The glass transition of thin polymer films: some questions, and a possible answer

A simple and predictive model is put forward explaining the experimentally observed substantial shift of the glass transition temperature, T-g, of sufficiently thin polymer films. It focuses on the limit of small molecular weight, where geometrical 'finite size' effects on the chain conformation can be ruled out. The model is based on the idea that the polymer freezes due to memory effects in the viscoelastic eigenmodes of the film, which are affected by the proximity of the boundaries. The elastic modulus of the polymer at the glass transition turns out to be the only fitting parameter. Quantitative agreement is obtained with our experimental results on short chain polystyrene (M-W = 2 kg/mol), as well as with earlier results obtained with larger molecules. Furthermore, the model naturally accounts for the weak dependence of the shift of Tg upon the molecular weight. It explains why supported films must be thinner than free standing ones to yield the same shift, and why the latter depends upon the chemical properties of the substrate. Generalizations for arbitrary experimental geometries are straightforward.