Journal of Colloid and Interface Science 343 (2010) 622–627

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier .com/locate / jc is

Nanoporous polymer – Clay hybrid membranes for gas separation

Guillaume Defontaine a, Anne Barichard a, Sadok Letaief a, Chaoyang Feng b, Takeshi Matsuura b,Christian Detellier a,*

a Centre for Catalysis Research and Innovation and Department of Chemistry, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5b Department of Chemical Engineering, University of Ottawa, Ottawa, Ontario, Canada K1N 6N5

a r t i c l e i n f o

Article history:Received 3 August 2009Accepted 20 November 2009Available online 26 November 2009

Keywords:Separation membranesInorganic membranesPolydimethylsiloxaneClay mineralsSepioliteMontmorilloniteSi-29 MAS NMRNanohybrid materialsThermal gravimetric analysisScanning electron microscopy

0021-9797/$ - see front matter � 2009 Elsevier Inc. Adoi:10.1016/j.jcis.2009.11.048

* Corresponding author.E-mail address: dete@uottawa.ca (C. Detellier).

a b s t r a c t

Nanohybrid organo–inorgano clay mineral-polydimethylsiloxane (PDMS) membranes were prepared bythe reaction of pure and/or modified natural clay minerals (Sepiolite and montmorillonite) with PDMS inhexane, followed by evaporation of the solvent at 70 �C. The membranes were characterized by means ofXRD, SEM, ATD-TG and solid state 29Si magic angle spinning (MAS) and cross-polarization (CP) CP/MASNMR. The morphology of the membranes depends on the content loading of clay mineral. For low con-tent, the membrane composition is homogeneous, with well dispersed nanoparticles of clay into thepolymer matrix, whereas for higher clay content, the membranes are constituted also of a mixture of welldispersed nanoparticles into the polymer, but in the presence of agglomerations of small clay particles.Quantitative 29Si MAS NMR demonstrated a strong correlation between the clay content of the membraneand the average length of the PDMS chain, indicating that the nanohybrid material is made of clay par-ticles covalently linked to the PDMS structure. This is particularly the case for Sepiolite with has a highdensity of Q2 silanol sites. The separation performances of the prepared membranes were tested for CO2/CH4 and O2/N2 mixtures. The observed separation factors showed an increase of the selectivity in the caseof CO2/CH4 in comparison with membranes made from PDMS alone under the same conditions.

� 2009 Elsevier Inc. All rights reserved.

1. Introduction

Membrane and polymer technologies have allowed industrialapplications of separation membranes [1–4]. Polymers presentusually high permeabilities to gases or vapors, and can easily pro-duce thin films. Silicon rubbers such as polysiloxanes, whoserepeat unit is given by [SiRR’O], are semi-inorganic rubbery poly-mers. They have larger temperature stability ranges than organicpolymers. As a consequence, they were largely used in commercialapplications, and are among the most studied polymers [5,6]. How-ever, polysiloxanes remain poorly selective. An alternative tomembranes made of organic polymers is the fabrication of pureinorganic membranes. Inorganic materials exhibit usually highthermal and chemical stability. Their mechanical and structuralproperties give them a real potential for applications inseparation processes [7–11]. However, these materials suffer fromlow processibility and higher costs than organic polymericmembranes.

These observations have lead to the development of polymer-inorganic nanohybrid materials, which should combine the advan-tages of the two types of materials for separation purposes; that isthe high processibility of polymers and the high selectivity and

ll rights reserved.

permeability of inorganic materials. Thus, various combinationsof polymers and inorganic materials have been extensively studiedin the past years. Among others, one can cite poly(caprolactone)/montmorillonite system [12], poly(amide)/dodecyl-sulfate mont-morillonite system [13] polyimide/layered aluminophosphate[14] or carbon molecular sieves/polyimide [15].

Among the clay minerals and clay-like materials used as inor-ganic component, montmorillonite, a member of the smectitegroup, is the most common material, because of its low cost andhigh availability, but also because montmorillonite presents a rel-atively high cationic exchange capacity and is easily expandable,which allows the intercalation of a wide range of organic species.Among other systems that have been developed with silicatematerials are for example zeolites or amorphous silica in poly-dimethylsiloxane [16].

