Desalination 231 (2008) 182–190

0011-9164/08/$– See front matter © 2008 Elsevier B.V. All rights reserved

Presented at The 4th IWA Conference on Membranes for Water and Wastewater TreatmentMay 15–17, 2007, Harrogate, UK

*Corresponding author.

Impact of aeration rates on particle colloidal fractionin the biofilm membrane bioreactor (BF-MBR)

Igor Ivanovic, TorOve Leiknes*

NTNU - Norwegian University of Science and Technology, Department of Hydraulic and Environmental Engineering,S.P. Andersensvei 5, N-7491 Trondheim, Norway

Tel. +47 7359 4758; Fax +47 7359 1298; email: torove.leiknes@ntnu.no

Received 17 May 2007; accepted revised 29 November 2007

Abstract

In BF-MBR, as in all sMBR systems, sufficient aeration is necessary to maintain local cross-flow conditions forsustainable membrane operation. Aeration rate is normally based on previous experiences and manufactures’recommendations. BF-MBR with an external submerged membrane unit, designed only for particle separation,could be relatively easily optimized in terms of membrane aeration. Combining observed fouling rates with measuredparticle size distribution and minimizing both parameters indicates a desirable sustainable aeration rate for givenwastewater characteristic, operating parameters (i.e. production/backwash flux, recovery) and membrane modulegeometry.

Keywords: Biofilm-MBR; Colloidal fraction; Fouling rate; Particle size distribution (PSD); Specific aeration demand– membrane area (SADm)

1. Introduction

Biofilm membrane bioreactor (BF-MBR) is ahybrid system for municipal wastewater treatment,where the moving-bed-biofilm reactor (MBBR)for biodegradation of soluble organic matter is

coupled with a submerged membrane reactor(sMBR). BF-MBR has the potential to utilize thebest characteristics of a biofilm process and mem-brane separation resulting in compact, efficientparticle removal schemes. As in all membrane pro-cesses, fouling is the main challenge that needsto be addressed. The efficiency of the process is

doi:10.1016/j.desal.2007.11.046

I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190 183

constrained by the accumulation of materials onthe surface of, or within the membrane resultingin a reduction in membrane permeability/produc-tion and the overall performance of the treatmentprocess.

Aeration in MBR systems (both for biologicaland membrane operational purposes) is the mostcost demanding factor in terms of energy con-sumption. Air scouring is necessary in submergedmembrane systems in order to generate localizedcross-flow conditions along the membrane sur-face thereby reducing cake deposit on the mem-brane. However, the relationship between aera-tion and flux or transmembrane pressure declineis still not fully understood. Aeration rates in MBRsystems are based on previous experience andnormally recommended by membrane suppliers[1].

Studies investigating membrane fouling inMBR processes have reported the significance ofcolloidal particles as an important factor contrib-uting to fouling development. Wisinewski andGrasmick [2] evaluated major components con-tributing to membrane fouling and estimated thatcolloids are responsible for 25% of total measuredfouling. Defrance et al. [3] estimated this value at30%, while Bouhabila et al. [4] reported 50% intheir results. Other studies have also shown theimportance of colloidal particles in membranefouling (i.e. pore blocking) for ultrafiltration mem-branes [5,6]. However, references in the litera-ture that report correlations between aeration ratesand effects on changes in colloidal particle char-acteristics as a consequence of membrane aera-tion can not be found. Reports have been pub-lished that show how more intensive aeration candamage floc structures and release foulants (i.e.EPS) bound in the floc structure [1,7,8].

Biofilm reactors produce an effluent with dif-ferent particle characteristics compared to acti-vated sludge, i.e. floc structure, particle size dis-tribution, etc. Previous studies have shown thataeration plays an important roll in particle brakageand promotes the formation of more colloidal

particles, an important component in membranefouling [5,9,10]. Also, it was observed in previ-ous work that aeration rates could be a functionof the configuration and geometry of the mem-brane reactor [11,12], though a clear connectionbetween these two factors has not yet been fullyinvestigated.

The aim of this study is to investigate the ef-fect of aeration rates in the biofilm-MBR processand how this affects the colloidal fraction in themembrane filtration unit and implications this mayhave on membrane fouling potential.

