Element residues in food contact plastics and their migration into food simulants, measured by inductively-coupled plasma-mass spectrometry
Food Additives and Contaminants v.12, n.5, 1995
PETER J. FORDHAM1, JOHN W. GRAMSHAW1, HELEN M. CREWS2 and LAURENCE CASTLE2,§
1 Procter Department of Food Science, University of Leeds, Leeds LS2 9JT, UK; 2 Ministry of Agriculture, Fisheries and Food, CSL Food Science Laboratory, Norwich Research Park, Colney, Norwich NR4 7UQ, UK § To whom correspondence should he addressed.
(Received 4 July 1994; revised 6 September 1994; accepted 16 September 1994)
Polymers intended for food contact use have been analysed for inorganic residues which can be attributed to a range of substances employed as polymerization aids (e.g. catalysts), or to additives incorporated into the polymer to fulfill a specific task (e.g. lubricants). The migration of these residues into food simulants was studied. Residues were determined by using the multi-element capability of Inductively-Coupled Plasma-Mass Spectrometry (ICP-MS). Semi-quantitative analysis was carried out on acid digests of polymer materials, obtained by microwave heating in sealed Teflon containers. Limits of detection in the polymer were generally less than I mg/kg. Migration experiments were carried out with three food simulants and under two sets of conditions. Analysis for element residues was performed directly or, in the case of olive oil, on an emulsion of the simulant. Migration of certain elements into aqueous simulants was observed: Zr from polystyrene (650 µg/kg), Sb from polyethylene terephthalate (4 µg/kg), and Mg from acrylonitrile/butadiene/styrene copolymer (50 µg/kg). In ail cases, where limits of detection permit, the levels which migrated from polymer to stimulant were less than proposed limits on migration.
Keywords: aids to polymerization, element residues, migration, ICP-MS, plastics
As part of the moves towards harmonization of regulatory controls governing materials and articles intended for food contact use within the European Community and, more specifically, those relating to the constituents of plastic materials and articles, interest has focused recently on the group of substances referred to hereafter as aids to polymerization. These are defined as substances necessary for the manufacture of polymers, e.g. catalysts, initiators and solvents, but which exclude monomers and other starting substances. Many papers have been published on the determination of additives in food contact polymer materials and their migration into foods and simulants (Bieber et al. 1984, Gilbert et al. 1986) plus some which include work on the migration of catalyst residues into food simulants (Ashby 1988) but more information is needed to assess possible schemes of control for these particular substances. The work reported here covers the determination of residual elements derived from aids to polymerization in polymers and food simulants.
A large number of catalysts within the polymer industry are organometallic or inorganic compounds, for example, the so-called Ziegler-Natta and chromium-based catalysts used for the co-ordination polymerization of polyolefins; and antimony-based catalysts used in the production of poly(ethylene terephthalate). Hence, for the initial wide-range screening for catalyst residues in polymers, ICP-MS was the preferred analytical technique because of its ability to generate multi-element semi-quantitative data (Ekimoff et al. 1989, Amarasiriwardena et al. 1990), and microwave digestion was adopted as a rapid method of sample pre-treatment. Microwave digestion of polymers is a relatively recent application (Marshall and Franks 1990, Marshall et al. 1990) of a preparative technique which has been used for many other types of sample.
A list of proposed limits for inorganic residues in polymer materials has been drawn up following a review of existing national requirements and to explore means for the control of residues of aids to polymerization (EEC 1992). Some of these recommendations are specified as limits on the composition of the polymer, but most are limits on the migration of residues into food simulants (see table l). Thus, many residues that are considered to be toxic but do not migrate into foodstuffs from the polymer may be allowed to be present in the polymer. Consequently, migration studies are an essential aspect of any research which seeks to formulate an appropriate system of control for aids to polymerization; semi-quantitative ICP-MS was employed as the analytical technique for analysis of food simulants following exposure to polymers under different conditions of temperature and time.
For many ICP-MS instruments, a plot of mass of analyte submitted, against sensitivity gives a fairly smooth response curve, once the isotopic abundance and degree of ionization have been taken into account. The resulting response curve can be used to calibrate the instrument and generate semi-quantitative data. In practice, the curve is usually defined by six to eight elements spread across the atomic mass range and a second order curve is fitted to the data which can then be stored. An internal standard positioned in the centre of the mass range (115ln in this case) corrects for variations in instrument sensitivity and the concentrations of a large number of elements can be determined in an unknown sample by reference to the stored response curve. However, the accuracy of the data produced is dependent on the element measured and the sample matrix, and can be variable, but if a degree of quality assurance is undertaken so that the performance is closely monitored then the method is highly suitable for survey work.
