Processing, structure, and properties of cellulose based nanocomposites

Abstract

On account of their abundance, renewable nature, biodegradability, low toxicity, high strength and stiffness, cellulose nanocrystals (CNCs) have emerged as an attractive nanofiller for polymeric materials. Their usually high aspect ratio, which depends on the source and method of extraction, and the large specific surface area are also important features. While on a laboratory scale nanocomposites made with CNCs have demonstrated significant improvement in mechanical properties over the neat polymers, the usefulness of CNCs as reinforcing filler on a commercial scale depends on how the lab scale results can be translated using technologically viable processes. Interestingly, relatively few systematic studies regarding the processing-structure-properties of nanocomposites have been conducted, and the understanding of technologically viable melt-processing methods, which permit upscale in the production of such nanocomposites is still very limited. The experimental research program presented in this thesis first sought to identify the effect of the melt-processing conditions on the morphology of CNCs in the final nanocomposites. This was achieved by using homogenously mixed, solution-cast poly(vinyl acetate) (PVAc)/CNC nanocomposites as reference materials and further melt-processing them in a low-shear roller-blade mixer (RBM) and a high-shear twin-screw extruder (TSE). The results reveal a relationship between processing methods (mainly involving shear force), CNC degradation, and mechanical properties of the materials. The high shear mixing environment using a TSE led to a significant degradation of the CNC length, and significant reduction of the mechanical reinforcement, while compounding with a lower-shear RBM mixer resulted in materials that are comparable to the solution cast reference samples. The results were used in a second step to explore the direct mixing of PVAc and CNCs in a RBM mixer and the results confirm that it is possible to create materials that are much stiffer than the neat matrix, but which did not quite reach the levels of the solution cast reference series. Mixing of unmodified, polar CNCs with hydrophobic polymer matrices such as low density polyethylene (LDPE), represents a particularly significant challenge because the incompatibility between the components stifles adequate CNC dispersion in the composite. When CNCs were directly melt-mixed with LDPE using either a RBM or TSE, significant aggregation of CNCs was observed. In order to develop a standard reference system containing LDPE/CNC, a previously established template process was successfully adapted to create homogenous LDPE/CNC nanocomposites with unmodified CNCs and without compatibilizer. CNC organogels were prepared by a solvent exchange process. Impregnation of the resulting organogel with a hot LDPE solution resulted in homogenous nanocomposites that display significant increase in mechanical properties (a 236% increase of the storage modulus and a 314% increase of the tensile strength were observed for an LDPE/CNC nanocomposite with 7.6% v/v CNCs) compared to the neat LDPE. It was also possible to reprocess these nanocomposites and dilute them with LDPE using conventional melt-processing techniques and retain mechanical characteristics. While the template process used to create reference LDPE/CNC nanocomposites in its present form may not directly be scalable for technological exploitation, the fact that the high level of dispersion is largely maintained upon compression molding films and also reprocessing and “diluting” such nanocomposites bodes well for the development of alternative mixing approaches that could be readily scalable. To explore the upscaling possibility and maximize the dispersion of CNCs in LDPE without chemical modification and/or relying on a compatibilizer, an organic solvent-free two-step process was developed. The first step involves pre-mixing an aqueous dispersion of the CNCs with LDPE powder to create slurry in water and evaporating the solvent. This mixture was melt-mixed using a conventional melt-processing method and the resulting product was compression molded into nanocomposite films. This approach permits both modified and unmodified CNCs to be pre-mixed with LDPE prior to melt-mixing and substantially improves their overall dispersion in the polymer matrix. The mechanical properties of nanocomposites thus obtained showed a significant increase of the storage modulus by a factor of 2.5 for 15% w/w nanocomposite compared to neat LDPE at room temperature. For several filler systems, the combination of filler particles having dissimilar shapes has been shown to exhibit synergistic effects. Extending this understanding to account for the different reinforcing behavior of different types of CNCs in a nanocomposite system is interesting. Based on the hypothesis that the different length distribution of nanorods exhibits a rich phase and structural behavior and in principle can lead to the formation of cosupporting networks, a notion that it could also be an effective means to manipulate the final material properties can be anticipated. In order to successfully test this hypothesis, a modelling approach was developed to predict the mechanical reinforcement achieved through a bimodal length distribution of nanorods. For the first time, this generalization approach addresses a more realistic prediction of the critical percolation volume for a reallife situation with bimodal length distribution of nanofillers in the nanocomposite system. The origin and nature of CNCs exert a significant influence on their reinforcing capability and the mechanical properties of the nanocomposites formed with a polymer of interest. CNCs with higher aspect ratio A such as those isolated from tunicates (tCNCs, A = ca. 76) reach the percolation threshold at a lower concentration and have higher reinforcing capability than CNCs with lower aspect ratio, for example those isolated from cotton (cCNCs, A = ca. 11). To realize/validate the earlier developed theoretical modelling, we investigated nanocomposites based on a poly(ethylene oxide-co-epichlorohydrin) (EO-EPI) matrix polymer reinforced with tCNCs, cCNCs, and mixtures thereof. The TEM images as well as the obtained mechanical reinforcement suggests that tCNCs and cCNCs form mixed percolating structures, reduce the aggregation of smaller cCNCs and result in the synergistic mechanical reinforcement. Considering the broad applicability, the mixture system offers an intriguing compromise between high aspect ratio, availability and mechanical reinforcement. In summary, this experimental thesis presents different aspects of processing of cellulose based nanocomposites and summarizes a systematic investigation of the compositionprocessing- structure-property relationship of these materials. Identifying the melt-processing conditions and the potential of CNCs as reinforcing filler for technologically relevant polymers is established with methodical studies. In addition, fundamental studies of the composition-structure-property for CNC based nanocomposites are established with the help of theoretical modelling and experimental evidences.