A recently developed statistical-mechanical model is
applied systematically
to estimate the fraction of tie-molecules (polymer chains linking
different crystals directly or via entanglements) in semicrystalline
polyethylene (PE) samples. The amorphous domains of the polymer are
divided into constrained interlamellar domains and “free”
outer-lamellar domains. A set of model parameters is assigned to each
sample by correlating previous experimental measurements and minimizing
the difference between the predicted solubility of pure hydrocarbons
in the sample and the experimental values. We show that the sorption
isotherms of multiple pure fluids in each sample can be described
by a single parameter set, proving that the polymer–solute
interactions (described accurately by the SAFT-γ Mie EoS) are
decoupled from the sample-specific properties of the polymer. We find
that ∼30% of the crystalline stems in the lamellae of PE are
connected to tie-molecules, within the bounds suggested by previous
theoretical and computational work. The transferability of the sample-specific
parameters is assessed by predicting cosolubility effects and solubility
at different temperatures, leading to good agreement with experimental
data.