Highly Filled Bionanocomposites from Functionalized Polysaccharide

May 30, 2008 - Ayodele FatonaRichard M. BerryMichael A. BrookJose M. Moran-Mirabal. Chemistry of Materials 2018 30 (7), 2424-2435. Abstract | Full Tex...
0 downloads 0 Views 200KB Size
1974

Biomacromolecules 2008, 9, 1974–1980

Highly Filled Bionanocomposites from Functionalized Polysaccharide Nanocrystals Youssef Habibi and Alain Dufresne* Ecole Franc¸aise de Papeterie et des Industries Graphiques, Institut National Polytechnique de Grenoble (EFPG-INPG) BP65, 38402 Saint Martin d’He`res Cedex, France Received March 4, 2008; Revised Manuscript Received April 14, 2008

Cellulose and starch nanocrystals obtained from the acid hydrolysis of ramie fibers and waxy maize starch granules, respectively, were subjected to isocyanate-mediated reaction to graft polycaprolactone (PCL) chains with various molecular weights on their surface. Grafted nanoparticles were characterized by X-ray diffraction analysis and contact angle measurements. We observed that the nanoparticles kept their initial morphological integrity and native crystallinity. Nanocomposite films were processed from both unmodified and PCL-grafted nanoparticles and PCL as matrix using a casting/evaporation technique. We showed that mechanical properties of resulting films were notably different. Compared to unmodified nanoparticles, the grafting of PCL chains on the surface results in lower modulus values but significantly higher strain at break. This unusual behavior clearly reflects the originality of the reinforcing phenomenon of polysaccharide nanocrystals resulting from the formation of a percolating network thanks to chain entanglements and cocrystallization.

Introduction A possible source of inspiration for the design of new high performance materials is the nature and its wonderful nanocomposite structures. Among biologically inspired nanocomposites, polysaccharides are probably among the most promising sources for the production of nanoparticles. These abundant biorenewable raw materials are increasingly used in nonfood applications. They can also be used for the preparation of crystalline nanoparticles with different geometrical characteristics providing a wide range of potential nanoparticles properties.1,2 Huge quantities of these nanoparticles are potentially available, often as waste products from agriculture. Moreover, polysaccharide surfaces provide potential for significant surface modification using well-established carbohydrate chemistry. The latter allows tailoring the surface functionality of the nanoparticles. In this field, the scientific and technological challenges to take up are tremendous. During the past decade, many attempts have been reported to mimic natural bionanocomposites by blending polysaccharide nanocrystals from different sources with polymeric matrices.3–14 The resulting nanocomposite materials display outstanding mechanical properties, in terms of both stiffness and thermal stability. It has been show that, regardless the source of nanoparticles, cellulose, chitin, or starch, the formation of a rigid percolating network, resulting from strong interactions between them was the basis of this phenomenon. However, gain in modulus is generally obtained at the expense of the ductility of the material. Any factor that affects the formation of this percolating network or interferes with it changes the mechanical performances of the composite. Until now, the processing of polysaccharide nanocrystals reinforced nanocomposites was mainly restricted to aqueous media. The possibility to obtain stable cellulose nanocrystals suspensions in dimethylformamide (DMF),8,15–17 dimethyl sulfoxide (DMSO), and N-methyl pyrrolidine (NMP)17 without any * To whom correspondence should be addressed. E-mail: alain.dufresne@ efpg.inpg.fr.

additive or surface modification was also reported. The surface chemical modification of nanocrystals constitutes another alternative and sound approach to broaden the number of possible polymeric matrices by allowing the processing of composite materials from an organic solvent instead of aqueous suspensions.18–20 However, a severe loss of mechanical performances was reported for nanocomposites processed from these modified nanocrystals because of the coating of the nanoparticles and resulting damaging of the percolating network of nanoparticles. A new and promising way of processing consists in transforming polysaccharide nanocrystals into a cocontinuous material through long chain surface chemical modification. It consists in submitting polysaccharide nanocrystals to surface chemical modification based on the use of grafting agents bearing a reactive end group and a long compatibilizing tail. In a recent work,21 we reported the possibility to successfully graft polymer chains onto starch nanocrystals. Grafting of these species on the surface of polysaccharide nanocrystals should allow enhancing the nonpolar nature of original nanoparticles and dispersion in organic media. In addition, providing that the length of the grafted chains is high enough, entanglements are expected to occur with the polymeric matrix. We report here the characterization of PCL-grafted polysaccharide nanocrystals by X-ray diffraction and contact angle measurements. Two different molecular weights PCL were grafted and both rod-like cellulose and platelet-like starch nanocrystals were used. The two reinforcement types were compared and the importance of grafting was evaluated. The thermal and mechanical properties of highly filled nanocomposite materials obtained from either unmodified or PCL-grafted nanoparticles reinforced PCL are also reported.

