The aim of this thesis has been to assess the potential of branched polymers as novel exfoliants for layered silicates and their application in polyurethanes. Three different types of –OH functional branched polymers have been studied: dendrimers, hyperbranched polymers (HBPs) and star branched polymers (SBPs), with a range of structures and molecular weights. The dendrimers and their HBP analogues were synthesized in-house, while the other HBPs and SBPs were obtained from commercial suppliers. Various types of montmorillonite (MMT) layered aluminosilicate clays have been used as modifiers. The different branched polymers were either melt processed or solution processed with the MMT, using water and THF as dispersants. Whether intercalated or exfoliated nanocomposites were obtained depended on the choice of MMT, the presence of solvent during processing as well as the characteristics of the interface between the polymer and the MMT, with exfoliation being favored when the polymer-MMT interactions were strong. Very high degrees of exfoliation were obtained from dendrimers and HBPs processed in water at loadings as high as 20 wt% unmodified Na+MMT, with intercalation only becoming dominant at higher loadings. The MMT layer spacing in the intercalated HBP/Na+MMT nanocomposites depended directly on the pseudo-generation number, that is, on the size of the HBP, at intermediate loadings, consistent with adsorption of a monolayer of unperturbed HBP molecules at the layer surface in aqueous suspension. Dispersions of the HBPs with a range of organically modified MMTs in THF, on the other hand, showed mainly intercalation on drying, with layer spacings of the order of 3.8 nm at intermediate clay contents, which were apparently independent of the pseudo-generation. The difference between the final MMT layer spacing and its value prior to mixing was found to increase significantly with the polarity of the organic modifier. The contrasting behavior of Na+MMT and the organically modified MMTs was partly attributed to the tendency of Na+MMT to exfoliate in aqueous dispersion. The HBPs and dendrimers are then able to coat the individual MMT layers and stabilize the exfoliated structure during drying. The organically modified MMTs, on the other hand, did not exfoliate in THF dispersions and intercalation consequently dominated. Predominantly intercalated nanocomposites were also obtained with the SBPs, although increasing degrees of exfoliation were observed as the strength of the interactions between the SBP and the clay increased, as a result of systematic modification of the polarity of the SBP. The rheological properties of a range of exfoliated and unexfoliated highly branched polymer/MMT nanocomposites were investigated in small strain oscillating shear. The exfoliated nanocomposites showed a sharp increase in shear viscosity at very low MMT contents, accompanied by a transition from Newtonian behavior to solid-like behavior, characterized in terms of a limiting shear modulus at low frequencies and a limiting viscosity at high frequencies, both of which were strongly dependent on MMT content. Unexfoliated nanocomposites showed a more gradual increase in viscosity with MMT content, although they showed qualitatively similar behavior to the exfoliated nanocomposites at sufficiently high loadings. The rheological behavior was analyzed in terms of models for conventional filled polymers, leading to estimates of the effective particle aspect ratio that were approximately consistent with direct morphological observations by transmission electron microscopy.To investigate the influence of the MMT on solid state properties, the HBP/MMT and SBP/MMT nanocomposites were incorporated into model polyurethane formulations. In the case of the HBP/MMT, a reactive low molar mass polyol or an un-reactive solvent (THF) was used to reduce the viscosity sufficiently for room temperature processing. Exfoliated HBP/Na+MMT led to a 60% increase in the rubbery plateau modulus relative to that of the corresponding unfilled matrix at clay contents as low as 0.5 vol%, whereas more modest relative increases were observed with unexfoliated or partly exfoliated MMT. The same materials also showed a 2-fold increase in stress at break and elongation at break in the presence of 0.5 vol% exfoliated Na+MMT, although no improvement was observed for the corresponding intercalated polyurethane nanocomposites obtained in the absence of HBP. Much higher exfoliated clay contents led to processing problems owing to their high viscosities. Polyurethanes containing unexfoliated MMT, such as obtained by using SBPs rather than HBPs as the matrix, could be prepared up to relatively high MMT contents, but the improvements in mechanical properties were modest at any given MMT content. The semi-empirical Halpin-Tsai model and Mori-Tanaka's average stress theory, which are both based on a classical approach to particle reinforcement, accounted well for the evolution of the mechanical properties of the polyurethanes as the MMT content was varied and they led to estimates of the particle aspect ratio that correlated well with those inferred from the rheological models and direct observations. This in turn allowed correlations to be established between the degree of exfoliation, the shear viscosity of the nanocomposite precursors and the mechanical properties of the final polyurethanes.