Pervaporation performance of polydimethylsiloxane membranes for separation of benzene/cyclohexane...

Pervaporation Performance of PolydimethylsiloxaneMembranes for Separation ofBenzene/Cyclohexane Mixtures

Jian Chen, Jiding Li, Yangzheng Lin, Cuixian Chen

Department of Chemical Engineering, Tsinghua University, Beijing, 100084, China

Received 19 June 2008; accepted 5 September 2008DOI 10.1002/app.29799Published online 18 February 2009 in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: Crosslinked polydimethylsiloxane/polye-therimide (PDMS/PEI) composite membranes were pre-pared, in which asymmetric microporous PEI membraneprepared with phase inversion method was acted as themicroporous supporting layer in the flat-plate compositemembrane. The different function composition of thePDMS/PEI composite membranes were characterized byreflection Fourier transform infrared (FTIR) spectroscopy.The surface and section of PDMS/PEI composite mem-branes were investigated by scanning electron microscope(SEM). The composite membranes prepared in this workwere employed in pervaporation separation of benzene/cyclohexane mixtures. Effects of feed temperature, feed

composition, concentration of crosslinking agent on theseparation efficiency of benzene/cyclohexane mixtureswere investigated experimentally. In addition, the swel-ling rate and stableness of composite membrane duringlong time operation were studied, which should be sig-nificant for practical application. The results demon-strated that the pervaporation method could be veryeffective for separation of the benzene/cyclohexane mix-tures. VVC 2009 Wiley Periodicals, Inc. J Appl Polym Sci 112:2425–2433, 2009

Key words: benzene/cyclohexane; polydimethylsiloxane;polyetherimide; composite membrane; pervaporation

INTRODUCTION

Organic mixtures have conventionally been sepa-rated by extractive distillation, extraction, andadsorption process while these separation technolo-gies need high capital investment and high energyconsumption.1 Pervaporation is a very promisingmembrane technology for separation of organic/or-ganic mixtures, among which separation of aromatichydrocarbons from aliphatic hydrocarbons is a veryimportant target.2 Especially, benzene/cyclohexanemixture is a tough system because both have veryclose boiling points (only 0.6�C difference) andapproximately equal molecular volumes.3 Conse-quently, a conventional distillation process is notpractical for separation of benzene and cyclohexaneon account of complexity, high energy consumption,and operating costs. Furthermore, high purity cyclo-

hexane is most often required by petrochemicalindustries, especially in the production of precursorsof nylon (adipic acid, caprolactam) and also inpaints and varnishes.4 For all these reasons, theindustry has always been eager to look for a viablealternative to the conventional benzene/cyclohexaneseparation processes.Permeability through polymeric membranes is

mainly dependent on solubility and diffusion differ-ences between the components of the mixture.Because of the insignificant difference between diffu-sion of benzene and cyclohexane, a polymer with ahigh affinity toward one component of the feed isgenerally preferred as the membrane material, as itis expected to offer a good selectivity.5 Therefore, inthe separation of the benzene/cyclohexane mixturesby pervaporation, many polymer membranes havebeen studied through various ways including newmaterial synthesis, membrane modification, usingcarriers to facilitate transport, and molecular dynam-ics simulation.6–13 Besides choosing suitable mem-brane material for separating a given systemeffectively, the preparation technology of mem-branes is another very important issue, even in mak-ing dense membrane, such as the casting solvent,additive, crosslinking, temperature and moisture,and so on,14,15 among which the casting solvent is

Journal ofAppliedPolymerScience,Vol. 112, 2425–2433 (2009)VVC 2009 Wiley Periodicals, Inc.

Correspondence to: J. Li ([email protected]).Contract grant sponsor: Foundation of the State Key

Laboratory of Chemical Engineering; contract grantnumber: SKL-ChE-08A01.

