Aristoula G. Karamani · Panagiotis G. Demertzis Konstantoula Akrida-Demertzi
A.G. Karamani · P.G. Demertzis* · K. Akrida-Demertzi Laboratory of Food
Chemistry, Department of Chemistry, University of Ioannina, P.O. Box 1186, 45110
Ioannina, Greece
* e-mail: pdemertz@cc.uoi.gr
Received: 7 August 2001 / Published online: 23 January 2002
Abstract
Refillable PET containers offer great advantages from an ecological point of view, i.e., to reduce waste from packaging materials. A major concern is the potential public health risks of PET refillable bottles due to possible misuse. In this study, a modified method for determining the chemical inertness of PET bottles involving the interaction of strips from bottle walls with a selection of model contaminants was investigated. A further object was to establish temporally shorter contamination conditions for PET strips, yielding comparable contamination levels as the 14 days/40 °C exposure conditions applied so far. Results showed that substantially shorter exposure times (2-3 days) at higher temperature (60 °C) can be applied to make the overall chemical inertness test procedure as simple and quick as possible.
Keywords Chemical inertness · Refillable PET bottles · Model contaminants
Introduction
For the past few years there has been a worldwide tendency to reduce the environmental influences caused by human activities. As a consequence there has been a higher interest in reusing and recycling packaging materials. A typical example of refillable soft drink packaging materials in the international market is that of PET bottles. Refillable plastic containers composed of polyethylene terephthalate (PET) polycondensate are widely used. PET is particularly suitable for this purpose because it has a very good resistance to CO2 permeation losses, due to biaxial orientation of the molecules compared to other plastic material [1]. A major concern is that plastic material can interact with contacting chemicals to a greater or lesser extent. As a consequence, refillable containers may be contaminated by migration of harmful substances into the plastic material due to misuse by consumers before return.
The sorption experiments with PET bottle material applied so far showed that significant amount of chemicals can be adsorbed into the plastic material if misused, thus establishing a remigration potential in the bottle material after refilling.
Moreover there are neither specific national or EU regulations nor a standard test available which could be applied by industry or enforcement laboratories to demonstrate that a certain inertness behavior requirement, which would have to be specified, is fulfilled by the tested bottles.
Many studies, some of them highly time-consuming, have been made for determining the transfer of pollutants into the food [2, 3, 4, 5, 6, 7, 8, 9, 10]. The potential public health risks in connection with the use of PET refillable bottles as a consequence of possible misuse have also been reviewed and summarized recently [11].
Consumers misuse the bottles by filling them with, for example, concentrated lemonade with a strong flavor, household chemicals, and even pesticides. The adsorbed compounds will not be fully removed during the cleaning and washing procedure as usually done industrially before refilling the bottles. The adsorbed chemicals could migrate into a newly packed product and present a potential quality or safety risk [12]. From the results of all misuse studies carried out so far and from probability considerations it is generally concluded that returnable PET bottles can be safely reused, provided that food manufacturing procedures, including visual and electronic inspection systems, are required to eliminate abused bottles. However, the results achieved so far, at tremendous cost by work and cost intensive contamination procedures and remigration testing of whole bottles, give a more or less statistical picture of the possible range for remigration from abused bottles into refilled soft drinks. In order to ensure the quality and safety-in-use of recycled and re-used plastics for food packaging with the aim to establish sanitation criteria, a practical, cost efficient, and relatively quick standard test procedure, which could also be used by surveillance laboratories, must be developed.
