Control of membrane fouling by Hydrogen Sulfide (H2S) diffusion ...

Control of membrane fouling by Hydrogen Sulfide (H 2 S) diffusionHasan Mahmud KhanWater and Environmental Engineering, Department of Chemical Engineering, Lund UniversityP.O. Box 124, SE-221 00, Lund, SwedenAbstractThe major problem of widespread applications of membrane bioreactors (MBRs) in water andwastewater treatment is quick membrane fouling and consequently it’s operation cost in term offouling mitigation although this technology ensures better treatment efficiency and high effluentquality. This work presents an experimental study on membrane fouling in MBR operation atdifferent concentrations of hydrogen sulphide (H 2 S). Nine samples have been collected fromMBRs to investigate the sludge and membrane properties before and after the application of 0,25 ppm, 50 ppm, and 100 ppm concentrations of H 2 S. It was found that protein, carbohydrate,total organic carbon (TOC) of eEPS and TOC of SMP decreased and pH increased with theincrease of H 2 S concentration in feed whereas zeta potential showed no trends.Key works : MBR, membrane fouling, eEPS, SMP, TOC, pH, H 2 S diffusionIntroductionMBR, the technology of membraneseparation of activated sludge, is theactivated sludge treatment together with thebiological sludge separation by micro- orultra-filtration membranes with pore size ofnormally 10 nm to 0.5µm to generateparticle-free effluent (Lesjean et al. 2007).MBR technologies have been potentiallyimproved after an intensive research in lastdecade. It is gaining popularity in water andwastewater treatment sector due to its betterquality of effluent compare to conventionaltreatment process, high treatment efficiency,less foot print and small reactorrequirements, good disinfection capabilityand higher volumetric loading (Liang et al.,2007 & Judd, S., 2006). It offers a completeseparation of solid-liquid in the biologicaldegradation process by activated sludge,truly physical retention of bacterial floc andother suspended solid within bioreactor (Le-Clech et al., 2006). However, its stillchallenging in widespread applications ofMBR for rapid fouling of membraneresulting quick decline of permeate flux andhigh maintenance cost in rigorous physicaland chemical cleaning (Meng et al. 2006).Foul elements deposit and form a cake layeron the membrane surface in consequence ofblocking and narrowing the membranepores. After a certain time trans-membranepressure increases sharply or permeates fluxdrops below the significant level underconstant pressure. This fouling propensity isthe major obstacle of MBR applications incommercial sector over conventionalsystems. The main objective of this work isto control the membrane fouling throughH 2 S diffusion.Extracellular polymeric substances (EPS)EPS are the composition of polymericmaterials consist of carbohydrates, proteins,nucleic acids and have been found at or outside the cell surface and in the intercellularopening of microbial aggregates. It is theconstruction material of biofilm, flocs.Carbohydrates and proteins are the foremostcomponents in extracted EPS and areresponsible for membrane fouling. EPSprovide a highly hydrated gel wheremicroorganisms are embedded and theyhamper flux to permeate in the MBRs.Specific EPS resistance could be estimatedfrom the filtration resistance (m -1 ) dividedby EPS density on the membrane surfaces(kgTOC/m 2 ) and it is between 10 16 and 10 17m/kg. Recent studies showed that proteinshave strong positive correlation tomembrane resistance whereas carbohydrates1