2. Methods

2.1. Experimental setup

A schematic of the process configuration andpilot plant study is shown in Fig. 1. The BF-MBRprocess investigated in this study consists of themoving-bed-biofilm reactor (MBBR) followed byan external submerged membrane filtration reac-tor. This setup is based on the concept in whichthe biofilm reactor is responsible for the biologi-cal treatment of municipal wastewater, while themembrane reactor is used as an enhanced particleseparation unit with relatively short HRT. Thisconfiguration was chosen as it has the potentialfor better fouling control due to the fact that al-most no biological activity takes place in the mem-brane reactor unit and that the membrane reactordesign can be focused on efficient particle re-moval.

Four moving-bed-biofilm (MBBR) reactorswere installed in series as shown in Fig. 1. Thevolume of each reactor was 65 L and each reactorwas filled with biofilm carriers type K1, suppliedby AnoxKaldnes. Filling fraction was 67% of re-actor volume giving a specific growth area forthe biofilm of 335 m2.m–3

reactor volume. The hydraulicretention time (HRT) in the complete biofilm re-actor was 4 h. During the whole experimentalperiod full nitrification was achieved. Real mu-nicipal wastewater was used during the experi-ments with pre-treatment consisting of a primary

184 I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190

Fig. 1. Flow diagram of the biofilm-MBR process configuration.

clarifier. The characteristics of the raw wastewa-ter and effluent quality parameters (average, maxi-mum and minimum) as well as the organic load-ing rates of the MBBR and are summarized inTable 1.

The membrane reactor was designed as an ex-ternal submerged membrane reactor with a vol-ume of 33 L. A ZW-10 pilot plant membranemodule supplied by Zenon Environmental Inc.was applied in this study. The membrane unit was

Table 1Loading rates and characteristic of raw wastewater and effluent from MBBR

Parameter Avg. Max Min SS, mg/L 114.45 286.11 45.79 COD, mg O2/L 274.5 382 184 FCOD – 1.2 µm 142.2 180 80.2 NH4-N, mg/L 22.4 48.1 9.13 Loading rate COD, g O2/m2?d 4.53 6.32 3.04

Raw wastewater

Loading rate FCOD, g O2/m2?d 2.35 2.98 1.32 SS, mg/L 156.75 438.33 62.22 COD, mg O2/L 204 224 140.1 FCOD – 1.2 µm 30.5 45.2 27.2

Effluent MBBR

NH4-N, mg/L <1

operated under a constant flux of 52 L.m–2.h–1 andit was periodically backwashed with 5.8% of thepermeate production. A constant backwash fluxof 58 L.m–2.h–1 was applied. During the operatingperiod wastewater temperature was between 8.2and 15.7oC. HRT of the membrane reactor wasless than 45 min. From the bottom of the mem-brane reactor retentate was extracted continuouslywith a flow rate ~2 L/h. Recovery in the mem-brane reactor was set to 95.6%. The filtration was

I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190 185

operated in a cyclic mode (5 min) consisting of4.75 min production followed by 0.25 min back-wash. Continuous coarse bubble aeration was ap-plied to promote local cross-flow conditions alongthe membrane surface. The aeration system is in-tegrated into the membrane module at the bottomof the membrane bundle, designed as four smallholes with diameter of 2 mm. Membrane modulespecifications and operating modes are summa-rized in Table 2.

2.2. Analysis

Particle size distribution (PSD) analysis of thewastewater was done by using laser diffractionspectroscopy (Beckman Coulter LS230). The de-velopment of transmembrane pressure (TMP) wasmeasured continuously using an online pressuretransducer connected to a data acquisition systemfrom National Instruments, FieldPoint (FP1000and FP-AI-110), with the LabVIEW data acqui-sition and analysis software. TMP and tempera-ture were logged for every two seconds. Data se-ries where then extracted from the raw data de-pending on analysis, e.g. for overall performancethe values from the beginning of a productioncycle were extracted and plotted over time. Sus-pended solids (SS) were analyzed by filteringthrough a Whatman GF/C 1.2 µm according tothe NS 4733. Chemical Oxygen Demand (COD)and ammonia (NH4-N) were measured with theDr Lange LCK 114, 314, 302 and 303 cuvette

Table 2Membrane module specifications and operating mode

Membrane reactor Operational characteristics Zenon ZW10 membrane pilot module

• Type of membrane – hollow fiber • Nominal pore size – 0.04 µm • Configuration – outside/in • Volume of membrane reactor – 33 L

