Ru

Wa

Bb

h

����

a

ARRAA

KGMPMR

1

udfdcl

0h

Journal of Hazardous Materials 235– 236 (2012) 128– 137

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials

j our na l ho me p age: www.elsev ier .com/ locate / jhazmat

emoval of phenol from synthetic waste water using Gemini micellar-enhancedltrafiltration (GMEUF)

enxiang Zhanga, Guohe Huanga,∗, Jia Weib, Huiqin Lia, Rubing Zhenga, Ya Zhoua

MOE Key Laboratory of Regional Energy and Environmental Systems Optimization, Resources and Environmental Research Academy, North China Electric Power University,eijing 102206, ChinaFaculty of Engineering and Applied Science, University of Regina, Regina, Saskatchewan, Canada S4S 0A2

i g h l i g h t s

Gemini surfactant micellar enhanced ultrafiltration was used to remove phenol.The effect of different hydrophilic head groups of surfactant was analyzed.SEM, ATR-FTIR and mercury porosimeter were applied to elucidate membrane fouling.Gemini surfactant had superior performance in comparing with conventional surfactant.

r t i c l e i n f o

rticle history:eceived 27 April 2012eceived in revised form 12 July 2012ccepted 14 July 2012vailable online 20 July 2012

eywords:emini surfactanticellar-enhanced ultrafiltration

henolembrane fouling

etention

a b s t r a c t

Comprehensive studies were conducted on the phenol wastewater ultrafiltration (UF) with thehelp of various concentrations of cationic Gemini surfactant (N1-dodecyl-N1,N1,N2,N2-tetramethyl-N2-octylethane-1,2-diaminium bromide, CG), conventional cationic surfactant (dodecyl trimethylammonium bromide, DTAB), anionic surfactant (sodium dodecyl sulfate, SDS) and nonionic surfactant((dodecyloxy)polyethoxyethanol, Brij35). A flat sheet module with polyethersulfone (PES) membranewas employed in this investigation. The effects of feed concentration (phenol and surfactant) on theretention of phenol and surfactant, permeate flux and membrane fouling by micelles were evaluated.The distribution coefficient (D), the loading of the micelles (Lm) and the equilibrium distribution con-stant (K) were also utilized to estimate the micellar-enhanced ultrafiltration ability for phenol. Scanningelectron microscope (SEM), Fourier transform infrared spectrometer with attenuated total reflectanceaccessory (ATR-FTIR) and mercury porosimeter were applied to analyze membrane surface morphology,

membrane material characteristics and membrane fouling for the original and fouled membranes. Basedon the above analysis, the performance of the selected Gemini surfactant was proved superior in thefollowing aspects: retention of phenol/surfactant (peak value is 95.8% for phenol retention), permeateflux and membrane fouling with respect to other conventional surfactants possessing equal alkyl chainlength. These results demonstrated that CG surfactant with exceptional structure has favorable prospectsin the treatment of phenol wastewater by the micellar-enhanced ultrafiltration.

. Introduction

Phenolic compounds are regarded as one of the major and mostndesirable water pollutants. They often come from waste waterischarged from a variety of industrial sources (e.g., the manu-acture of papers, antioxidants, plastics and dyes). Phenol and its

erivatives are toxic to most mammals and aquatic life, and alsoause objectionable taste and odor in drinking water even at veryow concentrations. The effects of phenol molecules are even more

∗ Corresponding author. Tel.: +86 61772018; fax: +86 10 51971284.E-mail addresses: huang@iseis.org, hgh201222@gmail.com (G. Huang).

304-3894/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.jhazmat.2012.07.031

© 2012 Elsevier B.V. All rights reserved.

poisonous when they become chlorinated [1]. At present, vari-ous techniques have been developed to treat phenolic wastewaterincluding extraction, adsorption, chemical oxidation, UV oxida-tion and biological treatment, etc. [2]. However, these techniqueshave obvious deficiencies, such as low efficiency, high cost, inferiorselection and stringent running conditions, which limit their wideapplications [3]. Particularly, it is uneconomical to deal with largevolumes of wastewater with low concentration of contaminants bysuch methods [3].

Membrane technology has been suggested as a promising toolfor water treatment because of its compact size, high efficiencyand superior selectivity [4]. Membrane technology can be clas-sified as microfiltration (MF), ultrafiltration (UF), nanofiltration

ous M

(NmpaacfBsmfttbatcststhtmlrot

sEtw(tnCdbfm[tmtssp

stimcaaeBrsiofnp

W. Zhang et al. / Journal of Hazard

NF) and reverse osmosis (RO) based on their pore size. RO andF are two popular membrane techniques for separating smallerolecular weight of pollutant in the aqueous phase. However, less

ermeate flux and higher transmembrane characteristics of ROnd NF make the purifying process much more expensive. Therere additional disadvantages related to these methods, such asoncentrated polarization, cake formation and severe membraneouling, all of which extremely restrict their extensive use [5].ecause UF has high permeate flux and low differential pressure, ithows great promises for wastewater treatment compared to otherembrane methods. Nevertheless, the ordinary UF process is inef-

ective in removing organic pollutants due to a larger pore size thanhat of the organic molecule. Presently, micellar-enhanced ultrafil-ration is aiming to improve the performance of UF membranesy capturing small size pollutants into large micelles, which hasttracted considerable attention [6]. In micellar-enhanced ultrafil-ration, when a surfactant is added into polluted aqueous above itsritical micellar concentration (CMC), micelles begin to form andeize pollutants [2]. As a consequence, the micelles bearing pollu-ants tend to be intercepted by a proper UF membrane whose poreize is substantially smaller than that of micelles [7]. Therefore, fur-her treatment of the retentate stream containing pollutants withigh concentration is effortless and much more economic than thereatment of the feed stream directly. In the long run, because

icellar-enhanced ultrafiltration exhibits such characteristics asow energy consumption, high removal efficiency and a small spaceequirement, it is more appropriate for removing trace amountsf organic pollutants compared to common membrane separationechniques.