have moderate correlation due to lowamount. According to the currentresearches, proteins are the most importantfactor of membrane fouling (Meng et al.,2006).Soluble microbial product (SMP)SMP are the large amount of solubleorganic substances that are produced frombacterial populations in bioreators. SMPlevel is extremely important for biologicaltreatment process as treatment efficiencyand effluent quality highly depends on SMPamount in treatment reactors. It is wellestablished that majority of soluble organicsubstances come from SMP in effluent inbiological treatment systems and itsconcentration set up the discharge level ofchemical oxygen demand (COD) and totalorganic carbon (TOC) in effluent.Furthermore, some SMP have certaincharacteristic like toxicity and metalchelating properties which affect themetabolic activities of microorganisms bothin receiving water and in treatment process.In some case, it lessens the specificrespiration rate of microorganisms.Therefore, for better efficiency of treatmentsystems, less concentration of SMP isdesirable (Liang et al., 2007).SMP could be again categorized into two asutilization associated products and biomassassociated products. Although humicsubstances (humic and fulvic acids),carbohydrates and proteins have beenidentified successfully as major componentsof SMP, it is still unclear of its precisecomposition (Liang et al., 2007).However, most of previous studiesconcentrated on SMP in conventionalbiological treatment plants but fewer havebeen focused to grasp the impact of SMP inMBR. It is found from study that dependingof the operation condition SMP are liablefrom 26 to 52% of membrane fouling inmicrofiltration and ultrafiltrationmembranes (Liang et al., 2007).Materials and MethodHydrogen Sulphide (H 2 S) diffusionAir and mixture of air & H 2 S was sprayedinto sludge for getting differentconcentration of H 2 S in sludge. Pressure andflow rate were 2 bars and 400 ml per minrespectively and was aerated for 15 min andafter then kept it for 10 min forhomogeneous distribution of H 2 S in sludge.High concentration of H 2 S (2000 mg/l) hasbeen diluted by mixing of compressed air toget a certain concentration of H 2 S. Two gasflow meters were used to control the air andH 2 S gas flow and connected to the gascylinder by plastic tube line. This mixture ofgases has been applied to feed by diffuserwhich was connected at the end of sourcetube. Concentration of 25 ppm (0.075 mgH 2 S/g biomass), 50 ppm (0.15 mg H 2 S/gbiomass) and 100 ppm (0.3 mg H 2 S/gbiomass) of H 2 S were sprayed in threesamples with flow rate of 400 ml/min andeach sample is amount of 2 liters. Only airwas pumped into the sludge to get zeroconcentration H 2 S sample for same durationwith flow rate 400 ml/l. These samples wereanalyzed for eEPS, SMP, ZP and pH.H2S TankAir TankRegulatorRegulatorH2SFlowmeterAirFlowmeter2 mtubeline2 L Glass CylinderFigure 1: Schematic setup for H 2 S diffusionDead End FiltrationThe whole experiment was consisted ofdead end filtration cell, reservoir,compressed air and computer program formeasuring weight of collected permeate.Around one liter of aerated sludge is pouredin about 2 l volume closed feed reservoirand pump to dead end filtration cell (110 mlvolume capacity, 1.59 X 10 -3 m 2 membrane2

area). PVDF membrane (0.189 cm 2 ) withpore size of 0.22 µm is placed at on porousmedia at the end of cell and rubber circularring is placed on the membrane to preventleakage. Dead end cell is completely airtight and connected to feed tank by rubbertube. Compressed air cylinder is connectedto the tank through pressure gauge andpressure stabilizer for pumping sludge fromtank to cell. Applied air pressure iscontrolled by EL PRESS P-702CM pressuretransducer installed in the line tubing, whichis then connected to an EL PRESSelectronic regulator for a more accuratecontrol of pressure. Filtrated supernatantweight is measured every 60s by computerlinked balance and computer programcalculate the flux rate.Air tankAutomatic pressureregulatorPressure ControllerReservoirDead-endfiltration cellMass balanceFig 2: Dead-end filtration set-upPermeat eComputerSource of biomassBiomass has been collected from NorthHead MBR and Malabar MBR which wasoriginally from North Head’s MBR. Bothsewages go to MBR after primary treatmentand are a mixture