October 18-21, 2015
The contamination from materials is coined as “migration” and refers to a combination of various mass transfer mechanisms under thermodynamic control: molecular or mutual diffusion, sorption and desorption phenomena linear with concentration or not. The contamination pathways are particularly complicated when they involve several components or steps and strong interactions between materials and the medium in contact. They have been studied and mathematically described with two possible purposes: i) simplified approaches with a goal of overestimating the real contamination and ii) more sophisticated approaches to optimize the design of polymer materials, packaging materials or to redesign plastic additives with lower migration risk and toxicological hazard (e.g. plasticizers, surfactant, etc.).
The simplified descriptions use conventional Fick equations and proper partition rules to enable decision making – Is the contamination acceptable or not? – without requiring a perfect understanding of underlying phenomena. The approach has been chosen in EU to regulate food contact materials and is used somehow for food contact notifications in US. It has been applied to the diffusion in various polymers and is currently under extension for elastomers and varnishes. A rigorous formulation using a probabilistic resolution of transport equations provides a justification of such intuitive assumptions while minimizing both contamination levels and adverse risks (i.e. discarding falsely a good material). These theoretical results demonstrated that migration in closed systems by materials was obeying to prescribed statistical laws and that it could be managed robustly through conventional risk assessment approaches . Special methodologies have been recently developed to cover the whole supply chain .
More sophisticated approaches seek relationships between the chemical structures of contaminants and polymers and their mass transfer properties. Several calculation procedures have been proposed and tested over the last decades including semi-empirical models, models derived from molecular thermodynamics and brute force molecular simulation [3,4]. An overview of current capabilities of prediction based on the sole chemical structure is presented. Among the most effective techniques, generalized Flory-Huggins approximations at atomistic scale enable tailored calculation of partition coefficients (or equivalently excess-chemical potentials) with reasonable accuracy for almost any contaminant, polymer or co-polymer and media in contact [5-9]. Similar breakthroughs have been proposed for diffusion coefficients based on a generalized free-volume theory  and solute classification . Experimental and theoretical results demonstrate that adding flexible segments and symmetry axes close to the center-of-mass in additives could reduce diffusion coefficients of several magnitude orders and accordingly migration. When no data are available on the initial compositions or the identity of possible migrants, rapid deconvolution techniques have been shown to be effective for plastic materials [12,13].
 AIChE Journal 2005, 51 (4), 1080-1095.  AIChE Journal 2013, 59 (4), 1183-1212.  J. Chem. Phys. 2010, 132(19): 194902  Crit. Rev. Food Sci. Nutrition 2015, doi: 10.1080/10408398.2013.849654.  I&EC 2009, 48 (11), 5285-5301.  I&EC 2010, 49 (16), 7263-7280.  Int. J. Chem. React. Eng. 2010, 8.  J. Polym. Sci. Part B: Polym. Phys. 2014, 52 (19), 1252-1258.  Nguyen, P.-M.; Guiga, W.; Dkhissi, A.; Vitrac, O., submitted to I&EC.  Macromol. 2013, 46 (3), 874-888.  J. Appl. Polym. Sci. 2006, 101 (4), 2167-2186.  J. Appl. Polym. Sci. 2010, DOI: 10.1002/app.32950.  I&EC 2015, 54 (10), 2667-2681.