emission

570 nm) for PN and red channel (excitation 633 nm, emission

647 nm) for PS. Z-stack images were obtained beginning at the

top of the floc with optical slices taken every 0.4 mm through

the floc.

2.4.2.4. Image analysis. The images were processed to mea-

sure the surface area covered by PN and PS using Leica

Confocal Software (LCS, version 2.61). Detailed information

describing image analysis, and the CLSM system used have

been described by Mahendran et al. (2012).

2.5. Standard wastewater parameters

Parameters, including MLSS, ESS, and sludge settled volume

(SSV), were analysed following the standard methods (APHA,

1998). Samples were analysed for soluble chemical oxygen

demand (sCOD), total ammonia-nitrogen (total NH4eN) and

total phosphorus (TP) by the HACH (DR 2800) method.

2.6. Statistical analysis

Linear regression analysis between y and x variables was

performed using the Sigma Plot 17. Multiple regression anal-

ysis was used to determine the correlations between SVI, CST

with other flocs parameters. The linear correlation was

assessed with regression coefficient (R2). R2 reflects statistical

significance between dependent and independent variables.

Differences between dependent and independent variables

were determined by analysis of variance (ANOVA) (p ¼ 0.05).

All the experiments were performed in triplicate.

3. Results and discussion

3.1. Overall reactor performance

The duration of the study including the operation of the LS

SBRs was over a 1.5 year period. DO was controlled in the LS

system (2.0e3.0 mg/L in the aerobic phase; 0.3e0.4 mg/L in the

anoxic phase) as was the food to microorganism (F/M) ratio

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7 109

controlled between 0.05 and 0.3 as shown in Table 1. The FS

system operated at F/M of <0.15. Each of the LS and FS reactors

demonstrated good performance in terms of COD, total

NH4eN and BOD removal (average removal was above 90%),

and about 50% removal of total P (Table 2).

A constant F/M ratio was maintained in the FS system

during the entire experimental phase. The FS system operated

as a continuous-flow stirred-tank reactor (CSTR). For the LS

system, in contrast, the F/M ratio was initially high at the start

of the cycle and lower at the end. The mixed liquor in the FS

reactor was well suspended by applying continuous aeration

from the surface by floatingmechanical aerators and in the LS

reactor by the fine bubble diffusers located at the bottom of

reactors. During the non-filamentous bulking phase, both the

LS and FS system were operated in discontinuous aeration

mode to obtain an aerobic phase followed by an anoxic phase.

Parameters including HRT, SRT and flow rate were similar in

all the reactors. Unlike CSTR systems used in previous studies

(Wil�en and Balm�er, 1999; Papadimitriou et al., 2009), the cyclic

operation of SBRs determines the dynamic nature of the SBR

process (Suresh et al., 2011). No significant difference was

observed on the floc properties at different reaction times in

SBRs, although the F/M ratio changed with time in each cycle

of the SBR operation, similar to a plug flow reactor, conditions

found in many full-scale wastewater treatment operations

(Liao et al., 2006).

The LS system experienced filamentous bulking during the

first 60 days of operations (filamentous index of 5 (Jenkins

et al., 1993)). During this time the operation of the reactors

included a feeding regime consisting of 2cycles/d with

continuous aeration through the full reaction cycle. The SVI

was increased from below 400 mL/g (seed sludge from FS

system) to 440e600 mL/g (Table 3). Any change in operation

that effectively increases the substrate concentration avail-

able to the activated sludge and introduces batch or plug-flow

characteristics to the aeration basin will help control low F/M

bulking (Richard and Daigger, 1993). These include: fed-batch

operation; intermittent feeding; and use of a selector. During

high F/M operation of LS system, themajority of the flocswere

pin flocs and the mixed liquor contained the high abundance

Table 4 e Morphological differences of mixed liquor of lab scal

Properties Lab scal

Floc size mm (% in range) <150 (25%)

Floc shape Round, com

Floc texture Firm

Filament abundance None/few

Filament effect on floc structure Little/none

Polysaccharide coating (India ink staining) Normal

Presence of Zoogloea Rare- finge

Dispersed growth Low/none

Higher life forms:

Amoeba þFlagellates þFree swimming ciliates þStalked ciliates þCrawling ciliates þRotifers þþ present; � absent.

of filamentous bacteria. Microthrix parvicella followed by Type

1863 was mainly observed. M. parvicella and Type 1863 are

commonly thought to be responsible for bulking in high

grease and oil environments (Richard and Daigger, 1993).

Martins et al. (2004) reported a variety of control measures:

installation of a skimmer to remove particulate substrate;

maintenance of a plug-flow regime in all the system; well-

defined aerobic/anoxic/aerobic stages; and the maintenance

of DO of 1.5 (mg/L) in the aerobic phase. Hence, on day 60 of

operation, F/M was further reduced to control the growth of

filamentous bacteria in the LS reactors.

