Separation and Purification Technology 134 (2014) 1–11

Contents lists available at ScienceDirect

Separation and Purification Technology

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

Dynamic changes in the characteristics and components of activatedsludge and filtrate during the pressurized electro-osmotic dewateringprocess

http://dx.doi.org/10.1016/j.seppur.2014.07.0191383-5866/� 2014 Elsevier B.V. All rights reserved.

⇑ Corresponding author. Tel.: +86 10 62336673; fax: +86 10 62336596.E-mail address: wangyilimail@126.com (Y.-L. Wang).

Jing Feng, Yi-Li Wang ⇑, Xue-Yuan JiCollege of Environmental Science and Engineering, Beijing Key Lab for Source Control Technology of Water Pollution, Beijing Forestry University, Beijing 100083, China

a r t i c l e i n f o

Article history:Received 26 August 2013Received in revised form 28 June 2014Accepted 4 July 2014Available online 18 July 2014

Keywords:Pressurized electro-osmotic dewateringActivated sludgeDynamic changeBound waterInner structure

a b s t r a c t

The performance and process of pressurized electro-osmotic dewatering (PEOD) technology for activatedsludge (AS) were investigated. In PEOD process, a single pressure of 600 kPa only removed a smallamount of free water and bound water in AS, whereas both kinds of water further decreased to0.24 g g�1 dry solid and 0.25 g g�1 dry solid after the application of 50 V voltage at electrical compression(EC) stage. During the PEOD process, the loosely network structure was gradually ruined, and a quantityof narrow and parallel slits generated in dewatered AS at EC stage. The contents of organic matters in fil-trate increased at all PEOD stages and humic acid-like organics formed at EC stage. Using the responsesurface method (RSM), the optimum dewatering conditions were determined as a combination of401 kPa pressure and 50 V voltage, which gave a dry solids content and energy consumption of 41.90%(wt%) and 0.153 kWh per kg removed water.

� 2014 Elsevier B.V. All rights reserved.

1. Introduction

The dewatering and disposal of sewage sludge are essential inmunicipal wastewater treatment plants, accounting for approxi-mately 50% of all costs in wastewater treatment. This high con-sumption is mainly caused by the high content of water in rawsludge. However, the removal of water from raw sewage sludgeto a favorable level for the following transport and disposal pro-cesses is difficult because of the colloidal nature of particles andthe gel-like structure of the sludge [1,2]. The commonly usedchemical and mechanical methods are only efficient for the reduc-tion of free water that has no interaction with sludge particles [2].Indeed, the bound water, which is defined as an immobilized waterthat is bound chemically or physically (or both) onto the flocs or isheld by the solid state either by sorption, is not fully separatedfrom the sludge particles.

Pressurized electro-osmotic dewatering (PEOD) technology isconsidered as one of the most efficient electro-osmotic dewateringmethods for enhancing dewatering efficiency, especially for theseparation of bound water from sludge [3–6]. In PEOD technology,conventional pressure is combined with electric field to improveliquid/solid separation, increase the final dry solids content, and

accelerate the dewatering process [7]. Due to the fixed negativesurface charge at the sludge biosolids interface, an electrical doublelayer (EDL) yields on solid–liquid interface and constitutes an ion-ized mobile region on the liquid side. Under the effect of the elec-tric field, the propagation of ionized mobile region is forced intoflowing and subsequently drives the neutral liquid in the centralchannel via the molecular viscosity, promoting water displace-ment. Therefore, an electro-osmotic flow occurs [8].

The dewatering effect of PEOD is mainly influenced by appliedvoltage (or current) and pressure [4,9,10]. Increase in voltageresults in a dryer sludge and a quicker dewatering speed. Increasein pressure has also been shown capable of improving the perfor-mance of PEOD. Ultimately, the application of higher voltage andpressure further enhances the dewatering efficiency of the PEOD,but increases energy consumption as well. The response surfacemethod (RSM) has been recently used to investigate the maineffects of the factors (voltage and pressure) on the dewatering effi-ciency and to determine the optimum operation conditions [7]. Thecombination of final dry solids content and energy consumption istreated as the response of the RSM, which provides the optimumconditions for such response. Using RSM, Mahmoud et al. [7] deter-mined that the optimum dewatering performance for AS can beachieved at voltage of 40 V and pressure of 728 kPa.

In the PEOD system, both the electro-osmosis and electrophore-sis depend on the surface charge of the sludge particles. The formerdrives the charged fluid of the diffusive double layer around the

Fig. 1. Schematic representation of the laboratory-scale PEOD device.

2 J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11

colloidal particles moving to the cathode, whereas the latter drivesthe colloidal particles transportation to the anode; water is electro-lyzed at the electrodes [11]. Combining these electrokinetic effects,numerous changes would occur not only in the water content inthe sludge, but also in the organic matters and metals. Tuan et al.[12] proposed that an application of electric field could inducethe migration of negatively charged organic matter from the cath-ode to the anode, and thus, the filtrate at the anode containedhigher levels of COD and total organic carbon (TOC), whereas lowerlevels of COD were observed at the cathode. Hwang and Min [9]and Tuan et al. [13] noticed a decrease in the content of heavymetal (Zn, Mn, Pb, Cd, and Ni) in the sludge after electro-dewater-ing. Moreover, the microorganisms in the sludge, such as coliform[14], can be inactivated by the application of electric field, as con-firmed by Huang et al. [15].

Most studies on the PEOD process have focused on its perfor-mance and the corresponding operating conditions. Indeed, theuse of an electric field is the main cause for the enhanced dewater-ing efficiency of sludge by reduction of bound water. However, thecontribution of the PEOD process to the reduction of free water andbound water requires further quantitative investigation. Moreover,little attention has been paid to the changes in the inner structureof the sludge during the PEOD process caused by the decrease inthe water content and the transitional distribution between freewater and bound water. The change in the porosity of sludge hasa significant influence on the dewatering flux of electroosmosis.

The current study investigates the dynamic changes in the char-acteristics and components of activated sludge (AS) matrix duringthe PEOD process under typical operational conditions. The differ-ential scanning calorimetry (DSC) was used to investigate the vari-ations in the distribution of free water and bound water in AS. Thesludge samples after each PEOD stage were paraffin sectioned andthe inner structure of the sludge was identified using micro-imageand fractal methods. The results are expected to provide supportfor investigations on the performance and mechanism of the PEODprocess.

2. Material and methods

2.1. Sludge samples

AS was sampled from a municipal wastewater treatment plantin Beijing, China, which treats 6.0 � 105 tons of wastewater dailyby using an Anaerobic–Anoxic–Oxic (A2/O) process. After sampling,AS was stored at 4 �C for a maximum five days to reduce its bio-chemical change. The main characteristics of AS are given inTable 1. In this study, due to the variability of original AS samplesand the strong heterogeneity among different AS, each PEODexperiment was performed at least in triplicate using three ormore AS samples that belonged to the same batch to ensure thereproduction of result. In addition, the volatile solids (VS), total sol-ids (TS), chemical oxygen demand (COD) and soluble chemical oxy-gen demand (SCOD) of the sludge were measured with thestandard method recommend by [16].

