halloysite nanotube nanocomposites: Structure and properties

Polymer 52 (2011) 4284e4295

Contents lists available at ScienceDirect

Polymer journal homepage: www.elsevier.com/locate/polymer

Water-assisted extrusion as a novel processing route to prepare polypropylene/ halloysite nanotube nanocomposites: Structure and properties B. Lecouvet a, M. Sclavons a, S. Bourbigot b, J. Devaux a, C. Bailly a, * a

Bio- and Soft Matter (BSMA), Institute of Condensed Matter and Nanosciences (IMCN), Université catholique de Louvain (UCL), Croix du Sud 1, Box 4, B-1348 Louvain-la-Neuve, Belgium b Unité Matériaux et Transformations, Ecole Nationale Supérieure de Chimie de Lille (ENSCL), 59652 Villeneuve d’Ascq. France

a r t i c l e i n f o

a b s t r a c t

Article history: Received 4 May 2011 Received in revised form 18 July 2011 Accepted 19 July 2011 Available online 23 July 2011

Naturally occurring halloysite nanotubes (HNTs) are used to prepare Polypropylene (PP)/HNTs nanocomposites using a novel “one step” water-assisted extrusion process with and without the use of a PPgraft-maleic anhydride (PP-g-MA) as compatibilizer. In order to analyze the influence of PP-g-MA and/or water injection on the HNTs dispersion and therefore on nanocomposite properties, structural analysis (SEM and TEM) is combined with rheological and thermo-mechanical experiments. The best clay dispersion is obtained when compatibilizer and water injection are combined together (synergistic effect). As a consequence, the linear viscoelastic properties and the dynamic storage modulus are dramatically enhanced for this system. A mechanism explaining the interaction between HNTs and PP-gMA in presence of water is proposed. The thermal stability and flame retardant property are also investigated. Thermal analyses reveal two opposite effects of HNTs on the thermal behaviour of PP. A surface catalytic action of the halloysite speeds up thermal degradation of PP. However, this effect is reduced with improved HNTs dispersion, presumably via an entrapment mechanism of the decomposition products inside the lumens. Finally, cone calorimeter results show that low flammability of nanocomposites is only achieved when combining water injection and PP-g-MA. In view of these results, PP/HNTs nanocomposites prepared using this novel processing route are promising candidates for flame retardant applications. Ó 2011 Elsevier Ltd. All rights reserved.

Keywords: Halloysite nanotubes Polypropylene Nanocomposites

1. Introduction Polymer nanocomposites based on inorganic clay minerals have drawn a great deal of attention during the last two decades. The major reason is related to the peculiar and fascinating properties of the polymer matrix that could be obtained at very low filler contents. Compared to the neat polymer matrix, clay-based polymer nanocomposites exhibit enhanced mechanical properties, reduced gas permeability and improved thermal stability and flame retardant behaviour [1e6]. The final properties of nanocomposites depend greatly on several factors including the chemistry of the polymer matrix, its compatibility with the clay filler, the geometry of the filler, its degree of dispersion and orientation inside the matrix and also the preparation method [7,8].

* Corresponding author. Tel.: þ32 10 478412; fax: þ32 10 451593. E-mail addresses: [email protected] (B. Lecouvet), christian.bailly@ uclouvain.be (C. Bailly). 0032-3861/$ e see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.polymer.2011.07.021

Among different preparation methods, melt-blending by extrusion remains the most usual way to produce polymer/clay nanocomposites. Indeed, it is cost-effective as well as environmentfriendly and thus quite attractive to the polymer industry [1,9]. Most of the polymer/clay nanocomposites prepared using this route are based on montmorillonite (MMT). These layered silicates are naturally stacked into aggregates and require the use of surfactants to increase the interlayer spacing in order to achieve a delaminated structure [2,10]. However, this organic modification presents some disadvantages: a decrease of the binding energy between polymer and filler, an acceleration of the thermooxidative degradation of the polymer matrix and a reduction of the mechanical performances of the material due to the plasticizer effect of the organic modifier [11,12]. Halloysite, another type of clay nanofiller combining the geometry of nanotubes and the chemistry of kaolinite are recently focusing a lot of attention. Halloysite is a naturally occurring aluminosilicate (Al2Si2O5(OH)4$2H2O) with a predominantly hollow tubular structure mined from natural deposits in different countries such as America, China, New Zealand, France and Belgium

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

[13]. Halloysite nanotubes (HNTs) exhibit a two-layer structure. The external surface of the tube has a tetrahedral sheet structure and is mainly composed of siloxane groups (SieOeSi), whereas the internal surface has an octahedral structure with aluminol groups (AleOH). The cylindrical shape results from a mismatch in the twolayered alignment of the tetrahedral sheet of silica bonded to the octahedral sheet of alumina [14]. Due to the multi-layer structure, HNTs exhibits unique surface chemical property with only a few hydroxyl groups located on the external surface [15]. As compared to other layered silicates, halloysite can thus be more easily dispersed in a polymer matrix due to the weak secondary interactions among the nanotubes via hydrogen bonds and van der Waals forces [16]. This lack of cohesion between HNTs opens the doors to their use in the field of polymer nanocomposite prepared by extrusion. The increasing number of publications on the preparation and characterization of halloysite nanocomposites with different polymer matrices such as polyamide, polypropylene, natural rubber, etc, shows its considerable scientific and technological potential [17e24]. Most of the research has so far focused on the mechanical properties and have reported enhanced tensile modulus, tensile strength and elongation at break at low HNTs loading [18e20,22]. In the specific case of polypropylene (PP), which is one of the most important commodity polymers, Du et al. have showed that organomodification of the HNTs surface using silane is required in order to achieve a uniform dispersion of the clay nanotubes in the polymer matrix, resulting in higher storage modulus. This result has been attributed to the improved interfacial adhesion between functionalized filler and matrix [16]. In another study, the same research group demonstrated that for these PP/HNTs nanocomposites, the thermal stability and resistance to flammability of PP are enhanced due to a combination of several factors including presence of iron in HNTs, barrier properties of the nanotubes and an entrapment mechanism of the initial degradation products inside the lumen structure of the nanotubes [25]. Similar observations have been reported in literature for other polymeric materials [18,24,26]. Specifically for PA 6/halloysite nanocomposites, Marney et al. have shown that the fire performance of PA 6 is mainly improved by the formation of an insulating barrier for heat and mass transport during the burning process, which increases the time to burn by more than a factor two [26]. In a previous work, we have also reported a good dispersion associated with reinforced thermal properties for unmodified HNTs in a medium polar matrix (PA 12) [21]. Nevertheless, all studies have shown that without any surface modification of the nanotubes, nanocomposites usually exhibit poor HNTs dispersion in apolar matrices with limited final properties. The main reason is the poor interfacial adhesion between the polymer matrix and the clay filler due to their huge polarity difference. Compatibility between polymer and clay platelets is well known to be crucial to obtain nanocomposites with high level of dispersion. For PP/clay nanocomposites, PP-graft-maleic anhydride (PP-gMA) is often used as compatibilizer in order to improve the affinity between the hydrophilic surface of clay and the hydrophobic PP [27e29]. However, even with compatibilizer, PP/layered silicate nanocomposites require organomodification of the clay for optimal exfoliation, which is still a practical limitation due to the poor thermal stability of the alkylammonium surfactants used [30e32]. In the field of HNTs based nanocomposites, Pasbakhsh et al. investigated the effect of an ethylene propylene diene monomer (EPDM) grafted with maleic anhydride (MAH-g-EPDM), as compatibilizer, on the structural and mechanical properties of EPDM/ unmodified HNTs nanocomposites [33]. After vulcanization by sulphur under a Moving Die Rheometer, a limited distribution of HNTs inside the EDPM matrix was reported at high HNTs loading