Contract grant sponsor: Postdoctor Science Foundationof China; contract grant number: 023201069.

the key factor. High efficiency and low productioncosts favor composite membranes made up of apolymer layer and a porous carrier. In the field ofenvironmental protection, polydimethylsiloxane(PDMS) membranes are used most often.16 They dis-play highly satisfactory properties to separate awide range of solvents, especially those of hydro-phobic nature.17 Bai et al.18 have studied separationof acetic acid/water mixtures by silicone rubber-coated polyetherimide membranes. It was found thatthe composite membrane could become either waterselective or acetic acid selective, depending on thepore size of the support membrane and the condi-tion of the silicone rubber coating. However, solubil-ity parameter was not used to explain this result.

PDMS has not been extensively investigated forthe separation of the benzene/cyclohexane mixturesin the previous research studies nevertheless, andfew works on crosslinking modification of PDMSmembranes to enhance the pervaporation perform-ance were reported. In this study, microporous poly-etherimide (PEI) ultrafiltration membranes wereused as support layer of the composite membranes,while PDMS and PEI both have AOA bond so theirinterface can cement closely. Subsequently, PDMS/PEI composite membranes were prepared andemployed to separate the benzene/cyclohexane mix-tures. The influences of concentration of PDMS,crosslinking temperature, concentration of crosslink-ing agent, and crosslinking time, on the separationefficiency of benzene/cyclohexane mixtures wereinvestigated experimentally to obtain more practicalmembrane preparation condition for the scale up ofmembrane separation technology, which was themost important point for practical application.

EXPERIMENTAL

Materials

PEI (UltemVR -1000), as shown in Figure 1, was pur-chased from General Electric (Shanghai, China). N-methyl pyrrolidone (NMP) was obtained from Bei-jing Yili Fine Chemicals, Beijing, China. n-Heptanewas obtained from China Medicine (Group) Shang-hai Chemical Reagent Corp. PDMS (viscosity 20 Pas), ethyl orthosilicate, dibutyltin dilaurate were pur-chased from Tianjin Chemical Company of China

for the preparation of PDMS membrane. All thechemicals used in the experiments were of analyticalgrade and were used without any furtherpurification.

Membrane preparation

PEI supporting membranes preparation

PEI was used as a membrane material after beingdried in vacuum at 150�C for 6 h. 20 wt % of PEIwas dissolved in NMP at 80�C by stirring for 12 hand then the casting solution was filtrated to get ridof impurity. To remove air bubbles, the casting solu-tion was kept at room temperature for 24 h undervacuum. After degassing, the casting solution wascast on a polyester nonwoven fabric with a scraperof 150 lm thickness. The nascent membrane wasdried for 10 s at 25 � 1�C. And then, it wasimmersed into DI water at 25�C. After the immer-sion, the precipitated membranes were washed for12 h to remove residues of solvent mixtures fromthe membranes.

Pdms/PEI composite membranes preparation

PDMS, crosslinking agent ethyl orthosilicate and cat-alyst dibutyltin dilaurate were dissolved in n-hep-tane at room temperature. After degassed undervacuum, the solution was cast onto the PEI mem-brane with a scraper of 20 lm thickness. The mem-brane was first vulcanized under room temperatureto evaporate the solvent, and then an automatic elec-tric oven was introduced to complete crosslinking.Controlling the PDMS concentration or the coatingamount could produce membranes with variable toplayer thickness. The thickness of the top skin layercould be determined by means of scanning elec-tronic microscopy (SEM) spectroscopy.19

Membrane characterization

SEM spectroscopy

To investigate the membrane structure, SEM charac-terization of the prepared membranes has been car-ried out. For this purpose, the membrane sampleswere fractured in liquid nitrogen and then coated

Figure 1 The chemical structure of PEI.

2426 CHEN ET AL.

Journal of Applied Polymer Science DOI 10.1002/app

with Au/Pd under vacuum conditions. The crosssection and surface membrane morphology weretaken by SEM (JSM-6301F SEM, JEOL, Japan).