Flux – 52 L.m–2.h–1, Backwash flux – 58 L.m–2.h–1 Total permeate production – 43.2 L/h Concentrate flow - 2 L/h Operating cycle: 4.75 min production 0.25 min backwashing Recovery = 95.6% Aeration rates: 0.84, 1.68, 3.37, 5.05, 6.74 Nm3m–2h–1

tests. For the filtered chemical oxygen demand(FCOD) samples were first filtered with WhatmanGF/C glass microfiber filters (1.2 µm). Sludge vol-ume index (SVI), capillary suction time (CST)and time-to-filtrate (TTF) were preformed accord-ing to the Standard Methods [13] in order to evalu-ate settling, dewatering and filtering characteris-tics of the retentate. The performance of the mem-brane filtration unit in the BF-MBR has been in-vestigated with five different aeration rates; 0.84,1.68, 3.37, 5.05 and 6.74 Nm3/m2h expressed asspecific aeration demand – membrane area(SADm) according to Judd [1]. The aeration is ex-pressed as SADm, however, as the pilot plant isbased on a small-scale unit the values of the SADmmeasured do not represent values one can expectfrom full-scale systems and the trends and rela-tionships of the values are therefore of more es-sence in this study.

3. Results and discusion

Different aeration rates did not affected thecharacteristics of permeate due to the fact thatstable permeate quality was observed during thewhole experimental period. Furthermore it wasnot possible to observe any effect of different aera-tion rates on other parameters of treated waste-water e.g., SS/VSS, CST, TTF, SVI etc. The mea-sured quality parameters for the retentate and per-meate from the membrane reactor are summarizedin Table 3.

186 I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190

Membrane performance, however, expressedas TMP development over time or permeabilitydecline (i.e. membrane fouling) showed a clearresponse when different aeration rates were ap-plied. Fig. 2 illustrates typical measurements ob-served for the TMP development within an oper-ating cycle of 5 min. A higher fouling rate for thelower aeration rates can be observed, however,the variation for the three higher aeration ratesapplied are not that significantly different. Simi-lar findings were reported by Le-Clech et al. [14].Permeability decline within one cycle on averagewas 116.2 LMH@20oC/bar for SADm = 0.84Nm3.m–2.h–1 and 75.0 LMH@20oC/bar for SADm= 6.74 Nm3.m–2.h–1, showing that higher aerationrates give less membrane fouling. For the mem-brane configuration used in this study and for thegiven flux of operation, the effect of aeration rateshigher than 3.37 Nm3.m–2.h–1 do not show signifi-cant reduction of membrane fouling (i.e. perme-ability decline) (Fig. 2).

The overall effect of varying aeration rates inthe membrane rector was investigated by operat-ing the pilot plant for five days where one aera-tion rate was applied per day; starting with thehighest rate. Typical results measured are illus-trated in Fig. 3. The overall permeability decline(i.e. fouling rate) for the first three aeration ratesis essentially the same. Based on this analysis,

Table 3Characteristic of retentate and permeate from membrane reactor

Parameter Avg. Max Min SS, mg/L 1211.39 2583.74 568.79 VSS, mg/L 289.07 636.40 127.67 COD, mgO2/L 1209.11 2057.1 511.2 CST, s 43.18 166.70 15.60 TTF, s 56 180 44

Retentate

SVI, ml/g 124.53 215.03 47.00 SS, mg/L 0 FCOD, mgO2/L 22.6 30.5 17.2 NH4-N, mgO2/L <1

Permeate

Turbidity, NTU <0.1

the performance of the membrane filtration is notsignificantly affected for SADm values higher then3.37 Nm3.m–2.h–1 (days 1–3). When the aerationrate is less than SADm values of 3.37 Nm3.m–2.h–1

the fouling rate increases significantly (days 4–5on Fig. 3). The results confirm the importance ofaeration as a means to mitigate fouling in im-mersed membrane systems, however, there ap-pears to be a practical limit above which the ef-fect of increasing aeration has a minor added ben-efit. These results are in agreement with findingsfrom other studies; Le-Clech et al. [14], Liu et al.[15], Psoch and Schiewer [16] and Ueda et al. [17].The limit found in this study is an artefact of themembrane reactor and module design and thussystem dependent. However, aeration rates mayalso affect the characteristics of the water in thereactor.

As fouling mitigation from aeration is com-monly attributed to the shear forces generated bythe air bubbles and the effect this has on the mem-branes, this will ultimately also affect the proper-ties of the particulate matter in the membrane re-actor. Too high shear forces will have a detrimen-tal affect on the suspended solids by breaking themup into smaller fragments. Given that membranefouling correlates to the submicron particles inthe reactor, excessive aeration may also have anegative effect on fouling mitigation. To investi-

I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190 187

100

120

140

160

180

200

220

240

260

280

15 30 45 60 75 90 105 120 135 150 165 180 195 210 225 240 255 270 285 300

Time [second]

Per

mea

bilit

y [L

MH

@20

o C/b

ar]

0,84 Nm3/(m2h)1,68 Nm3/(m2h)3,37 Nm3/(m2h)5,05 Nm3/(m2h)6,74 Nm3/(m2h)

Fig. 2. Permeability decline measured within a production cycle.