4285

using MAH-g-EDPM, with a combination of HNTs-poor and HNTsrich domains. This result was attributed to the low concentration of polar MAH-g-EPDM compared to the non-polar EPDM matrix. Associated mechanical properties were only partially improved using compatibilizer with a tensile modulus at 100% of elongation in the same order of magnitude for both EPDM and EPDM/MAH-gEPDM nanocomposites below 30 phr HNTs. The lack of efficiency of compatibilizing agents and the use of continuous melt-blending process like extrusion gave the opportunity to design new preparation methods for polymer/clay nanocomposites. Water-assisted extrusion was first used to prepare polyamide/ untreated clay nanocomposites [34]. Polyamide 6 (PA 6) and clay were melt mixed with water in a twin-screw extruder equipped with a water injection system and a degassing zone. Water was injected into the extruder downstream at high pressure and high temperature to remain in the liquid state and was removed at the end of the process by degassing using a vacuum pump. Exfoliated PA6/clay nanocomposites were produced using this technique without requiring any clay organomodification. Since this pioneering work, several investigations have been conducted on the water-assisted preparation of polymer/clay nanocomposites [35e39]. Recently, water-assisted extrusion was applied to prepare PP/layered silicate nanocomposites [40]. In this study, a dual role of water during the melt-blending step was reported. On one hand, water acts as a solvent and promotes the clay dispersion. On the other hand, it favours bonding between PP-g-MA and the organoclay surfactant through Fisher esterification. As a result, nanocomposites exhibit higher degree of clay exfoliation and remarkably enhanced rheological, mechanical and thermal properties. In this context, the present paper discusses the elaboration of PP/untreated HNTs nanocomposites using the previously described water-assisted extrusion process. The influence of PP-g-MA as compatibilizer and/or water injection on the structural, rheological, thermo-mechanical, thermal and fire properties of PP/HNTs nanocomposites is presented. A mechanism explaining the combined role of PP-g-MA and water injection at the different stages of PP/ HNTs nanocomposites preparation is also proposed. 2. Experimental 2.1. Materials A commercially available polypropylene (HH420FB) was purchased from Borealis (Belgium). PP-g-MA Polybond 3200 (Plb) has been provided by Chemtura (Belgium). Halloysite nanotubes were purchased from SigmaeAldrich (Germany) without any chemical modification. Typical specific surface area of this halloysite is 64 m2/g; cation exchange capacity of 8 meq/g; pore volume of 1.25 ml/g; and specific gravity of 2.53 g/cm3. 2.2. Sample preparation All references and composites were prepared in a co-rotating twin-screw extruder Krupp WP ZSK25. The extruder has a screw length of 1000 mm with a L/D ratio of 40. The extrusion process was conducted at 200  C at a screw speed of 200 rpm. The polymer and HNTs were introduced through the feeder in the melting zone using a throughput of 7 kg/h. The residence time in the extruder under these conditions is around 2 min. Water, when used, was pumped into the extruder at a 50 ml/min flow rate in the high compression zone. The screw is specifically tailored to allow the increase of the pressure in this zone up to 125 bar, which maintains water in the liquid state. Water is next degassed in the transport zone and fully

4286

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

Table 1 Composition of the PP references, Plb references and their respective nanocomposites (Polypropylene (PP), PP-g-MA (Plb), HNTs (H), Water injection rate (W)). Samples

PP [wt%]

Plb [wt%]

H [wt%]

W [ml/min]

PP PP-W PP-H4 PP-H8 PP-H16 PP-H4-W PP-H8-W PP-H16-W Plb Plb-W Plb-H30 Plb-H30-W

100 100 96 92 84 96 92 84 0 0 0 0

0 0 0 0 0 0 0 0 100 100 70 70

0 0 4 8 16 4 8 16 0 0 30 30

0 30 0 0 0 30 30 30 0 30 0 30

a Reichert Microtome. Ultrathin sections of approximately 95 nm in thickness were cut using an ultra-diamond knife with a cut angle of 35 (Diatome, Switzerland) and collected on 400 mesh copper grids. 2.5. Fourier Transform Infrared (FTIR) spectroscopy FTIR spectra were recorded on a Nicolet Nexus 870 FTIR spectrometer on compression-molded films in transmission mode by collecting 64 scans at a resolution of 2 cm1 over a spectral range from 4000 to 400 cm1. The spectra were normalized by setting the peak height of 1167 cm1 absorption band (eCH3 symmetric deformation of PP) to an absorbance of 1. 2.6. Melt rheology

removed using a vacuum pump. After melt mixing, samples were immediately quenched at room temperature and pelletized. Three kinds of composites were elaborated. In a first step, PP references (PP) and PP/HNTs nanocomposites (PP-H) were prepared at different HNTs loadings (4, 8 and 16 wt%). PP-g-MA references (Plb) and respective nanocomposites (Plb-H with 30 wt% HNTs) were also produced to analyze by FTIR spectroscopy the interactions between PP-g-MA and HNTs (Table 1). In a second step, PP/PP-g-MA 10wt% blends (PP-Plb) and respective nanocomposites (PP-Plb-H) were prepared at the same HNTs contents (4, 8 and 16 wt%) by a one-pot process (Table 2). In all cases, samples were prepared with and without water injection. The samples obtained were injection-molded using a KrausseMaffei type KM 80-160E injection-molding machine into standard dog-bone specimens (ASTM 527) for morphological, rheological and thermo-mechanical analyses. The barrel temperatures ranged within 170e200  C and the mold temperature was kept at 25  C. The specimens for the flammability measurements were prepared by compression molding at 240  C for 2 min. Before injection molding and compression molding, the pelletized samples were first dried in a vacuum oven at 80  C for 48 h. 2.3. Scanning electron microscopy (SEM)