Much less attention has been paid to another clay family, thepalygorskite–Sepiolite group. These clays are talc-like 2:1 layeredsilicates characterized by the inversion of one every six Si–O–Sibonds in their tetrahedral layer. The consequence of this inversionis that the octahedral layer is interrupted, although the silicon lay-ers are continuous, giving a tunnel-like structure, with a tunnelcross section of approximately 3.7 Å � 10.6 Å [17,18]. The coordi-nation sphere of the magnesium atoms located on the border ofthe octahedral layers is completed by coordination water mole-cules strongly bound to these magnesium atoms. The tunnels are

G. Defontaine et al. / Journal of Colloid and Interface Science 343 (2010) 622–627 623

filled completely with water molecules which can be easily re-moved by heating at 75 �C. Palygorskite and Sepiolite can be usedas catalysts [19–22] and are suitable for the formation of nanohy-brid materials resulting from the intercalation in their tunnels ofrelatively small organic molecules such as pyridine [23–25], meth-anol, ethanol [26] or acetone [27]. Palygorskite was used by theMayas to produce Maya Blue, a pigment made of indigo extractedfrom a local plant strongly bound to the clay mineral [28–32].

In this work, membranes were prepared from hybrid materialsmade of polydimethylsiloxane (PDMS) and Sepiolite and tested fortheir gas separation performance. As in the case of other hybridpolymer/inorganic membranes, Sepiolite is expected to increasethe selectivity of the membrane, whereas the permeability is ex-pected to be reduced compared to pure polysiloxane membranes.The effect of the presence of the clay particles on the structure ofthe macromolecular segments was also investigated by 29Si MASNMR. The separation results obtained with PDMS/Sepiolite mem-branes are improved in comparison to the PDMS membranes.

2. Experimental

2.1. Materials

Sepiolite (SepSp-1) and montmorillonite (SWy-1) were obtainedfrom the Source Clays Repository of the Clay Minerals Society (Pur-due University), with chemical compositions (%) of SiO2 (52.9), MgO(23.6), Al2O3 (2.56), Fe2O3 (1.22), FeO (0.3), MnO (0.13), K2O (0.05)for Sepiolite, and SiO2 (62.9), Al2O3 (19.6), TiO2 (0.090), Fe2O3

(3.35), FeO (0.32), MnO (0.006), MgO (3.05), CaO (1.68), Na2O(1.53), K2O (0.53), F (0.111), P2O5 (0.049), S (0.05) for montmorillon-ite, respectively. Polydimethylsiloxane (PDMS) was purchased fromgeneral electric (silicon compound RTV615A and cross-linking agentRTV615B). Tetramethyl ammonium bromide, sodium chloride andsilver nitrate were purchased from Aldrich. All chemicals were usedas received without further purification. Doubly deionized waterwas used.

2.2. Preparation of the clay fine fraction

The crude clays were purified according to previously reportedprocedures [33,34]. The Na+ homoionic samples were prepared bytreatment of clay dispersion with 1 N NaCl several times. The ex-cess of chloride was removed after a series of centrifugation andwashing using doubly deionized water. The resulting gel from cen-trifugation was dispersed again in distilled water and finally, toeliminate all residual chloride ions (determined by the silver ni-trate test), the suspension was membrane-dialyzed. The resultingmaterials material was then air-dried at 60 �C, ground and sievedthrough 100 mm mesh. Sepiolite fine fraction and montmorillonitefine fraction will be respectively labeled Sep and MMT.

2.3. Preparation of the organo-clay (TMA-MMT)

The Na+-exchanged montmorillonite was dispersed in a tetra-methylammonium (TMA) bromide solution and maintained understirring overnight at room temperature to force the exchange of thesodium cations. The organo-clay suspension was filtered and driedovernight at 60 �C ground and sieved through 100 mm mesh. Theintercalation of TMA was checked by XRD.

