Trans in d Ceram Soc 2014

VOL. 73 (3) JULY SEPTEMBER 2014 1

Trans. Ind. Ceram. Soc., vol. 73, no. 3, pp. 1-8 (2014). 2014 The Indian Ceramic SocietyISSN 0371-750X (Print), ISSN 2165-5456 (Online)http://dx.doi.org/10.1080/0371750X.2014.882246

Preparation and Characterization of Single LayerUltra Filtration Alumina Membrane Directly over

Porous Clay-Alumina Tubular and CapillarySupport for Textile Effluent Treatment

Sandeep Sarkar*, Sourja Ghosh, Priya Banerjee, Andr Larbot1, Sophie Cerneaux1,Sibdas Bandyopadhyay and Chiranjib Bhattacharjee2

CSIR-Central Glass and Ceramic Research Institute196, Raja S. C. Mullick Road

Jadavpur, Kolkata 700 032, India

2Department of Chemical Engineering, Jadavpur Universi ty,Kolkata 700032, India

1Institut Europen des Membranes, UMR 5635 ENSCM-CNRS-UMII,Place Eugene Bataillon CC47 34095 Montpellier Cedex 5, France

*Corresponding author; e-mail: [email protected]

[MS received August 21, 2013; Accepted January 07, 2014]

Clay-alumina porous tubular supports of 8/6 mm for outside diameter/inside diameter (OD/ID) were prepared byextrusion and sintering at 1450oC for 110 min to obtain an average pore size and porosity of 1 m and 40% respectively.Slip casting slurry was prepared by peptization of boehmite (AlO(OH)) powder with 1.5 N nitric acid and mixing itwith -Al2O3 powder of 40 nm particle size along with PVA and PEG in determined weight percentage. Clay aluminaporous support tubes of 3.1/1.9 and 8/6 mm (OD/ID) were coated with the slip casting slurry and sintered at 550oCfor 60 min to obtain -alumina membrane directly over the supports in a single step. The membrane layer preparedover both the support tubes were without any intermediate layer. The membranes presented a pore diameterdistribution centered at 7.0 nm on 3.1/1.9 mm (OD/ID) and 5.5 nm on 8/6 mm (OD/ID) supports, which was inultrafiltration (UF) range. The prepared UF membranes were studied for treatment of bio-treated synthetic and realtextile effluent in the side stream mode. About 98% decolorization of synthetic dye solution and 95.6% colourremoval of real textile effluent were obtained by the combined effect of UF membrane and bioreactor.

[Keywords: Crack-free, Single layer, Ultrafiltration, Clay-alumina, Support tube, Textile effluent]

IntroductionCeramic membrane technology is known for its

applications in various fields such as chemistry, food,biotechnology, wastewater treatment,1 gas transportsystem2 and membrane bioreactor (MBR).3 Ceramicmembrane has an edge over the polymeric membranesdue to the intrinsic properties of a ceramic material,4 i.e.high resistance to chemical abrasion, bio-inertness,excellent thermal stability, high-pressure resistance, longlife time5, 6 and stability over a wide range of pH.7

In spite of these advantages, ceramic membrane didnot find too much industrial applications due to its complexpreparation process and high cost as compared to theapplication of polymeric membrane.8 This has raisedinterest to prepare defect free ceramic ultrafiltration (UF)membrane at low cost.

Unlike polymeric membrane, microporous UF ceramicmembrane is prepared in a multilayered structurewith progressively decreasing pore sizes from the porousouter support to the inner microporous filtration layer.9, 10These membranes are being prepared by depositing

inorganic materials viz. alumina,1117 silica,18, 19 zirconia20or titania21, 22 through sol-gel process23 over a strongmacroporous tubular support prepared from high purityalumina, usually sintered at a high temperature superiorto 1500oC.24, 25 Attempts have been made to prepare lowcost support with natural minerals like clay along withalumina.24 However, the complexity and cost of preparingUF membrane over the support remains high. Thepreparation of ceramic UF membrane is a multi stepprocess. Several membrane layers, each with a particularpore size, are prepared one over another to reduce thepore size and achieve the required UF membrane. As aresult, processing of ceramic UF membrane becomescomplex and more expensive than processing polymericmembrane. Due to this, industrial applications of UFceramic membrane are limited, which otherwise mayeffectively be used in wastewater treatment includingvarious other separation applications, viz. MBR, gasseparation, water purification, contactor, etc consideringits advantages over polymeric membrane.47 Industrialwastewater, such as textile effluent, is usually treated byconventional activated sludge processes (CASPs), whichinvolve the natural biodegradation of pollutants byheterotrophic bacteria (i.e. activated sludge) in aeratedbioreactors. The treatment efficiency is usually limited forthe difficulties in separating suspended solids (SS).Ceramic UF membrane bioreactor has been used for

