influence of the clay content and the polysulfone molar mass on nanocomposite membrane properties

G

A

Im

Pa

Sb

a

ARRAA

KPMNMS

1

bdBsenpaep

hrtmomb

(

0h

ARTICLE IN PRESS Model

PSUSC-25037; No. of Pages 11

Applied Surface Science xxx (2013) xxx– xxx

Contents lists available at SciVerse ScienceDirect

Applied Surface Science

j our nal ho me p age: www.elsev ier .com/ loc ate /apsusc

nfluence of the clay content and the polysulfone molar mass on nanocompositeembrane properties

riscila Anadãoa,∗, Rafael Rezende Montesa, Nelson Marcos Laroccab, Luiz Antonio Pessanb

Department of Metallurgical and Materials Engineering, Engineering School – University of São Paulo, Av. Prof. Mello Moraes, 2463, CEP 05508-900, Cidade Universitária, São Paulo,P, BrazilDepartment of Materials Engineering, Federal University of São Carlos, Rod. Washington Luis, km 235, CEP 13565-905, São Carlos, SP, Brazil

r t i c l e i n f o

rticle history:eceived 13 April 2012eceived in revised form 13 January 2013ccepted 14 January 2013vailable online xxx

a b s t r a c t

Polysulfone/MMT nanocomposite membranes were prepared by a congruence of the wet-phase inversionand the solution dispersion techniques. Different clay contents and two kinds of polysulfone were used inorder to investigate the changes in the nanocomposite structure as well as in the thermal and mechanicalproperties. The increase in the basal spacing with the clay content increase was revealed by SAXRD. TEMimages depicted the presence of hybrid morphology and SEM images showed that the clay particles were

eywords:olysulfoneontmorilloniteanocompositeembrane

olution dispersion

trapped inside the cross-section pores. By increasing the clay content and polysulfone molar mass, theonset temperature of decomposition was increased and the mass loss was decreased. From DTA studies,it was observed that PSf P-1700 low clay content membranes had higher values of enthalpies and theenthalpy values of PSf P-3500 membranes did not present a regular behavior. Also in the tensile tests,the increase of the clay content up to 4.0 mass% promoted the increase of elongation at break and tensilestrength.

© 2013 Elsevier B.V. All rights reserved.

. Introduction

Ceramic materials have been used not only as fillers in polymers,ut also their nanometric properties have been explored to pro-uce polymer nanocomposites with enhanced performance [1–4].y using clay, the dispersed phase in the polymer matrix is theilicate which is constituted of particles of about 100 nm in diam-ter and 1 nm in thickness [5–7]. The interest in the polymer/clayanocomposite production stems from the large spectrum of theossible applications. Features such as higher thermal, mechanicalnd chemical resistances, barrier properties, low expansibility andase of processing can be achieved through this type of nanocom-osite [8–15].

In addition to the nanocomposite technology, much attentionas been directed to membrane separation technology, which inecent years has been studied for a growing number of applicationshat requires high-performance membranes [16–21]. A largely used

ethod to prepare them is the wet-phase inversion. This process

Please cite this article in press as: P. Anadão, et al., Influence of the clay coproperties, Appl. Surf. Sci. (2013), http://dx.doi.org/10.1016/j.apsusc.2013.

ccurs by polymer dissolution in an appropriate solvent or a solventixture, solution casting as a thin film and sorption of a non-solvent

y the atmosphere with posterior film immersion in a non-solvent

∗ Corresponding author. Tel.: +5511 3091 5236; fax: +5511 3091 2275.E-mail addresses: [email protected] (P. Anadão), [email protected]

R.R. Montes), [email protected] (N.M. Larocca), [email protected] (L.A. Pessan).

169-4332/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.apsusc.2013.01.102

bath. In this last step, solvent and non-solvent exchange occurssimultaneously to various phase transformations, resulting in thefinal membrane [22–27].

In the last five years, special attention was given tothe addition of a fourth component in this ternary system,polymer/solvent/non-solvent since the following advantages werereported: solution rheology variation, macrovoid suppression,mechanical strength improvement, longer lifetime and inter-connectivity pore improvement resulting in higher permeabilitycombined with unchanged retention properties [28–33]. Studiesdescribe the use of clay as a fourth component aiming to takeadvantage of the resulting features reported above as well asthe properties of the formed polymer/clay nanocomposite. Recentwork by Hashemifard and co-authors showed the use of differenttypes of montmorillonite to prepare a polyetherimide nanocom-posite membrane via dry/wet phase inversion method. The use of1.0 mass% Closite 15A produced a membrane with the best selectiv-ity and permeance in CO2/CH4 separation due to the great adhesionbetween polymer chains and clay platelets [34]. Hande et al.prepared sulfonated poly (ether ether ketone) (SPEEK)/reactiveorganoclay (Cloisite 30B) nanocomposite membranes and prop-erties such as oxidative, thermal and mechanical stabilities

ntent and the polysulfone molar mass on nanocomposite membrane01.102

were improved by nanocomposite formation without significantlychanging the proton conductivity in comparison with the purepolymer membrane, being useful for direct methanol fuel cell appli-cation [35]. Besides that, it is known that the use of different

dx.doi.org/10.1016/j.apsusc.2013.01.102


http://www.sciencedirect.com/science/journal/01694332

http://www.elsevier.com/locate/apsusc

mailto:[email protected]





ING Model

A

2 rface

mts

wpcpnpMtsm

tabtmtc

2

2

tsfwct

2

dadatsppaw

2

uwfiotlrst

tnfc

ARTICLEPSUSC-25037; No. of Pages 11

P. Anadão et al. / Applied Su

olar mass polymers can also modify membrane properties, due tohe polymer interactions with the other components of the ternaryystem [36,37].