At day 60 (phase 2) the feed flow was decreased and the

number of feeding cycles (i.e. reaction time was reduced) was

increased (from 2 cycles/d to 4 cycles/d). Over the next 60 days,

the sludge settleability improved and the SVI value decreased

from >440 mL/g to <170 mL/g. Filamentous bacteria were

effectively controlled in the reactor. Themode of aerationwas

changed from continuous aeration throughout the react stage

to a discontinuous cycle including an anoxic phase within the

react cycle (aerobic/anoxic/aerobic) during phase �3 of oper-

ation (>120 days). This resulted in a further improvement in

sludge settleability (SVI¼<70 mL/g) (Table 3). The SVI stabi-

lized to below 150 mL/g filamentous bacteria were not detec-

ted, and stable flocs were observed. Further, the floc size

increased and a granular structure appeared.

3.2. Floc properties

3.2.1. Morphological characteristics of flocThe morphological appearance of floc grown under non-

filamentous bulking condition in the LS system was ana-

lysed and compared with those from the FS system (Table 4).

The FS flocswere relatively small (60e85% of the floc consisted

of pin flocs (<50m diameter), irregular, weak with an open floc

structure, and containing low numbers of filamentous bacte-

ria (filament index of 1e2). It appears that the viscous bulking

observed was caused by an excess amount of EPS in the pin

flocs. Higher content of EPS leading to weak and pin flocs and

EPS could contribute to high water retention (Jenkins et al.,

1993). Unlike filamentous bulking, the cause of non-

e and full scale reactors.

e system Full scale system

; >150 (75%) <150 (85%); >150 (15%)

pact Irregular, compact

Weaker

Some/common

Little/none

Excessive/moderate

red type Always abundant e amorphous type

None

þ (mostly testate)

þþþ/� (rare)

þþ/�

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7110

filamentous bulking is less well understood (Peng and Zhu,

2006; Guo et al., 2009).

Due to the poor settleability of MLSS in the LS system, the

reactors experienced loss of biomass and a decline in biosolids

content. To correct for the loss of biomass, 1 L of mixed liquor

was drained from the reactors and it was replaced with fresh

mixed liquor (MLSS ¼ 2300 mg/L) from the FS system. During

phase-2, floc size gradually increased (50% of flocs had a dia.

�150 mm and filamentous bacteria were rarely observed, (fila-

ment index ¼ 1e2). During phase-3, filamentous bacteria were

not observed (Table 3), and the flocs were markedly larger in

size, spherical in shape, and compact and firm (Fig. 1c).

The bulking flocs in the FS system often appeared to bind

or join together to form large flocs which could be easily dis-

associated into small flocs, whereas granular flocs in the LS

system were naturally larger and settled well. Microscopic

examination by India ink reverse staining suggested the

bulking flocs contained a high level of zoogleal bacteria and

elevated levels of EPS (Fig. 1e). The granular floc, on the other

hand, were observed to have little or no zoogleal growth, less

EPS, and tightly packed cells (Fig. 1d). These findings were

further confirmed by fluorescent staining coupled with

confocal laser scanning microscopic (CLSM) analysis of flocs

(Fig. 2aed). The initial inoculum from the FS system contained

largely pin flocs which eventually transformed into a granular

structure in the LS system. Published descriptions of aerobic

granules indicate a wide range of sizes, approximately

0.5e4.0 mm in diameter (Dangcong et al., 1999; Linlin et al.,

2005; de Bruin et al., 2004). Selective pressure created by

means of decreasing sedimentation time and increasing

organic loading rate have been found to enhance the forma-

tion of aerobic granules (Wang et al., 2004).

3.2.2. Physical characteristics of flocsThe surface properties of flocs and their settling and dew-

atering characteristics are presented in Table 4. Samples were

taken during non-filamentous bulking conditions over a 1 year

period. The settling and dewatering properties measured as

sludge volume index (SVI) and capillary suction time (CST),

respectively, for FS sludge (SVI¼>170 mL/g and CST¼>38 s)

and LS sludge (SVI ¼ 70e150 mL/g and CST ¼ 16e26 s) are

shown in Table 4. The low CST and SVI values observed for LS

sludge indicated good sludge quality. These values are

consistent with those sludge of similar quality as reported by

others (Sobeck and Higgins, 2002; Kara et al., 2008).

The granular flocs developed in the LS system were mark-

edly larger in size (the volume basedmedian size of the flocs as

measured by master size analyzer was 150e440 mm). Granular

flocs from LS possessed an increased hydrophobicity (57.5%),

net negative surface charge (�3.24 meq./gVSS), and contained

more TB-EPS and less LB eEPS than the pin flocs from the FS

system (Tables 5 and 6). This is consistent with McSwain et al.

(2005) who reported that the aerobic granules in that study

were predominantly composed of tightly packed cells and EPS.