0

50

100

150

200

250

300

0 1 2 3 4 5

Time [days]

Perm

eabi

lity

[LM

H@

20o C

/bar

]

6,74 Nm3/(m2h) 5,05 Nm3/(m2h) 3,37 Nm3/(m2h) 1,68 Nm3/(m2h) 0,84 Nm3/(m2h)

Fig. 3. Overall permeability decline for different aeration rates.

188 I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190

gate such an affect on the system, analysis of thesuspended solids around the membrane was there-fore undertaken. Particle size distribution (PSD)analyses for the different operating conditions areillustrated in Fig. 4. PSD results show a clear in-crease of the differential number percentage ofparticles in the colloidal fraction and in particularthe fraction of particles below 0.1 µm with in-creasing aeration rates. The same trend is observedfor all samples analyzed where representative re-sults are shown in Fig. 4. In the example shown,the differential number percentage shows that thelargest fraction of particles present around themembranes has a diameter of approximately0.07 µm. This value increased from 8.49% to9.72% with increasing aeration rates showing aclear trend towards a larger number of smaller par-ticles. The size of the particles also decreased withthe largest number of particles having a diameteraround 0.064 µm (Fig. 4). This trend towardssmaller colloids and slightly smaller diameters isa function of the increased shear forces induced

0

1

2

3

4

5

6

7

8

9

10

0,01 0,1 1

Particle size [µm]

Part

icle

dis

trib

utio

n [n

umbe

r %] 6,74 Nm3/(m2h)

5,05 Nm3/(m2h)3,37 Nm3/(m2h)1,68 Nm3/(m2h)0,84 Nm3/(m2h)

Fig. 4. PSD as a function of aeration rates — zoom at submicron area below 0.1 µm.

during the increase in aeration intensity. Conse-quently, aeration in the membrane reactor has twocontradicting effects with regards to membranefouling mitigation. Higher aeration rates more ef-ficiently remove the fouling deposition on themembrane surface and simultaneously increasesthe fraction of components that have been identi-fied as major contributors to fouling. Optimalconditions for sustainable membrane operationmay therefore be when these two effects are bal-anced, thereby giving an upper limit for aerationrates for a specific mode of operation.

One approach to defining the optimal mode ofoperation with respect to aeration rates is to evalu-ate the relationship between membrane foulingrates and the effects of aeration on suspended sol-ids characteristics. Results from PSD analysescombined with measured fouling rates for differ-ent SADm values could potentially indicate thedesirable range for recommended aeration ratesfor the most efficient mode of operation for a sys-tem. The recommended SADm would ultimately

I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190 189

define the condition that provides the least en-ergy consumption for aeration while maintainingminimum fouling rates. Based on the results fromthis study, overall fouling rates measured as afunction of SADm have been plotted together withPSD analysis results for the same conditions(Fig. 5). From the presented results in Fig. 5 adesirable range for SADm should be estimated forvalues higher then 1.68 but lower than3.37 Nm3.m–2.h–1, for the given membrane reactordesign and chosen operating conditions.

It is important to note that these values are spe-cific for the system and operating conditions ap-plied in this study. However, the approach repre-sents and alternative strategy to define optimaloperating conditions for a system with respect toaeration rates. In combination with the flux-step-ping approach to define the optimal flux range ofoperation, analysis of PSD effects and foulingrates as a function of SADm may be applied tooptimize membrane aeration rates. Studies are cur-

Fig. 5. Fouling rate and PSD values of the most present submicron fraction for different aeration rates (SADm).

0 1 2 3 4 5 6 7 80,00

0,02

0,04

0,06

0,08

0,10

0,12

0,14

0,16

0,18

0,20

SADm [Nm3/m2h]

Foul

ing

rate

- dT

MP/

dt [b

ar/h

]

8

9

10

PSD

[num

ber %

]

rently being conducted to generalize this approachand validate this analysis as a means to definerecommended operating conditions with respectto aeration rates for a given system.