To this point in time, numerous efforts have been devoted to theelection of proper surfactants to remove corresponding pollutants.l-Abbassi et al. reported that the efficiency of UF treatment facili-ated by an anionic surfactant (SDS) for the removal of polyphenolsas desirable [5]; Fang et al. found that cetylpyridinium chloride

CPC) performed better for the removal of phenol than octadecylrimethyl ammonium bromide (OTAB) and cetyltrimethylammo-ium bromide (CTAB) [8]. Xu et al. studied the interaction betweenPC and phenol micelle and found that phenol molecules wereissolved into the water–micelle interface by the ion interactionetween C6H6O− and C5H5N+ [9]. In general, the cationic sur-actants are suitable to remove anionic metal ions and organic

atter, while the anionic surfactants are apt to remove cationic ions10]. Meanwhile, both the concentration and structure of surfac-ants could influence the efficient solubility of pollutants. However,

ost previous studies related to micellar-enhanced ultrafiltra-ion reported the efficiency of pollutants removal by conventionalurfactant systems [5,10–12]. Thus, exploring new and effectiveurfactant systems is significant for improving the efficiency andractical applicability of this potential technique.

Gemini surfactants are third generation surfactants that con-ist of two hydrophilic and hydrophobic head groups, connectinghrough a spacer group [13,14]. They have aroused considerablenterest due to their lower CMC values, minor Krafft point, and

uch greater efficiency in reducing surface tension than those oforresponding conventional surfactants. These surfactants show

very high bactericidal activity. Therefore, they have manifoldpplications in the fields of gene therapy, protein study, drugntrapment, soil remediation and enhanced oil recovery [14].esides, Gemini surfactants have low toxicity because of theeduced accessibility of the hydrophobic part in the stretchedpacer conformation which could limit the incorporation of Gem-nis surfactant into the cell membrane [15]. However, the effect

f micellar-enhanced ultrafiltration with the help of Gemini sur-actants contributing to the removal of organic compounds isot yet clear due to their unique structure and physiochemicalroperties.

aterials 235– 236 (2012) 128– 137 129

The objective of current work is to investigate the removalefficiency of phenol by micellar-enhanced ultrafiltration in thepresence of cationic Gemini surfactant and conventional sur-factants (anionic surfactant SDS, cationic surfactant DTAB andnonionic surfactant Brij35) with the same length of alkyl chain.A new method is also developed for phenol treatment, Geminimicellar-enhanced ultrafiltration (GMEUF). More detailed studieshave focused on: (1) analyzing the effect of different hydrophilichead groups of Gemini and conventional surfactants on the removalof phenol; (2) evaluating the enhanced removal efficiency of phe-nol by GMEUF; (3) elucidating the effects of feed concentration onretention, permeate flux and secondary resistance; and (4) com-paring the membrane fouling of different surfactants using severaltechniques including SEM, ATR-FTIR and mercury porosimeter. Theresults of this study will be useful to find out the operational effectand evaluate the viability of Gemini micellar-enhanced ultrafiltra-tion. At the same time, it is significant for providing support forexploring new and efficient surfactant systems for UF in wastewa-ter treatment.

2. Materials and methods

2.1. Materials

The cationic Gemini surfactant (CG), N1-dodecyl-N1,N1,N2,N2-tetramethyl-N2-octylethane-1,2-diaminium bromide was sup-plied by Chengdu Organic Chemicals Co. Ltd. Chinese Academy ofScience, with a purity of 98%. The nonionic surfactant (Brij35) withpurity 98% was obtained from Sigma–Aldrich. The cationic surfac-tant dodecyl trimethyl ammonium bromide (DTAB, analysis purity)and the anionic surfactant sodium dodecyl sulfate (SDS, analy-sis purity) were delivered by Tianjin Fushen and Kemel ChemicalReagent Company, China, respectively. The molecular structuresand properties of selected surfactants are given in Table 1. Phe-nol (Log Kw = 1.46, Sw = 8.3 g L−1 (20 ◦C)) [16] with analysis puritywas purchased from Beijing Chemical Reagent Company, China. Allreagents were used without further purification. Distilled waterwas used for solution preparation in all experiments.

2.2. Methods

2.2.1. Surface tension measurementsSurface tension (�) measurements were conducted by a model

BZY-2 surface tensionmeter using the Wilhelmy plate technique[17]. The concentration of the stock surfactant solutions was10.0 mM. Different concentration levels of surfactant solutionswere made and allowed to equilibrate for approximately 5 h beforethe measurement. Then the measurements were conducted untilconstant values indicating that equilibrium has been reached. Theaccuracy of measurements was within ±0.1 mN m−1. The CMC val-ues were defined by noting the inflection point in the plot ofsurface tension versus logarithm of surfactant concentration shownin supplementary data Fig. S1.

