Morphology and properties of polymer/organoclay nanocomposites based on poly(ethylene terephthalate) and sulfopolyester blends

43

9

Research ArticleReceived: 25 November 2011 Revised: 21 March 2012 Accepted: 24 March 2012 Published online in Wiley Online Library: 21 November 2012

(wileyonlinelibrary.com) DOI 10.1002/pi.4331

Morphology and properties ofpolymer/organoclay nanocomposites based onpoly(ethylene terephthalate) andsulfopolyester blendsAbbas Ghanbari,a Marie-Claude Heuzey,a∗ Pierre J Carreaua

and Minh-Tan Ton-Thatb

Abstract

Poly(ethylene terephthalate) (PET) nanocomposite films containing two different organoclays, Cloisite 30B (C30B) andNanomer I.28E (N28E), were prepared by melt blending. In order to increase the gallery spacing of the clay particles, asulfopolyester (PET ionomer or PETi) was added to the nanocomposites via a master-batch approach. The morphological,thermal and gas barrier characteristics of the nanocomposite films were studied using several characterization techniquessuch as scanning electron microscopy, transmission electron microscopy, X-ray diffraction, differential scanning calorimetry,dynamic mechanical analysis, rheometry and oxygen permeability. PET and PETi were found to form immiscible polymer blendsand the nanoparticles were preferentially located in the PETi dispersed phase. A better dispersion of clay was obtained fornanocomposites containing N28E with PETi. On the contrary, for nanocomposites containing C30B and PETi, the number oftactoids increased and the clay distribution and dispersion became worse than for C30B alone. Overall, the best propertieswere obtained for the PET/C30B nanocomposite without PETi. Higher crystallinity was found for all nanocomposite films incomparison to that of neat PET.c© 2012 Society of Chemical Industry

Keywords: poly(ethylene terephthalate); nanocomposite; organoclay; rheology; ionomer

INTRODUCTIONPoly(ethylene terephthalate) (PET) is one of the most widelyused engineering polymers due to its good processability, hightransparency and mechanical and gas barrier properties, as wellas its chemical resistance and low cost. This semi-crystallinethermoplastic has found wide applications in the form of fibersand non-fibers (such as food and beverage packaging, automotive,electrical devices and construction). For some applications itis desirable to improve specific characteristics such as barrierproperties for beverage and food packaging. Incorporation ofsilicate nanolayers into a PET matrix has been demonstrated toenhance various physical properties.1,2

Much effort has been devoted to achieve well-dispersedand delaminated silicate nanolayers in PET matrices. In moststudies, the coexistence of a PET microcomposite along withan intercalated structure has been observed.3 – 7 The processingtemperature of PET is around 260 ◦C while commercial organoclaysstart degrading around 200 ◦C. The decomposition of the organicmodifier leads to a collapse of the silicate nanolayers and to adecrease of the interlayer spacing, thus impeding the intercalationof polymer chains within the gallery spacing. Moreover, the lack ofa delaminated or exfoliated morphology in PET nanocomposites isoften due to the lack of compatibility of the organic modifiers usedin commercial organoclays with the PET macromolecules. For thesereasons, several studies have been performed to modify silicatenanolayers with various cationic surfactants,6,8 – 10 mostly aimed

at improving the thermal stability of commercial organoclays.Efforts have also been devoted to improve the compatibilitybetween the organoclay and polyesters. Chisholm and co-workers11 synthesized sulfonated poly(butylene terephthalate)copolymers (PBT ionomers) by the melt polymerization ofdimethyl terephthalate, dimethyl-5-sodiosulfoisophthalate and1,4-butanediol. They found that the incorporation of 5 mol%of–SO3Na groups into PBT had no significant effect on themorphology of composites containing pristine clay (i.e. Na-MMT). However, the presence of sulfonated groups in PBTionomer/organoclay nanocomposites led to a dramatic change inthe morphology, in comparison to the reference PBT/organoclaynanocomposites. It was suggested that the better intercalated anddelaminated morphology obtained in the presence of sulfonatedgroups could be attributed to the binding of these negatively

∗ Correspondence to: Marie-Claude Heuzey, Center for Applied Research onPolymers and Composites (CREPEC), Chemical Engineering Department, EcolePolytechnique de Montreal, PO Box 6079, Stn Centre-Ville, Montreal, Quebec,Canada H3C 3A7. E-mail: [email protected]

a Center for Applied Research on Polymers and Composites (CREPEC), ChemicalEngineering Department, Ecole Polytechnique de Montreal, PO Box 6079, StnCentre-Ville, Montreal, Quebec, Canada H3C 3A7

b Industrial Materials Institute, National Research Council Canada, 75 MortagneBoulevard, Boucherville, Quebec, Canada J4B 6Y4

Polym Int 2013; 62: 439–448 www.soci.org c© 2012 Society of Chemical Industry

44

0

www.soci.org A Ghanbari et al.