2.2. Experimental setup

Fig. 1 shows a laboratory scale setup for PEOD, which consists ofa cylindrical laboratory filtration/compression cell (diameter of

Table 1Characteristics of the AS samples.

Moisture content (%) TS (g/L) VS/TS (%) pH

99.11 ± 0.04 8.34 ± 0.08 67.97 ± 0.20 7.12 ± 0.02

70 mm and volume of 2 L) made of plexiglass, a DC power supply(DH1719A-4, Dahua Co., China), a beaker, and a precision balance(PL2002-IC, Mettler Toledo, Switzerland). In the filtration/compres-sion cell, a compression piston made of Teflon was designed tocompress the water from the anode side. The bottom of the cellwas covered with a polypropylene filter cloth with a 5 lm aper-ture. A dimensionally stable anode was fixed on the bottom ofthe piston, and a perforated disk cathode was placed under the fil-ter cloth and sieve plate. Titanium-coated mixed metal oxide wasused as the electrode to prevent electrical chemical corrosion.

2.3. Experimental procedure

2.3.1. ConditioningThe conditioning experiments were conducted using a six-pad-

dle stirring apparatus (JTY-6, Tangshan Dachang Chemicals Ltd.,China). The cationic polyacrylamide (CPAM) (WD4960, ShanghaiWeidu Water Treatment Technology Ltd.) was employed as theflocculant, which had an average molecular weight of 20–25 MDaand a charge density of 60%. Prior to conditioning, 2 L of AS wasmaintained in open air to let it reach room temperature. Then,1 L of AS (containing 8.34 ± 0.08 g dry solid) was transferred to astandard beaker and then 0.1% working solution of CPAM wasadded under agitation of 800 rpm. The mixture was subjected torapid mixing for 1 min, followed by gentle agitation at 62 rpmfor 5 min. The conditioned sludge was withdrawn from the beakerfor the CST test. The optimum dose of polyelectrolyte was deter-mined to be 4.00 (kg ton�1 dry sludge), at which the CST valuesof the conditioned sludge reached minimum. Then, conditionedsludge at the optimum dosage was decanted, and both the sedi-ment and the supernatant or filtrate was weighed. The dry solidscontent in the dewatered AS and the water removal percentagewere determined as follows:

Conductivity (mS/cm) CST (s) COD (mg/L) SCOD (mg/L)

1.39 ± 0.01 53.5 ± 3.10 6500 ± 510 1800 ± 290

J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11 3

Dry Solids Content ¼ ms �ms �W0

ms �mw� 100% ð1Þ

Percentage of Water Removed ¼ mw

ms �W0� 100% ð2Þ

where mw is the mass of supernatant or filtrate extracted in eachstage, ms is the total mass of the original sludge, and Wo is the ori-ginal moisture content of the AS.

2.3.2. Drainage under gravity (DG)The conditioned sludge was filtered under gravity in the PEOD

cell without voltage. To start the drainage, the valve of the filtratepipe was opened and the filtrate was collected using a beaker on abalance. Drainage continued until no more than two drops of fil-trate were collected within 10 min (2 h were necessary). The col-lected filtrate was weighed on the balance. The correspondingdry solids content of the dewatered AS and the percentage of waterremoval were determined using Eqs. (1) and (2), respectively.

2.3.3. Compression and electrical compression (EC)Compression and EC the two successive stages were performed

after the AS dewatering procedure of DG, and each stage proceededfor two hours [7]. In the compression stage, constant pressure of200, 400, or 600 kPa was imposed on the piston. In the EC stage,the selected operating voltage was combined with the appliedpressure. The filtrate at the cathode was collected using a beakerand then weighed on a balance every 10 min. The average dry sol-ids content of the dewatered AS and the percentage of waterremoval were determined as before.

2.4. Analytical methods

2.4.1. Methodology of RSM for the AS PEODIn order to limit the energetic cost of the dewatering process

and simultaneously obtain a satisfactory final dry solids content,the central composite design (CCD) of the RSM was used [7]. Thepressure and voltage were considered as two independent vari-ables, X1 and X2, respectively. X1 ranged from 200 kPa to 600 kPa,and X2 from 10 V to 50 V. The combination of final dry solids con-tent and energy consumption was used as the response (Y1 and Y2).The dewatering experimental design is presented in Table 2. A sec-ond-order polynomial model based on the principle of CCD wasused to describe the relationship of the response with the indepen-dent variables. The corresponding equation is as follows:

Y ¼ a0 þ a1X1 þ a2X2 þ a12X1X2 þ a11X21 þ a22X2

2 ð3Þ

where Y is the predicted response, a0, a1, a2, a12, a11, and a12 repre-sent the model regression coefficients, and X1 and X2 are the codedindependent variables. Design Expert 7.0 software was used to fit

Table 2Experimental design matrix and the observed and predicted responses.

No. Pressure (kPa) Voltage (V) Final dry solid

Exp. X1 x1 X2 x2 Y2 (Observed)

1 200 �1.00 50 1.00 20.00032 600 1.00 50 1.00 62.39273 600 1.00 10 �1.00 12.79604 400 0.00 50 1.00 41.63885 400 0.00 30 0.00 15.97946 600 1.00 30 0.00 16.34497 400 0.00 30 0.00 15.08358 400 0.00 30 0.00 15.93879 200 �1.00 30 0.00 13.1924

10 400 0.00 10 �1.00 12.620711 200 �1.00 10 �1.00 11.4144

the experimental data matrix and obtain the optimum pressure,voltage, and predicted final dry solids content and energy consump-tion. The quality of the fit was expressed using the determinationcoefficient R2, and the significance was estimated using the F-valueand p-value.

2.4.2. AS characterizationThe original AS samples were collected from a block of the thor-

oughly blended AS matrix. At each dewatering stage (conditioning,DG, compression and EC), the AS samples were taken from the cen-ter of the corresponding dewatered AS block or cake for the follow-ing characterization.

