Evaluation of plastics for food packaging
Food Additives and Contaminants v.11, n.2, 221-230 1994
O. G. Piringer
Fraunhofer Institut für Lebensmitteltechnologie and Verpackung, Schragenhofstr. 35, D-80992 München, Germany
One of the principal aims of the regulations for food contact materials and articles is the protection of the consumer. In order to fulfill this goal many analytical questions must be answered in the next few years. Of great help in the evaluation procedures can be theoretical predictions of migration based on empirical data for partition and diffusion coefficients. Using well established gas chromatographic methods for the detection of volatiles and new techniques involving supercritical gases for the analysis of minimally volatile constituents both specific and overall migration can be investigated. Further activities must be focused on the organoleptic properties and the average migration potential in different groups of materials and articles produced by the usual technology. Better correlations between rates of migration into food simulants and into real foodstuffs should be found in order to make the evaluation of plastics as realistic as possible.
Keywords: migration, diffusion, partition, supercritical fluid, chromatography
EVA Ethylene vinyl acetate LDPE Low density polyethylene HDPE High density polyethylene PP(isot) Polypropylene with high degree of isotacticity OPP Oriented polypropylene PET Polyethylenetherephtalate Tenax Polymer of 2,6-diphenyl-p-phenylene oxide BHT 2,6-Di-tert-butyl-4-methyl phenol Chimassorb 81 2-Hydroxy-4-n-octoxybenzophenon Irganox PS 800 Dilaurylthiodipropionate Irganox 1076 3(3,5-Di-tert-4-hydroxyphenyl) propionate Irganox 1010 Tetrakis [3-(3'5'-di-tert-butyl-4'-hydroxyphenyl) propionyloxymethyl]-methane Irgafos 168 tris-(2,4,di-tert,butylphenyl)-phosphite
The fundamental requirements for an evaluation of food packaging can be derived from the Framework Directive 89/ 109/EEC (CEC 1989). In Article 2 of this Directive, four items with corresponding practical consequences for food contact materials are mentioned. Three of these ideas are prohibitory and refer to the transfer of constituents to foodstuffs in quantities which:
- could endanger human health,
- bring about an unacceptable change in the composition of the foodstuffs,
- bring about a deterioration in the organoleptic characteristics of the foodstuffs.
The fourth item is the requirement for good manufacturing practice which is at the beginning of Article 2.
Presently, the first two prohibitions are specified in a more precise manner, especially for plastics, in Directive 90/ 128/EEC (CEC 1990). With a list of authorized substances and specific migration limits (SML) for certain substances, the transfer of constituents which could endanger human health is addressed. The overall migration limit is understood as a guarantee for a sufficient inertness of the plastic material or article to avoid an unacceptable change in the composition of the foodstuffs.
To account for the complexity of foodstuffs and the variety of conditions in which they come in contact with plastics, a series of conventions is necessary to make the quantification of specific and overall migration practical. To compare results from different laboratories, conventional rules for test conditions (e.g. temperature and time) and a selection of reproducible food simulating agents were established in Directive 82/711/EEC (CEC 1982). In Directive 85/572/EEC (CEC 1985), the assignment of simulating liquids to foodstuffs is given. Also included are some reduction factors to correlate the SML values into the simulants to migration into real foodstuffs.
In order to fulfill the numerous demands in the regulations there are many analytical questions which must be answered in the next few years. Here it should be emphasized that all chemical compounds not given by the positive list are principally forbidden. Two major requirements are:
the analytical methods should be practical, e.g. to be carried out in a reasonable time frame, and have sufficient sensitivity and reliability; and
the global amount of migrating substances must be separated into individual components for specific identification or identification according to their structural moieties.
Prediction of migration parameters
There exists already a plethora of data relating migration results to the theory of diffusion. These data address migration from a multitude of plastics under various conditions, into a variety of solvents, simulating liquids and foodstuffs. From the comparison between diffusion theory and practice it follows that, in many cases, the agreement is within the experimental data fluctuation.
For food-simulating liquids in contact with homogeneous plastics, the experimental results most closely correspond to the equations of the diffusion theory. All that is needed for a reasonable prediction of migration in many practical cases is the availability of data for two fundamental constants:
the partition coefficient (KP/L ) of a migrating solute (i) between the plastic (P) and
the simulating liquid (L) in contact with the plastic (P); and
- the diffusion coefficient (DP) of the solute (i) in the plastic (P).
