"A New Experimental Procedure for Incorporation of Model Contaminants in Polymer Hosts"
C.D. Papaspyrides1*, Y. Voultzatis1, S. Pavlidou1, C. Tsenoglou1, P. Dole2, A. Feigenbaum2, P. Paseiro3, S. Pastorelli3, C. de la Cruz Garcia3, T. Hankemeier4, S. Aucejo4
- Laboratory of Polymer Technology, School of Chemical Engineering, NationalTechnicalUniversity of Athens,Zographou, 15780 Athens, Greece
- INRA SquAlE, Moulin de la Housse, BP 1039, 51687 Reims, France
- Analytical Chemistry, Nutrition and Bromatology Department, Faculty of Pharmacy, University of Santiago de Compostela (USC), 15782 Santiagode Compostella, Spain
- Packaging Research Group, Analytical Sciences Division, TNO-Nutricion and Food Research Institute, POBOX 360, 3700 AJ Zeist, The Netherlands
* Corresponding Author: Tel. +30 210 772 3179, Fax +30 210 772 3180
E-mail:
ABSTRACT
A new experimental procedure for incorporation of model contaminants in polymers was developed as part of a general scheme for testing the efficiency of functional barriers in food packaging. The aim was to progressively pollute polymers in a controlled fashion up to a high level in order to monitor migration phenomena difficult to study in real conditions. To this end, a contamination recipe was initially formulated by a set of selected surrogates. The experimental procedure developed, led to satisfactory results as far as homogeneity and final concentration in the polymers concerned. High temperature data were also used in order to evaluate the efficiency of typical thermoforming processes in reducing possible volatile and non-volatile substances from recycled polymeric materials.
KEYWORDS: Surrogates, contamination procedure, mixing, functional barrier, diffusion
INTRODUCTION
Packaging has become an indispensable element in the food manufacturing process. Presently, plastics are preferentially used for packaging foodstuffs due to their outstanding usage properties. It is worth mentioning that more than 30 different plastics are currently being used for this purpose. Plastic packages are capable of retarding and even preventing detrimental changes in the packed material caused by oxygen, light and microorganisms. Plastics are also able to greatly reduce the loss of valuable components on behalf of the packed material, such as water or flavor (1).
Due to modern environmental concerns and associated legislation, the question of recyclability of used plastic packaging into new food packaging applications is of increasing interest. Consequently, recycled plastics have already been used in food-contact applications around the world, for several years. However, these cases were considered to have more of a pilot character than real market value and, in most cases, the mass fraction of recycled plastics in these applications was relatively low, due to blending with virgin plastics or sandwiching with layers of virgin plastic of relatively high thickness (1).
Plastics contain low molecular components, such as monomers and oligomers, and additives, such as plasticizers, lubricants, stabilizers and antioxidants, which are absolutely necessary for the processing and the stability of the final materials. The drawback in this case is the migration of the additives from the package into the packed material (2-6).
It must be said that when a polymer is in contact with a liquid, such as foods or chemicals, generally two mass transfers take place simultaneously: liquid enters the polymer, while polymer additives diffuse from the plastic into the liquid. The following drawback is obvious, a pollution of the stored liquid and a decrease in mechanical properties of the plastic. Thus, the packaging can itself represent a source of contamination, through the migration of substances from the packaging material into food. Hence, regulatory authorities around the world have recognized that it is necessary to control such contamination, and many have enacted extensive legislation (7). On the other hand, a large amount of research regarding the migration of volatiles, additives, monomers and oligomers from plastic packaging materials into food has been conducted (1, 4-16).
Approaches envisaged for producing food-packaging materials from post-consumer collected plastic packaging include washing, which is, however, incapable of extracting all the chemicals that have migrated into the plastic, and depolymerization, which is too costly to be considered applicable.