charged groups to the edges of the clay platelets that bear positivecharges. This attachment could facilitate the diffusion of mobilepolymer chain segments into the gallery spacing of the organoclay.Improvements on the processing and formulation sides havealso been considered. Ghasemi et al.12 prepared nanocompositescontaining 3 wt% Cloisite 30B (C30B) using a PET experimentalgrade of high viscosity (Selar PTX 295, DuPont) blended witha commercial-grade PET (9921, Eastman Chemical Company) ata ratio of 1 : 4. They studied the effect of feeding rate and screwspeed using two screw geometries with different mixing elements.For the best case (higher screw speed), 27% reduction in oxygenpermeability was achieved. In another study, Xu and co-workers13

prepared nanocomposites containing 2 wt% Nanomer I.28E

(N28E) and 6 wt% PET ionomer (PETi) using Selar PTX 295. Thebest improvement of the barrier properties achieved via theirprocedure was 19%.

In the work reported here, a commercial sulfopolyester (PETi)was used to improve the compatibility between a PET matrixand two different organoclays. Formulation and compoundingwere entirely based on a melt processing approach, as in ourprevious study.13 In the current work a thorough investigationof the state of miscibility between PET and PETi was carriedout. Furthermore we show that the dispersion and delaminationof the clays are dependent on the organomodifier chemistry.The interactions between the clays and the ionomer werestudied through rheological measurements in order to explainthe resulting morphologies.

EXPERIMENTALMaterialsA commercial-grade PET (PET 9921), with intrinsic viscosity of0.8 dL g−1, was obtained from Eastman Chemical Company. ThePETi used for the study, AQ55S, with an ionic content of 9 mol%and inherent viscosity of ca 0.3 dL g−1, was also kindly suppliedby Eastman Chemical Company. The chemical formula of PETiis shown in Fig. 1. The organically modified montmorillonites,C30B and N28E, were obtained from Southern Clay Products Inc.and Nanocor Inc., respectively. C30B is modified with methyl,tallow, bis-2-hydroxyethyl, quaternary ammonium, while N28E ismodified with octadecylammonium according to R&D of NanocorInc. It has also been proposed that the Nanomer and Cloisitemontmorillonites do not have the same origin, since they containdifferent amounts of Fe3+.14

Sample preparationThe PET, PETi and organoclays were vacuum dried at 80 ◦C for24 h before processing. For all blends the organoclay nominalcontent was 2 wt% of the total mass, while the amount of PETi was6 wt% in the nanocomposites containing the compatibilizer. Thenanocomposites containing PETi were prepared by a master-batchapproach: 25 wt% PETi, 8.3 wt% clay and 66.7 wt% PET were melt-blended using a small co-rotating twin-screw Leistritz extruder

Figure 2. Screw geometry of the Leistritz 34 mm twin-screw extruder.

Table 1. Nomenclature and compositions of the nanocompositefilms

Nomenclature PET (wt%) PETi (wt%) C30B (wt%) N28E (wt%)

Neat PET 100 0 0 0

PET/PETi 94 6 0 0

PET/C30B 98 0 2 0

PET/PETi/C30B 92 6 2 0

PET/N28E 98 0 0 2

PET/PETi/N28E 92 6 0 2

(screw diameter = 18 mm and L/D = 40). Then the preparedmaster-batch was diluted with the neat PET using a larger Leistritzextruder (screw diameter = 34 mm and L/D = 42). Figure 2 showsthe larger extruder screw geometry. The first and third mixingzones contained 10 and 15 kneading elements, respectively, with7.5 mm width kneading lobes with right-hand (positive) and left-hand (negative) 60◦ staggering angles. The second mixing zonewas composed of two left- and right-handed gear-type mixingelements. There were eight teeth around each circumference andfive gears in each block, while the total length of the mixing blockwas 30 mm. The screw speed and feed rate were kept at 200 rpmand 3 kg h−1, respectively. The extrusion temperature profile fromthe feed to the die was set between 240 and 265 ◦C. Table 1 givesthe nanocomposite compositions and nomenclature.

To prepare the PET-based films, a 20 cm wide slit die with a1.42 mm die gap was used. An air knife was mounted on bothsides of the die. To stretch the extrudate, chill rolls (20 ◦C) wereemployed with the distance between the die and the chill rollsbeing approximately 10 cm. The width of the films was 16 cm andthe average neck-in due to stretching was 20%. The draw ratiowas around 44 and the corresponding thickness of the films was40 µm.

Finally, to investigate the miscibility between PET and PETi,several PET/PETi blends containing 0, 10, 20, 30, 50, 70 and 100wt% PETi were prepared using a 30 mL Plasti-Corder internal

Figure 1. Repeat unit of PETi. The ionic comonomer is randomly incorporated within the structure.15.

wileyonlinelibrary.com/journal/pi c© 2012 Society of Chemical Industry Polym Int 2013; 62: 439–448

44

1

Polymer/organoclay nanocomposites www.soci.org

mixer (Brabender). Melt compounding was performed at 260 ◦Cand 100 rpm for 10 min under dry nitrogen. Also, PET/PETi blendscontaining 6 and 20 wt% PETi and 2 wt% C30B were preparedusing the same instrument and conditions (except at 250 ◦C) todetect preferential localization of clay particles.