of municipal andindustrial waste. Sludge age of Malabar is20 days and about 13L of sludge is wastedeach day to keep the sludge age at 20 daysand hydraulic retention time (HRT) is about10 hours. Volume of MBR is 250 l and thetotal membrane area is 2 m 2 with constantflux at 12 l/h.Nine samples have been used for theexperiment marked from A to I. Sampleshave been randomly selected for differenttests and observed the effect of H 2 S on thisparticular sludge characteristic.3SMP and EPS extractionHeating method has been followed toextract SMP and EPS. Sample is centrifugedfirst at 2000 rpm and supernatant is filtratedthrough 1.2 µm pore size filter paper forSMP extraction. The suspension was dilutedwith 0.98% NaCl solution, mixed for 5 minand kept it 80 0 C temperature for 10 min.This temperature has been controlled byhigh pressure stream sterilizer (ES-315).After then it was centrifuged at 2500 rpmand supernatant was again filtrated using 1.2µm pore size filter and collected for EPSextraction.Carbohydrate analysisPhono-Sulfuric acid method proposed byHanson and Philips, 1981 was used todetermine the carbohydrate concentration ofSMP and extracted EPS. In presence ofsulfuric acid, phenol and carbohydrate formcolored aromatic compounds and measuredit as an absorbance at 480 nm wave lengthby spectrophotometer. Concentration ofcarbohydrate then was determined from thevalue of absorbance using carbohydratecalibration curve which was developed fromstandard solution.Concentration, mg/lCarbohydrate Calibration353025y = 93.487xR 2 = 0.9891201510500 0.05 0.1 0.15 0.2 0.25 0.3 0.35Absorbance at 480 nmFig 3: Carbohydrate calibr. curve at 480 nmProtein analysisModified Lowry method has been used tomeasure the concentration of protein ofSMP and extracted EPS from filtrate whichwas modified by Peterson (Peterson, 1977).In this method, bonded proteins react tolowry reagent which contains copper ionsand Folic-Ciocalteu reagent composed ofphosphomolybdate and phosphotungstate.When lowey reagent is added to the filtrate,copper ions together with protein form a

complex alkaline solution, BuiretChromophore and further addition of Folin-Ciocalteu reagent reacted protein developsblue color that is detectable byspectrophotometer under the wavelengthfrom 500 to 750 nm. Protein concentrationis then calculated from calibration curve.mg/L Proteins150100500Protein calibrationy = 253.11xR 2 = 0.93260 0.1 0.2 0.3 0.4Absorbance 650Fig 4: Protein calibration curve at 650 nmZeta Potential (ZP)ZP is the electrical double layer that issurroundings the particles in a colloidalsolution or suspension (Lindquist, 2003).Ionic characteristics and dipolar features ofthe colloidal particles dispersed in thesolution are the causes of their electricallycharged. Particles are surrounded by theopposite charges which is called fixed layerand outside of fixed layer there is the cloudarea composed of positive and negativeions. This cloud area is called diffuse doublelayer. ZP is considered to be the electricpotential of this inner area including thisconceptual "sliding surface". As this electricpotential approaches zero, particles tend toaggregate.It has been measured of new and foulmembrane by PAAR EKA-Electro KineticAnalyzer RV 4.31C at different pH value.Total Organic Carbon (TOC)Sludge is centrifuged and filtrated throughmembrane with nominal pore size 0.45 µmin order to get particle free clear sample.Few drops of 98% H 2 SO 4 are added toremove any inorganic carbon fromsupernatant and TOC is then measured bytotal organic carbon analyzer (TOC-5000A).4Results and DiscussionEffect of H 2 S on ChemicalCharacteristics:pHpH876543210Effect of H 2 S on pHA C E F GSampleFig 5: Variation of pHNo H2S sparging25 ppm H2Ssparging50 ppm H2Ssparging100 ppm H2SspargingWastewater treatment process largelydepends on pH of biomass. It is observedfrom obtained result that pH valuemoderately increased with increase of H 2 Sconcentration. This