A third route appears as an interesting compromise both from the point of view of performance and cost. This consists of reusing old polymer packages in new packaging products by sandwiching them through co-extrusion in bi- or tri- layer polymer configurations, where the old polymer is located between two virgin polymer layers (1, 8, 12, 17-33). As it takes some time for the contaminant initially located in the old polymer to diffuse through the virgin layer, the virgin polymer layer acts as a shield to food pollution, and thus constitutes a “functional barrier”. During the co-extrusion process, the polymer is heated up to high temperatures before being cooled down by air. In the course of this process, contaminant diffusion accelerates and, therefore, the time of protection of the functional barrier is significantly reduced. Thus, the main problem is to determine the period of time over which the food is protected, either by experiments or calculations (19-31).
Considerable progress has been made from the scientific point of view in understanding and modelling diffusion processes of adventitious hazardous compounds, from a recycled plastic in direct contact with food or from a core layer, across a functional barrier. However, putting this knowledge to use by devising industrial solutions is still in waiting.
Currently available are guidelines developed by the Food and Drugs Administration (FDA) and the U.S. Food and Plastics Industry, which, however, were established on the basis of too conservative an approach, and require enormous efforts concerning the implementation of the underlying tests (7). These test schemes, also known as “challenge” tests, are simulating a recycling process by artificially introducing model contaminants, so-called surrogates, in the polymer; subsequently, the cleansing efficiency or surrogate removal potential of the process is checked (7).
European project FAIR-CT98-4318, dedicated one of its research tasks to the study of functional barriers. More specifically, one of the overall objectives of the project was to generate an advanced scientific understanding, along with modelling, of the physicochemical behaviour of chemical contaminants in recycled plastic layers buried by functional barrier polymers. This will eventually constitute the basis for evaluating the safety and defining design criteria for appropriate functional barrier protection against recycled plastics used for food packaging (34).
Previous experimental procedures aiming to generate “real” polluted material, proposed between 1990 and 1997, consisted in simulating pollution with cocktails of surrogates. Simple contact of the material to be contaminated with the surrogates was suggested for about two weeks at around 40°C (18, 35-39). The “real” polluted material thus obtained was characterized by:
- Low pollution levels (100 ppm).
- Surface pollution, primarily; this means that aqueous washing is in general very effective, even for low contact times (200 seconds) and high molecular weight compounds (32, 36, 37). This comprises a basic disadvantage of those earlier “model pollution” experiments since in real life pollutants are homogeneously dispersed throughout the polymer phase.
- Pollution level dependent on the nature of the surrogate (polarity, molecular weight).
In this project the aim was to pollute materials at a precontrolled and high level, of the order of 1000 ppm, in order to monitor migration phenomena difficult to quantify in real life situations. It should be emphasized here that the extrapolation to real cases has so far been made only by numerical simulations based on data corresponding to low levels of concentration (35-41). Furthermore, deep penetration and therefore, homogeneous surrogate distribution in the material was considered here.
EXPERIMENTAL
Materials
Polymers:
The contamination procedure was tested on different polymers in order to study a wide range of polymer matrix qualities. The following polymer hosts were supplied by Cryovac in the form of pellets: High Density Polyethylene (HDPE), Low Density Polyethylene (LDPE), Linear Low Density Polyethylene (LLDPE), Polypropylene (PP), Ethylene Propylene Copolymer (EP), Polyamide 6 (PA-6), Ethyl Vinyl Alcohol (EVOH) and Polyethylene Terepthalate (PET).
Contaminants:
The model set of solid and liquid contaminants, the so-called “surrogates”, represents all general categories of chemical compounds as described in the FDA regulations: volatile and non-polar, volatile and polar, non-volatile and polar and non-volatile and non-polar. Additionally, a wide range of functional groups was used, in order to reflect the different chemical and physical nature of real life contaminants. The three solid and five liquid contaminants used are presented in table 1, along with their characteristic properties and supply source.
[Insert table 1 about here]
Contamination Recipe
The major criterion in assessing a contamination recipe was its ability to provide an initial excess of solid and liquid surrogates, in order to accommodate the possibility of a worse case contamination scenario during usage, disposal or storage. Preliminary experiments were carried out to determine the initial concentrations required to obtain detectable amounts of surrogates in the final product (i.e. in the polymer after processing). With this in mind, it was decided to incorporate high amounts of volatile surrogates in the polymer hosts and to exclude Dimethylsulfoxide (DMSO) from the non-polar polymers, since this polar contaminant does not permeate in these polymers at all (42). The contamination recipe finally established is shown in table 2.