Fourier transform infrared spectra-attenuated totalreflection (FTIR-ATR)

Information about the presence of specific functionalgroups of the prepared membrane surfaces wasobtained by a Nicolet IR 560 spectrometer with hori-zontal ATR accessory equipped with a ZnSe crystal.For evaluation, a total of 32 scans were performed ata resolution of 4 cm�1 at temperature of 25 � 1�C.Meanwhile, FTIR spectra were recorded within therange of 4000 cm�1 to 400 cm�1.The software fromNicolet was used to record the spectra and for theselection of the corresponding backgrounds.

Swelling experiments

A piece of membrane sample was cut out and driedat the 60�C in vacuum drying oven for at least 48 huntil constant weight, when its mass was noted asm1. Then the piece of membrane was immersed intothe feed and soaked. When the sample kept constantweight, it was carefully blotted between filter papersto remove surface liquid, and then the weight of theswollen membrane was quickly measured and itsmass was noted as m2. All experiments wererepeated at least three times, and the results wereaveraged. The swelling rate g was defined as

g ¼ m2 �m1

m1� 100% (1)

Pervaporation experiments

Pervaporation experiment apparatus used in thisstudy was shown in Figure 2.20 The membrane was

positioned in the stainless steel permeation cell, andthe effective surface area of the membrane in contactwith the feed mixture in this cell was 22.4 cm.2 Thefeed solution was continuously circulated from afeed tank to the upstream side of the membrane inthe cell at the desired temperature by a pump, andthe feed temperature was monitored by a digitalvacuometer. Pervaporation experiments were carriedout by maintaining atmospheric pressure on oneside (feed) and about 200 Pa using a vacuum pumpon the other side (permeate). After a steady statewas obtained (about 1 h after start-up), the permea-tion was collected in the cold traps and condensedby liquid nitrogen. The compositions of the feed so-lution and permeate were analyzed by gas chroma-tography (SHIMADZU, GC-14C). The results forpervaporation were reproducible, and the errors in-herent in the pervaporation measurements were lessthan 2%. Separation performances of the membranescan be evaluated on the basis of total flux and sepa-ration factor.The permeate total flux J was determined by

measuring the weight of permeate collected in thecold trap and divided by time and the membrane’ssurface area as shown in eq. (2):

J ¼ DmA � Dt (2)

where Dm is the mass of permeate during the opera-tion time Dt at steady state, and A is the effectivemembrane area. Then, the selectivity of a membranein a binary system is obtained as,

a ¼ ya=ybxa=xb

(3)

where a is separation factor; x and y represent theweight fractions of corresponding solute in feed andpermeate, respectively; subscripts a and b refer tothe more permeable component (benzene) and theless permeable one (cyclohexane), respectively.

THEORY

The membrane material and its intrinsic propertiesare very pivotal for the separation applications.Especially for pervaporation process, solubility pa-rameter is effective to characterize the interaction in-tensity between solvent and membrane, so it is animportant way for membrane materials selection.Solubility parameter (d) was proposed first by Hilde-brand,21 which was defined as the square root of co-hesive energy for unit volume molecule. Thesolubility parameter which depended on chemicaland physical structure of material is important tocharacterize the interaction intensity among simpleliquids and closer solubility parameter results in

Figure 2 Scheme of the pervaporation apparatus.

PERVAPORATION PERFORMANCE OF POLYDIMETHYLSILOXANE 2427


higher attraction for permeation components inmembrane phase. The closer solubility parameterbetween two substances, the better it will be for theirmutual solubility. The d value of certain substancecan be represented by its three components: disper-sion power (dd), polarity power (dp) and hydrogenbond power (dh). The relation is expressed as

d2 ¼ d2d þ d2p þ d2h (4)

To a ternary system, including component A, com-ponent B and membrane, the component which isexpected to preferred permeation should exertstrong dissolution performance and have closer solu-bility parameter with polymer molecule. That is tosay, whether or not the membrane can fulfill its sep-aration goal depends on the relative permeationcapability of the membrane to components. Higherattraction results in the increased solubility for per-meation components in the membrane. However,ultra-strong affinity will lead to membrane swelling.In summary, it is a feasible way to evaluate the se-lectivity of the membrane by estimating the interac-tion between polymer and solvent molecule.