Table 1 Model compounds (mol. weight) applied in a previous study [13]
Set A Set B Set C Set D . “alcohol-type” compounds “ester/ketone-type” “hydrocarbon-type” chlorinated (strongly interactive) compounds compounds compounds Ethylene glycol(62) Ethyl acetate(88) Toluene(92)a Chlorobenzene (112) Phenol(94)a Cyclohexanone (100)a n-Heptane (100) 1, 1, 1-Trichloroethane (133)a n-Hexanol (102)a Iso-Amyl acetate p-Xylene (106)a – 2-Phenylethanol (122)a Benzophenone 182)a Limonene (136)a – Menthol (156)a Linalyl acetate (196)a Phenyl cyclohexane(160)a – 1, 2-Decandiol (174) Methyl stearate (299) Phenyl decane (218) – a Model contaminants, which were selected to compose the new single set of compounds
In a previous study [13] a relatively simple inertness test method which focuses on the interaction of PET bottle wall strips material with selected model contaminants has been proposed. The principle of the proposed procedure was that strips cut from PET bottle walls should be in contact with four sets (representing different chemical classes) of model contaminants (Table 1) and their uptake (sorption) by the PET material was measured. A simple gas chromatographic method using flame ionization detection was developed to allow quantification of the whole range of model contaminants.
The aim of the present study was to investigate the possibility of establishing a modified standard, quick test method for refillable PET bottles to evaluate their inertness concerning the uptake and subsequent release of chemicals and sensory active compounds. The proposed test method involves the interaction of strips from PET bottle walls with a single set of twelve (12) model contaminants selected from the previously proposed four sets, followed by measuring their sorption by the PET material through GC analysis. A second objective was to investigate whether temporally shorter contamination conditions for PET strips could be applied, yielding comparable contamination levels as the 14 days/40 °C exposure conditions applied so far.
Experimental
Materials
Refillable 1.5-l PET bottles supplied by Continental PET Europe (France) were investigated in this study. The bottles were made of virgin material and had not been previously filled or washed. All chemicals and solvents (model contaminants) used (Table 1) were of analytical grade (i.e., purity> 99%).
Methods
Gas chromatography (GC) analysis. The GC unit used was a Fisons GC 9000 series gas chromatograph equipped with an injector and a FID detector. The separation column was a 30 m longX—0.32 mm internal diameter fused silica capillary DB-1 with a film thickness of 3 µm. The following GC parameters were kept constant: detector temperature: 280 °C, injector temperature: 220 °C, injector mode: split with split ratio ca. 20 ml/min, injection volume 1 µl. The applied temperature program was 90 °C (2 min), from 90 °C at a rate of 15 °C/min to 270 °C (10 min).
Contamination of PET bottle wall strips. Sorption experiments with rectangular PET-strips (6X—1.2 cm) were carried out with model compounds mixture set B (Table 1) which was prepared by mixing equal weight parts of the components and the mixture was further diluted 1:5 with polyethylene glycol (PEG 400) to reduce the vapor pressure and aggressiveness of the model compounds. The set B contains the following model compounds: ethyl acetate, cyclohexanone, iso-amyl acetate, benzophenone, linalyl acetate, and methyl stearate. Contamination was performed by placing each PET-strip in 20-ml screw cap glass vials filled with sufficient volume (10 ml) of set B compounds mixture to contact the whole strip area (lying position of the vials). The vials were then stored for specific time intervals (3, 5, 7, 9, and 11 days) at 60 °C and 70 °C as well as at 40 °C for 14 days.
Determination of model contaminant adsorption into the strips.
After contamination, PET-strips were cleaned on the surface and then the edges from each strip were cut off. The strips were then cut into small pieces and swollen by contact with 1 ml HFIP for 24 h at 40 °C. The swollen PET material was extracted with 4 ml isopropanol for a further 24 h at 60 °C. After cooling down of the samples the extracts were withdrawn and cleaned by filtering through regenerated cellulose disposable filters with 0.2 µm pore size. The filtered extracts were GC analyzed using ethyl myristate as internal standard.
Results and discussion
Discussion on development of the PET inertness test The increased use of refillable PET bottles as a package for soft drinks has prompted us to investigate the effect of adsorption of chemical compounds onto the wall of refillable PET bottles, which have been misused to various extent. The degree of adsorption depends on temperature, chemical nature, and molecular weight of the chemical compounds. The food safety quality control of washed and refillable PET bottles before they go out again to the consumer is completely in hands of the bottle fillers. Authorities and enforcement labs do not have any access to compliance testing on a refillable PET bottle drawn from the market. Consequently, there is clearly a need to establish a validated, reproducible, simple, and quick test method for refillable PET bottles with respect to evaluation of their inertness concerning the uptake and subsequent release of chemicals and sensory active compound.