Melt rheological properties were obtained with the help of a Bohlin GEMINI II rotational rheometer at 200  C in a nitrogen atmosphere with a plateeplate geometry. Oscillatory dynamic measurements were performed at a constant strain of 4%, within the linear viscoelastic range of the materials as checked by strain sweeps. The frequency test range was between 0.01 and 100 rad/s. Samples used were disks with a diameter and a thickness of 25 and 2 mm, respectively. Before experiments, rheological test specimens were dried at 80  C for 48 h in vacuum oven to prevent degradation of PP during experiments. 2.7. Size exclusion chromatography (SEC) Weight-average molar mass (Mw) of references and nanocomposites was measured on a Waters Alliance GPCV 2000 equipped with two Styragel HT 6E and one Styragel HT 2 columns. A differential refractometer and a viscometer were used as detectors. The mobile phase was 1,2,4-trichlorobenzene (TCB). The sample concentration was 1.5 mg/mL in TCB and the dissolution was achieved by shaking at 160  C for 1 h. The injection volume was 215 mL and the temperature was held constant at 145  C throughout the analysis. Prior to analysis, the system was calibrated with PS standards according to ISO 16,014-2 specification. 2.8. Dynamic mechanical analysis (DMA)

The morphologies of PP/HNTs nanocomposites were observed on cryofractured surfaces with the help of a LEO 982 (Zeiss) scanning electron microscope.

Dynamic mechanical analysis measurements were performed using a DMA/SDTA 861e (Mettler-Toledo, Switzerland) on rectangular specimens prepared by compression molding with

2.4. Transmission electron microscopy (TEM) A transmission electron microscope LEO 922 (Zeiss) with a 200 kV acceleration voltage was used to study the dispersion of HNTs inside the polypropylene matrix. The specimens for TEM analysis were cut from bulk compounds at room temperature using

Table 2 Composition of the PP-Plb blends and nanocomposites (Polypropylene (PP), PP-gMA (Plb), HNTs (H), Water injection rate (W)). Samples

PP/Plb10wt% [wt%]

H [wt%]

W [ml/min]

PP-Plb PP-Plb-W PP-Plb-H4 PP-Plb-H8 PP-Plb-H16 PP-Plb-H4-W PP-Plb-H8-W PP-Plb-H16-W

100 100 96 92 84 96 92 84

0 0 4 8 16 4 8 16

0 30 0 0 0 30 30 30 Fig. 1. TEM micrograph of HNTs.

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

4287

Fig. 2. SEM micrographs of: (a) PP-H8, (b) PP-H8-W, (c) PP-Plb-H8 and (d) PP-Plb-H8-W nanocomposites at a magnification of 2000. Circles underline HNTs aggregates.

dimensions of 2.0  4.0  9.0 mm3. The measurements were performed under tension at a constant frequency of 1 Hz. The amplitude of tensile strain was 0.90% and the heating rate was 3  C/min. Measurements were performed in a temperature interval from 40  C temperature to 100  C. Before the experiments, the samples were dried at 80  C for 48 h under vacuum. 2.9. Thermal gravimetric analysis (TGA) The thermal stability of the references and nanocomposites was studied by TGA. The analyses were performed using a TGA/SDTA 851e (Mettler-Toledo, Switzerland). Solid samples (z13 mg) of around 1 mm thickness were heated from room temperature to 600  C at different constant heating rates (5, 10 and 20  C/min) under nitrogen atmosphere with a flow rate of 100 ml/min. TGA

results are presented as temperatures at 5% (T5), 10% (T10) and 50% (T50) weight loss. These temperatures are usually considered as the initial decomposition temperature, the onset and the midpoint of the degradation, respectively. TGA measurements were also performed in isothermal conditions at 360  C under nitrogen atmosphere (100 ml/min). The mass difference curve, D(T), between the experimental and the theoretical TG curves is computed as follow:    

MPP(T): TG curve of neat PP MH(T): TG curve of pristine HNTs MExp(T): TG curve of PP-H8 nanocomposite Mth(T): theoretical TG curve computed by linear combination between experimental TG curves of PP and HNTs: Mth(T) ¼ x  MH(T) þ (1  x)MPP(T)  D(T): mass difference curve: D(T) ¼ MExp(T)  Mth(T)

Fig. 3. SEM micrographs of PP-Plb-H16-W nanocomposite at a magnification of: (a) 500 and (b) 10,000. Circles underline HNTs aggregates.

4288

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

This kind of curve allows to show any potential increase or decrease in the thermal stability of the polymer matrix related to the presence of one or several additives [41].

discuss the relationship between the nanomorphology and the physical properties. For the best system, a study of the properties as a function of the filler content is further presented.

2.10. Fire properties

3.1. Morphological characterization

The cone calorimeter tests were carried out using an FFT (Fire Testing Technology) Mass Loss Calorimeter following the procedure defined in ASTM E 906. The equipment is similar to that used in oxygen consumption cone calorimetry. The only difference is the use of a thermopile at the top of the chimney to measure to heat release rate (HRR) instead of employing the oxygen consumption principle. The specimens (100 mm  100 mm  3 mm) were irradiated in a horizontal position by an external flux of 50 kW/m2 and the flame ignition was initiated using a spark igniter. The flammability parameters considered in this study are: heat release rate (HRR), peak of heat release rate (PHRR), total heat release (THR), residue and the time to ignition (TTI). The cone data reported in this paper are based on the average of three replicated experiments and are reproducible with variation less than 10%.