Therefore, considering both the solution dispersion and theet-phase inversion method and aiming to improve membraneroperties, in our previous study [38], sodium montmorillonitelay (MMT) was added to the ternary system composed ofolysulfone (PSf)/N-methyl-2-pyrrolidone (NMP)/water and theanocomposite membranes were successfully prepared from dis-ersions containing 25 mass% PSf Udel® P-1700/0.5 and 3.0 mass%MT. However, no relation was determined between clay con-

ent/molar mass of the polymer and final material properties in thattudy; therefore, an in-depth study of the interactions between clayineral platelets and PSf chains should be performed.Thus, our main purpose is to study the influence of clay con-

ent as well as PSf molar mass on the nanocomposite structurend final material properties. Nanocomposite structure was studiedy small-angle X-ray diffraction, Fourier transform infrared spec-roscopy, scanning electron microscopy and transmission electron

icroscopy. Additionally, the thermal properties were evaluated byhermogravimetry and differential thermal analysis and mechani-al properties by tensile tests.

. Material and methods

.1. Materials

Two kinds of commercial PSf (Udel® P-1700 and Udel® P-3500),he Mw* of which are 62,973 and 71,526, respectively, were kindlyupplied by Solvay Advanced Polymers. As a solvent, NMP, obtainedrom Arino’s Química Ltda, was used and as a non-solvent, distilledater was used. Sodium montmorillonite (MMT), which is a natural

lay, was from Sigma Chemical Co. and was used without furtherreatment.

.2. Nanocomposite preparation

Polysulfone nanocomposite membranes were prepared asescribed elsewhere [38], whose preparation method consists in

combination of the wet-phase inversion method and the solutionispersion technique. Different clay contents (0.5, 1.0, 2.0. 3.0, 4.0nd 5.0 mass%) were added to a 25/75 mass%/mass% PSf/NMP solu-ion and homogeneous dispersions were prepared by mechanicaltirring. Each dispersion was used to cast viscous film onto a glasslate by using a knife and was left to rest for 5 min. Then, the glasslate with the film was immersed in the non-solvent bath at 25 ◦Cnd after 1 min the membrane was formed. Finally, the membranesere dried in a vacuum oven at 80 ◦C for 6 h.

.3. Structural characterization of the PSf/MMT nanocomposites

Small-angle X-ray diffraction (SAXRD) analysis was used to eval-ate the basal spacing between the clay mineral platelets. SAXRDas performed with a Rigaku X-ray diffractometer using nickel-ltered Cu K�1 radiation source (� = 0.15418 nm) under a voltagef 40 kV and a current of 30 mA in the range of 1.5–25◦ (2�). In ordero verify if polymer chains and no water molecules were interca-ated between clay mineral layers, membranes were calcinated in aeducing atmosphere up to 300 ◦C, with a heating rate of 5 ◦C min−1,taying at 300 ◦C for 5 min and cooled at room temperature, then,heir SAXRD patterns were taken.

Transmission electron microscopy (TEM) was used to inves-


igate the clay structure in the PSf matrix. Thin films of theanocomposites were encapsulated in epoxy resins and were cut

rom these epoxy blocks at room temperature with an ultrami-rometer with a diamond knife. The samples were examined with

PRESSScience xxx (2013) xxx– xxx

a Philips CM120 transmission electron microscope operated at a120 kV acceleration voltage.

Aiming to observe clay dispersion as well as differences onmembrane surface and cross-section by adding clay or by usinga different polysulfone molar mass, scanning electron microscopy(SEM) images were collected by using a Quanta 600 FEG with a20 kV voltage. The surfaces were sputter-coated with carbon tomake the surface conductive by a Bal-tec SCD 050 metallizer.

The nanocomposite structure was also studied by Fourier trans-form infrared spectroscopy (FTIR). As the main aim of this studywas to analyze the differences between the bands of the mem-branes prepared with different clay contents and polysulfone molarmasses, membranes were prepared with the same thickness bycasting the different polymer dispersions onto a glass plate bymeans of custom made tubular casting knife (gap 250 �m) and then,this plate was immersed in the water and the samples were dried.FTIR study was performed by using a Bomem MB-10 interferome-ter for a scanning coverage from 4000 to 300 cm−1 at a resolutionof 4 cm−1 using 64 scans.

2.4. Thermal properties of the nanocomposite membranes

Thermogravimetry (TGA) and differential thermal analysis(DTA) curves were carried out on SDT-500 TA Instruments. Polymersamples, around 5 mg, were heated from 25 to 900 ◦C at 10 ◦C min−1

under flowing air (50 mL min−1).

2.5. Mechanical properties of the nanocomposite membranes

Fifteen specimens of each membrane were tested on an Instron5567 at room temperature with a crosshead speed of 50 mm min−1

according to ASTM D882-02 and tension strength and elongationat break average values as well as their standard deviations weredetermined.

3. Results and discussion

3.1. Structural characterization of the PSf/MMT nanocomposites

SAXRD is one of the most powerful techniques to study the dis-persion of clay mineral inside the polymer matrix. The basal latticeof the MMT, which was observed at 2� value of 6.0◦ (14.8 A), is typ-ical for this type of clay due to the presence of water moleculesinside the clay mineral galleries. This basal lattice shifted to highervalues when mixed with PSf, corresponding to basal spacings (d0 0 1)around 13 A (Fig. 1) as seen in Table 1.