2 TRANSACTIONS OF THE INDIAN CERAMIC SOCIETY

domestic wastewater treatment.26 Ceramic membraneremains chemically stable in a wide range of pH andtemperature.7 A MBR system associated with UF ceramicmembrane filtration is the most appropriate for treatmentof wastewater since it makes the in situ cleaning ofmembrane easy and convenient as these inorganicmembranes highly resist chemical abrasion andbiodegradation as stated earlier.

The objective of this work is to prepare a cost effectivecrackless -alumina UF membrane directly over porousclay-alumina capillary support of 3.1/1.9 mm (OD/ID),precisely for the first time and 8/6 mm (OD/ID) tubularsupport without any intermediate layer by newly designedcoating slurry. The performance efficiency of crossflow UFmembrane bioreactor (UMBR) with the prepared tubularUF membrane in comparison to capillary UF membranewas studied for textile effluent reclamation.

The first step consisted of preparation of single layer-alumina UF membrane over a low cost porous clay-alumina tubular support and capillary tubes. Method ofpreparation of capillary tubes has been discussedelsewhere.24 The tubular support was prepared byextrusion of ceramic paste, composed of clay and aluminapowder, and then sintering at a temperature below thesintering temperature of pure alumina porous support(1700oC).25 Newly developed slip casting slurry wasprepared from boehmite sol and -alumina powder alongwith organic additives for coating both the tubular andcapillary porous supports by dip coating technique. Thecoated tubes were then fired to obtain -alumina UFmembrane over the porous supports without anyintermediate layer. In the second step, the prepared UFmembranes were characterized and their applicability wasinvestigated by treating textile effluent in a side-streammembrane bioreactor process.

ExperimentalRaw Materials

Alpha alumina (-Al2O3; 99% purity), with a meanparticle size of 6 m was purchased from Hindalco IndiaLimited (Birla, India). Kaolin clay (China), mainly composedof kaolinite (Al2O3.2SiO2.2H2O), with a mean particle sizeof 7 m and an organic binder, Methocel (Dow ChemicalCompany), were used along with the alumina in preparationof porous capillary support tube as described elsewhere.24

Boehmite AlO(OH) powder (74% purity) with a specificsurface area of 250 m2.g1 (Pural SB, SASOL, Germany)and -Al2O3 with mean particle size of 40 nm (Presi, France)were used as starting raw materials for preparation of UFcoating slurry. Polyvinyl alcohol (PVA) (RHODOVIOL25/140) was used as binder and polyethylene glycol with300 Daltons molecular weight (Merck, India) was used asplasticizer along with nitric acid (70% purity, Merck, India) aspeptizing agent in the slurry preparation. Dolapix C64 wasused as dispersing agent for dispersing -Al2O3 powder.

Polyethylene glycol (SRL, India) of various molecularweights, starting from 8000 to 35000 Da was utilized to

execute the molecular weight cutoff (MWCO) experimentof the membrane. Permeate and feed concentrations weredetermined by liquid phase chromatography using anoptical differential refractometer, Waters R 401.

Stock solutions of commercially used textile reactivedyes were prepared in distilled water and autoclaved at121oC and 1 atm for 15 min. Reactive yellow dye suppliedby M/s Clariant Chemicals India Ltd was selected as amodel dye in this study. All the culture media, organic andinorganic compounds, and reagents used were ofanalytical grade obtained from Merck, Mumbai, India.

Preparation of Tubular Support TubesThe ceramic paste was obtained by mixing clay and

alumina raw powders in 55:45 ratio along with organicadditives and water.24 The paste was extruded by singlescrew extruder using an extrusion die leading to greentube with dimensions of 10/7 mm (OD/ID). The extrudedtube was then dried and fired in air at 1450oC for 100 minas per the schedule illustrated in Fig. 1 to obtain 8/6 mm(OD/ID) tubular porous supports.

25oC

300oC

2oC/min 60 min

25oC

1450oC

1000oC

Free cooling

2oC/min

2oC/min

5 min

100 min

Fig. 1 Firing schedule for 8/6 mm (OD/ID) tubular support

Preparation of Slip Casting SlurryThree types of slip casting slurries, namely, -alumina

slip (S1), boehmite sol slip (S2) and mixture of boehmitesol and -Al2O3 slip (S3), were prepared to coat the poroustubular support for comparative study.

