Availableonlineatwww.sciencedirect.comScienceDirectDESALINATIONELSEVIERDesalination202(2007)253-261www.elsevier.com/locate/desalHybrid low-pressure submerged membrane photoreactorforthe removal of bisphenol ASze Sze Chin, Tuti Mariana Limab, Ken Chiang, Anthony Gordon Fanea*aSchool ofCivil and Environmental Engineering,NanyangTechnological University,Block Nl,NanyangAvenue,Singapore639798,SingaporeTel.+65-6794-3801;Fax+65-6792-1291:email:a.fane@unsw.edu.aubInstitute of Environmental Science and Engineering,Innovation Centre,Block 2,Unit 237,18, Nanyang Drive,Singapore637723,Singapore°ARCCentreforFunctional Nanomaterials,School of Chemical Engineeringand Industrial Chemistry,UniversityofNewSouthWales,Sydney,NSW2052,AustraliaReceived31 July 2005; accepted23December 2005AbstractTheefficiencyof ahybrid systemcombiningphotocatalysisand membranefiltration ina singlemodulewasinvestigated. Low-pressure submerged hollow fibre membranes were used to retain the TiO,particles in thesystem.BisphenolA(BPA)wasused as amodelpollutant.Ninety-sevenper centphotodegradation and morethan90%photomineralization of 10 ppmof BPA were achievedafter90 and120min of UV illumination,respectively.Aeration wasapplied inthesubmergedmembranephotoreactor(SMPR)toprovidemixing,dissolved oxygen,mechanicalagitationtopreventagglomeration of TiO,particles as well asprovidingshearforcesto removeTiOparticles from themembrane surface.Theoptimumaerationrate used in our 0.8-Lreactor was 0.5L/min.It wasfound that intermittentpermeation enhanced the sustainability of the submerged membranes but showed no effecton thephotoactivityof the system.An intermittencefrequency(IF)of 0.1wassufficienttoreducethefoulingrateof the membrane under the experimental conditions. The SMPR appears to be very effective and can achieveremoval of low-concentrationorganics (such asBPA)in acompact, low-energy system.Keywords:Aeration; Low-pressure submerged membrane;Photocatalysis; Intermittent permeation;Bisphenol A1.Introductionof the most promising methods for the treatment oforganic pollutants at low concentrations [1]. TiO,Heterogeneous photocatalytic oxidation (PCO)in the form of anatase and/or rutile crystalsbasedmainlyontheuseofTiO,hasbecomeonebehaves as a classic photocatalyst. When incidentphotons of a wavelength less than 385 nm are*Corresponding author.Presented at the conference on Wastewater Reclamation and Reuse for Sustainability (WWRS2005),November&-ll,2005,Jeju,Korea.OrganizedbytheInternationalWaterAssociation(IWA)andtheGwangjuInstituteofScience and Technology (GIST).0011-9164/06/$-See front matter 2006Published byElsevierB.Vdoi:10.1016/j.desal.2005.12.062
Presented at the conference on Wastewater Reclamation and Reuse for Sustainability (WWRS2005), November 8–11, 2005, Jeju, Korea. Organized by the International Water Association (IWA) and the Gwangju Institute of Science and Technology (GIST). Hybrid low-pressure submerged membrane photoreactor for the removal of bisphenol A Sze Sze China , Tuti Mariana Lima,b, Ken Chiangc , Anthony Gordon Fanea * a School of Civil and Environmental Engineering, Nanyang Technological University, Block N1, Nanyang Avenue, Singapore 639 798, Singapore Tel. +65-6794-3801; Fax +65-6792-1291; email: a.fane@unsw.edu.au b Institute of Environmental Science and Engineering, Innovation Centre, Block 2, Unit 237, 18, Nanyang Drive, Singapore 637 723, Singapore c ARC Centre for Functional Nanomaterials, School of Chemical Engineering and Industrial Chemistry, University of New South Wales, Sydney, NSW 2052, Australia Received 31 July 2005; accepted 23 December 2005 Abstract The efficiency of a hybrid system combining photocatalysis and membrane filtration in a single module was investigated. Low-pressure submerged hollow fibre membranes were used to retain the TiO2 particles in the system. Bisphenol A (BPA) was used as a model pollutant. Ninety-seven per cent photodegradation and more than 90% photomineralization of 10 ppm of BPA were achieved after 90 and 120 min of UV illumination, respectively. Aeration was applied in the submerged membrane photoreactor (SMPR) to provide mixing, dissolved oxygen, mechanical agitation to prevent agglomeration of TiO2 particles as well as providing shear forces to remove TiO2 particles from the membrane surface. The optimum aeration rate used in our 0.8-L reactor was 0.5 L/min. It was found that intermittent permeation enhanced the sustainability of the submerged membranes but showed no effect on the photoactivity of the system. An intermittence frequency (IF) of 0.1 was sufficient to reduce the fouling rate of the membrane under the experimental conditions. The SMPR appears to be very effective and can achieve removal of low-concentration organics (such as BPA) in a compact, low-energy system. Keywords: Aeration; Low-pressure submerged membrane; Photocatalysis; Intermittent permeation; Bisphenol A 1. Introduction Heterogeneous photocatalytic oxidation (PCO), based mainly on the use of TiO2, has become one of the most promising methods for the treatment of organic pollutants at low concentrations [1]. TiO2 in the form of anatase and/or rutile crystals behaves as a classic photocatalyst. When incident photons of a wavelength less than 385 nm are *Corresponding author. Desalination 202 (2007) 253–261 0011-9164/06/$– See front matter © 2006 Published by Elsevier B.V. doi:10.1016/j.desal.2005.12.062