2.2.2. Micellar-enhanced ultrafiltration measurementsMicellar-enhanced ultrafiltration measurements were carried

out at 20 ◦C using a flat sheet module supplied by Xia MenTianquanxin Membrane Technology Co, Ltd., China. Polyethersul-fone (PES) flat sheet membranes with a molecular weight cut-offof 10 kDa and total effective areas 0.06 m2 were obtained fromAdvanced Membrane Corporation, America. The flow, head andpower of pump were 2 m3 h−1, 40 m and 1.1 kW, respectively. A

schematic diagram of the UF is presented in Fig. 1. In the prepa-ration of feed, phenol and surfactant solutions were mixed at therequired concentrations and stirred adequately to ensure that thesolutes were evenly dispersed before feeding to the membrane

130 W. Zhang et al. / Journal of Hazardous Materials 235– 236 (2012) 128– 137

Table 1The physicochemical properties of surfactants in experiment.

Surfactant Structure MW (g mol−1) CMCa (mM)

CG C12H25N+(CH3)2(CH2)2−N+(CH3)2C12H25·2Br− 614.67 0.8 [17]

Brij35 C12H25(OCH2CH2)23OH 1200 0.065 [18]SDS C12H25SO4Na 288.38 7.8 [7]DTAB C12H25(CH3)3NBr 308.4 15.0 [18]

msmPrcpptatfi2bw

2

tEtCaao

ftFrC

F(

a Error limits of CMCs are ±4%.

odule. The pressures and retentate flow rates were kept con-tant at 0.30 MPa and 5 L min−1. Then, the solution was fed to theembrane module. Retentate stream was returned to feed tank.

ermeate stream was collected into measuring cylinder, and theneturned to the feed tank to maintain a constant concentration. Pro-ess was stopped when total permeate streams reached 0.5 L. Theermeate fluxes were measured every 5 min. It was found that theermeate fluxes were almost constant after 0.5 h of operation. Afterhat, the retentate stream and permeate stream were collected tonalyze concentrations. The reported values were the average ofhree duplicate records. After each experiment run, tap water wasltered with 0.2 MPa to wash the exterior of membrane within0 min, and then the distilled water was used to rinse out the mem-rane at 0.30 MPa for 20 min. At last, the membrane permeabilityas recovered.

.2.3. AnalysisIn the synthetic solution of Brij35/phenol and SDS/phenol,

he concentration of Brij35, SDS and phenol were analyzed byxtinction Coefficient method [19,20] with UV-2102 PCS spec-rophotometer. In the synthetic solution of DTAB/phenol andG/phenol, the concentration of phenol was measured by UVbsorption at a wavelength of 270 nm. The concentrations of DTABnd CG were measured using a titrating method [21]. The viscositiesf surfactant solutions were measured by a viscometer (NDJ-5S/8S).

With the purpose of characterizing the microscopic effects byour types of surfactants on membrane fouling, scanning elec-

ron microscope (SEM, FEI QUANTA 200, FEI Company, USA),ourier transform infrared spectrometer with attenuated totaleflectance accessory (ATR-FTIR, Nicolet 560, Thermo Electronorporation, USA) and mercury porosimeter (AUTOPORE II 9220,

2

1

8

ig. 1. Schematic diagram of the ultrafiltration unit: (1) feed solution, (2) pump, (3) man7) retentate rotameter, (8) retentate.

Micromeritics, USA) were applied in this study. The measuredmembranes consisting of three layers-inner support layer, outeractive layer and intermediate porous connection layer were cut toseparate the flat-sheet active layer alone for the measurements bymercury porosimeter [22] and ATR-FTIR. Before the all analysis,membrane samples were dried by vacuum under 50 ◦C for 8 h toexamine the moisture.

3. Results and discussion

3.1. Critical micelle concentration properties of all selectedsurfactants

The surface tension isotherms for the studied individual surfac-tants (CG, DATB, SDS and Brij35) are illustrated in supplementarydata Fig. S1. The CMC values of all selected surfactants are deter-mined by the intersection point between the two fitted straightlines, the pre-CMC and post-CMC data [23]. They are complemen-tary with literature values, given in Table 1. Obviously, the CMCvalues of CG, as expected, are almost 18 orders of magnitude lowerthan DTAB with the same alkyl chain indicating the high efficiencyof micellization for Gemini surfactant. The spacer group of CGwith a chemical bonding effect weakens the electrostatic repul-sion between the hydration forces and polarity between layers [18],which enable the micellar molecules even more hydrophobic andhence micellization is more favorable with respect to conventional

surfactant. The lowest CMC value of Brij35 results in part from inex-istence of electric charge and electrostatic repulsive force. Anotherreason is the large size of hydrophilic groups of Brij35 which mayweaken the force between monomers with dissolvent.

3

4

5

67

ometer, (4) flat sheet module with PES membrane, (5) permeate, (6) manometer,

W. Zhang et al. / Journal of Hazardous M

012345640

50

60

70

80

90

100

CG 2CMC

CG 6CMC

CG 10CMC

R(%

)

[Phenol] (mM)

a

0123456

50

60

70

80

90

DTAB 2CMC

DTAB 6CMC

DTAB 10CMC

R (

%)

b

F

3s

3

oer

R

wi

iintocso

[Phenol] (mM)

ig. 2. Variations of the retention of phenol under different feed concentration.