4. Conclusions

The aim of this study was to define a proce-dure that can determine sustainable aeration ratesfor a BF-MBR process with a given membranereactor design, sustainable flux and wastewatercharacteristics. The results show the importanceof finding the relationship between sufficient aera-tion to minimize membrane fouling, while pre-venting formation of colloidal particles due toexcessive share forces caused by the aeration. Anapproach to define optimal operating conditionswith respect to aeration rates is proposed. Byevaluating overall fouling rates and PSD analysisbased on an assessment of the differential num-ber percent of submicron particle, recommenda-

190 I. Ivanovic, T. Leiknes / Desalination 231 (2008) 182–190

tions for a design aeration rate may be defined.Further studies are currently being conducted todevelop this approach and validate a general formand applicability as a design tool for optimal andsustainable operation of the membrane filtrationunit for a BF-MBR process.

Acknowledgment

AnoxKaldnes, Norway, for support with thebiofilm reactor and ZENON Environmental Inc.,Canada, for supplying the membrane modules.

References[1] S. Judd, The MBR Book – Principles and Applica-

tions of Membrane Bioreactors in Water And Waste-water Treatment, Elsevier, 2006.

[2] G. Wisniewski and A. Grasmick, Floc size distribu-tion in a membrane bioreactor and consequences formembrane fouling, Colloids Surfaces A:Physicochem. Eng. Aspects, 138 (1996) 403–411.

[3] L. Defrance, M.Y. Jaffrin, B. Gupta, P. Paullier andV. Geaugey, Contribution of various constituents ofactivated sludge to membrane bioreactor fouling,Bioresource Technol., 73 (2000) 105–112.

[4] E.H. Bouhabila, R. Ben Aïm and H. Buisson, Foul-ing characterisation in membrane bioreactors, Separ.Purif. Technol., 22 (2001) 123–132.

[5] R.M. Åhl, T. Leiknes and H. Ødegaard, Trackingparticle size distributions in a moving bed biofilmmembrane reactor for treatment of municipal waste-water. Water Sci. Technol., 53(7) (2006) 33–42.

[6] L.C. Juang, D.H. Tseng and H.Y. Lin, Membraneprocesses for water reuse from the effluent of in-dustrial park wastewater treatment plant: a study onflux and fouling of membrane, Desalination, 202(2007) 302–309.

[7] L. Ji and J. Zhou, Influence of aeration on micro-bial polymers and membrane fouling in submerged

membrane bioreactors, J. Membr. Sci., 276 (2006)168–177.

[8] J.-S. Park, K.-M. Yeon and C.-H. Lee, Hydrody-namics and microbial physiology affecting perfor-mance of a new MBR, membrane-coupled high-per-formance compact reactor. Desalination, 172 (2005)181–188.

[9] T. Leiknes and H. Ødegaard, The development of abiofilm membrane bioreactor Desalination, 202(2007) 135–143.

[10] E. Melin, T. Leiknes, H. Helness, V. Rasmussen andH. Ødegaard, Effect of organic loading rate on awastewater treatment process combining movingbed biofilm and membrane reactors. Water Sci.Technol., 51(6–7) (2005) 421–430.

[11] A. Bick, F. Yang, S. Shandalov and G. Oron, Dataenvelopment analysis for assessing optimal opera-tion of an immersed membrane bioreactor equippedwith a draft tube for domestic wastewater reclama-tion, Desalination, 204 (2007) 17–23.

[12] T. Leiknes, H. Bolt, M. Engmann and H. ØdegaardAssessment of membrane reactor design in the per-formance of a hybrid biofilm membrane bioreactor(BF-MBR). Desalination, 199 (2006) 328–330.

[13] Standard Methods for the Examination of Water andWastewater, 20th ed., American Public Health As-sociation/American Water Works Association/Wa-ter Environment Federation, Washington DC, USA.

[14] P. Le-Clech, B. Jefferson and S.J. Judd, Impact ofaerat(1998). ion, solids concentration and membranecharacteristics on the hydraulic performance of amembrane bioreactor, J. Membr. Sci., 218 (2003)117–129.

[15] R. Liu, X. Huang, Y.F. Sun and Y. Qian, Hydrody-namic effect on sludge accumulation over membranesurfaces in a submerged membrane bioreactor. Pro-cess Biochem., 39(2) (2003) 157–163.

[16] C. Psoch and S. Schiewer, Anti-fouling applicationof air sparging and backflushing for MBR. J. Membr.Sci., 283 (2006) 273–280.

[17] T. Ueda, K. Hata, Y. Kikuoka and O. Senio, Effectsof aeration on suction pressure in a submerged mem-brane bioreactor. Water Res., 31(3) (1997) 489–494.