254S.S. Chin et al. / Desalination 202 (2007) 253-261absorbedbyTiOz,electrons are promotedfrom theTheirreactorconsistedoftwocompartments-valence band to the conductance band.ThisMF and a PCO zone.An‘in-house'nanoparticleexcitation leaves positive holes in the valence bandof TiO, was used which enhanced the separationand electrons in the conduction band. The holesof TiO, and maintained high flux of the mem-can directly oxidize the organic compounds thatbrane due to its larger particle size compared toadsorbed onto the catalyst surface and/or reactDegussa P25. It was reported that a removal of73%of total organic carbon(TOC)was achievedwith the hydroxylated surface and adsorbed watermolecules to form hydroxyl radicals (OH), whichwithin2hofUV irradiation.can then oxidize and mineralize toxic organicThe objective of our study is to investigatethe efficiency of a low-pressure SMPR in thecompounds.Simultaneously,thefree electrons canreact with oxygen to produce superoxide radicalsdegradation of an endocrine disrupting chemicalusing commercial TiOz, Degussa P25. Theor reduce other species that are present ontheSMPR consisted of a low-pressure submergedcatalyst surface.However,these charge carriershollow fibre membrane module which was inmayalsorecombine and thusreduce the overalldirect contact with the PCO medium. Carefulefficiencyofthephotocatalyticprocess.Although TiO, is known to be a goodselection ofmembranematerialallowed integra-tion of the membrane and the reactor. Bisphenolphotocatalyst interms of itsactivity,physical andchemical stability, the recovery of the sub-A (BPA) was selected as the model pollutant.micron-sized photocatalysts is one of the keyThepresenceofBPAinlowconcentrationcanresult in hormonal imbalance, male infertilitychallenges forlarge-scaleapplication.Membranetechnology can address this issue by enhancingand breast cancers[6].Thus,effectiveremedia-the separation of particulateTiO,from the treatedtion technologies to destroy BPA have to bewater. In this case, a higher throughput of treateddeveloped. This study also investigated thewatermaybeachieved.Themembrane optionseffect of operating mode, namely continuousinclude high-pressure nanofiltration (NF) as usedand intermittent,ontheefficiencyoftheSMPR.by Molinari et al. [2], or low-pressure membranessuch asmicro-and ultrafiltration (MF andUF)The low-pressure MF and UF have an energy2.Experimentalbenefitbutarenotabletoretainlowmolecular2.1.Materialsweight species.Submerged membranes haveThe TiO, (P25)usedfor the experiments wasbeen widely used in MF and UF processes, suchsupplied byDegussa AG, Germany.High purityasmembranebioreactors(MBR)duetotheBPA (Merck-Schucharatt)was used and thepHrecognized advantages of lower cost of fabrica-of all suspensions was adjusted using sodiumtion and maintenance [3,4]. The key featureshydroxide and perchloric acid.Purified air wasofthe submerged membrane module includeused as a source of oxygen in this study. Unlessairbubbling as the main mechanical methodotherwise stated, all solutions were preparedtoprovidemembranecleaningactionand suctionusingMilliporeMilliQwaterwitharesistivityofto withdraw the permeate from the system to18.2MQcm.A U-shapedhollow fibre mem-prevent overpressure of the bioreactor.The non-brane module of area, 4.107 × 10- m*, was used.pressurized open system also reduces the overallThe MF membranes made of poly-vinylideneoperating cost of the submerged membranefluoride (PVDF) were supplied by Bluestarsystem. Fu et al. [5] have described a submergedChina, with a pore size of 0.22 μm and outermembranephotoreactor(SMPR)for thedegrada-diameter of 1mm.The choice of PVDFfor thetion of natural organicmatter (NOM), fulvic acid
254 absorbed by TiO2, electrons are promoted from the valence band to the conductance band. This excitation leaves positive holes in the valence band and electrons in the conduction band. The holes can directly oxidize the organic compounds that adsorbed onto the catalyst surface and/or react with the hydroxylated surface and adsorbed water molecules to form hydroxyl radicals (OH· ), which can then oxidize and mineralize toxic organic compounds. Simultaneously, the free electrons can react with oxygen to produce superoxide radicals or reduce other species that are present on the catalyst surface. However, these charge carriers may also recombine and thus reduce the overall efficiency of the photocatalytic process. Although TiO2 is known to be a good photocatalyst in terms of its activity, physical and chemical stability, the recovery of the submicron-sized photocatalysts is one of the key challenges for large-scale application. Membrane technology can address this issue by enhancing the separation of particulate TiO2 from the treated water. In this case, a higher throughput of treated water may be achieved. The membrane options include high-pressure nanofiltration (NF) as used by Molinari et al. [2], or low-pressure membranes such as micro- and ultrafiltration (MF and UF). The low-pressure MF and UF have an energy benefit but are not able to retain low molecular weight species. Submerged membranes have been widely used in MF and UF processes, such as membrane bioreactors (MBR) due to the recognized advantages of lower cost of fabrication and maintenance [3,4]. The key features of the submerged membrane module include air bubbling as the main mechanical method to provide membrane cleaning action and suction to withdraw the permeate from the system to prevent overpressure of the bioreactor. The nonpressurized open system also reduces the overall operating cost of the submerged membrane system. Fu et al. [5] have described a submerged membrane photoreactor (SMPR) for the degradation of natural organic