.2. Effect of feed concentration on retention, permeate flux andecondary resistance for all selected surfactant systems

.2.1. Phenol retentionThe retention of an organic pullutant is the [7] main objective

f micellar-enhanced ultrafiltration. The effectiveness of micellar-nhanced ultrafiltration process is represented by percentageetention (R) as follows [8]:

= 1 − Cp

Cf(1)

here Cp and Cf denote the concentration of phenol or surfactantn the permeate and the feed stream, respectively.

Fig. 2 shows that the retention of phenol ascends with thencrease of feed surfactant concentration, but diminishes with thencrease of feed phenol concentration. As the concentration of phe-ol increases, the amount of phenol in the micelles increases untilhe solubility of phenol is saturated, thus leading to the decrease

f phenol retention. Simultaneously, the elevated surfactant con-entration enhances the number of micelles, giving rise to higherolubility of phenol. The worst rejection rate (19.2%) of phenol isbserved by Brij35 because nonionics are not able to form ionic

aterials 235– 236 (2012) 128– 137 131

bonding with phenol resulting in less effect on the solubilizationof phenol. It is clearly observed that the ionic surfactant on thephenol solubilization is superior to nonionic surfactant, which maycorresponds with the rule of similarity and intermiscibility [8].Specifically, the phenol solubilization efficiency by cationic sur-factant is better than anionic surfactant, which can be explainedby electrostatic interaction. It is easy to find that the conventionalcationic surfactant DTAB has better removal efficiency (averageretention 54.31% and 64.84%) at the low feed surfactant con-centrations (2 CMC and 6 CMC). Yet, the best retention (above95.8%) is obtained by the 10 CMC CG surfactant and 1 mM phe-nol solution. The size of micelles studied herein is in the orderof CG > SDS > DTAB > Brij35, in accordance with their aggregationnumber [24,26–28]. Because the aggregation number of surfactantis an indicator of the size of micelles [24]. Thus, phenol tends todissolve in larger micelles than smaller ones because the formerincorporate a relatively greater number of solubilizing molecules[10]. Interestingly, the average aggregation number of Gemini sur-factant increases with increased concentration [24]. Therefore, thesmall CG micelles are formed at the lower feed concentrations,which may be responsible for inferior removal efficiency. When theconcentration of CG achieves 10 CMC, the formed large CG micellesprovide great stability. Additionally, aromatic organics may incor-porate into the palisade layer of the micelle and interact with thecationic surfactant head groups through a �-cationic effect [25]. ForCG, the �-cationic effect is much intense due to two head groupsand the increasing of alky length also leads to higher solubilizingpower.

3.2.2. Surfactant retentionThe effects of feed concentration on the retention of surfac-

tants are reported in supplementary data Fig. S2. The retention ofsurfactants decreases with the increasing of feed surfctant con-centration. A larger size of micelles than that of UF membranemaking them easily intercepted. The result agrees well with thestudy of Yenphan et al. [7]. Therefore, the elevated feed phenol con-centration promotes the concentration of solutes in the micellesbefore saturation, but has little effect on the surfactant retention.In this study, the sequence of the retention of surfactant is generallyCG > DTAB > SDS > Brij35. The great retention rate of CG surfactantis presumably because of its large micelle size, which are favorableto be trapped by membrane.

3.2.3. Permeate fluxThe permeate flux for flat sheet membranes can be defined as

[8],

J = Q

A�t(2)

where Q is the feed volume (L), A is the area of membrane (m2),J is the permeate flux (L m−2 h−1) and �t is time interval (h) ofmicellar-enhanced ultrafiltration process.

As can be seen from Fig. 3, the four selected surfactants allhave good performance on the permeate flux, which is reasonable.Selective directional adsorption is formed in the boundary betweenthe solution and the surface of membrane due to the effect ofhydrophilic and hydrophobic head groups, resulting in the changeof status and properties [29–31]. For Brij35, pyknotic hydrophobiclayer is formed in the boundary, improving the hydrophobic prop-erties of the membrane surface. Whereas, the electric charge of CG,DTAB and SDS exclude some substances which consist of the sameelectric charge through electrostatic interaction.

The permeate fluxes decreased with the addition of feed sur-factant except CG surfactant. With the increased concentrationof surfactant, more micelles are rejected by UF membranes andform the highly viscous boundary layer adjacent to the membrane

132 W. Zhang et al. / Journal of Hazardous Materials 235– 236 (2012) 128– 137

654321020

22

24

26

28

30

32

a CG 2CMC

CG 6CMC

CG 10CMC

Flu

x (

L m

-2 h

-1)

[Phenol] (mM)

654321010

15

20

25

b DTAB 2CMC

DTAB 6CMC

DTAB 10CMC

Flu

x (

L m

-2 h

-1)

[Phenol] (mM)

654321025

30

35

40

c SD S 2CMC

SD S 6CMC

SD S 10C MC

Flu

x (

L m

-2 h

-1)

[Phe nol ] ( mM)

65432102

4

6

8

10

d Br ij35 2C MC

Br ij35 6C MC

Br ij35 10CMC

Flu

x (

L m

-2 h

-1)

[Phe nol ] ( mM)

Fig. 3. Variations of the permeate flux under different feed concentration.

smmabitotmf

3

f

R

wca

urface known as a gel layer [6] bringing about sedimentation oficelles on the membrane pores or membrane surface. Therefore,ore serious membrane pollution and declined flux is produced

ttributed to the increased viscosities of feed solutions and easilylocked pores of membrane. However, for CG surfactant, it becomes

onized with powerful positive electricity in solution on account ofwo hydrophilic head groups, thus electrostatic repulsive force mayccur between CG surfactant and PES membrane with positive elec-ricity as well; secondly, the large micelles is not easy to block the

embrane pores. Therefore, CG surfactant is potential for reducingouling resistance and enhancing permeates flux of surfactant.

