The treatment of pharmaceutical wastewater using in a submerged membrane bioreactor under different sludge retention times

Author's Accepted Manuscript

The treatment of pharmaceutical wastewaterusing in a submerged membrane bioreactorunder different sludge retention times

Yasemin Kaya, Gamze Ersan, Ilda Vergili, Z.Beril Gönder, Gulsum Yilmaz, Nadir Dizge,Coskun Aydiner

PII: S0376-7388(13)00262-7DOI: http://dx.doi.org/10.1016/j.memsci.2013.03.059Reference: MEMSCI12047

To appear in: Journal of Membrane Science

Received date: 26 December 2012Revised date: 22 March 2013Accepted date: 26 March 2013

Cite this article as: Yasemin Kaya, Gamze Ersan, Ilda Vergili, Z. Beril Gönder,Gulsum Yilmaz, Nadir Dizge, Coskun Aydiner, The treatment of pharmaceu-tical wastewater using in a submerged membrane bioreactor under differentsludge retention times, Journal of Membrane Science, http://dx.doi.org/10.1016/j.memsci.2013.03.059

This is a PDF file of an unedited manuscript that has been accepted forpublication. As a service to our customers we are providing this early version ofthe manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting galley proof before it is published in its final citable form.Please note that during the production process errors may be discovered whichcould affect the content, and all legal disclaimers that apply to the journalpertain.

www.elsevier.com/locate/memsci

The treatment of pharmaceutical wastewater using in a submerged

membrane bioreactor under different sludge retention times

Yasemin Kaya a*, Gamze Ersan a, Ilda Vergili a, Z. Beril Gönder a, Gulsum Yilmaz a,

Nadir Dizge b, Coskun Aydiner b

a Istanbul University, Faculty of Engineering, Department of Environmental Engineering, 34320, Istanbul, Turkey

bGebze Institute of Technology, Faculty of Engineering, Department of Environmental

Engineering, 41400, Kocaeli, Turkey

Abstract

The performance of a lab-scale submerged membrane bioreactor system (SMBR) for treating

a process wastewater containing the pharmaceutical active compound (PhAc) etodolac with

four different microfiltration (MF) membranes (MP005, MV02, CA, and MCE) was

investigated at three different sludge retention times (SRTs) under constant pressure. In the

first phase of the study, the continuous bioreactor system was operated at SRTs of 15 and 30

days and without sludge wasting (WSW). After steady state conditions were reached, the

SMBR process was started as the second phase. Short-term filtration (24 h) tests were

conducted for each SRT. When the SRTs were increased, both permeate volumes and steady-

state flux values increased. The best etodolac removals were obtained in case of WSW for

both bioreactor system and SMBR. The etodolac removals achieved by the different

membranes for the period WSW were observed in the following order: MV02=

MP005>CA=MCE. In addition, the chemical oxygen demand (COD) removal efficiencies for

bioreactor system and SMBR were approximately 80±2% and 86±2%, respectively at all

SRTs. The COD removals at each of the three SRTs were similar for all of the membranes.

Extracellular polymeric substances (EPSs), soluble microbial products (SMPs) were analyzed

as fouling control factors. Increasing the SRT caused increases in sludge concentrations in the

SMBR as well as increased etodolac removal, while EPS and SMP protein and carbohydrate

concentrations decreased. Fouling on the pores and surfaces of the membranes were

characterized using a Scanning Electron Microscope (SEM), an Atomic Force Microscope

(AFM), a Fourier transform infrared spectroscopy (FT-IR), and contact angle measurements.

The resistance in series model was used to evaluate the flux decline caused by the gel layer,

cake resistance, and internal pore blocking in the MF membranes at the three different SRTs.

Research Highlights

1. The etodolac removals achieved by the different membranes for the period WSW:

MV02=MP005>CA=MCE.

2. COD removal efficiencies for bioreactor system and MBR were approximately 80±2% and

86±2%, respectively at all SRTs.

3. MP005 had the highest steady state flux for each of the different SRTs.

4. EPS and SMP protein and carbohydrate concentrations decreased with increasing the SRT.

5. Membrane fouling was observed with SEM, AFM, contact angle, and FT-IR analysis at all

SRTs

Keywords: Pharmaceutical wastewater; submerged membrane bioreactor; Sludge retention

time; Resistance in series model

* Corresponding author at: Istanbul University, Faculty of Engineering, Department of Environmental

Engineering, Avcilar Campus, Avcilar, 34320, Istanbul-TURKEY.

Tel.: +90 212 4737070; fax: +90 212 4737180.

E-mail address: [email protected] (Y. Kaya).

1. Introduction

Pharmaceutical wastewater is a major, complex and toxic industrial waste because it

contains a variety of products. This wastewater contains relatively high levels of suspended

solids and soluble organics, many of which are recalcitrant [1]. Generally, the removal of

these compounds by conventional wastewater treatment processes is not effective and

removal mechanisms are not well understood. Economic reasons make biological treatment

processes appropriate for the treatment of high strength industrial wastewater. However,

conventional treatment systems are not sufficient for meeting discharge limits and

pharmaceutical active compounds (PhAcs) are often not effectively removed from

wastewater. Only a few studies reported in the literature have dealt with ozonation, O3/H2O2,

O3/UV, Fenton, and adsorption processes [2�5]. In recent years, membrane bioreactors have

been developed as an alternative method to conventional activated sludge (CAS) systems for

the removal of these organic micropollutants.

Membrane bioreactors (MBRs) are increasingly being used for municipal and industrial

wastewater treatment [6]. The main advantages of MBRs over other biological systems are

not only that they are more useful for disinfection purposes but also that they have smaller

footprints, produce less sludge, result in better effluent qualities, have longer SRTs

independent of hydraulic retention times, and allow for the rapid start-up of biological

processes [6�12]. MBRs integrate the biological degradation of waste products with

membrane filtration, ensuring effective removal of organic contaminants and nutrients from

municipal and/or industrial wastewaters [13]. Most of these features result from the

uncoupling of the sludge retention times (SRT) and the hydraulic retention time (HRT), which

reflects the basic concept that solid/liquid (sludge/effluent) separation is provided through

highly efficient membrane filtration rather than the traditional gravity settling [13].

There have been several studies demonstrating the high removal of PhAcs during MBR

treatment of municipal wastewater [14�17]. In a recent study, Pérez and Barceló

demonstrated the 56% elimination of the human metabolite of diclofenac, 40-

hydroxydiclofenac, in a laboratory- scale MBR was achieved versus only 26% removal in

CAS treatment [18]. Radjenovic et al. [14] reported on performances of full-scale CAS

treatment and two pilot-scale MBRs in eliminating PhAcs belonging to different therapeutic

groups with different physico-chemical properties. Two pilot-scale MBRs exhibited enhanced

elimination of several pharmaceutical residues that were poorly removed by CAS treatment,

whereas in some cases the more stable operation of one of the MBR reactors at prolonged

SRTs proved to be detrimental to the elimination of some compounds. Clara et al. [19]

analyzed eight PhAcs, two polycyclic musk fragrances and nine endocrine disrupting

chemicals in several wastewater treatment plants (WWTPs). A pilot scale membrane

bioreactor was operated at different SRTs and the results were compared to those observed in

conventional activated sludge plants (CASP). Some compounds, such as the antiepileptic drug

carbamazepine, were not removed in any of the sampled treatment facilities and effluent

concentrations similar to the influent concentrations were measured. Other compounds, such

as bisphenol-A, the analgesic ibuprofen and the lipid regulator bezafibrate, were removed

(higher than 90% removal). The process wastewater used in this study contains an analgesic

PhAc namely etodolac. This PhAc is a member of the pyranocarboxylic acid group of non-

steroidal anti-inflammatory drugs. Etodolac is one of the most-produced PhAcs in Turkey. A

survey of the literature reveals that there is no research available for the occurrence and

removal of etodolac from water and wastewaters.

