New Nanocomposite Materials Reinforced with Cellulose Whiskers in

Yucheng Peng , Sergio A. Gallegos , Douglas J. Gardner , Yousoo Han .... Yuan Lu , Mario Calderón Cueva , Edgar Lara-Curzio , Soydan Ozcan ..... S. J...
0 downloads 0 Views 220KB Size
Biomacromolecules 2005, 6, 2732-2739

2732

New Nanocomposite Materials Reinforced with Cellulose Whiskers in Atactic Polypropylene: Effect of Surface and Dispersion Characteristics N. Ljungberg,† C. Bonini,† F. Bortolussi,‡ C. Boisson,‡ L. Heux,*,† and J.Y. Cavaille´ § Centre de Recherches sur les Macromole´ cules Ve´ ge´ tales (CERMAV-CNRS), Universite´ Joseph Fourier, BP 53, 38041 Grenoble, France, Laboratoire de Chimie et Proce´ de´ s de Polyme´ risation, CNRS/CPE, 69621Villeurbanne, France, and Groupe d’Etudes de Me´ tallurgie Physique et de Physique des Mate´ riaux, UMR CNRS 5510, INSA, 69621 Villeurbanne, France Received March 24, 2005; Revised Manuscript Received April 29, 2005

New nanocomposite films were prepared with atactic polypropylene as the matrix and either of three types of cellulose whiskers, with various surface and dispersion characteristics, as the reinforcing phase: aggregated without surface modification, aggregated and grafted with maleated polypropylene or individualized and finely dispersed with a surfactant. Films obtained by solvent casting from toluene were investigated by means of scanning electron microscopy, dynamic mechanical analysis, and tensile testing. In the linear region, the mechanical properties above the glass-rubber transition were found to be drastically enhanced for the nanocomposites as compared to the neat polypropylene matrix. These effects were ascribed to the formation of a rigid network with filler/filler interactions. In addition, interactions between the filler and the matrix as well as the dispersion quality were found to play a major role on the mechanical properties of the composites when investigation of the films was performed in the nonlinear region. Introduction It is well-known that the macroscopic properties of a composite material reinforced with nanometric fillers are determined by various factors, such as its composition, the characteristics of each component, the geometry of the filler, the filler dispersion, the filler/filler and filler/matrix interactions and in some cases, the modification of the characteristics of the matrix itself.1 These parameters have an important influence on the final properties of the nanocomposite and are strongly interconnected, and this interconnectivity makes it problematic to draw conclusions on relative influences of the individual parameters. Many different nanofillers have been studied, including silica, clay, carbon black, and carbon nanotubes, and as the latter, cellulose is of particular interest because of its rodlike shape. Nanometric monocrystals of cellulose, commonly referred to as whiskers, can be obtained from various sources including an animal cellulose called tunicin.2 These tunicin whiskers consist of slender parallelepiped rods with typical dimensions of 1520 nm width and 1 µm length. They thus exhibit a high aspect ratio and an important surface area. Tunicin whiskers have been employed as fillers in several kinds of polymeric matrixes, including poly(styrene-co-butyl acrylate) latex,3-5 poly(vinyl chloride),6 poly(hydroxyl alkanoate),7-9 starch,10 poly(oxyethylene),11 cross-linked unsaturated polyethers,12 and LiClO4-doped ethylene oxide* Corresponding author. Tel.: +33 4 76 03 76 08. Fax: +33 4 76 54 72 03. E-mail: [email protected]. † Universite ´ Joseph Fourier. ‡ CNRS/CPE. § UMR CNRS.

epichlorohydrin copolymers.13 The first studies on nanocomposites consisting of an amorphous latex filled with tunicin whiskers displayed that the mechanical properties of the latex in the rubbery state were drastically enhanced when reinforced with very low percentages of cellulose whiskers (from 1 to 12 wt %).3,4 This reinforcement was attributed to the formation of a percolating whisker network supported by hydrogen bonds. The existence of the network was confirmed by electrical measurements performed on nanocomposites containing whiskers coated with conductive polypyrrole.14 The formation of such a percolating network requires a statistical distribution of the whiskers and hence well controlled properties. The mechanical properties thus obtained have been extensively described for different systems in a recent review.15 So far, stable suspensions of cellulose whiskers have only been obtained in aqueous suspension or in polar solvents such as DMF. Because of the incompatibility with the hydrophilic whiskers, it has until now been impossible to efficiently add cellulose whiskers to most of the classical apolar polymer matrixes. To do so, the surface characteristics of the whiskers have to be changed. One possible route is the chemical grafting of the surface,16 but this type of chemistry remains difficult on anisometric particles. In the present study, we have taken advantage of a recent process using surfactants aimed at dispersing cellulose whiskers in apolar solvents17 to investigate the possibility of obtaining nanocomposites made of an apolar matrix and cellulose whiskers. Polypropylene was selected as a representative apolar matrix, as numerous studies have been devoted to its

