Separation and Purification Technology 123 (2014) 45–52

Contents lists available at ScienceDirect

Separation and Purification Technology

journal homepage: www.elsevier .com/ locate /seppur

Separation of benzene–cyclohexane mixtures by filled blend membranesof carboxymethyl cellulose and sodium alginate

1383-5866/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.seppur.2013.12.017

⇑ Corresponding author. Fax: +91 33 351 9755.E-mail address: samitcu2@yahoo.co.in (S.K. Ray).

Sunil Baran Kuila a, Samit Kumar Ray b,⇑a Department of Chemical Engineering, Haldia Institute of Technology, Haldia, West Bengal, Indiab Department of Polymer Science and Technology, University of Calcutta, Kolkata, West Bengal, India

a r t i c l e i n f o a b s t r a c t

Article history:Received 29 August 2013Received in revised form 4 December 2013Accepted 11 December 2013Available online 30 December 2013

Keywords:Sodium alginateCarboxymethyl celluloseBlendMembrane characterizationBenzene–cyclohexanePervaporation

Blend membranes were prepared by solution blending of sodium alginate (SA) and sodium carboxy-methyl cellulose (CMC) in water with varied compositions. One of the blend membranes containing25% SA and 75% CMC was found to show optimum flux and benzene selectivity for 19.6 wt% benzenein cyclohexane. This unfilled blend membrane designated as F0 was further filled with 2, 4, 6 and8 wt% organophilic bentonite filler to obtained four filled membranes designated as F2, F4, F6 and F8,respectively. These five (one unfilled and four filled) membranes were characterized by various conven-tional methods like FTIR, XRD, DTA (for unfilled blend membranes) and SEM (for filled membranes).These membranes were used for pervaporative separation of benzene from its mixtures with cyclohexaneover the concentration range of 0.5–20 wt% benzene. The filled F8 membrane showed lower flux(35.65 kg lm/m2 h) than the unfilled F0 membrane (77 kg lm/m2 h) but separation factor for benzeneof the F8 membrane (212) was much higher than the F0 membrane (88.7).

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

In pervaporation hydrophilic membranes are used for dehydra-tion while organophilic membranes are employed for removal oftraces of organics from water. This straightforward approach ofmembrane selection for preferential permeation of water or organ-ics from water-organic mixture is not applicable for organic–or-ganic separation. The choice of an organophilic membrane forselective removal of an organic from its mixture with another or-ganic depends on relative solubility parameter of the membranepolymer and the permeants. The permeant with solubility param-eter close to that of the membrane polymer will selectively perme-ate through the membrane. Among the various aromatic–aliphaticmixtures, membrane based separation of low concentration of ben-zene from its mixtures with cyclohexane is important. Cyclohex-ane is industrially produced by hydrogenation of benzene underNi or Pt catalyst. The removal of the unreacted benzene from thereaction mixtures is difficult by conventional distillation processsince benzene and cyclohexane have close boiling points (benzene80.1 �C and cyclohexane 80.7 �C). Presently, instead of normal dis-tillation, azeotropic or extractive distillation is practiced for sepa-ration of benzene–cyclohexane mixtures [1]. However, theseseparation processes are more complex and also very expensivebecause of requirement of high energy. As a low energy alternative

candidate membrane based separation of benzene–cyclohexanemixtures by pervaporation have been tried by many researchers.Unlike distillation, separation efficiency of pervaporation doesnot depend on relative volatility of the components to be sepa-rated. Instead, separation potential of pervaporation depends onpreferential sorption and diffusion of the desired component pres-ent in low concentration, i.e., benzene in the present case. Betweenbenzene and cyclohexane, benzene has smaller size (benzene89.4 cm3/mol, cyclohexane 108.7 cm3/mol) and collision diameter(benzene 89.4 nm, cyclohexane 0.606 nm) [1]. Thus, under thesame concentration gradient benzene is likely to show preferentialdiffusion by moving at a faster rate than cyclohexane. Being aro-matic with pi electron conjugation, benzene also shows electro-static interaction with polymer having reactive functional groups.Depending on this electrostatic interaction a good many numberof polymers show varied degree of swelling and even complete sol-ubility in benzene. Thus, highly benzene selective membrane maybe made from polymer having high affinity for benzene. Based onthis concept membrane made from organophilic polymers likepolypropylene, low density polyethylene, polyvinyledene fluoride(PVDF), blend of polyvinyl chloride and ethylene vinyl acetate[1,2], functional polymers having varied degree of swelling in ben-zene, i.e., polyvinyl alcohol and its blend with other polymers andadsorptive fillers [1,3–7], cellulose based polymers i.e. celluloseacetate, tosylcellulose butyryl cellulose [1,8,9], variouscopolymers, i.e., copolymers of acrylonitrile, polyacrylonitrile–polymethyl acrylate block polymer [1,10], acrylate and styrene

46 S.B. Kuila, S.K. Ray / Separation and Purification Technology 123 (2014) 45–52

grafted polyamide [11], polyether block amide (PEBA) [12], semi-fluorinated aromatic, poly(ether amide) [13], aromatic polyamide[14], polyurethane, polyoxytetramethylene based polyurethane[1,15], nafion-117 containing silver and sodium metal [16], substi-tuted polyacetylene [17], etc., and also various elastomers showingswelling in benzene like polydimethyl siloxane rubber [18], sty-rene butadiene rubber (SBR) [19], etc. were tried for separationof low concentration of benzene from its mixtures with cyclohex-ane. In the present work several blend membranes have been syn-thesized from CMC, SA and organophilic bentonite filler. Thesemembranes were used for pervaporative separation of benzenefrom its mixtures with cyclohexane over the concentration rangeof 0.5–20 wt% benzene in cyclohexane.

