Supra-Molecular EcoBioNanocomposites Based on Polylactide and Cellulosic Nanowhiskers: Synthesis and Properties

Article pubs.acs.org/Biomac

Supra-Molecular EcoBioNanocomposites Based on Polylactide and Cellulosic Nanowhiskers: Synthesis and Properties Birgit Braun,† John R. Dorgan,*,‡ and Laura O. Hollingsworth§ †

PolyNew Inc., Golden, Colorado 80401, United States now: Dow Chemical, Performance Plastics Characterization R&D, Freeport, Texas 77541, United States ‡ Department of Chemical Engineering, Colorado School of Mines, Golden, Colorado 80401, United States § PolyNew, Inc., Golden, Colorado 80401, United States ABSTRACT: Successful filler dispersion and establishment of good interfacial contact with the surrounding matrix are essential for optimized reinforcement in polymeric nanocomposites. In particular, in renewable-based composites this can be challenging, where hydrophilic attractions between nanofillers facilitate aggregation. Here an innovative approach to prepare cellulosic nanowhisker (CNW) reinforced polylactide (PLA) is presented. The lactide ring-opening polymerization is initiated from CNW surface hydroxyl groups after partial acetylation to control the grafting density. Grafting of PLA chains is verified by Fourier transform infrared spectroscopy. The resulting nanocomposites display exceptional properties; a heat distortion temperature of 120 °C is achieved at 10 wt % CNW loading and can be further enhanced to reach 150 °C at 15 wt % CNW. The formation of a percolating network is verified by comparison of modulus data with an established theoretical model. Additionally, nucleation by CNWs reduces the crystallization half-time to 15 s compared with 90 s for PLA. Melt-pressed films retain transparency indicating good filler dispersion.

■

INTRODUCTION In the past decades, potential climate effects of rising carbon dioxide levels, consequences of the finite nature of fossil resources on energy availability, and mounting solid waste streams have been of increasing concern. In the plastics industry, this has led to attempts to increase recycling rates and a growing focus on biobased macromolecules as replacement for traditional, fossil-based polymers. Polylactide (PLA) is an economically successful biobased polymer. Large-scale production was accomplished in 2001 via fermentation of sugars to lactic acid, followed by reactive distillation to the lactide monomer and subsequent ringopening polymerization.1,2 In addition to being cost competitive, PLA possesses a significantly lower nonrenewable energy content compared with various other common polymers.2 The mechanical properties allow fabrication of fibers and films, extrusion and thermoforming, and injection molding, and the majority of its current global production enters the packaging and fiber sector.3,4 The basic properties of PLA depend strongly on optical purity, molecular weight and structure, and degree of crystallinity but are generally similar to those of polystyrene.5 However, the relatively low heat distortion temperature (HDT) of the homopolymer is particularly limiting as it excludes the use of PLA for usetemperatures exceeding 75 °C. Incorporation of reinforcing fillers is an established approach for the adjustment of material properties. Whereas synthetic and mineral reinforcements have been successfully introduced into PLA for property enhancement,6−9 fillers derivable from © 2012 American Chemical Society

renewable resources are desired to maintain the ecologically friendly character of the material in accordance with the 12 Principles of Green Chemistry.10 Direct use of biofibers is demonstrated in the literature,11−13 but the hydrophilic nature renders good dispersion challenging in the hydrophobic polymer; relatively high fiber loadings are necessary for sufficient mechanical properties compromising color and transparency. Cellulose contains whisker-like regions of highly ordered polysaccharide chains with diameters between 5 and 30 nm,14−16 which can be isolated via acid hydrolysis by taking advantage of the lower hydrolysis rate for crystalline regions.17 These cellulosic nanowhiskers (CNWs) demonstrate impressive mechanical properties; the tensile modulus was determined to be 143 GPa for CNWs isolated from tunicin18 and between 57 and 105 GPa for cotton.19 When well-dispersed in a polymeric matrix, the formation of a percolating network among the rod-shaped filler particles results in exceptional mechanical reinforcement, even at low loading levels.20 Given the hydrophilic nature of native CNWs, hydrosoluble polymers are an obvious choice as host matrices,21−25 but repulsive forces lead to aggregation and poor interfacial contact in hydrophobic polymers. Attempts to overcome this obstacle include use of surfactants,26,27 silane treatment,28 direct polymer grafting29 and polymer grafting,30−32 and surface acylation33 and Received: January 28, 2012 Revised: May 26, 2012 Published: May 30, 2012 2013