10.1021/bm050222v CCC: $30.25 © 2005 American Chemical Society Published on Web 07/06/2005

New Nanocomposite Materials

reinforcement with nanofillers, both with and without surface modifications.18-22 In the present work, amorphous atactic polypropylene was chosen as it would be independent of the probable changes in the crystallinity often encountered in polypropylene nanocomposites. The aim of the present article was to identify the effects of the filler characteristics on the overall properties of the final nanocomposite. The dispersion quality and surface activity of the whiskers were varied; either by maintaining the original whisker surfaces without modification, by grafting maleated polypropylene to the whisker surfaces23 or by dispersing them with a surfactant according to a recently developed process.17 Dynamic mechanical measurements were performed to estimate the viscoelastic behavior of the materials as a function of temperature and the experimental results were compared with models commonly used for composites. The nanocomposites were also investigated by tensile testing. Experimental Section Synthesis of High Molecular Weight Atactic Polypropylene. High molecular weight atactic polypropylene (aPP) is not commercially available and was thus synthesized by polymerization with metallocene catalysis. The metallocene complex Me2-Si(9-Flu)2ZrCl2 was prepared under an argon atmosphere by the standard Schlenk technique according to the literature.24 Polymerizations were performed in a 0.5-L glass reactor, equipped with a stainless steel blade. A solution of cocatalyst (30 wt % methylalumoxane in toluene) in 250 mL of toluene was mixed with Me2-Si(9-Flu)ZrCl2 and transferred into the reactor under a stream of argon. Argon was then pumped out, and the reactor was charged with a mixture of propylene/ethylene (1% ethylene) and heated to 50 °C. The pressure was kept constant at 4 bar during the entire polymerization. The activities were calculated by measuring the consumption rate of the propylene/ethylene gas. The polymerization was terminated by venting off ethylene and adding acidic methanol. The prepared polymer was washed in 300 mL of methanol, filtered, and dried in a vacuum at 100 °C. The number average molecular weight, characterized by size exclusion chromatography in toluene using polystyrene standards, was found to be 12 500 g/mol and the weight average molecular weight 90 300 g/mol. A very small degree of crystallinity (approximately 1%) was determined by differential scanning calorimetry. Cellulose Whiskers from Tunicin. Cellulose monocrystals, commonly called whiskers, were chosen as the filler material. One of the best sources to obtain regular whiskers with large aspect ratios is from tunicin; an animal cellulose where there is no need to separate cellulose from hemicellulose and lignin as in wood. The whiskers were extracted from the sea animal tunicate and obtained as an aqueous suspension after an acid hydrolysis treatment.3,4 The suspension did not sediment or flocculate as a result of surface acid groups created during the hydrolysis,2 and when concentrated, it displayed typical liquid crystal characteristics.2,25 The whiskers were 10-20 nm in diameter with lengths ranging from 1 to several micrometers and thus had large aspect ratios of about 67 (estimated from transmission