2. Experimental

2.1. Materials

Natural polymer sodium alginate (SA, average molecular weight500,000 and degree of deacetylation 84%) was procured from E.Merck, Mumbai and used as it is without any further purification.Sodium salt of carboxymethyl cellulose (CMC, pH 6–8, degree ofsubstitution 1.8 and molecular weight 20,000) was also procuredfrom E. Marck, Mumbai and was used as obtained. Organophilicbentonite filler (API grade, residue of diameter >75 micron or 200mesh is 3.5%) was kindly given by Amrfeo pte. Ltd., Kolkata. Highpurity, HPLC grade benzene, cyclohexane and crosslinker glutaral-dehyde (25% aqueous solution) was purchased from s.d. fine-chem,Mumbai, India. For making polymer solution deionized water, hav-ing a conductivity of 20 lS/cm, was produced in the laboratory it-self from a reverse osmosis module.

2.2. Preparation, crosslinking and casting of blend membranes

Polymer solutions were made in deionised water in a 250 mlglass beaker by gradual addition of required amount of CMC or so-dium alginate (1% m/v) to boiling water in several intervals withconstant stirring to obtain a viscous clear polymer solution. Twosolutions were then mixed with addition of required amount of fil-ler. It was stirred with magnetic stirrer for 8 h to get filler incorpo-rated stable polymer dispersion. Crosslinking of CMC and sodiumalginate blend by glutaraldehyde was carried out at a certain pHmaintained by adding sulfuric acid (10 wt% in water), acetic acid(10 wt% in water) and methanol (50 wt% in water) in the blendsolution like crosslinking of polyvinyl alcohol [20]. Membranewas prepared by casting this aqueous dispersion of the blend withan applicator on a clean and smooth glass plate. It was kept over-night at room temperature and then dried at 60 �C for 2 h undervacuum. Membrane thickness was maintained at �50 lm. Thethickness was measured by Test Method ASTM D 374 using a stan-dard dead weight thickness gauge (Baker, Type J17).

2.3. Membrane characterization

2.3.1. Fourier transform infrared spectroscopy (FTIR)The FTIR spectra of the thin film (�10 cm) of the polymer sam-

ples were carried out in a Perkin Elmer (model-Spectrum-2, Singa-pore) spectroscope.

2.3.2. X-ray diffraction (XRD)The wide angle X-ray diffraction (XRD) of the polymer samples

was carried out in a diffractometer (model: X’Pert PRO, made byPANalytical B.V., The Netherlands). For XRD Ni-filtered Cu Ka radi-ation with wavelength (k) of 1.542 Å was used. The scanning ratewas fixed at 2 deg/s.

2.3.3. Differential thermal analysis and thermogravimetric analysis(DTA–TGA)

The DTA and TGA of the membrane samples were carried out ina Perkin Elmer instrument in nitrogen atmosphere at the scanningrate of 10 �C per minute in the temperature range of 60–600 �C.

2.3.4. Scanning electron microscopy (SEM)For studying morphology of the blend samples SEM (model No.

S3400N, VP SEM, Type-II, made by Hitachi, Japan) of the goldcoated polymer samples was carried out at accelerating voltageof 15 kV.

2.4. Total sorption and membrane phase concentration

Sorption experiments were carried out by immersing knownweight (md) of dry membrane samples in mixtures of benzeneand cyclohexane of varied compositions. The membranes im-mersed in the solvent mixtures were allowed to equilibrate for96 h at 30 �C. Each sample was weighed periodically until noweight change was observed. The swollen membranes were takenout from the solution and weighed (me) after the superfluous liquidwas wiped out with tissue paper. The total sorption (St) of benzeneand cyclohexane mixtures by the membranes (g/g dry membrane)is obtained as

St ¼me �md

mdð1Þ

Partial sorption of benzene and cyclohexane was obtained bymultiplying total sorption weight (St) with membrane phase sol-vent concentration viz. weight fraction of benzene and cyclohexanein swollen membranes. For determining concentration of benzeneand cyclohexane in swollen membrane, the total amount of sol-vents sorbed by the membrane (St) was collected in a trap im-mersed in liquid nitrogen and connected to a vacuum pump. Thecomposition of benzene and cyclohexane in membrane was ana-lyzed by a digital refractometer (Anton Paar, model – Abbemat-HP).

2.5. Permeation studies

Permeation studies of solvent mixtures were carried out by per-vaporation experiment in a stirred batch cell. The membrane wasinserted in the stainless-steel stirred cell. Effective membrane area(A) in contact with the feed solution was 19.6 cm2 and the feedcompartment volume was 150 cm3. Downstream pressure wasmaintained at �1 mmHg by a vacuum pump. The feed mixturesin contact with the membrane were allowed to equilibrate foraround 3 h for the first experiment and 1 h for the subsequentexperiments with different feed compositions. The pervaporationexperiment was performed at a constant temperature by circulat-ing constant temperature water around the jacket of the pervapo-ration cell for different feed compositions. The solvent vaporscoming out from the membrane on its downstream side at lowpressure were collected in traps immersed in liquid nitrogen. Per-vaporation experiments were reproducible and the errors inherentin the measurements were less than 1.0%.