The tendency of a plastic to adsorb contaminants and subsequently release them on refilling is closely related to the inertness of the plastic. The inertness of a PET plastic bottle depends on the intensity of the interaction between the packaged food and the packaging material during their direct contact. The uptake of constituents from the packaged food by the packaging material is based on a diffusion process which depends on the properties and the structure of the plastic material as well as on the diffusion and partition coefficients. The inertness testing of a PET material with the use of model contaminants could be simulated to the migration of chemical substances from the filled liquid into the bottle material, which could be a foodstuff or harmful chemicals. Consequently, the adsorbed amounts of model contaminants could be used to evaluate the inertness of a PET plastic bottle [14].
In order to determine the effect of the refilling system and the use of different refilling materials on the inertness of reusable bottles, the use of mixtures of model contaminants has been proposed to model the myriad of chemicals that could in principle contaminate returnable bottles due to consumer misuse which could result in a quality problem or even a health risk. For a systematic study on the adsorption and migration process for refillable PET bottles, four different sets of chemical model compounds have been proposed. The selection of the proposed contaminants was based on the chemical structure (molecular size, polarity, and functional groups) of them. An overview of these sets is presented in Table 1. Experiments were carried out with strips of PET bottle walls on a laboratory scale so that the inertness testing could be achieved in a most time saving and economic way. The strips were exposed to model contaminants and their adsorption by the PET material was measured by means of GC analysis. A detailed description of the experiments has been published recently [13].
The work done on contamination of bottle strips indicates that, under given standard contamination conditions, the adsorbed amounts and the sorption patterns obtained can be used to evaluate the inertness of a PET plastic bottle. This could be a key element of a future quick inertness test procedure and provides a very promising scientific/technical basis for further optimization (in the sense of simplification) to end up in a standardized and easy to apply method for general chemical inertness testing of refillable PET bottles. Only in this case, the method will be fit for purpose, i.e., be acceptable as a candidate standard method for chemical inertness testing of refillable PET bottles.
In order to optimize further the method by making it as simple and quick as possible, the current number of four sets could be reduced to only one set by, for instance, mixing up selected compounds from each set. Therefore, the next step should be the selection of the appropriate model contaminants comprising the new single set and the development of a GC analytical methodology for the simultaneous determination of all compounds in a GC run.
GC methodology
The developed GC method allowed complete separation of 12 substances from all of the 4 sets of compounds (marked with a in Table 1). The new single set is composed of four of the six compounds of set A, three of the six compounds of set B, four of the six compounds of set C, and one of the two compounds of set D. The GC profile of these compounds (approximately 10 ppm concentration for each individual component and 40 ppm concentration of the internal standard, using an isopropanol/ HFIP mixture as solvent) is given in Fig. 1. It can be seen that peak resolution is very satisfactory in all cases. The calibration curves for each contaminant were constructed within the concentration range 0-40 ppm applying the internal standard method. In this way, five different concentration levels of each contaminant in the solvent mixture were prepared, containing always a constant amount of the corresponding internal standard. Calibration curves were prepared by plotting the peak area ratios of each contaminant to the internal standard against the concentration ratios of the contaminant to the internal standard. As an example, the calibration curve obtained for menthol is presented in Fig. 2. Analogous curves were obtained for all contaminants. Calibration curves were submitted to linear regression analysis and correlation coefficients (R) greater than 0.99 were found in all cases, thus indicating very satisfactory linear relationship between contaminant concentration and FID detector response.
