Polymer additive migration to foods--a direct comparison of experimental data and values calculated from migration models for high density polyethylene (HDPE)

Food-packaging migration models: A critical discussion

The widely accepted and used migration models that describe the mass transport from polymeric packaging material to food and food simulants are confirmed here. A critical review of the most accepted models is presented in detail. Their main advantages and weak points, regarding their predictive accuracy, are discussed and weighted toward their usage extensiveness. By identifying the specific areas where using such models may not provide a strong correlation between theoretical and actual results, this work also aims in outlining some particular directions regarding further research on food - packaging interactions.

Determination of Partition Coefficients of Selected Model Migrants between Polyethylene and Polypropylene and Nanocomposite Polypropylene

Studies on nanoparticles have focused the attention of the researchers because they can produce nanocomposites that exhibit unexpected hybrid properties. Polymeric materials are commonly used in food packaging, but from the standpoint of food safety, one of the main concerns on the use of these materials is the potential migration of low molecular substances from the packaging into the food. The key parameters of this phenomenon are the diffusion and partition coefficients. Studies on migration from food packaging with nanomaterials are very scarce. This study is focused on the determination of partition coefficients of different model migrants between the low-density polyethylene (LDPE) and polypropylene (PP) and between LDPE and nanocomposite polypropylene (naPP). The results show that the incorporation of nanoparticles in polypropylene increases the mass transport of model migrants from LDPE to naPP. This quantity of migrants absorbed into PP and naPP depends partially on the nature of the polymer and slightly on the chemical features of the migrant. Relation (RPP/naPP) between partition coefficientKLDPE/PPand partition coefficientKLDPE/naPPat 60°C and 80°C shows that only BHT at 60°C has aRPP/naPPless than 1. On the other hand, bisphenol A has the highestRPP/naPPwith approximately 50 times more.

Feasibility study for the development of certified reference materials for specific migration testing. Part 2: Estimation of diffusion parameters and comparison of experimental and predicted data

This paper describes the second part of a project whose main objective was to develop the know-how to produce certified reference materials (CRMs) for specific migration testing. Certification parameters discussed are the diffusion coefficient, D(P), the respective polymer-specific coefficient, A(P), of the migrant polymer combinations and the partitioning coefficient, K(P,F), describing the partitioning of the migrant between the polymer and a food simulant. The parameters were determined for 16 preliminary candidate CRMs. Each parameter was determined by one laboratory. The six materials most suitable as reference materials were selected and the parameters then determined by four laboratories. The coefficients resulting from this small-scale interlaboratory comparison study can be regarded as the most reliable values available to date. These coefficients were applied for a comparison of experimental and predicted migration data. The experimental migration data arose from the same project and were determined by one laboratory for the first 16 materials and subsequently by four laboratories for the six materials selected in the second phase. Overall, experimental and predicted migration data fit together quite well. Roughly half of the predicted data were within +/-10%; almost all predicted data were within +/-40% compared with the experimental data.

Evaluation of migration models that might be used in support of regulations for food-contact plastics

Materials and articles intended to come into contact with food must be shown to be safe because they might interact with food during processing, storage and the transportation of foodstuffs. Framework Directive 89/109/EEC and its related specific Directives provide this safety basis for the protection of the consumer against inadmissible chemical contamination from food-contact materials. Recently, the European Commission charged an international group of experts to demonstrate that migration modelling can be regarded as a valid and reliable tool to calculate 'reasonable worst-case' migration rates from the most important food-contact plastics into the European Union official food simulants. The paper summarizes the main steps followed to build up and validate a migration estimation model that can be used, for a series of plastic food-contact materials and migrants, for regulatory purposes. Analytical solutions of the diffusion equation in conjunction with an 'upper limit' equation for the migrant diffusion coefficient, D(P), and the use of 'worst case' partitioning coefficients K(P,F) were used in the migration model. The results obtained were then validated, at a confidence level of 95%, by comparison with the available experimental evidence. The successful accomplishment of the goals of this project is reflected by the fact that in Directive 2002/72/EC, the European Commission included the mathematical modelling as an alternative tool to determine migration rates for compliance purposes.