SRT is an important factor affecting the performance of bioreactors, as the SRT can have a

significant influence on the biomass properties in an MBR system. The operation of WWTPs

with high SRTs was suitable for nitrogen removal (SRT › 10 days at 10 °C) and increased the

potential for the removal of selected micropollutants. As in conventional WWTPs, the

removal efficiency of MBR treatment depends on the SRT. A long SRT is considered an

advantage because the amount of sludge produced is reduced and the cost of sludge handling

and disposal is reduced. Since higher biomass concentrations results higher treatment

efficiencies, many MBRs are operated with higher SRTs. Nevertheless, at high SRTs, dead or

inactive microorganisms accumulate in the MBR, affecting sludge activity [20, 21]. Another

important factor related to SRT is membrane fouling, which causes significant declines in

flux, increased operating costs, and decreased membrane lifetimes [22].

Most research has been focused on the identification of MBR membrane fouling

mechanisms under different SRTs. In this study, removal efficiencies, sludge properties and

membrane fouling mechanisms in a lab-scale SMBR were investigated. This system, treating

a pharmaceutical process wastewater, was studied under constant pressure for SRTs of 15 and

30 days and WSW. Permeates from four different MF membrane modules under the constant

pressure were collected and analyzed. Chemical oxygen demand (COD) was measured to

evaluate the performance of the system. The microbial flock structures, extracellular

polymeric substances (EPSs), soluble microbial products (SMPs), and surface properties

(hydrophobicities) were analyzed as controlling factors. The hydrophobicity of the membrane

surface was analyzed by contact angle measurements. Membrane fouling on the pores and the

surface was observed with using Scanning Electron Microscope (SEM), Atomic Force

Microscope (AFM) and Fourier transform infrared spectroscopy (FT-IR) analysis. Zeta

potential measurements were used to determine the surface charges of substances in the

wastewater. In addition, the membrane fouling mechanism was investigated by resistance in

series model. In each phase, the HRT was kept constant at 24 hours (h) and a 1.5 kg COD m–3

d–1 organic loading rate (OLR) was applied.

2. Materials and methods

2.1. Experimental set-up

A laboratory-scale SMBR was used in this study. A schematic representation of the

experimental setup is shown in Fig. 1. Four flat sheet membrane modules were immersed in

the activated sludge system that consisted of an aeration tank and a sedimentation tank. The

total volume of cylindrical plexiglas aeration tank was 14 L (with an effective volume of 8.5

L) with an inner diameter 25 cm and height of 29.5 cm. Compressed air was supplied by a

membrane diffuser at 8–10 L/min. The SMBR tank was aerated at the bottom to supply

oxygen for the microorganisms and to scour the membrane surface. The dissolved oxygen

(DO) concentration in the bioreactor was 4.6 ± 1.1 mg L-1. A sedimentation tank was placed

after the aerobic tank to maintain a constant HRT. The treated water at the top of the settling

tank was collected and discharged from the tank as final effluent. The total volume of

plexiglas sedimentation tank was 8.5 L with 13 cm inner diameter and 27 cm height. The

sludge in the sedimentation tank was transferred to the aeration tank with a sludge pump

(Sisdoz-PRS6). The continuous-flow back-mixed bioreactor was tested at the three different

SRTs. After reaching steady state conditions, four flat sheet membrane modules were

immersed in the aeration tank during the 24 h experiments. The sedimentation tank was

cancelled at that time and permeates were not returned to the feed tank. The permeates were

drawn through the membranes using vacuum pumps and were collected in the vessels. The

permeate flow rate was measured every minute by an electronic balance and recorded by a

computer via a PCI card and an RS 232 line.

The four flat sheet membrane modules, made of kestamit, had a total area of 84 cm2 were

used and the four different microfiltration membranes (MP005, MV02, CA, MCE) were

placed on the modules. MP005 and MV02 membranes, with pore sizes of 0.05 μm and 0.2

μm, were supplied by Microdyn-Nadir GmbH. Nitrocellulose Mixed Ester (MCE) and

Cellulose Acetate (CA) membranes, with pore sizes of 0.22 μm and 0.45 μm, were supplied

by GE Osmonics. The characteristics of the membranes are summarized in Table 1.

2.2. Wastewater characterization

Pharmaceutical process wastewater was fed into the SMBR system with a feed pump (Cole

Parmer Masterflex). The activated sludge inoculum and raw wastewater were obtained from a

wastewater treatment plant in Gebze, Turkey. The characteristics of the raw wastewater,

which contained etodolac, are shown in Table 2. The chemical and physical properties of the

PhAc in the raw wastewater are shown Table 3. Nitrogen and phosphate sources were added

as NH4Cl and KH2PO4, respectively, to maintain a COD/N/P ratio of 100/5/1.

2.3. Operating conditions

The continuous bioreactor system was operated at SRTs of 15 and 30 days and WSW. All

other process conditions were the same for each of the experiments. The SMBR was operated

over a period of eight months to investigate the influence of different SRTs on sludge

characteristics and membrane fouling. The HRT of the SMBR was controlled at 24 h. A

temperature controller was used in the SMBR to maintain a constant temperature of 25±1 oC.

In each phase, a 1.5 kg COD m–3 d–1 organic loading rate (OLR) was applied. The pH of the

bioreactor was adjusted to approximately 7.3 ± 0.2 with a phosphate buffer. The membrane

system had a pressure gauge (Keller PA-21Y) capable of measuring in the range of 0�4 bar,

and the transmembrane pressure was kept constant at 200 mbar throughout the experiments.

2.4. Analytical methods

Total suspended solids (TSS), volatile suspended solids (VSS), and COD were analyzed

according to Standard Methods [23]. Effluent and permeate The pH, DO concentration and

temperature in the bioreactor were controlled by a Hach-Lange SC1000 control unit.

EPS extraction was performed based on the methods reported by Frolund et al. [24]. After

reaching steady state conditions, samples for extraction were taken from the aeration tank.