Experimental Section Materials. Amylopectin-rich waxy maize starch (trade name Waxylis 200) and pure ramie fibers were obtained from Roquette S. A. (Lestrem, France) and Stucken Melchers GmbH & Co. (Germany), respectively. Sulfuric acid (95%), toluene 2,4-diisocyanate (2,4-TDI, 95%), triethylamine (TEA, 99.5%), polycaprolactone (PCL, 10000 g · mol-1 and

10.1021/bm8001717 CCC: $40.75  2008 American Chemical Society Published on Web 05/30/2008

Bionanocomposites from Polysaccharide Nanocrystals 42500 g · mol-1), acetone (99%), toluene (anhydrous, 99.8%), and dichloromethane (99.5%) were all obtained from Sigma-Aldrich. All solvents were dried over molecular sieves (3 Å, 4-8 mesh beads, Sigma-Aldrich) for 48 h. Nanocrytals Preparation. Ramie fibers were cut into small pieces and treated with 2% NaOH at 80 °C for 2 h to remove residual additives. The purified ramie fibers were submitted to acid hydrolysis with a 65 wt % H2SO4 solution at 55 °C for 30 min and under continuous stirring. The suspension was washed with water by centrifugation and dialyzed to neutrality against deionized water. The obtained suspension was homogenized by using an Ultra Turrax T25 homogenizer for 5 min at 13500 rpm and then filtered in sintered glass No. 1 to remove unhydrolyzed fibers. The suspension was concentrated to constitute the stock suspension. Starch nanocrystals were prepared as described in detail elsewhere.22 Briefly, they were treated by acid hydrolysis of waxy maize starch for 5 days at 40 °C in a 3.16 M aqueous H2SO4 solution with continuous stirring. The resulting suspension was washed with water until neutrality by centrifugation. The final suspension was concentrated to constitute the stock suspension. Polycaprolactone Grafting. Chemical modification of polysaccharide nanoparticles was performed in a round-bottomed reaction flask under a nitrogen atmosphere while constantly stirring with magnetic stir bar. The reaction was carried out, as described in our previous publication,21 with a minor modification. Indeed, the polysaccharide nanocrystals were never dried before grafting, but solvent exchanged from water to toluene. For that, aqueous suspension with the desired amount of cellulose or starch nanocrystals (1 wt %) was solvent exchanged to acetone and then to dry toluene by several successive centrifugation and redispersion operations. Sonication was performed after each solvent exchange step to avoid aggregation. However, the suspension in toluene was not stable in time. Polycaprolactone grafting onto cellulose or starch nanocrystals involved a three-step process. The first step required the reaction of the polymer on isocyanate functionality of phenylisocyanate. The second step required the reaction of the polymer, now protected, with one isocyanate functionality of 2,4-TDI. During the third step, the unreacted second isocyanate functionality of 2,4-TDI was then reacted with the surface hydroxyl groups of the polysaccharide (cellulose or starch) nanocrystals to graft the polymer chain onto the nanoparticles. Details of the reaction are reported elsewhere.21 Each mixture was washed three times with toluene, with successive centrifugations at 10000 rpm and 10 °C for 45 min. It was then washed with dichloromethane with three successive centrifugations in the same conditions to remove ungrafted polymer chains. The washed modified nanocrystals were subsequently Soxhlet extracted with dichloromethane for 24 h before drying at 50 °C in a convective oven. Processing of Nanocomposite Films. The processing of nanocomposite films was done using a casting/evaporation involving dichloromethane, a solvent of PCL, as liquid medium. Suspensions of polysaccharide nanocrystals in dichloromethane were first prepared. Cellulose or starch nanocrystals, at various concentrations, were first dispersed in dichloromethane by stirring during 2 h at room temperature. PCL-grafted nanoparticles were easily dispersed, whereas it was more difficult for unmodified hydrophilic nanoparticles. Indeed, after drying the suspension of unmodified nanocrystals, strong aggregates of nanoparticles formed that were difficult to redisperse. The corresponding amount of PCL was then added to the suspension of polysaccharide nanocrystals and the resulting suspension was stirred again until complete solubilization of PCL. The nanocrystals content was varied from 0 to 50 wt %. It is worth noting that for PCL-grafted nanoparticles, this fraction also includes the grafted polymer chains layer. Solid films were obtained by casting these suspensions on Teflon plates and dichloromethane evaporation at room temperature. The thickness of final films was around 250 µm. X-ray Diffraction. Wide angle X-ray diffraction (WAXD) experiments were carried out at room temperature on powdered dry samples with a Panalytical X’Pert PRO MPD diffractometer in the Bragg-