Migration of aniline from polyamide cooking utensils into food simulants

Migration of aniline from polyamide cooking utensils was investigated. Aniline was found to originate from the application of black colorant in the polyamide 66 raw material. Analysis of polyamide raw material revealed an aniline concentration of 121 +/- 13 mg kg(-1). The aniline concentration in a cooking utensil (turner) manufactured with 70% polyamide raw material was 82 +/- 6 mg kg(-1). When testing migration from black coloured samples of turner, whisk and cooking spoon into water simulant at 100 degrees C for 30 min, the migration levels at the third exposure were 39 +/- 1, 11 +/- 4 and 37 +/- 4 mug dm(-2), respectively. Hence, these articles were not compliant with respect to the area-based limit for primary aromatic amines (3.33 mug dm(-2)) set by European Union Directive 2002/72/EC. Repeated testing of a cooking utensil (turner) showed that approximately 100 h use at 100 degrees C was required to reach legal (non-detectable) levels of aromatic amine migration.

Migration of Substances from Food Packaging Materials to Foods

The employment of novel food packaging materials has increased the number of occurring hazards due to the migration from packaging material to the packaged food. Although polymers have mainly monopolized the interest of migration testing and experimentation, recent studies have revealed that migration also occurs from "traditional" materials generally considered to be safe, such as paper, carton, wood, ceramic, and metal. The regulations and the directives of the EU tend to become stricter in this respect. The emphasis is on reaching a consensus in terms of food simulants and testing conditions for migration studies. Furthermore, the list of hazardous monomers, oligomers, and additives continues to augment in order to ensure that the consumer safety is in current agreement with the HACCP, which is continuously gaining ground.

Alternative fatty food simulants and diffusion kinetics of nylon 12 food packaging

The migration of laurolactam and cyclic di- and trimer of nylon 12 was assessed using three different films and five food simulants (olive oil, isooctane, 95% ethanol, 50% ethanol, water). Substitute test conditions for migration into olive oil according to European Union Directive EC/97/48 were applied using 95% ethanol and isooctane. Results showed that 95% ethanol overestimated while isooctane underestimated the respective migration into olive oil. Water was the best olive oil substitute, as migration of laurolactam into water and olive oil using the same temperature gave similar results. Additionally, diffusion kinetics of laurolactam were investigated by migration kinetic studies using isooctane and olive oil. Diffusion coefficients determined with isooctane were significantly higher than those found using olive oil. It was proved that isooctane had an interaction and olive oil was inert to the polymer. The diffusion conductance parameter, A(p), for nylon 12 determined using olive oil ranged from 0.3 to 0.6.

Evaluating the migration of ingredients from active packaging and development of dedicated methods: a study of two iron-based oxygen absorbers

The behaviour of two commercial oxygen-scavenging products with respect to migration of active ingredients into foodstuffs was investigated. Migrants were identified, and by using appropriate analytical methods, migration was determined in a variety of liquid, solid or gelled food simulants and foods. Simulants were chosen to cover a range of water activities and viscosities. Foods and the gelled food simulant agar were packed with and without vacuum, and with the oxygen scavenger in various locations relative to the packed food. The main migrants, as identified by X-ray fluorescence spectrometry, infrared spectroscopy and scanning electron microscopy with energy-dispersive spectrometry were Na(+) and Cl(-) in non-acidic aqueous simulants, and Na(+), Cl(-) and Fe(2+) in 3% acetic acid. Migration into aqueous simulants exceeded the current European Union limit for total migration from plastic materials (assumed to be currently applicable to these systems) and was probably excessive by any reasonable standard. However, neither oxygen scavenger appeared to release significant quantities of migrants into solid foods when the scavenger was properly located in the package and the packing process does not favour the contents becoming wet by water released from the food.

Diffusion of volatile compounds in fibre networks: experiments and modelling by random walk simulation

Predictive migration models for polymers are already so well established that the European Commission intends to allow the use of the models as one quality assurance tool in product safety assessment of plastic materials and articles for food contact. The inhomogeneity of fibre-based materials makes modelling difficult--thus, little research has been done in this area. The authors compare experiments on the diffusion of certain volatile compounds through laboratory kraft pulp sheets with computer simulations in which the fibre network structure is modelled explicitly. The major advantage of the present random walk simulation is that it gives an estimate of the effective diffusion constant for the fibre network. For most compounds, the agreement between the experiments and simulations is good. The experiments and simulations indicate that gas diffusion rate is very sensitive to sheet porosity.