2.5.1. Total and partial fluxFrom the permeation data, total flux (J, as g/m2h) was obtained

using the following equation

J ¼WAt

ð2Þ

Total molar flux (Jm, mole/m2 h) was obtained as

JMavg:

¼ Jm ¼J

xiMi þ ð1� xiÞMjð2aÞ

S.B. Kuila, S.K. Ray / Separation and Purification Technology 123 (2014) 45–52 47

Here x1 is mole fraction of solvent 1 in permeate while M1 is itsmolecular weight. Permeate composition of benzene and cyclohex-ane was determined in a similar way as membrane phase concen-tration during sorption experiments. Partial molar flux of solvent(benzene or cyclohexane) was obtained from Eq. (3)

J1 ¼WAt

xw1

M1ð3Þ

Here w1 and M1 are wt fraction and molecular weight of solvent 1,respectively.

2.5.2. Pervaporation selectivity, sorption selectivity and diffusionselectivity

Pervaporation selectivity of benzene or separation factor of ben-zene (aPV) was obtained as the ratio of molar concentration of ben-zene in permeate (Ybenzene) and feed (Xbenzene) by the following Eq.(4) [21]

apv ¼Ybenzene

Ycyclohexane

XbenzeneXcyclohexane

ð4Þ

Sorption selectivity of benzene (as) was obtained as the ratio ofmolar concentration of benzene in membrane (Ymbenzene) and feed(Xbenzene) using an equation similar to Eq. (4).

Diffusion selectivity of benzene (aD) was obtained from its per-vaporation and sorption selectivity as [22]

aD ¼ aPV=as ð5Þ

Table 1Flux and selectivity of the blend membranes for 19.6 wt% benzene in feed.

Membranename

Membranecomposition

Flux, g/m2 h

Permeate selectivity forbenzene

SA10CMC90 10% SA + 90%CMC 2680.10 48.13SA25CMC75⁄ 25%SA + 75%CMC 2233.67 57.90SA34CMC66 34% SA + 66%CMC 1550.51 39.34SA50CMC50 50% SA + 50%CMC 1938.75 36.55SA66CMC34 66% SA + 34%CMC 1712.25 34.53SA75CMC25 75% SA + 25%CMC 1321.56 29.51

⁄ SA25CMC75 showed the best results for flux and benzene selectivity.

2.5.3. Partial permeability and intrinsic membrane selectivityPartial permeability of solvent 1 (P1) through membranes was

obtained from its partial molar flux (J1), membrane thickness (l)and partial vapor pressure on feed and permeate side using the fol-lowing Eq. 6 [23]

P1 ¼ l� J1

ðxf1c1psat

f1� yp1

ppÞð6Þ

where xf1 and yp1are mole fraction of solvent 1 on feed (liquid

phase) and permeate (vapor phase) side, psatf1

is saturated vaporpressure of solvent 1 on feed side obtained using Antoine’s equationand pp is permeate pressure in pervaporation process. Ignoring lowpermeate pressure Eq. (6) may be simplified as

Fig. 1. FTIR of th

P1 ¼ l� J1

f1ð6aÞ

where fugacity f 1 ¼ xf1c1psat

f1ð6bÞ

Intrinsic membrane selectivity (amem) is defined as the ratio of per-meability of solvent i and j i.e.

amem ¼Pi

Pjð7Þ

3. Results and discussion

3.1. Synthesis of membranes

Both SA and CMC are water soluble polymers and hence mem-branes made from these two polymers are likely to be hydrophilic.In fact, hydrophilic blend membranes of SA and acrylamide graftedCMC have also been reported for dehydration of isopropyl alcoholby pervaporation [24]. However, in the present work the objectivewas to synthesize a benzene selective membrane. CMC showsswelling in benzene while sodium alginate does not show signifi-cant swelling in benzene. However, both of these polymers are in-ert to cyclohexane. Thus, the objective of the present study was tosynthesize alloy type [1] membrane from SA and CMC where SAwill give stability and integrity of the membrane while CMC willprovide permeability for benzene. From Table 1 it is observed thatwith increasing amount of CMC in the blend, flux increases. How-ever, above 75 wt% CMC the blends showed extensive swelling insolvent mixtures and the membranes were also not stable duringPV experiments. The membrane showing the best results for flux

e polymers.

0

200

400

600

800

1000

1200

0 20 40 60 80

Two theta

Inte

nsit

y

Sodium alginate (SA)CMCSA25CMC75 (F0)

Fig. 2. XRD of the polymer.

48 S.B. Kuila, S.K. Ray / Separation and Purification Technology 123 (2014) 45–52

and benzene selectivity viz. SA25CMC75 was filled with organo-philic bentonite filler. This adsorptive filler fills up the void spacein the blend interface and improve benzene selectivity of themembranes.