Assuming that data for KP/L and DP do exist or can be predicted with sufficient accuracy, a considerable reduction in analytical work, time and financial support would be possible.
The following considerations refer to the estimation of values for KP/L and DP and their use for prediction of migration from plastics (Piringer 1992). Special consideration is given to the comparison of predicted values with experimental results obtained from specific and overall migration testing from polyolefins into the simulating liquids olive oil, corn oil, the synthetic triglyceride HB 307, and 95-100% ethanol.
As a starting point, the following relationship results from the mass balance and the partition of a solute between plastic (P) and liquid (L) at equilibrium:
with the definition of a as:
mL,t, mL, and mP,0 represent the mass of solute migrated into L at the time t, at equilibrium (t = ) and in P at t = 0, respectively. VL and VP represent the volume of L and P. CL, and CP, are the equilibrium concentrations of the solute in L and P.
It can be supposed that in almost all practical cases VL/VP > 10. This means that from the value of Kp/L two border cases occur:
KP/L < 1. In this case 1 and mL, = mP,O. At equilibrium the whole mass of solute is transferred from P to L. Because cP,0 = mP,0/VP and VP can be considered as the product of the plastic surface area A in contact with L and the thickness dP of P, in this case the amount of solute mL,t/A migrated until the time t from P into L can be calculated approximately with the simplified equation:
|CP,O • (DP • t)1/2 CP,O • (DP • t)1/2||(3)|
as long as ML,t < 0 • 6mL,.
Kp/L 1. In this case < 1 and mL, < mP,O. The solute can be transferred only partially from P to L. Since DP is the same as in case I (as long as no considerable interaction between L and P occurs) the maximum migration is, in practice, attained much faster than in case I (figure 1).
Considering the partition of solutes between P and L, the simulating liquids defined in Directive 82/711/EEC (CEC 1982) reflect the two extremes of polarity of foodstuffs: the nonpolar fatty food is represented by olive oil and the polar food is represented by the aqueous simulants. For most food contact plastics (and especially for all the non-polar polyolefins), one can assume that the equilibrium concentration of a solute in P is approximately the same or smaller than in fatty foodstuffs or fatty-simulating liquids. That means, Kp/L 1 and consequently case I occurs. In aqueous foodstuffs and aqueous-simulating liquids the solubility of organic solutes is generally much lower than in fat or organic simulants and thus KP/L 1.
Even in the low molecular weight alcohols such as methanol and ethanol, which have much higher polarities than the polyolefins, the solubility is so high compared with that in water (Koszinowski and Piringer 1989), that they can be considered as fatty simulants with KP/L < 1. Comparisons between migration into olive oil, corn oil, synthetic triglycerides and ethanol have demonstrated the similarities of these liquids, especially for testing the migration behaviour of the polyolefins (Till et al. 1987, Figge 1988, Figge and Hilpert 1991, Barter et al. 1992).
Figure 1. The relative migration mL,t/mP,O of a solute as a function of t 1/2 for KP/L < 1 (case I, ethanol, oil) and KP/L 1 (case I I, water).
For general applications of plastics in contact with foodstuffs a prediction of migration for the worst case (case 1) is of special interest. From the experimental results of diffusion studies with normal alkanes (Koszinowski 1986) and a variety of other organic solutes in polyolefins during the last 10 years (Becker el al. 1983, Koszinowski and Piringer 1986a,b), the following formula can be developed for an estimation of the DP values (CM 2/S) as a function of the relative molar mass (Mr) of the solute and the temperature (T) (K):
log DLDPE = 6 • 8 - 4350 • - 0 • 0035Mr
DLDPE 10 • DHDPE 10DPP(isot)
Three experimental observations support this formula: (1) the logarithm of DP decreases approximately linearly with increasing Mr in the homologous series of n-alkanes; (2) a substance with the same Mr value as an (hypothetical) n-alkane has a DP value which does not exceed that of the alkane; (3) the exponential temperature dependence of DP.