2.4. Preparation of the membranes (PDMS/Sepiolite and PDMS/TMA-MMT)

Four grams of RTV615A and 0.4 g of RTV615B were mixed in abeaker, and Sepiolite or TMA-MMT was added in the paste. Three

milliliter of hexane (ACS reagent) was added to enhance dispersionof the clay in the paste. After evaporation of the solvent, the pastewas cast on a glass support and placed in an oven overnight at70 �C. Membranes were prepared with various clay contents; thesamples will be referred to as PSxx for Sepiolite or PTMxx, wherexx represents the weight percent of the clay in the case of Sepioliteand the organo-clay in the case of montmorillonite. Taking into ac-count that the TMA amount in the organo-clay is 5.5% (obtainedfrom TG measurements), the % amounts of MMT in the membraneremain close to the % amounts of the organo-clay. For the mem-branes PTM5, PTM10, PTM15, PTM20, the amount of MMT in themembranes is 4.73%, 9.45%, 14.2% and 18.9% respectively, allowingthe comparison between the two series of membranes, if onewanted to consider the clay content only.

A membrane with PDMS without clay particles was preparedand characterized under the same conditions.

2.5. Characterization

Solid state 29Si magic angle spinning (MAS) and cross-polariza-tion (CP) CP/MAS NMR spectra were recorded at 39.75 MHz, at roomtemperature on a Bruker ASX-200 spectrometer. Typical spinningrates of 4 kHz were used. A ramped CP pulse sequence was usedfor all 29Si cross-polarization experiments. The recycle delay timewas 2 s, and the proton 90� pulse was 4 ms. The contact time to allowthe transfer of magnetization between protons and 29Si nuclei was10 ms. The 29Si NMR signals were externally referenced to the�Si(CH3)3 resonance of tetramethylsilane (TMS). Thermogravimet-ric (TGA) and differential thermal analyses (DTA) were performedon a SDT 2960 Simultaneous DSC–TGA instrument under N2 flow(100 mL min�1) with a heating rate of 10 �C min�1. Scanning elec-tron microphotographs were taken on a SEM–EDX (Zeiss DSM960) scanning electron microscope. This technique was used to visu-alize the surface and the cross section of the membrane.

The XRD diffractogram and the solid state 13C NMR CP/MASspectrum of montmorillonite exchanged with tetramethylammo-nium (TMA-MMT) are given as supporting information (Figs. 1Sand 2S respectively).

2.6. Gas separation test

Gas permeation tests were performed with a constant pressuresystem equipped with three 3.7 cm diameter cells, under a feedpressure of 200 psig. The gas permeation rate through the mem-branes was measured with a bubble flowmeter. Before each mea-surement, gas permeation was continued for two 2 h until steadystate was reached. Four high purified gases were tested: oxygen(99% purity), nitrogen (99% purity), carbon dioxide (99% purity)and methane (99% purity). Eq. (1) was used to obtain the gas flowF values. These values were than converted for each gas in standardtemperature and pressure Two permeability ratios F(O2)/F(N2) andF(CO2)/F(CH4), called hereafter selectivities, are reported for eachmembrane.

FðgasÞ ¼ Q=t ð1Þ

Q represents the quantity of gas through the membrane (mol) dur-ing a period of time t (s).

3. Results and discussion

3.1. Scanning electron microscopy (SEM)

Scanning electron microscopy (SEM) was used to examine themorphology of the surface of the membranes as well as thedistribution of the clay nanoparticles into the polymer. The cross

0 200 400 600 800 1000

60

70

80

90

100

cde

ba

We

ight

(%

)

Temperature (ºC)

Fig. 2. TG curves of: (a) pure montmorillonite; (b) TMA-MMT intercalate; (c) PTM5membrane; (d) PTM10 membrane and (e) PTM20 membrane.

70

80

90

100

c

a

d

We

ight

(%

)

624 G. Defontaine et al. / Journal of Colloid and Interface Science 343 (2010) 622–627

section of different membranes with varying the amount of Sepio-lite content was visualized (Fig. 1).

