A

dsp(

pon©

K

1

1

spitBpt

bep

s

0d

Journal of Membrane Science 296 (2007) 131–138

Swelling of polymeric membranes in roomtemperature ionic liquids

P. Izak a,b, S. Hovorka c, T. Bartovsky c, L. Bartovska c, J.G. Crespo a,∗a REQUIMTE/CQFB, Dept. de Quımica, FCT, Universidade Nova de Lisboa, P-2829-516 Caparica, Portugal

b Institute of Chemical Process Fundamentals, Rozvojova 135, 165 02 Prague 6, Czech Republicc Department of Physical Chemistry, Institute of Chemical Technology, Technicka 5, 166 28 Prague 6, Czech Republic

Received 15 February 2006; received in revised form 1 March 2007; accepted 12 March 2007Available online 15 March 2007

bstract

This research work demonstrates how an optical technique, developed for studying the process of polymer swelling, is able to pro-uce highly reproducible data of swelling of dense polymeric membranes exposed to liquid mixtures. This technique was applied fortudying the swelling process of different polymeric membranes, from proton conducting polymers (Nafion) to non-charged hydrophobicolymers, such as a polyurethane–polybutadiene elastomer (PU/PBDO), polydimethylsiloxane (PDMS) and polyvinylidene fluoride membranePVDF).

Room temperature ionic liquids (RTILs) were chosen as solvents to be studied, due to the growing interest of using their unique physico-chemical

roperties in hybrid membrane processes. The results obtained show that the technique developed allows for studying the swelling and “deswelling”f polymeric membranes even when anisotropic phenomena take place. These results are discussed and interpreted from the perspective of theanoscale properties of the polymers.

2007 Elsevier B.V. All rights reserved.

Ionic

gtf

dttpGt

1

eywords: Membrane swelling; Swelling measurement; Anisotropic swelling;

. Introduction

.1. The swelling phenomenon

Swelling is the process of dissolution of a polymer in a definedolvent. At first, the solvent molecules slowly diffuse into theolymer to produce a swollen gel. If the polymer–polymerntermolecular forces are high, thanks to crosslinking, crys-allinity, or strong hydrogen bonding, this is all what happens.ut, if these forces are overcome by the introduction of strongolymer–solvent interactions, a second stage, the dissolution ofhe polymer, can take place [1].

The process of swelling results actually from the balance

etween repulsive and attractive phenomena. These phenom-na may include: the thermodynamic mixing between the netolymer and the solvent; the interaction between fixed charged

∗ Corresponding author. Tel.: +351 212948385; fax: +351 212948385.E-mail addresses: izak@icpf.cas.cz (P. Izak),

tepan.hovorka@vscht.cz (S. Hovorka), jgc@dq.fct.unl.pt (J.G. Crespo).

mswimh

376-7388/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.memsci.2007.03.022

liquids; Nafion

roups and free ions as happens in proton-exchange membranes;he elastic force of the polymer and also inter-chain attractiveorces.

The expansion of the polymer takes place due to the entropiciffusion of its constituent chains and their counterions [2]. Onhe other hand, swelling is countered by elastic forces withinhe polymer chain and inter-chain attractive forces. Overall,olymer–solvent systems tend to reach the minimum of theibbs energy of mixing, Gmix, which is the driving force of

he process.

.2. Solvent–solute–polymer interactions

The process of sorption and diffusion of solutes into poly-ers is characterized by the interaction of molecules of all

pecies among them and with the polymer. Their interaction

ith the polymer chains of the membrane can significantly mod-

fy the membrane properties [3–6]. Interaction among differentolecules in mixtures, in which the presence of one component

as a positive or a negative influence on the rate of transport of

1 brane

atmb

1

2

At

1s

tdswwXvs

cwatdfeaaartootnfi

uadopada

smat

1

whthwiiShRctiiadpwcIcmsmsi[

umamdahR

wb[pltp

32 P. Izak et al. / Journal of Mem

nother [7,8], may also have a significant effect on their sorp-ion and penetration into the polymer. During transport into the

embrane, two ways of mutual effecting of components maye considered [3]:

. A free volume effect, which generally leads to increasingof the diffusivity of components (i.e., macroscopically per-ceived as a plasticizing effect).

. A coupling effect, which results from interactions betweendifferent molecule species within the polymer that mayincrease or decrease their diffusivity within the membrane(i.e., retarded or accelerated diffusion due to interactioneffects).

careful choice of the solvent, solute and membrane material isherefore crucial for a successful and efficient transport process.