Forty-five food contact grade polymers were obtained from European manufacturers and were supplied in poly(ethylene) bags and then transferred to acid-washed poly(ethylene) containers for storage (21 were in pellet form and therefore suitable for migration studies, all others were in powder form). Concentrated sulphuric and nitric acids (Aristar grade. BDH Ltd) were used for polymer digestion, and deionized distilled water was used to carry out dilutions.
For the lower temperature migration experiments (10 days/40°C), acid-washed (5% nitric acid) screw-top low density poly(ethylene) containers of 125 ml capacity were used, and for the higher temperature work (2h/100°C), acid-washed screw-top poly(tetrafluoroethylene) conical flasks of 125 ml capacity were preferred. Olive oil of a suitable grade for overall migration testing (EEC 1985) was supplied by Pira International (Leatherhead), acetic acid (Aristar grade) was supplied by BDH Ltd and ethanol (absolute alcohol) was acquired from Hayman Ltd.
Table 1. Proposed limits for inorganic residues (EEC 1992).
Proposed limit in polymer Element (mg/kg) Gallium 20 Germanium 10 Iridium 20a Manganese 60 Osmium 20 Palladium 20 Platinum 20a Rhodium 20a Selenium 10 Proposed limit in food simulants (mg/kg) Aluminium 60 Antimony 0.05 Arsenic 0.01 Barium 0.2 Boron 12 Bromides (as Br) 0.5 Cadmium 0.005 Calcium 60 Cerium 1.0 Chromium (trivalent) 60 Chromium (hexavalent) nd Cobalt 0.1 Copper 30 Fluorides (as F) 0.5 Iodides (as 1) 1.0 Lead 0.01 Lithium 0.6 Magnesium 60 Mercury 0.005 Nickel 0.1 Phosphorus oxides 60 Rubidium 1.0 Silicon 60 Tin (inorganic compounds) 60 Tin (organic compounds) 0.05b Titanium 60 Vanadium 0.1 Zinc 60 Zirconium 01 a For silicone plastics only a limit of 80mg/kg is proposed. b 0.02 for di-n-octyltin compounds. nd, not detectable.
Table 2. ICP-MS operating parameters. ICP radiofrequency forward power 1350 W
Carrier gas flow 0 81/min Auxiliary gas flow 0-5 l/min Coolant gas flow 131/min Number of channels 2048 Number of scan sweeps 100 Dwell time 320 µs Detector mode Pulse Sample uptake rate 0.75 ml/min Nebulizer type Cross-flow Sample cone Platinum (I mm orifice) Skimmer Nicone® (0.75 mm orifice) Mass range 6-240 (skipped regions: 12-235, 27.5-41.5 and 79.5-80.5) Isotopes selected 24Mg, 27Al, 52Cr, 59Co, 63Cu, 66Zn, 72Ge, 90Zr, 121Sb and 208Pb for measurement
The indium standard used for semi-quantitative determinations was Spectrosol grade indium nitrate (1000 mg/l3) in nitric acid (0-5 mol/1) (BDH Ltd).
An MDS-8 I D Microwave Digestion System (CEM Corporation) was used to carry out digestion of polymer materials. The digestion vessels were fabricated from Teflon® PFA; screw tops fitted with safety valves allowed pressure release at 7 X 105 Pa, and the instrument was programmable for time and microwave power.
For the migration experiments, a Gallenkamp Oven 300 (Plus Series) was used to equilibrate vessels at the required temperatures.
A VG Plasmaquad PQ I instrument (with extended dynamic range) was used for all ICP-MS measurements. The conditions used are given in table 2. Measurements were made in the scanning mode, with post-run data manipulation being handled by the calculation software module and concentrations calculated using semi-quantitative analysis algorithms contained therein. Semi-quantitative data were produced by the software based on an indium ("In) internal standard and using the response curve generated from a multi-element solution containing 50 pg/1 of each of Be, Mg, Co, In, Ba and U.
Determination of polymer composition.