The samples, with known VSS levels, were centrifuged. All of the experiments were hold

cold (flasks and tubes were kept in a container filled with ice). After centrifugation, the

supernatant of the sludge sample, called the SMP, was stored at -20 °C [24]. The EPSs were

extracted from residual sludge, called pellet, using a cation exchange resin (CER). The CER

was DOWEX Marathon C, a strongly acidic 20�50 mesh resin, in the sodium form (Sigma-

Aldrich). Liquid phase samples of the extractable portions of the EPS were stored at 4 °C

until analysis of the protein and polysaccharide fractions. The concentrations of proteins and

the polysaccharide fractions of the SMPs and EPSs were determined using the methods

reported by Lowry et al. [25] and Dubois et al [26], respectively. For measurements of

proteins and polysaccharide contents, Bovine Serum Albumin (BSA) and Alginate were used

as standards, respectively. The results were expressed in mg equivalent of BSA per liter for

protein content and mg equivalent of Alginate per liter for polysaccharide content. All

samples were measured using a UV–vis spectrophotometer (PG Instruments T60U) at the

wavelength of 750 nm for proteins and 480 nm for polysaccharide concentrations. All of the

analyses were conducted in duplicates.

Membrane fouling on the pores and surfaces was observed with AFM, SEM, and FT-IR

analysis. AFM (NanoScope IV, Digital Instruments, USA) was used to examine the surface

morphology of the different membrane surfaces before and after fouling. Clean membranes

were stored in pure water and dried at room temperature before the AFM measurements were

performed. Fouled membranes were carefully removed from the filtration cells after

experiments and stored in 100 mL of a 0.5% sodium bisulfite solution. To protect the fouling

layer on the membrane surface, measurements were made within one week with dried

membranes at room temperature. Membranes were cut into squares of 1cm×1 cm and glued

on a glass substrate. The samples were then attached to a magnetic sample holder located on

top of the scanner tube. Generally, there are three definitions of the roughness parameters: the

mean roughness (Ra), the root mean square (Rrms) of the average height of the membrane

surface peaks, and the mean difference in the height between the five highest peaks and the

five lowest valleys (Rz). Ra, Rrms and Rz were calculated according to the formulas defined by

Kaya et al. [27]. SEM measurements were carried out both before and after the filtration

experiments. The samples of membranes were fractured using liquid nitrogen and were then

coated with gold to obtain an adequate contrast of the membrane structure. SEM images were

obtained using a JEOL/JSM- 6335F-INCA instrument with an accelerating voltage of 10.0

kV. An FT-IR (Perkin-Elmer Spectrum One FTIR Spectrometer) was used to identity the

functional groups of the clean and fouled membranes. FT-IR analysis was measured with with

the attenuated total reflection (ATR) method.

Membrane hydrophobicity was quantified by measuring the contact angle that was formed

between the membrane surface and water. Contact angles were measured using the sessile

drop method with a goniometer (KSV Instruments, CAM 101). The particle size distribution

was determined with a laser particle size analyzer (Malvern Master Sizer 2000, UK MATH

(microbial adhesion to hydrocarbons) methods were used for the surface properties

(hydrophobicity). Zeta potential of sludge by Malvern Zetasizer Nano ZS90 was performed

according to Laser Doppler and phase Doppler techniques. Etodolac analysis of the samples

was made by Accredited (ALS) Laboratories of the Czech Republic. The concentration of

etodolac in each sample was determined by using a HPLC-MS-MS method. 5 mL of

wastewater sample was filtered and then directly injected. The analytes were separated on a

RP-column (Gemini C18 from Phenomenex) using a gradient of a methanol-water-eluent.

Etodolac-D3 was added as internal standard to all samples prior analysis. The API 4000

(ABSiex,CA) in the MRM-mode with 2 MRM – transitions was used for each analyte. The

quantification was made by standard addition to the sample which eliminates possible matrix

effects. The detection limit of method is 0,05 μg/L.

Effluent and permeate obtained from continuous bioreactor system and SMBR, respectively

were stored at 4 °C. After reaching steady state conditions in the bioreactor, etodolac was

measured in the effluent for SRTs of 15 and 30 days and WSW. COD of effluent was

monitored daily in the bioreactor. COD and etodolac were measured in the permeate obtained

from four flat sheet membrane modules after 24 h. The reactor content was sampled weekly

for determination of TSS and VSS.

2.5. Flux decline and membrane resistance analysis

Membrane fouling plays a significant role in filtration processes. The presence of this effect

was confirmed by the declines in permeate flux with increased processing times. This fouling

could be due to several factors, such as cake formation, adsorptive fouling mechanisms, and

pore blocking [28].

The permeate flux is described by Darcy’s law [29, 30]:

J= 1 dVA dt

where J is the permeate flux; A, the membrane filtration area; V, the total volume of permeate

and t is the filtration time.

The membrane resistances were determined from the flux data as follows: (1) total resistance

(Rt) was estimated based on the final flux of studied MF membrane; (2) membrane resistance

(Rm) was determined by measuring the clean membrane flux with deionized water; (3) at the

end of the filtration experiments, the membrane surfaces were rinsed with deionized water

and cleaned with a sponge to remove the fouling cake layer, and the fluxes with deionized

water were measured to obtain the combined membrane resistance and resistance due to pore

blockage (Rm + Rp); (4) the pore blockage resistance (Rp) was calculated as the difference of

the values measured in steps 3 and 2 ((Rm + Rp)�Rm); and (5) the cake resistance (Rc) was

calculated as the difference of the values measured in steps 1 and 3 (Rt � (Rm + Rp)) [31].

3. Result and discussion

3.1. Effects of SRT on the removals of PhAc and COD

All of the permeate samples were collected simultaneously from the MP005, MV02, CA,

and MCE membranes at constant pressure after the reactor reached steady state conditions.

COD and etodolac removal efficiencies were determined to compare the performances of the

SMBR at different SRTs with different membrane materials (Fig. 2). The sufficient COD

removal was achieved in period WSW for all membranes. The MP005 and MV02 membranes

had slightly higher COD removal efficiencies than the CA and MCE membranes in case of

WSW. COD removals for all of the membranes were higher than 86% whereas only 80%

removal efficiency was reach in the conventional activated sludge bioreactor. As the SRT

decreased from 30 to 15 days, the COD removal efficiencies changed gradually for both

systems.

COD removals by the MP005 membrane were 85.3 and 84.0% at 15 and 30 day SRTs,

respectively, and 87.3% in the WSW. COD removals by the MV02 membrane were 85.3 and

86.0% at 15 and 30 days SRTs, respectively, and 87.3% in the WSW. COD removals by the

CA membrane were 86.6 and 82.6% at 15 and 30 day SRTs, respectively, and 86.6% in the

WSW. COD removals by the MCE membrane were 85.3 and 84.6% at 15 and 30 day SRTs,

respectively, and 86.6% in the WSW. As a result, different membrane materials did not have a

significant effect on COD removal efficiencies under the same operation conditions. COD

removals for bioreactor system were 78.6 and 77.3% at 15 and 30 days SRTs, respectively,

and 80.0% in the WSW. Consequently, COD removal efficiencies for bioreactor system were

lower than that of SMBR. The COD removal from supernatant was primarily due to

biological degradation in the bioreactor while COD removal in the permeates were due to

membrane filtration and biofilm (biofouling layer) on the membrane surfaces. However, it

may be due to blocking of membrane pores with some high molecular organics and become

smaller at the beginning of the filtration, and able to filtrate more organics, in turn improve

the quality of effluent water.