2.4.2.1. Inner structure of dewatered AS. The inner structure of theAS was obtained from a paraffin section experiment according tothe following procedure [17]. Several sludge blocks taken fromthe dewatered AS were fixed with formalin buffer (4% formalde-hyde solution diluted in phosphate buffer (PB) solution) for 24 hat 4 �C. Prior to dehydration, the blocks were washed with PB(pH value of 7.2) for 3 � 1 h. Then, the blocks were immersed inethanol/water solution (15 min at 50%, 75%, 90%, 95%, and 100%(v/v)) to replace the water in the blocks with ethanol. The ethanolin the blocks was then replaced with xylene by immersing theblocks in xylene/ethanol solution (15 min at 50%, 75%, 90%, and100% (v/v)). The xylene-filled blocks were maintained in severalembedding boxes filled with molten paraffin at 75 �C overnight.Finally, the embedding boxes were cooled to 25 �C with tap waterto solidify the paraffin-embedded cakes. A YD-202A rotary micro-tome (Jinhua YiDi Medical Appliance Co., China) was used to slicethe cakes into 10 lm sections. These sections were then moved tothe surface of a deionized water-covered glass slide that washeated using a low-power hair dryer until all sections were spreadthoroughly. Then, the slides were air dried after removing theexcess water they contain with bibulous paper [17]. Finally, thedried sections were dewaxed with xylene and coverslipped usingPermount TM Mounting Medium for image observation andrecording with Aigo digital microscope (GE-5, China).

To determine the geometric characteristics of the dewatered ASsection, the Image J 1.42q software (National Institutes of Health,USA) was first used to convert the RGB slice images into gray scale,and then Otsu’s method proposed by Chu et al. [17] was used toobtain the corresponding binary images. Then, the geometricparameters of the slice, such as the pore area AreaP (the white frac-tions) and the solid area AreaB (the black fractions), were deter-mined from the binary images by using the ImagePro Plus 5.0software (Media Cybernetics, USA). The porosity of the slice wascalculated to be the ratio of AreaP/(AreaP + AreaB). In addition, thefractal dimensions of the binary images, which represent the frac-tal dimensions of the dewatered AS slices, were estimated usingthe Box-Counting method provided in the Image J 1.42q software.

s content (%) Energy consumption (kWh/kgwater removed)

Y2 (Predicted) Y2 (Observed) Y2 (Predicted)

23.0457 0.1013 0.093059.1935 0.2909 0.2563

9.6200 0.0179 0.011841.7973 0.1097 0.152215.5796 0.1079 0.097022.7231 0.1131 0.153415.5796 0.1055 0.097015.5796 0.1062 0.0970

7.0808 0.0978 0.085412.7292 0.0176 0.003014.4831 0.0188 0.0390

4 J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11

2.4.2.2. Surface morphology of dewatered AS. A scanning electronmicroscope (SEM, Quanta 200, FEI, USA) was used to observe thesurface micro-morphology of the dewatered sludge.

2.4.2.3. Bound water content. Total water content in the sludge wasmeasured following the Standard Methods by drying a sample at105 �C for 24 h [16]. The bound water content was determinedby the DSC method permitting direct thermal analysis for phasechanges of free water [18]. Sample weight for a DSC test was keptat approximately 8 mg. To achieve uniform sample quality, thesludge was first vacuum filtered to remove most of the free waterand the wet filtered cake blended thoroughly for 15 min in a smalltwo-arm kneader. Since the bound water should be hard to removeby mechanical force, the blending action was assumed to have noeffect on the measurements. With the blending treatment, themeasurement deviation for each sample could be controlled towithin 15%. After that, a differential thermal analyzer (DSC204,Netzsch, Germany) was used to record the thermograms accordingto the following procedure: the sludge temperature first decreasedto �50 �C at a rate of 10 �C/min, and then rose to 25 �C at the samerate. Heat absorption was determined by integrating the peak areaunder the endothermic curve. Then, the bound water content wascalculated as follows:

WB ¼WT � DH=DH0; ð4Þ

where WB and WT are the bound water content and the total watercontent in the sludge samples (g g�1), respectively, DH is the DSCenthalpy of the sludge samples, and DH0 is the standard meltingheat of ice (334.7 kJ kg�1).

2.4.2.4. The contents of metal ions. The contents of metal ions insludge were measured by using an Inductive Coupled Plasma(ICP) spectrometer (Leeman Labs, USA).

2.4.3. Characteristics of the filtrateThe filtrate was collected during each dewatering stage under

the typical dewatering conditions. The conductivity of the filtratewas measured using a conductivity meter (EC215, HANNA, Italy),and the zeta potential was recorded using a Zetasizer (Nano Z sys-tem, Malvern Co., UK). Wahlberg et al. [19] observed that the SS ofthe filtrate was equal to the product of the turbidity and the massconversion factor for the turbidity to solids (1.2 mg SS/l/FTU),where the turbidity (FTU) was obtained from the absorbance ofthe centrifuged filtrate (2 min centrifugation at 3000 rpm) at650 nm according to a standard calibration curve. After removingthe particulates in the filtrate with the mixed cellulose ester mem-brane (0.45 lm), the total organic carbon (TOC) in the filtrate wasdetermined using a TOC analyzer (TOC-V CSN, Shimadzu, Japan)and its excitation emission matrix fluorescence spectrum (EEM)was obtained using luminescence spectrometry (F-7000, Hitachi,Japan). The EEM spectra were collected with subsequent scanningemission spectra from 220 nm to 600 nm at 1 nm increments byvarying the excitation wavelength from 200 nm to 400 nm at5 nm increments. Excitation and emission slits were maintainedat 5 nm and the scanning speed was set at 1200 nm/min for themeasurements. COD and SCOD of the filtrate were measured usingportable COD equipment (DR1010, HACH, USA).

3. Results and discussion

3.1. Effect of operational conditions on dewatering efficiency

In the PEOD process, the dry solids content increased from0.9 ± 0.1% (wt%) in the original AS to 6.0 ± 0.1% (wt%) in the condi-tioned AS, and further to 8.7 ± 0.1% (wt%) for the dewatered AS at

the DG stage. The corresponding percentage of water removedafter the conditioning stage and DG stage were 86.4 ± 0.8% (wt%)and 90.6 ± 0.3% (wt%), respectively. Fig. 2 shows the effect of oper-ational conditions (pressure and voltage) on the dewatering kinet-ics during the compression and EC stages. Compared with the DGstage, the dry solids content presented a slight increase duringthe compression stage, and the average dewatering speed changedfrom 0.017%/min to 0.020%/min as the applied pressure increasedfrom 200 kPa to 600 kPa. After the compression stage, both the per-centage of water removed and the dry solids content under 200,400, or 600 kPa reached 93% (wt%) and 10% (wt%), respectively.These results show that compression alone did not reach a suffi-cient dewatering efficiency, which was also obtained by Mahmoudet al. [7] in their research on the electro-dewatering of activatedwastewater sludge.