dx.doi.org/10.1021/bm300149w | Biomacromolecules 2012, 13, 2013−2019

Biomacromolecules

Article

Figure 1. Illustration of the synthesis approach to optimize the reinforcement potential of CNWs in polylactide by controlling the grafting density.

acetylation.34,35 Alternatively, nanocomposites can be fabricated via the formation of a 3-D CNW network of well-individualized nanowhiskers36,37 or following a papermaking-like process using microfibrillated cellulose,38 followed by impregnation with the polymer. In this work, the approach toward composite fabrication takes advantage of the mechanism through which the lactide ring-opening polymerization proceeds when stannous octoate is used as catalyst.39,40 In the absence of free initiators, only hydroxyl groups available on the CNW surface initiate the polymerization reaction. Contrary to previously published work,29 the molecular weight of the grafted chains is controlled through partial substitution of the accessible surface hydroxyls with acetate groups (Figure 1). The resulting supramolecular structures have molecular weights in the billions without the need of additional compounding. The combination of advances in biotechnology for the utilization of renewable resources to produce biopolymers with the ability of material property modification through nanotechnology allows for the creation of ecologically responsible materials, which can be referred to as ecobionanocomposites.

■

concentration of 17.5 M acetic and 0.027 M hydrochloric acid, and the mixture was stirred vigorously for 2 h. After quenching the reaction with ice water, the acid was removed by repeated centrifugation at 5200 rpm and replacement of the supernatant with DI water (three times). The suspension was diluted to twice the volume and blended in a Waring blender for 5 min, and the cellulose was washed until the pH reached ∼5 at which the supernatant remained turbid. After collection of the first turbid supernatant, the cake was resuspended in DI water and two additional turbid supernatants collected. After combination of the turbid supernatants, the CNW suspension was freeze-dried. C. CNW Characterization. A small amount of freeze-dried, acetylated cellulosic nanowhiskers (a-CNWs) was ground into a fine powder with potassium bromide (IR-grade from Sigma Aldrich) and pressed into a pellet to determine the extent of acetylation using Fourier transform infrared (FT-IR) spectroscopy. Spectra were recorded on a Thermo Nicolet Nexus 670 FTIR in the range of 400 to 4000 cm−1. The instrument settings were kept constant (120 scans, data spacing: 0.964 cm−1, velocity: 0.6329, aperture: 69). CNWs were also examined in a transmission electron microscope at 200 kV (CM200 TEM by Philips, now FEI, Hillsboro, OR). A drop of dilute aqueous suspension was allowed to dry on a carbon-coated copper grid. To enhance contrast, the grid was floated in a 2 wt % uranyl acetate solution for 3 min. Samples were dried under vacuum for 20 min at room temperature before examination. D. Solution Polymerization. Freeze-dried a-CNWs were placed under vacuum (635 mmHg) at 80 °C for at least 24 h. After purging with argon, the material was suspended in anhydrous toluene and sonicated in a Branson Ultrasonic Bath for 24 h to individualize nanocrystals. Recrystallized L-lactide was dried at 40 °C under vacuum (635 mmHg) for at least 48 h in a flame-dried, argon-purged roundbottomed flask. After sonication, the nanowhisker suspension was transferred to the round-bottomed flask, and the mixture purged with argon through a septum stopper. Stannous octoate was added at a ratio of 2500 monomer molecules per catalyst molecule. The flask was subsequently submersed in an oil bath at 90 °C, and the polymerization reaction was allowed to proceed for 68 h. At the end of the reaction time, the suspension was diluted with chloroform until it could be stirred with a magnetic stir bar, and poly(acrylic acid) dissolved in dioxane at 0.2 g/mL was added at a level of 0.25 wt % of initial lactide to deactivate the catalyst.41 After stirring for 24 h, the