RESULTS AND DISCUSSION

Membrane materials selection for the separation ofthe benzene/cyclohexane mixtures

To select suitable membrane material for benzene/cyclohexane system, the molar volumes, collisiondiameters, solubility parameters, and some represen-tative physical properties of benzene and cyclohex-ane are listed in the Table I by query.22,23 Thesolubility parameters, shown in Table I, indicate thatthe hydrogen bonding component (dh) of benzenesolubility is stronger than that of cyclohexane. It isexpected that this high hydrogen bonding compo-nent of benzene interacts with free polar groups inthe membrane, as explained by Yamasaki et al.24

who compared the retention time of benzene, n-hex-ane, and cyclohexane through a column filled withPVA polymer. The solubility parameter of PDMSmaterial is 21.01 J1/2/cm3/2,25 which is closer to ben-zene than that of cyclohexane, so PDMS material actas top layer of the composite membrane. Generally,top layer controls the flux and selectivity of the com-posite membrane. The porous support membrane is

to provide the mechanical strength for the selectivelayer in high operational pressures, so it does notinfluence the permeation data in this work controlsthe flux and selectivity of the composite membrane.Therefore, PDMS/PEI membrane can fulfill the sepa-ration goal.On the other hand, Huang et al. have reported

that the molecular size of permeate has stronglyinfluenced the permeation and separation of ben-zene/cyclohexane mixtures in pervaporation pro-cess.26 Comparing benzene molecule with thecyclohexane molecule, the former has smaller molec-ular size than the latter as listed in Table I. Conse-quently, it is expected that the permeability of thesmaller molecular size diffuse predominantly in themembrane.

SEM photographs of PDMS/PEIcomposite membrane

SEM permits imaging cross section and surfacemembrane morphology. The cross section morphol-ogy of the PDMS/PEI membrane was shown in Fig-ure 3. As demonstrated in the SEM photographs,there is a clear boundary between the PDMS toplayer and the PEI support layer. Meanwhile, thecross-sectional structure of the PDMS/PEI compositemembrane consisted of an ultra-thin skin layerand a porous finger-like structure. Moreover, the

TABLE IProperties of Benzene and Cyclohexane

bf (�C)Density at

25�C (g/mol)Molar volume(cm3/mol)

Collisiondiameter (nm)

Solubility parameter (J/cm3)1/2

dd dp dh d

Benzene 80.1 0.87 89.4 0.526 18.4 0 2.0 18.6Cyclohexane 80.7 0.78 108.7 0.606 16.8 0 0.2 16.8

Figure 3 The cross section morphology of the PDMS/PEIcomposite membrane.

2428 CHEN ET AL.


thickness of the PDMS top layer was determined tobe about 11 lm from the SEM photograph by thescale tab. The surface morphology of the PDMS/PEImembrane was shown in Figure 4. From this figure,the originally porous surface of the PEI substratewas covered by a flat, featureless PDMS layer, andthe top PDMS layer, functioning as the basis of se-lectivity, had a nonporous and tight structure. Thesurface of the PDMS/PEI composite is dense andthere is no any pinhole or crack, which is importantfor the practical application.

FTIR spectra of PDMS/PEI composite membrane

The attenuated total reflection FTIR spectroscopy isa commonly used method to characterize the chemi-cal structure of the surface.27 The ATR techniqueenables the identification of specific molecules andgroups located within 100 nm from the surface layer.To obtain detailed information about the structuralchanges of PDMS/PEI membranes resulting fromcrosslinking modification, FTIR spectra of the surfaceof PDMS/PEI membranes were recorded in Figure 5using the ATR technique. From Figure 5(a), a peakat 1260 cm�1 was assigned to CH3 (two CH3 ofSiACH3) symmetric deform. It also displayed threenew absorbance signals at 850 cm�1 to 730 cm�1