The dewatering efficiency was enhanced when voltage wascombined with pressure. The dry solids contents ascended withtime at each applied pressure and were accelerated by the com-bined voltage, as shown in Fig. 2a–c. At high levels of voltage of40 V or 50 V, the dry solids contents in the AS dramaticallyimproved. It can be seen that the thickness of the cake significantlyreduced from the beginning of the EC stage (about 55 mm) (seeFig. S1 in Supporting information). However, at the voltage below40 V, the significant electrical resistance induced small electric cur-rent throughout the cake and consequently a slight decrease of thecake thickness. After the EC stage under 50 V voltage, the dry solidscontents reached 20.0% (wt%) at 200 kPa, 41.6% (wt%) at 400 kPa,and 62.40% (wt%) at 600 kPa. A gradual increase in the dewateringratio (water removed by EC stage/water content after the compres-sion stage) was followed by a rapid increment as the combinedvoltage climbed from 10 V to 50 V, as shown in Fig. 2d–f. Theseresults showed that the percentage of water removed by the ECstage can exceed 6% (wt%) at 600 kPa combined with 50 V, andthe corresponding water content in the sludge was approximately37% (wt%). In the study reported by Mahmoud et al. [7], a similarwater content of dewatered sludge was obtained under a pressureand voltage of 1200 kPa and 50 V when the PEOD technology wasused to treat 2 kg AS and the corresponding dry solid contentreached more than 60%. Then a combination of 600 kPa pressureand 50 V voltage was selected as a typical operational conditionin EC stage during the following running of PEOD process.

3.2. Evolution of AS characteristics during different dewatering stages

3.2.1. Content and distribution of moisture and organics in ASBased on the DSC thermograms of AS samples during the PEOD

process (Fig. 3a–d), the bound water content in the AS samples canbe calculated using Eq. (4). Fig. 3e shows the moisture distributionin terms of free water content (F), bound water content (B), andtheir ratio (RB/F) in AS at different dewatering stages. The originalAS contained approximately 11.52 ± 3.32 g bound water and98.60 ± 3.32 g free water per g dry solid. After the conditioningand DG stages, the free water content in AS dramatically decreasedby approximately 92.35% (wt%), and the bound water content wasreduced to 3.23 ± 0.35 g g�1 dry solid. Chu et al. [20] measured thebound water content of activated sludge with a similar method,and the corresponding value was 4.3 g g�1 dry solid. In addition,the corresponding RB/F increased from 0.12 for the original AS to0.43 for the conditioned AS after the DG stage. In previous studies[20,21], they indicated that the chemical conditioning can reducethe bound water content of sludge either by replacing water mol-ecules adsorbed on the particle surface or by affecting the waterbinding capacity to the particle. Therefore, the dosing of CPAM alsoconverts a part of bound water to free water.

Both the bound water content and the free water content in theAS declined slightly with the application of 600 kPa pressure. At

0 20 40 60 80 100 120 140 160 180 200 220 2400

10

20

30

40

50

60

70

80

90

100(a)

EC stageCompression stage

200kPa

10V 20V 30V 40V 50V

Time (min)

Dry

sol

ids

cont

ent (

%)

0 20 40 60 80 100 120 140 160 180 200 220 2400

10

20

30

40

50

60

70

80

90

100

EC stageCompression stage

(b)400kPa

10V 20V 30V 40V 50V

Time (min)

Dry

sol

ids

cont

ent (

%)

0 20 40 60 80 100 120 140 160 180 200 220 2400

10

20

30

40

50

60

70

80

90

100

EC stageCompression stage

(c)600kPa

10V 20V 30V 40V 50V

Time (min)

Dry

sol

ids

cont

ent (

%)

10 20 30 40 500

1

2

3

4

5

6

7

8

9

10(d)

Perc

enta

ge o

f w

ater

(%

)

Voltage (V)

water remained in dewatered sludge water removed by EC stage

10 20 30 40 500

1

2

3

4

5

6

7

8

9

10(e)

Perc

enta

ge o

f w

ater

(%

)

Voltage (V)

water remained in dewatered sludge water removed by EC stage

10 20 30 40 500

1

2

3

4

5

6

7

8

9

10(f)

Perc

enta

ge o

f w

ater

(%

)

Voltage (V)

water remained in dewatered sludge water removed by EC stage

Fig. 2. Variations in the dry solids content and water in AS under different pressures and voltages (a–c): Dry solids content versus time; and (d–f): Percentages of waterremoved by EC or of water that remained in dewatered sludge at the end of the EC stage, + = water remained in sludge after compression.

-60 -50 -40 -30 -20 -10 0 10 20 30-25

-20

-15

-10

-5

0

5

10

15 (a)

-281.6 mJ·mg-1

297.0 mJ·mg-1

Endothermic peak

Exothermic peak

Hea

t flo

w (

mW

·mg-1

)

Temperature (°C)

-60 -50 -40 -30 -20 -10 0 10 20 30

-20

-15

-10

-5

0

5

10 (b)

-211.9 mJ·mg-1

214.5 mJ·mg-1

Endothermic peak

Exothermic peak

Hea

t flo

w (

mW

·mg-1

)

-60 -50 -40 -30 -20 -10 0 10 20-12

-10

-8

-6

-4

-2

0

2

4

6 (c)

-185.2 mJ·mg-1

Exothermic peak

184.7 mJ·mg-1

Endothermic peak

Hea

t flo

w (

mW

·mg

-1)

-60 -50 -40 -30 -20 -10 0 10 20-5

-4

-3

-2

-1

0

1

2

3 (d)

-53.93 mJ·mg-1

56.42 mJ·mg-1

Exothermic peak

Endothermic peak

Hea

t flo

w (

mW

·mg-1

)

Orininal AS DG Compression EC0

10

20

30

40

50

60

70

80

90

100

110

Bound water Freewater Ratio of B/F

Dewatering stages

Moi

stur

e C

onte

nt (

g.g-1

dry

soli

ds)

0.0

0.2

0.4

0.6

0.8

1.0

1.2

Ratio of B

/F

(e)

Original Conditioning DG Compression EC55

56

57

58

59

60

61

62

63

64

65

VS/

TS

(%)

Dewatering stages

(f)

Temperature (°C) Temperature (°C)

Temperature (°C)

Fig. 3. DSC thermogram of AS, moisture distribution, and VS/TS of AS at different PEOD stages: (a) original; (b) DG stage; (c) compression stage; (d) EC stage(600 kPa, 50 V);(e) moisture distribution; and (f) VS/TS.

J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11 5

the end of the compression stage, the free water content andbound water content reached 3.78 ± 0.20 g g�1 dry solid and2.07 ± 0.20 g g�1 dry solid, respectively, corresponding to anincrease of 0.12 in the RB/F value at the DG stage.

Thereafter, when voltage was combined with pressure duringthe EC stage, the bound water content and free water content fur-ther decreased to 0.24 ± 0.01 g g�1 dry solid and 0.25 ± 0.01

g g�1 dry solid, respectively. The corresponding removal ratiosreached 88.4% (wt%) and 93.4% (wt%) compared with the watercontent in the AS after compression.

Fig. 3f shows the transition of VS/TS values as the AS PEOD pro-gressed. After the conditioning stage, the ratio of VS/TS increasedfrom 61.4% to approximately 64.0% for original AS. Such anincrease can be attributed to both the CPAM addition and the

6 J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11

phase transfer of some solid matters from AS to its supernatant. Ingeneral, the VS in the dewatered AS remained constant during theDG and compression stages, whereas the organic matters in ASdecreased to approximately 60.0% during the EC stage.