.3. Methods for quantitative determination of membranewelling

Membrane swelling has been described in the literature fromwo different perspectives: “macroscopic swelling”, which isefined as a large scale expansion of polymer membranes mea-ured over an area of at least 1 mm2, and is used synonymouslyith the term “bulk swelling”, and “microscopic swelling”,here the change in dimension is inferred from small-angle-ray scattering (SAXS) data. The microscopic swelling obser-ation is then defined as the increase in the equivalent Braggpacing as a result of solvation [9,10].

The macroscopic swelling is usually measured gravimetri-ally, which in many situations is not very accurate, especiallyhen volatile (methanol and toluene) or viscous (room temper-

ture ionic liquids—RTILs) solvents are used. In the later casehe difficulties, when using gravimetric methods, arise from aeficient removing of excess solvent from the membrane sur-ace. In fact, several authors reported the difficulty of removingxcess RTILs from the surface of polymeric membranes, evenfter careful cleaning and rinsing. This behaviour has beenlso demonstrated by XPS studies of the membrane surfacefter contact with RTILs. Under these circumstances, gravimet-ic techniques are rather limited and inaccurate for studyinghe swelling behaviour of polymeric membranes. Therefore,ne of the major aims of our research work was the devel-pment and validation of an instrument, and an experimentalechnique, which allow for the continuous quantitative determi-ation of macroscopic swelling kinetics and equilibrium of thinlms.

Another advantage of the proposed technique over widelysed gravimetric methods is that all data are obtained “in situ”nd, therefore, there are no effects of the external atmosphereuring transfer of the samples to the balance and highly volatiler very viscous liquids can be used. This method is also able to

erceive rather small and slow membrane swelling, which is andded advantage over gravimetric methods where it is extremelyifficult to ensure exactly the same experimental conditions forll measured samples.

1

Science 296 (2007) 131–138

Finally, it is worth mentioning that in case of anisotropicwelling of the membrane, as previously described for Nafion®

embranes, this method offers the possibility to differentiatend quantify the swelling behaviour in the x and y directions athe membrane plane.

.4. Membranes and room temperature ionic liquids

Four membranes were investigated in this study: Nafion®,hich is a proton exchange membrane, and two non-chargedydrophobic membranes (a polyurethane–polybutadiene elas-omer (PU/PBDO) and polydimethylsiloxane (PDMS)) andydrophilic polyvinylidene fluoride (PVDF) membrane. In fact,e know from previous confocal Raman spectroscopic stud-

es that the PVDF membrane does not solubilise any of themidazolium-based ionic liquids we have been working with.till, we decided to perform swelling measurements withydrophilic PVDF membranes in contact with the “hydrophilic”TIL, [C4mim] [BF4]. These materials were chosen due to theirommercial application and different physico-chemical proper-ies. The Nafion membrane is widely used as a reference materialn fuel cell studies and is suitable for these experiments due tots high mechanical, thermal and chemical stability. These char-cteristics are coupled with a high ionic conductivity, whichepends strongly on the water content of the membrane. Thisolymer consists of a polytetrafluoroethylene (PTFE) backboneith sulphonic acid side-groups arranged in intervals along the

hain. These acid groups are aggregated into clusters [11,12].t is widely believed that the properties of Nafion membranesan be derived from the microscopic phase separation of ionicaterial of the fluorocarbon matrix. In addition, by performing

mall angle X-ray scattering (SAXS) experiments on Nafionembranes with different water contents, it has been demon-

trated that the discrete entities associated with ionic clustersncrease in size as the water content of the membrane increases12–14].

PU/PBDO, PDMS and are important hydrophobic polymerssed in several separation processes, such as in vapour per-eation and pervaporation processes, where organic solutes

re removed/recovered from vapours and diluted aqueousedia. PDMS has been also successively used in spray-

rying micro-coating to protect powders from the oxidativection of the atmosphere [15]. The PVDF membrane selectedas been used in supported liquid membrane studies withTILs.

This work is focused on the study of membrane swellinghen exposed to ionic liquids. This type of compounds haseen frequently recognised as new promising “green solvents”16,17] and their use in hybrid membrane processes has beenroposed recently, namely in pervaporation [18–20], supportediquid membranes [21,22], solvent extraction [23] and nanofil-ration [24] processes. Ionic liquids exhibit extremely interestingroperties:

. Due to their negligible vapour pressure they are not lostinto the environment making possible their repeatedly usein downstream processing.

P. Izak et al. / Journal of Membrane Science 296 (2007) 131–138 133

Table 1Refractive index n, density ρ and dynamic viscosity η of “dry” ionic liquids

Ionic liquid n (20 ◦C) ρ (kg m−3) (25 ◦C) η (Pa s) (25 ◦C) Water solubility in the RTIL(gwater/lRTIL) (25 ◦C)

[C4mim] [BF4] 1.420 1120 [23] 0.12 Water miscible[ [20][ [20]

2

3

4

Miiae

2

2

b([r[u([tmr

i

t[∼ad

2

pNi[Oaupalaa

wEf

V

dm

TS

I

c

c

M

PNPP

C4mim] [PF6] 1.409, 1.409 [23] 1360 [23], 1320C8mim] [PF6] 1.423, 1.423 [23] 1220 [23], 1190

. Ionic liquids can solubilize a large range of organic moleculesand transition metal complexes, which may present reducedsolubility in conventional solvents.