.2.4. Secondary resistanceThe secondary resistance (RS) caused by the surfactant layer

ormation (gelation) can be calculated as follows [5]:

S = �P

J(t)�− Rmem (3)

here �P and � are the transmembrane pressure (Pa) and the vis-osity of solution (Pa h), respectively (the viscosity of distilled watert temperature of 20 ± 1 ◦C equals to 1.01 × 10−3 Pa s).

When considering the resistance-in-series model [32], the totalresistance can be written as:

Rtot = Rmem + Rcp + Rf = Rmem + RS (4)

where the subscripts tot, mem, cp and f are the total, membrane,concentration polarization and fouling resistances, respectively; RSrepresents the secondary resistance. The most significant parame-ter affecting permeate flux decline is fouling resistance (Rf), whichcontributes to a great extent to the secondary resistance [5]. Thesecondary resistance caused by surfactant molecules is calculatedat t0 using Eq. (4) and results are presented by Fig. 4. The permeateflux measured with a transmembrane pressure of 0.30 MPa.

The RS of surfactant expresses a falling tendency with the enlarg-ing concentration of surfactant in the feed stream. As is well known,the accumulated micelles on the membrane surface account forthe increment of the polarized resistance and fouling of mem-brane surface. In this study, the RS decreases obviously in the orderof Brij35 > DTAB > CG > SDS. The extent of polarization resistance

and fouling resistance are in agreement with the RS. The SDS sys-tem has the lowest RS, probably because the lowest molecularweight of micelle entails the lowest accumulation of SDS micelleon the membrane surface [7]. Powerful hydrophilic groups of CG

W. Zhang et al. / Journal of Hazardous Materials 235– 236 (2012) 128– 137 133

543210.5

1.0

1.5

2.0

a

Rs×

10

13

(m

-1)

[Phenol] (mM)

CG 2CMC

CG 6CMC

CG 10CMC

54321

2

3

4

5

b DTAB 2CMC

DTAB 6CMC

DTAB 10CMC

Rs×

10

13

(m

-1)

[Phenol] (mM)

543210.0

0.3

0.6

0.9

1.2

1.5

c

Rs×

10

13

(m

-1)

[Phenol] (mM)

SDS 2CMC

SDS 6CMC

SDS 10CMC

543215

10

15

20

d

Rs×

10

13

(m

-1)

[Phenol] (mM)

Brij35 2CMC

Brij35 6CMC

Brij35 10CMC

istanc

sb

3

jdp

3

dpbbtttAeMas

obtained as the distribution coeffcient ranged between 5 and 25 forall selected surfactants. The D values in most of the CG surfactantsystem exceed 3, indicating high retentions of phenol. The high-est R (95.8%) and D (26) are obtained, when the concentration of

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

CG 2CMC

CG 6CMC

CG 10CMC

DTAB 2CMC

DTAB 6CMC

DTAB 10CMC

SDS 2 CMC

SDS 6 CMC

SDS 10 CMC

Brij35 2 CMC

Brij35 6 CMC

Brij35 10 CMC

Re

latio

nsh

ip r

ete

ntio

n (

R (

%)

Fig. 4. Variations of the secondary res

urfactant depress the interaction between micelles and hydropho-ic membrane, which is responsible for its low membrane fouling.

.3. Evaluation of micellar-enhanced ultrafiltration

The performance of micellar-enhanced ultrafiltration can beudged from the micelle loading, micelle binding constant and theistribution coefficient of solutes in micellar phase and aqueoushase obtained from distribution coefficient [12,22].

.3.1. Distribution coefficient of phenolFig. 5 illustrates the relationship of the retention R (%) versus the

istribution coefficient D for phenol. It is an indicator of the ratio ofhenol concentration in retentate and permeate stream. The distri-ution of the phenol in the pseudophases (micelles and aqueous) ofoth permeate and retentate stream is in equilibrium. It is evidenthat the increment of the feed surfactant concentration leads tohe increasing concentration of phenol and surfactant in the reten-ate stream in terms of D values for all sutfactants studied herein.n increase of D shows that more and more surfactant molecules

xist in the micellar pahase, binding more and more phenol [33].oreover, there is a one-one relationship between the retention

nd the distribution coefficient (Fig. 5), regardless of the types ofurfactants. As can be seen from Fig. 5, a high retention of phenol is

e under different feed concentration.

302520151050

Distribution coefficient (D)

Fig. 5. Relationship of retention R versus distribution coefficient D of phenol.

134 W. Zhang et al. / Journal of Hazardous M

65432100

1

2

3

CG 2CMC

CG 6CMC

CG 10CMC

DTAB 2CMC

DTAB 6CMC

DTAB 10CMC

SDS 2CMC

SDS 6CMC

SDS 10CMC

Brij35 2CMC

Brij35 6CMC

Brij35 10CMC

Lm (

M M

-1)

Cfttw

3

iwbf

L

wptocTasomlTmlaoistt

3

ettm

K

[Phenol] (mM)

Fig. 6. The micelle loading of phenol.

G surfactant in feed stream reaches to 10 CMC which is almostour times than that of DTAB (R = 84.8%, D = 6.757). The results fur-her demenstrate substantial solubilizing power of CG surfactantoward phenol in comparasion with other conventional surfactantith equal alkly chain length.