Floc surface relative hydrophobicity had a strong correlation to

SVI and CST (Table 5) indicating that increased hydrophobicity

generally leads to better settleability and dewaterability of

sludge (Sponza, 2004). Considering that higher cell surface

hydrophobicity promotes cell-to-cell interaction, the resultant

larger flocs are likely to demonstrate better sludge settleability

and dewaterability. This result is consistent with Liao et al.,

2001, but is in contrast to the study reported by Jin et al.

(2004) indicating a correlation between decreased dewater-

ability of sludge with the increasing values of hydrophobicity.

3.3. EPS content and composition of flocs

Extracted EPS content and composition of flocs are presented

in Table 6. The content of tightly bound (TB) and loosely bound

(LB) EPS of flocs derived at different settling conditions were

significantly different. Granular flocs developed in the LS

system possessed more TB-EPS and less LB -EPS (TB/LB-EPS

ratio range of 3.0e3.2) whereas FS pin flocs contained less TB-

EPS and more LB-EPS (TB/LB-EPS ratio range of 0.9). LS gran-

ular flocs contained a higher portion of proteins (PN/PS

ratio ¼ 1.4e1.6) in its total EPS fraction whereas FS flocs con-

tained higher quantities of polysaccharides (PN/PS ratio ¼ 0.5)

(Table 6). Differences were further highlighted when the PN/

PS ratio was calculated on the TB-EPS and LB-EPS fractions.

TB-EPS contains a higher portion of proteins (PN/PS

ratio ¼ 1.6e2.0) whereas LB-EPS contains a higher portion of

polysaccharide (PN/PS ratio ¼ 0.7e0.9).

Quantitative differences of composition of LB and TB -EPS

fractions in different settling flocs were observed. This was

further confirmed by microscopic observations as shown

above in Fig. 2. Protein was predominantly localized in the

core region of granular flocs where cells were compactly

packed (Fig. 2a and c). Pin flocs contained higher quantities of

polysaccharide, and cells were loosely packed (Fig. 2d and b).

Signal intensitywas proportional to the content of EPS and the

PN/PS ratiowas 1.4 for good settling granulated floc and 0.82 in

poor settling pin floc. PN/PS ratios generated by a two-step EPS

extraction method and generated by CLSM image analysis are

comparable (Fig. 3).

Polysaccharides are hydrophilic polymers (Morris, 2007;

Seviour et al., 2009), that reversibly absorb and exude water

or biological fluids (Shen et al., 2006; West et al., 2007), and

contribute to high water retention (Blanco et al., 2004). In-

creases in bound water content of flocs resulted in poor

settling and dewatering. Van Dierdonck et al. (2012) observed

deterioration in floc structure owing to an increase in water

soluble EPS of flocs grown under low organic loading condi-

tions in a laboratory scale activated sludge system. Proteins,

and the amino acid composition of proteins, contribute to the

hydrophobic character of flocs (Dignac et al., 1998; Raszka

et al., 2006). A strong correlation exists between the protein

content in the EPS fraction hydrophobicity, and good settling

sludge (Zhang et al., 2007). Since protein has a high content of

negatively charged amino acids, it is more involved than

sugars in electrostatic bonds with multivalent cations, a key

factor in stabilizing the aggregate structure (Laspidou and

Rittmann, 2002). The functions of protein include the aggre-

gation of bacterial cells, and the formation of an active gel-like

matrix that maintains cell cohesion (Dogsa et al., 2005).

3.4. Factors affecting sludge settleability anddewaterability

Granular flocs settled and dewatered faster than pin flocs, as

shown in Fig. 4. Strong positive correlations between floc size,

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7 111

and SVI (R2 ¼ 0.92) and CST (R2 ¼ 0.96), respectively, were

observed. Filaments were not a factor in the FS system. There

were periods when settling improved, and when there were

periods of poor settleability and dewaterability these

appeared to be attributed to non-filamentous bulking. Ye et al.

(2011) reported that the poor settleability and dewaterability is

caused by small floc formation, but Jin et al. (2004) reported

that sludge with large-sized flocs displayed poorer settle-

ability and dewaterability. It is important to note that very

small and big flocs have been shown to have a high SVI

(Sponza, 2002).

The FS system was continuously operated at F/M of �0.1,

while LS systems were well controlled with respect to the

target F/M (0.05e0.3). Better sludge settleability and dewater-

ability were observed at a higher F/M (�0.15) whereas it was

poorer at F/M � 0.1 (Fig. 5). Further, aerobic and anoxic phases

during the react stage in the LS system were well controlled

within a narrower range whereas it was highly variable in the

FS system (Table 1). The correlation between the sludge set-

tleability and dewaterability of LS system was stronger than

the FS system (Fig. 6). Wu and Huang (2009) found that the

MLSS concentration correlated differently with dewaterability

at different concentration ranges, i.e. when lower than 10 g/L,

MLSSwas significantly correlated; however, when higher than

10 g/L, the MLSS concentration had almost no effect on dew-

aterability (Wu and Huang, 2009). Note that in the present

study anMLSS concentration of 2250 ± 250 (mg/L) was used for

CST and SVI analysis. Studies reported by Alam and Fakhru'l-Razi (2003) and Liao et al. (2006) indicate that the bio-

flocculation and settling properties of activated sludge are

governed by sludge floc characteristics.