3.2. Characterization of membrane

3.2.1. Characterization of membranes by FTIRAs observed in Fig. 1, SA shows absorption band at 1107 and

1022 cm�1 corresponding to asymmetric and symmetric stretchingvibration of its C–O–C group [25]. The band at 1642 cm�1 and1402 cm�1 corresponds to asymmetric and symmetrical stretchingvibration, respectively, of its COO� groups [26]. The absorptionband at 3357 cm�1 is due to its hydroxyl (O–H) stretching. Thestretching vibration of carboxylic group of pure CMC is observedat 1606.5 cm�1 while its CH2 scissoring and OH bending vibrationis observed at 1419 and 1326 cm�1. 1, 4- b- D-glucoside stretching

(a)

-20

-15

-10

-5

0

5

10

15

-200 0 200 400 600 800

Temperature oC

mW

(c)

0

2

4

6

8

10

0 200 400 600 800

Temperature oC

mW

(e)

-15

-10

-5

0

5

0 200 40

Temperature (

mW

Fig. 3. DTA of the polymers. (a) Sodium alginate (SA), (b) carboxymethyl cellulose (

vibration of CMC is observed at 1083 cm�1. The broad absorptionbands from 3300 to 3500 cm�1 are due to stretching of large num-ber of hydroxyl groups of CMC [27,28]. The broad absorption bandsfrom 3300–3500 cm�1 is also observed in the blend while carbox-ylic symmetric stretching vibration of 1642 cm�1 of SA and1606.5 cm�1 of CMC is observed to overlap at around 1645 cm�1

in the blend. Similarly, 1402 cm�1 peak of SA corresponding itssymmetrical COO- groups and 1419 cm�1 peak of CMC correspond-ing its CH2 scissoring is also overlapped at one single peak of1420 cm�1 in the blend. The 1326 cm�1 peak of CMC due to OHbending vibration is also observed to shift to 1365 cm�1 in theblend. All of these overlapping and shifting of absorption peaksin the blend indicate electrostatic interaction among various func-tional groups of the constituent polymers of the blend.

3.2.2. XRD analysisFrom Fig. 2 it is observed that SA shows three crystalline peaks

at 14.3�, 21.3� and 37.1� as also reported elsewhere [29]. CMC alsoshows its crystalline peak at around 20 degree [30]. All of theseXRD peaks are shifted with much reduction in peak intensity inthe blend of CMC and SA. In fact, in the blend of CMC and sodiumalginate crystallinity of both CMC and alginate reduces because offormation of new chemical bonds between molecules of the twopolymers by crosslinking reaction and also because of strong elec-trostatic interaction between SA and CMC by intermolecularhydrogen bonding.

3.2.3. DTAFrom Fig. 3 it is observed that SA shows an endothermic peak at

90 �C corresponding to loss of water while the exothermic peak at257 �C is due to its degradation. Similar thermal characteristicpeaks were reported elsewhere [31]. As SA is blended with CMC,the characteristic peaks of alginate is shifted due to mutual inter-action of functional groups of CMC and sodium alginate. Thus,

(d)

-10

-5

00 200 400 600 800

Temperature oC

mW

(b)

-4

-2

0

2

4

6

8

Temperature OC

mW

0 600 800

oC)

0 200 400 600 800

CMC), (c) CMC25SA75 blend, (d) CMC50SA50 blend, (e) CMC75SA25 (F0) blend.

S.B. Kuila, S.K. Ray / Separation and Purification Technology 123 (2014) 45–52 49

endothermic peak of pure sodium alginate is observed to shift from90 �C (pure alginate i.e. 0% CMC) to 84 �C (25% CMC), 82 �C(50%CMC), 77 �C (75% CMC) and 75 �C (90% CMC), respectively. The exo-thermic degradation peak also similarly shifts from 257 �C (0%CMC) to 250 �C (25% CMC), 239 �C (50% CMC), 220 �C (75% CMC)and 217 �C (90% CMC) with much reduction in peak intensity.The reduction in peak intensity may be due to loss in crystallinityin the blends. Both of these polymers show crystallinity because ofhydrogen bonding among its hydroxyl and carboxylic functionalgroups [32]. However, in the blend crystallinity of alginate as wellas CMC reduces because of hydrogen bonding between these twopolymers through its functional groups. The shifting of the endo-thermic and exothermic peaks in the blends also indicates stronghydrogen bonding among functional groups of the two polymers.

3.2.4. SEM analysisSEM of unfilled blend F0 and also blend membrane filled with 2,

4, 6 and 8% filler (i.e. F2, F4, F6 and F8, respectively) are shown inFig. 4a–e, respectively. The unfilled blend morphology is shown at5000� magnification for better clarity of the two phases i.e.

Fig. 4. SEM of the membranes, (a)

alginate and CMC. The rod like morphology of the cellulose i.e.CMC and alginate [33] is evident from this figure. The distributionof filler in the blend is shown at a lower magnification i.e. 1000�. Itis observed that fillers are uniformly distributed in the blendthough at higher filler concentration agglomeration of the fillerparticles is observed.