Biomacromolecules, Vol. 9, No. 7, 2008

1975

Brentano geometry equipped with X’Celerator detector and operated with Cu KR anode (λ ) 0.15406 nm). The 2 θ ranged from 3° to 60° with 2 θ steps of 0.05° and a counting time of 60 s. Contact Angle Measurements. The dynamic contact angle of sessile drops of water on the cellulose and starch nanocrystal pellets was measured with an OCA20 automated and software-controlled VideoBased Contact Angle Meter (DataPhysics Instruments GmbH, Filderstadt, Germany). Smooth surface pellets were obtained by compacting the dried powder of cellulose and starch nanocrystals under a pressure of 10 t using IR press. Pellets were dried again under vacuum for 24 h to remove residual water. All measurements were conducted at room temperature (22 °C). Differential Scanning Calorimetry (DSC). DSC experiments were performed using a DSC Q100 differential scanning calorimeter from TA Instruments. Around 10 mg of samples were placed in a DSC cell in glovebox. Each sample was heated from -100 to 100 °C at a heating rate of 10 °C min-1. The glass transition temperature, Tg, was taken as the inflection point of the specific heat increment at the glass-rubber transition, while the melting temperature, Tm, was taken as the peak temperature of the melting endotherm. Dynamic Mechanical Analysis (DMA). DMA experiments were performed in tensile mode (RSA3, TA Instruments, U.S.A.). The measurements were carried out at a constant frequency of 1 Hz, strain amplitude of 0.05%, a temperature range of -100 to 70 °C, a heating rate of 2 °C · min-1, and a gap between jaws of 10 mm. The samples were prepared by cutting strips from the films with a width of 5 mm. Three samples were used to characterize each material. Tensile Measurements. The tensile mechanical behavior was analyzed with a RSA3 (TA Instruments, U.S.A.) with a load cell of 100 N. Experiments were performed with a cross head speed of 10 mm · min-1 at room temperature, 25 °C. The sample dimensions were 10 × 5 × 0.25 mm3, and the results were averaged on five measurements.

Results and Discussion Acid hydrolysis of native ramie cellulose fibers leads to aqueous suspensions of high aspect ratio rod-like nanocrystals. These nanoparticles exhibit an average diameter of 6-8 nm and a length ranging between 150 and 250 nm as estimated by transmission electron microscopy (micrographs not shown) and corroborate the results obtained for the similar cellulosic substrates hydrolyzed in the same conditions.23 Waxy maize starch nanocrystals consist of 5-7 nm thick platelet-like particles with a length ranging from 20 to 40 nm and a width in the range 15-30 nm. The detailed investigation on the structure of these platelet-like nanoparticles was reported elsewhere.24 It is worth noting that because of electrostatic repulsions between surface sulfate ester groups grafted during the sulfuric acid treatment, the nanocrystal suspensions did not flocculate.25 Grafting efficiency was verified in our preliminary report by Fourier transform infrared (FTIR) and X-ray photoelectron (XPS) spectroscopies, as well as elemental analysis.21 The most efficient grafting was obtained with PCL of lower molecular weight (10000 g · mol-1). For starch nanocrystals, we found that the fraction of hydroxyl groups that have reacted among those accessible was around 2.19 and 0.51 for PCL10000- and PCL42500-grafted nanoparticles, respectively. It corresponded to weight fractions of polymer of 54.45 and 55.25%, respectively.21 Further qualitative evidence can be obtained from contact angle measurements that allow estimating the change in hydrophobicity of the grafted substrates compared to unmodified polysaccharide nanocrystals. It was found that the water droplet was rapidly adsorbed on the surface of ungrafted cellulose or starch nanocrystals, whereas it remained on the surface for PCL-grafted nanocrystals for a much longer time

1976

Biomacromolecules, Vol. 9, No. 7, 2008

Habibi and Dufresne

Figure 2. X-ray diffraction patterns of (A) cellulose and (B) starch nanocrystals before and after PCL grafting: (0) ungrafted, (O) PCL 10000g · mol-1-grafted,and(3)PCL42500g · mol-1-graftednanocrystals. Figure 1. Water contact angle vs time for (A) cellulose and (B) starch nanocrystals before and after PCL grafting: (0) ungrafted, (O) PCL 10000g · mol-1-grafted,and(3)PCL42500g · mol-1-graftednanocrystals.