EXPERIMENTAL SECTION

A. Materials. Microcrystalline cellulose (Avicel PH-101) was purchased from Sigma Aldrich, as were glacial acetic acid, anhydrous toluene (99.98%), and stannous octoate. Stannous octoate was distilled under reduced pressure, and solutions in anhydrous toluene were prepared for each reaction. Poly(acrylic acid) with a molecular weight of 2000 g/mol was also purchased from Sigma Aldrich, and dissolved in 1,4-dioxane for transfer purposes. Solvents were reagent grade and used as received. L-Lactide was purchased from NatureWorks (Minnetonka, MN) and recrystallized from ethyl acetate. B. Isolation and Acetylation of CNWs from Microcrystalline Cellulose. Microcrystalline cellulose was dried at 130 °C for 1 h, followed by drying under vacuum (635 mmHg) for 24 h at room temperature. After blending at a concentration of 15 wt % in glacial acetic acid in a Waring blender for 15 min, the suspension was transferred to a 2 L jacketed glass reactor and heated to 95 °C. Deionized (DI) water and hydrochloric acid were added to obtain a 2014

dx.doi.org/10.1021/bm300149w | Biomacromolecules 2012, 13, 2013−2019

Biomacromolecules

Article

polymer composite was precipitated into 10-fold excess of methanol and dried under vacuum at 60 °C overnight. In the subsequent discussion, the filler loading levels refer to the CNW-to-lactide ratio employed to make the composite. E. Bulk Polymerization. Freeze-dried a-CNWs and L-lactide were dried as described above, and defined amounts were mixed manually as powders in a round-bottomed flask before purging with argon for 10 min. The flask was submersed in an oil bath at 130 °C, and the monomer was allowed to melt. The mixture was briefly stirred before the catalyst was added at a ratio of 2500 monomer molecules per catalyst molecule. The reaction was allowed to proceed for 4 h, which leads to a monomer conversion of 96%.40 The material was dissolved in chloroform, and poly(acrylic acid) dissolved in dioxane was added at a level of 0.25 wt % relative to lactide. The material was isolated, as described above. In the subsequent discussion, the filler loading levels refer to the CNW-to-lactide ratio employed to make the composite. F. Solution Blending. Freeze-dried a-CNWs were kept at 80 °C under vacuum (635 mmHg) for at least 24 h. After purging with argon, a-CNWs were suspended in chloroform and sonicated for 24 h. PLA purchased from NatureWorks (Grade 2000D) was dissolved in chloroform at a concentration of 5 wt % and combined with the aCNW suspension at various ratios. After stirring for 24 h, the composite was precipitated into an excess of methanol and dried as previously described. G. Compression Molding and Mechanical Testing. Samples for mechanical testing were prepared by vacuum-compression molding. The material was allowed to melt in the mold cavity at 180 °C before vacuum was introduced in a stepwise manner. The samples were kept at 635 mmHg vacuum for 5 min before they were transferred to a compression-molding machine preheated to 200 °C. A constant load of 5000 psi was applied for 3 min, during which the mold temperature reached a maximum value of 195 °C. After cooling to room temperature, the sample bars were removed from the mold cavities, crystallized at 110 °C for 3 h, and allowed to physically age for 24 h before testing. Mechanical properties were determined by dynamic mechanical thermal analysis (DTMA) using an ARES rheometer calibrated for normal force and torque. Samples were subjected to 0.1% strain at 1 Hz during a temperature ramp from 30 to 160 at 5 °C/min under tension (10% from 30 to 100 °C, 1% from 100 to 160 °C). The HDT under a load of 1820 kPa was determined according to the correlation established by Takemori (Young’s modulus equals 0.75 GPa).42 H. Differential Scanning Calorimetry. Samples after compression molding were analyzed for the glass-transition temperature (Tg) and degree and rate of crystallization in a Perkin-Elmer DSC-7. The machine was calibrated against an Indium standard twice, and a baseline was established on a daily basis. The DSC testing protocol was as follows: (a) heat from 5 to 200 at 10 °C/min, (b) hold at 200 °C for 5 min, (c) quench to 110 at 50 °C/min, (d) hold at 110 °C for 30 min, (e) cool from 110 to 5 at 5 °C/min, and (f) heat from 5 to 200 at 10 °C/min. Tg and amount of achievable crystallinity were determined on the second heating cycle. I. Proof of Covalent Attachment of PLA Chains. Covalent attachment of the PLA chains onto the CNW surface was verified by FT-IR spectroscopy. The composite was dissolved in chloroform and stirred for 24 h to facilitate desorption of noncovalently bonded PLA chains, and the solid fraction was separated by centrifugation at 8600g for 40 min. This fraction was resuspended in chloroform, and the procedure was repeated. The spectrum resulting from the isolated solids was compared with a control experiment assessing the extent of physisorption of PLA chains onto the CNW surface. The solutionpolymerization procedure was repeated using preformed, dried PLA (weight-average molecular weight, Mw = 167 000 g/mol) instead of lactide without catalyst addition. The solid fraction was isolated as previously described.