The etodolac removals by the MP005 and MV02 membranes were found same (31.7 and

46% for 15 and 30 day SRTs, respectively).In addition etodolac removals were found close

values (65 and 66.8%) for WSW. The etodolac removals by the CA membrane were 27.0 and

44.7% at 15 and 30 day SRTs, respectively, and 81.0% in the WSW. The etodolac removals

by the MCE membrane were 25.8 and 43.5% at 15 and 30 day SRTs, respectively, and 81.1%

in the WSW. Changes in operating conditions had important effects on the etodolac removal

efficiencies of the CA and MCE membranes. The etodolac removals for the bioreactor system

were 27.0 and 38.8% at SRTs of 15 and 30 days, respectively, and 55.0% in the WSW. Thus,

the membrane usage contributed to the etodolac removal efficiency. This contribution was

relatively lower at 15 and 30 SRTs (5%) while 25% contribution was obtained in case of

WSW. As the SRT decreased the etodolac removal efficiencies decreased appreciably for

both systems. These results indicate that different types of microorganisms may be present

under different SRTs and that more resistant microorganisms may arise in the WSW.

3.2. Effects of SRT on membrane fouling factors

Membrane fouling factors were investigated at different SRTs. To better understand

membrane fouling mechanisms, many different factors were evaluated, including microbial

floc structure (through EPS and SMP analysis), membrane surface fouling (through AFM

analysis), membrane properties (pore sizes, hydrophobicities, pore size distributions, and

membrane materials), solution properties (concentrations, the natures of the components, and

particle size distributions) and operating conditions (SRTs, HRTs, and fluxes). As SRT

increased, sludge concentration increased accordingly. VSS at SRT of 15 days maintained

1.65 g/L and reached 2.26 g/L at SRT of 30 days. VSS in the period WSW was 4.06 g/L.

3.2.1. SMP and EPS secretions at different SRTs

When SRTs were increased both permeate volumes and steady-state fluxes increased. This is

considered to be due to decreased membrane fouling. The total permeate volumes (VT),

initial fluxes (Jo), and pseudo steady-state fluxes (Jss) for each membrane are presented in

Table 4. As SRTs decreased, steady state fluxes decreased. Permeate volumes also decreased

with different SRT values. Among the different MF membranes, MP005 had the highest

steady state flux at each SRT. The permeate volume of MP005 was also higher than those of

the other membranes during the 24-hour period. At an SRT of 15 days, the permeate volumes

for the CA and MCE membranes were lower than those of the other membranes. The most

important cause of this may be changes in EPS and SMP fractions. Therefore, the effects of

SRTs on the concentrations of proteins and carbohydrate fractions of the EPSs were

investigated (Fig.3). EPSp and EPSc indicate protein and carbohydrate fractions of the

extracellular polymeric substances in the bioreactor. With decreasing SRTs, both the protein

and carbohydrate concentrations increased dramatically (86–87% increase for 15 days of SRT

compared to WSW). The amounts and characteristics of the EPSs in the aerobic sludge were

affected by many parameters, including substrate compositions, loading rates, aeration, and,

most importantly, SRTs [32, 33]. Lee et al. [34] found that, as SRTs increased, EPSp

concentrations also increased, while EPSc concentrations stayed stable [8].

The same behavior was observed in SMP fractions in the bioreactor. 92–94% increase in

SMP fractions was found for 15 days of SRT compared to WSW. The effects of SRTs on the

concentrations of SMPp and SMPc are shown in Fig. 4. SMPp and SMPc indicate protein and

carbohydrate fractions of the soluble microbial products in the bioreactor or membrane

permeates. For all three different SRT values, concentrations of protein fractions were

consistently higher than those of carbohydrate fractions in both EPS and SMP. Furthermore,

concentrations of EPSc and EPSp were consistently higher than those of SMPc and SMPp.

Overall, total EPS concentrations were generally one order of magnitude higher than those of

SMP [8].

The protein and carbohydrate fractions of the SMP from the four membranes supernatant

were also measured at the different SRTs (Fig. 5). The concentrations of SMPp were relatively

higher than those of SMPc for each membrane. SMP protein and carbohydrate concentrations

decreased with increasing SRTs.

3.2.2. Analytical measurements

The surface morphologies of clean and fouled membranes were characterized using AFM for

the evaluation of the fouling on the membrane surfaces. AFM images of the different

membrane surfaces before and after fouling at the different SRTs are shown in Fig. 6, and the

roughness parameters of the surfaces obtained from AFM measurements are presented in

Table 5.

Based on the Rrms values for clean membranes in Table 5, MP005 had the smoothest surface,

and the CA membrane had the roughest surface. The Rrms values for the fouled membranes

examined in the WSW were similar to those of the membranes at the other SRTs, and the Rrms

values obtained at SRTs of 15 and 30 days were almost the same.

The fouling of the membrane pores was characterized using SEM. The SEM images of the

membranes with the different SRTs are given in Fig. 7. The SEM images of the clean and

fouled membranes for an SRT of 15 days indicate that the pore structures of the MV02, CA

and MCE membranes changed. However, the pore structure of MP005 changed less than the

others. Changes in the hydrophilic/hydrophobic structures on the MF membranes were

evaluated through contact angle measurements. The contact angle values for clean and fouled

membranes are presented in Table 6.

As shown in Table 6, MP005 was the most hydrophobic membrane, based on contact angles,

whereas the others were highly hydrophilic. MP005 exhibits the lowest change in contact

angles between clean and fouled membranes. With increasing SRTs the contact angle values

increased dramatically for MV02, CA and MCE membranes. It can be said that the surfaces of

the membranes became hydrophobic in consequence of fouling. The contact angle values of

the MV02 and CA membranes were higher than the others at all SRTs.

Membrane fouling, sludge sedimentation properties and biomass size distributions were

evaluated through relative hydrophobicity and particle size distribution analysis. The results

of these analyses for the three different SRTs are given in Table 7. The initial sludge sizes

were the same because the same seed sludge was introduced. It is known that sludge size

decreases with extended runtimes and is affected by the SRT. The sludge sizes in the SMBRs

were 77.3 and 76.8 μm at SRTs of 15 and 30 days, respectively, and 65.3 μm in the WSW.

The sizes of sludge flocs in the SMBRs were related to the bonding forces among the bacteria

cells and the hydraulic shear forces due to aeration. The relative hydrophobicities changed

between 63% and 72% for different SRTs (Table 7). The hydrophobicities of the biomasses

decreased with increasing SRTs. Correspondingly, particle sizes decreased with increasing

SRTs. Zeta potential measurements were carried out to quantify the surface charges of the

materials in the wastewater. The zeta potentials in the SMBRs were -13.7 and -12.1 mV at

SRTs of 15 and 30 days, respectively, and -12.3 mV in the WSW (Table 7). The zeta

potentials at the 30 day SRT and in the WSW were very similar, while the biomass surface

charge at the 15 day SRT was higher.