elevation of pH valuemight be able to contribute positively onmembrane fouling propensity when originalpH of biomass is low. It is well known thatdrop of pH under certain level makes thewastewater treatment process difficult.Therefore, removal of H 2 S from activatedsludge could decrease the pH of sludge andmight lead higher fouling in MBR becausethe living species could not work properly atlow pH.Zeta Potential (ZP)ZP of membrane represents the surfacecharge of membrane. It is found from study,it has strong negative correlation withfouling resistance. Resistance increase whenZP drop and it plays a significant role inmembrane fouling (Meng et al., 2006).ZP of virgin and fouled membrane has beendetermined and results show that ZP offouled membrane is higher than virginmembrane but it was not found any impactof H 2 S that influence the ZP of membranesurface.From test result, it could be said that foulingdefinitely affect the ZP of membranesurface because clear difference has been

found between the virgin and all fouledmembranes but the effect of H 2 S in differentconcentration is ambiguous. ZP unevenlyvaries with H 2 S concentration.Effect of H 2 S on BiopolymericComposition of Biomass:Extracellular Polymeric Substances (EPS)Extracellular polymeric substances areconsisting of protein, carbohydrate andcarbon compound that are mainly liable formembrane fouling. Four samples have beentested to determine the concentration ofprotein, carbohydrate and total organiccarbon in extracted EPS. The results whichwere found from lab test are given below:ProteinProtein Concentration,mg/l300250200150100500Effect of H 2 S on eEPSpC D F GSampleNo H2S sparging25 ppm H2Ssparging50 ppm H2Ssparging100 ppm H2SspargingFig 6: Protein conc in eEPS at different H 2 SFrom the bar chart it is found that proteinconcentration of one set of sample (D) doesnot give any correlation with differentconcentration of H 2 S diffusion but in otherthree set of samples protein concentrationdecreases with increase of H 2 S diffusion insludge. According to Meng et al. (2006)fouling resistance increase with increase ofprotein concentration, sparging of H 2 S insludge could lead to decrease the foulingresistance and consequently increase thefilterability of membrane. As there is nostandard method for extraction of EPS, theobtained result might not be quite accuratebut in general it could be said that diffusionof H 2 S in activated sludge could enhancethe permeability of membrane. Dissolvedmatter is mainly responsible for irreversiblefouling (Yamato et al 2006) so thatapplication of H 2 S might lead to reduce theirreversible fouling.5CarbohydrateCarbohydrateConcentration, mg/l80706050403020100Effect of H 2 S on eEPScC D F GSampleNo H2S sparging25 ppm H2Ssparging50 ppm H2Ssparging100 ppm H2SspargingFig 7: Carbohydrate of eEPS at differentH 2 S conc.Concentration of carbohydrate in threesamples out of four samples dropped whenH 2 S was sparged in higher concentration.Carbohydrate has poor correlation withfouling resistance (Meng et al., 2006) sothat H 2 S diffusion is less effective techniqueto reduce the membrane resistance in termof carbohydrate. However, results give thepositive sign of filterability improvementwhen H 2 S was sparged in sludge.Total Organic CarbonTOC, mg/l250200150100500Effect of H 2 S on TOCD F GSampleNo H2S sparging25 ppm H2Ssparging50 ppm H2Ssparging100 ppm H2SspargingFig 8: TOC of eEPS at different H 2 S concConstituents of TOC like polysaccharides,humics are adsorbed on the membranesurface and increase the fouling resistanceby reducing the pore size. Adsorptiontendency is 46% for polysaccharides and16% for humics in the feed (Huber, 1998).It was found from lab results thatconcentration of TOC slightly drops withincrease of H 2 S diffusion in feed water.According to Huber’s (1998) study, it couldbe concluded that sparging of H 2 S improvethe permeability of membrane as it reducesTOC in feed.