3.3. Sorption isotherm of benzene and cyclohexane

From Fig. 5 it is observed that with increase in benzene concen-tration in feed partial sorption of benzene and cyclohexane in-creases almost linearly up to around 7% (molar) benzene in feedand thereafter the change in partial sorption with feed concentra-tion is marginal. These sorption isotherms resemble dual modetype-II sorption of Rogers, typical of filled membranes [34]. It isalso observed that partial sorption decreases with increase in fillerconcentration in membrane. Thus, for 0.6 M% benzene in feed par-tial sorption of 1 g of membrane (dry) decreases from0.058 � 10�3 mole/g for F0 to 0.036 � 10�3 mole/g for F8 mem-brane. Similarly, for cyclohexane it decreases from 1.33 � 10�3 -

F0, (b) F2, (c) F4, (d) F6, (e) F8.

0

3

6

9

12

15

0 5 10 15 20 25

Feed conc. of benzene (molar% in cyclohexane)

Part

ial b

enze

ne (

Bz)

sor

ptio

n (m

ol/g

dry

mem

bran

e) x

10

3

0

3

6

9

12

15

Part

ial c

yclo

hexa

ne (

CH

) so

rptio

n,

(mol

/g d

ry m

embr

ane)

x10

3

F0Bzsorp F2Bzsorp F4BzsorpF6Bzsorp F8Bzsorp F0CHsorpF2CHsorp F4CHsorp F6CHsorpF8CHsorp

Fig. 5. Variation of partial sorption with feed conc. of benzene (Bz) at 30 �C.

0

50

100

150

200

250

300

350

0 5 10 15 20 25

Feed conc. of benzene (molar% in cyclohexane)

PV

sel

ecti

vity

for

ben

zene

(-)

0

5

10

15

20

25

30

35

Dif

fusi

on o

r so

rpti

on

sel

ectiv

ity

for

benz

ene

(-)

F0PVsel F2PVsel F4PVsel F6PVselF8PVsel F0diffusionsel F2diffusionsel F4diffusionselF6diffusionsel F8diffusionsel F0sorpsel F2sorpselF4sorpsel F6sorpsel F8sorpsel

Fig. 7. Variation of benzene selectivity with feed conc. at 30 �C.

60

70

80

90

100ra

tion

of b

enze

ne

ar%

) 60

70

80

90

100

conc

. of

benz

ene

r%)

50 S.B. Kuila, S.K. Ray / Separation and Purification Technology 123 (2014) 45–52

mole/g for F0 to 0.535 � 10�3 mole/g for F8 membrane. Increase inpartial sorption with feed benzene concentration may be due toplasticization of the membranes at higher feed benzene concentra-tion. The void space and defects in the interphase of the blend isfilled up by the filler. Thus, partial sorption decreases with increasein filler concentration in the membrane.

0

10

20

30

40

50

0 20 40 60 80 100

Feed concentration of benzene (molar% in cyclohexane)

Perm

eate

con

cent

(mol

0

10

20

30

40

50

Mem

bran

e ph

ase

(mol

a

F0per F2per F4per F6per F8per VLEF6mem F8mem F0mem F2mem F4mem

VLE

Fig. 8. Variation of membrane phase and permeate concentration of benzene withits feed concentration at 30 �C.

1408

0.9

3.4. Pervaporation (PV)

3.4.1. Flux and selectivityVariation of partial flux with feed benzene concentration is

shown in Fig. 6. It is observed that with increase in feed concentra-tion partial flux increases linearly. It is also observed that above20% (molar) benzene in feed there is a sudden increase in fluxwhich indicates plasticization of the membranes. The partial ben-zene molar flux is observed to be much higher than partial cyclo-hexane molar flux for any feed concentration. Variation ofpervaporation (PV) selectivity, sorption selectivity and diffusionselectivity with feed benzene concentration is shown in Fig. 7. Itis observed form Fig. 7 that variation of PV, sorption or diffusionselectivity with feed concentration is not significant up to 20% ben-

0

5

10

15

20

25

30

35

40

45

0 5 10 15 20 25

Feed concentration of benzene (molar% in cyclohexane)

Part

ial b

enze

ne F

lux

(mol

e/m

2hr)

-1

1

3

5

7

9

11

13

15

Part

ial c

yclo

hexa

ne

F0BzFLUX F2BzFLUX F4BzFLUX F6BzFLUXF8BzFLUX F0CHFLUX F2CHFLUX F4CHFLUXF6CHFLUX F8CHFLUX

(CH

) Fl

ux (

mol

e/m

2 hr)

Fig. 6. Variation of flux with feed conc. of benzene (Bz) at 30 �C.

0

2

4

6

8

10

12

0 5 10 15 20 25

Feed concentration of benzene (molar% in cyclohexane)

Part

ial p

erm

eabi

lity

of

benz

ene

(mol

/m2sk

Pa)

x 1

00.10.20.30.40.50.60.70.8

Part

ial p

erm

eabi

lity

of

cycl

ohex

ane

(CH

)

(mol

/m2 sk

Pa)

x 10

8

F0Bz F2Bz F4Bz F6Bz F8BzF8CH F0CH F2CH F4CH F6CH

Fig. 9. Variation of partial permeability with feed concentration of benzene (Bz) at30 �C.

zene in feed. Above this feed concentration PV or diffusion selectiv-ity decreases almost exponentially. Diffusion selectivity is alsoobserved to be much higher than sorption selectivity at any feedconcentration which signifies domination of the overall solvent

0

2

4

6

8

10

12

14

0 5 10 15 20 25

Feed conc. of Bz (molar% in CH)

Fuga

city

(K

Pa)

Fugacity of Bz (Kpa)

Fugacity of CH (Kpa)

Fig. 10. Variation of fugacity with feed conc. of benzene (Bz) at 30 �C.