The morphology of the as received HNTs is shown in Fig. 1. The average dimensions of the tubular structure were obtained by measuring 20 randomly chosen nanotubes. The inner tube diameter is around 15 nm, the outer tube diameter ranges between 30 and 100 nm and the aspect ratio is between 10 and 40. SEM and TEM analyses were performed in order to investigate the effect of the compatibilizer and water injection on the structure of PP/HNTs nanocomposites. SEM micrographs show in Fig. 2 the morphology of the various PP/HNTs systems with 8 wt% HNTs. The lower level of HNTs dispersion is observed for PP-H8 nanocomposites (Fig. 2a) with a huge aggregation of HNTs and only few individual nanotubes (i.e. microcomposite). For both nanocomposites prepared using PP-g-Ma (PP-Plb-H8, Fig. 2b) or water injection (PP-H8-W, Fig. 2c), the scanning electron micrographs reveal an heterogeneous structure characterized by a combination of areas with well-dispersed HNTs (i.e. nanocomposite) and other ones with the clay nanotubes in their original aggregate state (i.e. microcomposite). The best clay dispersion is observed when combining PP-g-MA and water injection in PP-Plb-H8-W system, in which HNTs are dispersed individually in PP matrix without any

3. Results and discussion In this section, we mainly report a comparison between different PP nanocomposites containing 8 wt% HNTs and we

Fig. 4. TEM micrographs of: (a) PP-H8, (b) PP-H8-W, (c) PP-Plb-H8 and (d) PP-Plb-H8-W nanocomposites at a magnification of 2300.

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

4289

Fig. 5. (a) Storage modulus (G0 ) and (b) complex viscosity (h*) at 200  C as a function of frequency for references, PP-H and PP-Plb-H nanocomposites prepared with or without water injection.

indication of nanotube aggregation (i.e. nanocomposite). It is also noteworthy to mention that, for this system, the good HNTs dispersion is almost preserved at high HNTs loading (16 wt% HNTs), as presented in Fig. 3a. Indeed, only few aggregates are observed on the whole investigated area. Furthermore, for all systems, little cavities are present on the fracture surface, suggesting that all the nanotubes do not break during cold fracture but can be pulled out of the PP matrix (Fig. 3b). Such phenomenon has already been reported in our previous work [21] and is related poor interfacial interaction between pristine HNTs and PP. TEM micrographs in Fig. 4. corroborate SEM observations with a preferential orientation of HNTs by elongational flow in the extrusion direction. The single use of compatibilizer (Fig. 4b) or water injection (Fig. 4c) partially improved the HNTs dispersion compared to PP-H8 nanocomposites (Fig. 4a), while their combined use (Fig. 4d) give rise to a uniform dispersion of the clay nanotubes in the PP matrix. It must be stressed that the integrity of the nanotubes is partially destroyed due to cutting by the ultramicrotome. 3.2. Rheological properties Rheological characterization is usually used as an indirect method to evaluate the dispersion state of the filler in polymer nanocomposites. Indeed, the melt flow behaviour is directly related to the structure, the size and the shape of the dispersed phase. An increase of the storage modulus (G0 ) and the complex viscosity (h*) in the low frequency regime may be associated to attractive interfacial interactions between polymer chains and filler surface and/or to better dispersion of the filler in the polymer matrix [42].

Moreover, the formation of a percolated filler network structure in the material is usually characterized by the presence of a plateau for G0 at low frequencies combined with a continuous increase of h* [43,44]. Fig. 5a and b show G0 and h* curves of references and nanocomposites melt blended with 8 wt% HNTs, respectively. Compared to respective references, a more pronounced increase of G0 is observed in the low frequency region for PP-Plb-H nanocomposites. Furthemore, the use of water injection when producing PP-Plb-HW nanocomposites leads to the highest enhancement of the rheological properties. Fig. 5b shows similar trends for h* curves. Rheological measurements are in perfect agreement with the better HNTs dispersion observed in the morphological study for PPPlb-H-W nanocomposites. Indeed, the higher G0 and the smaller the slope, the better the dispersion of HNTs in PP [42]. Fig. 6a and b show the viscoelastic properties of the best system at different HNTs contents (4, 8 and 16 wt%). As the nanoclay content increases, the storage modulus and the complex viscosity decrease at high frequencies, while a plateau appear at low frequencies for G0 (Fig. 6a) combined with a continuous increase of h* (Fig. 6b). These results indicate that two mechanisms are in competition. On one hand, it is well known from literature that the viscoelastic properties of PP at experimentally accessible high frequencies are directly related to the polymer terminal relaxation and hence to the molecular weight of the matrix [45]. The SEC data listed in Table 3 show a significant decrease of the PP weightaverage molar mass ðMw Þ with increasing the halloysite content, while the use of water injection during extrusion does not modify significantly Mw of PP and nanocomposites. For instance, M w of PP is reduced by 25% upon adding 16 wt% HNTs. Therefore, compared

Fig. 6. (a) Storage modulus G0 and (b) complex viscosity (h*) at 200  C as a function of frequency for PP-Plb-H-W nanocomposites.

4290

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295 Table 3 Weight-average molar mass ðM w Þ values determined from SEC measurements for neat PP and PP-H nanocomposites. Samples

Mw [kg/mol]

PP PP-W PP-H4 PP-H16 PP-H16-W PP-H16-W (annealed)

404 413 390 290 302 301

to neat PP, nanocomposites have lower moduli and viscosity at high frequencies associated to shorter relaxation times of polymer chains. The M w decrease results from a thermo-mechanical degradation occurring during melt processing due to higher local shear stress applied by halloysite nanotubes on PP chains. Indeed, the hypothesis of pure catalytic thermal degradation during extrusion can be rejected as SEC measurements indicate that M w of PP-H16-W remains around 300 kg/mol after annealing during 1 h at the extrusion temperature. On the other hand, dynamic properties in the low frequency regime are directly related to the filler dispersion [42e44]. Therefore, for PP-Plb-H-W nanocomposites with well-dispersed HNTs, nanotube-nanotube interactions start to dominate as the halloysite content increases, leading in the formation of a percolated network structure starting from 16 wt% HNTs. The transition from a liquidlike to a solid-like behaviour is visible in rheological measurements by a frequency independent storage modulus associated to a continuous rise of the complex viscosity (Fig. 6). 3.3. Interfacial interaction between PP-g-MA and HNTs The presence of possible interactions between halloysite and the PP-g-MA (Plb) matrix has been investigated by FTIR. The infrared spectra of pure halloysite, virgin Plb, Plb-H30 and Plb-H-W30 nanocomposites are presented in Fig. 7. The HNTs (H) infrared spectrum (Fig. 7a) shows absorption bands around 1033 cm1 and 912 cm1, which are attributed to the SieO stretching vibrations and AleOH vibrations bands, respectively. Two main absorption bands are observed for Plb (Fig. 7b) in the carbonyl region: the maleic anhydride band around 1780e1790 cm1 and the carboxylic acid band around 1714 cm1. The latter one results from the hydrophilicity of the Plb anhydride moieties which are hydrolyzed by ambient moisture absorption into carboxylic diacid functions [46]. It can be noted from Fig. 7a that the differences between Plb spectra and nanocomposites spectra are due to the nanoclay addition. Moreover, no clear distinguishable shift in the characteristic peaks of HNTs is observed between the spectra of pristine