Soluble Microbial Products (SMP)SMP is soluble EPS in feed water. It ismainly composed of protein, carbohydrateand total organic carbon. In our experiment,four samples were used to determine theconcentration protein and carbohydrate andfive samples for total organic carbon.Obtained results are given below:ProteinTable 1: Protein conc. (mg/l) in SMPRun C D F GNo H 2 S 38.8 8.3 0 025 ppm H 2 S 25.3 4.3 1.5 4.250 ppm H 2 S 46.3 0 0 0100 ppm H 2 S 33.8 0 14.1 0Concentration of protein in SMP has beendetermined after sparing of H 2 S in fourdifferent concentrations. Most results showthat protein concentration is very low and itdoes not vary with H 2 S concentration.CarbohydrateTable 2: Carbohydrate conc. (mg/l) in SMPRun C D F GNo H 2 S 0 7.3 6.2 23.625 ppm H 2 S 10.8 11.0 4.8 14.450 ppm H 2 S 8.3 10.9 10.3 14.8100 ppm H 2 S 12.2 13.9 13.9 16.2After diffusion of H 2 S it was not found asignificant variation of carbohydrateconcentration in SMP sample and accordingto obtained result, effect of H 2 S onconcentration of carbohydrate in SMP is notcountable.Total Organic CarbonTOC, mg/l180160140120100806040200Effect of H 2 S on TOCD F G H ISampleNo H2S sparging25 ppm H2Ssparging50 ppm H2Ssparging100 ppm H2SspargingTOC has been measured for five set ofsamples and in two test results, TOC wasnot vary significantly when applied H 2 S insludge but other three set of results showthat TOC decreases with the increase of H 2 Sdiffusion. As total organic carbon play animportant role of in fouling resistance,application of H 2 S in feed water deduce theTOC and consequently help to increase thefilterability of membrane in MBRs.ConclusionThis research works have focused on theimpact of H 2 S diffusion on membranefouling in MBR operation. ZP, pH, SMP,eEPS have been used to investigate theeffect of H 2 S application on membranefouling behavior. From the obtained result,following specific conclusions could bedrawn out:1. Application of H 2 S in feed waterassisted significantly to decrease theconcentration of eEPS and SMP TOCin sludge. It was found other studiesthat eEPS and SMP have a vital roleto enlarge the fouling resistance(Liang et al., 2007). Therefore,diffusion of H 2 S enhancedpermeability of membrane in MBRoperation by limiting foulingresistance as it reduced theconcentration of eEPS and SMPcompounds.2. It is quite clear from obtained resultsthat pH of biomass increased withthe increase of H 2 S concentration.This elevated pH value facilitates theactivities of the livingmicroorganisms which couldimprove the whole treatmentperformance.3. ZP of virgin membrane is lower thanthe fouling membrane but the effectof H 2 S application on zeta potentialis dubious because it unevenlyvaried with different concentrationof H 2 S.Fig 9: TOC of SMP at different H 2 S conc6

List of AbbreviationsCOD - Chemical oxygen demandeEPS - Extracted extracellular polymericsubstancesHRT - Hydraulic retention timeMBR - Membrane bioreactorPVDF - Polyvinylidene fluorideSMP - Soluble microbial productsTOC - Total organic carbonZP - Zeta Potentialfouling in MBRs (MBRs) caused by membranepolymer materials, Journal of MembraneScience, 280, 911–919http://nition.com/en/products/zeecom_s.htm,(2007, June 10)AcknowledgementsI would like to thank to University of NewSouth Wales, Australia for assistance andmaterials support, and to Lund University,Sweden for partial financial and sincereacademic support to carry out this researchwork.ReferencesHuber S. A. (1998). Evidence for membranefouling by specific TOC constituents,Desalination 119, 229-234.Judd, S. (2006). The MBR Book: Principles andApplications of Membrane Bioreactors in Waterand Wastewater Treatment, Elsevier, Oxford.Le-Clech, P., Chen, V. and Fane, T.A.G. (2006).Fouling in MBRs used in wastewater treatment,Journal of Membrane Science 284, 17–53.Lesjean, B. and Judd, S. (2007). MBR Network,Facts on MBR tech, (online)Available:http://www.mbr-network.eu/mbrprojects/index.php(2007, February 04).Liang, S., Liu, C. and Song, L (2007). Solublemicrobial products in MBR operation:Behaviors, characteristics, and fouling potential,Water Research 41, 95 – 101.Lindquist, A. (2003). About Water Treatment,Helsingborg, Sweden : Kemira Kemwater.Meng, F., Zhang, H., Yang, F., Zhang, S., Li, Y.and Zhang, X. (2006). Identification of activatedsludge properties affecting membrane fouling insubmerged MBRs, Separation and PurificationTechnology 51, 95–103Yamato, N., Kimura, K., Miyoshi, T. andWatanabe, Y. (2006). Difference in membrane7