In a mixture of two components, the component with the lower surface energy will segregate preferentially to the surface. In most cases, this will be generally one monolayer, which, in the polymer case, can be quite large (because the polymer chains are quite large). However, if the two components are immiscible, phase separation will occur, and the surface will direct that phase separation into the film.
Very close to the boundary of phase separation, the balance between entropic mixing and thermodynamic repulsion is very fine, and spontaneous phase separation will be quite slow (so phase separation in the bulk will not be observable. This does not apply at the surface, which nucleates phase separation, and a wetting layer will start to grow. This layer starts as a series of droplets at the surface, which grow diffusively as the square root of time (think of the units of a diffusion coefficient to see why this is so). (Top figure.)
For more immiscible mixtures, there will be phase separation in the bulk of the film. This will compete with the wetting layer for material, essentially slowing wetting layer growth down. The square root growth law no longer applies, and a logarithmic growth ensues. (Middle figure.)
For very immiscible systems the wetting layer is unstable and breaks up, just like in the bulk. The growth law (1/3 power) is the same as that in the bulk. (Bottom figure.) The transition to surface droplets from a uniform wetting layer is known as a wetting transition. In experiments on a partially miscible polymer blend, we observed all of the above .
Above right: Diagram showing different wetting regimes, between two miscible polymers (top), intermediate state (middle) and two immiscible polymers (bottom).
 M. Geoghegan, H. Ermer, G. Jüngst, G. Krausch, and R. Brenn Phys. Rev. E 62 940-50 (2000).