Polymer Nanocomposites Reinforced by cellulose whiskers

Polymer Nanocomposites Reinforced by cellulose whiskers

(Parte 1 de 2)

Macromolecules 1996,28, 6365-6367 6365

Polymer Nanocomposites Reinforced by Cellulose Whiskers

V. Favier, H. Chanzy,* and J. Y. Cavaille*

Centre de Recherches sur les Macromolkcules Vkgktales, CNRS and Universitk Joseph Fourier, BP 53, 38041 Grenoble Cedex 09, France

Received April 7, 1995 Revised Manuscript Received June 20, 1995

There are numerous examples where animals or plants synthesize extracellular high-performance skel- etal biocomposites consisting of a matrix reinforced by fibrous biopolymers. lw3 Cellulose is a classical example of these reinforcing elements which occur as whiskerlike microfibrils that are biosynthesized and deposited in a continuous fa~hion.~-~ In many cases, this mode of biogenesis leads to crystalline microfibrils that are almost defect-free, with the consequence of axial physi- cal properties approaching those of perfect crystals. In the present study, we have attempted to mimic biocom- posites by blending cellulose whiskers from the mantles of tunicates with synthetic polymer latices. The films cast from such mixtures had a nanocomposite organiza- tion whose structure and mechanical properties are described in the present paper.

A batch of edible-grade tunicate Microcosmus fulca- tus, from the Mediterranean, was obtained from a local fish shop. These sea animals had an overall diameter between 5 and 10 cm with a 1 cm thick cellulose tunic. After anesthetizing with chloroform, the animals were gutted and their tunic was cut into small fragments that were deproteinized by three successive bleaching treat-

ments, following the method of Wise et a1.6 The bleached mantles (the "tunicin") were then disintegrated in water, first with a Waring blender (at a concentration of 5% by weight) and then with 15 passes through a Gaulin laboratory homogenizer operated at 400 bar (at a concentration of 1% by weight). The resulting aqueous tunicin suspension was mixed with HzSO4 to reach a final acidwater concentration of 5% weight fraction. Hydrolysis conditions were 60 "C for 20 min under strong stirring. A dispersion of cellulose whiskers resulted. After sonication, the suspension was neutral- ized and washed by dialysis. It did not sediment or flocculate as a consequence of surface sulfate groups created during the sulfuric acid treatment.' When concentrated by evaporation, the suspensions displayed typical liquid crystal characteristics.8

The suspensions of cellulose whiskers were homoge- neously mixed with polymer latices (Tg = 0 "C) resulting from the copolymerization of styrene (35% wlw), butyl acrylate (65% w/w), and a small amount of acrylic acid. The suspensions were poured into poly(tetrafluor0eth- ylene) molds and allowed to dry slowly for 1 month at room temperature. Homogeneous and bubble-free 2 m thick films resulted. These films, cut into strips 35 m long and 6 m wide, were analyzed with a DMTA

Metravib SA Mecanalyseur operating with a forced oscillation pendulum. The experiments were achieved at a fixed frequency of 0.1 Hz and in a temperature range 200-500 K.

Electron microscopy and electron diffraction analysis on cellulose whiskers was achieved with a Philips EM400T operated at 120 kV. For this, cellulose whisker suspensions were deposited on carbon-coated grids and observed under low-dose conditions, using a 20 pm

024-9297/95/2228-6365$09.0/0 objective aperture that gave bright-field diffraction contrast images. Microdiffraction on individual cel- lulose whiskers was performed by the diffraction method of Riecke and Ruska9 with a probe of 50 nm that required a C2 condenser aperture of 5 pm. Observations on the reinforced films were made on ultrathin sections obtained by cryosectioning. These sections were ob- served in bright-field diffraction contrast with a Philips CM200 CRY0 operated at 200 kV and equipped with a 15 pm objective aperture. The images obtained under low-dose conditions revealed (in black) the cellulose crystallites within the clear polymer matrix. An image of the latex particles was made under low-dose condi- tions by depositing drops of latex suspensions on carbon- coated grids.

A typical preparation of tunicin crystals is shown in

Figure 1A. This sample consists of parallelepiped rods with lengths ranging from 100'nm to several microme- ters for widths on the order of 10-20 nm. Upon testing by a microelectron diffraction technique (inset), each element gave a spot difiactogram that corresponded to a section of the reciprocal lattice of cellulose I@ (here the a*c* section) and persisted when the electron probe was scanned along a given rod. As the diffractogram indicated that the cellulose chain axis is along the long dimension of the rods, each rod is therefore a whisker- like tunicin crystal with no apparent defect. Figure 1B is a low-dose image of the latex particles that were used in this study. Each particle has a diameter of around 150 nm. Some of these particles are individual and have kept their initial spherical shapes. Most of them, however, are coalesced together due to the conditions of sample preparation, viz., room temperature which is 20 K above the Tg of the latex.