mechanical and thermal properties requires the optimization of multiple factors. Grafted polymer chains improve the compatibility of the filler with the surrounding matrix and prevent self-association of the whiskers, but sufficient molecular weight is vital to avoid the necessity to compound subsequently the modified filler with the homopolymer. Additionally, complete surface coverage is detrimental to the reinforcement potential of the filler because interparticle interactions are required to facilitate the stress transfer within a percolating network.20 The key is to control the number of hydroxyl groups on the CNW surface available for polymerization initiation through partial substitution before conducting the ring-opening polymerization of lactide monomer in the presence of these CNWs in solution and bulk. Partial surface functionalization of CNWs is achieved through a documented method that allows the acetylation of surface hydroxyl groups during the isolation of the nanocrystals from the source material;43,44 In this work, the source for CNWs is partially hydrolyzed, spray-dried cellulose pulp (microcrystalline cellulose). CNWs are isolated and dispersed in acetic acid through mechanical energy input. Accessible surface hydroxyls are subsequently acetylated via Fischer esterification at elevated temperatures, as verified by FT-IR analysis. The spectrum shown in Figure 2 of acetylated CNWs

Figure 2. Fingerprint region of native and acetylated CNW. The arrow marks the signal characteristic for the carbonyl carbon of the ester bond at 1736 cm−1. The photo demonstrates the increase in hydrophobicity of the functionalized CNWs; whereas native CNWs remain in the aqueous phase on top after rigorous mixing of a twophase system composed of chloroform and water (right), acetylated CNWs partition into the organic solvent on the bottom (left).

(a-CNWs) exhibits a peak centered at 1736 cm−1, which is characteristic for the ester carbonyl carbon of the acetate group45 but absent in unmodified CNWs. Decreased hydrophilicity is evident by partitioning of a-CNWs between immiscible phases and shown in the photograph included in Figure 2; whereas native CNWs remain in the aqueous phase on the top after rigorous mixing of the two-phase system and re-establishment of the equilibrium (right vial), a-CNWs prefer to migrate into the organic bottom layer, where they remain dispersed (left vial). The surface modification does not affect the average dimensions of the CNWs; they range between 200 and 500 nm in length with diameters around 10 nm based on

■

RESULTS AND DISCUSSION The successful synthesis of supramolecular structures composed of CNWs with grafted PLA chains that exhibit improved 2015

dx.doi.org/10.1021/bm300149w | Biomacromolecules 2012, 13, 2013−2019

Biomacromolecules

Article

TEM (Figure 3). The presence of aggregates is detected along with individual whiskers; it is unclear whether this is due to

Figure 4. FT-IR spectra of the bottom fractions isolated by centrifugation of the control experiment and in situ polymerized composites compared with polylactide and freeze-dried acetylated cellulosic nanowhiskers. Dotted lines indicate PLA signals that do not overlap with peaks originating from cellulose.