Another point that should be emphasized is that by increasingthe clay content in the hybrids with both types of PSf, the diffrac-tion peak intensity related to the clay mineral was also increased inagreement with Homminga et al. [39]. However, this increase wasnot regular, which could imply the existence of not only interca-lated structures, but also the possibility of having tactoids and/orexfoliated portions. The diffraction peaks situated at 2� value of12◦ and 17.5◦, which corresponded, respectively to the interplanardistances of 7 and 5 A intrinsic to polysulfone.

In order to confirm that polymer chains instead of watermolecules were intercalated between clay mineral layers, SAXRDpatterns of the calcinated hybrids were taken and basal spacingswere calculated. Since the hybrids were dried at 80 ◦C, if therewere no polymer chains between these galleries, the clay min-eral platelets would collapse and the new d0 0 1 spacing would be


around 10 A. Therefore, from the basal spacings in the calcinatedmembranes (Table 1), it can be confirmed that polymer chainswere intercalated between the clay mineral layers, since the d0 0 1spacings were around 14.0 A.


ARTICLE IN PRESSG Model

APSUSC-25037; No. of Pages 11

P. Anadão et al. / Applied Surface Science xxx (2013) xxx– xxx 3

Fig. 1. SAXRD patterns of the (a) sodium montmorillonite clay and of themembranes prepared from the dispersions containing: (b) 25 mass% PSf P-1P

dcoirciotc

oc

gle MMT layer is 1 nm and the thickness of the dark lines were

TB

700/0.5 mass% MMT, (c) 25 mass% PSf P-1700/4.0 mass% MMT, (d) 25 mass% PSf-3500/0.5 mass% MMT and (e) 25 mass% PSf P-3500/4.0 mass% MMT.

From these results, it is thought that the clay mineral layers wereispersed in NMP by the increase in entropy due to the disorganizedlay mineral platelets which overcame the clay mineral plateletrganizational entropy. Then the polymer chains were adsorbedn the delaminated clay mineral layers and when the solvent wasemoved, the clay mineral platelets were reunited, filled with PSfhains, forming a multilayer ordered structure, in other words, thentercalated morphology. Moreover, from the basal spacing valuesf these nanocomposites (around 13 A), it is possible to affirm thathere was the formation of monolayers of the intercalated polymerhains into the interlayer spaces [40].


Representative TEM photomicrographs showing the morphol-gy of some of the PSf nanocomposites containing different MMTontents are compared qualitatively in this section. Fig. 2 shows a

able 1asal spacings (d0 0 1) of the non-calcinated and calcinated membranes prepared with diff

Clay content(mass%)

d0 0 1 of the non-calcinatedmembranes prepared with PSfP-1700 (A)

d0 0 1 of the non-cmembranes prepP-3500 (A)

0.5 13.3 13.1

1.0 13.3 13.3

2.0 12.9 13.5

3.0 13.2 13.4

4.0 13.3 13.2

5.0 13.5 14.0

Fig. 2. TEM micrographs of the membranes prepared from the dispersions contain-ing: (a) 25 mass% PSf P-1700/1.0 mass% MMT, (b) 25 mass% PSf P-3500/1.0 mass%MMT and (c) 25 mass% PSf P-3500/5.0 mass% MMT.

series of low magnification TEM images in which clay content isvaried. Filled materials clearly showed a particular microstructure:layered silicates appear as dark lines, PSf matrix is the dark greypart and the light grey part is related to the pores of the mem-branes filled by epoxy resin. This overview demonstrates that allthe nanocomposites contained intercalated clay mineral layers andalso aggregated clay mineral layers, since the thickness of a sin-


about 100 nm [41]. Exfoliated particles (Fig. 2a and b) could also beseen in the samples with low clay contents, and a complex inter-calated/exfoliated morphology was formed. Therefore, a lack of

erent clay contents and PSf molar masses.

alcinatedared with PSf

d0 0 1 of the calcinatedmembranes preparedwith PSf P-1700 (A)

d0 0 1 of the calcinatedmembranes preparedwith PSf P-3500 (A)

14.0 14.114.0 13.813.8 14.514.0 14.413.8 14.114.0 13.9


ING Model

A

4 rface

rma

rtctTts

SpabPbAts

F2



egular increase of the intensity of the diffraction peak related to theontmorillonite can be ascribed both to the presence of tactoids,

s well as to exfoliated platelets.The visualization of these aggregated clay mineral layers can be

elated to several factors, such as: a not complete intercalation andhus a significant number of some-lamellae tactoids; the ultrami-rotoming direction may not be perfectly normal to the surface ofhe platelets and then clay mineral layers appeared tilted in theEM image or, still, there is the possibility of imperfect focusing ofhese unstable samples in TEM [42]. Furthermore, there were noignificant differences between the two types of PSf.