(CH3 out-of-plane bending and SiAC stretching),2890–3000 cm�1 and 1000–1150 cm�1 (SiAOHstretching) and 860–920 cm�1 (SiAOH angle bendingvibration), respectively. Comparing with the non-crosslinked membrane [from Fig. 5(a)], the spectra ofthe crosslinked membranes [from Fig. 5(b)] dis-played that absorbance signals of SiAOH evidentlyweakened. These changes were the evidences ofcrosslinking reaction of hydroxyl-terminated PDMS

with tetraethylorthosilicate under dibutyltin dilau-rate catalysis.28 The crosslinking reaction is shownas the following formula:

Results of swelling experiment

The swelling results of PDMS/PEI composite mem-branes in the different feed concentrations weregiven in Figure 6, where it could be observed thatthe rate of swelling increased with the increasingimmerging time and then reached the equilibriumslowly within 5 h. It has also been shown that theswelling rate of membrane increased as benzeneconcentration increased. These phenomena indicated

Figure 5 FTIR spectra of PDMS/PEI composite mem-brane. (a) Before corsslinking modification, (b) Afer cross-linking modification.

Figure 4 The surface morphology of the PDMS/PEI com-posite membrane.



that benzene was more easily dissolvable in PDMSmembrane, which was due to the stronger interac-tion between benzene and membrane according tothe solubility parameter. All these experimentalresults showed consistence with the fact that ben-zene mainly stimulated the membrane swelling fur-ther and crosslinking was an effective modificationway for PDMS membrane to reduce the swelling.Sun and Ruckenstein29 also showed that the swellingof different membranes was mainly due to the sorp-tion of benzene rather than cyclohexane. Moreover,it can be observed that the maximum swelling ratein the solution was not more than 5%, which shouldbe suitable for practical application. Thus, all swel-ling behaviors aforementioned were attributed tobetter affinity of PDMS/PEI membrane for benzenerather than cyclohexane.

Effect of feed temperature on pervaporationperformance

Figures 7 and 8 illustrate the effect of feed tempera-ture on total flux and separation factor, respectively.It can be observed clearly that both permeation fluxand separation factor changed significantly with therise of feed temperature from 50 to 80�C. It can bealso seen that the flux increased and separation fac-tor decreased with feed temperature increasing,which was consistent with common rule.30 Asexpected, all the experimental evidences confirmedan increase in permeability and a decrease in selec-tivity of membranes with increasing temperature.31

One reason is that the increased mobility of the per-meating molecules in the bulk feed solution with thefeed temperature increasing from 50 to 80�C, resultsin higher partial vapor pressure and providesgreater driving force for the permeating components.

Also, the increased mobility of the permeating mole-cules within the membrane will facilitate the transferof the components. The other reason is that the ther-mal motion of polymer chains became more violentat higher temperature and caused larger availablefree volume of polymer matrix for diffusion, leadingto more availed components permeating trough themembrane. In a word, the increase in total flux withtemperature was due to the increase of the mobilityof individual permeating molecules caused both bythe temperature and by the enhanced mobility of thepolymer segments. On the other hand, the increasein the degree of swelling of the membrane with tem-perature results in more cyclohexane transport,which led to the decrease of selectivity toward

Figure 7 Effect of feed temperature on total flux [prepa-ration condition: PDMS content (30 wt %), crosslinkingtemperature (100�C), crosslinking content (20 wt %), cross-linking time (20 h)].

Figure 8 Effect of feed temperature on separation factor[preparation condition: PDMS content (30 wt %), crosslink-ing temperature (100�C), crosslinking content (20 wt %),crosslinking time (20 h)].

Figure 6 Swelling behavior of the PDMS/PEI compositemembrane, [preparation condition: PDMS content (30 wt%), crosslinking temperature (100�C), crosslinking content(20 wt %), crosslinking time (20 h)].