A two-step digestion procedure was employed to digest polymers. An accurately-weighed sample of polymer (0.5 g) was placed into a digestion vessel. Concentrated sulphuric acid (8 ml) and nitric acid (3 ml) were added and the cap fastened to finger-tight only. Vessels were than placed on the microwave oven turntable until the carousel held a total of 12, and venting tubes were attached. The instrument was programmed for 15 min at 75% power for the first stage of digestion. The sample was allowed to cool to room temperature before an additional 5 ml of concentrated nitric acid was added and the cap replaced and tightened to 330 N to with the specifically designed capping station. The vessel was returned to the turntable for the second digestion stage: 30 min at 100% power with a full carousel, and allowed to cool to room temperature before venting and transferring to an appropriate container. Most polymer samples could be digested using this procedure but a small number of poly(vinyl chloride) (PVC) samples required a slightly modified procedure which differed from the one described above in that just two vessels were placed in the microwave oven and the first and second programs were: (i) 20 min at 40% power and (ii) 20 min at 40% power followed immediately by a third stage: (iii) 5 min at 50% power. For multi-element, semi-quantitative analysis, digests were diluted ten-fold in deionized distilled water and indium internal standard was added to give a concentration in the final diluted sample of 100µg/l, prior to introduction into the plasma. Reagent blanks were also taken through the entire procedure.
Migration experiments were carried out using three food simulants (olive oil, aqueous acetic acid (3% w/v) and aqueous ethanol (15% v/v)), under two sets of test conditions: 10 days exposure at 40°C and 2h exposure at 100°C. Of the 45 polymer materials supplied by industrial sources, 21 were in pellet form. These pellets were generally spherical or cylindrical in form and, since variability was minimal, the surface area of each pellet and hence of a sample of pellets could be calculated to a reasonable degree of accuracy. Thus, migration experiments could be performed with relative ease and data on migration generated for those polymer materials in pellet form.
The migration experimental procedure was as follows. Polymer material (3 g, or the maximum amount of material allowing good distribution and complete submergence of pellets within the simulant) was weighed into the appropriate plastic ware and 20 ml of food simulant added. Samples were then subjected to one of the two experimental conditions of temperature and time, following which simulants were analysed for metal element residues. Aqueous acetic acid (3% w/v) could be analysed directly by ICP-MS but aqueous ethanol (15%v/v) was diluted in deionized distilled water prior to analysis. In agreement with the findings of other workers (Longerich 1989), some dilution of the aqueous ethanol matrix was required to reduce suppression of signal intensity. For the olive oil simulant, a suspension method, used previously for determining inorganic residues in crude oil (Lord 1991), was adopted: 0.2g olive oil was weighed accurately into a 20 ml volumetric flask and mixed with 200 µl of tetralin (a co-solvent), plus 0.5 ml of the surfactant Triton X-100. The mixture was then diluted to 20 ml with distilled water to give an emulsion of short-term stability (approx. 1.5 h). Samples were shaken vigorously immediately prior to analysis to ensure homogeneity. An indium internal standard was added at 100,ug/I to all three simulants to enable the software to generate multi-element, semi-quantitative data, and background subtraction of simulant blanks was carried out. Simulant blanks were prepared by placing simulant in the appropriate container (see Materials) and exposing under the test conditions in the absence of any plastics test samples. Analysis of simulant before and after exposure in this way revealed that the migration containers released negligible element residues. Levels in simulants (following migration) were calculated from migration measured on a surface area basis (µg/dm2) and then converted to mg element per kg simulant values using the standard migration test ratio of 6 dm2 = 1 kg simulant.
A total of 73 defined isotopes covering 69 elements were monitored during analysis of the polymer digests in semi-quantitative mode. The data for those elements showing a signal above the limit of detection (taking the limit of detection to be three times the standard deviation of the response for five blank samples; see table 3) have been summarized in tables 4-10, where figures for the polymer content are expressed in mg element per kg polymer material. Ten PVC samples and two poly(vinylidene chloride) (PVdC) samples were analysed and found to contain no significant residual elements (except for 560 mg/kg aluminium in one PVdC sample). Hence, tables for these samples are not given.
Table 3. Limits of detection for elements present in polymers and in food simulants, following appropriate method of sample preparation.
