Interactions between hydroxypropylcelluloses and vapour/liquid water.

Understanding of the uptake of water vapour or liquid water by cellulose-based polymers is important because of the influence of these processes on many of the biologically or technologically relevant properties of these polymers. In this work we studied these processes in the cases of twelve hydroxypropylcelluloses with low or medium-high degrees of substitution (L-HPCs and HPCs, respectively), characterization of which showed significant differences in structural and physical parameters (substitution pattern, crystallinity, particle size, specific surface area, and intraparticular porosity). Water vapour sorption-desorption isotherms determined to characterize the uptake of water vapour were fitted well by the Young-Nelson model, the optimized parameters of which indicated that at all relative humidities the capacity to bind water vapour as a surface monolayer is greater for HPCs than L-HPCs, but the capacity to absorb water vapour internally is greater for L-HPCs than HPCs. Guggenheim-Anderson-deBoer (GAB) models fitted the sorption-desorption isotherms less well. Differential scanning calorimetry (DSC) experiments showed all sorbed water vapour to be held as non-freezing water. Isoperibol microcalorimetry experiments carried out to investigate interactions with liquid water showed enthalpies of hydration/dissolution of between -62.86 and -71.35 J g(-1) for L-HPCs and between -82.95 and -99.80 J g(-1) for HPCs, and DSC showed average numbers of non-freezing water molecules per polymer repeat unit of 2.65-4.19 for L-HPCs and 18.10-22.42 for HPCs. DSC characterization of the kinetics of the water uptake by 10 mg compacts obtained by direct compression of hydroxypropylcelluloses showed faster uptake by L-HPC compacts than by HPC compacts, among which there were significant differences in capacity for diffusive uptake. The explanations of the above differences in terms of the different substituent contents, particle sizes and porosities of the HPCs is supported by multiple linear regression analyses.