2430 CHEN ET AL.


benzene. By all given reasons, both mass transfercoefficients of components in the feed and sorptionof components into the membrane increase withfeed temperature,32 so total flux increased with tem-perature. Moreover, the continuous increase of feedtemperature weakened the difference of solubilityand diffusion velocity of benzene and cyclohexane,and caused the decrease of separation factor.33

Effect of feed composition onpervaporation performance

The effects of benzene content in feed on the totalflux and separation factor are depicted in Figure 9and Figure 10, respectively. The corresponding ben-zene contents in the feed under consideration werein the vicinity from 30 to 70 wt %. From Figure 9, itcan be found that the flux increased and separationfactor decreased with increasing benzene content infeed. When the benzene content in feed increased,an extensive swelling of the membrane occurred dueto the strong affinity of benzene to the membrane,which was attributed to the fact that benzene perme-ates more easily than cyclohexane at the same tem-perature. It is well-known that a remarkableswelling of polymer membranes leads to an openedmembrane structure and consequently an enhance-ment of permeation in the polymer membranes.Therefore, total flux increased due to the enhancedactivity of polymer chains and bonds.

From Figure 10, separation factor decreased withthe increase of concentration of benzene. It is to saythat the benzene permselectivity of the membraneswas lowered with increasing benzene content. Ingeneral, the permselectivity of liquid mixturesthrough polymer membranes by pervaporation

depends on both the differences in the solubility ofpermeants in polymer membranes (the sorption sep-aration process) and in the diffusivity of permeantsin polymer membranes (the diffusion separationprocess), namely the solution-diffusion theory.34

Accordingly, the permselectivity for the benzene/cyclohexane mixtures through the membranedepends on both solubility and diffusibility. In thefirst step, the benzene molecules which have ahigher affinity for the membranes than the cyclohex-ane molecules are preferentially sorbed into themembrane in the sorption process. Then, in the sec-ond step the diffusivity of these molecules in the dif-fusion separation process is significantly dependenton the molecular size and shape. In this case,increasing benzene content led to a higher permea-tion rate and lower separation factor.

Effect of concentration of crosslinking agent onpervaporation performance

PDMS material behaved stronger affinity to benzeneand this interaction increased the solubility of ben-zene in membrane to fulfill separation purpose. How-ever, the affinity was so strong that the membranewas prone to present excessive swelling. In ourexperiments, swelling of uncrosslinked membranewas too severe so that the crosslinking modificationwas quite necessary. Considering the practical appli-cation, what is more important for crosslinking modi-fication was that the swelling resistance of membraneand the steadiness of separation performanceimproved distinctly after crosslinking modification.Figures 11 and 12 reveal the effect of concentration

of crosslinking agent on pervaporation performance.As shown in Figure 11, flux decreased with the

Figure 9 Effect of concentration of benzene on total flux[preparation condition: PDMS content (30 wt %), crosslink-ing temperature (100�C), crosslinking content (20 wt %),crosslinking time (20 h)].

Figure 10 Effect of concentration of benzene on separa-tion factor [preparation condition: PDMS content (30 wt%), crosslinking temperature (100�C), crosslinking content(20 wt %), crosslinking time (20 h)].



increase of concentration of crosslinking agent withincrease from 10 to 30 wt %, and flux decreasedsharply with crosslinking agent changing between10 and 20 wt %. As shown in Figure 12, separationfactor increased at the certain range with theincrease of concentration of crosslinking agent. How-ever, separation factor decreased when concentrationof crosslinking agent exceeded 20 wt %. From FTIRspectrum results as shown in Figure 5, crosslinkinginfluenced the structure of PDMS membranes sincethe chemical connection occurred between macromo-lecules and reticular spatial structure formed, whichhas important influence on pervaporation perform-ance. Interchain-free volume lessened with the addi-tion of crosslinking agent, which led to the decreaseof permeation of feed.

Especially, exorbitant addition of crosslinkingagent brought about lower flux and membrane in-tensity decrease for solubility restrict to crosslinkingagent and these results were unfavorable for practi-cal application. For this reason, at the range of 15–20wt % of crosslinking agent, the flux and enrichmentfactor changed sharply since crosslinking agentreached equilibrium saturation. Furthermore, as theconcentration of crosslinking agent was up to 20 wt%, total flux decreased to ultra low level. Under the20 wt % of the crosslinking agent concentration con-ditions, flux and separation factor both exertedhigher level. In conclusion, 20 wt % of crosslinkingagent concentration was more practical and efficientsince ultra low level flux was unwanted for the scaleup of membrane technology.