Biomacromolecules, Vol. 6, No. 5, 2005 2733

electron microscopy) and significant interfacial areas of approximately 150 m2/g. Three types of whiskers with various surface qualities were prepared in order to probe the effect of the surface and dispersion characteristics. The first type was aggregated cellulose whiskers in toluene. The aggregated whiskers were obtained from individualized whiskers in water, which were freeze-dried and then redispersed in toluene with a mechanical mixing apparatus (T25 basic Ultra-Turrax) operating at 15 000 rpm for about 5 min. These whiskers had no surface modification, and the resulting dispersion was not stable and was found to flocculate at rest. These aggregated whiskers were denoted AGWH. The second type was whiskers grafted with maleated polypropylene (PPgMA) according to an experimental process described elsewhere.23 Epolene E43, a PPgMA with a number-average molecular weight of 3900 g/mol and a weight-average molecular weight of 9100 g/mol, was kindly supplied by Eastman. The percentage of maleic anhydride was found to be 7.5 wt % by acidic titration. The percentage of PPgMA grafted on the cellulose was determined by solidstate NMR to be 13.6 × 10-5 mol for each gram of cellulose. Owing to the large surface area of the whiskers, the percentage of grafting was 100 times higher than what was observed for macroscopic cellulose fibers.26 The grafted whiskers (GRWH) were dispersed in toluene by mechanical mixing with the Ultra-Turrax, and as for the aggregated whiskers, the dispersion flocculated at rest. The final type was a suspension in a nonpolar solvent. The suspension of cellulose in water was mixed with a phosphoric ester of polyoxyethylene(9) nonylphenyl ether (BNA, commercialized by Ceca ATO Co.) at a weight ratio of 4:1 of BNA to cellulose. After adjustment of the pH to 8 with aqueous sodium hydroxide, the suspension was freezedried and then redispersed in toluene.17 The excess surfactant was removed by centrifugation, and the resulting pellet could easily be redispersed in a desired amount of toluene by an ultrasonic treatment in a Branson Sonifier for a few tens of seconds. The amount of surfactant adsorbed on the cellulose was found to be 1.6 times the cellulose weight. The final suspensions of the surface coated whiskers (SUWH) in toluene did not precipitate nor flocculate at rest. Film Preparation. Films of the neat fillers were prepared by solvent casting. A neat whiskers film was cast from water and films of AGWH, GRWH, and SUWH were cast at 110 °C from toluene. All four films were pressed at 7 MPa during 5 min and heated during 18 h under vacuum at 90 °C to ensure total solvent evaporation. Nanocomposites were prepared by mixing solubilized atactic polypropylene in hot toluene (110 °C) with one of the three kinds of fillers (AGWH, GRWH, or SUWH) dispersed in toluene. The percolation threshold for tunicin whiskers in a polar matrix had previously been determined to 1 vol % ()1.5 wt %)3,4 and the weight fractions of whiskers in the present study were chosen above this value. Thus, film materials reinforced with 6 wt % of each of the three whisker types were obtained and in addition, a nanocomposite film containing 3 wt % SUWH was also prepared. The solvent was evaporated overnight at 110 °C

2734

Biomacromolecules, Vol. 6, No. 5, 2005

to avoid PP precipitation, and an additional 6h under vacuum at 110 °C was performed to ensure the total solvent evaporation. The aPP-composite films were then pressed at 150 °C during 20 min under a pressure of 7 MPa. Tensile Testing. Experiments were performed on an INSTRON 4300 tensile tester with a cross head speed of 1 mm/min at room temperature, 25 °C. The samples dimensions were 20 mm × 6 mm × 1 mm. The initial strain rate was (d/dt)ini) 8 × 10-4 s-1 for the films of neat fillers and 1.6 × 10-3 s-1 for the composites films. The force (F) was recorded as a function of the sample elongation (L - L0) and the stress, σ, and strain, , of the material could be calculated as σ)

F A

and )

L - L0 L0

respectively, where A was the cross-sectional area of the sample and L0 was the initial sample length. Scanning Electron Microscopy (SEM). The dispersion of whiskers in the composite films was characterized by SEM. The observations were performed with a JEOL SM 6100 microscope on films that had been fractured in liquid nitrogen and coated with gold/palladium on a JEOL-110E ion sputter coater. Dynamic Mechanical Analysis (DMA). Dynamic mechanical analysis (DMA) was performed on an RSAII from Rheometrics. The experiments were conducted in tensile mode under isochronal conditions at a frequency of 1 Hz. The value for the strain magnitude was set at 0.05% in order to be in the domain of the linear viscoelasticity of the materials. Curves displaying the storage (E′) modulus were recorded as a function of temperature between 180 and 310 K at a heating rate of 3 K/min. The shape of the film samples was rectangular, approximately 20 mm × 6 mm × 1 mm. Theoretical Predictions. Two mechanical approaches can be used when considering short fiber composites. The first one takes into account the reinforcement effect due to the presence of rigid inclusions in a soft matrix, without considering either filler/filler or filler/matrix interactions. This mean-field approximation is well illustrated by the HalpinKardos model,27 which correctly predicts a composite behavior containing fibers randomly oriented. It assumes that such materials are equivalent to many layers of unidirectional plies oriented at various angles to give a quasi-isotropic composite. The mechanical properties of each ply are derived from the micro-mechanic equations of Halpin-Tsai.28 One of the main interests of this approach is that it takes into account the anisotropy of the mechanical properties of the particle. The second approach is based on the consideration of filler/ filler interactions and has led to the percolation concept, well adapted to describing a percolating network of cellulose whiskers.3,4 Percolation is a statistical theory, which can be applied to any system involving a great number of species

Ljungberg et al.