Characterization of polymer blends and nanocompositesDynamic mechanical analysis (DMA) was carried out using a TAInstruments DMA 2980 from 40 to 120 ◦C at a heating rate of 5 ◦Cmin−1 and frequency of 1 Hz. Rectangular-shaped samples wereprepared via compression molding in a hot press at 265 ◦C for 9 minunder a nitrogen atmosphere, followed by quenching using a coldpress for 5 min. The curves of tan δ as a function of temperaturewere analyzed to obtain the glass transition temperature, Tg, ofthe PET/PETi blends.

XRD patterns were obtained using a Bruker D8 Discover withCu Kα radiation operating at an incident X-ray wavelengthλ = 0.15406 nm and a scan rate of 0.6◦ min−1. The spectrawere recorded over the 2θ range of 0.8–10◦.

The level of clay distribution and dispersion were determinedusing SEM and transmission electron microscopy (TEM) at themicro- and nano-level, respectively. SEM observations were carriedout using a cold field emission gun SEM instrument (Hitachi S4700)with an operation voltage of 2 kV. All the specimens were preparedby employing an Ultracut FC microtome (Leica) with a diamondknife followed by coating with platinum vapor. TEM images wereobtained using a JEOL JEM 2100F microscope operating at 200 kV.The samples were ultramicrotomed into ultrathin slices of about50–80 nm thickness at cryogenic temperature (−100 ◦C) usingthe aforementioned microtome system. To determine if PET andPETi were miscible, PET/PETi samples of various compositionswere etched in deionized hot water to selectively extract the PETidomains prior to SEM characterization.

The thermal behavior of the nanocomposites was analyzedusing DSC (TA Instruments Q1000). All the measurements wereperformed under a helium atmosphere. The samples were heatedfrom room temperature to 300 ◦C and held at that temperaturefor 3 min, then cooled to 30 ◦C and heated again to 300 ◦C at aconstant rate of 10 ◦C min−1.

Molten-state rheological measurements were performed usinga strain-controlled rotational rheometer Advanced RheometricExpansion System (ARES, TA Instruments) and a stress-controlledBohlin Gemini, both with a parallel plate flow geometry (25 mmdiameter, 1 mm gap). Dynamic linear frequency sweeps werecarried out for neat PET and PET-based nanocomposites at 265 ◦Cunder a nitrogen atmosphere. The time for the measurementswas restricted to 5 min to avoid a severe change of the molecularweight of the samples.

Finally, to determine oxygen transmission rates (OTRs), anOX-Tran model 2/21 apparatus (Mocon Inc.) with an oxygenpermeability MD module was employed. All measurements wereperformed at 23 ◦C under a pressure of 690 mmHg (92 kPa) of100% dry oxygen. The permeability coefficient (P, in L m−1 day−1

atm−1) was obtained from the OTR values using the followingformula:

P = OTR × L

p(1)

where L is the film thickness (m) and p is the testing pressure (atm).

RESULTS AND DISCUSSIONInvestigation of PET and PETi miscibilityThe compounding of two polymers may lead to miscible, partiallymiscible and immiscible blends.16 In the scientific literature,16 – 18

a single Tg for a polymer blend is generally recognized as a sign ofmiscibility (where the size of the domains is below 15 nm), while ablend is qualified as immiscible if it exhibits two or more Tg valuesat a given composition. To use the Tg approach, the content ofeach component in the blend should be more than 10 wt%, andtheir Tg values should differ by at least 10 ◦C. The melting pointdepression of a semi-crystalline polymer in the presence of anamorphous polymer is also used to examine the miscibility of ablend. Based on this method, a blend is considered miscible if themelting point Tm decreases with composition, while a constantTm is a characteristic of immiscible polymer blends.17,18 There are,however, several publications reporting that these two methodsmay lead to contradictory results regarding the miscibility state ofa blend.17,19 – 21 For example, for a blend exhibiting two Tg values,a melting point depression was observed,20 and a single Tg withno melting point depression has been reported as well.21 On thebasis of the above observations, it seems that a conclusion onthe miscibility of two polymers based on Tg and melting pointdepression may be uncertain. In this work, the miscibility betweenPET and PETi was investigated by the former two methods, as wellas via a selective solvent dissolution technique followed by SEM.