. Due to selective solute-RTIL interactions it is possible toremove/recover defined target solutes from reaction media,while keeping other reactants or products.

. Due to their good thermal stability, reactive processes maytake place at high temperatures (up to around 250 ◦C), whichleads to faster kinetics in the case of endothermic reactions;

ost RTILs solubilise water in defined amounts and therefore its not possible to avoid some water dissolution during their usen most processes. For this reason it is extremely important tossess the effect of water content in RTILs, when studying theirffect on membrane swelling.

. Experimental

.1. Chemicals

Room temperature ionic liquids: [C4mim] [BF4] (1-n-utyl-3-methyl-imidazolium tetrafluoroborate), [C4mim] [PF6]1-n-butyl-3-methyl-imidazolium hexafluorophosphate), andC8mim] [PF6] (1-n-octyl-3-methyl-imidazolium hexafluo-ophosphate) were prepared using generally reported procedures25]. These ionic liquids were selected, because they are widelysed and they exhibit different physical and chemical propertiessee Table 1). The ionic liquid [C4mim] [BF4] is, in contrast toC4mim] [PF6] and [C8mim] [PF6], totally miscible with water;herefore, it was possible to measure swelling kinetics of the

ixture [C4mim] [BF4] and H2O in the whole concentrationange.

RTILs were dried under vacuum at 70 ◦C prior to experimentsn order to reach a minimum water concentration, as reported in

i8ta

able 2welling equilibrium data of membranes exposed to “pure” RTILs at 25 ◦C

onic liquid [C4mim] [BF4]

IL,usedH2O (ppm) 1190IL,usedH2O /c

IL,sat.H2O Water miscible

embrane [C4mim] [BF4] [C4m

�X (%) �Y (%) �X (%

U/PBDO 0.14 0.25 0.39afion 4.59 3.68 2.14DMS 0.05 0.00 –VDF 0.14 0.31 –

0.24 [20] 27.84 [20]0.58 [20] 15.73 [20]

he Table 2. According to the literature, the hydrophilic [C4mim]BF4] is considered “dried”, if its water content is kept below4500 ppm [25], whereas the more hydrophobic [C4mim] [PF6]

nd [C8mim] [PF6] are considered “dried” if the water contentoes not exceed ∼600 and ∼400 ppm, respectively [25].

.2. Membranes

The flat non-porous PU/PBDO membrane (Fig. 1a) was pre-ared at the Department of Material Science of the Universidadeova de Lisboa, by mixing PU (Mw = 3500 g mol−1) and PBDO

n a 3:2 ratio in toluene (60 wt% of polymer) at room temperature26]. Both PU and PBDO were acquired from Sigma–Aldrich.ne drop of cross-linking agent, dibutyltin dilaurate was also

dded. After 0.5 h of mixing with a rotation speed at 60 rpm,nder nitrogen atmosphere, the solution was cast onto a glasslate at room temperature with a Gardner knife moving atcontrolled speed of v = 5 mm s−1. To complete the cross-

inking, the polymer film was cured for 3.5 h in a vacuum ovent 75 ◦C and then for 3 days at room temperature at ambienttmosphere.

The PDMS membrane (Fig. 1b) was prepared in a similaray by mixing a solution of RTV 615A and RTV 615B (Generallectric) in a 10:1 ratio in toluene (60 wt% of polymer) at 60 ◦C

or 0.5 h. The dense membranes prepared were 60 ± 4 �m thick.The PVDF membrane was provided by Pall Corporation (FP-

ericel membrane).The Nafion® 112 membrane (Fig. 1c) provided by DuPont

e Nemours, USA, was used as received. According to theanufacturer the membrane has the following properties: nom-

nal thickness 51 �m, density 2000 kg m−3, ionic conductivity.3 S m−1, acid capacity 0.89 mequiv. g−1 and initial water con-ent 5 wt%. This membrane was specifically referred to exhibitnisotropic swelling [27].