The morphology of the hybrid material surfaces was observed byEM (Fig. 3). By increasing the clay content in the membranes pre-ared with both kinds of PSf, the surfaces tended to be denser with

lower number of pores, which also had lower pore diameter. Thisehavior was less pronounced in the membranes prepared withSf P-3500, which present a slight difference in both parametersetween the unloaded and the 2.0 mass% MMT loaded membranes.


good dispersion of the clay particles, which are represented byhe small light lines, into the PSf matrix could also be verifiedince they were spread randomly and homogeneously on the whole

ig. 3. Some examples of SEM images of the surfaces of the membranes prepared from.0 mass% and (c) 4.0 mass% MMT.


surface. The formation of more dense surfaces is believed to be dueto the adsorption of polymer and solvent onto the clay mineral layersurface, hence there was a lower number of solvent molecules thatsolvate the formed structure. Consequently, when in contact withwater, the structure was precipitated faster, forming a dense layer.

Regarding cross-section SEM images (Fig. 4), the addition ofclay did not promote changes to the cross-section morphology,which was maintained as a cellular morphology. The only differ-ence is the presence of imprisoned clay particles within membranepores and the presence of bulbous macrovoids on the SEM imagesof the membranes prepared from dispersions containing at least2.0 mass% MMT. The presence of imprisoned clay particles can beexplained by the formation mechanism of the membrane using thewet-phase inversion method. For solutions with 10 mass% or higherpolymer concentrations, phase separation is known to occur bynucleation and growth of the poor polymer phase, which is the for-mation mechanism of the membranes presented herein [43–45]. Asthe growth of the nucleated phases was performed, clay particles


were excluded from the rich polymer phase and were imprisonedin the poor polymer phase, which originated the pores. Further-more, with the increase of clay content in the dispersion, dispersion

the dispersions containing 25 mass% PSf P-1700 or P-3500 and: (a) 0.0 mass%, (b)





F epared(

vss[

nau1s[otis3v

spfFd

ig. 4. Some examples of the SEM images of the cross-sections of the membranes prb) 2.0 mass% and (c) 4.0 mass% MMT.

iscosity was increased and formed a barrier preventing non-olvent molecules from entering. Hence, diffusion of the non-olvent was reduced and there was delayed local phase separation46].

Fig. 5 presents FTIR spectra of the MMT, pure PSf andanocomposite membranes. Characteristic bands of polysulfonere: C H stretching of the methyl groups in the isopropylidenenit at 2969 cm−1, benzene ring skeletal stretching mode in the600–1475 cm−1, sulfone bands at 1325 and 1154 cm−1 and anti-ymmetric C O stretching frequencies at 1257 and 1014 cm−1

47]. All these bands had high intensities in the FTIR spectrumf the membrane prepared with PSf P-1700 and from this spec-rum, the other membranes were studied. In relation to the MMT,ts spectrum presented the following bands: antisymmetric Si Otretching at 1043 and 481 cm−1, O H stretching at 3631 and420 cm−1, Al OH deformation vibration at 916 cm−1 and Si O Alibration at 797 and 524 cm−1 [48–51].

In all nanocomposite membranes, the band related to the C Htretching of the methyl groups in the isopropylidene unit disap-


eared, suggesting that this part of the PSf molecule had restrictedreedom of movements in the nanocomposite structures. In theTIR spectrum of the membrane prepared from 0.5 mass% MMTispersion, the other PSf bands had lower intensities and the MMT

from the dispersions containing 25 mass% PSf P-1700 or P-3500 and: (a) 0.0 mass%,

bands at 3631, 918, 796 and 465 cm−1 had low intensities, mostprobably due to the low MMT content and not because of theinteractions between the PSf chains and clay mineral platelets.The membrane prepared from 1.0 mass% MMT showed MMT bandswith higher intensities, including the band in the 1030 cm−1 region,on the other hand, the PSf bands had lower intensities. The mem-branes prepared from the 2.0, 3.0 and 4.0 mass% MMT dispersionspresented the same behavior, which was proportional to the claycontent. Lastly, the membrane prepared from the highest MMT con-tent dispersion also showed MMT bands with increased intensities;moreover, the referring spectrum did not have any PSf band, pos-sibly due to the high MMT content which restricted the PSf chainvibration.

Comparing the PSf P-1700 and P-3500 membranes, it is pos-sible to observe that the intensity of the C H stretching of themethyl groups in the isopropylidene unit band is lower in the PSf P-3500 spectrum, while the other PSf bands had pronounced higherintensities. Curiously, this is the same band that had its intensityreduced in the presence of clay, demonstrating that this group


has a fundamental role in the intermolecular interactions. SincePSf P-3500 has a higher molar mass, PSf chains are larger and,therefore, hinder this group stretching as observed when clay isadded.


Please cite this article in press as: P. Anadão, et al., Influence of the clay content and the polysulfone molar mass on nanocomposite membraneproperties, Appl. Surf. Sci. (2013), http://dx.doi.org/10.1016/j.apsusc.2013.01.102



6 P. Anadão et al. / Applied Surface Science xxx (2013) xxx– xxx

Fig. 5. FTIR spectra of the pure MMT and the PSf P-1700 and P-3500 membranes prepared from the dispersions containing: (a) 0.0 mass%, (b) 0.5 mass%, (c) 1.0 mass%, (d)2.0 mass%, (e) 3.0 mass%, (f) 4.0 mass% and (g) 5.0 mass% MMT.





Fig. 6. TGA curves of the (a) pure MMT and the PSf P-1700 membranes prepared from the dispersions containing: (b) 0.0 mass%, (c) 0.5 mass%, (d) 1.0 mass%, (e) 2.0 mass%,(

brpoipiFMt3tspba

f) 3.0 mass%, (g) 4.0 mass% and (h) 5.0 mass% MMT.