3.2.2. SEM morphology of ASThe morphology of the dewatered AS flocs was observed using

SEM at the end of each dewatering stage. Fig. 4 shows the corre-sponding SEM images. AS became increasingly compact as thePEOD progressed. As can be seen in Fig. 4a, the loose original AScontained many pores. After the conditioning stage, the poresbecame smaller, and the application of the following DG and com-pression only resulted in a slight reduction in pore size. However,the electric field had a significant effect on the structure of AS flocs,and the large pores were hardly presented in Fig. 4e. Barton et al.[22] indicated that during the sludge electro-dewatering process,the beneficial effects due to electrophoresis were minimal, theprincipal mechanism for sludge almost certainly was electro-osmosis. Therefore, the compact structure of AS after the EC stagecan be attributed to the synergistic effects of electro-osmosis andcompression.

In Fig. 4a, Filamentous, rod-shaped, and coccoid microorgan-isms were sparsely distributed in the sludge. After the condition-ing, DG and compression stage, the distribution density increased(Fig. 4b–d). Fig. 4a–d illustrate that the cells remain intact, whereasFig. 4e shows some cell debris, which are marked with an ellipse. Ingeneral, the conditioning-DG-compression operation can removelarge quantities of water in AS, but it cannot rupture the cells.According to previous reports [23–25], when an electric field wasexposed to sludge cells, a transmembrane potential caused bythe presence of opposite charges on either side of the cell mem-brane can be induced. Then, a compressive pressure on the cellmembrane yields due to the attraction between these oppositecharges, which causes irreversible electroporation on the cellmembrane and thus disrupts the cell at the transmembrane poten-tial higher than about 1 V. The combined thermal effect can

Fig. 4. SEM morphology of the dewatered AS at different stages: (a) origina

accelerate the rupture of cells. The dewatering ratio was improvedby the release of the cytoplasm as soon as the cell was ruptured.

3.2.3. Inner structure of the AS matrixThe inner structures of the AS matrix at different PEOD stages

were characterized by using the paraffin section technology com-bined with the image analysis. For brevity, only the images of themiddle paraffin sections of these AS matrix were presented(Fig. 5b–e). As can be seen in the figures, the AS matrix consistedof many cores and pores that formed a network-like topology.The cores were concentrated biomass [26], and numerous poreswere distributed among the biomass cores. As can be seen inFig. 5a, the slice of original AS presented a loose structure withpores of different sizes. Some fiber-like matters scattered aroundthe biomass particles were assumed to be filamentous bacteria.The average area of the pores in the section was calculated to be7.4 � 10�5 mm2, and the corresponding slice porosity was 60.07%(Table 3). Fig. 5b shows the inner structure of a typical conditionedAS aggregate. The slice porosity decreased to 38.00%, and the aver-age area of the pores in the section increased to 11.6 � 10�5 mm2.In the DG stage, the conditioned AS aggregates were further com-pressed by gravity and the slice porosity declined to 29.85%. As canbe seen in Fig. 5d, the structure of the AS matrix became more uni-form at the compression stage. The slice porosity was 29.19%, andthe average area of the pores in the section increased from4.0 � 10�5 mm2 at the DG stage to 5.1 � 10�5 mm2, implying thatthe number of pores in the AS matrix at the compression stagedecreased and that a larger single passage for water flow formed.With the evolution of the PEOD process before the EC stage, theAS matrix maintained the similar network-like structure, althoughthe biomass core was compact and distributed more uniformly.However, this network-like structure vanished when a combina-tion of voltage and pressure was induced in the PEOD process.More compact and larger biomass cores occurred in the AS matrix,and a quantity of narrow and parallel slits was distributed accord-ingly. In addition, the slice porosity sharply decreased to 12.87%and the average area of the slits in the section increased to

l; (b) conditioning; (c) DG; (d) compression; and (e) EC (600 kPa, 50 V).

Fig. 5. Images of the AS paraffin sections at different dewatering stages: (a) original; (b) conditioning; (c) DG; (d) compression; and (e) EC (600 kPa, 50 V).

Table 3Characteristics of the inner structure of the AS flocs at different dewatering stages.

Stage Fractaldimension

Sliceporosity (%)

Average area of pores in thesection (mm2)

Original 1.76 60.07 7.4 � 10�5

Conditioning 1.81 38.00 11.6 � 10�5

DG 1.91 29.85 4.0 � 10�5

Compression 1.92 29.19 5.1 � 10�5

EC 1.96 12.87 6.8 � 10�5

J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11 7

6.8 � 10�5 mm2, providing a much larger and straighter passagefor water flow. Chu et al. [27] investigated the porosity of theanaerobic sludge cake during electro-osmosis dewatering andfound that the dewatered sludge cake had porosities between64% and 84% under a horizontal electrical field without mechanicalpressure. In this study, the combination of high mechanical pres-sure and voltage resulted in a more compact dewatered sludgecake after EC stage than that reported by Chu et al. [27]. The fractaldimensions (DBC) can be determined to quantify the ‘‘disappearingrate’’ of the pores in the dewatered AS matrix. As can be seen inTable 1, as the dewatering progressed, the overall steady upwardtrend of DBC indicated that the AS matrices became more compact.This result is in accordance with the variation in slice porosity.Moreover, in the former three stages, the fractal dimension dra-matically increased from 1.76 for the original AS to 1.91 for theAS matrix at the DG stage. Then, the fractal dimension slightlyincreased to 1.96 at the end of the EC stage.

Urbain et al. [28] considered the extracellular polymeric sub-stances (EPS) and moisture, including free water and bound water,as the primary contributors to the maintenance of the network-likestructure of the AS. When AS was treated during the conditioningand DG stages, their DSC tests indicated that some bound waterwas converted to free water [20,21], and most free water and anamount of the EPS were removed. Furthermore, a more compactnetwork structure was observed in the dewatered AS than in the

original AS. As the bound water and EPS left from the AS at thecompression stage, an amount of small pores in the AS collapsedand large pores appeared, and thus, the network structure in theAS became more compact. Nevertheless, under the combinedeffects of the 600 kPa pressure and 50 V voltage during the ECstage, the propagation of ionized mobile liquid on a solid–liquidinterface drove the neutral liquid in the central channel throughviscous momentum transfer [8]. In previous study of Yuan et al.[29], EPSs were destroyed by electrolysis reactions, releasingbound water into the filtrate. This phenomenon was in consistentwith the observation in this study. However, the EPS polymer arenegatively charged molecular [30], they must migrate toward theanode. These processes cause the destruction of the network struc-ture of the AS matrix and the formation of some parallel narrowslits. Simultaneously, significant changes in the inner morphologyof the AS matrix occurred.