Figure 2 corresponds to an ultrathin cry0 cross section of a nanocomposite film containing 6% (w/w) of tunicin within the latex. In this image, recorded under low- dose and diffraction contrast conditions, the cellulose crystals are in black within the clear polymer matrix. This image reveals that the cellulose whiskers are distributed throughout the structure, without segrega- tion or association.

When reinforced by a small percentage of tunicin whiskers, the polymer films showed improved mechan- ical properties which were particularly striking when the films were heated above the glass transition of the polymer. This is illustrated in Figure 3, where part A shows a plot of the shear modulus G as a function of temperature for various whiskerflatex compositions ranging from 0 to 14% (wlw). The curve corresponding to the pure matrix is typical for a thermoplastic: below

Tg, G remained constant at around 1 GPa and dropped rapidly to 1 MPa during the crossing of the glass- rubber transition temperature. Above this temperature, it behaved as a viscous liquid, with G decreasing rapidly with increasing temperature. The films that contained cellulose whiskers had a slight increase in their G value below Tg, but the drop in G value above Tg was dramatically reduced: only from 1 to 0.1 GPa for a film reinforced by 6% (wlw) whiskers. Above Tg, the rein- forced films behaved as rubbers as their G value stayed constant over a wide temperature range. This is il- lustrated in Figure 3B that corresponds to a film reinforced by 6% (w/w) whiskers for which G keeps a value of 0.1 GPa all the way to 500 K, a temperature at which cellulose starts to decompose.

The variation of the shear modulus G, taken at 325 K (Le., 50 K above Tg) as a function of the whisker

0 1995 American Chemical Society

8566 Communications to the Editor

Figure 1. (A) Electron micrograph of rcdlike cellulose mi- crystals extracted from the mantle of Microcosmus fulcatus. Scale bar: 0.5 pm. Inset: typical electron diffractogram recorded on one microcrystal. oriented with its axis vertically. The indexation corresponds to the unit cell of cellulose I5 defined by Sugiyama et QL** (B) Low-dose image of the latices that were used for this work. Scale bar: 0.5 pm.

Figure 2. Low-dose images of a cryosection of a nanomm- posite film consisting of 6% (w/w) tunicin whiskers in a latex as in Figure 1B. The image, recorded under difiaction con- trast bright-field conditions, revealed in black the cellulose whiskers. Scale bar: 0.5 pm.

content, is plotted in Figure 4. The observed values of

G (Figure 4) are much higher than those (dashed line in Figure 4) predicted from a classical mean-field mechanical model developed for short-fiber composite. In such an approach, following Halpin and Kardos," the modulus and the geometry of the fibers are accounted for, but one assumes that there is no interaction between the fibers. In particular, the mean-field model is based on the concept that a material made of short fibers, homogeneously dispersed in a continuous matrix, is mechanically equivalent to a superposition of four plies. Within each ply, the fibers are parallel to one another and the mutual orientation of the plies is O", +45', +go', and -45". The mechanical properties of each ply are derived from the miemmechanic equations of Halpin-Tsai."

Macmnwkules, Vd. 28, No. 18.1995 10

Y' B 6

- 200 500 m m

T.np IK)

Figure 3. Reinforcement effect as a function of temperatm below and above the latex T (A) Log of the shear modulus (Pa) for composites reinforced by weight fractions of cellulose whiskers from 0 to 14%, as a function of temperature in the range 200-350 K. (B) As in A but for temperatures up to 500 K and for a specimen reinforeed by 6% (w/w) cellulose whiskers. This specimen is compared with a control sample that does not contain any whisker.

- 0 o,oa a04 0.w om *1 rrN.L.nooM

Figure 4. Plot of the log of the shear modulus G (Pa) at 325 K as a function of the volume fraction of cellulose whiskers.

Comparison between the experimental (black dots) and cal- culated data with two different mechanical models: a mean- field model (dashed line) and a percolation model (continuous line).

As seen in Figure 4, the mean-field approach is obviously not able to account for the observed data. In order to explain the unusually high G values of the reinforced films, one needs to invoke (i) a strong interaction between the whiskers and (i) a percolation effed. The influence of such an effect on the mechanical properties of the films can be calculated following the method of Ouali et aL'* in their adaptation of the percolation concept to the classical parallel-series model of Takayanagi et al.I3 The shear modulus of the composite is then given by the equation:

(1 - 2V + VXJG'&', + (1 - X&G:

(1 - X,)G, + aU, - V)G, G= where the subscripts s and r refer respectively to the soft and rigid phases; X is the volume fraction of

Macromolecules, Vol. 28, No. 18, 1995 whiskers. In the Takayanagi et al. model,13 q is an adjustable parameter. In the Ouali et al. model12 which was used here, q corresponds to the volume fraction of the percolating rigid phase. It can be calculated with a simple prediction based on the percolation concept:12

Communications to the Editor 6367

In this paper, laboratory-scale nanocomposite struc- tures were obtained by reinforcing latices with whiskers extracted from rather rare cellulose samples. There are many more common native cellulose or chitin specimens that could also be used to produce similar reinforcement. Both polysaccharides can provide systems of microfibrils with an axial ratio surpassing the above whiskers. These elements could be used directly as reinforcing agents provided that they could be separated from one another and dispersed efficiently within polymer ma- trices., If this could be achieved, one would expect an even better reinforcing effect at a much lower cellulose concentration.