3.5. Floc structure and composition and observations ofchanges in settleability and dewaterability

Figs. 7 and 8 illustrate the relationship between the type,

content, and composition of floc EPS and the settleability and

dewaterability of the sludge samples studied. Estimations of

SVI and CST will be influenced by a number of factors and

importantly concentration. Sludge samples where flocs

possessed a high content of TB-EPS exhibited better sludge

settleability and dewaterability. The opposite was true for LB-

EPS dominant flocs (Figs. 7a and 8a). The PN/PS ratio (Figs. 7b

and 8b) and TB-EPS/LB-EPS ratio (Figs. 7c and 8c) also had

strong correlations with SVI and CST. Protein appears to be an

imported key component contributing to the changes in the

structure of flocs that demonstrated improved settleability

and dewaterability. The opposite was true for the FS system

where flocs were found to have a relatively higher proportion

of polysaccharide rich LB-EPS. This is consistent with other

studies describing an excess amount of EPS polysaccharides in

the flocs leading to poor effluent quality and poor sludge

dewaterability (Martinez et al., 2000; Park et al., 2006; Wil�en

Fig. 1 e Comparative morphologies of mixed liquor flocs

from full scale (a) and lab scale system (b, c), and reverse-

staining of flocs with India ink showing appearance of

extracellular matrix in lab scale SBRs flocs (d), and full scale

system (e).

Fig. 2 e CLSM images of the structure of flocs from lab (a,c) and full (b,d) scale system respectively: Cells-top panel; EPS-

bottom panel; PN- protein (green) and PS- polysaccharide (red) in bottom panel.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7112

et al., 2008). EPS production is amicrobial response to external

environmental conditions (Li and Yang, 2007). The results of

this study and other relatively recent reports supports an

emphasis on closer examination of the role of TB and LB-EPS

role in bioflocculation (Ye et al., 2011), and the changes in

composition of the EPS on sludge properties (Yang and Li,

2009).

Table 5 e Physical and functional characteristics of flocs.

Parameter Units Lab scalesystem

Full scalesystem

GSS GSGS PSS

Floc size (Diameter) mm 92.9 150e440 36.9

MLSS mg/L 2850 2750 3013

Hydrophobicity % 35.0 57.5 18.6

Zeta-potential mV �1.39 �3.24 �1.3

Settleability (SVI) mL/g <150 <70 >170Dewaterability (CST) Sec. 26 16 38

GSS- good settling flocs (SVI <150); GSGS- good settling granular

flocs (SVI <70); PSS - poor settling sludge (SVI >170).Values represent means of three determinations; standard de-

viations are lower than 10%.

Confocal imaging combined with extracted EPS analyses

indicates a bi-layer structure in the granular flocs derived in

this study: i) an extraction resistant condensed core layer, and

ii) dispersed periphery layer consisting of tightly bound EPS

surrounded by a loosely bound layer. Protein was predomi-

nantly localized in the core where cells were densely packed.

EPS from this layer was not readily available to extract. The

region at the periphery of the core was likely the layer

extracted as TB-EPS given that this component of the EPS

contained protein and polysaccharides in equal amounts or

where protein content was slightly higher. However, unlike

the granular flocs, pin flocs did not have a compact core,

possessed a mono-layer structure where higher quantities of

polysaccharide were observed, and cells were loosely packed

throughout the entire floc structure. Protein was predomi-

nantly localized in the core region of granular flocs where cells

were compactly packed. This reduces the number of sus-

pended cells and small particles present in the sludge, a factor

that has been shown previously to make sludge easier to

dewater (Karr and Keinath, 1978). The increase in EPS protein

appears to have an important impact on floc structure. This

likely reduces the electrostatic repulsion between cells, owing

to increased hydrophobicity and net-negative surface charge

Table 6 e Content and composition of extracted sludgeEPS at different settling condition.

