Cellulose nanocrystals: Preparation, surface modification, and application in nanocomposites

Cellulose nanocrystals were prepared from bacterial cellulose by hydrolysis with sulfuric acid. Crystals, prepared under different conditions, were characterized by sulfate content and thermal degradation behavior. Sulfate content increased with acid concentration, acid-to-cellulose ratio, and hydrolysis time. Increased sulfate content caused a decrease in degradation temperature and increased the char fraction. Activation energies were lower at higher sulfate content suggesting a catalytic effect on the degradation reactions. Cellulose crystals with increasingly hydrophobic surfaces were prepared by silylation with hexamethyldisilazane in formamide. Silicon content, morphology, and thermal degradation behavior were analyzed. Based on kinetics, a two-step reaction mechanism was suggested. Degradation temperatures in surface-trimethylsilylated crystals were higher than in unreacted crystals possibly due to suppression of dehydration reactions. Excessive silylation resulted in reduced X-ray crystallinity index and crystallite dimensions, and lower degradation temperatures. Composites of cellulose acetate butyrate (CAB) and both native and surface-trimethylsilylated cellulose nanocrystals were investigated by differential scanning calorimetry and dynamic mechanical analysis. Differences in matrix crystallinity and double-melting behavior in the composites indicating improved filler-matrix compatibility for the silylated fillers. The CAB glass transition temperature, T_g, and heat capacity increment at T _g were not affected by either of the fillers. Both storage modulus, E′, and loss modulus, E ″, in the composites increased with increasing filler content. The increase in E′ and E ″ was larger with native fillers than with silylated ones. From E″ data the interphase thickness was calculated using the strain-magnification concept. The tan δ peaks shifted to higher temperatures and reduced in magnitude with increasing filler content. The apparent activation energies, E_a, and entropies of activation, ΔS‡, for the glass transition were determined from the frequency dependence of the tan δ maximum. Both E_a and ΔS ‡ decreased with increasing filler content. The decrease in damping peak height, E_a, and ΔS ‡, was more pronounced for the silylated fillers implying stronger filler-matrix interactions.