For lower contents of clay (Fig. 1A), the surface of the mem-brane is relatively soft. The nanoparticles of Sepiolite are well dis-persed into the polymer giving a homogeneous membrane. Incontrast, for the membrane of higher Sepiolite loading (Fig. 1B),the surface of the membrane shows more roughness. Despite thepoor contrast between the organic (PDMS) and inorganic (clay)phases, one can observe that the membrane is not homogeneous:it is a mixture of well dispersed nanoparticles into the polymerand an agglomeration of small clay particles.

3.2. Thermal analysis

The thermal properties and stability of the prepared PDMS/Sepiolite and PDMS/TMA-MMT membranes were studied by ther-mal gravimetry (TG) techniques and compared to the thermogramof the pristine clays (Sepiolite and montmorillonite). TG curves ob-tained for the pristine montmorillonite, TMA-MMT and for threePDMS/TMA-MMT membranes are shown in Figs. 2 and 3. TheTGA profile of TMA-MMT (Fig. 2b) is similar to that of the startingmontmorillonite (MMT) (Fig. 2a) with an additional loss around300 �C. The first loss of material that occurs in a relatively broadrange of temperature from 50 to 100 �C, is attributed to the lossof water physically adsorbed on the surface. The second loss be-tween 300 and 400 �C is attributed to the release and decomposi-tion of the TMA from the interlayer space of clay. The third loss at640 �C corresponds to the dehydroxylation of the layers of mont-morillonite. In the case of PDMS/TMA-MMT membranes, it was ob-served that they are thermally stable up to 300 �C. They decomposein several steps. The first weight loss between 350 and 450 �C cor-responds to the release and decomposition of TMA and also to thedecomposition of a part of the PDMS like in the case of the purepolymer (see Fig. A supplementary material). The second losswhich takes place in a broad range of temperature (450–700 �C)corresponds to the decomposition of the PDMS and to the dehydr-

Fig. 1. SEM pictures of cross section of two PDMS/Sepiolite membranes: (A) PS25and (B) PS50.

0 200 400 600 800 1000

60b

Temperature (ºC)

Fig. 3. TG curves of: (a) pure Sepiolite; (b) PS5 membrane; (c) PS15 membrane and(d) PS20 membrane.

oxylation of the layers. Due to the high amount of organic loadingin the membrane, the dehydroxylation of the layers is not as wellobserved as in the case of MMT or TMA-MMT.

The same behavior was observed for PDMS/Sepiolite mem-branes. They are thermally stable up to 300 �C. Contrary to the caseof pure Sepiolite where four weight losses were observed (Fig. 3a)at 90, 260, 510 and 820 �C corresponding respectively to the loss ofzeolitic water, the first structural water, the second structuralwater, and to the dehydroxylation of the internal Mg–OH, the ther-mograms of the PDMS/Sepiolite membrane exhibit mainly twoweight losses (Fig. 3b–d). The first one, between 320 and 450 �C,is due to the removal of the first structural water and a decompo-sition of part of the PDMS. The second one, between 450 and 600 �Ccorresponds to the removal of the second structural water and thedecomposition of the remaining PDMS.

3.3. Gas separation

Fig. 4 gives the results of gas permeation experiments for aseries of PDMS/Sepiolite and PDMS/TMA-MMT membranes. For

0 5 10 15 200

2

4

6

8

10

12

14

Gas

Sel

ectiv

ity

Clay Content (% weight)

α (O2/N

2) (PDMS/Sep)

α (CO2/CH

4) (PDMS/Sep)

α (O2/N

2) (PDMS/TMA-MMT)

α (CO2/CH

4) (PDMS/TMA-MMT)

Fig. 4. Evolution of the membrane selectivity (CO2/CH4 and O2/N2) as a function ofthe clay content in the membrane.

50 0 -50 -100 -150

50 0 -50 -100 -150

50 0 -50 -100 -150

ppm

ppm

A

B

C

ppm

Fig. 5. 29Si CP/MAS NMR spectra of: (A) pure Sepiolite; (B) PDMS; (C) PDMS/Sepiolite membrane, 10% clay content (PS10).

G. Defontaine et al. / Journal of Colloid and Interface Science 343 (2010) 622–627 625

the PDMS membrane without adding Sepiolite, the selectivity was2.65 for O2/N2 and 5.75 for CO2/CH4. These values are comparableto those reported in the literature under similar conditions [35–37].