0

20

40

60

80

100

120

140

0 5 10 15 20 25

Feed concentration of benzene (molar% in cyclohexane)

Intr

insi

c m

embr

ane

sele

ctiv

ity f

or b

enze

ne (

-)

F0 F2 F4 F6 F8

Fig. 11. Variation of intrinsic membrane selectivity for benzene with feed concen-tration of benzene at 30 �C.

S.B. Kuila, S.K. Ray / Separation and Purification Technology 123 (2014) 45–52 51

transport through the membranes by diffusion. In fact, diffusionrate of benzene is much higher than diffusion rate of cyclohexanebecause of smaller size and kinetic diameter of benzene [1]. This

Table 2Comparison of flux and selectivity of the present membranes with reported membranes.

Membrane Feed conc. benzene (wt%) Temperature �C Flux (kg/m2 h)/thi

PANPMA 50 30 0.07/60PVDC–PVC 10 50 0.009 /49PVC–EVA 10 30 0.028/20PANSTY 50 30 0.052/50PEMA–EGDM 10 40 0.19 /50PEA 50 50 0.62/42F0 13.3 30 1.54/50F8 13.3 30 0.713/50

may be the reason for higher diffusion selectivity of benzene.Sorption selectivity of benzene may be due to increased interactionof its aromatic pi electrons with membranes [1]. Permeate concen-tration of benzene is observed to be much higher than its corre-sponding feed concentration as observed in Fig. 8. From Fig. 8 itis also observed that permeate and membrane phase (obtainedfrom sorption experiments) benzene concentration is much higherthan its vapor – liquid equilibrium data (VLE). Permeate benzeneconcentration is also observed to be much higher than its mem-brane phase concentration which may be ascribed to high diffusionrate of benzene through the membranes. The filler may also causepreferential benzene adsorption and thus membrane phase as wellas permeate benzene concentration increases with increase in fillerwt% in membrane.

3.4.2. Permeability and intrinsic membrane selectivityEffect of feed molar concentration of benzene on partial perme-

ability of benzene and cyclohexane is shown in Fig. 9. From Fig. 9 itis observed that like partial flux benzene permeability is also muchhigher than cyclohexane permeability at any feed concentrationsignifying benzene selectivity of the membranes. However, partialpermeability shows a different trend than partial flux. It is ob-served that benzene permeability increases with feed concentra-tion up to around 7% benzene and thereafter the change ofpermeability with concentration is not very significant. However,above around 20% benzene in feed the rate of increase of benzenepermeability is very high. On the other hand variation of cyclohex-ane permeability up to this feed concentration is marginal butabove this concentration (>20%) cyclohexane permeability also in-creases at a much higher rate indicating plasticization of the mem-branes. Permeability is the ratio of flux and fugacity (Eq. 6a). FromFig. 10 it is observed that fugacity of benzene increases with feedconcentration. Initially up to around 7% benzene in feed, the rateof increase of flux more than offset the rate of increase of fugacity.Hence, benzene permeability increases. At higher feed concentra-tion rate of increase of benzene flux decreases and hence thechange of permeability with feed concentration is not significant.However, above 20% benzene in feed, partial flux of both benzeneand cyclohexane increases significantly (Fig. 6). Thus, permeabilityalso increases at a much higher rate. Further, fugacity of cyclohex-ane decreases with increase in feed benzene concentration as ob-served in Fig. 10. Hence, the variation of its permeability up toaround 20 M% feed concentration is also not very significant sincethe rate of increase of cyclohexane flux up to this concentrationrange is marginal as observed in Fig. 6. Intrinsic membrane selec-tivity for benzene is defined as the ratio of permeability of benzeneand cyclohexane (Eq. (7)). The variation of intrinsic membraneselectivity of the membranes with feed concentration is shown inFig. 11. It is observed that the membrane selectivity remains mar-ginally constant up to 20% benzene in feed. However, above thisconcentration, membrane selectivity decreases drastically indicat-ing plasticization of the membranes. Similar to flux and permeabil-ity, membrane selectivity is also observed to increase with increase

ckness of membrane (lm) Separation factor of benzene (–) References

10.5 [10]6 [35]24 [36]22.5 [37]6.7 [38]7.1 [39]88.7 Present work212 Present work

52 S.B. Kuila, S.K. Ray / Separation and Purification Technology 123 (2014) 45–52

in filler concentration from F0 (containing 0% filler) to F8 (8% filler)membrane.

3.4.3. Comparison of present work with reported dataThe flux and benzene selectivity of the present membranes are

compared with other reported membranes in Table 2. From the gi-ven data present in Table 2 it is evident that the present unfilled F0and filled F8 membranes show much higher flux and benzeneselectivity than most of the reported membranes.

4. Conclusion

Several blend membranes were prepared by solution blendingof varied compositions of CMC and SA. It was found that flux in-creases with increase in concentration of CMC in the blend and ablend membrane containing 25 wt% sodium alginate and 75 wt%CMC yielded optimum flux and selectivity for19.6 wt% benzenein cyclohexane. The blends containing CMC above 75 wt% werenot stable during pervaporation experiment. This blend membranewas further filled with 2%, 4%, 6% and 8% organophilic filler to ob-tain four filled membranes and these four filled membranes andthe unfilled membranes were used for separation of benzene fromits mixture with cyclohexane over the concentration range of 0.5–20 wt% benzene in cyclohexane. The unfilled membrane showedhigher flux than the filled membranes. However, the filled mem-branes showed much higher benzene selectivity without muchsacrifice in flux.