3.3. Variations in the characteristics and components of the filtrate

3.3.1. Organics3.3.1.1. COD/SCOD and TOC. The variations in the COD, SCOD, andTOC of the filtrate during the PEOD process are presented in Table 4.These three parameters declined after the conditioning stage, andthen exhibited ascending trends in the following stages. The signif-icant increases of COD, SCOD and TOC after the DG and compres-sion stages indicated that some colloids were also introducedinto the filtrate at the compression stage, and contributed to theorganics in filtrate. After the EC stage, the three parameters wereapproximately 6.67, 7.42, and 7.41 times higher compared withtheir values after the compression stage. Tuan et al. [12] showedthat an electric field application induced the migration of nega-tively charged organic matter from the cathode to the anode, andit was removed from sludge by the water flow in pressure-drivenexperiments. Therefore, COD and TOC concentrations in the filtratewere higher at the anode side and lower at the cathode side. How-ever, in this study the filter was collected only from the cathode.Some dissolved organic matters in water flow could induce a rise

Table 4COD/SCOD and TOC in the filtrate at different PEOD stages.

Stage COD (mg/L) SCOD (mg/L) TOC (mg/L) pH Conductivity (mS/cm)

Original 128 ± 6.5 77 ± 3.5 40.71 ± 5.1 6.91 1.41Conditioning 76 ± 2.5 61 ± 5.9 28.64 ± 3.4 6.90 1.41DG 252 ± 10.3 173 ± 23.6 83.86 ± 15.6 7.44 1.43Compression 824 ± 21.2 681 ± 11.4 215.8 ± 13.8 7.80 1.61EC 5500 ± 102.4 5050 ± 112.3 1600 ± 80.4 10.73 3.03

8 J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11

of organic content in the filtrate after EC compared with the formerstages. Moreover, the discrepancy between COD and SCOD inTable 4 shows that the filtrate contained several organic particles.Except for the conditioning stage, the ratio of COD to SCODdecreased from 1.66 to 1.09 during the PEOD process, implyingthat soluble organics gradually became the main constituents ofCOD in the filtrate. In addition, the decrease of 52 mg/L in theCOD and 16 mg/L in the SCOD of the filtrate after the conditioningstage indicated that more organic particles in the AS suspensionwere flocculated and removed from the supernatant. However, lesssoluble organics could be reduced by CPAM flocculation. In addi-tion, Table 4 also presents both the pH and conductivity of the fil-trate at the end of each dewatering stage during the PEOD process.It can be seen that pH and conductivity of the filtrate slightlyincreased in the former four stages, but sharply increased in theEC stage. The detailed analysis was shown in SI.

3.3.1.2. EEMs. Fig. 6 shows the three-dimensional EEM fluores-cence spectrometry of the filtrates at different dewatering stages.As can be seen in the EEMs in Fig. 6a, the supernatant of the origi-nal AS contained protein-like (Peaks A and B), visible fulvic acid-like (Peak C), and ultraviolet fulvic acid-like (Peak D) substances[31–33], which were identified by the following situations at (Ex/Em): 280/340–350 nm, 230/340–350 nm, 320–360/370–450 nm,and 240–260/430–440 nm, respectively. In the following

Em (nm)250 300 350 400 450 500 550 600

Ex

(nm

)

200

250

300

350

4000 500 1000 1500 2000

Em (250 300 350 400

Ex

(nm

)

200

250

300

350

400

Em (nm)250 300 350 400 450 500 550 600

Ex

(nm

)

200

250

300

350

4000 1000 2000 3000 4000 5000 6000 7000 8000

Em (

250 300 350 400

Ex

(nm

)

200

250

300

350

400

(a) (b)

)e()d(

Peak A Peak A

Peak A Peak C

Peak B

Peak C

Peak D Peak B

Peak B Peak D

Peak C

Fig. 6. EEM fluorescence spectra and the peak intensity variations of the filtrate at differe(600 kPa, 50 V); and (f) fluorescent intensity.

conditioning, DG, and compression stages, more than three kindsof organic matters acted as the main components in the filtrate.Moreover, the peak intensity in the EEMs increased as the dewater-ing progressed to the compression stage, as shown in Fig. 6f. Moreidentified organics remained in the water phase of the AS matrix asthe physical dewatering process progressed [34]. In the compres-sion stage, the intensities of the peak A, B, C and D reached 8174,5110, 3567, and 3638, respectively. Nevertheless, aside from thevisible fulvic acid-like substance that appeared in the filtrate asthe dewatering process progressed to the EC stage, a new peak cor-responding to a humic acid-like substance (marked as E) appearedat (Ex/Em) 390/444 nm. This peak could be induced by the applica-tion of voltage. The intensities of both peaks C and E reached morethan 9900, as shown in Fig. 6f. Thus, the humic acid-like and thevisible fulvic acid-like substances appeared.

3.3.2. Content and zeta potential (ZP) of the colloidsThe content and ZP of the colloids in the filtrate were measured

after each dewatering stage, and the results are presented inTable 5. As can be seen in Table 5, the content of colloids (as SS)was reduced from 198.37 mg/L to 68.73 mg/L in the water phaseof the AS matrix via CPAM conditioning, and the ZP value of thecolloid was reversed from �9.51 mV to 1.81 mV. After the DGstage, the content of colloids in the filtrate slightly decreased.Under the gravity, the structure of the sludge was not destroyed,

nm)450 500 550 600

0 500 1000 1500 2000 2500

Em (nm)250 300 350 400 450 500 550 600

Ex

(nm

)

200

250

300

350

4000 1000 2000 3000 4000 5000

nm)

450 500 550 600

0 2000 4000 6000 8000 10000

OriginalConditioning DG Compression EC0

2000

4000

6000

8000

10000 Peak A:Protein-like Peak B:Protein-like Peak C:Visible fulvic acid-like Peak D:Urtraviolet fulvic acid-like Peak E:Humic acid-like

Inte

nsity

Dewatering Stage

(f)

(c)

Peak A Peak C

Peak D Peak B

Peak C

Peak D

Peak E

nt dewatering stages: (a) original; (b) conditioning; (c) DG; (d) compression; (e) EC

Table 5Content and ZP of colloids in the filtrate at different dewatering stages.

Stage SS (mg/L) Zeta potential (mV)

Original 198.37 �9.51Conditioning 68.73 1.81DG 62.09 �10.15Compression 563.19 �10.82EC 792.33 �20.76

J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11 9

and the filtrate was still the supernatant of the conditioning sludge.Moreover, the application of filter cloth trapped the colloids in thesupernatant. The ZP value of the colloids changed to a negativevalue of �10.15 mV, implying the transfer of some soluble organicsubstances with negative charge from the sludge phase to theresidual supernatant of the conditioned AS matrix. Then, the con-tent of colloids sharply increased to 563.19 mg/L at the end ofthe compression stage, and the ZP value of the colloids slightlychanged. The significant increase of colloid content suggested thatcompression ruined the physical structure of sludge, whichinduced numerous colloids into the filtrate. After the EC stage,the content of colloids in the filtrate increased to 792.33 mg/L,and the corresponding ZP value of the colloids was reduced to�20.76 mV. As can be seen in Table 2, the SS in the water phaseof the AS matrix became the main contributor to the differencebetween COD and SCOD as the PEOD progressed. Moreover, thesignificant decline in the ZP value of the colloids after the EC stagecan be related to the cell rupture, cytoplasm release, and pHincrease [35] in the filtrate (As given in Supporting information).