Content EPS composition(mg/g MLSS)

Mixed liquor floc

Lab scale system Full scalesystem

GSS GSGS PSS

LB-EPS PN 35.3 45.0 60.9

PS 48.3 50.2 135.8

Total 83.6 95.2 196.7

PN/PS 0.7 0.9 0.4

%LB-EPS 25.0 23.9 51.4

TB-EPS PN 168.5 185.3 73.5

PS 82.6 118.0 112.1

Total 251.0 303.3 185.7

PN/PS 2.0 1.6 0.7

%TB-EPS 75.0 76.1 48.6

Total-EPS PN 203.8 230.3 134.4

PS 130.9 168.2 248.0

Total 334.6 398.5 382.4

PN/PS 1.6 1.4 0.5

TBEPS/LBEPS 3.0 3.2 0.9

Food to microorganism ratio 0.1 0.2 0.1

SVI (mL/g) 81.0 134.0 199.0

GSS- good settling sludge (SVI <150); GSGS- good settling granular

sludge (SVI <70); PSS - poor settling sludge (SVI >170).EPS-extracellular polymeric substances; LB-loosely bound; TB-

tightly bound; PN-protein; PS-polysaccharides.

Values represent means of three determinations; standard de-

viations are lower than 5%.

Fig. 3 e Comparison of protein to polysaccharide content of

flocs derived at different settling condition from full scale

and lab scale system through EPS extraction and image

analysis methods.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7 113

of the cell surface, resulting in a more tightly compact struc-

ture and the development of granular flocs.

Fig. 4 e Relationship between floc size, and the settleability

and dewaterability of a) granular flocs derived from lab

scale system, and b) pin flocs derived from full scale

system.

4. Conclusions

This study provides an improved understanding of the

physico-chemical properties of flocs, in particular, the protein

and polysaccharide content in tightly bound (TB) and loosely

bound (LB) fractions of EPS and their influence on the floccu-

lation, settleability, and dewaterability of activated sludge.

Establishing parallel laboratory scale (LS) bioreactors

permitted the opportunity to address the settling and dew-

atering problems of a full scale (FS) activated sludge treatment

plant receiving wastewater from a meat rendering facility

having a long history of poor settling sludge and poor sludge

dewaterability. FS flocs were generally pin flocs, and settled

poorly. These structures were found to contain elevated levels

of EPS, which was associated with non-filamentous bulking,

often causing the deterioration of sludge handling. The

biomass developed over time in the LS system was markedly

larger in size, spherical in shape, and compact and firm. The

results of the study indicated the initial inoculum composed

of flocs had transformed into a granular structure. Specific

conclusions from this study are as follows:

1) The LS systemwaswell controlled within a narrower range

of aerobic (DO ¼ 2.0e3.0 mg/L) and anoxic phases

(DO ¼ 0.3e0.4 mg/L), and F/M (0.05e0.3). Better sludge

Fig. 5 e Relationship between F/M and sludge settleability.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7114

settleability and dewaterability were observed at higher F/

M (�0.15) whereas it was poorer at lower F/M (�0.1). The LS

system developed granular flocs. Initial transformation in

floc structure was observed following the initial 60 day

period of operating the LS system with granular structures

evident beyond 120 days when the aeration cycle was

Fig. 6 e Relationship between SVI and CST: a) Lab scale

system b) Full scale system.

Fig. 7 e Relationship between EPS content and

composition, and settleability of activated sludge: a), SVI

versus LB-EPS and TB-EPS content; b), SVI versus PN/PS

ration of LB-EPS, TB-EPS and Total EPS content; c), SVI

versus TB-EPS/LB-EPS.

changed. DO and F/M were identified as important pa-

rameters controlling the development of granular flocs.

2) Granular flocs observed in the LS system possessed a net

negative surface charge and increased hydrophobicity, and

contained more TB-EPS and less LB -EPS. The opposite was

observed in FS flocs.

Fig. 8 e Relationship between EPS content and

composition, and dewaterability of activated sludge: a),

CST versus LB-EPS and TB-EPS content; b), CST versus PN/

PS ration of LB-EPS, TB-EPS and Total EPS content; c), CST

versus TB-EPS/LB-EPS.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7 115

3) TB-EPS contained a larger amount of protein, whereas LB-

EPS possessed a larger amount of polysaccharide. Protein

was predominantly localized in the core region of granular

flocs where cells were compactly packed. Pin flocs

contained higher quantities of polysaccharide, and cells

were loosely packed.

4) Protein rich TB-EPS exhibited granular structures that

demonstrated better sludge settleability and dewater-

ability. The opposite was true for pin flocs containing a

higher proportion of polysaccharide rich LB-EPS.

Acknowledgements

Financial support was provided by the Mathematics of Infor-

mation Technology and Complex Systems (MITACS) and

Rothsay, A Division of Darling International Canada Inc,

through an Accelerate Program grant (M. Basuvaraj). We are

grateful to the Natural Sciences and Engineering Research

Council of Canada (NSERC) for funding that helped support

this study. The authors wish to acknowledge Rothsay, Dun-

das, ON, Canada for providing in-house laboratory facilities

and the operators of the wastewater treatment plant, for their

assistance on the reactor maintenance and sampling.

r e f e r e n c e s

Ahmed, Z., Cho, J., Lim, B.R., Song, K.G., Ahn, K.H., 2007. Effects ofsludge retention time on membrane fouling and microbialcommunity in a membrane bioreactor. J. Membr. Sci. 287,211e218.