Detection limits . Aq. acetic Polymer (mg/kg a) acid Aq. ethanol Olive oil Element Method 1b Method 2b (µg/kg)c (µg/kg)c (mg/kg)c Mg 1-5 12 0.1-0.3 223-4000 0.2-13 Al 03-44 25 0.08-0.6 4-168 1-193 Cr 0.2-0.3 1 03-3 54-1714 03-30 Mn 0.04-0.1 0.3 0.02-0.08 0.5-2 0.06-3 Co 0.03-0.06 0.1 0.01-0.02 0.1-1 0.005-0.01 Cu(63Cu) 0.1-0.2 0.6 0.04-0.08 0.5-3 0.04-0.6 Cu(65CU) - - 0.6 8 7 Zn(66Zn) - - 0.2-0.4 2 - Zn(68Zn) 0.7-1 0.8 0.3-0.8 4-9 0.5-7 Ge 0.8-2 12 0-04-0.07 0.5-2 0.02-0.05 Zr 0.02-0.09 0.05 0.01-0.03 0.09-0.4 0.01-0.02 Sb 0.02-0.2 0.09 0.007-0.01 0.05-0.2 0.004-0.2 Pb 0.3-0.04 0.06 0.009-0.03 0.1-0.4 0.005-0.2 a mg element per kg polymer. b The two methods of polymer digestion are described in the experimental section. c µg (or mg) element per kg simulant (calculated from migration measured on a surface area basis (µg/dm2) and then converted to mg element per kg simulant values using the standard migration test ratio of 6dm2= I kg simulant.
Interpretation of the accumulated data required particular care as a number of the isotopes scanned had very high apparent background levels, because of the presence of species derived from the nitric and sulphuric acids used in the digestion procedure and to residual carbon derived from the digested polymer matrix. The occurrence of interference peaks was noted in the very earliest publications on plasma source mass spectrometry and more detailed accounts are now available (Tan and Horlick 1986). Awareness of these problems and their origins is an essential aspect of the interpretation of ICP-MS data. In particular, the presence of polyatomic ion interferences derived from the sulphuric acid and the degradation of the nickel sampler cone (which is the usual material for use in the interface regions of the ICP-MS) are two of the chief reasons why the use of sulphuric acid for sample preparation procedures prior to analysis by ICP-MS is generally avoided (Jarvis et al. 1989). However, complete digestion of polymers, which are generally designed to be resistant to chemical attack, was found to necessitate the use of both sulphuric and nitric acids. The use of a more corrosion-resistant platinum cone allowed the analysis of such samples without having severe detrimental effects on the instrument. Subtraction of the background spectrum for a reagent blank was applied to all measurements and is an effective correction for minor interferences but problems arise when the intensity of the analyte is of a similar magnitude to the intensity of the interference, reducing the detection limit. Hence, to avoid these pitfalls, selection of the isotope for measurement requires care and, in certain cases, where such problems are severe useful data cannot be obtained. Figures for calcium, iron and titanium are omitted owing to interferences coinciding with the isotopes monitored.
Table 4. Elements in low density poly(ethylene) and their migration into food simulants.
Mg Al . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 1 LDPE* <1 3.9 # <340 # # 0.41 2.1 1 <4 # # 2 LLDPE* 23 2.1 # <250 # # 7.2 0.89 # 3.7 # # 3 LDPE* 1.5 0.59 # <230 # # 8.3 0.48 # <3 # # 4 LDPE* <1 0.33 # <210 # # 1.2 0.49 # <3 # # 5 LDPE* <1 0.87 # <190 # # <2 1.3 # 3.1 # # I: 40ºC/10 days, II: 100°C/2 h, *: polymers in pellet form, #: experiment not performed. a LDPE-low density poly(ethylene); LLDPE-linear low density poly(ethylene).
Table 5. Elements in high density poly(ethylene) and their migration into food simulants.
Al Cr . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 6 HDPE* 0.51 4.2 <0.06 14 <200 <130 3.7 2.2 11-4 <390 <5 <3 7 HDPE 13 - - - - - <1 - - - - - 8 HDPE <12 - - - - - <1 - - - - - 9 HDPE 16 - - - - - <1 - - - - - 10 HDPE 34 - - - - - <1 - - - - - 11 HDPE <12 - - - - - 2.5 - - - - - 12 HDPE <12 - - - - - 4.3 - - - - - Zr . Migration into food simulants . 3% acetic acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) . Sample typea (mg/kg) I II I I II 6 HDPE* <0.08 <0.01 <0.02 <0.4 <0.02 <0.02 7 HDPE 0.28 - - - - - 8 HDPE 1.9 - - - - - 9 HDPE 29 - - - - - 10 HDPE 42 - - - - - 11 HDPE 0.03 - - - - - 12 HDPE 0.03 - - - - - I: 40ºC/10 days, II: 100°C/2 h, *: polymers in pellet form. a HDPE-high density poly(ethylene).
Table 6. Elements in poly(propylene) and their migration into food simulants.