Figure 3. TEM image of acetylated cellulosic nanowhiskers isolated from microcrystalline cellulose.

incomplete separation during the isolation or caused by drying of the a-CNW suspension on the grid. Following the approach outlined in ref 43, the average degree of substitution for surface hydroxyl groups was estimated to be around 60 to 70%.43 Nanocomposites are subsequently synthesized via lactide polymerization in the presence of freeze-dried a-CNWs dispersed in toluene. The resulting precipitate consists of CNWs with PLA chains covalently grafted to the surface (CNW-g-PLA) embedded in free polymer molecules initiated by adventitious water present during the synthesis. Grafting is verified using FT-IR analysis by comparison of the spectrum of washed CNW-g-PLA with a control experiment assessing the affinity of PLA chains to adsorb irreversibly onto the a-CNW surface. The resulting spectra are shown in Figure 4. Nonoverlapping PLA absorbance peaks originating from CH3 stretching (3017 cm−1), in-plane deformation (1458 cm−1), and rocking vibration (1187 cm−1) as well as increases in the signals characteristic for the CO stretching vibration (1760 cm−1) and in-plan deformation (762 cm−1)46 are observed for CNWg-PLA but absent in the control experiment. The presence of surface grafted PLA chains has a profound impact on the material properties. In Figure 5, the HDT under a load of 1820 kPa is plotted against the a-CNW loading level, showing tremendous improvements for solution-polymerized nanocomposites. Whereas the effect is small at lower loadings, the HDT increases to 120 °C from 78 °C at 10 wt % CNW and is further increased to 150 °C at 15 wt % filler loading. However, increasing the a-CNW fraction to 20 wt % does not further enhance the heat resistance; as the number of initiating hydroxyls increases with a-CNW loading, it is valid to attribute this effect to a decrease in the average molecular weight of the polymer chains. When conducting the synthesis in bulk rather

Figure 5. Heat distortion temperature as a function of a-CNW weight loading for nanocomposites prepared by in situ polymerization in solution and bulk compared with solution-blended nanocomposites.

than in toluene, improvements in HDT occur to a lesser extent; at 15 wt % a-CNW, the heat resistance is 115 °C. Insufficient individualization of a-CNWs due to the lack of solvent and energy input (sonication) is the likely to be the cause. It is evident that aggregate breakup is not the only factor contributing to the superior mechanical improvement in solution polymerized nanocomposites because solution blending of a sonicated a-CNW suspension with preformed PLA in chloroform does not result in similar HDT values despite comparable Hansen solubility parameters for toluene and chloroform. Therefore, only the presence of grafted PLA chains preventing the reaggregation of nanoparticles embedded in the matrix during isolation and sample preparation can explain the impressive HDT increases. The origin of the exceptional increases in heat resistance in solution-polymerized nanocomposites is the formation of a 2016

dx.doi.org/10.1021/bm300149w | Biomacromolecules 2012, 13, 2013−2019

Biomacromolecules

Article

percolating network of filler elements resulting in substantial modulus increases over the entire temperature range examined (40 to 160 °C). Whereas improvements in the glassy modulus are significant, the reinforcement effect is particularly remarkable above the glass-transition temperature Tg when compared with other studies with unmodified CNWs dispersed in PLA.38,47,48 Adaptation of the percolation concept to the classical parallel-series allows the calculation of the shear modulus Gc for percolated nanocomposites.49,50 Gc =

(1 − 2ΨX r)GsGr + (1 − X r)ΨGr2 (1 − X r)Gr + (X r − Ψ)Gs

contacts; the good agreement with the experimental data corroborates that the approach of limiting the PLA grafting density through acetylation of a fraction of the CNW surface hydroxyls successfully allows the formation of such a network. The value of the aspect ratio used in the model is 15.6, which is slightly higher than the value of 10.5 previously reported for CNWs isolated from microcrystalline cellulose,37 and likely to be due to the more involved fractionation procedure used in this study. The modulus of the rigid, reinforcing network takes a value of 22.8 GPa in the percolation model. Compared with a previous publication, this is about two orders of magnitude higher than the modulus achieved using the template approach36 for nanocomposite preparation with unmodified CNWs isolated from microcrystalline cellulose.37 However, values reported in the literature for the mechanical properties of individual CNWs19 and their networks span a wide range,36 and many parameters such as interfacial compatibility, type of polymer matrix, CNW size distribution, and cellulose source material influence these properties. Nanocomposite films retain their transparency, as demonstrated in the photograph in Figure 7. This is further