likely to be connected. In this approach, the whiskers are believed to be linked by strong hydrogen bonds. These interactions induce the mechanical percolation of the fibers and the mechanical properties of the films can be predicted following the method of Ouali et al.29 The concept of percolation can be adapted to the classical series-parallel model of Takayanagi et al.,30 where the tensile modulus of the composite, E, is given by the equation E)

(1 - 2ψ + ψXr)EsEr + (1 - Xr)ψEr2 (1 - Xr)Er + (Xr - ψ)Es

The subscript s and r refer to the soft and rigid phases, respectively, and Xr is the volume fraction of whiskers. In the Takayanagi et al model,30 ψ is an adjustable parameter, but when following the model of Ouali et al., ψ corresponds to the volume fraction of the percolating rigid phase. It can be calculated with a prediction based on the percolation concept29 ψ ) 0 for Xr e Xc ψ ) Xr

(Xr - Xc)b (1 - Xc)b

for Xr > Xc

where b is the percolation exponent predicted to be 0.4 in a three-dimensional system29 and Xc is the critical volume fraction needed to reach the geometrical percolation of the whiskers. This percolation threshold depends on the aspect ratio and the orientational distribution of the whiskers. In the present study, Xc ) 1 vol % that corresponds to 1.5 wt %.3,4 When Xr > Xc, an infinite rigid cluster appears resulting in a substantial change in the mechanical properties. The experimental results have also been compared to the basic calculation of the Voigt superior boundary. This model assumes a parallel association of the two phases and thereby a constant deformation for an applied stress. Thus, the modulus of the composite is given by E ) (1 - Xr)Es + XrEr For both the Voigt and the percolation models, the relevant parameter is the cohesion of the fiber network and both models take into account the modulus of a film of neat fillers, i.e., Er. Results and Discussion Characterization of Films of Neat Fillers. To understand the mechanical behavior of the nanocomposites, films of the neat fillers needed to be characterized. Indeed, the filler/ filler interaction is an important factor, which influences the properties of the composite. In addition, the filler characteristics, particularly the Young modulus of the films of the neat fillers, are required for the theoretical approach based on the percolation and the Voigt models. The two films obtained from the suspensions in which whiskers were individualized in water (neat whiskers) or in toluene (SUWH) were transparent. This suggests that the

Biomacromolecules, Vol. 6, No. 5, 2005 2735

New Nanocomposite Materials Table 1. Density, Solids Fraction, and Young Modulus for Films of Neat Whiskers, AGWH, GRWH, and SUWH films of neat whiskers

density (g/cm3)

solids fraction (wt %)

Young modulus (GPa)

neat whiskers AGWH GRWH SUWH

1.20 0.97 0.86 1.50

76 61 54 95

5 1.3 0.5 1.4

whiskers remained uniformly dispersed in the final film. The other two films, on the other hand, obtained from the dispersions of AGWH and GRWH in toluene were opaque, similar to a paper sheet. This opacity suggests the presence of aggregates. The apparent density and the solids fraction were evaluated by weight-dimension relationships for each of the four whisker films and the data can be seen in Table 1. Film casting of whiskers from the nonflocculating suspensions (neat whiskers and SUWH) resulted in films that were more compact, i.e., showing higher density values, than when obtained from aggregated and grafted cellulose (AGWH and GRWH). These observations were in agreement with the idea that compacting a pile of stick-like objects is easier than compacting a pile of aggregates. The AGWH film was found to have a slightly higher density than the film of the grafted whiskers (GRWH). Thus, the presence of PPgMA-grafts on the whisker surface contributed to making the aggregate network even less compact. Figure 1 presents the stress vs strain curves from tensile testing experiments for the four neat filler films. The stress at break was much higher for films of SUWH and neat whiskers as compared to AGWH and GRWH films. This indicates that the cohesion between aggregates was weaker than the cohesion between percolating individualized whiskers. For similar reasons, the strain at break was found higher for the films obtained from the individualized whiskers. Nevertheless, by incorporating PPgMA-grafts on the cellulose, the elongation at break could be increased by a factor 2.5 for GRWH as compared to AGWH. The very low stress and strain at break obtained for the latter (i.e., AGWH) was likely to be due to the fragility of the compacted bare aggregates. The Young moduli of the four films, obtained from the slope of the stress vs strain curves, are also given in Table 1. Two factors seemed to be predominant on the films’ cohesion: the cellulose surface and the aggregation of whiskers. In the case of the neat whisker film, the hydrogen bonds between the cellulose whiskers ensured a very good mechanical strength. When the cellulose surface was modified with surfactant (SUWH), the mechanical strength was conserved, but the links between the whiskers became weaker resulting in a decrease of the Young modulus by a factor 4. For the same reasons, the Young modulus was decreased by a factor 3 between AGWH and GRWH. The higher modulus found for AGWH was likely due to the establishment of stronger filler-filler interactions mediated by hydrogen bonds. In the case of SUWH and neat whiskers, the cohesion of the films was ensured by a percolating network of whiskers, whereas in the case of AGWH and

Figure 1. Stress vs strain curves for films of neat whiskers.