(Figure 1). In addition, a significant increase in contact angle values was observed for PCL-grafted substrates for both cellulose or starch nanocrystals compared to unmodified substrates, clearly showing the more hydrophobic nature of the polymer-modified substrates. The contact angle values obtained for PCL-grafted substrates corroborate the values found in the literature.26 The highest contact angle value observed for PCL 10000 g · mol-1-grafted nanocrystals is ascribed to the highest grafting efficiency reported for the lowest molecular weight grafted polymer.21 It also appears that the value of the contact angle depends on the nature of the grafted substrate, cellulose or starch. Wide angle X-ray diffraction patterns obtained for both ungrafted and PCL-grafted nanoparticles are shown in Figure 2. Unmodified cellulose nanocrystals display four well-defined diffraction peaks at 14.8° (0.60 nm), 16.5° (0.54 nm), 22.6° (0.39 nm), and 34.5° (0.258 nm), which are typical of cellulose I.27,28 The diffractogram recorded for unmodified starch nanocrystals obtained from waxy maize starch displays typical peaks of A-type amylose allomorph.29 It is characterized by two weak peaks at 2 θ ) 10.05° (0.87 nm) and 11.5° (0.76 nm), a strong peak at 15.3° (0.57 nm), a double strong peak at 17.2° (0.52 nm) and 18.0° (0.49 nm), and a strong peak at 23.5° (0.39 nm). These characteristic peaks observed for either cellulose or starch nanocrystals also appear in the diffractogram obtained for PCLgrafted nanoparticles. This is an indication that the isocyanatemediated grafting reaction does not change the main crystalline

Figure 3. DSC thermograms of PCL10000-grafted cellulose (A) and starch nanocrystals (B).

form of nanocrystals. In addition, for PCL-grafted starch nanocrystals (Figure 2B), a new diffraction peak around 21.3° is clearly evidenced. It should correspond to the nanocrystalinduced crystalline structure of PCL.30 Indeed, extensive crystallization of stearate moieties grafted onto starch nanocrystals surface, forming a crystalline hydrophobic shell around the hydrophilic nanoparticles, was reported.31 This peak is much less defined for PCL-grafted cellulose nanocrystals because a strong diffraction peak due to cellulose I occurs in the same diffraction angle range. The position of this sharp signal is in agreement with the PCL crystallization diffraction pattern. From these results, it is most probable that the covalently linked PCL

Bionanocomposites from Polysaccharide Nanocrystals

Biomacromolecules, Vol. 9, No. 7, 2008

1977

Table 1. Thermal Characteristics of Polysaccharide Nanocrystals Reinforced PCL Nanocomposites Using Data Obtained from the DSC Curvesa filler none cellulose nanocrystals

filler modification none PCL10000-grafted

PCL42500-grafted

starch nanocrystals

none PCL10000-grafted

PCL42500-grafted

filler content (wt %)

Tg (°C)

Tm (°C)

∆Hm (J/g)

χcb

0 10 20 30 10 20 30 40 50 10 20 30 40 50 10 20 30 10 20 30 40 50 10 20 30 40 50

-61.9 -60.3 -60.6 -60.2 -61.5 -60.7 -60.7 -60.9 -60.7 -61.0 -60.4 -60.9 -60.4 -61.0 -60.3 -60.9 -60.3 -60.0 -60.1 -59.1 -59.0 -61.0 -60.9 -60.2 -60.0 -59.2 -60.9

65.6 65.4 65.3 64.6 65.1 65.5 65.0 63.7 63.1 66.0 65.9 65.0 65.3 63.8 64.0 62.5 62.5 65.8 65.6 65.6 65.7 64.4 65.4 66.4 64.3 65.1 64.5

88.2 83.4 84.8 63.9 83.5 80.0 60.4 65.8 59.3 87.5 76.7 71.5 68.9 52.8 98.3 95.3 90.2 86.9 81.2 67.5 60.3 61.1 86.4 84.6 69.2 63.0 50.4

0.56 0.60 0.67 0.58 0.60 0.63 0.62 0.70 0.75 0.62 0.61 0.65 0.73 0.67 0.62 0.60 0.57 0.61 0.64 0.61 0.64 0.77 0.61 0.67 0.63 0.67 0.64

a Glass-rubber transition temperature (Tg); melting temperature (Tm), associated heat of fusion (∆Hm), and degree of crystallinity (χc). b χc ) ∆Hm/ w∆Hm, where ∆Hm ) 157 J · g-1 is the heat of fusion for 100% crystalline PCL and w is the weight fraction of polymeric matrix material in the composite.