DMA data for neat PET, neat PETi and their blends (70/30,50/50 and 30/70 wt%) are reported in Fig. 3 in terms of theloss tangent (tan δ). Tg of neat PET and neat PETi are 88.1 and73.0 ◦C, respectively. A single Tg is observed for all blends, and thisvalue is composition dependent, suggesting miscibility of the twopolymers. DSC data are shown in Fig. 4, illustrating the meltingbehavior of PET in the presence of the amorphous PETi. Themelting point drops monotonically with increasing PETi content:however, at the largest PETi concentration (70 wt%), the meltingpeak disappears completely. It seems that a less stable crystallinephase, which melts at lower temperature, is formed in the presenceof PETi. The observation of a single Tg for all blend compositions(Fig. 3) along with the melting point depression shown in Fig. 4strongly suggest miscibility of the two polymers. However, afterfurther investigation by solvent extraction, the blends show phase-separated domains as illustrated in the SEM micrographs ofFig. 5 for PET/PETi blends containing various amounts of PETi.Separated droplet-like domains, of the order of several hundredsof nanometers to micrometers, are observed, indicating the non-miscibility of PET and PETi at this scale. In blends containing 50wt% PETi a co-continuous structure is generated during the phaseseparation process, while in blends containing 20 and 30 wt% PETi

Figure 3. Tan δ versus temperature for PET, PETi and several blendcompositions of PET and PETi.

Polym Int 2013; 62: 439–448 c© 2012 Society of Chemical Industry wileyonlinelibrary.com/journal/pi

44

2


Figure 4. Melting behavior of PET/PETi blends containing 0 to 70 wt% PETi.

phase separation leads to a droplet-matrix type of morphology.As expected, droplets are larger for blends containing 30 wt% PETiin comparison to those containing 20 wt% PETi. To explain theapparent contradiction between the DMA and SEM results, it isnecessary to have a closer look at Fig. 3. The broad peak associatedwith pure PETi in the DMA data suggests that it may be a two-phase system itself, with separation of the ionic copolymer (withthe sulfonated groups in Fig. 1) from the PET homopolymer inPETi. Hence, there may be the possibility that the phase-separatedportion of PETi interferes with the crystallization of the other PETin the blends, leading to the melting point depression observed inFig. 4.

Morphology of PET/PETi nanocompositesAs the organoclays have positive charges on the edges whilePETi exhibits negative charges at the sulfonated groups,3,11 andas higher affinity of silicate nanolayers to polar polymers hasbeen reported in the literature,22,23 it is anticipated that theorganoclays will have greater affinity with the PETi phase. Thusthe phase separation behavior of PET and PETi discussed inthe previous section may cause migration and partitioning ofthe clay particles. Figure 6 shows TEM micrographs at differentlocations of nanocomposites containing 2 wt% C30B and 20 and6 wt% PETi, respectively. This figure reveals that the C30B clayparticles are indeed preferentially localized in the PETi domainsrather than in the PET matrix. As expected, the greater PETiconcentration of 20 wt% generates larger domains of PETi inthe nanocomposites (Figs 6(a)–(d) versus Figs 6(e) and (f)). As aconsequence, the clay density in the separated (PETi) domains isindeed expected to be higher in the lower PETi content system(6 wt%), since the overall clay concentration remains the same inboth cases.

The preferred affinity of the organoclay for the PETi phase mayimpede its efficient distribution within the bulk phase. A ‘gooddistribution’ refers to a case where single layers and/or multilayerclay particles show a uniform presence within the whole material,both the domains and the matrix, even if clay aggregates might beobserved. On the other hand, the term ‘dispersion’ refers to howwell the clay layers are delaminated and form single layers at thenanoscale.24 Fig. 7 shows SEM micrographs of various PET-basednanocomposites with and without PETi, at a nominal content of 2wt% of organoclay C30B or N28E.

In the nanocomposite containing C30B in the absence ofionomer (Fig. 7(a)), the observed clay density is low and the particlesize is smaller than for samples containing the PETi (Fig. 7(b)). Thelow density of clay particles may be due to the limitation of

Figure 5. SEM micrographs of PET/PETi blends initially containing (a, b) 50 wt% PETi at different magnifications, (c) 30 wt% PETi and (d) 20 wt% PETi. Allmicrographs were obtained after extraction of the PETi phase.


44

3


Figure 6. TEM micrographs of nanocomposites containing 2 wt% C30B and (a-d) 20 wt% PETi and (e, f) 6 wt% PETi.

Figure 7. SEM micrographs of (a) PET/C30B, (b) PET/PETi/C30B, (c) PET/N28E and (d) PET/PETi/N28E.

SEM for sub-micrometer levels while smaller aggregates suggestthat clays are well dispersed in the PET matrix. By adding PETi innanocomposites containing C30B, more and larger clay particlesare visible in the SEM image and there are several ‘empty’spaces neighboring large agglomerates (Fig. 7(b)), indicating apoor distribution and dispersion of C30B particles in this system.The presence of large aggregates is due to concentration ofparticles in the PETi domains as discussed earlier. However, it isinteresting to note that this phenomenon is not observed in thenanocomposites containing N28E (Figs 7(c) and (d)).