Migration modelling as a tool for quality assurance of food packaging

The current potential for the use of migration modelling for studying polyolefin packaging materials (low- and high-density polyethylene and polypropylene) is summarized and demonstrated with practical examples. For these polymers, an upper limit of migration into foodstuffs can be predicted with a high degree of statistical confidence. The only analytical information needed for modelling in such cases is the initial concentration of the migrant in the polymer matrix. For polyolefins of unknown origin or newly developed materials with new properties, a quick experimental method is described for obtaining the characteristic matrix parameter needed for migration modelling. For easy handling of both the experimental results and the diffusion model, user-friendly software has been developed. An additional aim of the described method is the determination of the migrant partition between polymer and food or food simulant and the specific contribution of the migrant molecular structure on the diffusion coefficient. For migration modelling of packaging materials with multilayer structures, a numerical solution of the diffusion equation is described. This procedure has been also applied for modelling the migration into solid or high viscous foodstuffs.

Practical experience in the use of mathematical models to predict migration of additives from food-contact polymers

To reduce the amount of compliance-specific migration testing for food-contact polymers, the use of migration modelling has been evaluated. The paper describes experimental work carried out on a range of plastics and compares measured migrations against predictions obtained using mathematical models. A large number of experimental migration data have been obtained and used to evaluate a Fickian-based migration model in the prediction of specific migration of additives into olive oil. All tests were conducted using olive oil, representing the most severe case for fatty foods with test conditions including 2h at 121 degrees C, 6h at 70 degrees C, 2h at 70 degrees C, 2h at 60 degrees C and 10 days at 40 degrees C, representing short-term exposures at high temperatures and room temperature storage. Predicted migrations were calculated by inputting the measured initial concentration of additive in the polymers (Cp,0) into the equations together with known variables such as additive molecular weight, temperature and exposure time. The results indicate the Piringer migration model, using the 'exact' calculations of the Migratest Lite program, predicted migrations into olive oil that were close to or in excess of the experimental results and gave an overestimation for > 95% of the migrations generated here.

Influence of solvent absorption on the migration of Irganox 1076 from LDPE

The effect of solvent absorption on additive migration was studied by relating the diffusion coefficient (D) of Irganox 1076 to the maximum solvent absorption of different solvents in low-density polyethylene (LDPE) film. Solvents tested were ethanol, isopropanol, isooctane, ethylacetate, cyclohexane, tributyrin, tricaprylin and olive oil. Diffusion and partition coefficients were determined by fitting the migration curves, i.e. the concentration of Irganox 1076 in solvent as a function of time, with Fick 's diffusion equation. The results for the low molecular weight solvents show that with increasing maximum solvent absorption, D of Irganox 1076 is increasing as well. This trend is not observedfor the two triglycerides and olive oil. In spite of absorption, no increase in D was observed. The obtained result is the basis of an extended predictive migration model that, besides migrant and polymer properties, is also based on the maximum solvent absorption in the polymer.

Polymer additive migration to foods--a direct comparison of experimental data and values calculated from migration models for polypropylene

To reduce the amount of compliance testing for food contact polymers the use of migration modelling is under discussion and being evaluated by an EU Commission funded project (Evaluation of Migration Models No. SMT4-CT98-7513). The work reported in this paper was exclusively funded by industry to provide data for the independent evaluation of a diffusion based model using eight different samples of polypropylene (PP) covering the range of polymers specification and five commonly used plastics additives. One hundred and fifty experimental migration data have been obtained in triplicate and used to evaluate a Fickian-based migration model in the prediction of specific migration of five additives into olive oil. All tests were conducted using olive oil, representing the most severe case for fatty foods, with test conditions of 2h at 121 degrees C, 2h at 70 degrees C and 10 days at 40 degrees C, representing short term exposures at high temperatures and room temperature storage. Predicted migration values were calculated using the Piringer 'Migratest Lite' model by entering the measured initial concentration of additive in the polymers(Cp.0) in to the equations together with known variables such as additive molecular weight, temperature and exposure time. Where necessary the data generated in this study have been used to update the model. The results indicate the Piringer migration model, using the 'exact' calculations of the Migratest Lite program, predicted migration values into olive oil close to, or in excess of, the experimental results for > 97% of the migration values generated in this study. For all measurements, the predicted migration from the Migratest Lite program was greater than 70% of the observed value. This study has identified the possibility, that random co-polymers of propylene and ethylene give higher migration than other grades of polypropylenes and could be treated as a separate case. However, further work on more samples of random co-polymers is required to confirm this finding.