Mg Al . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 13 PP <44 - - - - - 22 - - - - - 14 PP <44 - - - - - 26 - - - - - 15 PP <44 - - - - - 26 - - - - - 16 PP <44 - - - - - l3 - - - - - 17 PP/PE* 23 0.45 <0.1 <270 <12 <8 115 1.0 <0.1 <3 <170 <120 Cr . Migration into food simulants . 3% acetic acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) . Sample typea (mg/kg) I II I I II 13 PP <1 - - - - - 14 PP <1 - - - - - 15 PP <1 - - - - - 16 PP 1.3 - - - - - 17 PP/PE* <1 <3 3 <470 <5 <5 I: 40ºC/10 days, II:100°C/2h, *: polymers in pellet form a PP-poly(propylene); PP/PE-propylene/ethylene copolymer.
Table 7. Elements in poly(styrene) and their migration into food simulants.
Mg Al . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 18 PS* <1 <0.2 <0.1 3300 <4 0.26 19 <0.4 <0.1 <10 <35 <2 19 PS* <1 <0.2 <0.1 4500 <4 <0.1 0.60 <0.3 <0.1 9.5 <33 3.2 20 PS* 23 <0.2 1.1 5200 <5 <0.1 3.3 <0.5 <0.1 <13 <46 3.9 21 PS* <1 <0.2 <0.1 4700 <4 <0.1 0.33 <0.3 <0.1 <9 <32 <2 22 PS* <1 <0.2 <0.1 6900 <5 <0.2 <0.3 <0.5 <0.1 18 <47 <3 23 PS* <1 <0.2 <0.1 6100 <5 <0.2 <0.3 <0.4 <0.1 14 <43 <3 24 PS* <1 0.25 <0.1 <4000 <6 <0.2 <0.3 <0.5 <0.1 <13 <49 <3 Co Cu . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 18 PS* 0.60 <0.01 <0.01 0.55 <0.01 <0.01 16 4.2 6.0 28 <0.4 <0.03 19 PS* 0.06 <0.01 <0.01 <0.3 <0.01 <0.01 3.5 1.8 13 16 <0.4 <0.03 20 PS* < 0.06 <0.01 <0.01 <0.4 0.02 <0.01 11 0.14 <0.05 <0.5 <0.6 <0.03 21 PS* < 0.06 <0.01 <0.01 <0.3 <0.01 <0.01 1.4 0.10 0.05 036 <0.4 <0.03 22 PS* < 0.06 <0.01 <0.01 0.45 <0.01 <0.01 7.4 0.24 0.07 <0.5 <0.6 <0.04 23 PS* < 0.06 <0.01 <0.01 <0.3 <0.01 <0-01 3.6 0.22 0.07 0.74 <0.5 <0.04 24 PS* < 0.06 <0.01 <0.01 <0.4 <0.01 <0.01 8.1 0.28 0.22 0.65 <0.6 <0.4 Zn . Migration into food simulants . 3% acetic acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Sample typea (mg/kg) I II I I II 18 PS* 67 4.2 5.0 82 <5 <0.3 19 PS* 70 10 <0.2 <4 <5 <0.3 20 PS* 3.0 35 50 650 <7 <0.3 21 PS* 91 <0.5 <0.2 18 <5 <0.3 22 PS* 94 <0.8 <0.3 31 <7 <0.5 23 PS* 110 <0.7 <0.3 18 <6 <0.4 24 PS* 97 <0.8 <0.3 23 <7 <0.5 I: 40ºC/10 days, II: 100ºC/2h. *: polymers in pellet form. a PS--poly(styrene)
Table 8. Elements in poly(ethylene terephthalate) and their migration into food simulants.
Mg Al . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 25 PET* <1 0.51 <0.1 <210 <13 <10 0.66 0.78 <0.1 <160 <200 <140 26 PET* 5.9 2.8 <0.1 <220 <0.9 <8 620 0.75 <0.1 <170 <1 <220 Co Ge . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 25 PET* 58 0.08 0.05 0.13 <0.01 <0.01 0.95 <0.07 <0.05 <0.2 <0.1 <0.09 26 PET* 33 0.24 0.15 0.24 <0-01 <0.01 14 0.25 <0.05 <0.2 <0.1 <0.07 Sb . Migration into food simulants . 3% acetic acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Sample typea (mg/kg) I II I I II 25 PET* 160 2.7 3.9 23 <0.01 <0.01 26 PET* 230 1.2 2.6 1.1 <0.01 <0.01 I: 40°C/10 days, II: 100°C/2h, *: polymers in pellet form. a PET-poly(ethylene terephthalate).
Table 9. Elements in poly(methyl methacrylate) and their migration into food simulants.