With the equations (3) and (4) a rapid estimation of the maximum mass transfer of a solute (with the molecular weight Mr) from a polyolefin into a foodstuff or a fat simulant liquid at a given time t and a given temperature T is possible.
Table 1. Estimated and measured migration values for some additives from polyolefins into fat simulants.
mL,t/A CP,O T t (mg/dm2 Additive Simulant Polymer (mg/g) (ºC) (days) Calc. Exp. Ref. BHT Olive oil LDPE 5 40 1 17 8•1 1 M, = 220 Ethanol LDPE 5 40 1 17 8•8 1 Olive oil HDPE 5 40 10 17 4•7 1 Corn oil HDPE 0•1 40 10 0•34 0•06 2 HB 307 HDPE 0•1 40 10 0•34 0•10 2 HB 307 PP 2 40 10 6•9 2•9 1 Olive oil PP 5 40 10 5•4 2•0 1 Chimassorb 81 Olive oil LDPE 5 40 l 11 12 1 M, = 326 Olive oil HDPE 5 40 10 11 6•6 1 Olive oil PP 5 40 10 3•5 3•5 1 Irganox PS 800 Olive oil LDPE 5 40 1 5•2 9•3 1 M,=515 Olive oil HDPE 5 40 10 5•2 4•2 1 Olive oil PP 5 40 10 1•7 3•4 1 Irganox 1076 Ethanol LDPE 0•8 40 2 1•1 0•44 3 M,=530 HB 307 HDPE 1 40 10 0•98 0•77 4 HB 307 PP 1 40 10 0•31 0•50 4 Irganox 1010 Olive oil LDPE 2 40 1 0•46 1•2 1 M,= 1178 Corn oil LDPE 0•25 49 10 0•09 0•11 2 Corn oil EVA* 0•215 4 10 0•006 0•006 2 Olive oil HDPE 5 40 10 0•36 0•32 1 Olive oil PP 5 40 10 0•11 0•15 1 References: 1-Figge and Hilpert (1991); 2-Little (1983); 3-Franz et al. (1992); 4-Figge (1988). * = copolymer of ethylene and 12% vinyl acetate.
From published (Little 1983, Figge 1988, Figge and
Hilpert 1991) and unpublished experimental results (Franz et al. 1992), a
estimated and measured values of migration into liquid at time t over area A (mL,t/A) can be made. This is shown in table 1.
As the densities of the polyolefins are > 0•9, the conversion CP,O (mg/cm3) = CP,O (mg/g) was neglected. For additives of higher molecular weight often an effect of bloom is observed, especially at lower temperatures (Till et al. 1987). The consequence of bloom has a considerable effect on the mL,O/A value at t = 0, and results in corresponding higher values of migration at a given time t than predicted.
Determination of the migrant concentration in plastics by supercritical fluid extraction and chromatography
For a rapid estimation of the migration potential of a plastic, the concentrations CP,O for individual solutes and the total amount of extractable matter must be determined. This can be done using different traditional procedures or more sophisticated instrumental-analytical methods (van Battum and van Lierop 1988). For very volatile solutes the well established techniques of headspace gas chromatography can be used. This is applicable to some monomers and solvents from printing inks. For less and very low volatile solutes an efficient technique is the extraction with gases in their supercritical state using supercritical fluid extraction (SFE, e.g. with CO2). The entire extract can be resolved via a flame ionization detector (FID), which is highly carbon sensitive, but of low selectivity for certain moieties (figure 2). With such a detector (FID 1 in figure 2), the extractable and potentially migrating matter from the plastic sample can be determined in about 1-2 h. To avoid an underestimation of the extractable amount, a calibration of the FID with some phenolic antioxidants (which are less sensitive for this detector than hydrocarbons) can be made.
Supercritical carbon dioxide can also be used as the mobile phase for chromatographic separation in supercritical fluid chromatography (SFC) of less volatile compounds within fused silica capillary columns in combination with an FID as the detector (figure 2). The extraction and separation steps can be coupled on-line. For optimal sensitivity and separation power, the extracted compounds are first collected in a trap (with C02 decompression) and then separated under supercritical conditions. The SFC can also be performed off-line following an extraction step accomplished via other techniques such as Soxhlet extraction or contact of the plastic with a swelling solvent. Samples of the extract solution can be introduced directly into the SFC apparatus.