In order to complement the information provided by the SEMmicrographs, XRD and TEM techniques were employed to give abetter understanding of the nanocomposite morphology. Inves-tigation of XRD patterns is an important task for evaluating theoverall morphology of nanocomposites. Morphological informa-tion on the nanocomposites can be inferred from the position,shape and intensity of the reflected X-ray beam. Figure 8 shows theXRD patterns of nanocomposites containing N28E and C30B. Thebasic interlayer spacing of N28E is 2.4 nm. The interlayer spacingof the organoclay in the PET/N28E nanocomposite increases up


44

4


Figure 8. XRD patterns of PET nanocomposites containing (a) 2 wt% N28Eand (b) 2 wt% C30B. Nanocomposites without ionomer (thick solid curves)and nanocomposites containing ionomer (narrow dotted curves).

to 3.4 nm (Fig. 8(a)). This increase is assumed to be the result ofthe diffusion of PET macromolecules within the silicate layers, andsuggests an intercalated morphology. For the N28E nanocompos-ite containing PETi, no peak is observed. This could be due tothe delamination of silicate nanolayers within the PET matrix anddisruption of the well-ordered structure of the nanolayers. Theweaker second-order peak also disappears for nanocompositescontaining PETi, which is another sign that the high degree of theperiodic order of N28E does not exist in these nanocomposites. Onincorporation of C30B into the PET matrix, the interlayer spacingof this organoclay increases from 1.8 to 3.6 nm (Fig. 8(b)). This is inagreement with the findings of Ghasemi et al.12,25 The incrementis here also attributed to the intercalation of PET chains within theclay galleries. The C30B nanocomposites containing PETi show nodiffracted peak, possibly because of clay dilution due to aggregateformation as illustrated by Fig. 7(b). However, the disappearanceof a diffraction peak cannot always be directly attributed to anexfoliated morphology. For these reasons, XRD analysis is not suf-ficient for morphological characterization of the nanocomposites,and TEM imaging is required for a better understanding of theinternal structure.

Figure 9 shows TEM micrographs of PET-based nanocompositescontaining 2 wt% N28E. For these nanocomposites, a betterdistribution of clay particles is achieved for samples containingno PETi (Fig. 9(a)) as the particles have a larger area in whichto be distributed. In contrast, in nanocomposites containing

PETi the clay particles are trapped in the restricted PETi phaseas discussed earlier and their distribution is hence constrained(Fig. 9(c)). However, the presence of PETi results in a betterdispersion of the clay layers due to a better affinity betweenPETi and clay, as mentioned before. A larger gallery spacing anda greater clay layer disorder due to the addition of PETi canbe observed by comparing the TEM images of nanocompositeswith and without PETi (Figs 9(d) and (b), respectively). It seemsthat a semi-exfoliated morphology is achieved for nanocompositescontaining PETi and N28E (Fig. 9(d)). For these systems, a significantamount of individual clay layers can be observed in the TEMmicrographs.

TEM micrographs of nanocomposites containing C30B areshown in Fig. 10 at various magnifications. The TEM observationsare in good agreement with the SEM ones. Figure 10(a) reveals agood distribution of the C30B clay particles within the PET matrix.At high magnifications, single layers, double layers and tactoidscontaining a few layers can be observed (Fig. 10(b)) indicating agood level of dispersion and exfoliation. However, after addingPETi, the clay particles are not distributed as well as in the PETmatrix alone (Fig. 10(c)). This is related to the lack of miscibilitybetween the PETi and PET phases and greater affinity of clay forthe PETi phase as discussed above. The TEM observations alsodemonstrate that the morphology of the N28E systems (Figs 9(c)and (d)) is very different from that of the C30B systems (Figs 10(c)and (d)), in which the PETi domains are larger and less uniform.As clays are mainly concentrated in the PETi domains in bothcases, it can explain the appearance of large clay aggregates in theC30B/PETi systems and not in the N28E/PETi ones.

To evaluate quantitatively the effect of the compatibilizer onthe PET-based nanocomposites containing the two organoclays,the number of platelets per clay particle was manually countedusing the TEM micrographs. Around 400 particles were counted toensure statistical validity of the analysis. The corresponding dataare shown in Fig. 11. In the case of PET/C30B nanocomposites,the count for single- and double-layer particles is the highestamong all of the nanocomposites (around 80%), indicating thatthis organoclay is both well dispersed and distributed in thePET matrix. In the presence of PETi, the amount of C30B clayaggregates (five or more platelets per particle) increases due to theconfinement of the clay particles within the small available area ofthe PETi phase. On the contrary, for the PET/N28E nanocompositesthe count for single- and double-layer particles is the lowest, andthis kind of organoclay is neither well dispersed nor distributedin the PET matrix alone (Figs 9(a) and (b)). However, by addingthe PETi to the nanocomposites containing N28E, the counts forsingle- and double-layer particles increase. While it is seen thatthe addition of PETi leads to C30B aggregate formation resultingin a worse morphology, the use of a more polar polymer enhancesthe gallery spacing in the case of N28E and results in more single-and double-layer particles. The better dispersion of C30B overN28E in PET could be due to the presence of the two ethoxy(or ethyloxy) groups in the intercalant of the former that canfacilitate the interaction with the carboxylic and hydroxyl groupsof PET molecules via hydrogen bonds or even chemical bonds(etherification or esterification). The better dispersion of N28Eover C30B in the PETi domains could be explained by the betteraffinity of PETi with N28E over C30B. The primary amine used inthe intercalant in N28E should provide a less crowded surface asopposed to the quaternary amine used in the intercalant in C30B,thus allowing a better interaction between the clay surface andPETi.