Sorption measurements
With a view to making the method easier and quicker, another step should be made. This step should target on reducing the time of the contamination procedure. Currently the method used in the previous study [13] applied standard contamination conditions of 14 days at 40 °C to achieve sufficient uptake of contaminants by the PET material. These conditions are generally accepted as standard contamination conditions for adsorption experiments with model contaminants and PET plastics. To reduce this time it will be necessary to investigate the question at which shorter time period somewhat higher temperatures, e.g., 60 °C or even higher, will cause the same contamination levels. If, for example, it would be possible to halve the contact time to only 7 days, considerable progress would have been achieved. Moreover, the inertness test method could involve migration measurements into EU-official food simulants, which takes 10 days at 40 °C. Here, in an analogous way, the time to measure this migration potential must be shortened significantly, if possible to 2-3 days. As a whole, the test could be shortened in this way from currently 24 days to less than 10 days, which definitely would justify labeling it a quick inertness test.
Kinetic sorption experiments were carried out at two different temperatures (60 °C and 70 °C) for specified time periods (3, 5, 7, 9, and 11 days) using set B as guide contaminants. Results obtained are shown in Tables 2 and 3. The results were compared with the sorption levels obtained by applying the standard contamination conditions of 14 days/40 °C, presented in Table 4. It is obvious from Tables 2 and 3 that the adsorbed amounts (given in mg/dm2) of all the contaminants onto the PET strips were increased as the contact time with the model compound solution (set B) was increased. This was originally observed for both 60 °C and 70 °C. However, at 70 °C the increase lasted until the seventh day of exposure and then a considerable decrease in sorbed amount was observed for all compounds. This possibly indicates that equilibrium contamination levels were reached within 7 days at 70 °C and thereafter desorption was beginning. From this point of view the temperature of 70 °C seems to be rather unsuitable for our contamination experiments. It is clear that the adsorption of the chemical compounds by PET strips depends on the contact time as well as on temperature. Van Lune et al. [1] have shown that temperature seems to influence the adsorption level and suggested that this is caused by a change in the structure of the polymer. Moreover, it has shown that the adsorption rate was temperature-time dependent [6, 15].
Fig. 1 GC profile of the new single set of model contaminants
Fig. 2 The calibration curve for menthol
Raising the temperature from 60 °C to 70 °C led to the increase of the adsorbed amounts of all of the model contaminants. This will partly be due to an increase in the diffusion coefficients of the contaminants with increasing temperature. The crystallinity of the plastic might decrease and the free volume can increase at increased temperatures resulting in molecules being adsorbed more easily. This means that, besides temperature, the composition and the crystallinity of the plastic bottle material influence the adsorption levels.
Another very important finding (Tables 2, 3, and 4) is that at 60 °C and for all contaminants similar sorption levels as with the 40 °C/14 days exposure conditions were obtained after 3, 5, and 7 days of exposure. At 70 °C, similar or even higher sorption levels were observed at 3 days of exposure. This can justify the point of making the inertness test procedure substantially shorter, using elevated temperature. More specifically, the temperature of 60 °C seems to be the most appropriate for sorption (contamination) experiments.
Ethyl acetate was the model contaminant, which has been sorbed in a significant manner in the bottle wall strips. This compound is generally used as a solvent in coatings and plastics. The high sorptive capacity of ethyl acetate can be attributed to its chemical affinity to the plastic material. Isoamyl acetate, cyclohexanone, and benzophenone have been sorbed in moderate amounts by the PET material. In contrast, linalyl acetate has been sorbed in very small amounts (usually 0.02-0.09 mg/dm2). Results obtained concerning this substance are therefore not reliable and should not be taken into account. Safa and Bourelle [10] have also reported that linalyl acetate penetrates slightly onto the walls of the bottles. This behavior can be attributed to the poor solubility of linalyl acetate in PET.