(1)

In eq 1, Xr is the filler volume fraction, Gs and Gr are the shear moduli of the soft (polymer) and rigid (filler) phases, and Ψ is the percolating fraction of the rigid filler; whereas the latter is considered to be an adjustable parameter in the model developed by Takayanagi, Uemura, and Minami,49 it can be calculated according to the following equation based on the percolation concept. X r < Xc

Ψ=0

X r > Xc

⎛ X r − Xc ⎞0.4 Ψ = X r⎜ ⎟ ⎝ 1 − Xc ⎠

(2)

Xc is the critical volume fraction of rigid filler to reach percolation and can be related to the particle aspect ratio f according to Xc = 0.7/f.51 Whereas f is often determined experimentally by analysis of a number of TEM images, it is treated as an adjustable parameter in this study, along with the rigid phase shear modulus. Experimentally determined shear moduli of a-CNW nanocomposites synthesized by solution polymerization are compared with the percolation model in Figure 6 and are in excellent agreement for loading levels up to 15 wt %; for the reason mentioned above, the sample containing 20 wt % filler is not considered in this analysis. An underlying assumption of the percolation model is that the stress transfer within the percolating filler network occurs efficiently through interparticle

Figure 7. Melt-pressed films of the nanocomposite materials retain transparency (shown: 10 wt % a-CNW).

verification that PLA grafting onto the a-CNW surface facilitates good dispersion of the filler in the matrix. In addition, the material demonstrates improved heat stability compared with alternative fabrication routes; nanocomposites prepared by PLA grafting onto CNWs isolated via sulfuric acid hydrolysis showed significant discoloration upon injection molding.29 The highest possible degree of crystallinity and the time required to achieve it are important processing parameters for semicrystalline polymers like PLA. Optically pure PLA is notoriously slow to crystallize, which limits the practicality of injection molding and other processing operations.9 Additionally, the mechanical properties of PLA composites depend significantly on crystallinity, particularly at temperatures above the glass transition.48 Nanosized filler particles can serve as nucleating agents for PLA crystallization when well-dispersed9 and result in notably reduced crystallization half-times. The effect of dispersed a-CNWs on the crystallization kinetics is shown in Figure 8 when nanocomposites prepared by solution polymerization are quenched to 110 °C from the melt. The crystallization half-time at 5 wt % CNW loading level is just 45 s, which is half the time needed for neat PLA synthesized under

Figure 6. Experimentally determined shear storage moduli for nanocomposite materials prepared by in situ solution polymerization compared with the fitted percolation model. 2017