S1 slurry was prepared by dispersing -Al2O3 powderof 40 nm average particle size in Dolapix C64 and watersolution (0.01 wt% in 18 M distilled water) by ultrasonicbath for 15 min, where Dolapix acted as dispersant. PVAsolution (12 wt%; binder and pore former) and PEG(plasticizer) were added to the dispersed solution by slowstirring (100 rpm) in a magnetic stirrer for 30 min to obtaina well dispersed -Al2O3 slip casting slurry.

The second slurry, S2, was prepared from boehmitesol (-AlOOH sol). The sol was prepared by mixing 10 wt%boehmite (-AlO(OH)) powder in 18 M distilled water byslow stirring and adding 0.5% nitric acid of 1.5 N strengthuntil a translucent sol was obtained. An average sol particlesize of 14 nm (approx) was obtained after peptization withnitric acid. The sol was then mixed with PEG 300 and PVAsolution and mixed for 30 min by slow stirring (100 rpm) toobtain a uniform slip casting slurry from beohmite sol.

VOL. 73 (3) JULY SEPTEMBER 2014 3

The third slurry, S3, was prepared by adding -Al2O3powder in boehmite sol (10 wt%) and stirring (150 rpm) itfor 10 min for even dispersion of powder in the sol.Thereafter, PVA solution and PEG were added into themixture and stirred in a conical flux for 10 min at 100 rpmto minimize bubble formation. Table I illustrates thecomposition of three different slip casting slurries forcoating porous support tube.

Table I : Composition of different slip castingslurries in weight percentage

Sl.no.

S1 S2 S3(-alumina) (boehmite sol) (mixture)

1 Boehmite sol 66 64(10 wt%)

2 - alumina 1.5 2

3 Dolapix C64 solution 59.5 (0.01 wt%)

4 PVA solution 35 30 30(12 wt%)

5 PEG 4 4 4

Component Coating slurry

Preparation of UF Capillary and Tubular MembraneClay-alumina porous tubular and capillary support

tubes were coated with the prepared UF coating slurry bydip coating method. It is very difficult to coat the insidewall of a porous tube with the viscous coating slurry byconventional dip coating method especially with smallerinner diameter such as 2 mm or less in case of capillarysupport tube. Proper masking of the outer surface beforedipping the tube into the coating slurry is difficult for suchsmall diameter of capillary tube. Moreover masking mayhinder the required capillary suction pressure inside thetube wall for cake formation during the tube dipping intothe slurry. Pouring the solution from the top to bottom ofthe capillary tube may cause formation of unwantedbubbles along the inner wall, which may form defects afterfiring. Therefore, the tubes of 200 mm length and 3.1/1.9and 8/6 mm (OD/ID) were coated by introducing the slurryfrom bottom to top according to the communicating vesselsprocess based on the principle of inverted siphon, i.e. liquidflows from higher to lower level due to lower potentialenergy of lower location. The technique eliminated theabove said tube dip coating problems. Another advantageof this technique is to trap the unwanted denser particleof the slurry, such as undissolved coating powder, binderor other organic additives which may create defects in themembrane after firing. The coating time for both the supporttubes was 2 min. The tubes were then dried for 24 h atroom temperature before firing at a temperature dependingon their composition. The porous tube, coated with S1,

S2 and S3 slurries, were fired at 700o, 500o and 550oCrespectively. Figure 2 illustrates a typical firing scheduleof porous tube coated with S3 slurry.

25oC 60 min

200oC 1oC/min

25oC

550oC

Natural cooling 1

oC/min

90 min

Fig. 2 Firing schedule applied to coated capillary and tubularsupport to prepare UF membrane with S3 slurry

Membrane CharacterizationStatic Characterization

The UF 3.1/1.9 mm (OD/ID) and 8/6 mm (OD/ID)membranes were characterized by different techniques.The mechanical resistance of the prepared tubular supportwas measured by three-point bending method (universalLLOYD Instruments, LRX apparatus) using a span of30 mm and a cross-head speed of 10 mm.min1. The poresize of the support was examined by mercury porosimetry(Micromeritics autopore) whereas the UF top membranelayer pore size distribution was determined by nitrogensorption (Quantachrome Corp.) and the pore diameter wasestimated by the BJH (Barrett-Joyner-Halenda) method. Thestructural morphology, surface quality and thickness of thesupport and UF membrane layers were examined byscanning electron microscopy (JEOL Microscope, Japan).