matter (NOM), fulvic acid. Their reactor consisted of two compartments — a MF and a PCO zone. An ‘in-house’ nanoparticle of TiO2 was used which enhanced the separation of TiO2 and maintained high flux of the membrane due to its larger particle size compared to Degussa P25. It was reported that a removal of 73% of total organic carbon (TOC) was achieved within 2 h of UV irradiation. The objective of our study is to investigate the efficiency of a low-pressure SMPR in the degradation of an endocrine disrupting chemical using commercial TiO2, Degussa P25. The SMPR consisted of a low-pressure submerged hollow fibre membrane module which was in direct contact with the PCO medium. Careful selection of membrane material allowed integration of the membrane and the reactor. Bisphenol A (BPA) was selected as the model pollutant. The presence of BPA in low concentration can result in hormonal imbalance, male infertility and breast cancers [6]. Thus, effective remediation technologies to destroy BPA have to be developed. This study also investigated the effect of operating mode, namely continuous and intermittent, on the efficiency of the SMPR. 2. Experimental 2.1. Materials The TiO2 (P25) used for the experiments was supplied by Degussa AG, Germany. High purity BPA (Merck-Schucharatt) was used and the pH of all suspensions was adjusted using sodium hydroxide and perchloric acid. Purified air was used as a source of oxygen in this study. Unless otherwise stated, all solutions were prepared using Millipore MilliQ water with a resistivity of 18.2 MWcm. A U-shaped hollow fibre membrane module of area, 4.107 ´ 10–4 m2 , was used. The MF membranes made of poly-vinylidene fluoride (PVDF) were supplied by Bluestar China, with a pore size of 0.22 µm and outer diameter of 1 mm. The choice of PVDF for the S.S. Chin et al. / Desalination 202 (2007) 253–261
255S.S. Chin et al. / Desalination 202 (2007) 253-261byfour8Wblack light fluorescentlamps(NECmembranewasmadefollowing extensivetestson membrane stability inthepresence of UVFL8 BL-B) located on the four sides of the photo-and UV-H,O, [7].reactor.About 800 mL of TiO, suspension wasprepared by adding pre-determined amounts ofTiO, and BPA to a known volume of water fol-2.2.Effectofaerationlowed by pH adjustment.Thefinal solution con-Batch photocatalytic tests were carried out totaining10ppmBPAwas sonicatedfor10minstudy the effect of aeration on the kinetics ofbefore charging into the photoreactor. Duringphotodegradation of BPA.The air was spargedthe experiment, a constant flux of permeate wasto the system through a ceramic bubble diffuserwithdrawnfromthesystemthroughthe suctiontoprovidegoodmixing and adesirablelevel ofprovidedbyaperistalticpump.Fortheintermit-dissolved oxygen (DO). The aeration rate wastent permeationstudy,thesuction pressure wasstudied in the range of 0.2-4 L/min. Samples atapplied intermittently in order to withdraw per-predefined times were taken and filtered throughmeate in a stepwise manner. The transmem-0.22 μm filters for analysis.brane pressure (TMP)in all experimentswasmeasured by a pressure gauge placed in the per-meate line. A level sensor was installed and it2.3.Submergedmembranephotocatalyticcontrolled an inlet pump which pumped a solu-reactor (SMPR)tionof10ppmBPAintothesysteminordertoAphotoreactor made of borosilicate glasskeep the total volume of the reactor constant.wasusedto studythephotocatalyticdegradationTheoptimumaerationrate(0.5L/min),obtainedofBPA.Aschematic diagramofthe submergedfromtheaeration study,wasapplied insubse-hollow fibre MF system is shown in Fig. 1. Itquent experimentations. Samples were taken atconsisted of a borosilicate glass photoreactorpredefined times fromthe permeateline andanalysed for BPA and organic carbon concentra-withaportablehollowfibremembranemodulein the centre. UVA irradiation was providedtions.For the critical flux determination, a fluxstepping method was used whereby the SMPRpermeate flux was increased in increments ofPres20L/m*huptoafluxof 100L/m*h.EachfluxAastepwasmaintainedforadurationof25minTherateof increase of TMPwas used todetermine the critical flux.Panmeattank2.4.AnalysisBalanceThe concentration of BPA in samples wasmeasured using high performance liquid chro-matograph (Waters Alliance)interfaced with aUV blacklight bluephotodiode array detector. The separation wasmodulecarried out using a Waters Xterra RP18 HPLCcolumn(3.9mm×150mm,5μm)withamobileBubblephase consisting of water-acetonitrile mixtureAir supplydiffuser(50%:50%by volume).TheTOC present in thesamples was determined by a TOC analyzerFig. 1. Schematic diagram of the experimental setup
255 membrane was made following extensive tests on membrane stability in the presence of UV and UV-H2O2 [7]. 2.2. Effect of aeration Batch photocatalytic tests were carried out to study the effect of aeration on the kinetics of photodegradation of BPA. The air was sparged to the system through a ceramic bubble diffuser to provide good mixing and a desirable level of dissolved oxygen (DO). The aeration rate was studied in the range of 0.2–4 L/min. Samples at predefined times were taken and filtered through 0.22 µm filters for analysis. 