.3.2. The micelle loading of phenolThe measurement of the effectiveness of a particular surfactant

n solubilizing a given solute is the micelle loading of phenol Lm

hich is characterized as the number of moles of phenol solubilizedy one mole of micellized surfactant [18] and expressed by theollowing equation:

m = [PH]R − [PH]P

[S]R − CMC[M M−1] (5)

here [PH]p and [PH]R denote the concentration of phenol in theermeate and the retentate stream, respectively; [S]R representshe concentration of surfactant in the retentate stream. When otherrganic pollutants are present in feed stream, the removal efficien-ies rely on the micelles capacity of solubilizing organic pollutants.he other organic pollutants will enter micelles, if the micelles havebility to solubilize more organic matters [12]. On the contrary, theolubilization for phenol may be decreased by the existence of otherrganic pollutants while the micelles are unable to solubilize moreatters. As can be seen from Fig. 6, in the feed solution, the micelle

oading is approximately linear with the concentrations of phenol.he increase of the phenol concentration produces more phenololecules solubilized by micelles in the solution, so the micelle

oading of phenol enhances. The linearity suggested that micellesre still able to solubilize more phenol molecules. It is also clearlybserved that the optimal micelle loading of phenol following thencrease order: 10 CMC > 6 CMC > 2 CMC for all selected surfactantystems. Moreover, the micelle loading of CG is more than that ofhe other surfactants, due to its low CMC and better solubilizationoward phenol.

.3.3. Equilibrium distribution constant of phenolTo further characterize the effectiveness of solubilization, the

quilibrium distribution constan KPH, which is defined as distribu-ion of the mole fraction of phenol between surfactant micelles and

he aqueous phase may be calculated from experimental measure-

ent by using the following formula [18]:

PH = [PH]M

[PH]W S[M−1] (6)

aterials 235– 236 (2012) 128– 137

where the subscripts M and W are the micelle and bulk water (freephenol molecules), respectively, and S is the concentration of thesurfactant which is presented in micelle form. Essentially, [PH]M =[PH]R − [PH]P , [PH]W = [PH]P and S = SR − SP [32,33].

Effect of the feed surfactant and phenol concentration onthe equilibrium distribution constant Log KPH is depicted insupplementary data Fig. S3. The results show that the equilib-rium distribution constants of all selected surfactant systemsdecrease evidently with an increase of surfactant concentra-tion in feed stream, because of the great amount of micellesformed [3,33].

3.4. Characterization of membrane surface morphology andproperties altered by micelle for all selected surfactant systems

3.4.1. Membrane surface morphologyIt is clear from the appearances (Fig. 7), new membrane is

dark (A), whereas other membranes appear white spots indicat-ing they are fouled by all surfactant micelles studied herein. Thereare widespread gel layers on the fouled membrane by micelles.Obviously, cake layers can be observed in images C and D demon-strating that pollutants which are mainly micelles easily depositon the surface of C and D membranes due to the block of micellesby DTAB and Brij35, which bring about serious fouling resistanceof membrane. Consequently, the results illustrate that the mem-brane fouling by DTAB and Brij35 are much serious in comparingwith CG and SDS, which are consistent with the results discussed inSection 3.2.4.

3.4.2. Membrane material characteristicsThe changes of porosity in terms of pore volume per unit mass of

membrane material with pore size are illustrated in Fig. 8. The dis-tribution of pore diameter of fouling membranes is narrower thanthat of new membrane to a certain extent, meaning the existence ofpore blockage fouling. The sequence of the porosity of membraneis: new membrane > fouled membrane of SDS > fouled membrane ofCG > fouled membrane of DTAB > fouled membrane of Brij35, whichis in tune with the order of pore diameter of membrane. This coin-cides well with the degree of membrane fouling. When micellesdeposit on the surface of membrane, smaller pores may be easilyblocked by smaller size of micelles. The micelles of CG and SDS havemore dominant advantage than Brij35 and DTAB for the changesof porosity, presumably due to their larger micelle size. In otherwords, the smaller micelles of Brij35 and DTAB may become themain reason of pock blockage. Although most of the small pores areblocked, the numbers of porosity at smaller pores increase ratherthan decrease. This may be ascribed to some big pores that alsobe partial blocked, thus the small pores become the predominantaccumulation spot for micelles.

3.4.3. Analysis of membrane foulantsFig. 9 shows the FTIR spectrograms of new and fouled

membrane. Four characteristic peaks (837/872 cm−1, aro-

matic hydrocarbons 1,4-replacementR R'

;

1149/1298/1321 cm−1, R–SO2–R′; 1485/1578/3070 cm−1,; 1660 cm−1, aromatic hydrocarbons compounds) of all membrane

material(

SO2On )

are clearly observed.In comparison with the new membrane, the peaks of alkyl group

(2856/2923 cm−1, CH2) on the fouled membrane demonstratethe existence of alkyl-like substances on the fouling layer ofmembrane, which are surfactant micelles. The peaks of Brij35 at2800–3000 cm−1 are strengthened and broadened, meaning that

W. Zhang et al. / Journal of Hazardous Materials 235– 236 (2012) 128– 137 135

F memf

tsomcR

ig. 7. SEM images of membrane surface (20,000×). (A) New membrane; (B) fouledouled membrane of SDS.