The chemical bonds at the surfaces of the MF membranes were characterized using FT-IR

spectroscopy. FT-IR images of the clean and fouled membranes are shown in Fig. 8. In the

FT-IR spectrum of the clean MP005 membrane (Fig. 8a), there are some typical

polysaccharide absorption peaks from the membrane material. The band at 3294 cm-1 is the

characteristic band of the –OH group. C=O stretching was observed at 1655 cm-1 on the

original membrane surface. This band decreased after experiments at the different SRTs.

Aromatic bands at 1578 and 1486 cm-1 are characteristic of the PES membrane and decreased

after fouling at an SRT of 15 days. The band at 1535 cm-1 is characteristic of N-H and was

visible after experiments in the period WSW and 30 day SRTs. These bands were not altered.

The clean membrane exhibited bands at 1241 and 1151 cm-1 that were not affected by the

experiments. However, the band at 1105 cm-1 disappeared after fouling at an SRT of 15 days.

A band at 1045 cm-1, indicating the existence of organic matter and carbohydrates, appeared

after fouling. The peaks at 1034 cm-1 represents sulfone group (S=O) stretching in the clean

membranes. This band disappeared when the membranes were fouled at 15 and 30 day SRTs.

The band moved toward 1030 cm-1 after fouling in the WSW. The band at 923 cm-1 present in

the clean membranes was not observed in the fouled membranes. There was a band at 671 cm-

1 present for the membranes fouled at 15 day SRTs, indicating degradation of the polymer

chains [35].

There were typical polysaccharide absorption peaks in the FT-IR spectrum of the clean

MV02 membrane (Fig.8b) at 3301, 1404, 1182, 1074, 1092, 975, 880, 840, 795 and 766 cm-1.

The broad band from 3600 to 2800 cm-1 may be attributed to O-H and C-H stretching, and the

brand from 3200 cm-1 to 3570 cm-1 may be attributed to hydrogen bonds [33, 34]. Three

characteristic bands at 1404, 1182, and 975 cm-1 were attributed to C-H out-of-plane

deformation vibrations. The bands at 1074 and 1092 cm-1 were due to the structure of PVDF,

and were associated with C-F stretching. The bands at 840 and 795 cm-1 are attributed to CH2-

rocking vibrations and the band at 766 cm-1 represents CF2 bending. The band results around

1535 cm-1 showed N-H bending and appeared at each of three different SRTs. The bands at

1629, 1632 and 1629 cm-1 may be assigned to amine groups on the surface of the fouled

membranes and were present for WSW, 30 and 15 day SRTs, respectively. C-F stretching at

1074 and 1092 cm-1 showed slipping in the fouled membranes. The bands at 975, 797 and 762

cm-1 disappeared after experiments at each of the three different SRTs, and the band at

approximately 678 cm-1 in the fouled membrane indicated degradation of the polymer chain.

In the FT-IR spectrum of the clean CA membrane (Fig.8c), bands were observed at 3390,

2931, 1746, 1645, 1434 1370, 1321, 1237, 1160, 1051 and 904 cm-1. A broad band from 3600

to 2800 cm-1 may be attributed to O-H stretching. The band at approximately 1750 cm-1 is

characteristic of ester stretching. The C=O band at 1740 cm-1 and the C-O band at 1228 cm-1

indicate a carboxylate group. The band at 1168 cm-1 indicates a C-O-C bond. Bands at 2334

and 2276 cm-1 appeared in the fouled membranes after experiments at each of the three SRTs.

The C=O band at 1655 cm-1 on the surface of the original membrane showed slipping after

experiments at each of the three SRTs. The bands at 1535 and 1541 cm-1 are assigned to N–H

bending. The band at 1434 cm-1 emerged from an asymmetric metal C-H bend, and the band

at 1370 cm-1 resulted from symmetric metal C-H bending. After the membrane was fouled,

the asymmetric C-H bending band showed slipping for WSW and 30 day SRTs but was not

altered for the 15 day SRT. The three bands at 1237, 1160 and 1051 cm-1 were attributed to C-

O stretching at 1000-1300 cm-1. The bands at 1237 and 1051 cm-1 showed slipping in the

fouled membranes, and the band at 1160 cm-1 utterly disappeared in the fouled membranes.

The band at 904 cm-1 may be attributed to C-H out-of-plane bending.

In the FT-IR spectrum of the clean MCE membrane (Fig.8d), bands were observed at 3386,

1743, 1650, 1429, 1373, 1280, 1159,1062, 843, 751 cm-1. A broad band from 3800 to 3000

cm-1 in the clean MCE membrane was attributed to O-H stretching. The band at

approximately 1750 cm-1 is characteristic of ester stretching and the C=O band at 1743 cm-1

indicates a carboxylate group. The N-H band at 1650 cm-1 on the surface of the original

membrane showed slipping for all of the different SRTs. As the bands at 1429 cm-1 resulted

from asymmetric metal C-H bending, and the band at 1373 cm-1 resulted from symmetric

metal C-H bending. After membrane fouling, the symmetric metal C-H bending disappeared

for each SRT. The three bands at 1280, 1159 and 1062 cm-1 were attributed to C-O stretching

at 1000-1300 cm-1. The band at 843 cm-1 indicated C-H bending and disappeared after fouling

with an SRT of 30 days. The band at 751 cm-1 represented C-H bending and disappeared after

fouling with each of the three different SRTs.

3.3. Resistances in series model results

The resistance in series model was applied using the flux-time graphics at constant pressure

for each of the three different SRTs. Membrane filtration resistance (Rt) in the SMBR was

classified as membrane pore blocking resistance (Rp), gel layer resistance (Rg) and cake layer

resistance (Rc). Rp and Rg were mainly caused by soluble and colloid substances, while Rc

was mainly due to the deposition of flocs [33, 34].

The Rm, Rp and Rc values and their relative percentages of Rt for the different membrane

materials with different pore sizes are presented in Table 8. Each membrane had different

surface properties and morphologies. Therefore, either the type of microbial population

initially attached to the membrane surfaces or the orientation of the attachment or packing of

the flocs to the membranes were likely to be different, causing variations in the porosities of

the cake layer near the surfaces of the membranes [13]. At the different SRT values, the pore

resistances of the MP005 membrane were the lowest. The reductions of the total resistance

values for all of the membranes with increased SRTs were due to the decreases in cake

resistance. Increases in pore resistances were observed. In the MP005 membrane, the

reduction in cake layer resistance constituted a large portion of the total resistance decrease

with decreasing SRTs. For the other membranes, the pore resistances were higher than the

cake layer resistances at the different SRTs. A comparison of pore and cake resistances

reveals that the majority of the fouling occurred due to cake formation on the membrane

surfaces for all of the membranes.

The cake layer resistance (Rc) formed the majority of the total resistance under the WSW.

This situation was thought to be caused by the reduction of the EPS fractions under high

SRTs. A membrane pore resistance (Rp) increase was observed for 15 and 30 day SRTs. For

the CA and MCE membranes under the WSW, Rp did not have a greater effect on the total

resistance and was more effective than Rc under WSW and 15 day SRTs. The membrane pore

resistance of the MP005 membrane was not found to be as effective as the cake layer

resistance.