With respect to PDMS/Sepiolite membranes the selectivity ratioF(O2)/F(N2), remained constant near a value of 2.5, regardless of theSepiolite content. However, the selectivities measured for the pairCO2/CH4 displayed a different tendency. Even for low clay contents(samples PS05 and PS10) the value of the selectivity of the mem-brane prepared from Sepiolite was higher than the value observedfor the PDMS membrane. The selectivity then increased withincreasing clay content, to reach a value of 14 for the samplePS20. An increase of the clay content in the membrane thereforeenhances the separation performance of the membrane. Above acontent of 20% of clay mineral, the membrane became too brittleto obtain any reliable results.

The results obtained in the case of the microporous TMA-MMTmaterial are less straightforward to interpret. At high clay content(15–20%), the observed gas selectivity is close to the one obtainedin the case of Sepiolite. However, one observes a drop of the selec-tivity at low TMA-MMT content. A plausible interpretation for thisbehavior is the perturbation of the local structuration of the poly-meric chain of PDMS by TMA-MMT, resulting in loss of the gasselectivity displayed by PDMS. At higher clay content, the gasselectivity is governed mainly by the microporous organo-clay ina way similar to the Sepiolite case. The materials obtained by theincorporation of Sepiolite and of TMA-MMT in the polymeric net-work are structurally different: in one case, Sepiolite, the clay min-eral is covalently embedded in the polymeric network, as shown bythe 29Si NMR data (see below); in the other case, the organo-clay isphysically mixed with the polymer. This structural difference re-sults in different gas selectivity behaviors.

3.3.1. 29Si NMR solid state29Si CP/MAS NMR and quantitative 29Si MAS NMR were used to

estimate the influence of the clay content on the polymer structureand to characterize the connectivity of the silicon atoms, either inthe clay or in the polymer chain. The 29Si CP/MAS NMR spectrum ofSepiolite (Fig. 5A) has been reported and interpreted in the litera-ture. The four 29Si NMR peaks correspond to the four differenttypes of silicon atoms in the structure as indicated in Scheme 1(see S.I. section) [25,38].

The Q2 silanol groups at the edges of the structure give a 29SiNMR peak located near �85 ppm. The three different Q3 siliconatoms in the layer give peaks at �92 ppm, �94.3 ppm and�97.8 ppm, respectively for Si located at near edge (2), center (3)and edge (1), as indicated on Scheme 1 (see S.I. Section) [25,38].

Fig. 5B gives the 29Si MAS NMR spectrum of PDMS. The spec-trum displays only one sharp peak located at �22.4 ppm, whichcorresponds to the silicon of the repeat unit –(CH3)2Si–O– ofthe polymer. Since MAS NMR experiments were performed underconditions to give quantitative data, the signal of the terminal sil-icons of the chains is not observed. The number of terminal sili-cons is overwhelmed by the amount of silicons inside thepolymer chains.

The 29Si CP-MAS NMR spectrum recorded for the sample PS10 isgiven on Fig. 5C. It is representative of the other spectra obtainedfor various Sepiolite contents in the membrane. The spectrumcan be divided into two regions. The first region, from +50 ppmto �50 ppm, contains the signals corresponding to the different sil-icon atoms of the polymer structure (unit D). The peak located at+11.5 ppm is attributed to the terminal (CH3)3Si– groups of thepolymer chains. The peak at �3.3 ppm is attributed to the terminalHO(CH3)2Si– groups of the polymer chains [39–42]. The spectra ex-hibit a sharp peak at �22.4 ppm close to a broad band at�19.0 ppm. The sharp signal corresponds to the silicon in the re-peat units of the polymer –(CH3)2Si–O–. The broad one is due alsoto the silicon units in the repeat units of the polymer but close tothe terminal silicons, ((CH3)3Si–O–Si(CH3)2O–).