he number of surfactant micelles of Brij35 deposit on membraneurface or enter pores of membrane is the most in comparison with

ther selected surfactants [22]. Comparing with other surfactants,embrane fouling of Brij35 is the most serious of all, which is

orresponding with the minimum permeate flux and maximumS.

brane of CG; (C) fouled membrane of DTAB; (D) fouled membrane of Brij35; and (E)

4. Summary

Micellar-enhanced ultrafiltration can be regarded as a promis-ing method to dispose phenol from an aqueous stream. Thepresent study investigated the removal efficiency of phenol dur-ing micellar-enhanced ultrafiltration in the presence of cationic

136 W. Zhang et al. / Journal of Hazardous M

1000100100.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.10

New membran e

Fouled membrane of CG

Fouled membrane of DT AB

Fouled membrane of SDS

Fouled membrane of Brij35

Po

re v

olu

me p

er

mass o

f m

em

bra

ne m

ate

ria

l (m

l g

-1)

Pore diameter ( nm)

Fig. 8. Porosity changes of new and fouled membrane.

4000350030002500200015001000500

0.0

0.2

0.4

0.6

0.8

New membran e

Fou led membrane of CG

Fou led membrane of DTA B

Fou led membrane of SDS

Fou led membrane of Brij35

Ab

so

rban

ce

-1

GBftatsleotasrmi

A

Nvf

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

Wavenumber ( cm )

Fig. 9. ATR-FTIR spectrums of new and fouled membrane.

emini surfactant (CG) and conventional surfactant such as DTAB,rij35 and SDS. It is found that the removal efficiency of sur-

actants is in the sequence of CG > DTAB > SDS > Brij35 indicatinghe CG surfactant has great solubilizing capabilities of phenol onccount of its innovative structures and physicochemical proper-ies. Meanwhile, in terms of surfactant retention, permeate flux andecondary resistance, the selected CG surfactant also exhibits excel-ent performance with respect to conventional surfactant havingqual alkyl chain. Besides, distribution coefficient (D), the loadingf the micelles (Lm) and equilibrium distribution constant (K) fur-her prove that CG surfactant is useful for improving hydrophilicitynd permeability of the membranes through modifying the surfaceituation of polysulfone UF membranes. The systematic analysisesults demonstrate the potential utilization of CG surfactant inicellar-enhanced ultrafiltration which has a distinct superiority

n organic pollutant removal of process.

cknowledgements

This research was supported by the Major Project Program of the

atural Sciences Foundation (51190095), and the Program for Inno-ative Research Team in University (IRT1127). We are also gratefulor anonymous reviewers for their helpful suggestions and advices.

[

[

aterials 235– 236 (2012) 128– 137

Appendix A. Supplementary data

Supplementary data associated with this article can be found,in the online version, at http://dx.doi.org/10.1016/j.jhazmat.2012.07.031.

References

[1] S. Rengaraj, S.H. Moon, R. Sivabalan, B. Arabindoo, V. Murugesan, Agriculturalsolid waste for the removal of organics: adsorption of phenol from waterand wastewater by palm seed coat activated carbon, Waste Manag. 22 (2002)543–548.

[2] G.M. Zeng, K. Xu, J.H. Huang, X. Li, Y.Y. Fang, Y.H. Qu, Micellar enhanced ultrafil-tration of phenol in synthetic wastewater using polysulfone spiral membrane,J. Membr. Sci. 310 (2008) 149–160.

[3] S.R. Jadhav, N. Verma, A. Sharma, P.K. Bhattacharya, Flux and retention analysisduring micellar enhanced ultrafiltration for the removal of phenol and aniline,Sep. Purif. Technol. 24 (2011) 541–557.

[4] M.M. Husein, A. Deriszadeh, T.G. Harding, Experimental and modeling study ofMEUF removal of naphthenic acids, Desalination 273 (2011) 352–358.

[5] A. El-Abbassi, M. Khayet, A. Hafidi, Micellar enhanced ultrafiltration process forthe treatment of olive mill wastewater, Water Res. 45 (2011) 4522–4530.

[6] M.A.M. El Zeftawy, C.N. Mulligan, Use of rhamnolipid to remove heavy met-als from wastewater by micellar enhanced ultrafiltration (MEUF), Sep. Purif.Technol. 77 (2011) 120–127.

[7] P. Yenphan, A. Chanachai, R. Jiraratananon, Experimental study on micellarenhanced ultrafiltration (MEUF) of aqueous solution and wastewater contain-ing lead ion with mixed surfactants, Desalination 253 (2010) 30–37.

[8] L. Fang, G.M. Zeng, J.H. Huang, C. Zhang, Y.Y. Fang, Y.H. Qu, X. Li, D. Li, C.F. Zhou,Effect of groups difference in surfactant on solubilization of aqueous phenolusing MEUF, J. Hazard. Mater. 173 (2010) 455–461.

[9] K. Xu, H.Q. Ren, G.M. Zeng, L.L. Ding, J.H. Huang, Investigation of interactionbetween phenol and cetylpyridinium chloride micelle in the absence and inthe presence of electrolyte by 1HNMR spectroscopy, Colloid Surf. A 356 (2010)150–155.

10] J. Lee, J.S. Yang, H.J. Kim, K. Baek, J.W. Yang, Simultaneous removal of organicand inorganic contaminants by micellar enhanced ultrafiltration with mixedsurfactant, Desalination 184 (2005) 395–407.