In figure 3, an SFC chromatogram obtained with on-line combination of SFE/SFC with a sample of commercial OPP film is shown (Bücherl et al. 1993). Besides a mixture of oligomeric compounds, an antistatic additive (a phosphite together with its oxidized reaction product and two phenolic antioxidants) could be identified by coupling SFC with mass-spectrometry (MS). In table 2, estimated and measured migration values for three additives and the measured overall migration from the OPP-film to ethanol are shown (Franz et al. 1992). The CP,O values for the additives were obtained from SFC separation of a concentrated Soxhlet extract from the film. The overall extractable amount was measured gravimetrically from the Soxhlet extract and with SFE using the monitor FID 1 (figure 2). The estimated values for mL,t/A were calculated with equations (3) and (4), assuming DOPP DLDPE/100.
Figure 2. A simplified scheme of an SFE/SFC apparatus. 1,2 and 3 represent valves which allow SFE in combination with FID 1 and the combination SFE/SFC with FID 2.
Figure 3. SFE/SFC analysis of an OPP film. Carlo-Erba SFC 3000 Chromatograph with a fused silica column 7 m, 0•05 mm i.d. and DB5 as stationary phase: 140°C isotherm; pressure programme: 10•6 MPa (10 min), 1•5 MPa/min to 35 MPa (20 min).
Table 2. Estimated and measured specific migration from OPP into ethanol and overall migration to ethanol and olive oil, 10 days, 40°C.
CP,O mL,t/A (mg/dm2) Additive (mg/g) (mg/dm2) Calc. Exp. Irgafos 168 0•52 0•16 0•10 0•038 Irganox 1076 0•043 0•014 0•013 0•007 Irganox 1010 0•32 0•089 0•008 0•011 Overall 7•2 2•1 - 0•6 + 0-1 (Soxhlet) (ethanol) 6•0 - 1•0 + 1 (SFE) (olive oil)
In the SFC separations described above the retention times (tR) of the solutes increase approximately in proportion to their relative molar masses (Mr ). In order to estimate the overall migration from polyolefins into fat simulants, the Mr values of separated peaks and even peak groups can be estimated, in principle, via calibration of the time scale with a mixture of n-alkanes. Using the prediction in equation (4), the overall migration value can be estimated for any temperature and time with the following formula:
where n is the number of individual and/or peak groups, for which corresponding Mr values are estimated and used in the prediction with equation (4).
Consequences for the evaluation of plastics
A general observation from investigations with different commercial polypropylene films is that the initial overall concentration of extractable materials (CP,O) is < 1% (w/w). Since the majority of the OPP films in use have a thickness < 100 µm, the overall migration from such films cannot exceed 10 mg/dm2. Thus, measuring the overall migration in such cases is pointless.
A second conclusion results from investigations on plastic packages designed for repeated use in microwave ovens. The volatile fractions of constituents in such articles can easily be measured using temperature resistant sorbents such as the polymer Tenax (figure 4) (Fuchs et al. 1991a,b). High resolution gas chromatography allows the detection of individual constituents at low concentrations (CP,O) of about 1 µg/g without too much difficulty. If these substances are of unknown origin one can employ, in an initial investigation, a threshold concentration (figure 4) as a toxicological target. Compounds above this threshold can be investigated with an MS-detector in order to identify their chemical nature. From a 1 mm thick plastic sample the maximal migrating amount of one component with an initial concentration (CP,O of 1 ppm is <0•01 mg/dm2. This corresponds to about 0•06 mg/kg foodstuff. However, after repeated use of the plastic (figure 4) the amount of migrating substances decreases rapidly below the threshold line.
Figure 4. Migration of volatiles from a commercial
article of unknown origin to Tenax during the first and the third heating at 175oC
for 2 h. For conditions of analysis see Fuchs et al. (1991).
........... Component threshold concentration line. The two numbers in the figure represent the overall migration of volatiles into Tenax after the first and third heating, respectively.