HNTs and nanocomposites, even using water injection (Plb-H30W). From these observations, there is no clear evidence of chemical reactions and/or physical interactions between filler and Plb matrix. Similar research reported in literature indicates that water injection seems to have a dual role on the organic filler dispersion [40]. Water injected at high pressure into the extruder during the melt-blending step remains liquid and acts as a processing aid, improving the HNTs distribution. Moreover, injected water in the high compression zone hydrolyzes anhydride functions of PP-g-MA into carboxylic diacid functions and thus should enable the formation of interfacial hydrogen bonds with the siloxane groups (SiO) located at the external surface of the halloysite nanotubes and the aluminol groups (Al-OH) situated on the edges of the tubes. The potential presence of these favorable interactions during extrusion between HNTs and Plb should also contribute to generate a more homogeneous dispersion of the clay nanotubes into molten polymer. The absence of interfacial hydrogen bonds in the FTIR spectra, while morphological and rheological characterizations clearly show a synergistic effect of Plb and water injection on the HNTs dispersion could thus be explained schematically by a three steps mechanism (Fig. 8). In the first melting zone (Fig. 8a), HNTs and polymer are introduced simultaneously and melt mixed together. In the high compression zone (Fig. 8b), water is injected and should enable H-bonds between carboxylic diacid functions of PP-g-MA and HNTs surface. A good dispersion of the nanotubes is then achieved at the end of the second zone before degassing. In the terminal plastification zone (Fig. 8c), after water degassing, anhydride functions are reformed with no hydrogen bonds anymore between HNTs and Plb, while the good dispersion of the filler in the polymer matrix still remains.

3.4. Thermo-mechanical properties In order to investigate the influence of the HNTs dispersion state on the thermo-mechanical properties of nanocomposites, DMA measurements were performed and Fig. 9a shows the temperature dependence of the storage modulus (E0 ) of references and different nanocomposites with 8 wt% HNTs. Irrespective of the mixing strategy used, the storage modulus increases at all temperatures compared to its respective virgin polymer. The extent of the thermo-mechanical properties enhancement of PP/HNTs nanocomposites mainly depends on three dominant factors: the dispersion state of the nanotubes inside the polymer matrix, their high intrinsic stiffness and the interfacial adhesion between filler and matrix. In this work, only the first two factors have to be taken into account due to the lack of

Fig. 7. FTIR spectra of HNTs (H), Plb, Plb-H30 and Plb-H30-W samples between: (a) 1150 and 950 cm1 and (b) 1820 and 1680 cm1.

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

4291

Table 4 Thermo-mechanical properties of references and PP/HNTs nanocomposites: glass transition temperature (Tg), storage modulus (E0 ) below and above Tg, and the reinforcement factor (E0 composite/E0 matrix) at respective temperatures.

Fig. 8. Scheme of water-assisted extrusion process of Plb-H-W nanocomposites: (a) melting zone, (b) high compression zone and (c) plastification zone.

Samples

Tg [ C]

E0 (40  C) [MPa]

E0 c/E0 m (40  C) [%]

E0 (25  C) [MPa]

E0 c/E0 m (25  C) [%]

PP PP-Plb PP-H8 PP-H8-W PP-Plb-H8 PP-Plb-H4-W PP-Plb-H8-W PP-Plb-H16-W

9 9 6 6 6 6 6 7

3854 3301 4315 4741 4171 3719 5085 5451

1 1 1.12 1.23 1.26 1.13 1.54 1.65

1944 1532 2028 2229 2001 1688 2209 2597

1 1 1.04 1.15 1.31 1.10 1.44 1.70

a function of filler loading. As the HNTs content increases, E0 increases gradually over the whole range of temperatures. However, the lower relative increase of E0 at higher HNTs content (16 wt%, Table 4) is probably due to the presence of few aggregates in the polymer matrix (see Fig. 3a). The glass transition temperature (Tg) is usually defined as the peak in the loss modulus versus temperature curves. Tg values for references and nanocomposites are also listed in Table 4. Irrespective of the mixing process used, a slight decrease of Tg is observed for all nanocomposites compared to neat polymer. This effect has been already reported in the literature and is attributed to the reduced entanglements and interactions among PP chains due to the presence of the nanotubes, enhancing the motion of the polymer chains [21,47,48]. 3.5. Thermal stability

primary interaction between the pristine nanotubes and the polymer chains, as suggested previously in FTIR study. Moreover, as described in the morphological (SEM and TEM) and rheological analyses, compared to the simplest PP-H system, a gradually improved HNTs dispersion is achieved using only compatibilizer or water injection, with the best HNTs exfoliation obtained for PP nanocomposites prepared using both Plb and water-assisted extrusion. Consequently, the latter nanocomposite exhibits the highest enhancement of E0 compared to the other systems. The storage modulus and the reinforcement factor below (40  C) and above (25  C) the glass transition temperature are summarized in Table 4. For instance, the storage modulus of PP-Plb-H8-W nanocomposite at 40 and 25  C is improved by 54 and 44%, respectively, while only limited improvement (12 and 4%, respectively) is observed at the same HNTs loading without using PP-g-MA and water injection. Fig. 9b shows the temperature dependence of the storage modulus of PP-Plb matrix and PP-Plb-H-W nanocomposites as

The thermal behaviour of the different composites based on 8 wt% HNTs was studied by TGA under inert atmosphere and correlated with the structural investigations. The TGA curves and the onset degradation temperatures (T5%) are reported in Fig. 10a for a constant heating rate of 10  C/min. Under nitrogen, PP degrades through a radical chain process with an onset degradation temperature of 407  C. Regardless of the extrusion strategy and the heating rate, the introduction of halloysite causes a decrease of all PP decomposition temperatures. For 8 wt% HNTs, T5% of PP is reduced by approximately 40  C compared to pristine polymer. However, the thermal behaviour is improved for PP nanocomposites prepared with water injection and/or compatibilizer. Furthermore, the highest thermal stability is observed for PP-Plb-H8-W nanocomposite extruded using compatibilizer and water injection. This result is consistent with the better clay dispersion reported by microscopy and rheology. For the latter composite, HNTs protect the polymer in the early

Fig. 9. Storage modulus (E0 ) curves of: (a) references and different nanocomposites at 8 wt% HNTs; (b) PP-Plb-H-W nanocomposites as a function of HNTs content.