Acknowledgement

The authors are grateful to DST-SERC for sponsoring the works.

References

[1] J.P. Garcia, A. Villaluenga, T. Mohammadi, A review on the separation ofbenzene/cyclohexane mixtures by pervaporation processes, J. Membr. Sci. 169(2000) 159–174.

[2] Q.F. An, J.W. Qian, H.B. Sun, L.N. Wang, L. Zhang, H.L. Chen, Compatibility ofPVC/EVA blends and the pervaporation of their blend membranes for benzene/cyclohexane mixtures, J. Membr. Sci. 222 (2003) 113–122.

[3] D. Sarkhel, D. Roy, M. Bandyopadhyay, P. Bhattacharya, Studies on separationcharacteristics and pseudo-equilibrium relationship in pervaporation ofbenzene//cyclohexane mixtures through composite PVA membranes on PANsupports, Sep. Purif. Technol. 30 (2003) 89–96.

[4] L. Lu, F. Peng, Z. Jiang, J. Wang, Poly(vinyl alcohol)/chitosan blend membranesfor pervaporation of benzene/cyclohexane mixtures, J. Appl. Polym. Sci. 101(2006) 167–173.

[5] L. Lu, H. Sun, F. Peng, Z. Jiang, Novel graphite-filled PVA/CS hybrid membranefor pervaporation of benzene/cyclohexane mixtures, J. Membr. Sci. 281 (2006)245–252.

[6] H. Sun, L. Lu, F. Peng, H. Wu, Z. Jiang, Pervaporation of benzene/cyclohexanemixtures through CMS-filled poly(vinyl alcohol) membranes, Sep. Purif.Technol. 52 (2006) 203–208.

[7] F. Peng, C. Hub, Z. Jiang, Novel poly(vinyl alcohol)/carbon nanotube hybridmembranes for pervaporation separation of benzene/cyclohexane mixtures, J.Membr. Sci. 297 (2007) 236–242.

[8] T. Uragami1, K. Tsukamoto1, T. Miyata1, T. Heinze, Permeation and separationcharacteristics for benzene/cyclohexane mixtures through tosylcellulosemembranes in pervaporation, Cellulose 6 (1999) 221–231.

[9] T. Uragami1, K. Tsukamoto1, T. Miyata1, Performance of butyrylcellulosemembranes for benzene/cyclohexane mixtures containing a low benzeneconcentration by pervaporation, Biomacromolecules 5 (2004) 2116–2121.

[10] Q.F. An, J. Wen, Polyacrylonitrile-block-poly(methyl acrylate) membranes 2:swelling behavior and pervaporation performance for separating benzene/cyclohexane, J. Membr. Sci. 313 (2008) 60–67.

[11] M. Yoshikawa, M. Matsumoto, Y. Igarashi, Specialty polymeric membranes.XIV. Pervaporation of benzene/cyclohexane mixtures through modifiedpolyamide membranes, J. Appl. Polym. Sci. 88 (2003) 183–188.

[12] A.E. Yildirima, N.D. Hilmioglub, S. Tulbentci, Separation of benzene/cyclohexane mixtures by pervaporation using PEBA membranes,Desalination 219 (2008) 14–25.

[13] S. Maji, S. Banerjee, N.C. Pradhan, Separation of benzene/cyclohexane mixtureusing semifluorinated aromatic poly(ether amide) membranes with andwithout cardo unit in the main chain, Sep. Purif. Technol. 70 (2009) 128–135.

[14] Y.C. Wang, C.L. Li, J. Huang, C. Lin, K.R. Lee, D.J. Liawd, J.Y. Lai, Pervaporation ofbenzene/cyclohexane mixtures through aromatic polyamide membranes, J.Membr. Sci. 185 (2001) 193–200.

[15] W. Grabczyk, Effect of the hard segment domains on the permeation andseparation ability of the polyurethane-based membranes in benzene/cyclohexane separation by pervaporation, J. Membr. Sci. 282 (2006) 225–236.

[16] S.T. Kao, F. Joshua, W. Shingjiang, J. Lue, Sorption, diffusion, and pervaporationof benzene/yclohexane mixtures on silver–Nafion membranes, Desalination149 (2002) 35–40.

[17] T. Sakaguchi1, K. Yumoto1, G. Kwak, M. Yoshikawa, T. Masuda1, Pervaporationof ethanol/water and benzene/cyclohexane mixtures using novel substitutedpolyacetylene membranes, Polym. Bull. 48 (2002) 271–276.

[18] J. Chen, J. Li, Y. Lin, C. Ch, Pervaporation performance of polydimethylsiloxanemembranes for separation of benzene/cyclohexane mixtures, J. Appl. Polym.Sci. 112 (2009) 2425–2433.

[19] H. Benguergoura, S. Moulay, Styrene–butadiene rubber membranes for thepervaporative separation of benzene/cyclohexane mixtures, J. Appl. Polym. Sci.123 (2012) 1455–1467.