The non-volatile cyclic trimer (Castle et al. 1989, Begley and Hollifield 1990) in PET cannot be determined in the way described above with Tenax. But this trimer represents about 90% of the total amount of oligomers in PET and the concentration (CP,O) of the trimer in many investigated PET articles designed for microwave use is approximately 0•34% w/w. For a PET vessel with a thickness of ~ 1 mm, the maximal migrating amount of trimer mL,/A) is then 40 mg/dm2. Assuming migration during the first use to be much higher than 10 mg/dm2, then after a third use only migration less than 10 mg/dm2 is possible. On the contrary, if during the first use less than 10 mg/dm2 migrates, the overall migration limit cannot be reached. A time-consuming measurement of trimer migration is not necessary. The initial (maximum) concentration of the trimer in PET is easily determined by total extraction (e.g. solving the PET sample in hexa-fluorisopropanol and reprecipitation with propanol) followed by HPLC, MS or SCF analysis of the extracted trimer. With the knowledge of its diffusion constants at different temperatures, a sufficiently accurate estimation of migration is possible.
Similar considerations as with the PP films and PET package examples lead to the necessity for an in-depth investigation of the problem of good manufacturing practice. This is the first idea expressed in Article 2 of Directive 89/109/EEC (CEC 1989). A collection of data regarding the main composition and overall extractable amount of plastic constituents can help with the estimation of migration. This can be a considerable asset to both the producers of such articles and for quality control laboratories. Much time and money may be also be saved if studies are made in the evaluation of laminates containing layers of recycling material with unknown impurities which can migrate through the virgin plastic layer (functional barrier) in contact with food.
Another important problem is the organoleptic properties of plastics for food contact. At present, there are only recommendations at the national level of some member states (e.g. a BGA recommendation for materials and articles for microwave ovens). Even with materials such as PET with a very low migrating potential, off-odour problems can occur after long contact with water at room temperature as a result of an increased amount of acetaldehyde.
A final remark in connection with the evaluation of plastics for food contact is the necessity of reconsideration of Directive 85/572/EEC (CEC 1985). According to this Directive, for some categories of foods (dry products) no migration measurement is required. Or, for milk, measuring migration in aqueous simulants only is necessary. Whereas the migration of additives to dry foodstuffs is now an established fact (Schwope and Reid 1988), it must also be mentioned that the partition coefficients of many compounds between many plastics and milk with 3 - 5% fat are similar to 50% ethanol in water (Koszinowski and Piringer 1989, Franz et al. 1993).
LITTLE, A. D., 1983, Final summary report: a study of indirect food additive migration. FDA Contract 223-77-2360, July.
SANER, A., BIEBER, W., FIGGE, K., FRANZ, R., and PIRINGER, O., 1992, Alternative fatty food simulants for migration testing of polymeric food contact materials. Food Additives and Contaminants, 9, 137-266.
BECKER, K., KOSZINOWSKI, J., and PIRINGER, O., 1983, Permeation von Riech- and Aromastoffen durch Polyolefine. Deutsche Lebensmittel-Rundschau, 79, 257-266.
BEGLEY, T. H., and HOLLIFIELD, H. C., 1990, Evaluation of polyethyleneterephthalate cyclic trimer migration from microwave food packaging using temperature-time profiles. Food Additives and Contaminants, 7, 339-346.
BÜCHERL, T., SANER, A., and PIRINGER, O., 1993, Extraktion and Trennung migrierfähiger Komponenten aus Bedarfsgegenständen mit Hilfe superkritischen Kohlendioxides. Deutsche Lebensmittel-Rundschau, 89, 69-71.
CASTLE, L., MAYO, A., CREWS, C., and GILBERT, J., 1989, Migration of poly(ethyleneterephthalate) (PET) oligomers from PET Plastics into foods during microwave conventional cooking and into bottled beverages. Journal of Food Protection, 52, 337342.
CEC, 1982, 82/71 I/EEC. Council directive of 23 October 1982 laying down the basic rules necessary for testing migration of the constituents of plastic materials and articles intended to come into contact with foodstuffs. Official Journal of the European Communities. No L 296, 23.10.82, 26-30.
CEC, 1985, 85/572/EEC. Council directive of 19 December 1985 laying down the list of simulants to be used for testing migration of constituents of plastic materials and articles intended to come into contact with foodstuffs. Official Journal of the European Communities. No L 296, 23.10.82, 26-30.