Mg Zn . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 27 PMMA* 5.2 0.53 0.26 <160 <0.7 <9 <1 4.0 4.0 7.6 <0.7 <0.8 28 PMMA <1 - - - - - <0.8 - - - - - 29 PMMA <1 - - - - - <0.8 - - - - - I: 40°C/10 days, II: I00°C/2h, *: polymers in pellet form. a PMMA-poly(methyl methacrylate).
Table 10. Elements in various copolymers and mixed polymers and their migration into food simulants.
Mg Co . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 42 AA/AN/S* 130 31 13 230 <0.6 <10 <0.03 <0.01 <0.01 <0.1 <0.01 <0.01 43 ABS* 220 51 19 <180 <0.7 <11 <0.03 <0.01 <0.01 <0.1 <0.01 <0.01 44 MA/AN/S* <5 0.87 0.18 <60 <0.3 <4 <0.03 <0.01 <0.01 <0.03 <0.01 <0.01 45 PS/PPO* <5 1.5 033 <140 <0.5 <8 0.42 0.01 <0.01 <0.08 <0.01 <0.01 Cu Zn . Migration into food simulants Migration into food simulants . 3% acetic 3%acetic acid 15%ethanol Olive oil acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Content (µg/kg) (µg/kg) (mg/kg). Sample typea (mg/kg) I II I I II (mg/kg) I II I I II 42 AA/AN/S* <2 0.92 <0.04 <0.09 <0.06 <0.2 <1 <0S 0.8 <3 <0.7 <1 43 ABS* <2 1 0.12 <1 <0-08 <0.2 1.5 <0.6 <0.3 <3 <0.8 <1 44 MA/AN/S* <2 0.32 <0.01 <0.3 <0.03 <0-09 <1 <0.2 <0.1 <1 <0.3 <0.4 45 PS/PPO* 3.60 0.79 <0.03 <0.8 <0.06 <0.2 50 <0.5 <0.2 <2 <0.6 <0.8 Sb . Migration into food simulants . 3% acetic acid 15%ethanol Olive oil Polymer Content (µg/kg) (µg/kg) (mg/kg) Sample typea (mg/kg) I II I I II 42 AA/AN/S* <0.02 0.010 <0.01 <0.03 <0.01 <0.01 43 ABS* <0.02 <0.01 <0.01 <0.04 <0.01 <0.01 44 MA/AN/S* <0.02 0.010 0.020 <0.01 <0.01 <0.01 45 PS/PPO* 0.10 <0-01 <0.01 <0.03 <0.01 <0.01 I: 40°C/10 days, II: 100ºC/2h, *: polymers in pellet form. a AA/AN/S-acrylic acid ester/acrylonitrile/styrene copolymer; ABS-acrylonitrile/butadiene/styrene copolymer; MA/AN/S-methacrylic acid ester/acrylonitrile/styrene copolymer; PS/PPO-poly(styrene)/poly(phenylene oxide) mixed polymer.
Of the interferences derived from residual carbon of the digested polymer, the occurrence of 40A12C, which coincides with the 52Cr isotope, is of most concern. The limit of detection for Cr in the polymer was increased to 1 mg/kg for this reason following the observation of isotope ratios, and reported figures for polymer composition will therefore be slightly elevated because of the presence of 40Ar12C; monitoring of the 53Cr isotope (lower abundance (9.5%)), which is free of carbon-related interferences, might be considered for future work.
Analytical performance. Because of the absence of suitable polymer-based certified reference materials the performance of the method was determined from the analysis of a polymer reference material of known elemental composition and a certified non-polymer reference material (NBS 1572 Citrus Leaves). Figures for the calculated precision and bias are given in tables 11 and 12, respectively. In general, the bias figures were found to be within a factor or two of the certified/known values for most of those elements which were not subject to interference.
Analysis of food simulants, following migration experiments, was also performed in semi-quantitative mode for the same range of isotopes and elements and, again, those elements showing a significant signal above background (taking the limit of detection to be defined as three times the standard deviation of the response for five blank samples) are presented in tables 4-10. Limits of detection for the two aqueous simulants were generally low but, because of the heavy dilution factor necessary in preparing suspensions, the detection limits for olive oil were rather higher and are expressed in mglkg for this reason. Of the 45 polymers for which compositional data were obtained, only 21 were in pellet form and therefore suitable for migration experiments (all others were in powder form). It was considered that migration experiments for ethanol at 100°C/2 h were unnecessary as these are conditions that are unlikely to be encountered in real use.
