We have developed a simple "breakthough time" method for determining across-the-plane translational diffusion coefficients of dye molecules in thin and ultrathin polymer films. this method relies on fluorescence resonance energy transfer in conjuction with the production of trilayer polymer films in which the bottom layer at the substrate interface contains trace levels of a fluorescence "acceptor" dye, thi middle layer contains no dye, and the top layer contains traces levels of a fluorescence "donor" dye covalently attached to the polymer. Diffusion measurements are done at temperatures near the glass transition temperature, resulting in only the diffusion of the acceptor dye across the trilayer film. When an acceptor dye comes with 2-3 nm of a donor dye, energy transfer from the excited-state donor dye to the acceptor dye results. By observing the time scale of at which the donor fluorescence intensity decreases and noting the thickness of the middle layer, it is possible to calculate the Fickian diffusion coefficient which is proportional to the square of the thickness of the middle layer divided by the breakthrough time.

This simple experimental method has been applied to measurement of the across-the-plane translational diffusion coefficients of two dyes, decacyclene and Disperse Red 1, in polystyrene at a temperature 3 K above the bulk Tg of PS. For trilayer film thicknesses in excess of 200 nm, both dyes exhibit diffusion coefficients that are independent of thickness, with the decacylene dye having a diffusion coefficient two orders of magnitude smaller than the Disperse Red 1 dye. This results from the fact that the decacylene dye is much larger than Disperse Red 1, meaning that its translational motion cannot couple to as much of the heterogeneous cooperative segmental mobility of the polymer as can Disperse Red 1. Most stunning is the fact that when the trilayer film is decreased below 200 nm, there is only a weak reduction of the decacylene diffusion coefficient while that of Disperse Red 1 is reduced by at least an order of magnitude. This effect can be explained by the effect of confinement on the distribution of heterogeneous dynamics in polymer films. In particular, the translational diffusion coefficient of Disperse Red 1 is strongly indicative of the fast-time side of the distribution of cooperative segmental relaxation times which becomes truncated with confinement of thin films.