In the FTIR spectra of the PSf P-3500 nanocomposite mem-ranes, the C H stretching band was not present. Moreover, theest of the spectra of the nanocomposite membranes did notresent a regular behavior in relation to the decrease and increasef the intensities of MMT bands and of PSf bands, respectively. Tollustrate this lack of regularity, the FTIR spectrum of the membranerepared from the 0.5 mass% MMT dispersion can be mentioned,

n which the intensities of the PSf bands were very low and in theTIR spectra of the membranes prepared with 1.0, 2.0 and 5.0 mass%MT, the intensities of the PSf bands were very high in relation to

he intensities of the MMT bands. The membrane prepared with.0 mass% MMT showed lower intensities of the PSf bands whilehe intensities of the MMT bands were increased and in the FTIRpectrum of the membrane prepared from 4.0 mass% MMT dis-


ersion, this behavior was more pronounced. Hence, in order toetter understand this behavior, NMR and Raman spectroscopiesre intended to be performed in the future.

3.2. Thermal properties of the nanocomposite membranes

TGA curves (Figs. 6 and 7) of the PSf and nanocomposite mem-branes presented a two-stage mass loss event, which began at450 ◦C and ended around 650 ◦C, ascribed to the PSf decomposition[52]. For both polysulfones, the thermal stability of nanocompositemembranes was higher than that of PSf membrane.

In the first stage of degradation, a moderate improvement inthermal stability was found in the nanocomposites. Table 2 showeda 2–55 ◦C increase in the initial degradation temperature of PSfP-1700 nanocomposites and a 13–53 ◦C increase in the initial degra-dation temperature of PSf P-3500 nanocomposite with the increaseof MMT content. This may be due to kinetic effects, with a highercontent of platelets retarding the diffusion of oxygen into the poly-


mer matrix. Insignificant differences were observed for the onsettemperature for the second stage as well as for the final degradationtemperature.


Please cite this article in press as: P. Anadão, et al., Influence of the clay content and the polysulfone molar mass on nanocomposite membraneproperties, Appl. Surf. Sci. (2013), http://dx.doi.org/10.1016/j.apsusc.2013.01.102



8 P. Anadão et al. / Applied Surface Science xxx (2013) xxx– xxx

Fig. 7. TGA curves of the PSf P-3500 membranes prepared from the dispersions containing: (a) 0.0 mass%, (b) 0.5 mass%, (c) 1.0 mass%, (d) 2.0 mass%, (e) 3.0 mass%, (f)4.0 mass% and (g) 5.0 mass% MMT.

Table 2Onset and final temperatures of decomposition, mass losses and enthalpies (�H1 and �H2) of the decomposition event of the PSf P-1700 and P-3500 nanocompositemembranes.

Clay content (%) Onset temperature ofdecomposition (◦C)

Final temperature ofdecomposition (◦C)

Mass loss (%) �H1 (J/g) �H2 (J/g)

P-1700 P-3500 P-1700 P-3500 P-1700 P-3500 P-1700 P-3500 P-1700 P-3500

0.0 450 450 657 658 97.4 93.8 −12,157 −8881 −10,452 −88960.5 452 463 652 655 92.7 92.8 −7151 −9457 −8842 −91831.0 470 475 680 662 92.2 88.8 −7942 −5926 −7431 −83102.0 469 480 660 670 81.0 83.8 −7507 −9372 −8903 −10,4583.0 479 487 659 653 86.0 79.9 −8428 −8254 −10,058 −10,7964.0 490 490 688 654 89.3 78.1 −9203 −7527 −9783 −98975.0 505 503 685 665 86.3 76.3 −8428 −9445 −10,058 −9393





Ff

smhmbpedcptbse

tbia

wtaPspHsiilmawpe

3

tbih4

ical properties were found to be higher in the PSf P-3500membranes, since its polymer chain length is higher which resultsin a higher number of intermolecular bonds, thus increasing bothproperties.

ig. 8. Typical DTA curve for a nanocomposite membrane, in this case, preparedrom a dispersion containing 25% PSf P-3500/2.0 mass% MMT.

In relation to the mass losses, the PSf P-3500 membranes pre-ented a lower mass loss than the PSf P-1700 membranes, whicheant that the interactions between the polymer chains with

igher extension were bigger than that of between the lower molarass polymer chains. The mass loss was also found to be lower

y increasing clay content in both PSf membranes due to therotection of the polymer chains by a higher number of clay min-ral platelets towards decomposition. Besides that, a pronouncedecrease of the mass loss was observed when this parameter isompared between the pure PSf membranes and the membranesrepared from the 1.0 and 2.0 mass% MMT dispersions, showinghat not only the amount of clay was responsible for the mass loss,ut also the nanocomposite morphology, as these membranes pre-ented an exfoliated structure which hindered the mass transferfficiently.

From the DTA curves (Fig. 8), it is possible to observe that thewo-stage mass loss event shown in the TGA curves is constitutedy two exothermic processes. The relation between the clay content

n the membrane and the enthalpies of these exothermic processesre also presented in Table 2.

In the case of the PSf lower molar mass membranes, enthalpyas increased in the membranes with lower clay contents and

hen, decreased again in the membranes with higher clay contents,lthough their values were still lower than the enthalpy of the pureSf membrane. This phenomenon could be explained as follows:ince clay mineral layers protect polymer chains, the exothermalrocess is smoother and therefore, the released energy is lower.owever, in the membranes with higher clay contents, decompo-

ition takes longer to occur, having a located heat concentrationn the clay mineral layers which favor the occurrence of a morentense burn of the PSf chains in relation to the membranes withower clay contents in their composition. Even so, a portion of poly-

er chains is protected by the clay platelets; thus, their enthalpiesre higher than that of the pure PSf membrane. Irregular behavioras found in the PSf P-3500 membranes, possibly by the longerolymer chains and, therefore, the heat concentration was not soffective to produce significant differences on enthalpy values.

