Analysis of phthalate plasticizer migration from PVDC packaging materials to food simulants using molecular dynamics simulations and artificial neural network

Based on the experimental data of gas chromatography-mass spectrometry, an improved artificial neural network was first established to predict the migration of 2-ethylhexyl phthalate (DEHP) plasticizer from poly(vinylidene chloride) (PVDC) into food simulants (ie., heptane, ethanol and water). The sensitivity analysis indicated that temperature acted as a crucial factor influencing the migration values of DEHP. Then, a combined experimental and molecular dynamic (MD) simulation was performed to understand the migration kinetics and the mechanism of DEHP. Hansen solubility parameters of three component ( δ d , δ p , δ h ) were simplified into two-component solubility parameters ( δ vdW , δ e ), and the tuple was successfully applied to describe the interactions between PVDC and food simulants. The MD results showed that high interaction energy and fractional free volume in PVDC/DEHP/food simulant systems accelerated the migration of DEHP. These fundamental studies would provide significant insights into the migration of environmental contaminants.