Acknowledgment. The authors are indebted to P.

Smith for stimulating discussions. The help of B. Ernst from Elf Atochem for providing the latices is also acknowledged. This work was supported by the French Ministry of Agriculture.

References and Notes (1) Neville, A. C. Biology of Fibrous Comnosites. Develonment q=o Xr5Xc where b is the percolation exponent. According to several studies based on the percolation concepts,14J5 b takes the value of 0.4 in a three-dimensional system. X, is the percolation threshold: here X,= 1% (v/v) that corresponds to 1.5% (w/w).

The role of percolation of cellulose fibers in paper making is well documented.16-17 It is, in particular, established that the high mechanical properties of a paper sheet result from the hydrogen-bonding forces that hold the percolating network of the fibers. We believe that it is also the hydrogen-bonding system that is responsible for the unusual mechanical properties of our system when the percolation threshold is reached. A calculated curve based on the percolation theory (solid line in Figure 4) follows the dashed line up to a whisker volume fraction of 1%. Above this percentage, the solid line curve precisely fits the observed G values. We believe that it is also the whisker network that is responsible for the stabilization of G over a large temperature range above Tg. It will be only when the whiskers start to decompose at around 500 K that this stabilization will disappear, inducing a catastrophic decrease of G.

There are a number of literature data that deal with the reinforcement of synthetic polymers with fibrous elements based on cellulose.18-20 Unlike our nanocom- posites, these reinforced structures do not contain homogeneous dispersions of individual cellulose crystals of high axial ratio. Large quantities of cellulosic fillers are therefore required to reach properties comparable to ours.19 Another approach to the reinforcement of plastics by cellulose is that developed by the group of Nishio,21 who have prepared networks of regenerated cellulose gels interpenetrated with various water-soluble monomers that are subsequently polymerized. They obtain a significant reinforcement of their composites, even at a relatively low concentration in cellulose. However, their regenerated cellulose is poorly crystal- line and does not allow a thermally stable plateau to be reached as that shown in Figure 3B. In addition, their films are water sensitive as opposed to ours which are hydrophobic.

beyond the Cell Membrane; Cambridge University Press: Cambridge, U.K., 1993. (2) Preston, R. D. The Physical Biology of Plant Cell Walls;

Chapman and Hall: London, 1967. (3) Atkins, E. D. T.: Keller. A. Structure ofFibrous Bionohmers.

Colston Paper No. 26;'Buttenvorths~ London, 1675." (4) Itoh, T.; Brown, M., Jr. Planta 1984, 160, 372. (5) Benziman, M.; Haigler, C. H.; Brown, R. M., Jr.; White, A.;

Cooper, K. M. Proc. Natl. Acad. Sci. U.S.A. 1980, 7, 6678. Wise. L. E.: Mumhv. M.: D'Addiecco. A. A. Pan Trade J", I

1946; 122, 35.

Marchessault, R. H.: Morehead. F. F.: Walter. N. M. Nature 1959, 184, 632. ' Revol, J. F.; Bradford, H.; Giasson, J.; Marchessault, R. H.; Gray, D. G. Znt. J. Biol. Macromol. 1992, 14, 170. Riecke, W. D.; Ruska, E. In Electron Microscopy; Ueda, R., Ed.; Maruzen Co. Ltd.: Tokyo, 1966; p 19 and 20. Halpin, J. C.; Kardos, J. L. J. Appl. Phys. 1972, 43, 2235.

Tsai, S. W.; Halpin, J. C.; Pagano, N. J. Composite Materials Workshop; Technomic: Stanford, CT, 1969. Ouali, N.; Cavaille, J. Y.; Perez, J. Plast., Rubber Comp. Process. ADW~. 1991. 16. 5. Takayanaz, M.; Uemura, S.; Minami, S. J. Polym. Sci., Part

C 1964. 5. 113. Stauffer, D. Introduction to Percolation Theory; Taylor and Francis: London and Philadelphia, 1985. de Gennes, P.-G. Scaling Concepts in Polymer Physics; Cornel1 University Press: Ithaca, NY, 1979. Batten, G. L., Jr.; Nissan, A. H. TAPPZ 1987, 70, 119. Nissan, A. H.; Batten, G. L., Jr. TAPPZ 1987, 70, 128. Felix, J. M.; Gatenholm, P. J. Appl. Polym. Sci. 1991, 42,

(Parte 1 de 2)