44

5


Figure 9. TEM micrographs of (a, b) PET/N28E and (c, d) PET/PETi/N28E at various magnifications.

To quantify further the degree of layer dispersion in thenanocomposites, the free-path spacing measurement introducedby Luo and Koo is employed.26 According to this method, adimensionless dispersion value D0.1 is calculated based on thedistribution of the free-path spacing distances between the claylayers according to the following equation:

D0.1 = 1.1539 × 10−2 + 7.5933 × 10−2(µ

σ

)

+ 6.6838 × 10−4(µ

σ

)2 − 1.9169 × 10−4(µ

σ

)3

+ 3.9201 × 10−6(µ

σ

)4(2)

where µ is the mean spacing between the clay layers and σ is thestandard deviation.

A value below 4% for the dimensionless dispersion parameterD0.1 suggests an immiscible system or microcomposite; anintercalated nanocomposite displays a dispersion value between 4and 8%; for an exfoliated nanocomposite this value is above 8%.26

While the D0.1 value increases from 4.3 to 6.3% for PET/N28E andPET/PETi/N28E, respectively, for samples containing C30B a reversetrend is observed. The D0.1 values for PET/C30B and PET/PETi/C30Bare 7.5 and 4.6%, respectively. PET/C30B has the highest D0.1 valueof 7.5% which is close to that of a fully exfoliated system. Thisvalue is in agreement with the one obtained by Ghasemi et al.12

for PET nanocomposites containing C30B obtained under the bestprocessing conditions. For the N28E nanocomposites, the additionof PETi transforms a near microcomposite (D0.1 = 4.3%) to an

intercalated system (D0.1 = 6.3%), in agreement with previousobservations.

Rheological properties of PET/PETi nanocompositesFigure 12 shows the complex viscosity and storage modulus ofneat PET and its nanocomposites. The various nanocompositesshow a shear-thinning behavior while neat PET displays a pseudo-Newtonian behavior for the whole frequency range. The lowestviscosity and storage modulus are observed for the nanocompositecontaining C30B and PETi. However, at high frequency, where thebehavior of the matrix is dominant, all the nanocomposites show alower viscosity and lower storage modulus than the neat polymer,presumably due to the degradation induced by the presence ofthe organic modifiers. The addition of PETi in the nanocompositescontaining N28E increases the melt viscosity and storage modulusfor the whole frequency range, in comparison to the PET/N28Esamples. This is most probably due to the better dispersion of theN28E particles in the nanocomposites containing PETi. In contrast,the presence of PETi in the nanocomposites containing C30Bdecreases the melt viscosity and storage modulus for the wholefrequency range in comparison to the PET/C30B samples. Thismay be the result of two factors: the poor dispersion of the C30Bparticles in the presence of the ionomer, and a plasticization effect.

In order to understand better the interactions between PETi andthe two nanoclays, the complex viscosity and storage modulus ofthe neat PETi and its corresponding composites containing 6 wt%N28E and C30B were measured. This large concentration of claywas used to amplify the effect on the rheological properties


44

6


Figure 10. TEM micrographs of (a, b) PET/C30B and (c, d) PET/PETi/C30B at various magnifications.

Figure 11. Number of platelets per particle histogram. The total numberof counted particles was around 400 for each nanocomposite.

of PETi. The results are presented in Fig. 13. Both the neatPETi and the PETi/6 C30B samples display a pseudo-Newtonianbehavior, while the sample containing N28E exhibits a markedshear-thinning behavior, which is one of the characteristics ofnanocomposites with a favorable morphology. In the case of

PETi/6 C30B, most probably C30B particles form large aggregatesleading to a pseudo-Newtonian behavior as for the neat PETi.In contrast, it seems that N28E particles form an interconnectednetwork-like structure, which is obvious at low frequency andsuggests a strong interaction between the PETi and N28E. Thelower complex viscosity of the PETi/6 C30B sample in comparisonto the neat PETi is largely due to the degradation of the matrixinduced by the organic modifier used for the C30B particles. Inaddition, it may suggest a weak interaction between this clay andthe ionomer, which is already evident from the SEM (Fig. 7(b)) andTEM (Figs 10(c) and (d)) micrographs.