Table 6Changes in the contents of metal elements in filtrate at different PEOD stages (ppm).

Stages Na Ca Mg Fe Cu

Original 102.43 71.30 65.62 0.21 0.0Conditioning 101.43 72.44 62.26 0.14 0.0DG 105.11 74.99 56.99 0.13 0.0Compression 105.67 76.51 59.56 0.13 0.0EC 269.35 262.21 126.22 41.63 0.2

Table 7Changes in the contents of metal elements in dewatered AS at different PEOD stages (ppm

Stages Na Ca Mg Fe Cu

Original 3307.8 24656.6 7459.9 64654.7 54Conditioning 4369.1 30685.9 9392.4 77774.0 62DG 3819.1 30293.5 9094.1 77773.3 62Compression 3517.5 30075.1 8924.1 77772.9 62EC 1397.0 28010.9 7930.4 77445.2 62

Table 8ANOVA for the quadratic model for the final dry solids content.

Source Sum of squares df Mean squa

Model 2416.83 5 483.37A 367.02 1 367.02B 1267.33 1 1267.33AB 420.47 1 420.47A2 1.16 1 1.16B2 345.80 1 345.80Residual 117.62 5 23.52

Lack of fit 117.11 3 39.04Pure error 0.51 2 0.26

Cor total 2534.45 10 –

R2 = 0.9536; R2adj ¼ 0:9072.

3.3.3. Metal ionsTables 6 and 7 show the variations in the contents of metal ele-

ments in the filtrate and the AS during the PEOD process. In gen-eral, the contents of detected metals ions in the filtrate barelychanged during the former four stages, and experienced sharpboosts at the EC stage. The opposite variation trend of aforemen-tioned metal elements can be observed in the dewatered AS. Thedetailed analysis was shown in SI.

3.4. Analysis of the energy consumption

The combination of final dry solids content and energy con-sumption was used as the response (Y1 and Y2), and the datamatrixes (X1, X2, Y(observed)) were taken to fit Eq. (3) by usingExpert Design 7.0. Results of optimum PEOD conditions based onthe CCD and the corresponding analysis of variance (ANOVA) werepresented in Tables 2, 8 and 9, respectively. The following second-order fitting polynomial equations were obtained after the data fit-ting, as shown in Eqs. (5) and (6).

Y1 ¼ 32:472� 0:024�X1 � 2:051�X2 þ 2:563E� 003�X�1X2

� 1:694E� 005X21 þ 0:029X2

2 ð5Þ

Y2 ¼ 0:106� 6:356E� 004�X1 þ 1:874E� 003�X2 þ 1:191E

� 005�X1X2 þ 5:605E� 007X21 � 4:845E� 005X2

2 ð6Þ

As presented in Tables 8 and 9, good fitting results were obtainedbecause all the determination coefficients were greater than 0.79.The p-value less than 0.05 indicated that the model terms were sig-

Zn Mn Cr Pb Ni Co

0 0.03 0.00 0.02 0.07 0.00 0.000 0.00 0.00 0.02 0.10 0.00 0.000 0.00 0.00 0.02 0.08 0.00 0.000 0.00 0.00 0.02 0.12 0.00 0.007 1.43 0.09 0.06 0.15 0.24 0.10

).

Zn Mn Cr Pb Ni Co

9.3 556.2 188.8 90.5 56.8 49.7 6.43.8 625.7 118.0 73.1 65.3 102.6 2.63.8 625.7 118.0 73.0 64.9 102.6 2.63.8 625.7 118.0 72.9 64.6 102.6 2.61.6 614.4 117.3 72.5 63.4 100.7 1.8

re F value p-value Prob > F

20.55 0.0024 Significant15.60 0.0109 –53.87 0.0007 –17.87 0.0083 –0.049 0.8328 –14.70 0.0122 –– – –152.52 0.0065 Significant– – –– – –

Table 9ANOVA for the quadratic model for the energy consumption.

Source Sum of squares df Mean square F value p-value Prob > F

Model 0.051 5 0.010 8.83 0.0160 SignificantA 6.936E�003 1 6.936E�003 5.99 0.0581B 0.033 1 0.033 28.83 0.0030AB 9.073E�003 1 9.073E�003 7.83 0.0380A2 1.274E�003 1 1.274E�003 1.10 0.3424B2 9.514E�004 1 9.514E�004 0.82 0.4063Residual 5.791E�003 5 1.158E�003

Lack of fit 5.788E�003 3 1.929E�003 1266.49 0.0008 SignificantPure error 3.047E�006 2 1.523E�006

Cor total 0.057 10

R2 = 0.8983; R2adj ¼ 0:7966.

10 J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11

nificant. Therefore, the polynomial equation for the data fittingabove is a reliable model that can be used to predict the optimumdewatering conditions for AS. The optimum dewatering conditionswere determined as a combination of 401 kPa pressure and 50 Vvoltage for the experiment set and the initial mass of sludge usedin this study, which gave a dry solids content and energy consump-tion of 41.90% (wt%) and 0.153 kWh per kg removed water.

4. Conclusions

The electric field greatly removed the bound water and had asignificant effect on the AS dewatering efficiency of the PEOD pro-cess. As the PEOD progressed, a more compact and uniform struc-ture was observed. At the end of EC stage, the ruptured bacteriacells and some narrow and parallel slits appeared in AS. In addi-tion, the contents of organic matters in filtrate increased withthe progress of the PEOD. And the electrochemical reactionoccurred near the electrodes of the PEOD cell at EC stage, resultingin the formation of humic acid-like and visible fulvic acid-like sub-stances in filtrate.

Acknowledgments

This work was supported by the National Natural Science Foun-dation of China (Nos. 21177010 and 51078035), the FundamentalResearch Funds for the Central Universities (Nos. JC2011-1 andTD2010-5), the Ph.D. Programs Foundation of the Ministry of Edu-cation of China (No. 20100014110004), and the Technology Foun-dation for Selected Overseas Chinese Scholar, Ministry of Personnelof China.

Appendix A. Supplementary material

Sludge samples; images of the AS matrix at different dewateringstages; pH and conductivity; metal ions.

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.seppur.2014.07.019.

References

[1] D. Dursun, S. Dentel, Toward the conceptual and quantitative understanding ofbiosolids conditioning: the gel approach, Water Sci. Technol. 59 (2009) 1679–1685.