Fewer interferences occurred for the three food simulants when compared with the polymer digests and the blank spectra of the simulants were broadly similar with the following notable interferences: 40Ar1H2(42K,42Ca), '2C 1602(44Ca), 12C16O21H(45SC), 40A12C(52Cr), 40Ar16O(56Fe) and 40Ar16OH(57Fe). However, the relative severity of these interferences did vary considerably between simulants.
Food simulants (following the appropriate sample preparation) were spiked with multi-element standards at 100 µg/l and used to evaluate the performance of the method. Replicate analysis (n = 4) in each of the three simulants gave precision figures of between 1 and 12% for 19 of the 27 elements listed in the second part of table 1 (proposed migration limits) where interferences were not gross. Aqueous acetic acid (3%), aqueous ethanol (3%, following dilution) and heavily diluted oil suspension showed no marked influence on the precision and accuracy when compared with aqueous standards (again, where interferences were not excessive) and the performance of the semi-quantitative calculations generated was monitored using an aqueous multi-element standard. The accuracy for 24 elements across the mass range varied from - 44% to + 200% (n = 8).
Table 11. Precision and bias for uncertified Polymer Reference Standard (low density poly(ethylene) (n = 3)) following microwave digestion with H2SO4/HNO3 and semi-quantitative measurement by ICP-MS.
Known Isotope value Mean Std devn Element measured (mg/kg) (mg/kg) (mg/kg) %RSD Magnesium 24Mg 567 137.2 1.2 1 Aluminium 27Al 40.7 83-5 1.4 2 Chromium* 52Cr 33-6 763 1.3 2 Cobalt 59Co 31.8 64.2 1.1 2 Zinc 68Zn 44.0 49,0 0.7 2 Antimony 121Sb 63.4 50.9 2.8 5 Lead 208Pb 68.3 603 3.5 6 *Subject to interference.
Table 12.Precision and bias for NBS Reference Standard 1572 Citrus Leaves (n = 4).
Certified Certified Isotope value Mean Std devn element measured (mg/kg) (mg/kg) (mg/kg) %RSD Aluminium 27Al 92 182 25 14 Arsenic 75As 3.1 1.75 0.28 18 Barium 138Ba 21 19.2 2.6 14 Calcium* 44Ca 3.15% 6.93% 3.3% 47 Caesium 113Cs (0.098) 0074 0021 28 Copper 63Cu 16.5 29.0 5.7 20 Iron* 57Fe 90 529 193 36 Lanthanum 139La (0.19) 0.138 0040 29 Lead 208Pb 133 13.6 6.0 44 Magnesium* 24Mg 0.58% 204% 0.1% 4 Manganese 55Mn 23 493 6.8 14 Rubidium 85Rb 4.84 582 0.54 9 Strontium 88Sr 100 136 35 26 Zinc 68Zn 29 21.4 8.7 41 Values in parentheses are uncertified. * Subject to interference.
Comparison of data with proposed limits
Of those elements for which compositional limits have been put forward, germanium was found to be present at approximately the proposed limit in a single PET sample and all other elements were present at less than proposed limits. Of the other residues observed in the polymers, aluminium, magnesium and zinc compounds are generally of low toxicity and therefore of little concern, provided that they do not exceed the overall migration limit of 60 mg/kg. Of the remaining residues present, antimony, cobalt and zirconium are more toxic and have lower recommended migration limits (see table l). Hence, the levels of migration of these residues are of more interest.
Of the three simulants used in migration experiments, the two aqueous simulants are most likely to extract metal element residues and the results generally confirm this supposition.
Residues of aluminium, magnesium, chromium and zinc were present at much less than the migration limit of 60mg/kg proposed as the limit for these substances, under all five sets of conditions and where limits of detection permit. Migration levels of cobalt, copper, zirconium and antimony were also below their respective proposed limits. The high levels of antimony and cobalt observed in poly(ethylene terephthalate) showed little tendency to migrate. The sensitivity required of the methods to meet the proposed migration limits is sufficient for most of the elements listed in table l. The exceptions include Cr(VI) (in aqueous ethanol and olive oil), which has a much reduced migration limit compared with Cr(III), and aluminium in aqueous ethanol and olive oil for which limits of detection also exceed the proposed migration limits. Aluminium is a notoriously difficult element to measure by ICP-MS at low levels, particularly in organic matrices where CNH is an interference which gives rise to poor limits of detection.
Origins of element residues
In many cases, the residues observed in the polymers analysed can be tentatively attributed to the catalyst residues expected for each type of polymer, or else to other processing aids. Suspected origins are given below.