In addition, the AFM and contact angle results were supported by the Rc values for all

membranes. For example, the highest Rc value was found for the CA and MCE membranes at

an SRT 15 days, as were the highest values for Ra, Rrms and Rz and contact angle. Correlations

between AFM, SEM images and contact angle values had reasonable relationships with the Rp

and Rc values of fouled membranes obtained from the resistance in series model.

4. Conclusions

This study compared removal efficiencies, sludge properties and membrane fouling

mechanisms in a lab-scale SMBR system under constant pressure in the WSW, 30 and 15 day

SRTs. Specific conclusions can be summarized as follows:

� COD removal efficiencies for bioreactor system and SMBR were approximately 80±2%

and 86±2%, respectively at all SRTs. The COD removals for the different membranes

were very similar among the different SRTs.

� The etodolac removal of the bioreactor system was 86.6% in the WSW. The MP005 and

MV02 membranes achieved 86.6% etodolac removal efficiencies in period the WSW. CA

and MCE membranes achieved 87.3% etodolac removals in the WSW. The membrane

materials contributed 5�25% of the etodolac removal. As the SRT decreased from 30 to

15 days, the overall etodolac removal efficiencies decreased much faster than the etodolac

concentrations in the discharge.

� To investigate the flux decline resulting from fouling, experiments were carried out at

different SRTs. As SRTs decreased, the steady state flux values through the membranes

decreased. Permeate volumes decreased with different SRTs. Among the different MF

membranes, MP005 had the highest steady state flux for each of the different SRTs. When

the SRT was increased, both the permeate volumes and the steady-state fluxed increased.

This result is due to decreased membrane fouling. The most important reason for this

could be the change of EPS and SMP fractions with changing SRTs. The concentrations

of protein fractions were consistently higher than those of carbohydrate fractions in both

EPS and SMP under each of the different SRTs. These contents decreased with increasing

SRTs, leading to decreased filtration resistance.

� The SEM, AFM imaging and contact angles demonstrated that the membrane surfaces

were covered with compact gel layers formed by organic substances. The SEM and AFM

measurements of clean and fouled membranes also indicated that the Ra, Rrms and Rz and

contact angle values were supported by the Rg values for different SRTs.

� The fouled membranes were analyzed using a FT-IR spectrometer to obtain information

about organic fouling in the gel layer. In the case of a hydrophobic membrane, alcoholic

compounds, aliphatic amides and polysaccharides were the main contributors to

significant membrane fouling due to adsorption. FT-IR observations lead to the

conclusion that the microfiltration membranes do not allow any further retention of the

investigated substances due to size exclusion.

� Flux decline was also evaluated using the resistance in series model. Rp, Rg and Rc were

calculated according to this model to clarify the effects of different SRT values. The Rt of

the MP005 membrane showed the most significant decrease. The Rp of the MP005

membrane was also lower than those of the other membranes at the different SRTs. In

addition, Rt values decreased for all of the membranes with increasing SRTs because of

the decreases in Rc. As a result, when Rp and Rc were compared it was observed that

major fouling occurred due to cake formation on the membrane surfaces for all of the

membranes. Furthermore, the correlations between AFM, SEM images and contact angle

values had reasonable relationships with the Rp and Rc values of the fouled membranes

that were obtained using the resistances in series model.

Acknowledgment

This work was supported by the Research Fund of the Istanbul University (Project number:

4239).

Nomenclature

AFM atomic force microscope

BSA bovine serum albumin

CA cellulose acetate

CAS conventional treatment system

CASP(s) conventional treatment system plants

CER cation exchange resin

COD chemical oxygen demand (mg L-1)

DO dissolved oxygen (mg L-1)

EPS extracellular polymeric materials

FT-IR fourier transform infrared

HRT hydraulic retention time (h)

J permeate flux (L m-2 h-1)

Jo the initial flux at time t = 0 for studied or fouled

membrane (L m�2 h�1)

Jss pseudo steady-state flux (L m-2 h-1)

MCE nitrocellulose mixed ester

MF microfiltration

mV milivolt

OLR organic loading rate (kg COD m–3 d–1)

PhAc pharmaceutical active compound

PES polyetersulfon

PVDF polyvinlidenflorid

R rejection (%)

Rt resistance of total

Ra the mean roughness on membrane surface (nm)

Rp resistance of por blockage (m�1)

Rc resistance of cake layer (m�1)

Rm resistance of the membrane (m�1)

Rrms the root mean square of average height of membrane surface peaks (nm)

Rz the mean difference between five highest peaks and lowest valleys (nm)

SEM scanning electron microscope

SMBR submerged aerobic membrane bioreactor

SMP soluble microbial product

SRT(s) sludge retention time (s)

t filtration time (min)

TSS total suspended solid

� contact angle (0)

Vt total volume of the permeate (L)

VSS volatile suspended solid

μm micrometre

WSW without sludge wasting

WWTP(s) wastewater treatment plant(s)

Subscripts

a mean roughness

c concentrate

d day

f feed

g gel

in internal

F factor

H hour

m membrane

o initial

p permeate

rms root mean square

References

[1] H. F. Schroeder, Substance-specific detection and pursuit of non-eliminable compounds

during biological treatment of waste water from the pharmaceutical industry, Waste

Manage. 19 (2) (1999) 111–123

[2] I. A. Balcioglu, M. Otker, Treatment of pharmaceutical wastewater containing

antibiotics by O3 and O3/H2O2 processes, Chemosphere, 50(1) (2003) 85–95

[3] S. Baumgarten, H. F. Schröder, C. Charwath, M. Lange, S. Beier, J. Pinnekamp,

Evaluation of advanced treatment technologies for the elimination of pharmaceutical

compounds, Water Sci Technol. 56 (5) (2007) 1–8

[4] H. Tekin, O. Bilkaya, S. S. Ataberk, T. H. Balta, H. Ceribasi, F. D. Sanin, F. B. Dilek,

U. Yetis, Use of Fenton oxidation to improve the biodegradability of a pharmaceutical

wastewater, J Hazard Mater., 21 (2006) 258–265.

[5] T. Urase, T. Kikuta, Separate estimation of adsorption and degradation of

pharmaceutical substances and estrogens in the activated sludge process, Water Res., 39

(7) (2005) 1289–1300

[6] J. Zhang, H.C. Chuaa, J. Zhoub, A.G. Fane, Factors affecting the membrane

performance in submerged membrane bioreactors, J Membr Sci., 284 (2006) 54–66

[7] T. Stephenson, S. Judd, B. Jefferson, K. Brindle, Membrane Bioreactors for Wastewater

Treatment, IWA Publishing, London, UK, (2000)

[8] N. O. Yigit, I. Harman, G. Civelekoglu, H. Koseoglu, N. Cicek, M. Kitis, Membrane

fouling in a pilot-scale submerged membrane bioreactor operated under various

conditions, Desalination 231 (2008) 124–132

[9] P. Le-Clech, V. Chen and T.A.G. Fane, Fouling in membrane bioreactors used in

wastewater treatment, J Membr Sci., 284 (2006) 17–53.