[C4mim] [PF6] [C8mim] [PF6]

1020 1840

0.048 0.139

im] [PF6] [C8mim] [PF6]

) �Y (%) �X (%) �Y (%)

0.35 1.35 1.221.29 3.75 1.26– – –– – –

134 P. Izak et al. / Journal of Membrane Science 296 (2007) 131–138

F , (b) Pt eight

2m

kotTactmmub

tets

nsmtTt

ig. 1. (a) Preparation of PU/PBDO (60%/40%) membranes adapted from [26]he ratio of polar to non-polar material within the polymer (i.e. the equivalent w

.3. Apparatus for measurement of the swelling ofembranes

A new optical apparatus for measurement of the swellinginetics of flat polymer membranes in liquid mixtures was devel-ped and constructed at the Department of Physical Chemistry ofhe Institute of Chemical Technology, Prague, Czech Republic.he scheme of the apparatus is shown in Fig. 2. This equipmentllows for taking images in short sequences by using an opti-al digital camera with high resolution. The main advantage ofhis apparatus lies in the possibility of measuring very slow and

inute swelling of flat polymers. The accuracy of this opticalethod is also better than that of gravimetric methods, usually

sed for swelling equilibrium determination. On the other hand,y this method it is not possible to measure the change of the

dtct

DMS membrane, and (c) Nafion®, where m is equal to unity and n determines).

hickness of the polymer (z direction). This limitation is not rel-vant if changes of thickness (z) are identical to those in theransverse direction (x and y directions) as happens in isotropicwelling processes.

The digital camera Olympus Camedia 5050 (C) and the mag-ifying glass (MG-2×) are fixed at a stand with holders (SH). Aquare (either 0.5 cm × 0.5 cm or 1.0 cm × 1.0 cm) of a dry flatembrane is pierced by the spike of a circle Teflon cell (TCS),

hreaded onto the spike and put on the bottom of the hollow.o prevent lifting of the membrane after addition of the liquid,

he membrane is slightly tacked by two stainless steel wires that

o not prevent the extension of the membrane. The cell withhe membrane is then placed into a thermostated vessel (TV),ontrolled by a thermostat (Medingen U8). The camera (C) ishen switched on, together with its infra-red remote control and

P. Izak et al. / Journal of Membrane

Fig. 2. Schematic diagramme of the apparatus developed for determi-nation of swelling kinetics of membranes—C: digital camera OlympusCmt

ta8im

ifiwewtttliotmhlwtr

clmat�

tmodTo±

3

3

tuPmrtpats

Ncl[stms

oclwrctp

Fig. 3 shows the swelling equilibrium behaviour of a Nafionmembrane when exposed to a binary mixture of deionized waterand the RTIL [C4mim] [BF4]. The first striking observation isthe fact that when exposed to “pure” water the Nafion mem-

amedia 5050; MG: magnifying glass; TCS: circle Teflon cell; TVL: ther-ostated vessel; SH: stand with holders; RCT: infra-red remote control and

imer.

imer (RCT). The remote control and timer allow to take imagest regular intervals. Time intervals can be chosen from 10 to0 s. For each experiment, the time interval was chosen accord-ng to the swelling rate of the particular membrane and liquid

ixture.To exclude the influence of the liquid meniscus distortion

n the Teflon cell, blank experiments with a square Teflonlm (which does not swell in the liquid mixtures under study)ere performed for each mixture and pure solvents. The high-

st distortion caused by the meniscus was observed withater, which behaved like a magnifying glass. We measured

he apparent change of the width x and of the length y ofhe square Teflon film caused by pouring the liquid mix-ure into the cell. This change was then subtracted from theength and the width of the wet membrane (typically 2 pixels,.e. ±0.6%, where 1 pixel is 13 �m). After taking an imagef the dry membrane, from which we determined the ini-ial width x0 and length y0 of the polymer, 1 ml of liquid

ixture was added with a micropipette 200–1000 �l into theollow of the Teflon cell. The time between addition of theiquid and the first image of the wet membrane in the liquidas recorded. The images of the membrane extension were

hen taken automatically until the swelling equilibrium waseached.

The analysis of the data acquired was performed by using theommercial software Zoner Media Explorer 6, Czech Repub-ic. Dimensions of the membrane in dry and wet state were

easured in pixels and the extension was evaluated as a rel-tive change of the width �x and of the length �y comparedo the initial “dry” membrane (�x = ((x′ − x0)/x0) × 100% and

x = ((y′ − y0)/y0) × 100%, where x′ and y′ are dimensions ofhe swelled membrane). The absolute error in determining the

embrane extension is 1 pixel i.e. ±0.3% or ±13 �m. Theverall absolute error of the measurements including software

etermination and subtraction of the blank experiment witheflon square can be as high as 3 pixels i.e. ±0.9%. On thether hand, reproducibility was always better than ±0.9% i.e.39 �m in the x and y direction.