Alam, M.Z., Fakhru'l-Razi, A., 2003. Enhanced settleability anddewaterability of fungal treated domestic wastewater sludge byliquid state bioconversion process.Water Res. 37 (5), 1118e1124.

APHA., 1998. Standard Methods for the Examination of Watersand Wastewaters. American Public Health Association,Washington DC.

Blanco, M.D., Olmo, R., Teijon, J.M., 2004. Hydrogels. In:Swarbrick, James (Ed.), Technology, E.o.P, second ed. MarcelDekker, New York, pp. 235e259.

Bye, C.M., Dold, P.L., 1998. Sludge volume index settleabilitymeasures: effect of solids characteristics and test parameters.Water Environ. Res. 70 (1), 87e93.

Buzatu, P., Zsirai, T., Aerts, P., Judd, S.J., 2012. Permeability andclogging in an immersed hollow fibre membrane bioreactor. J.Membr. Sci. 421, 342e348.

Chao, A.C., Keinath, T.M., 1979. Influence of process loadingintensity on sludge clarification and thickeningcharacteristics. Water Res. 13, 1213e1223.

Comte, S., Guibaud, G., Baudu, M., 2006. Relations betweenextraction protocols for activated sludge extracellularpolymeric substances (EPS) and EPS complexation properties:part I. Comparison of the efficiency of eight EPS extractionmethods. Enzyme Microb. Technol. 38 (1e2), 237e245.

Dangcong, P., Bernet, N., Delgenes, J.P., Moletta, R., 1999. Aerobicgranular sludge-a case report. Water Res. 33 (3), 890e893.

de Bruin, L.M., de Kreuk, M.K., van der Roest, H.F., Uijterlinde, C.,van Loosdrecht, M.C., 2004. Aerobic granular sludgetechnology: an alternative to activated sludge? Water Sci.Technol. 49 (11e12), 1e7.

de Kreuk, M.K., Kishida, N., van Loosdrecht, M.C., 2007. Aerobicgranular sludge-state of the art. Water Sci. Technol. 55 (8e9),75e81.

Dignac, M.F., Urbain, V., Rybacki, D., Bruchet, A., Snidaro, D.,Scribe, P., 1998. Chemical description of extracellular

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7116

polymers: implication on activated sludge floc structure.Water Sci. Technol. 38, 45e53.

Dogsa, I., Kriechbaum, M., Stopar, D., Laggnerz, P., 2005. Structureof bacterial extracellular polymeric substances at different pHvalues as determined by SAXS. Biophys. J. 83, 2711e2720.

Fan, F., Zhou, H., 2007. Interrelated effects of aeration and mixedliquor fractions on membrane fouling for submergedmembrane bioreactor processes in wastewater treatment.Environ. Sci. Technol. 41 (7), 2523e2528.

Flemming, H., Wingender, C., 2001. Relevance of microbialextracellular polymeric substances (EPSs). Part II. Technicalaspects. Water Sci. Technol. 43 (6), 9e16.

Flemming, H.-C., 1996. The forces that keep biofilms together. In:Sand, W. (Ed.), Biodeterioration and Biodegradation, DechemaMonographs, vol. 133. VCH Verlagsgesellschaft, Weinheim,pp. 311e316.

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

Fryer, H.L., Davis, G.E., Manthorpe, M., Varon, S., 1985. Lowryprotein assay using an automatic microtiter platespectrophotometer. Anal. Biochem. 153, 262e266.

Gabarr�on, S., G�omez, M., Monclus, H., Rodrıguez-Roda, I.,Comas, J., 2013. Ragging phenomenon characterisation andimpact in a full-scale MBR. Water Sci. Technol. 67 (4),810e816.

Goodwin, J.A.S., Forster, C.F., 1985. A further examination into thecomposition of activated sludge surfaces in relation to theirsettlement characteristics. Water Res. 19, 527e533.

Guo, J., Peng, Y., Wang, S., Zhang, Y., Huang, H., Wang, Z., 2009.Long-term effect of dissolved oxygen on partial nitrificationperformance and microbial community structure. Bioresour.Technol. 100 (10), 2796e2802.

Jenkins, D., Richard, M., Daigger, G., 1993. Manual on the Causesand Control of Activated Sludge Bulking and Foaming.Ridgeline Press, Lafayette, CA.

Jin, B., Wilen, B.M., Lant, P., 2003. A comprehensive insight intofloc characteristics and their impact on compressibility andsettleability of activated sludge. Chem. Eng. J. 95, 221e234.

Jin, B., Wilen, B.M., Lant, P., 2004. Impact of morphological,physical and chemical properties of sludge flocs ondewatering of activated sludge. Chem. Eng. J. 98, 115e126.