Stable investigation of PDMS/PEI membranes

A desired PDMS/PEI composite membrane was pre-pared according to the above optimal preparationcondition of membrane, and the effect of long timeoperation on pervaporation performance of themembrane was shown as Figure 13. It was indicatedfrom Figure 13 that separation performance of themembrane was stable during the 180-h-long timeoperation of pervaporation. The results revealed thatthis kind of membrane had the property of resistingpollution and the composite membrane prepared inthis article was steady. After some time of run, swel-ling balance of the membrane reached to stable fluxand separation factor, which was consistent with theswelling results shown in Figure 6 where the maxi-mum of the swelling rate of the membrane in thefeed was no more than 5%. To sum up, average sep-aration factor of the prepared membrane came to13.2, with the corresponding flux of 218 g/(m2h)during the long time operation of 180 h. From all

Figure 11 Effect of concentration of crosslinking agent ontotal flux [preparation condition: PDMS content (30 wt %),crosslinking temperature (100�C), crosslinking time (20 h)].

Figure 12 Effect of concentration of crosslinking agent onseparation factor [preparation condition: PDMS content (30wt %), crosslinking temperature (100�C), crosslinking time(20 h)].

Figure 13 Effect of long time operation on pervaporationperformance [preparation condition: PDMS content (30 wt%), crosslinking temperature (100�C), crosslinking content(20 wt %), crosslinking time (20 h)].

2432 CHEN ET AL.


above results, the PDMS/PEI composite membranecould be used for separation of benzene/cyclohex-ane mixtures effectively.

CONCLUSIONS

Crosslinked PDMS/PEI composite membranes wereprepared, characterized by FTIR, SEM, andemployed in pervaporation separation of benzene/cyclohexane mixtures. The swelling experimentresults of the membranes demonstrated: the degreeof swelling increased with the increasing immergingtime and benzene concentration; the maximum ofthe swelling rate in the solution was no more than5%. Moreover, it was found that the increase in feedtemperature yielded higher total fluxes and lowerselectivity to benzene, and experiments also showedthat PDMS membrane was more selective to benzenethan to cyclohexane for the difference in solubilityparameter. Experimental results indicated that, withthe increase of the content of benzene in feed, thetotal flux increased correspondingly while the sepa-ration factor of benzene decreased simultaneously.

Crosslinking modification was important for theswelling resistance of membrane so that pervapora-tion performance depended on the condition ofcrosslinking. After crosslinking modification, notonly the separation factor increased distinctly, butalso the membrane performance was stable, whichwas the most significant point for practical applica-tion. Experimental results demonstrated that 20 wt% of crosslinking agent was optimal as far as fluxand separation factor were concerned. Finally, adesired PDMS/PEI composite membrane wasobtained with average separation factor of 13.2 andflux of 218 g/(m2h) during the long time operationof 180 h. From all the earlier results, separation ofbenzene/cyclohexane mixtures can be conductedeffectively by pervaporation using crosslinkedPDMS/PEI composite membranes.

References

1. Sun, E.; Ruckenstein, E. J Membr Sci 1995, 99, 273.2. Acharya, H. R.; Stern, S. A.; Liu, Z. Z.; Cabasso, I. J Membr Sci

1988, 37, 205.