chains crystallize at the nanoparticles surface. They are believed to take a crystalline brushlike structure from the nanoparticle surface outward. Indeed, it is worth noting that, after the grafting reaction, each mixture was washed several times with toluene and dichloromethane with successive centrifugations to remove ungrafted polymer chains. The grafted nanoparticles were then Soxhlet extracted with dichloromethane. Then, the remaining PCL can only be covalently bonded to the surface of the nanocrystals. The chemical grafting of polymer chains of chemical structure similar to the one of the matrix on the surface of nanocrystals is expected to create a near perfect interface when using these nanoparticles to process nanocomposites using a matrix with similar chemical nature. The PCL-grafted polysaccharides nanocrystals prepared above have been therefore incorporated in a PCL matrix to obtained nanocomposite materials with high filler content. The thermal properties of both ungrafted and grafted nanocrystals and resulting nanocomposite materials were investigated by DSC. The thermograms obtained for PCL10000-grafted nanoparticles are shown in Figure 3. They display a low magnitude but well defined melting endotherm around 50 and 47 °C for cellulose and starch nanoparticles, respectively, that was not observed for ungrafted nanocrystals. It is obviously attributed to grafted polymeric chains that probably form a crystalline structure at the surface of the nanoparticles in agreement with WAXD experiments. Compared to the typical values of the melting point reported for PCL (around 60 °C), the low temperature position of this melting endotherm is an indication of the restricted size of these crystallites. Similar results were obtained for PCL 42500-grafted nanoparticles. The thermograms (not shown) obtained for the neat matrix, as well as for ungrafted and PCL-grafted nanocomposites, were conventional and very similar among them. A glass transition and a melting phenomenon were observed as a specific heat

increment and endothermic peak, respectively. The thermal data obtained from DSC curves are reported in Table 1. Neither the presence of the nanoparticles nor the grafting of polymer chains on the surface induces any change of the glass transition temperature that remains roughly constant between -61 and -59 °C. It suggests that the filler and its surface morphology do not affect the onset of macromolecular translational and rotational backbone motions of the PCL matrix. It is an indication that there is most probably no direct contact between amorphous PCL domains of the matrix and nanoparticles. Again, the probable crystallization of PCL chains at the nanoparticles surface and crystalline brushlike structure is suggested in agreement with WAXD experiments. The melting point of the PCL matrix is found to remain constant or to slightly decrease upon nanocrystals addition regardless of its surface modification. This is an indication that the size of crystalline PCL domains tends to globally remains identical or slightly decreases. This last point could be related to the expected increase of the viscosity of the polymer solution ascribed to the presence of nanoparticles. This increased viscosity may induce an increase of the activation energy for diffusion of the chains. The crystalline brushlike structure of grafted PCL chains cannot be observed for nanocomposite materials. This is due to the fact that the corresponding melting endotherm is hindered by the main melting endotherm of the PCL matrix. Another possible explanation is the probable cocrystallization of covalently linked chains at the nanoparticles surface with those from the PCL matrix. The addition of starch nanocrystals results in a slight decrease of the degree of crystallinity from 62% for the unfilled matrix to 57% with 30 wt % nanocrystals (Table 1). For ramie whiskers, it seems that the addition of nanoparticles first leads to an increase of the degree of crystallinity of the matrix and then to a progressive decrease for higher filler content. The former phenomenon can be most probably ascribed to an

1978

Biomacromolecules, Vol. 9, No. 7, 2008

Habibi and Dufresne

Figure 4. Logarithm of the storage tensile modulus E′ versus temperature at 1 Hz for PCL-based nanocomposites reinforced with cellulose (Ai) and starch nanocrystals (Bi) with different content (•, 0; O, 10; 3, 20; ], 30; 0, 40; and ×, 50 wt %; i ) 1, unmodified; i ) 2, PCL10000-grafted; and i ) 3, PCL42500-grafted nanocrystals).

anchoring effect of the cellulosic nanoparticles, probably acting as nucleating agents for PCL and favoring its crystallization, as already reported for tunicin whiskers reinforced sorbitol plasticized starch.32 At higher loading level, an inverse dependence with the filler addition is observed and the former phenomenon probably competes with the increase in viscosity. When using PCL-grafted nanoparticles, it seems that the degree of crystallinity of the PCL matrix regularly increases and reaches high values up to 77% for the PCL film reinforced with 50 wt % of PCL10000-grafted starch nanocrystals. It could be ascribed to the easier crystallization process of PCL chains from the matrix with grafted moieties. The mechanical properties of PCL-based nanocomposites reinforced with either unmodified or PCL-grafted cellulose or starch nanocrystals have been investigated in the linear (DMA) and nonlinear range (tensile tests). For unmodified nanoparticles

reinforced PCL films, only low filler content nanocomposites have been tested because of the extreme brittleness of these materials. Figure 4 shows the isochronal evolution of the logarithm of the storage tensile modulus (E′) as a function of temperature. As it is well-known, the exact determination of the glassy modulus depends on the precise knowledge of the sample dimensions. To minimize this influence, the glassy elastic tensile modulus, E′, at -100 °C was normalized at 2.6 GPa for all the samples. It corresponds to the observed average value regardless the composition. The unfilled PCL displays a typical behavior of semicrystalline polymer with four distinct zones. At low temperature, the modulus slightly decreases with temperature but remains roughly constant. It is due to the fact that in this temperature range, the polymer is in the glassy state and molecular motions are largely restricted to vibration and short-