4292

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

Fig. 10. TGA curves under N2 atmosphere of references, PP-H8 and PP-Plb-H8 nanocomposites prepared with or without water injection: (a) at a constant heating rate of 10  C/min; (b) in isothermal conditions at 360  C.

Fig. 11. Mass difference TG curve of PP-H8 nanocomposite (10  C/min, N2 flow).

stage of weight loss (T5%), whereas most of the thermal degradation takes place at a lower temperature during the main volatilization step. The residue at the end of the thermal volatilization process corresponds the initial halloysite content of the composites (8 wt%). The overall effect of halloysite on the PP thermal behaviour can be explained by a dual role of the nanotubes. On one hand, halloysite exhibits a catalytic action on PP pyrolysis due to the presence of hydroxyl groups (AleOH) that act as Brönsted active sites in catalysis. The catalytic effect of clay nanoparticles on the thermal degradation of PP has already been reported in the literature [49]. On the other hand, a good dispersion of the nanotubes in the

polymer matrix favours: (i) the formation of a tortuous diffusion path for the released degradation products and (ii) the entrapment of volatile products at the surface and inside the hollow tubular structure of HNTs [21]. Both effects hinder the out-diffusion of the volatile decomposition products, resulting in an effective delay of the mass transfer and an increased thermal stability of the polymer matrix. The combined role of HNTs was highlighted in isothermal TGA experiments at 360  C under nitrogen atmosphere (Fig. 10b). The initial volatilization rate of PP-H8 nanocomposites dramatically increases compared to neat PP due to the domination of the catalytic effect. On the other hand, the good HNTs dispersion in PP-Plb-H8-W nanocomposites dominates the HNTs catalytic effect, resulting in a lower initial decomposition rate compared to PP. The mass difference curve is usually plotted to investigate the eventual interactions between filler and matrix during degradation (Fig. 11). As the experimental curve is lower than the theoretical one with a negative mass difference curve between 300 and 450  C, it strongly reinforces the hypothesis that the presence of the halloysite nanotubes leads to a thermal destabilization of the polypropylene.

3.6. Fire properties The small-scale cone calorimeter was used to evaluate the characteristics of references and composites on combustion. The heat release rate (HRR) is one of the most important parameters to describe the fire behaviour of a material and is defined as the heat release per unit surface area of burning sample. HRR curves for

Fig. 12. Heat release rate as a function of time for: (a) references and composite systems with 8 wt% HNTs; (b) PP-Plb-H-W nanocomposites as a function of HNTs content (heat flux ¼ 50 kW/m2).

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295 Table 5 Cone calorimeter data of references and PP/HNTs nanocomposites. Tests were conducted at a heat flux of 50 kW/m2. Samples

Residue [%] (1)

TTIa [s]

PHRR [kW/m2]

THR [MJ/m2]

PP PP-Plb PP-H8 PP-H8-W PP-Plb-H8 PP-Plb-H4-W PP-Plb-H8-W PP-Plb-H16-W

0 0 7.9 8 7.2 4.1 8.5 16.6

49.5  2 52.5  1 44  2 45.5  1 48  2 49  2 46  2 42.5  1

622  26 620  21 495  8 451  3 495  15 507  18 367  17 219  7

74.5  2 70.5  3 68.5  1 66.5  1 67  3 66.5  1 60.5  1 55  1

a TTI, time to ignition was defined as the time at which the HRR first increased to 25 kW/m2.

each sample are presented in Fig. 12a. Other fire parameters such as the time to ignition (TTI), the peak of heat release rate (PHRR), the total heat released (THR) and the residue were also measured and are listed in Table 5. For all composite systems, data show a decrease in the TTI, the PHRR and the THR compared to the neat polymer. The reduced TTI of composites may be attributed to three combined phenomena. First, it may be explained by the higher viscosity of the condensed phase at the initial step of the composite burning process, resulting in a decrease of the relative thermal conductivity. Indeed, lower convection in the melting nanocomposite decreases the apparent heat conductivity. Therefore, more heat is accumulated at the surface of the sample and the ignition temperature of the material is reached faster [50]. The decrease in the thermal stability observed for composites (see Fig. 10a) may also result in early release of volatile products that accelerate the flame ignition. Finally, it can be also due to a change in radiation absorption of incident radiation [51]. Indeed, compared to PP in which incident radiation tends to be absorbed inside of PP matrix perpendicularly to the sample surface, halloysite nanotubes may scatter isotropically incident radiation. Therefore, composite samples may tend to absorb and/or scatter incident radiation mainly near the sample surface, reducing the time required to reach the ignition temperature. Further investigations are required to evaluate the influence of these different effects on the reduced TTI of PP/HNTs composites and to determine whether one is predominant or if it is the superimposition of the three effects. The reduction in the THR with increasing the halloysite amount usually means that not all of the organic sample burns. However, as the residue corresponds to the initial inorganic weight percentage,

4293

it indicates that the polymer has combusted completely and that the clay barrier can only delay the volatilization of degradation products but not prevent it. Moreover, compared to other systems, the PHRR is strongly reduced by about 40% for PP-Plb-H-W8 nanocomposite. It corroborates the structural (SEM and TEM) investigation and indicates that the fire retardancy enhancement is directly related to the improvement of the clay dispersion with the best nanomorphology obtained combining PP-g-MA and water injection. Fig. 12b shows the HRR curves for PP-Plb-H-W nanocomposites as a function of the halloysite content. TTI, PHRR and THR are reduced with increasing the HNTs content in PP-Plb matrix and the residue is proportional to the filler loading (Table 5). Moreover, while the curve for neat PP-Plb is very sharp, corresponding to an intensive burning with quick heat release, a flat curve is obtained at 16 wt% HNTs characterized by a low and broad peak. The suggested mechanism of action of halloysite nanotubes involves their migration at the surface of the condensed phase and the formation of an inorganic residue playing a dual role. The residue insulates the material from external flame due to its low thermal conductivity and also acts as a mass transport barrier reducing the amount of volatile degradation products available for flame [52,53]. Compared to TGA experiments, the encapsulation process of the degradation products does not play any role in the combustion process since more volatilization products are produced in a short time in the cone calorimeter and are preferentially transfered to the surface of the condensed phase. Fig. 13 shows the residue aspect of virgin PP and nanocomposites after combustion. While pure PP is totally burnt (Fig. 13a), an inorganic residue is observed for nanocomposites. The residue is very compact and strong for PP-Plb-H8-W nanocomposite (Fig. 13c) while in the case of PP-H8 sample (Fig. 13b) it is more fragile and cracked. This result confirms that improved residue strength associated to homogeneous nanoclay dispersion enables to form an effective physical barrier for mass and energy transport with consequent reduced the flame intensity. Furthermore, as the HNTs content increases, the inorganic layer becomes tougher and does not break, resulting in the production of a smaller fire and thus an increased burning time (Fig. 12b). Similar results have already been reported by other researchers for PP nanocomposites based on modified HNTs [25]. It must be stressed that in this work PP nanocomposites with low flammability were prepared using pristine clay nanotubes, replacing the clay surface modification by a cheap and environmentally friendly water-assisted extrusion process.