.3. Mechanical properties of the nanocomposite membranes

The elongation at break, a measure of ductility, is related withhe clay content in Fig. 9. The presence of clay is usually known


y the nanocomposites becoming more brittle, causing a decreasen elongation at break. Nevertheless, significant improvementsave been seen in this mechanical property up to clay loadings of.0 mass% for both types of polysulfone, which corresponded to a

Fig. 9. Elongation at break for PSf nanocomposite membranes.

maximum increase of 56% in the PSf P-1700 nanocomposite mem-brane and of 106% in the PSf P-3500 nanocomposite membrane. Areason for this could be the rearrangement of clay platelets in thedirection of the deformation, allowing greater deformations.

The tensile strength values were determined as the maximumstress values. Fig. 10 presents the tensile strength as a function ofthe MMT content. Enhancements were also observed in the mem-branes prepared from dispersions containing up to 4.0 mass% MMT,corresponding to a maximum increase of 34% in the PSf P-1700nanocomposite membrane, and of 23% in the PSf P-3500 nanocom-posite membrane. Therefore, the enhancement of both mechanicalproperties is due to the insertion of the polymer chains betweenthe silicate layers which leaded to an increase in the surface areaof interaction between clay and polymer matrix.

Both PSf membranes prepared from the 5.0 mass% MMT dis-persions presented a decrease in these two mechanical properties.Several possible reasons can be responsible for this decrease. First ofall, higher clay content can cause material embrittlement. Besidesthis effect, the existence of flaws such as the weak boundariesbetween particles, which may increase with the volume frac-tion of the filler, can also decrease the final mechanical features.Another possible reason is the heterogeneity of the samples sincethe dispersions used to prepare these membranes had the highestviscosities which could cause inhomogeneity and a higher numberof macrovoids.

Moreover, comparing both types of polysulfone, these mechan-


Fig. 10. Tensile strength for PSf nanocomposite membranes.


ING Model

A

1 rface

sctpiwmf

4

twmttpabctccmiwtcabp

masfrhb

A

st

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[


0 P. Anadão et al. / Applied Su

The error bars related to both mechanical properties are con-iderable and reasonable different from each other whose behavioran be explained by some reasons. The first one is related to the facthat the maximum stress is very close to the stress at the breakingoint. Secondly, the material did not break because it had reached

ts plastic limit but rather because it probably contained a defect,hich is very common since these samples were porous and hadacrovoids in their substructure, which could cause mechanical

ailure when subjected to deformation.

. Conclusions

By a combination of the wet-phase inversion method and solu-ion dispersion technique, PSf/MMT nanocomposite membranesere prepared and the influence of MMT content as well as PSfolar mass was evaluated. The increase of the basal spacing with

he increase of clay content, detected by SAXRD, could be caused byhe intercalation of PSf chains. TEM photoimages of all nanocom-osite membranes showed intercalated clay mineral layers andggregated clay mineral layers while exfoliated portions coulde seen in TEM photoimages of the membranes with lower clayontents. SEM images of the surfaces depicted the decrease inhe number of pores and pore diameter with the increase of clayontent for both types of PSf and the cross-section SEM photomi-rographs showed that the cellular morphology of the pure PSfembranes was not altered by clay addition. Preliminary FTIR stud-

es presented the interactions between polymer and clay mineralhose results will be further investigated by Raman and NMR spec-

roscopies. Thermal resistance was improved by using a higher clayontent and polysulfone molar mass, since mass loss was decreasednd the onset temperature of decomposition was increased. Finally,y increasing clay content up to 4.0 mass% and by using higherolysulfone molar mass, mechanical resistance was enhanced.

In the next study, the influence of clay content and polysulfoneolar mass will be related to the membrane morphology as well

s to its filtration potential. Ternary phase diagrams will be con-tructed, kinetics of the quaternary system will be measured byollowing the distance of precipitation front over time and will beelated to the rheology of the dispersions. Furthermore, membraneydrophilicity will be quantified. Hence, filtration performance cane completely understood.

cknowledgements

Priscila Anadão would like to acknowledge Fapesp for her PhDcholarship and Solvay is also thanked by the authors for providinghe polysulfone Udel® P-1700 and Udel® P-3500.

eferences

[1] C.E. Corcione, A. Cataldi, M. Frigione, Measurements of size distributionnanoparticles in ultraviolet-curable methacrylate-based boehmite nanocom-posites, Journal of Applied Polymer Science, in press.

[2] C.E. Corcione, F. Freuli, A. Maffezzoli, The aspect ratio of epoxy matrix nanocom-posites reinforced with graphene stacks, Polymer Engineering & Science, inpress.

[3] C.E. Corcione, M. Frigione, UV-cured polymer-boehmite nanocomposite as pro-tective coating for wood elements, Progress in Organic Coatings 74 (2012)781–787.

[4] A. Greco, A. Maffezzoli, G. Buccoliero, F. Caretto, G. Cornacchia, Thermal andchemical treatments of recycled carbon fibers for improved adhesion to poly-meric matrix, Journal of Composite Materials 0 (2012) 1–9.