Dynamic CharacterizationsAll the dynamic characterizations were performed on a

laboratory experimental setup using a 200 mm longmembrane of both 3.1/1.9 mm (OD/ID) capillary and8/6 mm (OD/ID) tubular supports. Determination of purewater permeability of the membranes was performed withdeionized water (18 M.cm). The molecular weight cutoff(MWCO) of the membranes was determined usingsolutions containing polyethylene glycol with molecularweights of 8000 to 35000 Da. The concentration of eachsolution was fixed at 103 mol.L1 and all the experimentswere carried out at the room temperature (25o to 35oC) ata fixed pressure of 5 bar, and a solution velocity of 1 cm.s1for both the membranes. The retention rate (R) of eachsolute was estimated by the classical relation below:

R(%) = 100(1 Cp/Cf)

Cp and Cf being the solute concentration in the permeateand in the feed solution, respectively.

Application of the UF Membrane in MBRThe prepared UF membranes were applied in MBR for

waste water treatment. Activated sludge was collected froma common effluent treatment plant in sterile collection

4 TRANSACTIONS OF THE INDIAN CERAMIC SOCIETY

bottles. Biodegradation of textile effluent was studied intwo different systems: one was synthetic dye solution ofreactive yellow dye and the other was coloured textileeffluent collected from a local dyehouse. In the first step,batch experiments were run at increasing concentration ofthe synthetic dye solution as well as the industrial effluentwith variable pH in order to determine the optimumconditions of attaining the maximum bacterial activity. Inthe second step, the untreated effluent without dilution wasadded directly to the bioreactor in order to study theacclimatization time of the microorganisms. For thesynthetic dye solution, an initial concentration of 50 mg.L1was selected for this purpose.

The pH range suitable for microbial growth wasdetermined, and the final pH was adjusted between 6 and10. Bacterial growth was measured at 620 nm by aspectrophotometer.27

To measure decolourization of dye solution and effluentduring biological treatment, sampling was done at 12 hintervals from the inoculated synthetic solution as well asfrom textile effluent. The samples were clarified bycentrifuging at 5000 rpm for 4 min in order to prevent theabsorbance interference from the cellular or othersuspended debris. Uninoculated culture media with andwithout added dyes were used as negative controls. Thedecolorization efficiency of different isolates wasexpressed as:

Decolourization (%) = (A0A)/A0 100

where A0 is the initial absorbance and A is the absorbanceof medium after decolourization at the max (nm) of eachdye. The max was determined by scanning the solution inthe entire ultraviolet-visible (UV-Vis) range and identifyingthe wavelength corresponding to maximum absorption.The clarified samples from the decolourization media wereused for determining the possible changes in theabsorption spectra of the dye in the UV-Vis range againsta baseline defined by the absorbance of clarified samplesfrom dye free media.28

The prepared UF membranes were utilized to purifythe bio-treated synthetic and real textile effluents in thehome-made pilot unit. Figure 3 illustrates a schematicdrawing of the home-made pilot unit for treating the bio-treated textile effluents. The flow rate and the pressurewere independently controlled by regulating the speed(rpm) of the pump motor by variable frequency drive andintroducing nitrogen gas to the feed tank through nitrogeninlet valve. The flow rate and pressure were maintainedat 1.5 L.min1 and 2 bar, respectively during all filtrationexperiments. Permeate samples were collected at specificintervals and analyzed in the spectrophotometer to observethe decolourization efficiency of the MBR process.

The colour concentration of the treated and untreatedsolutions were estimated by measuring the optical densitycorresponding to max of 418.1 and 515 nm for reactiveyellow dye solutions and textile effluent respectively, witha spectrophotometer (Cary 50 Bio, UV-visiblespectrophotometer of Varian).