Kara, F., Gurakan, G.C., Sanin, F.D., 2008. Monovalent cations andtheir influenceonactivated sludgefloc chemistry, structure, andphysical characteristics. Biotechnol. Bioeng. 100 (2), 231e239.

Karr, P.R., Keinath, T.M., 1978. Influence of particle size on sludgedewaterability. J. WPCF 50 (8), 1911e1930.

Kerri, K., Dendy, W., Brady, J., Crooks, W., 2003. Operation ofWastewater Treatment Plants, sixth ed., vols. 1 & 2. CaliforniaState University, Sacramento Foundation.

Laspidou, C.S., Rittmann, B.E., 2002. A unified theory forextracellular polymeric substances, soluble microbialproducts, and active and inert biomass. Water Res. 36,2711e2720.

Lee, C.H., Liu, J.C., 2001. Sludgedewaterability and floc structure indual polymer conditioning. Adv. Environ. Res. 5, 129e136.

Li, X.Y., Yuan, Y., 2002. Collision frequencies of microbialaggregates with small particles by differential sedimentation.Environ. Sci. Technol. 36, 387e393.

Li, X.Y., Yang, S.F., 2007. Influence of extracellular polymericsubstances (EPS) on the flocculation, sedimentation anddewaterability of activated sludge. Water Res. 41, 1022e1030.

Li, Z.H., Kuba, T., Kusuda, T., 2006. The influence of starvationphase on the properties and the development of aerobicgranules. Enzyme Microb. Technol. 38, 670e674.

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 Res. 35, 339e350.

Liao, B.Q., Droppo, I.G., Leppard, G.G., Liss, S.N., 2006. Effect ofsolid retention time on structure and characteristics of sludgeflocs in sequencing bactch reactors. Water Res. 35, 339e350.

Lin, H.J., Xie, K., Mahendran, B., Bagley, D.M., Leung, K.T.,Liss, S.N., Liao, B.Q., 2009. Sludge properties and their effectson membrane fouling in submerged anaerobic membranebioreactors (SAnMBR). Water Res. 43 (15), 3827e3837.

Lin, Y., de Kreuk, M., van Loosdrecht, M.C.M., Adin, A., 2010.Characterization of alginate-like exopolysaccharides isolatedfrom aerobic granularsludge in pilot-plant. Water Res. 44,3355e3364.

Linlin, H., Jianlong, W., Xianghua, W., Yi, Q., 2005. The formationand characteristics of aerobic granules in sequencing batchreactor (SBR) by seeding anaerobic granules. Process Biochem..40, 5e11.

Liu, H., Fang, P., 2002. Characterization of electrostatic bindingsites of extracellular polymers by linear programminganalysis of titration data. Biotechnol. Bioeng. 80, 806e811.

Mahendran, B., Lishman, L., Liss, S.N., 2012. Structural,physicochemical and microbial properties of flocs andbiofilms in Integrated Fixed-film Activated Sludge (IFFAS)systems. Water Res. 46 (6), 5085e5101.

Martinez, F., Torres, E.F., Gomez, J., 2000. Oscillations ofexopolymeric composition and sludge volume index innitrifying flocs. Appl. Biochem.. Biotechnol. 87, 177e188.

Martins, A.M.P., Pagilla, K., Heijnen, J.J., van Loosdrecht, M.C.M.,2004. Filamentous bulking sludge-a critical review. Water Res.38, 793e817.

Masuko, T., Minami, A., Iwasaki, N., Majima, T., Nishimura, S.I.,Yuan, C., Lee, Y.C., 2005. Carbohydrate analysis by phenol-sulfuric acid method in microplate format. Anal. Biochem. 339(1), 69e72.

McSwain, B.S., Irvine, R.L., Hausner, M., Wilderer, P.A., 2005.Composition and distribution of extracellular polymericsubstances in aerobic flocs and granular sludge. Appl. Environ.Microbiol. 71 (2), 1051e1057.

Mikkelsen, L.H., Keiding, K., 2002. Physico-chemicalcharacteristics of full scale sewage sludges with implicationsto dewatering. Water Res. 36 (10), 2451e2462.

Morris, V.J., 2007. In: Belton, Gels.P. (Ed.), The Chemical Physics ofFood. Blackwell Publishing Ltd, Oxford, pp. 151e191.

Papadimitriou, C.A., Samaras, P., Sakellaropoulos, G.P., 2009.Comparative study of phenol and cyanide containingwastewater in CSTR and SBR activated sludge reactors.Bioresour. Technol. 100 (1), 31e37.

Park, C., Muller, C.D., Abu-Orf, M.M., Novak, J.T., 2006. The effectof wastewater cations on activated sludge characteristics:effects of aluminum and iron in floc. Water Environ. Res. 78(1), 31e40.

Peng, Y.Z., Zhu, G.B., 2006. Biological nitrogen removal withnitrification and denitrification via nitrite pathway. Appl.Microbiol. Biol. 73, 15e26.