Bionanocomposites from Polysaccharide Nanocrystals

Biomacromolecules, Vol. 9, No. 7, 2008

1979

Table 2. Rubbery Tensile Storage Modulus Estimated at 0 °C (E′0) and High Strain Tensile Mechanical Properties of Polysaccharide Nanocrystals/PCL Nanocomposite Filmsa filler none cellulose nanocrystals

filler modification none PCL10000-grafted

PCL42500-grafted

starch nanocrystals

none PCL10000-grafted

PCL42500-grafted

filler content (wt %)

E′0 (MPa)

E (MPa)

σb (MPa)

b (%)

0 10 20 30 10 20 30 40 50 10 20 30 40 50 10 20 10 20 30 40 50 10 20 30 40 50

518 776 660 524 690 847 879 889 1340 605 671 982 1019 1028 638 425 651 829 856 868 936 596 527 703 853 991

231 ( 25 338 ( 30 300 ( 27 254 ( 18 252 ( 15 345 ( 20 356 ( 22 372 ( 20 380 ( 17 255 ( 16 294 ( 18 328 ( 21 390 ( 29 442 ( 31 292 ( 18 243 ( 20 261 ( 14 300 ( 27 337 ( 20 350 ( 22 360 ( 25 280 ( 21 326 ( 30 300 ( 24 320 ( 25 384 ( 32

22.0 ( 2.0 16.7 ( 1.5 13.9 ( 1.7 7.64 ( 1.1 20.2 ( 2.0 21.6 ( 1.6 20.0 ( 1.2 18.3 ( 1.8 16.8 ( 1.3 20.0 ( 1.9 15.1 ( 1.5 15.7 ( 1.5 17.6 ( 2.0 18.7 ( 2.5 10.6 ( 1.1 6.0 ( 0.7 18.2 ( 2.0 17.3 ( 1.9 18.2 ( 2.1 18.4 ( 2.5 15.7 ( 2.0 21.2 ( 2.7 17.8 ( 1.8 17.0 ( 2.0 18.7 ( 2.7 15.5 ( 2.1

637 ( 40 395 ( 27 8.0 ( 1.0 4.0 ( 0.6 531 ( 25 420 ( 23 185 ( 12 64.0 ( 3.0 28.0 ( 1.7 598 ( 34 200 ( 23 13.5 ( 1.5 12.0 ( 1.2 8.60 ( 0.8 6.30 ( 1.0 2.2 ( 0.4 535 ( 24 83.5 ( 13 11.4 ( 1.1 8.2 ( 0.7 6.0 ( 0.5 284 ( 17 13.0 ( 1.1 8.0 ( 0.9 7.7 ( 1.3 5.0 ( 0.7

a Young modulus (E), strength (σb), and strain at break (b). Data into square brackets correspond to nanocomposite films processed from “not dried” nanoparticles.

range rotational motions. Around -60 °C, the modulus sharply drops. This temperature range corresponds to the main relaxation process of amorphous PCL domains associated to the glass transition of the polymer in agreement with DSC observations. The magnitude of this modulus drop is relatively weak because of the high degree of crystallinity of the material. In the temperature range from -60 to 40 °C, the storage tensile modulus slightly decreases due to the progressive melting of PCL. In this temperature range, amorphous rubbery domains coexist with crystalline domains, which play the role of both filler particles and physical cross-links for the system. At higher temperatures, the modulus drops sharply and irremediably due to the complete melting of the crystalline zones of PCL around 50 °C. When adding low content (i.e., 10 wt %) of ungrafted cellulose or starch nanocrystals (Figure 4, panels A1 and B1) to the PCL matrix, the rubbery storage modulus E′ increases. The rubbery modulus values estimated at 0 °C of the PCLbased films are collected in Table 2. Enhanced rigidity can be ascribed, at least partially, to the increase in the degree of crystallinity of the matrix observed from DSC measurements. It could also be ascribed to the reinforcing effect of the nanoparticles. This reinforcing effect is associated with filler-filler interactions through hydrogen bonding as already reported.5 For high nanocrystal content (above 10 wt %), the behavior is different. Both the rubbery modulus and the thermal stability of the material decrease. It probably results from the poor dispersability of the highly hydrophilic nanoparticles within the hydrophobic matrix. This aggregation phenomenon of the nanoparticles within the polymeric matrix could be most probably due to the processing of the nanocomposites. Indeed, ungrafted nanoparticles reinforced PCL films were obtained by a casting/evaporation method using dichloromethane as medium. For PCL-grafted nanoparticles reinforced nanocomposites, the rubbery modulus was found to continuously increase upon filler addition (Figure 4, panels A2,3 and B2,3, and Table 2). Again,