N2 Cylinder Pressure gauge

Positive displacement Pump

N2 Cylinder

Feed Tank

Membrane

Permeate Inlet Pressure Gauge

Outlet Pressure Gauge

Variable frequency drive for Pump-motor rpm control

Fig. 3 Schematic diagram of the home-made pilot unit

During the experimentation, analyses were carried outon the most common control parameters of wastewatertreatment on raw textile effluents and synthetic dyecontaining effluents, biotreated effluents, and permeatesamples of ultrafiltration. The chemical oxygen demand(COD) and total suspended solids (TSS) contents weredetermined according to the standard methods of APHA.Total Kjeldhal nitrogen (TKN) concentration was evaluatedby a TKN analyzer (Kelplus, Pelican, India) and totalorganic carbon (TOC) was measured by a TOC analyzer(Shimadzu, Japan). pH, dissolved oxygen (DO) and totaldissolved solids (TDS) concentrations were measuredusing respective meters (YSI, Model 55, USA). The numberof coliform organisms present was enumerated by the mostprobable number (MPN) test. The MPN was estimated bydetermining the number of tubes in each set of 5 tubes thatshows gas formation following the incubation period. Thistest was carried out following APHA defined method.29

Results and DiscussionClay-Alumina Support

Clay was used as a partial substitute of alumina forthe preparation of membrane support, which led to dualbenefits: Firstly, clay being cheaper than alumina, it reduces the

raw material cost. Secondly, clay contains fluxing impurities, which forms

a low melting phase for bringing down the sinteringtemperature to below 1500oC, thereby reducing theprocessing cost.Tubular clay-alumina support (Fig. 4) prepared in this

work obtained a porosity of about 40% and an averagepore size of 1 m which is at par with the pore characteristicof the capillary support.24 Figure 5 illustrates the pore sizedistribution of tubular support determined by mercuryintrusion porosimetry. Porous tubular support preparedfrom pure alumina sintered at 1600oC has a pore size ofmore than 2 m (average).24 It is difficult to prepare UF

N2 inletvalve

VOL. 73 (3) JULY SEPTEMBER 2014 5

membrane directly over such support without anyintermediate layer (micro filtration membrane) due to itslarge pore size. Introduction of silica (clay) in the aluminamatrix of clay-alumina tubular support may reduce itsaverage pore size to 1 m, facilitating the preparation ofUF membrane directly over the support. The mechanicalresistance of the tubular support was found to be morethan 65 MPa by three-point bending method, which issufficient considering their use for ultrafiltration applicationunder pressure lower than 10 bar.

Fig. 4 Photographs of 8/6 and 3.1/1.9 mm (OD/ID) clay-aluminasupports

Alumina Membrane over Clay-Alumina SupportThe support tube was dip coated with S1, S2 and S3

slurry and fired as discussed earlier. Figure 6 illustratesthe SEM photograph of S1 membrane prepared overporous support with S1 slurry. Though a membranethickness of 8 m (approx) was formed over the support(Fig. 6a), a top view of the membrane (Fig. 6b) revealsthe occurrence of microcracks on the S1 membranesurface. Hence the membrane is not suitable for UFapplications. Result was more severe for S2 membraneprepared over porous support with S2 slurry. Themembrane (S2) was cracked as evident from the SEMphotograph (Fig. 7). A crackless membrane was preparedwith the S3 slip casting slurry. Probable reason for theabsence of cracks in S3 membrane is the presence of-Al2O3 powder along with boehmite sol which may act asa seeding agent during heat treatment.

-0,1

0

0,1

0,2

0,3

0,4

0,5

0,6

0,01 0,1 1 10 100

Pore size diameter / m

Log

Diffe

rent

ial in

trusio

n of

Hg

PORE DIAMETER (m)

LOG

DIF

FER

ENTI

AL

INTR

USI

ON

OF

Hg

Fig. 5 Pore size distribution of 8/6 mm (OD/ID) support tube

(b)

Fig. 6 SEM picture of S1 membrane over porous clay-aluminasupport: (a) cross section, (b) top view exposing the microcracks

(a) (b)

Fig. 7 Microcracks on S2 membrane over porous clay-aluminatubular support (SEM picture): (a) cross section, (b) top view

Single Layer UF MembraneBoth the capillary and tubular supports were coated

with S3 slip casting slurry. After sintering at 550oC for 90min the crackless -Al2O3 layer (S3 membrane) presenteda pore diameter in the nanometer range (Figs. 8a and 8b)and an average thickness of 7.3 m (Fig. 9a) for capillaryand 15.8 m for tubular supports (Fig. 9b). The variation

Fig. 8 SEM images of (a) defect-free UF -alumina membraneover macroporous clay-alumina tubular support and(b) UF -alumina membrane revealing the void (pore) in thenanometer range

(a)

(a) (b)

Fig. 9 SEM cross sections of UF -alumina membrane overmacroporous (a) 3.1/1.9 mm (OD/ID) capillary and (b) 8/6 mm (OD/ID)tubular clay-alumina support revealing its membrane thickness

(b)(a)