[5] C.E. Corcione, A. Maffezzoli, Glass transition in thermosetting clay-


nanocomposite polyurethanes, Thermochimica Acta 485 (2009) 43–48.[6] D.R. Paul, L.M. Robeson, Polymer nanotechnology: nanocomposites, Polymer

49 (2008) 3187–3204.[7] V. Mittal, Polymer layered silicate nanocomposites: a review, Materials 2 (2009)

992–1057.

[


[8] C.E. Corcione, G. Mensitieri, A. Maffezzoli, Analysis of the structure and masstransport properties of nanocomposite polyurethane, Polymer Engineering andScience 49 (2009) 1708–1718.

[9] A. Greco, M. Rizzo, A. Maffezzoli, Effect of the addition of organically modifiednanofiller on the relaxation behavior of a thermoplastic amorphous matrix,Thermochimica Acta 534 (2012) 226–231.

10] A. Greco, A. Maffezzoli, E. Calò, C. Massaro, R. Terzi, An investigation into sinter-ing of PA6 nanocomposite powders for rotational molding, Journal of ThermalAnalysis and Calorimetry 109 (2012) 1493–1502.

11] F. Santiago, A.E. Mucientes, M. Osório, C. Rivera, Preparation of compositesand nanocomposites based on bentonite and poly(sodium acrylate). Effect ofamount of bentonite on the swelling behavior, European Polymer Journal 43(2007) 1–9.

12] H.Y. Huang, T.C. Huang, T.C. Yeh, C.Y. Tsai, C.Y. Lai, M.H. Tsai, J.M. Yeh, Y.C.Chou, Advanced anticorrosive materials prepared from amine-capped anilinetrimer-based electroactive polyimide–clay nanocomposite materials with syn-ergistic effects of redox catalytic capability and gas barrier properties, Polymer52 (2011) 2391–2400.

13] Q.Y. Soundtrarajah, B.S.B. Karunaratne, M.G. Rajapakse, Mechanical propertiesof poly(vinyl alcohol) montmorillonite nanocomposites, Journal of CompositeMaterials 44 (2010) 303–311.

14] M. Bhattacharya, S. Biswas, A.K. Bhowmick, Permeation characteristics andmodeling of barrier properties of multifunctional rubber nanocomposites, Poly-mer 52 (2011) 1562–1576.

15] S.S. Ray, M. Okamoto, Polymer/layered silicate nanocomposites: a review frompreparation to processing, Progress in Polymer Science 28 (2003) 1539–1641.

16] P. Anadão, Ciência e Tecnologia de Membranas, Artliber Editora, São Paulo,2010.

17] W.C. Chao, S.H. Huang, Q. Na, D.J. Liaw, Y.Y.C. Huang, K.R. Lee, J.Y. Lai, Novelinterfacially-polymerized polyamide thin-film composite membranes: stud-ies on characterization, pervaporation, and positron annihilation spectroscopy,Polymer 52 (2011) 2414–2421.

18] A.I.I. Sodeye, T. Huang, S.P. Gido, J.W. Mays, Polymer electrolyte membranesfrom fluorinated polyisoprene-block-sulfonated polystyrene: membranestructure and transport properties, Polymer 52 (2011) 1963–1970.

19] J.T. Arena, B. McCloskey, B.D. Freeman, J.R. McCutcheon, Surface modificationof thin film composite membrane support layers with polydopamine: enablinguse of reverse osmosis membranes in pressure retarded osmosis, Journal ofMembrane Science 375 (2011) 55–62.

20] E. Gérard, E. Bessy, C. Salvagnini, V. Rerat, M. Momtaz, G. Hénard, P. Marmey, T.Verpoort, J. Marchand-Brynaert, Surface modifications of polypropylene mem-branes used for blood filtration, Polymer 52 (2011) 1223–1233.

21] T. Sakaguchi, S. Tominaga, T. Hashimoto, Synthesis and gas permeability of estersubstituted poly(p-phenylene)s, Polymer 52 (2011) 2163–2169.

22] S. Loeb, S. Sourirajan, High-flow semipermeable membranes for separation ofwater from saline solutions, Advances in Chemistry Series 38 (1962) 117–132.

23] S. Strathmann, Development of new membranes, Desalination 35 (1980)39–58.

24] S.C. Pesek, W.J. Koros, Aqueous quenched asymmetric polysulfone membranesprepared by dry/wet phase separation, Journal of Membrane Science 81 (1993)71–88.

25] L. Yilmaz, A.J. McHuch, Modelling of asymmetric membrane formation. I.Critique of evaporation models and development of a diffusion equation for-malism for the quench period, Journal of Membrane Science 28 (1986) 287–310.

26] M.J. Han, D. Bhattacharya, Changes in morphology and transport characteristicsof polysulfone membranes prepared by different demixing conditions, Journalof Membrane Science 98 (1995) 191–200.

27] P. van de Witte, P.J. Dijkstra, J.W.A. Berg, J. Feijen, Phase separation processesin polymer solutions in relation to membrane formation, Journal of MembraneScience 117 (1996) 1–31.

28] G. Arthanareeswaran, T.K.S. Devi, M. Raajenthiren, Effect of silica particles oncellulose acetate blend ultrafiltration membranes: part I, Separation and Purifi-cation Technology 64 (2008) 38–47.

29] G. Arthanareeswaran, T.K.S. Devi, D. Mohan, Development, characterizationand separation performance of organic-inorganic membranes Part II. Effect ofadditives, Separation and Purification Technology 67 (2009) 271–281.