2.3. Submerged membrane photocatalytic reactor (SMPR) A photoreactor made of borosilicate glass was used to study the photocatalytic degradation of BPA. A schematic diagram of the submerged hollow fibre MF system is shown in Fig. 1. It consisted of a borosilicate glass photoreactor with a portable hollow fibre membrane module in the centre. UVA irradiation was provided by four 8 W black light fluorescent lamps (NEC FL8 BL-B) located on the four sides of the photoreactor. About 800 mL of TiO2 suspension was prepared by adding pre-determined amounts of TiO2 and BPA to a known volume of water followed by pH adjustment. The final solution containing 10 ppm BPA was sonicated for 10 min before charging into the photoreactor. During the experiment, a constant flux of permeate was withdrawn from the system through the suction provided by a peristaltic pump. For the intermittent permeation study, the suction pressure was applied intermittently in order to withdraw permeate in a stepwise manner. The transmembrane pressure (TMP) in all experiments was measured by a pressure gauge placed in the permeate line. A level sensor was installed and it controlled an inlet pump which pumped a solution of 10 ppm BPA into the system in order to keep the total volume of the reactor constant. The optimum aeration rate (0.5 L/min), obtained from the aeration study, was applied in subsequent experimentations. Samples were taken at predefined times from the permeate line and analysed for BPA and organic carbon concentrations. For the critical flux determination, a flux stepping method was used whereby the SMPR permeate flux was increased in increments of 20 L/m2 h up to a flux of 100 L/m2 h. Each flux step was maintained for a duration of 25 min. The rate of increase of TMP was used to determine the critical flux. 2.4. Analysis The concentration of BPA in samples was measured using high performance liquid chromatograph (Waters Alliance) interfaced with a photodiode array detector. The separation was carried out using a Waters Xterra RP18 HPLC column (3.9 mm ´ 150 mm, 5 µm) with a mobile phase consisting of water–acetonitrile mixture (50%:50% by volume). The TOC present in the samples was determined by a TOC analyzer UV blacklight blue Air supply Membrane module Bubble diffuser Balance Permeate tank Pressure gauge Feed tank Fig. 1. Schematic diagram of the experimental setup. S.S. Chin et al. / Desalination 202 (2007) 253–261
256S.S.Chinetal./Desalination202(2007)253-26110M90.2L/min508765432*0.5L/mindoe01L/min303204L/min100Aeration rate (L /min)1Ot0105060702030408090Time (min)Fig.2.Effect of aeration rates on (a) batch photocatalytic degradation of BPA (b) 80% removal of BPA [0.5 g/L TiO2initialpH=4,initialbisphenolA(BPA)concentration=10ppm).(Shimadzu TOC-VcsH). The particle size distri-methyl orange and its effectreached a plateaubution and turbidity of samples were measuredaboveacertainflowrate.by a Brookhaven ZetaPals particle sizer and aThe mean size of the TiO,particle aggregatesHACH2100AN IS turbidity meter,respectivelyafter1.5hof aerationareshown inTable1.Theaverage50% size distribution of0.2and 4L/minaeration rate are0.75 and 0.31 μm, respectively.Higher aeration rates produced greater shear3.Resultsand discussionrateswhichdispersedtheparticleswellandpre-3.1.Effectofaerationratevented agglomeration of particles.With lessagglomeration occurring, more surface areaAerationisimportantforsubmerged mem-wouldbeavailableforthedegradationofBPA.brane systems and photocatalytic reactions.TheSoparajee et al. [9] also reported the same nega-formerprocess requires the mechanical agitationtive effect of TiO, particles agglomeration onfrom aeration to reduce the fouling of thethephotocatalytic activity.membraneas well astokeepthemedium in sus-For a three-phase reactor, the bubbling ratepension; whereas the later process requireshas to be carefully chosen.Bubbling can increasedissolved oxygen for slowing down electrontheliquid film mass transfercoefficientaroundholerecombination reactions.Fig.2shows thatthe reaction rate increased with an increase inTable 1bubblingrateand reached amaximum valueat aAverage particle size of TiO,at different aeration ratebubblingrateof o.5L/min,beyond whichnofurther enhancement was observed.The insetAeration rate (L/min)Mean particle size (nm)Fig.2b shows the timefor80%removal of BPA0.2757 ± 147.1(to). Lee et al. [8], who performed experiments0.5477±49.5in a fluidized photocatalytic reactor, reported a1491 ±52.3similar observation.In theirexperiment,bub-4309 ±32.6bling strongly affected the photodestruction of
256 (Shimadzu TOC-VCSH). The particle size distribution and turbidity of samples were measured by a Brookhaven ZetaPals particle sizer and a HACH 2100AN IS turbidity meter, respectively. 3. Results and discussion 3.1. Effect of aeration rate Aeration is important for submerged membrane systems and photocatalytic reactions. The former process requires the mechanical agitation from aeration to reduce the fouling of the membrane as well as to keep the medium in suspension; whereas the later process requires dissolved oxygen for slowing down electron hole recombination reactions. Fig. 2 shows that the reaction rate increased with an increase in bubbling rate and reached a maximum value at a bubbling rate of 0.5 L/min, beyond which no further enhancement was observed. The inset Fig. 2b shows the time for 80% removal of BPA (t80). Lee et al. [8], who performed experiments in a fluidized photocatalytic reactor, reported a similar observation. In their experiment, bubbling strongly affected the photodestruction of methyl orange and its effect reached a plateau above a certain flow rate. The mean size of the TiO2 particle aggregates after 1.5 h of aeration are shown in Table 1. The average 50% size distribution of 0.2 and 4 L/min aeration rate are 0.75 and 0.31 µm, respectively. Higher aeration rates produced greater shear rates which dispersed the particles