[Insert table 2 about here]
Experimental Procedure
A “contamination broth” was prepared, by blending all eight surrogates together in quantities indicated by the contamination recipe mentioned above, with a view to a 2 kg final product. The broth was mixed with 400 g of virgin polymer in a reactor vessel, which was heated in an oil bath. The temperature in the bath was set between 60 °C and 80 °C, depending on the nature of the polymer. A vertical condenser was adjusted to the vessel to prevent volatile surrogate losses. The “master-batch” produced was continuously stirred for 1 day and then a 1-week “contaminant penetration period” followed, during which, the master-batch was periodically stirred (~ 1 h / day). After this week, the master-batch was extruded under conditions shown in Table 3, and the extrudates were pelletized and mixed with the rest of the virgin polymer host (1600 g). Afterwards, a second extrusion took place to improve the penetration of the surrogates into the core of the material. The final product was then chemically analysed.
The aforementioned experimental procedure was applied to all polyolefins, i.e. LDPE, HDPE, LLDPE, PP, and EP without any problems. The extrusion conditions are given in table 3.
[Insert table 3 about here]
On the contrary, contaminated pellets of PET and EVOH presented severe problems during the extrusion process due to their solubility in the cocktail of surrogates employed. The pellets of these polymers were swollen in the presence of this liquid mixture under the high temperature process conditions, causing congestion in the feeding zone of the extruder.
This problem led to a modification of the contamination procedure. In the case of PET, where DMSO was causing most of the problem, this particular surrogate was excluded from the broth, in order to allow processing of the master-batch. It was incorporated in the polymer in the second step of the mixing process, by adding DMSO containing polymer to the masterbatch instead of virgin.
The problems were more severe with EVOH, since this polymer is soluble to all liquid surrogates. In this case, a master-batch with lower surrogate concentrations was prepared by mixing the same quantities of surrogates with a larger quantity of pellets (800 g), in order to make master-batch extrusion possible. As far as DMSO is concerned, the same procedure as in the case of PET was followed.
The extrusion conditions for these polymer hosts are also given in table 3.
Chemical Analysis
The masterbatch of each polymer was analysed at the INRA Laboratory. In the case of PP, both the masterbatch and the final product were analysed at three different laboratories. The analytical procedures used for chemical analysis by each laboratory are described below. Each laboratory analysed five samples in order to evaluate the homogeneity of the product.
INRA Laboratory
Extruded polymer samples were ground to powder before extraction. Then, 1 g of each one of them was soaked with 10 ml of a swelling solvent at 40 °C for 5 days. This solvent is dichloromethane for all polymers, except for EVOH where methanol was used. Afterwards, the extraction liquid was analysed by means of Gas Chromatography with Flame Ionization Detector (GC-FID) with polar and non-polar column.
TNO Laboratory
Polymer samples were cut in thickness less than 1 mm and 250 mg of each were extracted with 10 ml of dichloromethane, containing dichloroethane and dichlorobenzene as internal standards, by shaking at room temperature for 24 h. Afterwards, the extraction liquid was analysed by means of Gas Chromatography – Mass Spectrometry (GC-MS) and High Performance Liquid Chromatography (HPLC).
USC Laboratory
Samples were cut in slices with thickness less than 500 μm and 250 mg of each were extracted with 10 ml of dichloromethane at 40oC for 24 hours in a hermetically sealed glass tube. The extract is subsequently analysed by means of GC-FID and HPLC.
