Specific migration testing - a comparison of experimental data and values predicted by a migration model

Migration modelling is a cost-efficient alternative to experimental testing when allowed by legislation. For migration modelling, two fundamental pieces of information are needed: the diffusion coefficient, and the coefficient describing a migrant's partitioning between polymer and food. For the latter, a worst case approach can easily be implemented assuming a high solubility of the migrant in the food. The diffusion coefficient however, depends on several factors, e.g. the molecular weight of the migrant, the temperature, and the matrix (polymer). For the first time, diffusion coefficients were determined in a small scale interlaboratory comparison study, meaning that the data established here can be considered the most reliable values relevant for this purpose available to date. A validation of the migration modelling approach itself was possible as it was possible to compare reliable experimental migration data on one side and predicted migration data on the other. The experimental and predicted data of materials fit together quite well. Of the 28 migration values predicted, 93% were in the range of ±40% compared to the experimental migration values. Deviations observed are likely to originate from analytical variations or in different solubilities of the migrants in oil (applied during specific migration testing) and food simulants (applied during diffusion testing).