well and prevented agglomeration of particles. With less agglomeration occurring, more surface area would be available for the degradation of BPA. Soparajee et al. [9] also reported the same negative effect of TiO2 particles agglomeration on the photocatalytic activity. For a three-phase reactor, the bubbling rate has to be carefully chosen. Bubbling can increase the liquid film mass transfer coefficient around 0 1 2 3 4 5 6 7 8 9 10 0 10 20 30 40 50 60 70 80 90 Time (min) BPA concentration (ppm) 0.2 L/min 0.5 L/min 1 L/min 4 L/min 0 10 20 30 40 50 0 12345 Aeration rate (L /min) t80 Fig. 2. Effect of aeration rates on (a) batch photocatalytic degradation of BPA (b) 80% removal of BPA [0.5 g/L TiO2, initial pH = 4, initial bisphenol A (BPA) concentration = 10 ppm]. Table 1 Average particle size of TiO2 at different aeration rate Aeration rate (L/min) Mean particle size (nm) 0.2 757 ± 147.1 0.5 477 ± 49.5 1 491 ± 52.3 4 309 ± 32.6 S.S. Chin et al. / Desalination 202 (2007) 253–261
257S.S. Chin et al. / Desalination 202 (2007) 253-26the aggregates [10] but it may also provide a bub-showsthatthe concentrationof BPAwasble cloud that can attenuate UV light transmissionreduced by97%after90minofUV illuminationin a photoreactor. The attenuation of UV lightand remained atthis valueuntil theend of themight be thereason for the slower 80% removalexperiment.rate of BPA which occurred with 4 L/min aerationThe main objective of a photocatalytic(33 min) compared to 0.5 L/min (28 min) as seenprocess is the complete mineralization of allin Fig.2b (inset).The balance of the competingorganiccarbontoensurethatthesubstrateandeffectsbetweenmasstransferandlightattenua-any intermediates formed have been degraded.tion leads to an optimal bubbling rate.It was found byChiang etal.[6] that someofThus, an aeration rate of 0.5 L/min has beenthe intermediatesgeneratedduringthephotoused in subsequent experiments. The use of 4L/mindegradation of BPA were more toxic than theaeration was avoided due to concern with theparent compound, especially at low pHs.There-aeration energy consumption. A strategy wasfore, it is always important to monitor thedevelopedtoavoidfoulingunderconditionsofmineralization ofanyPCOreaction.AscanbeseenfromFig.3,over90%oflowaeration (see section3.4)BPA was mineralized after 120 min of UVillumination and thereafter the concentration of3.2.Effectivenessofsubmerged membraneTOC remained at about 7% throughout thephotocatalyticreactorexperimental duration.At 60 min, the concen-Fig. 3 shows the photodegradation and photo-trationofTOCwas about3ppmwhile themineralization efficiency of BPA using theconcentration of BPA was about 1.6 ppmcontinuous SMPR.The permeate flux used was(equivalent to TOC of 1.26 ppm). The differ-100L/m’h, which was equivalent to a BPAence between the concentration of TOC andBPA concentrations indicates the presence ofretentiontimeof2h.Turbidity analysis of the permeate (similar tointermediates during the PCOreaction.The MF membrane was not able to retainMQwaterturbidityvalues)indicated thattherewas no leakage of TiO, from the SMPR.Fig.3BPA and its intermediate products in the10-PermeateBPAPermeateTOCooeor6420501001502002503000Time (min)Fig.3.Photodegradation and photomineralization of bisphenol A(BPA)in the submerged membrane photoreactor(SMPR) (Flux = 100 L/m* h, 0.5 g/L TiO, initial pH = 4, initial BPA concentration = 10 ppm, 0.5 L/min aeration rate)
257 the aggregates [10] but it may also provide a bubble cloud that can attenuate UV light transmission in a photoreactor. The attenuation of UV light might be the reason for the slower 80% removal rate of BPA which occurred with 4 L/min aeration (33 min) compared to 0.5 L/min (28 min) as seen in Fig. 2b (inset). The balance of the competing effects between mass transfer and light attenuation leads to an optimal bubbling rate. Thus, an aeration rate of 0.5 L/min has been used in subsequent experiments. The use of 4L/min aeration was avoided due to concern with the aeration energy consumption. A strategy was developed to avoid fouling under conditions of low aeration (see section 3.4). 3.2. Effectiveness of submerged membrane photocatalytic reactor Fig. 3 shows the photodegradation and photomineralization efficiency of BPA using the continuous SMPR. The permeate flux used was 100 L/m2 h, which was equivalent to a BPA retention time of 2 h. Turbidity analysis of the permeate (similar to MQ water turbidity values) indicated that there was no leakage of TiO2 from the SMPR. Fig. 3 shows that the concentration of BPA was reduced by 97% after 90 min of UV illumination and remained at this value until the end of the experiment. The main objective of a photocatalytic process is the complete mineralization of all organic carbon to ensure that the substrate and any intermediates formed have been degraded. It was found by Chiang et al. [6] that some of the intermediates generated during the photodegradation of BPA were more toxic than the parent compound, especially at low pHs. Therefore, it is always important to monitor the mineralization of any PCO reaction. As can be seen from Fig. 3, over 90% of BPA was mineralized after 120 min of UV illumination and thereafter the concentration of TOC remained at about 7% throughout the experimental duration. At 60 min, the concentration of TOC was about 3 ppm while the concentration of BPA was about 1.6 ppm (equivalent to TOC of 1.26 ppm). The difference between the concentration of TOC and BPA concentrations indicates the presence of intermediates during the PCO reaction. The MF membrane was not able to retain BPA and its intermediate products in the 0 2 4 6 8 10 0 50 100 150 200 250 300 Time (min) Concentration (ppm) Permeate BPA Permeate TOC Fig. 3. Photodegradation and photomineralization of bisphenol A (BPA) in the submerged membrane photoreactor (SMPR) (Flux = 100 L/m2 h, 0.5 g/L TiO2, initial pH = 4, initial BPA concentration = 10 ppm, 0.5 L/min aeration rate). S.S. Chin et al. / Desalination 202 (2007) 253–261