30] L.Y. Yu, Z.L. Xu, H.M. Shen, H. Yang, Preparation and characterization of PVDF-SiO2 composite hollow fiber UF membrane by sol–gel method, Journal ofMembrane Science 337 (2009) 257–265.

31] P. Aerts, S. Kuypers, I. Genne, R. Leysen, J. Mewis, I.F.J. Vankelecom, P.A. Jacobs,Polysulfone–ZrO2 surface interactions. The influence on formation, morphol-ogy and properties of zirfon-membranes, Journal of Physical Chemistry B 110(2006) 7425–7430.

32] F. Ferrando, C. Torras, R. Garcia-Valls, J. Paltakari, Performance, morphologyand tensile characterization of activated carbon composite membranes for thesynthesis of enzyme membrane reactors, Journal of Membrane Science 282(2006) 149–161.

33] J.H. Choi, J. Jegal, W.N. Kim, Fabrication and characterization of multi-walledcarbon nanotubes/polymer blend membranes, Journal of Membrane Science284 (2006) 406–415.

34] S.A. Hashemifard, A.F. Ismail, T. Matsuura, Effects of montmorillonite nano-clay


fillers on PEI mixed matrix membrane for CO2 removal, Chemical EngineeringJournal 170 (2011) 316–325.

35] V.R. Hande, S.K. Rath, S. Rao, M. Patri, Cross-linked sulfonated poly(ether etherketone) (SPEEK)/reactive organoclay nanocomposite proton exchange mem-branes (PEM), Journal of Membrane Science 372 (2001) 40–48.


ING Model

A

rface

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[



36] S. Loske, M.C. Gonc alves, B.A. Wolf, Fractionation of cellulose acetate for theinvestigation of molecular weight influences on the morphology of membranes,Journal of Membrane Science 214 (2003) 223–228.

37] L.P. Cheng, Y.S. Huang, T.H. Young, Effect of the temperature of polyurethanedissolution on the mechanism of wet-casting membrane formation, EuropeanPolymer Journal 39 (2003) 601–607.

38] P. Anadão, L.F. Sato, H. Wiebeck, F.R.V. Díaz, Montmorillonite as a componentof polysulfone nanocomposite membranes, Applied Clay Science 48 (2010)127–132.

39] D. Homminga, B. Goderis, S. Hoffman, H. Reynaers, G. Groeninckx, Influenceof shear flow on the preparation of polymer layered silicate nanocomposites,Polymer 46 (2005) 11624–11634.

40] G. Lagaly, M.F. Gonzalez, A. Weiss, Problems in layer-charge determination ofmontmorillonites, Clay Minerals 11 (1976) 173–187.

41] J.W. Cho, D.R. Paul, Nylon 6 nanocomposites by melt compounding, Polymer42 (2001) 1083–1094.

42] L.F. Valadares, C.A.P. Leite, F. Galembeck, Preparation of naturalrubber–montmorillonite nanocomposite in aqueous medium: evidencefor polymer–platelet adhesion, Polymer 47 (2006) 672–678.


43] Y. Yip, A.J. McHugh, Modeling and simulation of nonsolvent vapor-inducedphase separation, Journal of Membrane Science 271 (2006) 163–176.

44] H.J. Lee, B. Jung, Y.S. Kang, H. Lee, Phase separation of polymer castingsolution by nonsolvent vapor, Journal of Membrane Science 245 (2004)103–112.

[

PRESSScience xxx (2013) xxx– xxx 11

45] D.T. Lin, L.P. Cheng, Y.J. Kang, L.W. Chen, T.H. Young, Effects of precipitation con-ditions and the membrane morphology and permeation characteristics, Journalof Membrane Science 140 (1998) 185–194.

46] I.M. Wienk, R.M. Boom, M.A.M. Beerlage, A.M.W. Bulte, C.A. Smolders, H. Strath-mann, Recent advances in the formation of phase inversion membranes madefrom amorphous or semi-crystalline polymers, Journal of Membrane Science113 (1996) 361–371.

47] M. Mohr, D.R. Paul, I. Pinnau, W.J. Koros, Surface fluorination of polysulfoneasymmetric membranes and films, Journal of Membrane Science 56 (1991)77–98.

48] J. Hrachová, P. Komadel, V.S. Fajnor, The effect of mechanical treatment on thestructure of montmorillonite, Materials Letters 61 (2007) 3361–3365.

49] H. Lu, Y. Hu, M. Li, Z. Chen, W. Fan, Structure characteristics and thermal prop-erties of silane-grafted-polyethylene/clay nanocomposite prepared by reactiveextrusion, Composites Science and Technology 66 (2006) 3035–3039.

50] A. Rehab, N. Salahuddin, Nanocomposite materials based on polyurethaneintercalated into montmorillonite clay, Materials Science and Engineering: A399 (2005) 368–376.

51] T. Sugama, Polyphenylenesulfied/montomorillonite clay nanocomposite


coatings: their efficacy in protecting steel against corrosion, Materials Letters60 (2006) 2700–2706.

52] G. Lisa, E. Avram, G. Paduararu, M. Irimia, N. Hurduc, N. Aelenei, Thermalbehavior of polystyrene, polysulfone and their substituted derivatives, PolymerDegradation and Stability 82 (2003) 73–79.


influence of the clay content and the polysulfone molar mass on nanocomposite membrane properties

Documents