Fig. 13. Optical photographs of (a) neat PP, (b) PP-H8 and (c) PP-Plb-H8-W nanocomposites after combustion in the cone calorimeter.

4294

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295

4. Conclusions In this work, PP nanocomposites based on unmodified halloysite nanotubes have been successfully prepared through a waterassisted extrusion process. Different mixing strategies have been tested using compatibilizer and water injection separately or simultaneously. The morphological characterization reveals that a uniform dispersion of HNTs in PP is only achieved when PP-g-MA and water-assisted extrusion are combined together. This unexpected result can be explained by the dual role of water during extrusion: (i) water remains liquid during the melt-blending step and creates an aqueous suspension of halloysite which promotes the clay dispersion; (b) water hydrolyzes anhydride functions of PPg-MA into carboxylic diacid functions and enables the formation of H-bonds with the external SiO surface of the nanotubes. Structureproperty relationships were also investigated and confirm the synergistic effect of compatibilizer and water injection. DMA and rheological measurements show further enhancement in the storage modulus and linear viscoelastic properties for PP-Plb-H-W nanocomposites. Thermal analyses carried out by TGA reveal that addition of HNTs into PP matrix has two opposing effects on the thermal stability of the polymer. However, the unfavourable catalytic action of the nanotubes is considerably reduced with better dispersion via the formation of a tortuous diffusion path combined to an encapsulation process of the initial degradation products inside the tubules. Finally, mass loss calorimetry shows that PP-PlbH-W nanocomposites exhibit the lowest flammability due to the formation of a rigid and stable inorganic residue acting as a potential barrier to heat and mass transport. Acknowledgments The authors are very grateful to the experimental support of Ir. P. Van Velthem for DMA and TGA experiments, Mrs. S. Bebelman for assistance on FTIR and Mrs. P. Lipnik for TEM support. References [1] Ray SS, Okamoto M. Polymer/layered silicate nanocomposites: a review from preparation to processing. Prog Polym Sci 2003;28(11):1539e641. [2] Alexandre M, Dubois P. Polymer-layered silicate nanocomposites: preparation, properties and uses of a new class of materials. Mater Sci Eng 2000;28(1, 2):1e63. [3] Usuki A, Hasegawa N, Kato M. Polymer-clay nanocomposites. Adv Polym Sci 2005;179:135e95. [4] Paul DR, Robeson LM. Polymer nanotechnology: nanocomposites. Polymer 2008;49(15):3187e204. [5] Kawasumi M, Hasegawa N, Kato M, Usuki A, Okada A. Preparation and mechanical properties of polypropylene-clay hybrids. Macromolecules 1997; 30(20):6333e8. [6] Gilman JW. Flammability and thermal stability studies of polymer layeredsilicate (clay) nanocomposites. Appl Clay Sci 1999;15(1, 2):31e49. [7] Vaia RA, Giannelis EP. Polymer melt intercalation in organically-modified layered silicates: model predictions and experiment. Macromolecules 1997; 30(25):8000e9. [8] Weon JI, Sue HJ. Effects of clay orientation and aspect ratio on mechanical behavior of nylon-6 nanocomposite. Polymer 2005;46(17):6325e34. [9] Pavlidou S, Papaspyrides CD. A review on polymer-layered silicate nanocomposites. Prog Polym Sci 2008;33:1119e98. [10] Vaia RA, Teukolsky RK, Giannelis EP. Interlayer structure and molecular environment of alkylammonium layered silicates. Chem Mater 1994;6(7): 1017e22. [11] Fornes TD, Hunter DL, Paul DR. Nylon-6 nanocomposites from alkylammonium-modified clay: the role of alkyl tails on exfoliation. Macromolecules 2004;37(5):1793e8. [12] Fornes TD, Yoon PJ, Paul DR. Polymer matrix degradation and color formation in melt processed nylon 6/clay nanocomposites. Polymer 2003;44(24): 7545e56. [13] Joussein E, Petit S, Churchman J, Theng B, Righi D, Delvaux B. Halloysite clay minerals e a review. Clay Miner 2005;40(4):383e426. [14] Singh B. Why does halloysite roll? A new model. Clays Clay Miner 1996;44(2): 191e6.