258S.S. Chin et al. /Desalination 202 (2007)253-261system.Therefore, it would take time for BPAchoiceof 2minclosedand5minopenwasto be degraded at start-up and it is recommendedarbitraryandfurtheroptimizationisplannedthat the permeate line only be opened when theWhen suction was stopped and no permeate wascollected, there was a period for the aeration toeffluent water hasmet specification.exert shear rate on the membrane surfacetofacilitate the detachment of TiO, particles.This3.3.Intermittentpermeationshouldprevent an accumulation of TiO,Fig. 4 shows the determination of the criticalparticles on the membrane. The advantages offlux for the SMPR TiO, suspension and opera-intermittent permeation have also been reportedting conditions.The TMP was constantforby Hong et al. [12] and Howell et al. [4] duringfiltration fluxes below 80 L/m’h.Definingthe operation of their MBR. Hong et al. [12]critical flux as the flux below which fouling ofsuggested that the use of intermittent suctionmembrane does not occur [11], the critical fluxoperation could be economically feasible forfound in this study was 80 L/m h.Above thiscontrolling fouling in small-scale MBR pro-cesses treating wastewaters with high foulingvalue,the TMP wasnot steady.The operationof the SMPR at a flux of 100 L/m h resulted inpotentials.For the TiO,filtration process, inter-fouling, where the TMP steadily increased tomittency should be even more effective as theabout 12.5 kPa from the initial 10 kPa as shownTiO, deposits are more reversible than MBRin Fig. 5.deposits.One way to increase critical flux is toAccumulation on the membrane indirectlyindicated the loss of suspended TiO in theincreaseaeration.However,this would involvemore energy usage and could also attenuate UVSMPR and thus reducing overall active sites ofTiO, available for PCO reaction. By mass bal-transmission.Therefore,inordertomaintainthehigh flux operation at low aeration, the effect ofance,approximately38%(±14%)of theTiO,intermittentpermeation was investigatedandthewasfoundonthesurfaceofthemembraneafterresults shown in Fig.5.In this case, the perme-5h of continuous filtrationcompared to 7%ation was closed for 2minfor every5min.The(+2%)ofTiO,for the intermittentfiltration12120L/m2h10100 L/m?h8680L/m2h(660L/m2h40L/m2h4210020406080100120140Time (min)Fig.4.Thechange in transmembranepressurefor different filtration flux (0.5g/LTiO2,0.5L/min bubbling)
258 system. Therefore, it would take time for BPA to be degraded at start-up and it is recommended that the permeate line only be opened when the effluent water has met specification. 3.3. Intermittent permeation Fig. 4 shows the determination of the critical flux for the SMPR TiO2 suspension and operating conditions. The TMP was constant for filtration fluxes below 80 L/m2 h. Defining critical flux as the flux below which fouling of membrane does not occur [11], the critical flux found in this study was 80 L/m2 h. Above this value, the TMP was not steady. The operation of the SMPR at a flux of 100 L/m2 h resulted in fouling, where the TMP steadily increased to about 12.5 kPa from the initial 10 kPa as shown in Fig. 5. One way to increase critical flux is to increase aeration. However, this would involve more energy usage and could also attenuate UV transmission. Therefore, in order to maintain the high flux operation at low aeration, the effect of intermittent permeation was investigated and the results shown in Fig. 5. In this case, the permeation was closed for 2 min for every 5 min. The choice of 2 min closed and 5 min open was arbitrary and further optimization is planned. When suction was stopped and no permeate was collected, there was a period for the aeration to exert shear rate on the membrane surface to facilitate the detachment of TiO2 particles. This should prevent an accumulation of TiO2 particles on the membrane. The advantages of intermittent permeation have also been reported by Hong et al. [12] and Howell et al. [4] during the operation of their MBR. Hong et al. [12] suggested that the use of intermittent suction operation could be economically feasible for controlling fouling in small-scale MBR processes treating wastewaters with high fouling potentials. For the TiO2 filtration process, intermittency should be even more effective as the TiO2 deposits are more reversible than MBR deposits. Accumulation on the membrane indirectly indicated the loss of suspended TiO2 in the SMPR and thus reducing overall active sites of TiO2 available for PCO reaction. By mass balance, approximately 38% (±14%) of the TiO2 was found on the surface of the membrane after 5 h of continuous filtration compared to 7% (±2%) of TiO2 for the intermittent filtration. 0 2 4 6 8 10 12 0 20 40 60 80 100 120 140 Time (min) TMP (kPa) 40 L/m2 h 60 L/m2 h 80 L/m2 h 100 L/m2 h 120 L/m2 h Fig. 4. The change in transmembrane pressure for different filtration flux (0.5 g/L TiO2, 0.5 L/min bubbling). S.S. Chin et al. / Desalination 202 (2007) 253–261