dx.doi.org/10.1021/bm300149w | Biomacromolecules 2012, 13, 2013−2019

Biomacromolecules

■

REFERENCES

(1) Lunt, J. Polym. Degrad. Stab. 1998, 145−152. (2) Vink, E. T. H.; Rabago, K. R.; Glassner, D. A.; Gruber, P. R. Polym. Degrad. Stab. 2003, 403−419. (3) Dorgan, J. R., Braun, B., Wegner, J. R., Knauss, D. M. In Degradable Polymers and Materials; Scholz, C., Khemani, K.; ACS Symposium Series 939; American Chemical Society: Washington, DC, 2006; pp 102−125. (4) Jem, K. J., Pol, J. F., de Vos, S. In Plastics from Bacteria: Natural Functions and Applications; Springer: Heidelberg, 2010. (5) Dorgan, J. R.; Lehermeier, H. J.; Palade, L. I.; Cicero, J. A. Macromol. Symp. 2001, 55−66. (6) Chiu, W. M.; Chang, Y. A.; Kuo, H. Y.; Lin, M. H.; Wen, H. C. J. Appl. Polym. Sci. 2008, 3024−3030. (7) Lu, X.; Lv, X.; Sun, Z.; Zheng, Y. Eur. Polym. J. 2008, 2476−2481. (8) Pluta, M.; Galeski, A.; Alexandre, M.; Paul, M.-A.; Dubois, P. J. Appl. Polym. Sci. 2002, 1497−1506. (9) Sobkowicz, M. J.; Dorgan, J. R. J. Polym. Environ. 2008, 2, 131− 140. (10) Anastas, P. T., Warner, J. C. Green Chemistry: Theory and Practice; Oxford University Press: New York, 1998; p 148. (11) Braun, B.; Dorgan, J. R.; Knauss, D. M. J. Polym. Environ. 2006, 1, 49−58. (12) Huda, M. S., Mohanty, A. K., Misra, M., Drazal, L. T., Schut, E. ANTEC 2004 Plastics: Annual Technical Conference, SPE ANTEC: Detroit, Michigan, 2004; Conference Proceedings, Vol. 2, pp 1614− 1618. (13) Oksman, K.; Skrifvars, M.; Selin, J. F. Compos. Sci. Technol. 2003, 9, 1317−1324. (14) Habibi, Y.; Lucia, L. A.; Rojas, O. J. Chem. Rev. 2010, 3479− 3500. (15) Lynd, L. R.; Weimer, P. J.; van Zyl, W. H.; Pretorius, I. S. Microbiol. Mol. Biol. Rev. 2002, 3, 506−577. (16) Eichhorn, S. J. Soft Matter 2011, 303−315. (17) Rånby, B. G. Discuss. Faraday Soc. 1951, 158−164. (18) Šturcová, A.; Davies, G. R.; Eichhorn, S. J. Biomacromolecules 2005, 1055−1061. (19) Rusli, R.; Eichhorn, S. J. Appl. Phys. Lett. 2008, 0331111− 0331113. (20) Favier, V.; Canova, G.; Shrivastava, S.; Cavaille, J. Polym. Eng. Sci. 1997, 10, 1732−1739. (21) Angles, M. N.; Dufresne, A. Macromolecules 2000, 8344. (22) Mathew, A.; Dufresne, A. Biomacromolecules 2002, 609−617. (23) Morin, A.; Dufresne, A. Macromolecules 2002, 2190−2199. (24) Nair, K. G.; Dufresne, A. Biomacromolecules 2003, 666−674. (25) Paillet, M.; Dufresne, A. Macromolecules 2001, 19, 6527−6530. (26) Heux, L.; Chauve, G.; Bonini, C. Langmuir 2000, 8210−8212. (27) Oksman, K.; Mathew, A.; Bondeson, D.; Kvien, I. Compos. Sci. Technol. 2006, 2776−2784. (28) Goussé, C.; Chancy, H.; Excoffier, G.; Soubeyrand, L.; Fleury, E. Polymer 2002, 2645−2651. (29) Goffin, A.-L.; Raquez, J.-M.; Duquesne, E.; Siqueria, G.; Habibi, Y.; Dufresne, A.; Dubois, P. Biomacromolecules 2011, 2456−2465. (30) Araki, J.; Wada, M.; Kuga, S. Langmuir 2001, 21−27. (31) Majoinen, J.; Walther, A.; McKee, J. R.; Kontturi, E.; Aseyev, V.; Malho, J. M.; Ruokolainen, J.; Ikkala, O. Biomacromolecules 2011, 2997−3006. (32) Kloser, E.; Gray, D. G. Langmuir 2010, 16, 13450−13456. (33) Yuan, H.; Nishiyama, Y.; Wada, M.; Kuga, S. Biomacromolecules 2006, 696−700. (34) Tingaut, P.; Zimmermann, T.; Lopez-Suevos, F. Biomacromolecules 2010, 454−464. (35) Tomé, L. C.; Pinto, R. J. B.; Trovatti, E.; Freire, C. S. R.; Silvestre, A. J. D.; Neto, C. P.; Gandini, A. Green Chem. 2011, 419− 427. (36) Capadona, J. R.; Berg, v. d. O.; Capadona, L. A.; Schroeter, M.; Rowan, S. J.; Tyler, D. J.; Weder, C. Nat. Nanotechnol. 2007, 765−769. (37) Capadona, J. R.; Shanmuganatan, K.; Trittschuh, S.; Seidel, S.; Rowan, S. J.; Weder, C. Biomacromolecules 2009, 712−716.