it can be ascribed to both the increase in the degree of crystallinity of the polymeric matrix as shown by DSC measurements for some formulations (Table 1) and to a reinforcing effect of the nanoparticles. Also, it is worth noting that the nanocrystal content refers to the weight fraction of grafted nanoparticles and that the effective nanocrystal content is lower compared to unmodified. In addition the thermal stability of the nanocomposite film increases and the temperature for which the modulus drops irreversibly is shifted toward higher temperatures. All these observations lead to the conclusion that a strong adhesion exists between the reinforcing phase and the matrix, resulting in an efficient dispersion of the nanoparticles and efficient stress transfer at the interface. Tensile tests were performed on PCL-based films up to the break. The tensile modulus, strength and elongation at break values are reported in Table 2. The unfilled PCL matrix is found to be as well-known a quite ductile polymer, able to undergo large deformations before break. Its tensile modulus, around 200 MPa, is relatively modest and characteristic of semicrystalline polymers above their Tg. The addition of unmodified cellulose or starch nanofillers into PCL contributes to improving its tensile modulus but, at the same time, reduces drastically all the other mechanical properties, especially the elongation at break. The elongation at break decreases from 637% down to 8.0 or 2.3% when adding 20 wt % of cellulose or starch nanocrystals, respectively. The lack of intimate adhesion between the filler and the matrix leads to numerous irregularly shaped microvoids or microflaws in the composite structure. Because of these microflaws the stress transfer from the matrix to the filler is poor, and the mechanical properties of the nanoparticles are not fully utilized. The brittleness of the material is accentuated by the probable aggregation of the unmodified nanoparticles that leads to the formation of weak points. The addition of PCL-grafted ramie whiskers or waxy maize starch nanocrystals results in a continuous increase of the tensile modulus in agreement with DMA experiments. Compared to

1980

Biomacromolecules, Vol. 9, No. 7, 2008

unmodified nanoparticles reinforced PCL, the decrease in both the strength and the elongation at break is much more gradual. For instance, the elongation at break decreases from 637% down to 420 or 83% when adding 20 wt % of PCL10000-grafted cellulose or starch nanocrystals, respectively. It is most probably ascribed to the better dispersion of the filler within the polymeric matrix, induced by the grafting of PCL chains on the surface of the nanoparticles. The possibility of chain entanglements between grafted chains and PCL chains from the matrix could probably play an important role in this phenomenon. The difference between both kinds of nanoparticles is most probably ascribed to the poorer dispersion ability in the PCL matrix of starch nanocrystals24 and resulting lower grafting efficiency. Indeed, they are generally observed in the form of aggregates having an average size around 4.4 µm, as measured by laser granulometry.24 When using PCL42500-grafted nanoparticles, the ultimate properties decrease more rapidly as a function of filler loading. Even if the possibility of chain entanglements between grafted chains and PCL chains from the matrix should be higher for the higher molecular weight grafted chains, the lower efficiency of the grafting seems to dominate the effect.

Conclusions Rod-like cellulose nanocrystals extracted from ramie fibers and starch nanocrystals extracted from waxy maize starch were decorated with PCL chains. PCLs with two different molecular weights were grafted. Both unmodified and decorated nanoparticles were used to process highly filled nanocomposites by a casting/evaporation method using PCL as matrix. Grafted polymeric chains are organized as a semicrystalline shell on the surface of the nanoparticles, as evidenced from WAXD and DSC experiments. Nanocomposite materials processed from modified nanoparticles display both a high modulus and a good ductility (elongation at break) contrarily to unmodified nanocrystal-based composites. The transformation of polysaccharide nanocrystals into a cocontinuous material through long chain surface chemical modification consists in a new and promising way for the processing of nanocomposite materials. Acknowledgment. The authors acknowledge ADEME (Agence Franc¸aise de l’Environnement et de la Maıˆtrise de l’Energie) for the financial support (Convention 0401C0011) and Ste´phane Coindeau (CMTC-ENSEEG) for his help with X-ray diffraction analysis.

References and Notes (1) Azizi Samir, M. A. S.; Alloin, F.; Dufresne, A. Biomacromolecules 2005, 6, 612–626.