259S.S.Chinetal./Desalination202(2007)253-26117 5 min open, 2 min close15(IF = 0.4) Continuously13(ed>)11I050150200100250Time (min)Fig.5.Transmembranepressure(TMP)changefor continuous1ooL/mh and intermittentpermeation operation (inter-mittencyof2minclose/5minopen).The intermittence frequency (IF) is definedresulted in a decrease in water production byas the ratio of the off time to on time. From aabout 30% (Fig. 6, IF = 0.4). Therefore, anpractical point of view, a lower IF will mean aexperimentwithIFof 0.1(1min close,10minhigher production of water. Although intermit-open)was carried out.tent operation of 2 min off and 5min on wasIt is worth noting that the TOC photominer-shownto reducethefoulingtendencyofthealizationwithanIFofo.1wassimilartothatmembrane as indicated in Figs. 5 and 6, itobtainedwithanIFof0.4andwithcontinuous140ChangeinTMP120Efficiency ofTOC degradationEfficiency of production of water1008060140 20ContinuousIF (0.1)IF (0.4)Fig. 6.Comparison of the transmembrane pressure (TMP) change, total organic carbon (TOC) degradation and produc-tionofwaterbetweencontinuousand intermittentpermeationoperation
259 The intermittence frequency (IF) is defined as the ratio of the off time to on time. From a practical point of view, a lower IF will mean a higher production of water. Although intermittent operation of 2 min off and 5 min on was shown to reduce the fouling tendency of the membrane as indicated in Figs. 5 and 6, it resulted in a decrease in water production by about 30% (Fig. 6, IF = 0.4). Therefore, an experiment with IF of 0.1 (1 min close, 10 min open) was carried out. It is worth noting that the TOC photomineralization with an IF of 0.1 was similar to that obtained with an IF of 0.4 and with continuous 5 7 9 11 13 15 17 0 50 100 150 200 250 Time (min) TMP (kPa) 5 min open, 2 min close (IF = 0.4) Continuously Fig. 5. Transmembrane pressure (TMP) change for continuous 100 L/m2 h and intermittent permeation operation (intermittency of 2 min close/5 min open). 0 20 40 60 80 100 120 140 Continuous IF ( 0.1) IF (0.4) Percentage (%) Change in TMP Efficiency of TOC degradation Efficiency of production of water Fig. 6. Comparison of the transmembrane pressure (TMP) change, total organic carbon (TOC) degradation and production of water between continuous and intermittent permeation operation. S.S. Chin et al. / Desalination 202 (2007) 253–261
260S.S. Chin et al. / Desalination 202 (2007) 253-261operations i.e.94%BPAreduction at250min.AcknowledgementThe water production efficiencies with IF of 0.4The authors would like to acknowledgeand 0.1 would be approximately 70 and 90%,financial support from ASTAR(Singapore)forrespectively.Reducing the IF from 0.4 to 0.1the Temasek ProfessorshipProgramme onhas increased the water production rate by aboutMembraneTechnologyforSustainableWater.20%. Most importantly, it was found that theTMP of the system was decreased byfivefoldwith IF of 0.1 compared to the continuous oper-Referencesation. As TMP is proportional to the filtration[]]D.S.Bhatkande, V.G. Pangarkar and A. Beenackers.energy cost [13], intermittent operation wouldPhotocatalytic degradation for environmental appli-reduce the operating cost of the continuouscations-—a review,J.Chem.Tech.Biotechnol.,77SMPR.The use of intermittent permeation has(2001)102-111.[2] R.Molinari, L.Palmisano,E.Drioli and M.Schia-made the operation of the SMPR at fluxes abovevello,Studies on various rector configurations forcritical flux possible, taking into account of thecoupling photocatalysis and membrane processesreversibility of the cake deposits.in waterpurification,J.Memb.Sci.,206(2002)399-415.[3]A.G. Fane, S. Chang and E. Chardon, Submerged4.Conclusionshollow fibre membrane module-design optionsAeration in a photocatalytic reactor enhancedand operational considerations,Desalination,146thephotodegradationrate byprovidinghigher(2002)231-236[4] J.A.Howell,H.C.Chua and T.C.Arnot, In situconcentrations of dissolved oxygen (DO) andmanipulation of critical flux in a submerged mem-improving the mass transfer processin thebrane bioreactor using variable aeration rates,andsystem. Photoactivity was found to increaseeffectsof membranehistory,J.Memb.Sci.,242with aeration rate up to a limit of 0.5 L/min(2004)13-19.Beyond 0.5 L/min, no significant improvement[5] J. Fu, M. Ji and L. Jin, Submerged Membranewas observed in the photocatalyticprocess.ThePhotocatalysis Reactor forFulvic Acid RemovalSMPRwasabletodegrade97%of 10ppmofin DrinkingWaterbyUsingNanoparticleTiO,BPA after 70 min of UV illumination.Further-Proceedings 3rdIWA Leading-EdgeConferenceandExhibitiononWaterandWastewaterTreatmore,itwas also capable ofmineralizing morement Technologies, Sapporo, Japan,6-8 Junethan90%of theTOC after100minof UV2005.illumination. It was found that an intermittentK. Chiang, T.M. Lim, L. Tsen and C.C. Lee.[6]filtration process reduced fouling through thePhotocatalytic degradation and mineralization ofremoval of the TiO, layer from the membranebisphenol A by TiO, and platinized TiOz, Appl.surface intermittentlyand thusallowed highfluxCatal.AGen.,261 (2)(2004)225-237.operationoftheSMPR.ByvaryingtheIF,the[7]S.S.Chin,K.Chiang and A.G.Fane,The stabilityofpolymericmembranesinaTiO,photocatalysisfouling of thesubmerged membranecanbecon-process, J.Memb.Sci.,inPress.trolled. It should be possible to optimize the IF[8]J.H. Lee, W. Nam, M. Kang,G.Y. Han, K.J.to obtain maximum production of water withYoon, M.-S.Kim, K. Ogino, S. Miyata and S.-J.sustainability of membrane operation.An IF ofChoung,Designof twotypesof fluidized photo0.1wasfoundtobesufficientto stabilizethereactorsandtheirphoto-catalyticperformancesconditions at afluxof 100L/mh.The compactfor degradation of methyl orange, Appl. Catal.Aand low-pressure SMPR has shown to be effec-Gen.,244 (1) (2003)49-57.tiveand attractive for the photodegradation of[9]K.Soparajee,S.AQasim,S.BasakandK, Rajeshwar, An integrated flow reactor-membranelow concentrations of organics in water