[2] A. Mahmoud, J. Olivier, J. Vaxelaire, A.F. Hoadley, Advances in MechanicalDewatering of Wastewater Sludge Treatment, Wastewater Reuse andManagement, Springer, 2013, pp. 253–303.

[3] J.E. Lee, J.K. Lee, H.K. Choi, Filter press for electrodewatering of waterworkssludge, Drying Technol. 25 (2007) 1649–1657.

[4] A. Mahmoud, J. Olivier, J. Vaxelaire, A.F. Hoadley, Electrical field: a historicalreview of its application and contributions in wastewater sludge dewatering,Water Res. 44 (2010) 2381–2407.

[5] P.A. Tuan, S. Mika, Fractionation of macro and trace metals due to off-timeinterrupted electrodewatering, Drying Technol. 28 (2010) 762–772.

[6] M. Citeau, J. Olivier, A. Mahmoud, J. Vaxelaire, O. Larue, E. Vorobiev,Pressurised electro-osmotic dewatering of activated and anaerobicallydigested sludges: electrical variables analysis, Water Res. 46 (2012) 4405–4416.

[7] A. Mahmoud, J. Olivier, J. Vaxelaire, A.F. Hoadley, Electro-dewatering ofwastewater sludge: influence of the operating conditions and theirinteractions effects, Water Res. 45 (2011) 2795–2810.

[8] Z. Yang, X.F. Peng, D.J. Lee, Electroosmotic flow in sludge flocs, Int. J. Heat MassTransf. 52 (2009) 2992–2999.

[9] S.C. Hwang, K.S. Min, Improved sludge dewatering by addition of electro-osmosis to belt filter press, J. Environ. Eng. Sci. 2 (2003) 149–153.

[10] H. Saveyn, P. Van Der Meeren, G. Pauwels, R. Timmerman, Bench-and pilot-scale sludge electrodewatering in a diaphragm filter press, Water Sci. Technol.54 (2006) 53–60.

[11] P.A. Tuan, S. Mika, I. Pirjo, Sewage sludge electro-dewatering treatment—areview, Drying Technol. 30 (2012) 691–706.

[12] P.A. Tuan, V. Jurate, S. Mika, Electro-dewatering of sludge under pressure andnon-pressure conditions, Environ. Technol. 29 (2008) 1075–1084.

[13] P.A. Tuan, I. Pirjo, S. Mika, Effect of polyelectrolyte conditioning and voltageson fractionation of macro and trace metals due to sludge electro-dewatering,Sep. Sci. Technol. 47 (2012) 788–795.

[14] A. Esmaeily, M. Elektorowicz, S. Habibi, J. Oleszkiewicz, Dewatering andcoliform inactivation in biosolids using electrokinetic phenomena, J. Environ.Eng. Sci. 5 (2006) 197–202.

[15] J. Huang, M. Elektorowicz, J.A. Oleszkiewicz, Dewatering and disinfection ofaerobic and anaerobic sludge using an electrokinetic (EK) system, Water Sci.Technol. 57 (2008) 231–236.

[16] APHA, AWWA, WPCF, Standard methods for the examination of water andwastewater, 22th ed, Amer Public Health Assn, Washington DC, USA,2005.

[17] C.P. Chu, D.J. Lee, X.F. Peng, Structure of conditioned sludge flocs, Water Res. 38(2004) 2125–2134.

[18] D.J. Lee, S.F. Lee, Measurement of bound water content in sludge: the use ofdifferential scanning calorimetry (DSC), J. Chem. Technol. Biotechnol. 62(1995) 359–365.

[19] E.J. Wahlberg, T.M. Keinath, D.S. Parker, Influence of activated sludgeflocculation time on secondary clarification, Water Environ. Res. 66 (1994)779–786.

[20] C.P. Chu, D.J. Lee, Moisture distribution in sludge: effects of polymerconditioning, J. Environ. Eng. 125 (1999) 340–345.

[21] J. Vaxelaire, P. Cézac, Moisture distribution in activated sludges: a review,Water Res. 38 (2004) 2215–2230.

[22] W.A. Barton, S.A. Miller, C.J. Veal, The electrodewatering of sewage sludges,Drying Technol. 17 (1999) 498–522.

[23] C.Y. Lee, G.B. Lee, J.L. Lin, F.C. Huang, C.S. Liao, Integrated microfluidic systemsfor cell lysis, mixing/pumping and DNA amplification, J. Micromech. Microeng.15 (2005) 1215.

[24] Y.H. Lin, G.B. Lee, An optically induced cell lysis device using dielectrophoresis,Appl. Phys. Lett. 94 (2009). 033901-033901-033903.

[25] Y.H. Lin, G.B. Lee, An integrated cell counting and continuous cell lysis deviceusing an optically induced electric field, Sens. Actuators B: Chem. 145 (2010)854–860.

[26] D.H. Li, J.J. Ganczarczyk, Structure of activated sludge floes, Biotechnol. Bioeng.35 (1990) 57–65.

[27] C.P. Chu, D.J. Lee, Z. Liu, W.H. Jin, Morphology of sludge cake at electroosmosisdewatering, Sep. Sci. Technol. 39 (2005) 1331–1346.

[28] V. Urbain, J.C. Block, J. Manem, Bioflocculation in activated sludge: an analyticapproach, Water Res. 27 (1993) 829–838.

[29] H.P. Yuan, N.W. Zhu, L.J. Song, Conditioning of sewage sludge with electrolysis:effectiveness and optimizing study to improve dewaterability, Bioresour.Technol. 101 (2010) 4285–4290.

[30] L.H. Mikkelsen, K. Keiding, Physico-chemical characteristics of full scalesewage sludges with implications to dewatering, Water Res. 36 (2002)2451–2462.

J. Feng et al. / Separation and Purification Technology 134 (2014) 1–11 11

[31] A. Baker, Fluorescence excitation–emission matrix characterization of somesewage-impacted rivers, Environ. Sci. Technol. 35 (2001) 948–953.

[32] W. Chen, P. Westerhoff, J.A. Leenheer, K. Booksh, Fluorescence excitation–emission matrix regional integration to quantify spectra for dissolved organicmatter, Environ. Sci. Technol. 37 (2003) 5701–5710.

[33] A. Baker, R. Inverarity, Protein-like fluorescence intensity as a possible tool fordetermining river water quality, Hydrol. Process. 18 (2004) 2927–2945.

[34] F. Wu, D. Kothawala, R. Evans, P. Dillon, Y. Cai, Relationships between DOCconcentration, molecular size and fluorescence properties of DOM in a stream,Appl. Geochem. 22 (2007) 1659–1667.

[35] M. Citeau, O. Larue, E. Vorobiev, Influence of salt, pH and polyelectrolyte on thepressure electro-dewatering of sewage sludge, Water Res. 45 (2011) 2167–2180.