Fmc

Science 296 (2007) 131–138 135

. Results and discussion

.1. Swelling equilibrium

Table 2 presents the swelling equilibrium data measured forhe selected membranes, when exposed to the different RTILsnder study. In the case of the dense polymeric membranes ofU/PBDO, PDMS and PVDF it was not possible to observe aeasurable swelling behaviour. The results obtained for their

elative swelling are below 1%, or slightly above, i.e. withinhe experimental error. These results were confirmed by inde-endent confocal Raman spectroscopy studies using a PVDFnd PDMS [27] membrane and various RTILs. It was shownhat RTILs are not able to penetrate and “dissolve” inside thetructure of non-charged, PDMS and PVDF films.

However, when we evaluate the swelling behaviour of theafion membrane exposed to the different RTILs, we can con-

lude that swelling is taking place even if extremely dried ioniciquids are used. From confocal Raman spectroscopy studies28] we know that the cation of the RTILs can penetrate andolubilise within the structure of the Nafion membrane, wherehey interact with negatively charged sulfonic groups and water

olecules. Oppositely, the corresponding anions are not able toolubilise due to an efficient Donnan exclusion mechanism.

Additionally, as it can be observed from Table 2, the degreef swelling of the Nafion membrane seems also to include aontribution from the residual water present in these dried ioniciquids. In fact, the relative swelling observed may be relatedith the water activity of the ionic liquids used (given by the

atio between the water concentration in a given RTIL and theorresponding water saturation concentration), which controlshe competitive partitioning of the residual water from the RTILhase to the membrane.

ig. 3. Swelling equilibrium data of a Nafion membrane exposed to binaryixtures of H2O and [C4mim] [BF4] at 25 ◦C. “y” corresponds to the membrane

asting direction.

136 P. Izak et al. / Journal of Membrane Science 296 (2007) 131–138

Table 3Swelling equilibrium data of membranes exposed to binary mixtures of [C4mim] [BF4] and hexyl acetate at 25 ◦C

Membrane Hexyl acetate

1.5% (w/w) 7% (w/w) 100% (w/w)

�X (%) �Y (%) �X (%) �Y (%) �X (%) �Y (%)

PU/PBDO 1.06 1.53 2.79 2.81 42.0 44.1P 5.10N 12.3

bsdtihtmv

o∼wttccNiosBwpdbPiv

aoostt

aovemNc

smwrmbewcm[pissotsTtc

t

TS

M

PN

DMS 1.49 1.58afion 12.3 7.00

rane exhibits a clear anisotropic swelling, i.e. it swells muchignificantly in the transverse direction (x) than in the castingirection (y). Interestingly, when exposed to “pure” ionic liquidhe membrane swells isotropically. To understand this behaviourt is necessary to know the structure of a Nafion membrane andow that structure may regulate the process of solute penetra-ion and solvation [29]. Oppositely, the PU/PBDO and PDMS

embranes swell in an isotropic ways; irrespectively to the sol-ent/solute mixtures they are exposed to (see Tables 2–4).

Gierke et al. [30] concluded that the water-swollen morphol-gy of Nafion was best described by a model of ionic clusters40 A in diameter that were approximately spherical in shapeith an inverted micellar structure. Taking into consideration

he high ionic permselectivity and the requirement of a percola-ion pathway for ionic transport in Nafion membranes, the ioniclusters were further proposed to be interconnected by narrowhannels ∼10 A in size [29]. This presence of ionic clustering inafion was supported on SAXS and neutron scattering exper-

ments [12]. Using the SAXS data of Gierke and co-workersver a limited range of water contents, Litt showed that the dpacing are proportional to the volume of absorbed water [31].ased on this observation, a lamellar model, which is consistentith the bilayer structure suggested by Starkweather [32] wasroposed for the morphology of Nafion. In this model, the ionicomains are defined as hydrophilic “micelle” layers separatedy thin, PTFE-like crystallites. As water absorbs among theseTFE domains and separates them, then the increase in d spac-

ng between ionic domains is expected to be proportional to theolume fraction of water in the polymer.

The Nafion membranes were fabricated in a way that inducesn anisotropic distribution of ionic domains inside the membraner their anisotropic shape (lamellae or cylinders) and preferential

rientation in the casting (machine) direction y. Therefore, thewelling of Nafion is remarkably higher in the transverse direc-ion x (where elastic forces are weaker), which is perpendicularo the casting direction y.

shme

able 4welling equilibrium data of membranes exposed to binary mixtures of [C4mim] [BF

embrane Hexyl acetate

0% (w/w) 1.5% (w/w)