RESULTS AND DISCUSSION
The results of the analysis of the contaminated products are viewed in terms of the achieved homogeneity and retention of the surrogates. “Retention” of a surrogate is defined as the ratio of surrogate concentration in the masterbatch or the final product, divided by the surrogate concentration of the contamination recipe. This ratio is a measure of the overall efficiency of the impregnating process, as far as material utilization is concerned, and depends on physicochemical and operational parameters. In that respect, it is worth mentioning here the main factors affecting additive entrapment and homogeneity into a polymer (43, 44):
- Ability of the additive to survive the extrusion run; this is controlled by its volatility under the specific processing conditions. A measure of this is the difference between the boiling point of the additive and the processing temperature.
- Thermodynamic compatibility between the various species.
- Extent and intensity of the mixing process vis a vis its ability to bring the components in intimate contact: a) to generate interfacial area for liquid penetration and, b) to initially de-aggregate and consequently distribute the solids in a uniform fashion.
- Diffusivity of the substrates through the interface and into the polymer mass.
First of all it must be emphasized that the experimental procedure developed in all cases led to an acceptable homogeneity of the sample. The average concentration data in the masterbatch are displayed in table 4. These results are used to calculate the retention as previously described. table 5 displays the percentage retention of the surrogates in all masterbatch grades made.
[Insert table 4 about here]
[Insert table 5 about here]
As mentioned already in the Introductory Part, a major parameter influencing retention is the volatility of the surrogate, directly related with the boiling point in the case of the liquid contaminants. Therefore, it seemed worthwhile to correlate the aforementioned data with this particular parameter. According to figure 1 such a correlation proves valid, indicating that the retention of the liquid surrogates is an exponential function of the boiling point of the surrogate. This means that the most essential factor for the retention of volatile surrogates is the temperature difference between their boiling point and the temperature of the process applied. Therefore, potential use of processes involving high temperatures (e.g. melt recycling) is, as expected, a very efficient way to eliminate volatile substances.
[Insert figure 1 about here]
It is also evident that there is a distinct difference between the behaviour of the polyolefins and the behaviour of the polar polymers. It is well known that polyolefins generally do not have efficient barrier properties, compared with polar polymers like PET, PA-6 or EVOH. This is confirmed by the higher levels of retention in the case of polyolefins, in contrast with the data for the polar polymers (fig.1).
Turning to the solid surrogates, in figure 2, retention data are compared for the masterbatch of each polymer. Again, the very high levels of retention for the case of polyolefins show their poor barrier properties. However, surrogate retention is also high for the polar polymers, since the high temperatures used throughout the contamination procedure are not sufficient to remove these non-volatile substances.
[Insert figure 2 about here]
In particular for the case of PP, analysis of the final product was also carried out, in addition to the analysis of the masterbatch. This was part of a “ring test” between the three analytical laboratories involved (INRA, TNO, USC), in order to compare results and therefore evaluate reproducibility and the efficiency of the experimental procedure in terms of achieving homogeneity. The average concentration data are demonstrated in Table 6, indicating first of all a satisfactory dispersion of the surrogates throughout the masterbatch and the final grade. In other words, the experimental procedure applied resulted in dispersing all surrogates in the final product quite successfully and at a high level of homogeneity.
[Insert table 6 about here]
The aforementioned results were again converted to percentage retention and in figure 3, for the liquid surrogates, retention versus the boiling point of the surrogate is plotted for both the masterbatch and the final material. It is evident that in both cases the retention of the liquid surrogates is again an exponential function of their boiling point. However, the levels of retention in the final product are quite lower, a fact obviously attributed to the second extrusion step that is necessary to receive the final product from the masterbatch. By extrapolation, it seems possible that repeated passages through the extruder could further reduce or practically eliminate volatile substances in melt recycling.
[Insert figure 3 about here]
This, however, is not the case for the solid surrogates, which are not affected under similar conditions by the extrusion process. This is demonstrated in figure 4 by the high levels of retention both in the masterbatch and in the final PP grade. Thus, it is safe to assume that non-volatile substances remain in the recycled material even after repeated processes involving high temperatures. If this is the case, it becomes necessary to estimate the time needed for the substance to diffuse out of the polymer and into the food, or to define a toxicologically acceptable level of the substance in the recycled polymer.