It has been well known for over a decade that confinement at the nanoscale of low molecular weight and polymeric glass formers can lead to major changes in dynamics associated with the glass transition temperature (Tg). In particular, studies using a supported polymer film geometry have shown that major reductions in average Tg across the films can be observed with confinement for polymers that do not exhibit significant attractive interactions with the substrate (e.g., polystyrene supported on silica). In contrast, when attractive polymer-substrate interactions are present (e.g., hydrogen bonding between hydroxyl groups naturally on the surface of silica and nitrogen atoms in poly(2-vinyl pyridine)), confinement of polymer films can lead to major increases in average Tg relative to the bulk Tg value.

In 2003, a multilayer fluorescence method was developed (Ellison and Torkelson, Nature Materials 2002, 2, 695) which allowed for determination of Tg values in both free-surface layers and substrate layers of polymer films supported on silica and other substrates. This method provided the first demonstrations of how Tg is reduced as a function of confinement within a free surface layer or is increased as a function of confinement within a substrate interfacial layer when attractive interactions are present.

Here we extend this approach to bilayer or multilayer polymer films that are made up of different polymer species (e.g., polystyrene (PS) surface layer sitting atop a poly(2-vinyl pyridine) (P2VP) layer) and in which only one layer has a fluorescence label that reports Tg or physical aging response. This study reveals that for sufficiently thin substrate or free-surface layers (e.g., ~ 10 nm thick), the Tg of the interfacial layer can strongly on the nature and thickness of the adjoining species. For example, when a 14-nm-thick PS surface layer is atop a P2VP layer that is at least 45 nm thick, the Tg of the PS surface layer is within error equal to that of bulk PS. However, when a 14-nm-thick PS surface layer is atop a PS layer that is at least 45 nm thick, the Tg of the PS surface layer is ~32 K below that of bulk PS. The complexity of this Tg response is made evident from the fact that the Tg of the PS surface layer can depend on the thickness of the underlying substrate layer, meaning that both the type and amount of adjoining material can affect the Tg response of the interfacial layer. That is, when the underlying P2VP layer in sufficiently thin and the attractive substrate interactions are broken, the PS surface layer exhibits a major reduction in Tg relative to bulk PS.

These results indicate that the cooperative segmental dynamics within an ultrathin layer of one polymer in interfacial contact with a second polymer species can have its dynamics very strongly perturbed by the dynamics of the second polymer, even though the interfacial layer between the two polymers may be only 3 nm in thickness. Furthermore, the perturbation to glass transition dynamics caused by the adjoining polymer can be as strong or stronger than the perturbations caused by polymer-air and attractive-substrate interfaces. The implications of these results for achieving better understanding of the heterogeneous dynamics of glass forming polymers and potential novel properties of nanostructured polymer blends will be discussed.