Habibi and Dufresne (2) de Souza Lima, M. M.; Borsali, R. Macromol. Rapid Commun. 2004, 25, 771–787. (3) Angellier, H.; Molina-Boisseau, S.; Dole, P.; Dufresne, A. Biomacromolecules 2006, 7, 531–539. (4) Berglund, L. In Natural Fibers, Biopolymers, and Biocomposites; Mohanty, A. K., Misra, M., Drzal, L. T., Eds.; CRC Press: Boca Raton, FL, 2005; p 807. (5) Dufresne, A. J. Nanosci. Nanotechnol. 2006, 6, 322–330. (6) Kristo, E.; Biliaderis, C. G. Carbohydr. Polym. 2007, 68, 146–158. (7) Kvien, I.; Oksman, K. Appl. Phys. A: Mater. Sci. Process. 2007, 87, 641–643. (8) Marcovich, N. E.; Auad, M. L.; Bellesi, N. E.; Nutt, S. R.; Aranguren, M. I. J. Mater. Res. 2006, 21, 870–881. (9) Morin, A.; Dufresne, A. Macromolecules 2002, 35, 2190–2199. (10) Oksman, K.; Mathew, A. P.; Bondeson, D.; Kvien, I. Compos. Sci. Technol. 2006, 66, 2776–2784. (11) Paillet, M.; Dufresne, A. Macromolecules 2001, 34, 6527–6530. (12) Petersson, L.; Kvien, I.; Oksman, K. Compos. Sci. Technol. 2007, 67, 11–12. (13) Sriupayo, J.; Supaphol, P.; Blackwell, J.; Rujiravanit, R. Polymer 2005, 46, 5637–5644. (14) Wu, Q.; Henriksson, M.; Liu, X.; Berglund, L. Biomacromolecules 2007, 8, 3687–3692. (15) Azizi Samir, M. A. S.; Alloin, F.; Sanchez, J. Y.; El Kissi, N.; Dufresne, A. Macromolecules 2004, 37, 1386–1393. (16) Ljungberg, N.; Cavaille´, J. Y.; Heux, L. Polymer 2006, 47, 6285– 6292. (17) van den Berg, O.; Capadona, J. R.; Weder, C. Biomacromolecules 2007, 8, 1353–1357. (18) Ljungberg, N.; Bonini, C.; Bortolussi, F.; Boisson, C.; Heux, L.; Cavaille´, J. Y. Biomacromolecules 2005, 6, 2732–2739. (19) Nair, K. G.; Dufresne, A.; Gandini, A.; Belgacem, M. N. Biomacromolecules 2003, 4, 1835–1842. (20) Roman, M.; Winter, W. T. Cellulose Nanocomposites: Processing, Characterization and Properties; ACS Symposium Series 938; American Chemical Society: Washington, DC, 2006; pp 99-113. (21) Labet, M.; Thielemans, W.; Dufresne, A. Biomacromolecules 2007, 8, 2916–2927. (22) Angellier, H.; Choisnard, L.; Molina-Boisseau, S.; Ozil, P.; Dufresne, A. Biomacromolecules 2004, 5, 1545–1551. (23) Habibi, Y.; Foulon, L.; Aguie-Beghin, V.; Molinari, M.; Douillard, R. J. Colloid Interface Sci. 2007, 316, 388–397. (24) Putaux, J. L.; Molina-Boisseau, S.; Momaur, T.; Dufresne, A. Biomacromolecules 2003, 4, 1198–1202. (25) Angellier, H.; Putaux, J. L.; Molina-Boisseau, S.; Dupeyre, D.; Dufresne, A. Macromol. Symp. 2005, 221, 95–104. (26) Biresaw, G.; Carriere, C. J. J. Polym. Sci., Part B: Polym. Phys. 2001, 39, 920–930. (27) Nishiyama, Y.; Langan, P.; Chanzy, H. J. Am. Chem. Soc. 2002, 124, 9074–9082. (28) Sugiyama, J.; Vuong, R.; Chanzy, H. Macromolecules 1991, 24, 4168– 4175. (29) Bule´on, A.; Ge´rard, C.; Riekel, C.; Vuong, R.; Chanzy, H. Macromolecules 1998, 31, 6605–6610. (30) Chen, T. K.; Tien, Y. I.; Wei, K. H. J. Polym. Sci., Part A: Polym. Chem. 1999, 37, 2225–2233. (31) Thielemans, W.; Belgacem, M. N.; Dufresne, A. Langmuir 2006, 22, 4804–4810. (32) Mathew, A. P.; Dufresne, A. Biomacromolecules 2002, 3, 609–617.

BM8001717