Thermal propertiesThe influence of clay platelets on Tg, Tm, cold crystallizationtemperature Tcc, hot crystallization temperature Thc and crystalcontent of PET nanocomposite films was investigated using DSC.The thermal characteristics of the various films are given in Table 2.Tg of the nanocomposite films is somewhat lower than thatof neat PET, with a slightly larger reduction for nanocompositescontaining PETi, which suggests thermal degradation of PET chainsand/or plasticization. No significant difference is observed in themelting point of the nanocomposites in comparison with neat PET.For all the nanocomposites, Tcc and Thc decrease and increase,respectively, in comparison to neat PET. This reveals that clayparticles act as heterogeneous nucleating agents and promotenucleation of the semi-crystalline PET. The crystal content ofthe nanocomposite films is evaluated according to the following


44

7


Table 2. DSC results for neat PET and nanocomposite films containing C30B and N28E

Sample Tg (◦C) Tcc (◦ C) Thc (◦ C) Tm (◦ C) Crystalline fraction (%)

Neat PET 78.7 ± 0.1 135.7 ± 0.1 174.8 ± 0.1 243.1 ± 0.1 8.6 ± 0.5

PET/C30B 76.5 ± 0.2 126 ± 0.1 193.8 ± 1 244 ± 0.1 13.0 ± 0.3

PET/PETi/C30B 74.8 ± 0.1 129.2 ± 0.2 194 ± 2 244 ± 0.1 12.9 ± 0.1

PET/N28E 77.5 ± 0.1 133.3 ± 0.1 197 ± 3 244.2 ± 0.1 11.1 ± 0.5

PET/PETi/N28E 75.2 ± 0.1 125.9 ± 0.1 198 ± 0.3 244 ± 0.2 14.3 ± 0.3

Figure 12. (a) Complex viscosity and (b) storage modulus as functions offrequency for neat PET and its nanocomposites at 265 ◦C.

equation:

Xc = (�Hm − �Hcc)/α

�H0m

(3)

where α is the PET weight fraction, �Hm the enthalpy of melting,�Hcc the enthalpy of cold crystallization and �H0

m the enthalpyof melting of 100% crystalline PET (140 J g−1).27 As evident fromTable 2, the crystalline content of the nanocomposite films isslightly larger than that of neat PET films, which is anotherindication of a promoted crystallization in the presence of clay.

Barrier propertiesThe improvement in barrier properties of nanocomposites isgenerally explained by the tortuous path model.28,29 The presenceof silicate nanolayers, which are assumed to be impermeablelamellar fillers with a high aspect ratio, induces a more tortuous

Figure 13. (a) Complex viscosity and (b) storage modulus as functions offrequency for neat PETi and its composites with 6 wt% nanoclay at 250 ◦C.

gas path. Longer diffusion pathways are created, hence preventingthe gas from passing directly through the material.28,29 Crystallinityis another important factor, which can promote barrier propertiesof a material. Figure 14 shows the measured oxygen permeabilityof neat PET and its nanocomposites. For all the nanocompositefilms, the permeability is decreased in comparison to neat PET dueto both the presence of clay particles and a higher crystallinity. Thebest barrier properties are obtained for the nanocomposite filmcontaining C30B, for which a reduction of oxygen permeabilityup to 25% is observed in comparison to neat PET, confirming theresults of Ghasemi et al.12 By comparing Figs 9(a) and 10(a), it canbe concluded that the better distribution of C30B within the PETmatrix and the larger D0.1 value, as compared to N28E, explain thepermeability results. These data clarify the strong influence of thenanoclay distribution on the barrier properties. However, in thecase of clay particles like N28E, which cannot be well dispersed,


44

8


(1) (2) (3) (4) (5) (6)

Per

mea

bilit

y [µ

L m

−1 d

ay−1

atm

−1)]

4

5

6

7

8(1): Neat PET(2): PET/C30B(3): PET/PETi/C30B (MB)(4): PET/N28E(5): PET/PETi/N28E (MB)(6): PET/PETi

Figure 14. Oxygen permeability of neat PET and its nanocomposite films.

employing a more polar polymer (the sulfopolyester) can behelpful to achieve a better dispersion level. Miscibility betweenthe more polar polymer and polymer host matrix plays a key roleas the properties of nanocomposites are also controlled by thelevel of clay distribution. In this work, a fundamental study wasperformed to investigate the state of miscibility between PET andPETi and clarify the effect of PETi on the migration and partitioningof two different clay particles within the PET matrix, two aspectswhich were not covered in our previous work.13 Unfortunately theimmiscibility of PETi in PET does not permit a good distribution anddispersion of the clay particles outside of the PETi domains. Thusthe advantage of the good clay dispersion in an immiscible systemsuch as PET/PETi does not lead to improved barrier properties.

