| Abstract: | The migration of three additives from polypropylene samples into a series of five liquid food simulants was investigated at 60 °C in a repeated migration regime. This regime was simulated by 20 migration periods each of 8 h and using initially additive free food simulants. Between these migration periods the bare samples were stored in an air conditioned room at around 23 °C for 16 h at working days and for 64 h during weekends. An asymptotic decrease of the additive level in the food simulants in the successive migration periods, from a maximum values towards zero, was observed when the additives easily solve in the food simulants. In contrast when the additives solve poorly in the food simulants this additive level decrease is almost linear and shows a slow decrease rate. In the first case the leaching/migration of the additives is controlled mainly by their diffusion in the matrix of the polymer. In the latter case this process is controlled mainly by the partitioning of the additives between the polymer and food simulants. Two approaches were used to model theoretically these results.In the first approach the mass transport equation was solved by using as input parameters overestimated diffusion and partitioning coefficients as estimated for polypropylene homopolymer, the used food simulants and the three additives according to the recommendations of a practical guidance document in support of migration modelling as stipulated in the Regulation (EU) 10/2011. As expected the results of these calculations overestimate the vast majority of the migration levels determined experimentally. However there are also situations in which this approach leads to underestimations of the real migration levels.In the second approach numerical fitting was applied with the same equation and using the diffusion and partition coefficients as adjustable parameters to obtain a best-fit between experimental and calculated results. It was found that the so obtained “realistic” diffusion coefficients are all lower than the overestimated ones mentioned above. Moreover these “realistic” coefficients depend not only on the nature of the additive but also on the nature of the food simulant in contact with the polymer. This suggests that interaction of food simulants with the polypropylene occurs, which changes the properties of the polymer and thus the mobility of the additive molecules in its matrix. For the “realistic” partition coefficients it was found that they all are higher than the conservative ones mentioned above. Using these “realistic” parameters a good fit between calculated and experimental migration results was found. |