�X (%) �Y (%) �X (%) �Y (%

U/PBDO 1.02 1.02 1.99 1.71afion 4.78 3.47 13.8 7.30

4.70 33.8 33.17.20 15.8 4.40

Clearly, the Nafion membrane used in this work behavesnisotropically. Cleghorn et al. [33] also reported observationf a similar anisotropic swelling of a Nafion membrane in waterapour. The anisotropic swelling of the Nafion membrane, whenxposed to water, may be explained by the fact that small waterolecules are able to fill up the ionic lamellae domains of theafion network much easily than larger and bulkier ionic liquid

ations, such as [C4mim]+.Another interesting feature of the water–[C4mim] [BF4]

welling equilibrium is the maximum swelling observed in theid concentration range (Fig. 3). The maximum swelling occurshen the contribution of both the [C4mim]+ cation and water are

elevant. Taking into consideration the structure of the Nafionembrane we may assume that the ionic liquid cation is solvated

y water molecules and together they form large clusters thatxpand the Nafion network. The swelling equilibrium studiesere also extended to the use of mixtures of a RTIL with a non-

harged, hydrophobic solute able to solubilise in non-chargedembranes (hexyl acetate). As the liquid binary mixtures of

C4mim] [BF4] and hexyl acetate are only partially miscible werepared solutions with a maximum concentration of 7% (w/w)n hexyl acetate. The results obtained show a high and isotropicwelling of the hydrophobic membranes due to the solubili-ation of hexyl acetate, which confirms previous observationsbtained during vapour permeation and pervaporation studies;hey also show that hexyl acetate swells the Nafion membraneignificantly, as it would be expected, in an anisotropic mode.his result could be anticipated if we take into consideration

hat hexyl acetate is a relatively small and mobile molecule, inomparison with the ionic liquids tested.

This experimental work was performed repeatedly and, forhe specific situations referred, temperature seems to have no

ignificant effect on the observed swelling behaviour. Still, weave to keep in mind that temperature changes may induce aultitude of effects, which may add or cancel between them:

ffect on the activity coefficients of the feed components, effect

4] and hexyl acetate at 60 ◦C

7% (w/w) 100% (w/w)

) �X (%) �Y (%) �X (%) �Y (%)

2.91 2.25 46.2 45.314.0 7.98 15.9 5.65

P. Izak et al. / Journal of Membrane Science 296 (2007) 131–138 137

Fh

ona

3

snibfao

aANtse“

Fi

Fi“

ipbimttFNiex

ermc

ig. 4. Isotropic swelling kinetics of a PU/PBDO membrane exposed to pureexyl acetate at 60 ◦C.

n the viscosity and diffusion coefficients of the feed compo-ents (less important when equilibrium conditions are attained)nd effects on the membrane polymer matrix.

.2. Swelling kinetics

As mentioned before, one of the interesting advantages of thewelling method developed is the fact that it allows to performot only equilibrium measurements but also in situ kinetic stud-es. The first observation from Fig. 4 where a PU/PBDO mem-rane is exposed to pure hexyl acetate is the fact that swelling isast (from a process point of view), being completed in 2–3 min,nd isotropic. This figure is also a good example of the qualityf the experimental data that this method allows to achieve.

As can be seen in Fig. 5 the Nafion membrane swellsnisotropically when exposed to water, as previously discussed.

fully reversible process is supposed to occur inside theafion membrane during dehydration, which was experimen-

ally confirmed. From Fig. 5 it may be concluded that thewelling process is fully reversible when the membrane isxposed to “deswelling” conditions; as it can be observed, afterdeswelling” the �x and �y values came to zero.

ig. 5. Anisotropic swelling and “deswelling” kinetics of a Nafion membranen H2O at 25 ◦C. “y” corresponds to the membrane casting direction.

EsctwadApRaN

4

caora

ig. 6. Anisotropic swelling and “deswelling” kinetics of a Nafion membranen a binary mixture of [C4mim] [BF4] and H2O (50%/50% (w/w)) at 25 ◦C.Deswelling” was not observed during the timeframe of the experiments.

However, the kinetics of “deswelling” observed is extremelynteresting: in first place it is worth to notice that the “deswelling”rocess is much slower than swelling, which can be explainedy the fact that the water inside the Nafion membrane is partiallynvolved in solvation processes inside ionic clusters making their

olecules more bounded and structured; secondly, it seems clearhat “deswelling” in the x direction starts immediately, while inhe y direction it starts only when �x ≈ �y (see grey arrow inig. 5). Actually, this behaviour suggests that contraction ofafion in the casting (y) direction is only possible when the

nitial membrane spatial coherence is reacquired. Due to thelastic forces present in the Nafion network, contraction in thedirection proceeds when it starts in the y direction.