8.02 m

7.10 m

12.3 nm

7.22 nm6.94 nm

7.34 m 15.8 m

6 TRANSACTIONS OF THE INDIAN CERAMIC SOCIETY

in the formation of UF membrane layer thickness found incapillary and tubular clay-alumina supports with the samecoating slurry and process (including the coating time)depends on the thickness of the support. This may bedue to the higher capillary forces occurring in a larger porevolume concerning the tubular support than the capillarysupport. Higher capillary force increases the formation ofcake thickness during the dip coating process. The-alumina membrane layer obtained by the sol-gel processpresents a pore diameter distribution centered at 7.0 nmfor capillary and 5.5 nm for tubular supports, respectively(Fig. 10). These values, determined by nitrogen adsorption-desorption isotherms, confirm the prepared membrane asa low UF membrane. The resulting crack-free single layersUF membrane were achieved by adjusting the slipcomposition, particularly the boehmite sol to -aluminapowder in 32:1 weight ratio and maintaining the coatingtime of 2 min for both clay-alumina supports with similarpore morphology.

Pure Water Flux and MWCOThe pure water flux of the support was found to be

different for different support configurations as shown inFig. 11. At 5 bar pressure, the pure water flux of capillaryand tubular supports were 2514 and 2104 L.m2.h1,respectively. This small variation in flux for the two supportsmay be explained by the fact that their supports havedifferent wall thicknesses, which leads to differentmembrane resistances. The capillary support has higherflux than the tubular support due to its lower wall thickness.Similar trend has been found in the flux characteristics ofthe UF membrane also (Fig. 12). The pure water flux ofthe UF capillary and tubular membrane was found to be107 and 95 L.m2.h1 at 5 bar pressure, respectively.

The MWCO of both UF capillary and tubularmembranes were evaluated at 5 bar pressure using ahome-made pilot by administrating nitrogen gas forincreasing the pressure. The feed flow rate was maintainedat 1.5 L.min1. PEG solution (1 g in 1 L water) of differentmolecules was utilized as the polymer solution to obtain

0500

1000150020002500300035004000

0 2 4 6 8

Pressure (Bar)

Flux

(L/m

2 /hr)

3.1/1.9 mm (OD/ID) support

8/6 mm (OD/ID) support

Fig. 11 Pure water flux of the 3.1/1.9 (OD/ID) capillary and6/8 mm (OD/ID) tubular clay-alumina

PRESSURE (bar)

FLU

X (L

.m2

.h1

)

0

20

40

60

80

100

120

140

0 1 2 3 4 5 6 7Pressure (Bar)

Flux

(L/m

2 /h)

8/6 mm (OD/ID) UF membrane

3.1/1.9 mm (OD/ID) UF membrane

Fig. 12 Pure water flux of the 3.1/1.9 (OD/ID) capillary and6/8 mm (OD/ID) tubular UF membrane

PRESSURE (bar)FL

UX

(L.m

2.h

1)

the cutoff. The cutoff of both the UF membrane, determinedat 90% rejection, was found to be 35 KDa (Fig. 13) atroom temperature. The home-made pilot unit (Fig. 3) wasalso utilized to carry out membrane flux, MWCO and MBRside stream filtration studies.

0102030405060708090

100

0 10000 20000 30000 40000

Mol. Wt. of PEG

Rej

ectio

n (%

)

3.1/1.9 mm (OD/ID) UF membrane

8/6 mm (OD/ID) UF membrane

Fig. 13 PEG rejection percentage of UF tubular membrane over3.1/1.9 (OD/ID) capillary and 8/6 mm (OD/ID) tubular support

Textile Effluent Treatment by UF MembraneTable II illustrates the characterization of untreated

textile effluent samples and treated samples using boththe UF membranes (capillary and tubular) basedbioreactor. Samples 1 to 4 denote simulated textile effluentprepared with synthetic reactive yellow dye solutions

REJ

EC

TIO

N (%

)

PEG MOL. WT. (Da)

10 100 1000-0.002

0.000

0.002

0.004

0.006

0.008

0.010

0.012

0.014

0.016

0.018

0.020

Des

orpt

ion

Dv(

log

d) [c

c/g]

Diameter ()

3.1/1.9mm (OD/ID) membrane tube 8/6mm (OD/ID) membrane tube

Fig. 10 Nitrogen adsorption-desorption isotherms of UF-alumina membrane on 3.1/1.9 (OD/ID) capillary and 8/6 mm(OD/ID) tubular supports

DIAMETER ()

DE

SOR

PTIO

N D

v (lo

g d)

(cm

3 .g1

)

VOL. 73 (3) JULY SEPTEMBER 2014 7

Table II : Characterization of various samples obtained from the MBR study

having initial dye concentration of 50 mg.L1 with a saltconcentration of 10 g.L1, biotreated synthetic dye solution,permeate of biotreated synthetic dye solution collectedusing UF 8/6 tubular and 3.1/1.9 capillary membranes,respectively. While samples 5 to 8 denote untreated textileeffluent, biotreated effluent, permeate of biotreated effluentcollected using UF tubular and capillary membranes,respectively.

