A Bayesian Inference Approach to Accurately Fitting the Glass Transition Temperature in Thin Polymer Films.

Both fitting methods find that very thin films with h ≈ 9 nm exhibit larger sample-to-sample variability in T g ( h ). Interestingly, the trend in the transition width w ( h ) that is reliably obtained from the Bayesian fits shows an unexpected trend below h = 20 nm that strongly indicate a persistent decrease in the transition width.

also in 73 other papers