The second region ranges from �60 ppm to �130 ppm, and cor-responds to the signals of the various silicon in the clay structure.The three peaks observed in Fig. 5A at �92 ppm, �94.8 ppm and�98.3 ppm correspond to the Q3 silicons in the clay structure[25,38]. In the case of PDMS/Sepiolite (Fig. 5C), the spectrum pre-sents mainly three differences compared to the Sepiolite spectrum:

0 5 10 15 20 25 304.0

4.5

5.0

5.5

6.0

PDMS+Sep PDMS+TMA-MMT

Clay content (% w/w)

P1/P

2 (P

DM

S+

Sep

)

4

8

12

16

20

P1 /P

2 (PD

MS

+T

MA

-MM

T)

Fig. 7. Evolution of the ratio P1/P2 as function of the clay content in the membranesPDMS/Sepiolite and PDMS/TMA-MMT.

626 G. Defontaine et al. / Journal of Colloid and Interface Science 343 (2010) 622–627

(i) The decrease of the intensity of the peak at �94.8 ppm isattributed to the high sensitivity of this peak to the Hart-mann–Hann condition

(ii) The peak at �85 ppm, corresponding to the Q2 edge silanolgroups has disappeared.

(iii) A new peak has emerged at �110 ppm, which can be attrib-uted to Q4 Si atoms [43].

The presence of Q4 Si could result from the presence of quartzimpurities [44] but the crude Sepiolite spectra (Fig. 5A) do notshow any trace of quartz impurities. Furthermore, the peak posi-tion does not correspond to the position usually found for othersilicated minerals such as feldspath. A more plausible interpreta-tion is that these Q4 Si signals correspond to polymeric siliconsassociated to the clay structure during the polymerization reaction.This hypothesis is supported by the absence of the Q2 signal corre-sponding to the silanol moieties. This is a strong indication that Si–O–Si bonds are created between the Sepiolite particles and thepolymer chains, producing an hybrid polymeric material with inor-ganic and organic parts covalently linked, which further leads to anetwork of polymer chains and clay particles.

In order to test this hypothesis, quantitative 29Si MAS NMR wasperformed on samples with increasing clay content (Fig. 6). In thechemical shift range where the peaks of the silicon in the polymerare located (+30 ppm to �30 ppm), the peak attributed to theterminal Si–OH of the chains do no longer appear (�3.3 ppm),indicating that the amount of this terminal Si is negligible comparedto the number of (CH3)3Si– groups. The two peaks located at +11 and�22.4 ppm corresponding respectively to (CH3)3Si– and –(CH3)2Si–O– groups are still present on the pattern. The ratio P1/P2, where P1

and P2 represent the integration of the signals at �22.4 and+11 ppm respectively, gives a measure of the polymer chainslengths: the ratio decreases with decreasing chain length, sincethe relative amount of terminal (CH3)3Si– groups increases com-

30 20 10 0 -10 -20 -30 30 20 10 0 -10 -20 -30

30 20 10 0 -10 -20 -30 30 20 10 0 -10 -20 -30

30 20 10 0 -10 -20 -30

ppm

P2

P2

P2

P2

P2

P1

P1

P1

P1

P1

ppm

ppm

PS25

PS20

PS10

PS15

PS05

ppm

ppm

Fig. 6. Quantitative 29Si MAS NMR spectra of PDMS–Sepiolite membranes in the+30 ppm to �30 ppm chemical shift range.

pared to the number of repeating unit silicon. Fig. 7 shows the evo-lution of the P1/P2 ratio with increasing clay content in the PDMS/Sepiolite and PDMS/TMA-MMT membranes for which similar 29SiMAS NMR studies were made. In both cases, the length of the silox-ane chain decreases with an increase of the amount of clay mineralin the membrane. It drops from a ratio of the siloxane repeating unitsover terminal silicons of 18–6 in the case of amounts of modifiedmontmorillonite increasing from 5% to 20% weight. This is inter-preted by an increase of the available edge silanol groups found atthe defaults of the montmorillonite structure forming bonds withthe growing polymer chain. The situation is quite different in thecase of Sepiolite. The length of the polymer chain is shorter, beingclose to only 6% for 5% of clay content, and decreasing further to lessthan 5% for 25% of clay content. This is interpreted by the high den-sity of edge silanol groups present in the Sepiolite structure, far moreabundant than in the case of montmorillonite. These observationsindicate strong cross-linking of the polymeric chain with the Sepio-lite structure, a conclusion which is confirmed by the strong reduc-tion of the intensity of the Q2 silanol group signal on the 29Si NMRspectra, coupled with the appearance of Q4 signals.