The results indicated that 98% decolorization ofreactive yellow dye solution was achieved starting froman initial dye concentration of 50 mg.L1 in the membranebioreactor treatment.

For textile effluent, about 95.6% colour removal wasobtained by the combined effect of UF and bioreactor.Permeate samples in both the cases were free fromcoliforms. The process resulted in considerable removalof COD and TSS. The separation efficiency of both themembranes was found acceptable for reuse purpose. Theflux was found to be slightly better for 3.1/1.9 capillarymembrane than 8/6 tubular membrane for both thesynthetic dye solutions and the real effluent. As statedearlier, this may be due to the higher membrane resistanceof 8/6 tubular membrane owing to higher membranethickness than the 3.1/1.9 capillary membrane.

Batch scale setup of ultrafiltration coupled membranebioreactor was run for 6 days till collection of adequatequantity of permeate for chemical analysis. During this

operation, flux was measured at every 12 h. A decline influx was obtained after 48 h which may have occurreddue to fouling or clogging by sediments. Keeping otherparameters unaltered, flux levels were restored byproviding a backwash whenever necessary.

Conclusions The 8/6 mm (OD/ID) porous tubular support was

prepared by extrusion of a ceramic paste made withclay and alumina powders after firing at 1450oC. Thesupport had an average pore size of 1 m and 40%porosity, which were similar to the reported values ofclay-alumina capillary supports.

The defect-free UF -alumina membrane layer wassuccessfully deposited directly over the clay-aluminacapillary and tubular support. The UF membrane layerwas prepared from S3 slip casting slurry composed ofbeohmite sol and -alumina powder inside the supporttube by communicating vessels dip coating techniquein one-step.

Though the pore size and porosity of both the capillaryand the tubular support were similar, coating thicknessof the UF ceramic membranes was different due todifferent wall thicknesses, which generated differentcapillary forces in cake formation during coating of thetubes.

Sl. Samples MPN Residual COD DO pH Turbidity TSS TDS TOC TKN Membraneno. index/ colour (mg.L1) (mg.L1) (NTU) (mg.L1) (g.L1) (mg.L1) (mg.L1) flux

100 mL (%) (L.m2.h1)(P = 2 bar)

1 Synthetic dye 2 100 2750 2.3 7.6 9.8 73.3 10.6 12.0 1.5 solution (simulated)

2 Biotreated synthetic 920 96.8 400 4.6 7.4 2.3 400 10.2 17.6 1.3 dye solution

3 Permeate of 2 97.2 60.6 6.3 7.5 0.8 53.3 9.9 9.5 0.9 78.6bio-treated syntheticdye solution by UFtubular membrane

4 Permeate of 2 98 50 7.2 7.2 0.6 43.3 9.9 8.1 0.8 84.5bio-treated syntheticdye solution by UFcapillary membrane

5 Real textile effluent 46 100 3700 1.7 11.6 44.3 1593.3 79.5 170.0 2.4

6 Biotreated 1600 7.3 500 3.9 7.3 16.6 2103.4 12.3 14.2 1.3 real effluent

7 Permeate of 2 6.8 99.4 5.9 7.2 0.8 72.8 12.3 8.3 0.7 50.9bio-treated realeffluent by UFtubular membrane

8 Permeate of 2 4.4 65.7 6.8 7.2 0.5 62.6 12.1 7.9 0.7 58.8bio-treated realeffluent by UFcapillary membrane

8 TRANSACTIONS OF THE INDIAN CERAMIC SOCIETY

The membrane obtained was a -alumina membranewith a MWCO of 35 KDa directly over low cost clay-alumina capillary and tubular supports. This eliminatesthe requirement of intermediate layer between thesupport and the actual membrane, which furtherreduces the cost and complexity in development of UFceramic membrane, which can be utilized in variousgas and liquid filtration applications. The next part ofthe work will be devoted to the preparation of costeffective multichannel UF membrane for filtrationapplications.