On the other hand, “deswelling” of the Nafion membranexposed to a binary mixture of water and [C4mim] [BF4] is noteversible at all (see Fig. 6). In this case the term “swelling”ust be used with care because, in fact, an ion-exchange pro-

ess is taking place when Nafion uptakes the ionic liquid cation.ven after a period of 100 days under “deswelling” conditions,welling of the Nafion membrane was not reversed. Taking intoonsideration the knowledge gathered in this work, we interprethis result as a consequence of the highly bounded situation of theater inside the membrane matrix. As we know RTILs exhibit

n extremely low vapour pressure, which explains why swellingue to RTIL solubilisation inside the membrane is not reversible.dditionally, it seems that all water uptaken during the swellingrocess is most likely strongly solvating the cation from theTIL inside the Nafion structure. The interactions establishedre sufficiently strong and stable to impede water to leave theafion membrane.

. Conclusions

This work demonstrates that the equipment designed andonstructed for measurement of membrane swelling is a reli-

ble and powerful tool to follow the kinetics and equilibriumf solute/solvent uptake by flat membranes. This techniqueeveals to be particularly adequate for studying situations wherenisotropic swelling occurs; other techniques based on the over-

1 brane

aa

mtulptscoo

ipm

A

aotCf

(nFPtaU

cC

R

[

[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

[

[

[

[

[(1997) 80.

38 P. Izak et al. / Journal of Mem

ll volume or mass change are not able to account for this effectnd to quantify it.

The results obtained, when studying the swelling of Nafionembranes, show also how the process of molecular solubilisa-

ion and swelling can be influenced by the fabrication conditionssed during the manufacturing process. This aspect is particu-arly important if we take into consideration that the swellingrocess itself may also reflect on the transport properties ofhe polymer. When considering the Nafion membranes undertudy, with applications in low resistance devices such as fuelells and transducers, their transport properties are obviouslyf high importance and may determine their performance underperating conditions.

The understanding of the swelling behaviour of thin filmss therefore essential for a comprehensive design of membranerocesses, from a better control of the nanoscale properties ofaterials towards membrane macroscale applications.

cknowledgements

The authors would like to thank to the Prof. M.H. Godinhond Dr. A.C. Trindade from the Department of Material Sciencef the Universidade Nova de Lisboa, for helpful discussions ando Prof. C.A.M. Afonso, L.C. Branco from the Department ofhemical Enginnering of IST/Universidade Tecnica de Lisboa,

or ionic liquid preparation.P. Izak would like to acknowledge the post-doc grant

SFRH/BPD/9470/2002) from Fundacao para a Ciencia e a Tec-ologia (FCT), Portugal, and also a Marie Curie Intra-Europeanellowships within the 6th European Community Frameworkrogramme (MEIF-CT-2004-010927). Funding of this project

hrough grant POCTI/EQU/35437/2000, FCT, Portugal, is alsocknowledge. P. Izak also thanks to Prof. Kragl from Rostockniversity, Germany for helpful cooperation.The co-authors from Czech Republic acknowledge the finan-

ial support from the grant from Ministry of Education of thezech Republic (MSM 6046137307).

eferences

[1] F.W. Billmeyer Jr., Textbook of Polymer Science, J. Wiley & Sons, NewYork, 1984.

[2] T. Tanaka, D.J. Fillmore, Kinetics of swelling of gels, J. Chem. Phys. 70(1979) 1214.

[3] R.Y.M. Huang, Pervaporation Membrane Separation Processes, Elsevier,Amsterdam, 1990.

[4] J.G. Crespo, K.W. Boddeker, Membrane Processes in Separation and Purifi-cation, Kluwer Academic Publisher, Dordrecht, 1994.

[5] J. Welti-Chanes, J.F. Velez-Ruiz, G.V. Barbosa-Canovas, Transport Phe-nomena in Food Processing, CRC Press, London, 2003 (Chapters 17 and247).

[6] L. Bartovska, T. Bartovsky, Dimensional changes of polymeric foils in liq-uid media, in: Proceedings of the 36th Microsymposium on High-swellingGels, Prague, 1995.

[7] J.P. Brun, C. Larchet, R. Melet, G. Bulvestre, Modeling of the pervaporation

of binary mixtures through moderately swelling non-reactive membranes,J. Membr. Sci. 23 (1985) 57.

[8] F. Morton, R.Y.M. Huang, V.J.C. Lin, Diffusion coefficients of liquids inpolymer membranes by a desorption method, J. Appl. Polym. Sci. 14 (1970)523.

[

[

Science 296 (2007) 131–138

[9] G. Gebel, Structural evolution of water swollen perfluorosulfonatedionomers from dry membrane to solution, Polymer 41 (2000) 5829.

10] M. Fujimura, T. Hashimoto, H. Kawai, Small-angle X-ray scattering studyof perfluorosulfonated ionomers, Macromolecules 15 (1982) 136.