[10] S. Liang, C. Liu and L. Song, Soluble microbial products in membrane bioreactor

operation: behaviors, characteristics, and fouling potential, Water Res., 41 (2007) 95–

101

[11] Z. Wang, S. Tang, Y. Zhu, Z. Wu, Q. Zhou, D. Yang, Fluorescent dissolved organic

matter variations in a submerged membrane bioreactor under different sludge retention

times, J Membr Sci., 355 (2010) 151–157

[12] C. Visvanathan, R.Ben Aim, K. Parmeshwaran, Membrane separation bioreactors for

wastewater treatment, Crit. Rev. Environ. Sci. Technol., 30 (1) (2000) 1–48

[13] A. Pollice, G. Laera, D. Saturno, C. Giordano, Effects of sludge retention time on the

performance of a membrane bioreactor treating municipal sewage, J Membr Sci., 317

(2008) 65–70

[14] J. Radjenovic´, A. Jelic´, M. Petrovic´, D. Barcelo´, Determination of pharmaceuticals

in sewage sludge by pressurized liquid extraction (PLE) coupled to liquid

chromatography-tandem mass spectrometry (LC–MS/MS). Anal Bioanal Chem., (2007)

[15] B. Lesjean, S. Rosenberger, C. Laabs, M. Jekel, R. Gnirss and G. Amy. Correlation

between membrane fouling and soluble/colloidal organic substances in membrane

bioreactors for municipal wastewater treatment. Water Sci. & Tech., 51 (6–7) (2005) 1–

8

[16] A. Göbel, C.S. McArdell, A. Joss, H. Siegrist and W. Giger, Fate of sulfonamides,

macrolides, and trimethoprim in different wastewater treatment technologies. Sci Total

Environ., 372 (2007) 361–371

[17] K. Kimura, H. Hara and Y. Watanabe, Elimination of selected acidic pharmaceuticals

from municipal wastewater by an activated sludge system and membrane bioreactors.

Environ. Sci. Technol., 41 (2007) 3708–3714

[18] S. Pérez and D. Barceló, First evidence for occurrence of hydroxylated human

metabolites of diclofenac and aceclofenac in wastewater using QqLIT-MS and QqTOF-

MS. Anal. Chem., 80 (2008) 8135–8145.

[19] M. Clara, N. Kreuzinger, B. Strenn, O. Gans, H. Kroiss, The solids retention time—a

suitable design parameter to evaluate the capacity of wastewater treatment plants to

remove micropollutants. Water Res., 39 (1) (2005) 97–106.

[20] X. Huang, P. Gui, Y. Qian, Effect of sludge retention time on microbial behaviour in a

submerged membrane bioreactor, Process Biochem, 36 (10) (2001) 1001–1006.

[21] S. S. Han, T. H. Bae, G. G. Jang, T. M. Tak, Influence of sludge retention time on

membrane fouling and bioactivities in membrane bioreactor system, Process Biochem,

40 (7) (2005) 2393–2400.

[22] Y. Kaya, H. Barlas, S. Arayici, Evaluation of fouling mechanisms in the nanofiltration

of solutions with high anionic and nonionic surfactant contents using a resistance-in-

series model, J Membr Sci., 367 (2011) 45–54.

[23] APHA–AWWA–WEF, Standard Methods for Examination of Water and Wastewater,

20th ed., APHA/AWWA/WEF, Washington, DC, 1998.

[24] B. Frolund, R. Palmgren, K. Keiding, P.H. Nielsen. Extraction of extracellular polymers

from activated sludge using a cation exchange resin. (1996). Water Res., 30 (8) 1749–

1758.

[25] O.H. Lowry, N.J. Rosebrough, A.L. Farr, R.J. Randall, Protein measurement with the

folin phenol reagent, J. Membr. Sci., 193 (1951) 265–275.

[26] M. Dubois, K.A. Gilles, J.K. Hamilton, P.A. Rebers, F. Smith, Calorimetric method for

determination of sugar and related substances, Analy. Chem., 28 (1956) 350.

[27] Y. Kaya, H. Barlas, S. Arayici, Nanofiltration of Cleaning-in-Place (CIP) wastewater in

a detergent plant: effects of pH, temperature and transmembrane pressure on flux

behavior, Sep Purif Technol., 65 (2) (2009) 117–129

[28] Y. Kaya, Z.B. Gönder, I. Vergili, H. Barlas, The effect of transmembrane pressure and

pH on treatment of paper machine process waters by using a two-step nanofiltration

process: Flux decline analysis, Desalination, 250 (2010) 150–157

[29] Y. Lee, M.M. Clark, Modeling of flux decline during crossflow ultrafiltration of

colloidal suspensions, J Membr Sci., 149 (1998) 181–202

[30] M. Rahimi, S. S. Madaeni, K. Abbasi, CFD modeling of permeate flux in cross-flow

microfiltration membrane, J Membr Sci., 255 (1/2) (2005) 23–31

[31] N. Dizge, D. Y. Koseoglu-Imer, A. Karagunduz, B. Keskinler. Effects of cationic

polyelectrolyte on filterability and fouling reduction of submerged membrane bioreactor

(MBR), J Membr Sci., 377 (2011) 175– 181

[32] P. Le-Clech, V. Chen and T.A.G. Fane, Fouling in membrane bioreactors used in

wastewater treatment, J Membr Sci., 284 (2006) 17–53

[33] M. E. R. Hernandez, R. Van Kaam, S. Schetrite and C. Albasi, Role and variations of

supernatant compounds in submerged membrane bioreactor fouling, Desalination, 179

(2005) 95–107

[34] W. Lee, S. Kang and H. Shin, Sludge characteristics and their contribution to

microfiltration in submerged membrane bioreactors, J Membr Sci., 216 (2003) 217–227

[35] A.R. Badireddy1 and S. Chellam, Bismuth dimercaptopropanol (BisBAL) inhibits

formation of multispecies wastewater flocs, Journal of Appl Microbiol, 110 (2011)

1426–1437

Fig. 1. Schematic structure of submerged membrane bioreactor (SMBR)

Fig. 2. Removal efficiency of etodolac and COD in the different SRTs and membrane

materials

Fig. 3. Effect of SRTs on EPS fractions in the bioreactor

Fig. 4. Effect of SRTs on SMP fractions in the biorector

Fig. 5. SMP concentrations in the bioreactor and SMBR for different SRTs

Fig. 6. AFM images of clean and fouled membrane surfaces for different SRTs

Fig. 7. SEM images of clean and fouled membrane surfaces for different SRTs

Fig. 8. FT-IR spectra of (a) MP005 (b) MV02 (c) CA (d) MCE membranes

Table 1. Characteristics of MF membranes used in the experiments

a obtanied from our cross-flow membrane system; 0-2 bar, 25 oC, pH 7.5–8

Supplier Membranes Materyal Nominal

pore size

(µm)

Max.