The UF capillary and tubular membranes wereeffectively used in MBR process for textile wastewaterpurification; UF capillary membrane obtained better fluxdue to smaller membrane thickness.

References1. K. K. Cham and A. M. Brownstein, Am. Ceram. Soc. Bull.,

70, 703-707 (1991).

2. R. S. A. de Lange, K. Keizer and A. J. Burggraaf, J. Membr.Sci., 104, 81-100 (1995).

3. C. H. Xing, E. Tardieu, Y. Qian and X.-H. Wen, J. Membr.Sci., 177, 73-82 (2000).

4. M. R. Weir, E. Rutinduka, C. Detellier, C. Y. Feng, Q. Wang,T. Matsuura and R. LeVan Mao, J. Membr. Sci., 182, 41-50(2001).

5. K. A. DeFriend and A. R. Barron, J. Membr. Sci., 212, 29-38(2003).

6. B. Elyassi, M. Sahimi and T. T. Tsotsis, J. Membr. Sci., 316,73-79 (2008).

7. D. A. White and M. Asaadi, Fouling Behavior in InorganicTubular Membranes, p. 143 in: Proc. Int. Technical Conf.Membrane Separation Process, Brighton, UK (1989).

8. S. Khemakhem, A. Larbot and R. Ben Amar, Ceram. Int.,35, 55-61 (2009).

9. K. Li, Ceramic Membrane for Separation and Reaction,p. 1, John Wiley&Sons, Chichester, UK (2007).

10. T. Van Gestel, C. Vandecasteele, A. Buekenhoudt,C. Dotremont, J. Luyten, R. Leysen, B. Van der Bruggenand G. Maes, J. Membr. Sci., 207, 73-89 (2002).

11. A. F. M. Leenaars, K. Keizer and A. J. Burggraff, J. Mater.Sci., 19, 1077-1088 (1984).

12. A. F. M. Leenaars and A. J. Burggraaf, J. Colloid Interf. Sci.,105, 27-40 (1985).

13. M. A. Anderson, M. J. Gieselmann and Q. Xu, J. Membr.Sci., 39, 243-258 (1988).

14. T. Okubo, M. Watanabe, K. Kusakabe and S. Morooka,J. Mater. Sci., 25, 4822-4827 (1990).

15. T. Okubo, K. Haruta, K. Kusakabe and S. Morooka, J. Membr.Sci., 59, 73-80 (1991).

16. M. I. D. de Albani and C. P. Arciprete, J. Membr. Sci., 69,21-28 (1992).

17. A. Larbot, J. E. Fabre, C. Guizard and L. Cot, J. Membr.Sci., 39, 203-212 (1988).

18. A. Kaiser and H. Schmidt, J. Membr. Sci., 22, 257-268 (1985).

19. L. C. Klein and D. Gallagher, J. Membr. Sci., 39, 213-220 (1988).

20. J. Etienne, A. Larbot, A. Julbe, C. Guizard and L. Cot,J. Membr. Sci.,86, 95-102 (1994).

21. M. A. Anderson, M. J. Gieselmann and Q. Xu, J. Membr.Sci., 39, 243-258 (1988).

22. N. Seffaj, S. Alami Younsi, M. Persin, M. Cretin, A. Albizaneand A. Larbot, Ceram. Int.,31, 205-210 (2005).

23. X. R. Huang, G. L. Meng, Z. T. Huang and J. M. Geng,J. Membr. Sci., 133, 145-150 (1997).

24. S. Sarkar, S. Bandyopadhyay, A. Larbot and S. Cerneaux,J. Membr. Sci., 392-393, 130-136 (2012).

25. I. Voigt, G. Fischer, P. Puhlfur, M. Schleifenheimer andM. Stahn, Sep. Purif. Technol., 32, 87-91 (2003).

26. X. Wen, Tsinghua Sci. Technol., 5, 283-287 (2000).

27. A. Gopal, N. P. Shah and H. Roginski, Milchwissenschaft,51, 619-622 (1996).

28. S. Asad, M. A. Amoozegar, A. A. Pourbabaee, M. N.Sarbolouki and S. M. M. Dastgheib, Bioresour. Technol.,98, 2082-2088 (2007).

29. A. D. Eaton, L. S. Clesceri, E. W. Rice and A. W. Greenberg,Standard Methods for the Estimation of Water andWastewater , American Public Health AssociationPublications, Washington, USA (2005).