4. Conclusion

Hybrid materials were prepared from the polymerization ofPDMS in the presence of clay mineral particles, Sepiolite and tetra-methylammonium intercalated montmorillonite. In both cases theexternal surfaces silanol groups react with the growing chains ofthe polymers and a relationship was established between thelength of the polymer chain and the clay content in the hybridmaterials. An hybrid polymeric material with covalently linkedinorganic and organic parts is obtained which further leads to alarge polymeric network of PDMS oligomeric chains attached toclay particles.

Separation membranes were prepared and tested for CO2/CH4

mixtures. Particularly in the case of Sepiolite they show separationfactors increasing with the clay content, well above the separationfactor observed in the case of PDMS alone, under the sameconditions.

Acknowledgments

This work was financially supported by a Discovery Grant of theNatural Sciences and Engineering Research Council of Canada(NSERC). The Canada Foundation for Innovation and the Ontario

G. Defontaine et al. / Journal of Colloid and Interface Science 343 (2010) 622–627 627

Research Fund are gratefully acknowledged for infrastructuregrants to the Center for Catalysis Research and Innovation of theUniversity of Ottawa. Dr. Glenn A. Facey and Ms. Sheryl McDowelare acknowledged for experimental help for 29Si solid state NMRexperiments.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.jcis.2009.11.048.

References

[1] Y.S. Kang, S.W. Kang, H. Kim, J.H. Kim, J. Won, C.K. Kim, K. Char, Adv. Mater. 19(2007) 475.

[2] Y.K. Kim, H.B. Park, Y.M. Lee, J. Membr. Sci. 243 (2004) 9.[3] R.W. Baker, Ind. Eng. Chem. Res. 41 (2002) 1393.[4] C.M. Zimmerman, A. Singh, W.J. Koros, J. Membr. Sci. 137 (1997) 145.[5] J.E. Mark, Acc. Chem. Res. 37 (2004) 946.[6] I. Pinnau, Z. He, J. Membr. Sci. 244 (2004) 227.[7] Y.L.A. Leung, K.L. Yeung, Chem. Eng. Sci. 59 (2004) 4809.[8] Q.K. Wang, T. Matsuura, C.Y. Feng, M.R. Weir, C. Detellier, E. Rutinduka, R.L. Van

Mao, J. Membr. Sci. 184 (2001) 153.[9] M.R. Weir, E. Rutinduka, C. Detellier, C.Y. Feng, Q. Wang, T. Matsuura, R. Le Van

Mao, J. Membr. Sci. 182 (2001) 41.[10] V.A. Tuan, J.L. Falconer, R.D. Noble, Ind. Eng. Chem. Res. 38 (1999) 3635.[11] W.J.W. Bakker, F. Kapteijn, J. Poppe, J.A. Moulijn, J. Membr. Sci. 117 (1996) 57.[12] J. Hrobarikova, J.L. Robert, C. Calberg, R. Jerome, J. Grandjean, Langmuir 20

(2004) 9828.[13] Y.-C. Wang, S.-C. Fan, K.-R. Lee, C.-L. Li, S.-H. Huang, H.-A. Tsai, J.-Y. Lai, J.

Membr. Sci. 239 (2004) 219.[14] H-K. Jeong, W. Krych, H. Ramanan, S. Nair, E. Marand, M. Tsapatsis, Chem.

Mater. 16 (2004) 3838.[15] D.Q. Vu, W.J. Koros, S.J. Miller, J. Membr. Sci. 211 (2003) 311.[16] H. Yang, Q.T. Nguyen, Z. Ping, Y. Long, Y. Hirata, Mater. Res. Innov. 5 (2001) 81.[17] K. Brauner, A. Preisinger, Miner. Petr. Mitt. 6 (1956) 120.[18] B.F. Jones, E. Galán, Sepiolite and palygorskite, in: S.W. Bailey (Ed.), Hydrous,

Phyllosilicates, Reviews in Mineralogy, vol. 19, Mineralogical Society ofAmerica, Washington, DC, 1998. p. 631.