[15] Frost RL, Shurvell HF. Raman microprobe spectroscopy of halloysite. Clay Clay Miner 1997;45(1):68e72. [16] Du M, Guo B, Cai X, Jia Z, Liu M, Jia D. Morphology and properties of halloysite nanotubes reinforced polypropylene nanocomposites. e-Polymers 2008;130:1e14. [17] Liu M, Guo B, Du M, Lei Y, Jia D. Natural inorganic nanotubes reinforced epoxy resin nanocomposites. J Polym Res 2008;15(3):205e12. [18] Ismail H, Pasbakhsh P, Fauzi MNA, Bakar AA. Morphological, thermal and tensile properties of halloysite nanotubes filled ethylene propylene diene monomer (EPDM) nanocomposites. Polym Test 2008;27(7):841e50. [19] Handge UA, Hedicke-Höchstötter K, Altstädt V. Composites of polyamide 6 and silicate nanotubes of the mineral halloysite: influence of molecular weight on thermal, mechanical and rheological properties. Polymer 2010; 51(12):2690e9. [20] Guo B, Zou Q, Lei Y, Jia D. Structure and performance of polyamide 6/halloysite nanotubes nanocomposites. Polym J 2009;41(10):835e42. [21] Lecouvet B, Gutierrez JG, Sclavons M, Bailly C. Structureeproperty relationships in polyamide 12/halloysite nanotube nanocomposites. Polym Degrad Stab 2011;96(2):226e35. [22] Hedicke-Höchstötter K, Lim GT, Altstädt V. Novel polyamide nanocomposites based on silicate nanotubes of the mineral halloysite. Compos Sci Technol 2009;69(3, 4):330e4. [23] Ning NY, Yin QJ, Luo F, Zhang Q, Du R, Fu Q. Crystallization behavior and mechanical properties of polypropylene/halloysite composites. Polymer 2007; 48(25):7374e84. [24] Jia ZX, Luo YF, Guo BC, Yang BT, Du ML, Jia DM. Reinforcing and flameretardant effects of halloysite nanotubes on LLDPE. Polym-Plast Technol Eng 2009;48(6):607e13. [25] Du M, Guo B, Jia D. Thermal stability and flame retardant effects of halloysite nanotubes on poly(propylene). Eur Polym J 2006;42(6):1362e9. [26] Marney DCO, Russel LJ, Wu DY, Nguyen T, Cramm D, Rigopoulos N, et al. The suitability of halloysite nanotubes as a fire retardant for nylon 6. Polym Degrad Stab 2008;93(10):1971e8. [27] Liu X, Wu Q. PP/clay nanocomposites prepared by grafting-melt intercalation. Polymer 2001;42(25):10013e9. [28] Reichert P, Nitz H, Klinke S, Brandsch R, Thomann R, Mulhaupt R. Poly(propylene)/ organo-clay nanocomposite formation: influence of compatibilizer functionality and organo-clay modification. Macromol Mater Eng 2000;275:8e17. [29] Park JH, Lee HM, Chin I-J, Choi HJ, Kim HK, Kang WG. Intercalated polypropylene/clay nanocomposite and its physical characteristics. J Phys Chem Solids 2008;69(5e6):1375e8. [30] Xie W, Gao Z, Pan WP, Hunter D, Singh A, Vaia R. Thermal degradation chemistry of alkyl quaternary ammonium montmorillonite. Chem Mater 2001;13(9):2979e90. [31] Hedley CB, Yuan G, Theng BKG. Thermal analysis of montmorillonites modified with quaternary phosphonium and ammonium surfactants. Appl Clay Sci 2007;35(3, 4):180e8. [32] Shah RK, Paul DR. Organoclay degradation in melt processed polyethylene nanocomposites. Polymer 2006;47(11):4075e84. [33] Pasbakhsh P, Ismail H, Ahmad Fauzi MN, Abu Bakar A. Influence of maleic anhydride grafted ethylene propylene diene monomer (MAH-g-EPDM) on the properties of EPDM nanocomposites reinforced by halloysite nanotubes. Polym Test 2009;28(5):548e59. [34] Korbee RA, Van Geneen AA. Process for the preparation of a polyamide nanocomposite composition. WO; 1999:029767. [35] Hasegawa N, Okamoto H, Kato M, Usuki A, Sato N. Nylon 6/na-montmorillonite nanocomposites prepared by compounding nylon 6 with namontmorillonite slurry. Polymer 2003;44(10):2933e7. [36] Fedullo N, Sclavons M, Bailly C, Lefebvre J-M, Devaux J. Nanocomposites from untreated clay: a myth? Macromol Symp 2006;233(1):235e45. [37] Fedullo N, Sorlier E, Sclavons M, Bailly C, Lefebvre J-M, Devaux J. Polymerbased nanocomposites: overview, applications and perspectives. Prog Org Coat 2007;58(2, 3):87e95. [38] Mainil M, Urbanczyk L, Calberg C, Germain A, Jerome C, Bourbigot S, et al. Morphology and properties of san-clay nanocomposites prepared principally by water-assisted extrusion. Polym Eng Sci 2010;50(1):10e21. [39] Molajavadi V, Garmabi H. Water assisted exfoliation of PA6/clay nanocomposites using a twin-screw extruder: effect of water contact time. J Appl Polym Sci 2011;119(2):736e43. [40] Rousseaux DDJ, Sallem-Idrissi N, Baudouin AeC Devaux J, Godard P, Marchand-Brynaert J, Sclavons M. Water-assisted extrusion of polypropylene/clay nanocomposites: a comprehensive study. Polymer 2011;52(2):443e51. [41] Delobel R, Le Bras M, Ouassou N, Alistiqsa F. Thermal behaviours of ammonium polyphosphate-pentaerythritol and ammonium pyrophosphatepentaerythritol intumescent additives in polypropylene formulations. J Fire Sci 1990;8(2):85e108. [42] Shenoy AV. Rheology of filled polymer systems. Dordrecht: Kluwer Academic Publishers; 1999. [43] Cai G, Dasari A, Yu ZZ, Du X, Dai S, Mai YW, et al. Fire response of polyamide 6 with layered and fibrillar nanofillers. Polym Degrad Stab 2010;95(5):845e51. [44] Pötschke P, Abdel-Goad M, Alig I, Dudkin S, Lellinger D. Rheological and dielectrical characterization of melt mixed polycarbonate-multiwalled carbon nanotube composites. Polymer 2004;45(26):8863e70. [45] Hsieh AJ, Moy P, Beyer FL, Madison P, Napadensky E. Mechanical response and rheological properties of polycarbonate layered-silicate nanocomposites. Polym Eng Sci 2004;44(5):825e37.

B. Lecouvet et al. / Polymer 52 (2011) 4284e4295 [46] Rousseaux DDJ, Sclavons M, Godard P, Marchand-Brynaert J. Carboxylate clays: a model study for polypropylene/clay nanocomposites. Polym Degrad Stab 2010;95(7):1194e204. [47] Guo B, Zou Q, Lei Y, Jia D. Structure and performance of polyamide 6/halloysite nanotubes nanocomposites. Polym J 2009;41(10):835e42. [48] Du M, Guo B, Cai X, Jia Z, Liu M, Jia D. Morphology and properties of halloysite nanotubes reinforced polypropylene nanocomposites. e-Polymers 2008;130: 1e14. [49] Marcilla A, Gomez A, Menargues S, Ruiz R. Pyrolysis of polymers in the presence of a commercial clay. Polym Degrad Stab 2005;88(3): 456e60.

4295

[50] Bourbigot S, Samyn F, Turf T, Duquesne S. Nanomorphology and reaction to fire of polyurethane and polyamide nanocomposites containing flame retardants. Polym Degrad Stab 2010;95(3):320e6. [51] Kashiwagi T, Grulke E, Hilding J, Groth K, Harris R, Butler K, et al. Thermal and flammability properties of polypropylene/carbon nanotube nanocomposites. Polymer 2004;45:4227e39. [52] Zhu J, Morgan AB, Lamelas FJ, Wilkie CA. A model study for polypropylene/ clay nanocomposites. Chem Mater 2001;13:3774e80. [53] Zanetti M, Kashiwagi T, Falqui L, Camino G. Cone calorimeter combustion and gasification studies of polymer layered silicate nanocomposites. Chem Mater 2002;14(2):881e7.