CONCLUSIONSMorphological, rheological, thermal and gas barrier properties ofPET nanocomposites containing C30B and N28E were studied.The effect of a sulfopolyester (PETi) on the morphology of thenanocomposites was investigated. Based on SEM and TEM results,PETi was found to be immiscible with PET, although the PETbackbone of this copolymer influences the crystallization of thePET matrix and leads to a melting point depression. Nanoclayparticles have greater affinity with the PETi phase due to theformation of favorable electrostatic interactions between theionic groups of this copolymer and nanoclay particles, whichlocalize themselves preferentially into the PETi domains. Althoughmigration of clay particles into droplets of the PETi phase restrictsdistribution of the particles, XRD, TEM and rheological results showthat for nanocomposites containing N28E, PETi acts as an effectiveexfoliation agent and improves the dispersion of the clay particles.

Higher crystallinity was observed for all nanocomposite filmsin comparison to that for neat PET. Finally, it was found that thegas barrier properties of the nanocomposites were improved bya good distribution of clay particles within the matrix. The best

nanoclay distribution and barrier properties were obtained forC30B in the absence of PETi.

ACKNOWLEDGEMENTSThe authors thank Drs H Ghasemi and X-F Xu for their valuablehelp in producing the nanocomposite films. They are also gratefulto W Leelapornpisit for preparing SEM and TEM micrographs.Financial support from NSERC (Natural Science and EngineeringResearch Council of Canada) in the context of the NRC-NSERC-BDCNanotechnology Initiative is gratefully acknowledged.

REFERENCES1 Bhattacharya SN, Kamal MR and Gupta RK, Polymeric Nanocomposites:

Theory and Practice. Hanser Gardner, Munich (2007).2 Todorov LV and Viana JC, J Appl Polym Sci 106:1659–1669 (2007).3 Ammala A, Bell C and Dean K, Compos Sci Technol 68:1328–1337

(2008).4 Calcagno CIW, Mariani CM, Teixeira SR and Mauler RS, Polymer

48:966–974 (2007).5 Chung JW, Son SB, Chun SW, Kang TJ and Kwak SY, Polym Degrad Stab

93:252–259 (2008).6 Stoeffler K, Lafleur PG and Denault J, Polym Degrad Stab 93:1332–1350

(2008).7 Shen YC, Harkin-Jones E, Hornsby P, McNally T and Abu-Zurayk R,

Compos Sci Technol 71:758–764.8 Jung MH and Chang JH, Polymer-Korea 31:518–525 (2007).9 Monemian SA, Goodarzi V, Zahedi P and Angaji MT, Adv Polym Technol

26:247–257 (2007).10 Mun MK, Kim JC and Chang JH, Polym Bull 57:797–804 (2006).11 Chisholm BJ, Moore RB, Barber G, Khouri F, Hempstead A, Larsen M,

et al., Macromolecules 35:5508–5516 (2002).12 Ghasemi H, Carreau PJ, Kamal MR and Chapleau N, Int Polym Process

26:219–228 (2011).13 Xu XF, Ghanbari A, Leelapornpisit W, Heuzey MC and Carreau PJ, Int

Polym Process 26:444–445 (2011).14 Cole KC, Macromolecules 41:834–843 (2008).15 Barber GD, Calhoun BH and Moore RB, Polymer 46:6706–6714 (2005).16 Sepe MP, Thermal Analysis of Polymers. Rapra Technology, Shrewsbury

(1997).17 Utracki LA (ed.), Polymer Blends Handbook, 1st edition. Springer, Berlin

(2002).18 Yang F, Qiu ZB and Yang WT, Polymer 50:2328–2333 (2009).19 Cheung YW and Stein RS, Macromolecules 27:2512–2519 (1994).20 Prud’homme RE, Polym Eng Sci 22:90–95 (1982).21 James M and Jonza RSP, Macromolecules 19:1946–1951 (1986).22 Martın Z, Jimenez I, Gomez-Fatou MA, West M and Hitchcock AP,

Macromolecules 44:2179–2189 (2011).23 Kontopoulou M, Liu YQ, Austin JR and Parent JS, Polymer

48:4520–4528 (2007).24 Abbasi S, Carreau PJ, Derdouri A and Moan M, Rheol Acta 48:943–959

(2009).25 Ghasemi H, Carreau PJ, Kamal MR and Tabatabaei SH, Polym Eng Sci

52:420–430 (2012).26 Luo ZP and Koo JH, Polymer 49:1841–1852 (2008).27 Soon KH, Harkin-Jones E, Rajeev RS, Menary G, McNally T, Martin PJ,

et al., Polym Int 58:1134–1141 (2009).28 Bharadwaj RK, Macromolecules 34:9189–9192 (2001).29 Russo GM, Simon GP and Incarnato L, Macromolecules 39:3855–3864

(2006).


Documents

Morphology and properties of polymer/organoclay nanocomposites based on poly(ethylene terephthalate) and sulfopolyester blends