3. Villaluenga, J. P.G.; Tabemohammadi, A. J Membr Sci 2000,169, 159.

4. Pandey, L. K.; Saxena, C.; Dubey, V. J Membr Sci 2003, 227, 173.5. Ray, S. K.; Sawant, S. B.; Joshi, J. B.; Pangarkar, V. G. Ind Eng

Chem Res 1997, 36, 5265.6. Tanihara, N.; Umeo, N.; Kawabata, T.; Tanaka, K.; Kita, H.;

Okamoto, K. J Membr Sci 1995, 104, 181.7. Debra, L. B.; Richard, D. N.; Carl, A. K. J Membr Sci 1997, 127,

161.8. Lue, S. J.; Wang, F. J.; Hsiaw, S. Y. J Membr Sci 2004, 240, 149.9. Pan, F.; Peng, F.; Jiang, Z. Chem Eng Sci 2007, 62, 703.10. Bai, Y. X.; Qian, J. W.; Sun, H. B.; An, Q. F. J Membr Sci 2006,

279, 418.11. Inui, K.; Okazaki, K.; Miyata, T.; Uragami, T. J Membr Sci

1998, 143, 93.12. Wolınska-Grabczyk, A. J Membr Sci 2006, 282, 225.13. An, Q. F.; Qian, J. W.; Sun, H. B.; Wang, L. N. Zhang, L.

Chen, H. L. J Membr Sci 2003, 222, 113.14. Modrzejewska, Z.; Korus, I.; Owczarz, P. J Membr Sci 2001,

181, 229.15. Tsai, H. A.; Li, L. D.; Lee, K. R.; Wang, Y. C.; Li, C. L.; Huang,

J.; Lai, J. Y. J Membr Sci 2000, 176, 97.16. Smitha, B.; Suhanya, D.; Sridhar, S.; Ramakrishna, M. J Membr

Sci 2004, 24, 11.17. Boscaini, E.; Alexander, M. L.; Prazeller, P.; Mark, T. D. Int J

Mass Spectrom 2004, 239, 179.18. Bai, J.; Founda, A. E.; Matsuura, T.; Hazlett, J. D. J Appl

Polym Sci 1993, 48, 999.19. Conesa, A.; Gumı, T.; Palet, C. J Membr Sci 2007, 287, 29.20. Qi, R.; Wang, Y.; Li, J.; Zhao, C.; Zhu, S. J Membr Sci 2006,

280, 545.21. Hildebrand, J. H.; Scott, R. L. The Solubility of Nonelectro-

lytes, 3rd ed.; Reinhold: New York, 1950.22. Stichlmair, J. G.; Fair, J. R. Distillation, Principles and Practice;

Wiley/VCH: New York, 1998, pp 246–249.23. Rautenbach, R.; Albrecht, R. Membrane Processes; Wiley: New

York, 1989.24. Yamasaki, A.; Shinbo, T.; Mizoguchi, K. J Appl Polym Sci

1997, 64, 1061.25. Lin, L.; Kong, Y.; Wang, G.; Qu, H.; Yang, J.; Shi, D. J Membr

Sci 2006, 285, 144.26. Huang, R. Y. M.; Lin, V. J. C. J Appl Polym Sci 1968, 12, 2615.27. Hillborg, H.; Gedde, U. W. IEEE Trans Dielectr Electr Insul

1999, 6, 5.28. Zhao, C.; Li, J.; Jiang, Z.; Chen, C. Eur Polym J 2006, 42, 615.29. Sun, F.; Ruckenstein, E. J Membr Sci 1995, 99, 273.30. Inui, K.; Tsukarnoto, K.; Miyata, T.; Uragami, T. J Membr Sci

1998, 138, 67.31. Wang, L.; Li, J. Lin, Y.; Chen, C. J Membr Sci 2007, 305, 238.32. Jiraratananon, R.; Chanachai, A.; Huang, R. Y. M.; Uttapap, D.

J Membr Sci 2002, 195, 143.33. Lin, L.; Wang, G.; Qu, H.; Yang, J.; Wang, Y.; Shi, D.; Kong, Y.

J Membr Sci 2006, 280, 651.34. Miyata, T.; Iwamoto, T.; Uragami, T. J Appl Polym Sci 1994,

51, 2007.



Pervaporation performance of polydimethylsiloxane membranes for separation of benzene/cyclohexane...

Documents

Transcript of Pervaporation performance of polydimethylsiloxane membranes for separation of benzene/cyclohexane...