[20] N.A. Peppas, S.I. Wright, Crosslinking of polyvinyl alcohol membrane, Eur. J.Pharm. Biopharm. 46 (1998) 15–29.

[21] P. Luis, J. Degreve, B. VanderBruggen, Separation of methanol-n-butyl acetatemixtures by pervaporation: potential of 10 commercial membranes, J. Membr.Sci. 429 (2013) 1–12.

[22] P. Shao, R.Y.M. Huang, Polymeric membrane pervaporation, J. Membr. Sci. 287(2007) 162–179.

[23] R.W. Baker, J.G. Wijmans, Y. Huang, Permeability, permeance and selectivity: apreferred way of reporting pervaporation performance data, J. Membr. Sci. 348(2010) 346–352.

[24] M. Saraswathi, B. Viswanath, Separation of water-isopropyl alcohol mixtureswith novel hybrid composite membranes, J. Appl. Polym. Sci. 126 (2012)1867–1875.

[25] M. Ionita, M.A. Pandele, I. Horia, Sodium alginate/graphene oxide compositefilms with enhanced thermal and mechanical properties, Carbohyd. Polym. 94(2013) 339–344.

[26] M.S. Islama, M.R. Karim, Fabrication and characterization of poly(vinylalcohol)/alginate blend nanofibers by electrospinning method, Colloid Surf.A: Physicochem. Eng. Aspects 366 (2010) 135–140.

[27] M. Wang, L. Xu, H. Hu, M. Zhai, J. Peng, Y. Nho, J. Li, G. Wei, Radiation synthesisof PVP/CMC hydrogels as wound dressing, Nucl. Instrum. Methods B 265(2007) 385–389.

[28] C. Chang, B. Duan, J. Cai, L. Zhang, Superabsorbent hydrogels based on cellulosefor smart swelling and controllable delivery, Eur. Polym. J. 46 (2010) 92–100.

[29] Y. Dong, W. Dong, Y. Cao, H.Z. Yanan, Z. Ding, Preparation and catalytic activityof Fe alginate gel beads for oxidative degradation of azo dyes under visiblelight irradiation, Catal. Today 175 (2011) 346–356.

[30] S. Kaihara, Y. Suzuki, K. Fujimoto, In situ synthesis of polysaccharidenanoparticles via polyion complex of carboxymethyl cellulose and chitosan,Colloid Surf. B 85 (2011) 343–348.

[31] P.M. Davidovich, P.H. Bianco, Physical and structural characteristics ofacrylated poly(ethylene glycol)–alginate conjugates, Acta Biomater. 7 (2011)2817–2825.

[32] M.S. Kim, S.J. Park, B.K. Gu, C.H. Kim, Ionically crosslinked alginate–carboxymethyl cellulose beads for the delivery of protein therapeutics, Appl.Surf. Sci. 262 (2012) 28–33.

[33] J.Y. Qian, W. Chen, W.M. Zhang, H. Zhang, Adulteration identification of somefungal polysaccharides with SEM, XRD, IR and optical rotation: a primaryapproach, Carbohyd. Polym. 78 (2009) 620–625.

[34] C.E. Rogers, Permeation of gases and vapors in Polymers, in: J. Comyn (Ed.),Polymer Permeability, Elsevier Applied Sciences, New York, 1985, pp. 11–73.

[35] M. Yoshikawa, S. Takeuchi, Specialty polymeric membranes 13. Separation ofbenzene/cyclohexane mixtures through poly (vinyldene chloride-co-vinylchloride)-graft-poly (butyl methacrylate) membranes, Sep. Sci. Technol.35 (2000) 1878–1883.

[36] Q.F. An, J.W. Qian, H.B. Sun, L.N. Wang, L. Zhang, H.L. Chen, Compatibility ofPVC/EVA blends and the pervaporation of their blend membranes for benzene/cyclohexane mixtures, J. Membr. Sci. 222 (2003) 113–122.

[37] S.K. Ray, S.B. Sawant, J.B. Joshi, Vi.G. Pangarkar, Development of new syntheticmembranes for separation of benzene–cyclohexane mixtures bypervaporation: a solubility parameter approach, Ind. Eng. Chem. Res. 36(1997) 5265–5276.

[38] Y. Wang, S. Hirakawa, H. Wang, K. Tanaka, H. Kita, K. Okamoto, Pervaporationproperties to aromatic/non-aromatic hydrocarbon mixtures of cross-linkedmembranes of copoly(methacrylates) with pendant phosphate andcarbamoylphosphonate groups, J. Membr. Sci. 199 (2002) 13–27.

[39] S. Maji, S. Banerjee, Synthesis and characterization of new meta connectingsemifluorinated poly(ether amide)s and their pervaporation properties forbenzene/cyclohexane mixtures, J. Membr. Sci. 360 (2010) 380–388.

本文献由“学霸图书馆-文献云下载”收集自网络，仅供学习交流使用。

学霸图书馆（www.xuebalib.com）是一个“整合众多图书馆数据库资源，

提供一站式文献检索和下载服务”的24 小时在线不限IP

图书馆。

图书馆致力于便利、促进学习与科研，提供最强文献下载服务。

图书馆导航：

图书馆首页 文献云下载 图书馆入口 外文数据库大全 疑难文献辅助工具