GRWH, the cohesion arose from the interactions between aggregates and was consequently weaker. Characterization of Dispersion Quality in the Nanocomposites. In addition, the nanocomposite films reinforced with SUWH were practically transparent, whereas the films reinforced with AGWH and GRWH were opaque, most likely because of the presence of aggregates of micrometric sizes and/or voids. In any case, this suggests that the quality of the whisker dispersion in the nanocomposite film was greatly improved by the presence of surfactant on the cellulose surface. SEM analysis was performed in an attempt to visualize the dispersion quality in the composite films. The three film types were fractured in liquid nitrogen and Figure 2 shows the fracture surfaces of (a) aPP + 6 wt % SUWH, (b) aPP + 6 wt % AGWH, and (c) aPP + 6 wt % GRWH as observed by SEM. In the case of the film containing the surfactant-modified whiskers (SUWH), it was not possible to see individualized whiskers inside the matrix (Figure 2a), because of their very small size. However, some of them appeared as white points at the surface of the sample, thus corresponding to whiskers in the perpendicular plane of the film. No aggregates were visible at the experimental sensibility in this case. The quality of the whisker dispersion was essentially the same between films of aPP reinforced with AGWH and GRWH (Figure 2, panels b and c). In both cases, the whiskers were assembled as aggregates, with a thickness of 0.5-1 µm, which seemed to percolate throughout the composite. Thus, despite the surface modification obtained by the PPgMA grafts, which should induce better compatibility with aPP than AGWH, the dispersion quality was not improved for the GRWH-composite as compared to the AGWH-one. Dynamic Mechanical Analysis (DMA) and Comparison with Theoretical Predictions. Nanocomposites with 3 and 6 wt % SUWH were analyzed by DMA. Figure 3 gives the evolution of the storage modulus, E′, at 1 Hz versus temperature for the reinforced material as well as for neat aPP. The unfilled matrix displayed the typical behavior of an amorphous polymer material. For temperatures below Tg the polymer was in the glassy state and the modulus slightly decreased with temperature but remained roughly constant, around 3 GPa. A significant drop of the tensile modulus by more than three decades, corresponding to the glass-rubber transition, was then observed. Subsequently, the elastic

2736

Biomacromolecules, Vol. 6, No. 5, 2005

Ljungberg et al.

Figure 3. Temperature dependence of storage modulus curves from DMA runs comparing neat aPP and composites reinforced with 3 and 6 wt % SUWH.

Figure 4. Temperature dependence of storage modulus curves from DMA runs comparing experimental data for neat aPP and aPP reinforced with 6 wt % SUWH with predicted data from the Voigt, percolation and Halpin-Kardos models. Figure 2. Cryofracture of aPP composite films reinforced with (a) 6 wt % SUWH, (b) 6 wt % AGWH, and (c) 6 wt % GRWH.

modulus then became lower and lower with temperature until the experimental equipment failed to measure it as a consequence of irreversible chain flow. Below Tg, the moduli of the nanocomposite materials were roughly equivalent to the neat aPP matrix. In this temperature range, the difference between the elastic tensile modulus of the fibers and that of the aPP matrix was not large enough to distinguish a reinforcement effect with only 3 or 6 wt % of fillers. Above Tg, on the other hand, there was a significant mechanical reinforcement that increased with the percentage of whiskers (Figure 3). The drop in storage modulus was dramatically reduced as compared to neat aPP; for instance, the modulus of the composite containing 6 wt % whiskers presented a modulus 50-fold that of the neat matrix. This augmentation in storage modulus at high temperatures may arise from filler/filler interactions and the density of a cellulosic network increasing with the percentage of the filler.