Figure 8. Time required to obtain the maximum degree of crystallinity when the materials are quenched from the melt to 110 °C. The inset shows the crystallization half-time as a function of a-CNW concentration. (SP refers to materials prepared by solution polymerization, the number to the a-CNW loading level used to prepare the composite.).

the same conditions. Additionally, a slight increase in the achievable crystallinity to 58% compared with 49% for PLA is detected. This is noticeably higher than the values Goffin et al. reported for PLA/CNW-g-PLA materials with a filler loading of 4% (∼33%). This is not surprising given that many commercial grade PLAs are not macromolecules composed of the L-lactide stereoisomer but contain small amounts D-lactide, which decreases the achievable degree of crystallinity. At 15 wt % CNW, a further reduction of the half-time to 15s is observed. Similar to mechanical properties, the crystallization kinetics of nanocomposites prepared by bulk polymerization or solution blending do not exhibit the same degree of improvement.

■

CONCLUSIONS In this work, the synthesis and properties of supramolecular structures with molecular weights in the billions composed of CNWs isolated from biomass and PLA chains grafted directly to the surface via ring-opening polymerization are discussed. A well-designed approach to produce the materials is used to balance competing effects; blocking a fraction of surface hydroxyl groups available for ring-opening polymerization initiation leads to high-molecular-weight chains covalently attached to the filler surface, while maintaining sufficient interparticle contact for efficient stress transfer within a percolating network. Exceptional property improvements are the result; at only 10 wt % CNW, the HDT reaches 120 °C, and the material is significantly faster to crystallize. This versatile methodology allows control over the interfacial interactions between filler and matrix overcoming hydrophobic/hydrophilic incompatibilities and will help ultimately to optimize the final properties of polymer composites reinforced with fillers derived from renewable resources.

■

Article

AUTHOR INFORMATION

Notes

The authors declare no competing financial interest. 2018

dx.doi.org/10.1021/bm300149w | Biomacromolecules 2012, 13, 2013−2019

Biomacromolecules

Article

(38) Nakagaito, A. N.; Fujimura, A.; Sakai, T.; Hama, Y.; Yano, H. Compos. Sci. Technol. 2009, 1293−1297. (39) Kowalski, A.; Duda, A.; Penczek, S. Macromolecules 2000, 7359− 7370. (40) Witzke, D. R.; Narayan, R.; Kolstad, J. J. Macromolecules 1997, 7075−7085. (41) Kolstad, J. J., Witzke, D. R., Hartmann, M. H., Hall, E. S., Nangeroni, J. U.S. Patent 6,353,086, 2000. (42) Takemori, M. T. Polym. Eng. Sci. 1979, 15, 1104−1109. (43) Braun, B.; Dorgan, J. R. Biomacromolecules 2009, 2, 334−341. (44) Sobkowicz, M. J.; Braun, B.; Dorgan, J. R. J. Green Chem. 2009, 5, 680−682. (45) Adabajo, M. O.; Frost, R. L.; Kloprogge, J. T.; Kotot, S. Spectrochim. Acta, Part A 2006, 448−453. (46) Kister, G.; Cassanas, G.; Fabrègue, E.; Bardet, L. Eur. Polym. J. 1992, 10, 1273−1277. (47) Petersson, L.; Kvien, I.; Oksman, K. Compos. Sci. Technol. 2007, 2535−2544. (48) Suryanegara, L.; Nakagaito, A. N.; Yano, H. Compos. Sci. Technol. 2009, 1187−1192. (49) Takayanagi, M.; Uemura, S.; Minami, S. J. Polym. Sci. C 1964, 113−122. (50) Ouali, N.; Cavaillé, J. Y.; Perez, J. Plast., Rubber Compos. Process. Appl. 1991, 55−60. (51) Samir, M. A. S. A.; Alloin, F.; Dufresne, A. Biomacromolecules 2005, 612−626.

2019

dx.doi.org/10.1021/bm300149w | Biomacromolecules 2012, 13, 2013−2019