[19] J.B. d’Espinose de la Caillerie, J.J. Fripiat, Catal. Today 14 (1992) 125.[20] L. Zhu, S. Letaief, Y. Liu, F. Gervais, C. Detellier, Appl. Clay Sci. 43 (2009)

439.[21] R. Tao, S. Miao, Z. Liu, Y. Xie, B. Han, G. An, K. Ding, Green Chem. 11

(2009) 96.[22] P. Aranda, R. Kun, M.A. Martin-Luengo, S. Letaief, I. Dekany, E. Ruiz-Hitzky,

Chem. Mater. 20 (2008) 84.[23] G.A. Facey, W. Kuang, C. Detellier, J. Phys. Chem. B 109 (2005) 22359.[24] W. Kuang, G.A. Facey, C. Detellier, B. Casal, J.M. Serratosa, E. Ruiz-Hitzky, Chem.

Mater. 15 (2003) 4956.[25] M.R. Weir, W. Kuang, G.A. Facey, C. Detellier, Clays Clay Miner. 50 (2002) 240.[26] W. Kuang, C. Detellier, Structuration of organo-minerals: Nanohybrid

materials resulting from the incorporation of alcohols in the tunnels ofpalygorskite, in: Studies in Surface Science and Catalysis, in: M. Jaroniec, A.Sayari (Eds.), Nanoporous Materials IV, vol. 156, Elsevier B.V., 2005, pp. 451–456.

[27] W. Kuang, G.A. Facey, C. Detellier, J. Mater. Chem. 16 (2006) 179.[28] G. Chiari, R. Giustetto, J. Druzik, E. Doehne, G. Ricchiardi, Appl. Phys. A: Mater.

Sci. Process. 90 (2008) 3.[29] B. Hubbard, W. Kuang, A. Moser, G.A. Facey, C. Detellier, Clays Clay Miner. 51

(2003) 318.[30] M. Jose-Yacaman, L. Rendon, J. Arenas, M. Carmen Serra Puche, Science 273

(1996) 223.[31] H. Van Olphen, Science 154 (1966) 645.[32] R. Gettens, Am. Antiq. 27 (1962) 557.[33] L. Mercier, C. Detellier, Environ. Sci. Technol. 29 (1995) 1318.[34] G. Villemure, C. Detellier, A.G. Szabo, Langmuir 7 (1991) 1215.[35] N.M. Jose, L.A.S. De-A Prado, M.A. Schiavon, S.U.A. Redondo, I.V.P. Yoshida, J.

Polym. Sci., Part B: Polym. Phys. 45 (2006) 299.[36] U. Senthilkumar, B.S.R. Reddy, J. Membr. Sci. 232 (2004) 73.[37] T.C. Merkel, V.I. Bondar, K. Nagai, B.D. Freeman, I. Pinnau, J. Polym. Sci., Part B:

Polym. Phys 38 (2000) 415.[38] P.F. Barron, R.L. Frost, Amer. Mineral 70 (1985) 758.[39] A.S. Brandolini, D.D. Hills, NMR Spectra of Polymers and Polymer Additives,

Marcel Dekker, Inc., 2000.[40] F. Babonneau, New J. Chem. 18 (1994) 1065.[41] T. Iwamoto, K. Morita, J.D. Mackenzie, J. Non-Cryst. Solids 159 (1993) 65.[42] S. Diré, F. Babonneau, C. Sánchez, J. Livage, J. Mater. Chem. 2 (1992) 239.[43] J.F. Stebbins, Nuclear magnetic resonance spectroscopy of silicates and oxides

in geochemistry and geophysics, in: T.J. Ahrens (Ed.), Handbook of PhysicalConstants, vol. 2, American Geophysical Union, Washington, DC, 1995, p. 303.

[44] S.J. Chipera, D.L. Bish, Clay Clay Miner. 49 (2001) 398.