11] X. Li, G.M. Zeng, J.H. Huang, D.M. Zhang, L.J. Shi, S.B. He, M. Ruan, Simultane-ous removal of cadmium ions and phenol with MEUF using SDS and mixedsurfactants, Desalination 276 (2011) 136–141.

12] M. Bielska, J. Szymanowski, Removal of methylene blue from waste water usingmicellar enhanced ultrafiltration, Water Res. 40 (2006) 1027–1033.

13] M.J. Meziani, H. Benalla, J. Zajac, S. Partyka, D.J. Jones, Adsorption of a cationicGemini surfactant from aqueous solution onto aluminosilicate powders of theMCM-41 type: effect of pore size and co-adsorption of phenol, J. Colloid Inter-face Sci. 262 (2003) 362–371.

14] W.W. Song, N.B. Li, H.Q. Luo, Gemini surfactant applied to the heparin assay atthe nanogram level by resonance Rayleigh scattering method, Anal. Biochem.422 (2012) 1–6.

15] A. Brun, G. Brezesinski, H. Möhwald, M. Blanzat, E. Perez, I. Rico-Lattes, Inter-action between phospholipids and new Gemini catanionic surfactants havinganti-HIV activity, Colloid Surf. A 228 (2003) 3–16.

16] J. Fan, W.B. Yang, A.M. Li, Adsorption of phenol, bisphenol A and nonylphe-nol ethoxylates onto hypercrosslinked and aminated adsorbents, React. Funct.Polym. 71 (2011) 994–1000.

17] J. Wei, G.H. Huang, C.J. An, H. Yu, Investigation on the solubilization of polycyclicaromatic hydrocarbons in the presence of single and mixed Gemini surfactants,J. Hazard. Mater. 190 (2011) 840–847.

18] J. Wei, G.H. Huang, H. Yu, C.J. An, Efficiency of single and mixed Gem-ini/conventional micelles on solubilization of phenanthrene, Chem. Eng. J. 168(2011) 201–207.

19] M.K. Purkait, S. DasGupta, S. De, Separation of aromatic alcohols using micellarenhanced ultrafiltration and recovery of surfactant, J. Membr. Sci. 250 (2005)47–59.

20] S. Tripathi, D. Brown, Effects of linear alkylbenzene sulfonate on the sorptionof Brij30 and Brij35 onto aquifer sand, Environ. Sci. Technol. 42 (2008) 1492–1498.

21] Y.H. Sun, Y.J. Feng, H.W. Dong, Z. Chen, Adsorption of dissymmetric cationicGemini surfactants at silica/water interface, Surf. Sci. 601 (2007) 1988–1995.

22] C.H. Wei, X. Huang, R. Ben Aim, K. Yamamoto, G. Amy, Critical flux and chemicalcleaning-in-place during the long-term operation of a pilot-scale submergedmembrane bioreactor for municipal wastewater treatment, Water Res. 45(2011) 863–871.

23] M. Bielska, J. Szymanowski, Micellar enhanced ultrafiltration of nitrobenzeneand 4-nitrophenol, J. Membr. Sci. 243 (2004) 273–281.

24] R. Zana, H.L. Papoutsi, G. Beinert, Micellization of two triquaternary ammoniumsurfactants in aqueous solution, Langmuir 11 (1995) 3694–3698.

25] O. Zheng, J.X. Zhao, Solubilization of pyrene in aqueous micellar solutionsof Gemini surfactants C12 S C12·2Br, J. Colloid Interface Sci. 300 (2006)749–754.

26] A.R. Tehrani-Bagha, J. Kanbratt, J.E. Lofroth, K. Holmberg, Cationic ester-containing Gemini surfactants: determination of aggregation numbers by

ous M

[

[

[

[

[

[

W. Zhang et al. / Journal of Hazard

time-resolved fluorescence quenching, J. Colloid Interface Sci. 376 (2012)126–132.

27] L.P.J. Dianne, I.J. Wesley, M.S. Jessica, H.R. Amy, An 1HNMR investigationinto the loci of solubilization of 4-nitrotoluene, 2,6-dinitrotoluene, and 2,4,6-trinitrotoluene in nonionic surfactant micelles, Colloid Surf. A 375 (2011)12–22.

28] A. Jangher, P.C. Griffiths, A. Paul, S.M. King, R.K. Heenan, R. Schweins, Polymericmicelle disruption by cosolvents and anionic surfactants, Colloid Surf. A 391(2011) 88–94.

29] A.A. Kavitskaya, Separation characteristics of charged ultrafiltration mem-branes modified with the anionic surfactant, Desalination 184 (2005) 409–414.

[

aterials 235– 236 (2012) 128– 137 137

30] S.J. Kim, P. Noeon, S.Y. Lee, J.W. Cho, Membrane characterizations for mitigationof organic fouling during desalination and wastewater reclamation, Desalina-tion 238 (2009) 70–77.

31] X.F. Lu, S.Y. Chen, C.Z. Li, Study on the surface modification of ultrafiltrationmembrane by surfactants, Membr. Sci. Technol. 17 (1997) 67–74.

32] J. Sabate, M. Pujola, E. Centelles, M. Galan, J. Liorens, Determination of equilib-

rium distribution constants of phenol between surfactant micelles and waterusing ultrafiltering centrifuge tubes, Colloid Surf. A 150 (1999) 229–245.

33] V.D. Karate, K.V. Marathe, Simultaneous removal of nickel and cobalt fromaqueous stream by cross flow micellar enhanced ultrafiltration, J. Hazard.Mater. 157 (2008) 464–471.