However, contrary to what was observed for nanocomposites with neat whiskers,3 a noticeable negative slope was observed with increasing temperature above Tg for the nanocomposite containing the surfactant modified fillers. Indeed, the filler-filler interactions became weaker when the hydrogen bonds were prevented by the presence of the surfactant. This has also been observed in the case of amorphous latex and weakly interacting whiskers such as chitin extracted from squid pen31 or from Riftia tubes,32 in composites materials of cellulose acetate butyrate filled with native or surface-trimethylsilated cellulose nanocrystals33 or in nanocomposites with low whiskers contents.3,4 Figure 4 gives the experimental tensile modulus curves for neat aPP and the nanocomposite containing 6 wt % SUWH, and compares them with theoretical predictions. The Halpin-Kardos mean field model, which considers that the reinforcement was due to the presence of rigid inclusions in a soft matrix without filler/filler interactions, was calculated

New Nanocomposite Materials

for 6 wt % of SUWH and can be seen in Figure 4. Details of the calculations are reported elsewhere.34 The fiber characteristics were the following: the aspect ratio, L/d ) 67, the Poisson’s ratio, νf ) 0.3 (cellulose being crystalline throughout the temperature range), the stiffness in fiber direction, E11f ) 150 GPa, the stiffness perpendicular to the fiber direction, E22f ) 15 GPa, and the in-plane shear modulus, Gf ) 5 GPa. The percolation and the Voigt models were also calculated for 6 wt % of whiskers by taking into account that the modulus of a SUWH sheet was 1.4 GPa as determined in the previous section. Detailed calculations can be found elsewhere.3,4 In the glassy state, the reinforcement effect was very low as predicted by all of the models. However, the HalpinKardos model strongly underestimated the modulus of the composite above Tg. This observation indicates that the glassy modulus of SUWH-filled aPP was well described by a meanfield approach, but at higher temperatures, strong interactions between the whiskers and a percolation effect needed to be taken into account. The same kind of behavior is classical for these kinds of nanocomposites and has already been observed for nanocomposites of latex reinforced with cellulose whiskers. However, it is clear from Figure 4 that the percolation model underestimated the experimental modulus of the SUWH nanocomposite. This confirmed the presence of a rigid percolating network of interacting particles, and that just above the Tg, the predominant parameter was the filler network cohesion. Thus, the discrepancy could be due either to a whisker distribution that may not be entirely isotropic3,4 or to an extra contribution from the filler/matrix interaction. Also the Voigt boundary model (Figure 4), despite the fact that it naturally overestimated the experimental modulus, can be seen as a confirmation that the filler/filler interactions were a determining factor in the mechanical properties of the nanocomposites. As the temperature increased, and the filler/matrix interaction weakened, the modulus tended to reach the percolation limits. Dynamic mechanical analysis was also performed on the polypropylene matrix filled with aggregated whiskers (AGWH) and whiskers grafted with PPgMA (GRWH). Figure 5 compares the evolution of the storage modulus versus temperature for films of aPP reinforced with 6 wt % of AGWH and 6 wt % of GRWH. The curves portraying neat aPP and 6 wt % SUWH are recalled for comparison. As a consequence of the differences in surface modification for the three filler types, they should give rise to various interfaces with the matrix and different filler/filler interactions. At temperatures below the Tg of aPP, there was no noticeable increase in the modulus by the incorporation of the fillers. Above the glass transition temperature, on the other hand, the mechanical reinforcement was significant. The composites containing AGWH and GRWH displayed mechanical behavior very similar to that already observed for the SUWH composite, i.e., the filled films presented storage moduli with values approximately 50 times higher than the neat matrix. The resemblance of the rubbery moduli of the composites suggests that, also in the case of AGWH

Biomacromolecules, Vol. 6, No. 5, 2005 2737

Figure 5. Temperature dependence of storage modulus curves from DMA runs comparing neat aPP and composites reinforced with 6 wt % AGWH, SUWH, and GRWH.

Figure 6. Stress vs strain curves for films of neat aPP and composites reinforced with 6 wt % AGWH, SUWH, and GRWH.

and GRWH, a percolation of the aggregates throughout the sample is possible. The use of models aimed to describe the behavior of isolated rods is useless in these last cases. However, these two nanocomposites are also near their Voigt boundary. Thus, just above the glass-rubber transition temperature, the predominant parameter is the filler network cohesion whatever the dispersion may be. It is worth noting that the lower modulus levels obtained for the grafted and surfactant modified systems are related to the lower cohesion between the fillers. Tensile Testing. The nonlinear mechanical behavior of films of the neat matrix as well as the composites reinforced with 6 wt % of the three filler types was investigated by tensile testing. Figure 6 shows the stress-strain curves of the four film materials. The tensile modulus, E, tensile strength, σB, and elongation at break, B, were determined from the curves and the results are presented in Table 2. The tensile modulus was measured from the slope of the low-strain region. It can be observed that the unfilled matrix displayed a nonlinear elastic behavior. As a result of its amorphous nature, the tensile strength for neat aPP, was found to be extremely low. By incorporating 6 wt % of one of the fillers, both the tensile modulus and the tensile strength were

2738

Biomacromolecules, Vol. 6, No. 5, 2005

Ljungberg et al.