Raszka, A., Chorvatova, M., Wanner, J., 2006. The role andsignificance of extracellular polymers in activated sludge. PartI: literature review. Acta Hydrochim. Hydrobiol. 34 (5),411e426.

Richard, M.G., Daigger, G., 1993. Causes and Control of ActivatedSludge Bulking and Foaming. In: Jenkins, D. (Ed.), second ed.Lewis Publishers, Boca Raton, FL.

Seviour, T., Yuan, Z., van Loosdrecht, M.C.M., Lin, Y., 2012.Aerobic sludge granulation: a tale of two polysaccharides?Water Res. 46 (15), 4803e4813.

Seviour, T., Pijuan, M., Nicholson, T., Keller, J., Yuan, Z., 2009.Understanding the properties of aerobic sludge granules ashydrogels. Biotechnol. Bioeng. 102, 1483e1493.

Shen, W., Zhang, K., Kornfield, J.A., Tirrell, D.A., 2006. Tuning theerosion rate of artificial protein hydrogels through control ofnetwork topology. Nat. Mater. 5, 153e158.

wat e r r e s e a r c h 8 2 ( 2 0 1 5 ) 1 0 4e1 1 7 117

Sobeck, D.C., Higgins, M.J., 2002. Examination of three theories formechanisms of cation-induced bioflocculation. Water Res. 36(3), 527e538.

Sponza, D.T., 2002. Extracellular polymer substances andphysicochemical properties of flocs in steady and unsteady-state activated sludge systems. Process Biochem. 37, 983e998.

Sponza, D.T., 2003. Investigation of extracellular polymersubstances (EPS) and physicochemical properties of differentactivated sludge flocs under steady-state conditions. EnzymeMicrob. Technol. 32, 375e385.

Sponza, D.T., 2004. Properties of four biological flocs as related tosettling. J. Environ. Eng. 130 (11), 1289e1300.

Suresh, S., Tripathi, R.K., Gernal Rana, M.N., 2011. Review ontreatment of industrial wastewater using sequential batchreactor. Int. J. Sci. Technol. Manag. 2 (1), 64e84.

Thompson, G., Forster, C., 2003. Bulking in activated sludge planttreating paper mill wastewaters. Wat. Res. 37 (11), 2636e2644.

Tsuneda, S., Aikawa, H., Hayashi, H., Yuasa, A., Hirata, A., 2003.Extracellular polymeric substances responsible for bacterialadhesion onto solid surface. FEMS Microbiol. Lett. 223,287e292.

Van Dierdonck, J., Van den Broeck, R., Vervoort, E., D’haeninck, P.,Springael, D., Van Impe, J., Smets, I., 2012. Does a change inreactor loading rate affect activated sludge bioflocculation?Process Biochem. 47, 2227e2233.

Van Dierdonck, J., Van den Broeck, R., Vervoort, A., Van Impe, J.,Smets, I., 2013. Microscopic image analysis versus sludgevolume index to monitor activated sludge bioflocculation: acase study. Sep. Sci. Technol. 48, 1433e1441.

Wang, Q., Du, G., Chen, J., 2004. Aerobic granular sludgecultivated under the selective pressure as a driving force.Process Biochem. 39, 557e563.

West, E.R., Xu, M., Woodruff, T.K., Shea, L.D., 2007. Physicalproperties of alginate hydrogels and their effects on in vitrofollicle development. Biomaterials 28, 4439e4448.

Wil�en, B.M., Balm�er, P., 1999. The effect of dissolved oxygenconcentration on the structure, size and size distribution ofactivated sludge flocs. Water Res. 33 (2), 391e400.

Wil�en, B.M., Jin, B., Lant, P., 2003. The influence of key chemicalconstituents in activated sludge on surface and flocculatingproperties. Water Res. 77 (9), 2127e2139.

Wil�en, B.M., Lumleyb, D., Mattssonb, A., Minoc, T., 2008.Relationship between floc composition and flocculation andsettling properties studied at a full scale activated sludgeplant. Water Res. 4 (2), 4404e4418.

Wu, J., Huang, X., 2009. Effect of mixed liquor properties onfouling propensity in membrane bioreactors. J. Membr. Sci.342 (1e2), 88e96.

Yang, S.F., Li, X.Y., 2009. Influences of extracellular polymericsubstances (EPS) on the characteristics of activated sludgeunder non-steady-state conditions. Process Biochem. 44,91e96.

Ye, F., Ye, Y., Li, Y., 2011. Effect of C/N on extracellular polymericsubstances (EPS) and physicochemical properties of activatedsludge floc. J. Hazard. Mater. 188, 37e43.

Zhang, L., Feng, X., Zhu, N., Chen, J., 2007. Role of extracellularprotein in the formation and stability of aerobic granules.Enzyme Microbial. Technol 44 (5), 551e557.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具