Conclusions
A direct conclusion of principal character following from this discussion is the potential of using a single set of model contaminants to determine in a reproducible, simple, and comparable way the interactivity of a given PET material with the model contaminants. The benefits from the establishment of such a method will mainly be:
Table 2 Sorbed amounts (given in mg/dm2) of set B compounds in the bottle wall strips after exposure at 60 °C
Compound 3 days 5 days 7 days 9 days 11 days Ethyl acetate 2.2±0.1 2.38±0.2 2.24±0.2 2.57±0.28 2.59±0.02 Isoamyl acetate 0.95±0.007 0.94±0.02 1.11±0.2 1.27±0.17 1.10±0.04 Cyclohexanone 1.07±0.01 1.07±0.009 1.36±0.12 1.33±0.07 1.22±0.18 Linalyl acetate 0.06±0.009 0.09±0.01 0.05±0.01 0.04±0.002 0.05±0.01 Benzophenone 0.69±0.008 0.72±0.03 1.06±0.02 0.94±0.01 0.92±0.18
Table 3 Sorbed amounts (given in mg/dm2) of set B compounds in the bottle wall strips after exposure at 70 °C
Compound 3 days 5 days 7 days 9 days 11 days Ethyl acetate 2.26±0.25 2.28±0.01 3.88±0.34 3.27±0.09 1.68±0.4 Isoamyl acetate 1.13±0.016 1.49±0.05 2.48±0.1 1.63±0.035 0.87±0.2 Cyclohexanone 1.19±0.13 1.43±0.05 2.47±0.13 1.69±0.14 0.85±0.16 Linalyl acetate 0.02±0.005 0.04±0.01 0.27±0.01 0.05±0.01 0.05±0.01 Benzophenone 0.89±0.017 1.037±0.02 1.73±0.05 1.19±0.025 0.67±0.16
Table 4 Sorbed amounts (given in mg/dm2) of set B compounds in the bottle wall strips after exposure at 40 °C
Compound 14 days Ethyl acetate 2.04±0.02 Isoamyl acetate 1.24±0.23 Cyclohexanone 1.19±0.1 Linalyl acetate 0.03±0.007 Benzophenone 0.85±0.2
Regarding the possibility for optimization of the exposure conditions, results obtained are supporting the suggestion that shorter exposure times (2-3 days) at elevated temperature (60 °C) can be applied to make the test procedure as simple and quick as possible. Acknowledgements The financial support for this work as part of the project SMT4-CT96-2129 from DGXII of the Commission of the European Community (Brussels, Belgium) as well as materials from Continental PET (Bierne, France) are acknowledged.
References
1. Van Lune FS, Nijssen LM, Linssen JPH (1997) Packag Technol Sci 10:221-227
2. Landsberg JD, Bodyfelt FW, Morgan ME (1977) J Food Prot 40:772-777
3. Casaway JM (1978) J Food Prot 41:851-862
4. Casaway JM (1978) J Food Prot 41:965-973
5. Cassiday MD, Streu RJ, Wence RL, Delassus PTJ (1990) Plast Film Sheet 6:268-275
6. Begley TH, Hollified HC (1993) Food Technol 47:109-112
7. Nielsen TJ (1994) J Food Sci 59:227-230
8. Franz R, Huber M, Piringer OG (1994) Food Addit Contam 11:479-496
9. Feron VJ, Jetten J, De Kruijf N, Van den Berg F (1994) Food Addit Contam 11:571-594
10. Safa HL, Bourelle F (1999) Packag Technol Sci 12:37-44
11. TNO Nutrition and Food Research (1993) Code of Practice: Guidelines for an Industrial Code of Practice for refillable PET bottles, edn 1
12. Jetten J, de Kruijf N, Castle L (1999) Food Addit Contam 16:25-36
13. Demertzis PG, Johansson F, Lievens C, Franz R (1997) Packag Technol Sci 10:45-58
14. Palzer G, Franz R (1998) Kunststoffe 88:862-865
15. Tawfik MS, Devlieghere F, Huyghebaert A (1998) Food Chem 61:157-162
16. EC (1994) Directive 94/62/EC on packaging and packaging waste. Off J Eur Commun L365:10-23
|
If you have come to this page from an outside location click here to get back to mindfully.org |