Table 2. Tensile Modulus, E, Tensile Strength, σB, and Elongation at Break, B, for Films of Neat APP and Composites Reinforced with 6 wt % AGWH, GRWH, and SUWH sample

E (MPa)

σB (MPa)

B

neat aPP 6 wt % AGWH 6 wt % GRWH 6 wt % SUWH

0.4 18.5 6.1 8.9

0.026 0.58 0.32 0.34

0.49 0.07 0.20 0.70

drastically changed. All three composite films displayed significant increases in the tensile strength as compared to the neat matrix, with the largest increase being obtained for the film with aggregated fillers (AGWH). However, the much lower B-value of this material indicated that aPP lost almost all of its flexibility when reinforced with AGWH. The composite with 6 wt % GRWH also showed an increase in the tensile strength, though less than for the AGWH composite. On the other hand, its elongation at break was larger, even though it did not reach the initial value of the neat matrix. The SUWH composite not only displayed a large increase in tensile strength as compared to the neat aPP, but it also maintained, and even surpassed, the elongation at break of the neat matrix. The results from the tensile tests indicated large differences in the mechanical behavior of the three composite materials. When comparing with the DMA results it could be observed that the tensile modulus (Table 2) displayed the same trend as the storage modulus from the DMA curves (Figure 5), i.e., the AGWH composite showed the highest modulus, followed by the SUWH composite that had a modulus only slightly higher than the GRWH composite. However, the considerable differences in ductility could not be deduced from the DMA data. This could be explained by the fact that the dynamic mechanical measurements involved weak stresses, thus allowing the interactions between percolating cellulose whiskers, whether individualized or aggregated, to be maintained throughout the experiments. During tensile testing, higher stresses were involved, and the filler/filler interactions were now at least partially destroyed. Subsequently the filler/matrix interactions and dispersion quality grew in importance. As a consequence of the aggregates in the AGWH and GRWH composites, zones with accentuated fragility were created, which led to a higher brittleness in these materials. The GRWH composite was less fragile than the AGWH one as a result of the PPgMA grafts, enabling better filler/matrix compatibility. The better dispersion obtained for the surfactant-modified whisker composite was responsible for the considerable improvement in the mechanical properties of atactic polypropylene. It thereby appears that the mechanical properties of apolar polymer matrixes can be enhanced by incorporating cellulose whiskers modified by surfactant. Conclusions Nanocomposite films including cellulose whiskers with a high dispersion level were for the first time obtained in an apolar matrix by incorporating surfactant modified whiskers (SUWH) in atactic polypropylene (aPP). Two other types

