Abstract
The understanding of solvent ingress into polymers is of vital importance in a plethora of applications that embrace such diverse technologies as dental resins and food systems. Two limiting regimes of small molecule diffusion in polymers are widely acknowledged: Fickian and Case II. Fickian diffusion is associated with solvent uptake proceeding with the square root of time and with smooth concentration profiles. Case II ingress is characterised by uptake that is linear with time and by concentration profiles with sharp fronts. This thesis describes new insight into both these transport mechanisms. Binary mixtures of good (methyl ethyl ketone) and bad (ethanol) solvents ingressing polystyrene with Fickian dynamics are investigated. Using both 2H magnetic resonance imaging (MRI) with selectively deuterated solvents and 13C-1H cyclic cross-polarisation chemically selective imaging of normal solvents, the individual components are separately measured. The two solvents are found to ingress together but they have different spatial concentration profiles. These results are explained in terms of a simple model. In a further study using both MRI and ellipsometry, no polymer molecular weight dependence is observed for Fickian solvent ingress. Powerful evidence of a new rate-limiting step to give Case II diffusion is demonstrated. We show that the rate-limiting step can be the solvent flux at the polymer surface as well as the visco-elastic polymer swelling. Numerical simulations of a simple phenomenological model demonstrating these effects are presented. They are supported by experimental measurements of liquid and vapour toluene ingress into polystyrene using MRI and stray field imaging. One problem with ellipsometry for measuring Fickian ingress is that it is difficult to unambiguously fit the and data to the smooth refractive index profile through the swelling sample. Two methods have been implemented to overcome this problem: a Born approximation and a Bayesian inference technique. The latter has shown that ellipsometry can now, in theory, be described as a model independent technique.