11] M.A.F. Robertson, H.L. Yeager, in: M.R. Tant, K.A. Mauritz, G.L. Wilkes(Eds.), Ionomers, Blackie, London, 1992 (Chapter 7).

12] P.J. James, T.J. McMaster, J.M. Newton, M.J. Miles, In situ rehydration ofperfluorosulfonate ion-exchange membrane studied by AFM, Polymer 41(2000) 4223.

13] P.J. James, J.A. Elliott, J.M. Newton, A.M.S. Elliott, S. Hanna, M.J. Miles,Hydration of Nafion (R) studied by AFM and X-ray scattering, J. Mater.Sci. 35 (2000) 5111.

14] J.A. Elliott, S. Hanna, A.M.S. Elliott, G.E. Cooley, Interpretation of thesmall-angle X-ray scattering from swollen and oriented perfluorinatedionomer membranes, Macromolecules 33 (2000) 4161.

15] P. Gronchi, R. Del Rosso, P. Centola, R.F. Cosentino, Microencapsulationof menadione sodium bisulphite with polydimethylsiloxane by the spray-drying process: characterization by thermal analysis, J. Microencapsul. 9(1992) 183.

16] M. Freemantle, New horizons for ionic liquids, C&EN (2001) 21.17] P. Scovazzo, J. Kieft, D.A. Finan, C. Koval, D. Dubois, R. Noble, Gas

separations using non-hexafluorophosphate [PF6]-anion supported ionicliquid membranes, J. Membr. Sci 238 (2004) 57.

18] L. Gubicza, N. Nemestothy, T. Frater, K. Belafi-Bako, Enzymatic esteri-fication in ionic liquids integrated with pervaporation for water removal,Green Chem. 5 (2003).

19] P. Izak, N.M.M. Mateus, C.A.M. Afonso, J.G. Crespo, Enhanced ester-ification conversion in a room temperature ionic liquid by integratedwater removal with pervaporation, J. Sep. Purif. Technol. 41 (2005)141.

20] T. Shafer, L.C. Branco, R. Fortunato, P. Izak, C. Rodrigues, C.A. Afonso,J.G. Crespo, Opportunities for membrane separation processes using ionicliquids, in: ACS Series 902 on Ionic Liquids III, 2005.

21] R. Fortunato, C.A.M. Afonso, M.A.M. Reis, J.G. Crespo, Supported liquidmembranes using ionic liquids: study of stability and transport mechanisms,J. Membr. Sci. 242 (2004) 197.

22] L.C. Branco, J.G. Crespo, C.A.M. Afonso, Highly selective transport oforganic compounds by using supported liquid membranes based on ionicliquids, Angew. Chem. Int. 41 (2002) 2771.

23] R. Kertesz, M. Simo, S. Schlosser, Membrane-based solvent extraction andstripping of phenylalanine in HF contactors, J. Membr. Sci 257 (2005) 37.

24] J. Krockel, U. Kragl, Nanofiltration for the separation of nonvolatile prod-ucts from solutions containing ionic liquids, Chem. Eng. Technol. 26 (2003)1166.

25] J.G. Huddleston, A.E. Visser, W.M. Reichert, H.D. Willauer, G.A. Bro-ker, R.D. Rogers, Characterization and comparison of hydrophilic andhydrophobic room temperature ionic liquids incorporating the imidazoliumcation, Green Chem. 3 (2001) 156.

26] A.C. Trindade, M.H. Godinho, J.L. Figueirinhas, Shear induced finite ori-entational order in urethane/urea elastomers, Polymer 45 (2004) 5551.

27] J.A. Elliott, S. Hanna, J.N. Newton, A.M.S. Elliot, G.E. Cooley, Elimina-tion of orientation of perfluorinated ionomer membranes, Polym. Eng. Sci.(2006) 228.

28] T. Schafer, R. di Paolo, R. Franco, J.G. Crespo, Elucidating interactionsof ionic liquids with polymer films using confocal Raman spectroscopy,Chem. Commun. 20 (2005) 2594.

29] K.A. Mauritz, R.B. Moore, State of understanding of Nafion, Chem. Rev.104 (2004) 4535.

30] T.D. Gierke, G.E. Munn, F.C. Wilson, The morphology in Nafion perfluo-rinated membrane products, as determined by wide- and small-angle X-raystudies, J. Polym. Sci., Polym. Phys. 19 (1981) 1687.

31] L. Litt, H. Morton, A reevaluation of Nafion morphology, Polym. Prepr. 38

32] H.W. Starkweather Jr., W. Howard, Crystallinity in perfluorosulfonic acidionomers and related polymers, Macromolecules 15 (1982) 320.

33] S. Cleghorn, J. Kolde, W. Liu, Handbook of Fuel—Cells FundamentalsTechnology and Application, vol. 3, 2003, p. 566.