of films were prepared with aggregated whiskers without surface modification (AGWH) or aggregated whiskers grafted with maleated polypropylene (GRWH). The nanocomposite films reinforced with AGWH and GRWH were found to be opaque whereas the SUWH film was transparent. These observations were in accordance with scanning electron microscopy results, indicating that AGWH and GRWH formed aggregates while SUWH was finely dispersed in the aPP matrix. The mechanical behavior of the nanocomposite films was evaluated in the linear range by dynamic mechanical analysis and compared to theoretical models. It was found that the rubbery modulus was drastically enhanced with the incorporation of whiskers, and this resulted from filler/filler interactions. However, the global reinforcement effect was not dependent on the dispersion quality. Mechanical behavior in the nonlinear range, as investigated by tensile testing, portrayed drastic increases in the tensile strengths of the nanocomposite films as compared to the neat aPP. In respect to the ductility and elongation at break of the materials, it was found that when aPP was reinforced with either of the two aggregate-forming whisker types, i.e., AGWH or GRWH, most of its ductility was lost. With the incorporation of SUWH, on the other hand, the elongation at break could be maintained. It thus became evident that the mechanical properties of the composites were dependent not only on the filler/filler interactions, but also on the quality of the dispersion. This new method of obtaining nanocomposites with surfactant modified whiskers is thus a promising tool for widening the use of this type of nanometric filler to most of the common apolar matrixes. The weakening of the interaction between the whiskers due to the presence of the surfactant also permits the brittleness usually encountered in nanocomposites made with bare cellulose whiskers to be avoided. Acknowledgment. Elf-Atochem is gratefully acknowledged for financial support. Many thanks are also expressed to Danielle Dupeyre for performing the SEM analyses and interpretations. References and Notes (1) Pukanszky, B.; Fekete, E. In Mineral Fillers in Thermoplastics I; Springer: New York, 1999; Vol. 139, p 109. (2) Marchessault, R. H.; Morehead, F. F.; Walter, N. M. Nature 1959, 184, 632. (3) Favier, V.; Chanzy, H.; Cavaille, J. Y. Macromolecules 1995, 28, 6365. (4) Favier, V.; Canova, G. R.; Cavaille, J. Y.; Chanzy, H.; Dufresne, A.; Gauthier, C. Polym. AdV. Technol. 1995, 6, 351. (5) Favier, V.; Canova, G. R.; Shrivastava, S. C.; Cavaille, J. Y. Polym. Eng. Sci. 1997, 37, 1732. (6) Chazeau, L.; Cavaille, J. Y.; Canova, G.; Dendievel, R.; Boutherin, B. J. Appl. Polym. Sci. 1999, 71, 1797. (7) Dufresne, A.; Kellerhals, M. B.; Wittold, B. Macromolecules 1999, 32, 7396. (8) Dubief, D.; Samain, E.; Dufresne, A. Macromolecules 1999, 32, 5765. (9) Dufresne, A. Compos. Interfaces 2000, 7, 53. (10) Angles, M. N.; Dufresne, A. Macromolecules 2001, 34, 2921. (11) Azizi, S. M. A. S.; Alloin, F.; Sanchez, J.-Y.; Dufresne, A. Polymer 2004, 45, 4149. (12) Azizi, S. M. A. S.; Alloin, F.; Sanchez, J.-Y.; Dufresne, A. Macromolecules 2004, 37, 4839. (13) Schroers, M.; Kokil, A.; Weder, C. J. Appl. Polym. Sci. 2004, 93, 2883. (14) Flandin, L.; Bidan, G.; Brechet, Y.; Cavaille, J. Y. Polym. Compos. 2000, 21, 165.

New Nanocomposite Materials (15) Azizi, S. M. A. S.; Alloin, F.; Dufresne, A. Biomacromolecules 2005, 6, 612. (16) Gousse´, C.; Chanzy, H.; Excoffier, G.; Soubeyrand, L.; Fleury, E. Polymer 2002, 43, 2645. (17) Heux, L.; Chauve, G.; Bonini, C. Langmuir 2000, 16, 8210. (18) Liang, J.-Z.; Li, R. K. Y. Polymer 1999, 40, 3191. (19) Wah, C. A.; Choong, L. Y.; Neon, G. S. Euro. Polym. J. 2000, 36, 789. (20) Maiti, S. N.; Mahapatro, P. K. J. Appl. Polym. Sci. 1991, 42, 3101. (21) Stricker, F.; Bruch, M.; Mulhaupt, R. Polymer 1997, 38, 5347. (22) Kawasumi, M.; Hasegawa, N.; Kato, M.; Usuki, A.; Okada, A. Macromolecules 1997, 30, 6333. (23) Felix, J. M.; Gatenholm, P. J. Appl. Polym. Sci. 1991, 42, 609. (24) Resconi, L.; Jones, R. L.; Rheingold, A. L.; Yap, G. P. A. Organometallics 1996, 15, 998. (25) Revol, J. F.; Bradford, H.; Giasson, J.; Marchessault, R. H.; Gray, D. G. Int. J. Biol. Macromol. 1992, 14, 170.

Biomacromolecules, Vol. 6, No. 5, 2005 2739 (26) Gauthier, R.; Joly, C.; Coupas, A. C.; Gauthier, H.; Escoubes, M. Polym. Compos. 1998, 19, 287. (27) Halpin, J. C.; Kardos, J. L. J. Appl. Polym. Sci. 1972, 43, 2235. (28) Halpin, J. C.; Tsai, S. W. AFML-FR, Ed., 1969; p 67. (29) Ouali, N.; Cavaille, J. Y.; Perez, J. Plast. Rubber. Compos. Proc. Appl. 1991, 16, 55. (30) Takayanagi, M.; Mianami, S.; Uemura, S. J. Polym. Sci. Polym. Symp. 1964, 5, 113. (31) Paillet, M.; Dufresne, A. Macromolecules 2001, 34, 6527. (32) Morin, A.; Dufresne, A. Macromolecules 2002, 35, 2190. (33) Grunert, M.; Winter, W. T. In 223rd ACS National Meeting; American Chemical Society: Orlando, FL, 2002. (34) Helbert, W.; Cavaille, J. Y.; Dufresne, A. Polym. Compos. 1996, 17, 604.

BM050222V