260 operations i.e. 94% BPA reduction at 250 min. The water production efficiencies with IF of 0.4 and 0.1 would be approximately 70 and 90%, respectively. Reducing the IF from 0.4 to 0.1 has increased the water production rate by about 20%. Most importantly, it was found that the TMP of the system was decreased by fivefold with IF of 0.1 compared to the continuous operation. As TMP is proportional to the filtration energy cost [13], intermittent operation would reduce the operating cost of the continuous SMPR. The use of intermittent permeation has made the operation of the SMPR at fluxes above critical flux possible, taking into account of the reversibility of the cake deposits. 4. Conclusions Aeration in a photocatalytic reactor enhanced the photodegradation rate by providing higher concentrations of dissolved oxygen (DO) and improving the mass transfer process in the system. Photoactivity was found to increase with aeration rate up to a limit of 0.5 L/min. Beyond 0.5 L/min, no significant improvement was observed in the photocatalytic process. The SMPR was able to degrade 97% of 10 ppm of BPA after 70 min of UV illumination. Furthermore, it was also capable of mineralizing more than 90% of the TOC after 100 min of UV illumination. It was found that an intermittent filtration process reduced fouling through the removal of the TiO2 layer from the membrane surface intermittently and thus allowed high flux operation of the SMPR. By varying the IF, the fouling of the submerged membrane can be controlled. It should be possible to optimize the IF to obtain maximum production of water with sustainability of membrane operation. An IF of 0.1 was found to be sufficient to stabilize the conditions at a flux of 100 L/m2 h. The compact and low-pressure SMPR has shown to be effective and attractive for the photodegradation of low concentrations of organics in water. Acknowledgement The authors would like to acknowledge financial support from ASTAR (Singapore) for the Temasek Professorship Programme on Membrane Technology for Sustainable Water. References [1] D.S. Bhatkande, V.G. Pangarkar and A. Beenackers, Photocatalytic degradation for environmental applications — a review, J. Chem. Tech. Biotechnol., 77 (2001) 102–111. [2] R. Molinari, L. Palmisano, E. Drioli and M. Schiavello, Studies on various rector configurations for coupling photocatalysis and membrane processes in water purification, J. Memb. Sci., 206 (2002) 399–415. [3] A.G. Fane, S. Chang and E. Chardon, Submerged hollow fibre membrane module — design options and operational considerations, Desalination, 146 (2002) 231–236. [4] J.A. Howell, H.C. Chua and T.C. Arnot, In situ manipulation of critical flux in a submerged membrane bioreactor using variable aeration rates, and effects of membrane history, J. Memb. Sci., 242 (2004) 13–19. [5] J. Fu, M. Ji and L. Jin, Submerged Membrane Photocatalysis Reactor for Fulvic Acid Removal in Drinking Water by Using Nanoparticle TiO2, Proceedings 3rd IWA Leading-Edge Conference and Exhibition on Water and Wastewater Treatment Technologies, Sapporo, Japan, 6–8 June 2005. [6] K. Chiang, T.M. Lim, L. Tsen and C.C. Lee, Photocatalytic degradation and mineralization of bisphenol A by TiO2 and platinized TiO2, Appl. Catal. A Gen., 261 (2) (2004) 225–237. [7] S.S. Chin, K. Chiang and A.G. Fane, The stability of polymeric membranes in a TiO2 photocatalysis process, J. Memb. Sci., in Press. [8] J.H. Lee, W. Nam, M. Kang, G.Y. Han, K.J. Yoon, M.-S. Kim, K. Ogino, S. Miyata and S.-J. Choung, Design of two types of fluidized photo reactors and their photo-catalytic performances for degradation of methyl orange, Appl. Catal. A Gen., 244 (1) (2003) 49–57. [9] K. Soparajee, S.A Qasim, S. Basak and K. Rajeshwar, An integrated flow reactor-membrane S.S. Chin et al. / Desalination 202 (2007) 253–261
261S.S.Chin et al. /Desalination 202 (2007)253-261filtration system for heterogeneous photocatalysis.[1]] J.A. Howell, Sub-critical flux operation, J. Memb.Part Il: experiments on the ultrafiltration unit andSci., 107 (1995) 165-171.combined operation, J. Appl. Electrochem., 29[12] S.P. Hong, T.H. Bae, T.M. Tak, S. Hong and(1999) 11111118.A.Randall, Fouling control in activated sludge[10]Y.Jia, R Wang and A.G.Fane, Atrazine adsorptionsubmerged hollow fiber membrane bioreactors,in water by powdered activated carbon-improvedDesalination, 143 (3) (2002) 219-228.mass transferby bubbling, Chem.Eng.J.,116[13] IM. Cheryan, Ultrafiltration and Microfiltration(2006) 53-59.Handbook,Technomic,Lancaster,PA,1998
261 filtration system for heterogeneous photocatalysis. Part II: experiments on the ultrafiltration unit and combined operation, J. Appl. Electrochem., 29 (1999) 1111–1118. [10] Y. Jia, R. Wang and A.G. Fane, Atrazine adsorption in water by powdered activated carbon-improved mass transfer by bubbling, Chem. Eng. J., 116 (2006) 53–59. [11] J.A. Howell, Sub-critical flux operation, J. Memb. Sci., 107 (1995) 165–171. [12] S.P. Hong, T.H. Bae, T.M. Tak, S. Hong and A. Randall, Fouling control in activated sludge submerged hollow fiber membrane bioreactors, Desalination, 143 (3) (2002) 219–228. [13] M. Cheryan, Ultrafiltration and Microfiltration Handbook, Technomic, Lancaster, PA, 1998. S.S. Chin et al. / Desalination 202 (2007) 253–261