Temperature

(oC)

pH Water

Flux

(L m-2

h-1)

Microdyn-

Nadir

MP005 Polyethersulfone (PES) 0.05 95 0–

14

450a

Microdyn-

Nadir

MV02 Polyvinylidenefluoride

(PVDF)

0.2 95 2–

11

485a

GE

Osmonics

MCE Nitrocellulose Mixed

Ester

0.22 180 3–

11

470a

GE

Osmonics

CA Cellulose Asetate 0.45 135 3–

11

460a

Table 2. Characteristics of the pharmaceutical wastewater

Items Value

pH 7.6

COD (mg L-1) 18000

TKN (mg L-1) 32

N-NH4 (mg L-1) <5

P-PO4 (mg L-1) 2.7

N-NO3 (mg L-1) 37

Etodolac (mg L-1) 511

Table 3. Chemical and physical properties of etodolac

Molecular Structure

Chemical Formula C17H21NO3

Chemical Name 1.8-Diethyl-1.3.4.9-tetrahydropyrano[3.4-b]indole-1-

acetic acid

Appearance White crystalline

Therapeutic class Analgesic, nonsteroidal anti-inflammatory

Chemical group pyranocarboxylic acid

Molecular Weight

(g mol--1)

287.36

Water Solubility

(mg mL-1)

0.0392 (20oC)

Dissociation constant

(pKa)

4.65

Table 4. Values of permeate volume, initial and steady-state flux values in the SMBR

Membrane SRT: 15 day SRT: 30 day WSW

Vt

(mL)

Jo

(L m-2 h-1)

Jss

(L m-2 h-1)

Vt

(mL)

Jo

(L m-2 h-1)

Jss

(L m-2 h-1)

Vt

(mL)

Jo

(L m-2 h-1)

Jss

(L m-2 h-1)

MP005 1450 45 9 2075 120 13 2970 166 30

MV02 1050 17 5 1875 85 10 2850 85 24

CA 850 30 3 1180 67 8 2525 106 19

MCE 745 20 2.9 1050 39 6 2360 115 15

Tab

le 5

. Rou

gnes

s par

amet

ers o

btai

ned

with

AFM

of c

lean

and

foul

ed m

embr

anes

at d

iffer

ent S

RTs

Mem

bran

e C

lean

mem

bran

e Fo

uled

mem

bran

e

SRT

: 15

day

SRT

: 30

day

WSW

Ra

(nm

)

Rrm

s

(nm

)

Rz

(nm

)

Ra

(nm

)

Rrm

s

(nm

)

Rz

(nm

)

Ra

(nm

)

Rrm

s

(nm

)

Rz

(nm

)

Ra

(nm

)

Rrm

s

(nm

)

Rz

(nm

)

MP0

05

18.3

4 24

.00

75.9

8 24

.47

28.1

3 77

.12

25.9

2 31

.03

85.2

3 25

.03

27.3

9 82

.75

MV

02

31.7

4 40

.79

60.2

8 40

.69

59.3

8 89

.27

45.3

7 77

.89

95.2

1 36

.59

50.5

3 65

.43

CA

80

.53

88.3

5 15

6.6

81.6

4 97

.42

106.

7 88

.99

99.6

5 17

3.4

89.3

1 96

.36

172.

1

MC

E 25

.57

29.1

6 64

.14

55.4

8 67

.21

187.

0 54

.91

71.0

2 23

2.4

23.7

4 36

.34

113.

2

Table 6. Contact angle values of clean and fouled membranes

Membrane

Clean

membrane

(� o)

Fouled membrane (� o)

SRT: 15 days SRT: 30 days WSW

MP005 48.5 61 64 56

MV02 <10 69 71 69

CA <10 69 73 67

MCE <10 66 68 50

Table 7. Relative hydrophobicity, particle size distribution and zeta potential analysis

Analysis

SRT

15 days 30 days WSW

Relative hydrophobicity (%) 72 68 63

Zeta potential (mV) -13.7 -12.1 -12.3

Particle size (�m) 77.3 76.8 65.3

37

Table 8. Estimated resistances due to membrane (Rm), pore blocking (Rp), cake formation

(Rc) and total resistance (Rt)

Membrane

SRT: 15 days SRT: 30 days WSW Rt

(x101

2)

Rm(x101

2)(%)

Rp(x101

2)(%)

Rc(x101

2)(%)

Rt(x101

2)

Rm(x101

2)(%)

Rp(x101

2)(%)

Rc(x101

2)(%)

Rt(x101

2)

Rm(x101

2)(%)

Rp(x101

2)(%)

Rc(x101

2)(%)

MP005 7.99 0.16 (2.1)

1.89 (23.6

)

5.94 (74.3

)

5.53 0.16 (2.9)

0.87 (15.7

)

4.50 (81.4

)

2.39 0.16 (6.7)

0.17 (7.2)

2.06 (86.1

) MV02 14.37 0.15

(1.1) 8.32 (57.8

)

5.90 (41.1

)

7.19

0.15 (2.1)

2.51 (34.9

)

4.53 (63.0

)

2.99 0.15 (5.0)

0. 27 (9.1)

2.57 (85.9

)CA 23.95 0.15

(0.6) 14.2 (59.4

)

9.60 (40)

8.98 0.15 (1.7)

4.98 (55.4

)

3.85 (42.9

)

3.78 0.15 (4.0)

0.36 (9.5)

3.27 (86.5

) MCE 24.76 0.16

(0.6) 14.2 (57.3

)

10.4 (42.1

)

11.99 0.16 (1.3)

6.38 (53.2

)

5.45 (45.5

)

4.80 0.16 (3.2)

0.47 (9.8)

4.17 (87.0

)

Fig. 1

Blower

Wastewater

Peristaltic pump

Diaphram pump

Flow meter

Excess sludge

Returned sludge

Vacuum pump Vacuum tank

Membrane modules Sedimentation

tank

Electronic balances Computer

Permeate

Effluent

020406080100

WSW

30

15

WSW

30

15

WSW

30

15

WSW

30

15

WSW

30

15

SRT

(day

)

Removal Efficiency (%)

Etod

olac

CO

D

Bio

reac

tor

M

P005

M

V02

C

AM

CE

Fig.

2

0

10

20

30

40

50

60

70

80

90

100

15 30 WSW

SRT (day)

EPSp EPSc

EPS

frac

tions

in th

e bi

orea

ctor

(mg/

gVSS

)

Fig. 3

0

10

20

30

40

50

60

70

80

90

15 30 WSWSRT (day)

SMPp SMPc

SMP

frac

tions

in th

e b

iore

acto

r (m

g/gV

SS)

Fig. 4

0

20

40

60

80

100

120

140W

SW 30 15

WS

W 30 15

WS

W 30 15

WS

W 30 15

WS

W 30 15

SRT (day)

SM

P fra

ctio

ns o

f sup

erna

tant

(mgL

-1)

SMPpSMPc

Bioreactor MP005 MV02 CA MCE

Fig. 5

Fig.

6

Mem

bran

e C

lean

Mem

bran

e

Foul

ed m

embr

ane

SRT

: 15

days

SR

T: 3

0 da

ys

WSW

M

P005

MV

02

CA

MC

E

Fig.

7

Mem

bran

e C

lean

mem

bran

e

Foul

ed m

embr

ane

SRT

: 15

days

SR

T: 3

0 da

ys

WSW

M

P005

MV

02

CA

MC

E

(a

) MP0

05 m

embr

ane

(b

) MV

02 m

embr

ane

(c

) CA

mem

bran

e

(d) M

CE

mem

bran

e

Fig.

8