CEC, 1989, 89/109/EEC. Council directive of 21 December 1989 on the approximation of the laws of the Member States relating to materials and articles intended to come into contact with foodstuffs. Official Journal of the European Communities. No L 40, 11.2.89, 38-44.
CEC, 1990, 90/128/EEC. Council directive of 19 December 1985 laying down the list of simulants to be used for testing migration of constituents of plastic materials and articles intended to come into contact with foodstuffs. Official Journal of the European Communities. No L 75, 21.3.1990, 14-40.
FIGGE, K., 1988, Migration-Theorie and praktische Beispiele. Migration bei Kunsistoffverpackungen, edited by G. H. Hauschild and E. Springler (Stuttgart: Wissenschaftliche Verlagsgescllschaft mbH), pp. 33-92.
FIGGE, K., and HILPERT, H. A., 1991, Migration of different additives from polyolefine specimens into ethanol 95% by vol., test fat HB307 and olive oil-a comparison. Deutsche Lebensmittel-Rundschau, 87, 1-4.
FUCHS, M., KLUGE, S., RÜTER, M., WOLFF, E., and PIRINGER, O., 1991a. Methode zur Bestimmung der Globalmigration and Prüfung der sensorischen Eigenschaften von Bedarfsgegenst9nden zur Er9rmung von Lebensmitteln im Mikrowellenofen. 1. Teil. Deutsche Lebensmittel-Rundschau, 87, 273-276.
FUCHS, M., KLUGE, S., RÜTER, M., WOLFF, E., and PIRINGER, O., 1991b. Methode zur Bestimmung der Globalmigration and Prüfung der sensorischen Eigenschaften von Bedarfsgegenst9nden zur Erw9mung von Lebensmitteln im Mikrowellenofen. 2. Teil. Deutsche Lebensmittel-Rundschau, 87, 311-316.
FRANZ, R., LEE, K. T., KNEZEVIC, G., WOLFF, E. and PIRINGER, O. 1992, Measuring and evaluation of the global migration from food contact materials into food: a comparison between official EC-techniques and alternative methods. International Zeitschrift fur Lebensmittel-Technik, Marketing, Verpackung and Analytik, 43, 291-296.
FRANZ, R., O'NEILL, E., and TUOHY, J. J. 1993, Migration of styrene monomer from a polymeric packaging material in contact with ethanol/water mixtures and milk. International Dairy Journal (in press).
KOSZINOWSKI, J., 1986a, Diffusion and solubility of hydroxy compounds in polyolefines. Journal of Applied Polymer Science, 31, 2711-2720.
KOSZINOWSKI, J., 1986, Diffusion and solubility of n-alkanes in polyolefines. Journal of Applied Polymer Science, 31, 18051826.
KOSZINOWSKI, J., and PIRINGER, O., 1986, Influence of the packaging material and the nature of the packed goods on the loss of volatile organic compounds-permeation of aroma components. Food Packaging and Preservation, edited by M. Mathlouthi (London: Elsevier Applied Science Publishers), pp. 325-340.
KOSZINOWSKI, J., and PIRINGER, O, 1989, Der Einfluß von Verteilungsvorg9ngen in System Pack Packstoff/Füllgut auf die Qualität kunststoffverpackter, flüssiger Lebensmittel and Kosmetika. Verpackungs-Rundschau, Technisch-wissenschaftliche Beilage, 40, 39-44.
PIRINGER, O., 1992, Verpackungen für Lebensmittel-Eignung, Wechselwirkungen, Sicherheit (New York, Basel, Cambridge, VCH 1993 Weinheim).
SCHWOPE, A. D., and REID, R. C., 1988, Migration to dry foods. Food Additives and Contaminants, 5, Supplement No. 1, 445454.
TILL, D., SCHWOPE, A. D., EHNTHOLT, D. J., SIDMAN, K. R., WHELAN, R. H., SCHWARTZ, P. S., and REID, R. C., 1987, Indirect food additive migration from polymeric food packaging. CRC Critical Reviews in Toxicology, 18, 215-243.
VAN BATTUM, D., and VAN LIEROP, J. B. H., 1988 Testing of food contact materials in the Netherlands. Food Additives and Contaminants, 5, Supplement No. 1, 381-395.
If you have come to this page from an outside location click here to get back to mindfully.org