Ethylene may be polymerized by a high pressure, high temperature process, which employs a simple organic initiator compound to give a low density poly(ethylene) (LDPE) material. Hence, inorganic catalyst residues would not be expected to be present in the final product. Linear low density poly(ethylene) (LLDPE) and high density poly(ethylene) (HDPE) are both produced commercially by co-ordination polymerization under moderate temperature and pressure conditions. Catalysts employed are usually either Ziegler-Natta catalysts (which are complexes formed by the interaction of alkyls of metals of groups I-III in the Periodic Table with halides and other derivatives of transition metals in groups IV-VII) or supported metal oxide catalysts (where typical catalyst compositions include oxides of chromium, molybdenum, cobalt and nickel metals, supported on silica, alumina, titania, zirconia, or activated carbon). Hence, residues in samples 2, 6 and 7-12 may be attributed to the catalysts and catalyst supports described above. Poly(propylene) (PP) is produced by co-ordination polymerization processes similar to those employed for HDPE and LLDPE. Hence, the residues observed in samples 13-17 will have similar origins. Poly(styrene) is manufactured by a range of processes based on the addition polymerization of styrene. Many organic initiators and organic processing aids are used in production, but the use of inorganic compounds is likely to be confirmed to stearates (samples 2t1--24 are known to contain zinc stearate as a lubricant) and a few other substances. Residual magnesium in sample 20 may be derived from processing aids. Poly(ethylene terephthalate) (PET) is usually produced by condensation polymerization of terephthalic acid and ethylene glycol, which is carried out in two stages. A catalyst is added (usually antimony oxide, antimony acetate or germanium oxide) for the second, polycondensation stage and a cobalt compound is also added, as an accelerator, during production. Hence, the high levels of antimony and cobalt present in samples 25 and 26 can be assumed to be residues derived from catalyst and accelerator, respectively. As for PS, poly(methyl methacrylate) (PMMA), poly(vinyl chloride) (PVC) and poly(vinylidene chloride) (PVdC) are manufactured by addition polymerization which generally does not involve aids to polymerization containing metal elements and, consequently, element residue levels are low (the high level of aluminium present in one of the PVdC samples (not tabulated) is considered to be a processing aid residue). Other polymers: magnesium and zinc in copolymer samples 42, 43 and the mixed polymer sample, 39, are also thought to be residues derived from processing aids.
In the main, the method described achieves the performance required to enable rapid multi-element screening, of a semi-quantitative nature, for elements in polymers and food simulants to meet the sensitivity required by the proposed limits on food contact plastics. Particular care has been taken to assess the quality of the data generated. The accuracy of the method, as reflected in the bias figures, is not ideal but deployment of this ICP-MS approach is vindicated when one considers the volume of useful data generated for such a wide range of elements. Set in this context the use of semi-quantitative ICP-MS is to be recommended for the task at hand and with the advent of high resolution ICP-MS instrumentation, the problems of isobaric overlap from interferences may be eliminated in the future. If the proposed limits become regulations then, for enforcement purposes, more affordable analytical techniques, such as flame atomic absorption spectroscopy, would need consideration.
Residual elements were observed in many of the polymers analysed but most showed little tendency to migrate into food simulants. In all cases, where limits of detection permit, the levels of residual elements in the polymers analysed, and their migration into food simulants, were less than the proposed composition and migration limits, respectively.
The authors gratefully acknowledge the assistance of those European manufacturers which supplied the polymer samples required to undertake this work.
AMARASIRIWARDENA, C. J, GERCKEN, B., ARGENTINE, M. D., and BARNES, R. M., 1990, Semi-quantitative analysis by inductively coupled plasma mass spectrometry. Journal of Analytical Atomic Spectrometry, 5, 457-462.
ASHBY, R., 1988, Migration from poly(ethylene terephthalate) under all conditions of use. Food Additives and Contaminants, 5, 485-492.
BIEBER. W.-D., FREYTAG. W., FIGGE, K., and VOM BRUCK. C. G., 1984, Transfer of additives from plastic materials into foodstuffs and into food simulants-a comparison. Food and Chemical Toxicology, 22,737-742.
EEC, 1985, European Community Council Directive No. 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 intended t0 come into contact with foodstuffs. Official Journal of the European Communities, L372/14, 31 December.
EEC, 1992, Draft resolution on control of aids to polymerisation (technical adjuvants) for plastic materials intended t0 come into contact with foodstuffs. Adopted by the Committee of Ministers at the 482nd meeting on 19/10/92.
EKIMOFF, D., VAN NORSTRAND, A. M., and MOWERS, D. A., 1989, Semiquantitative survey capabilities of ICP-MS. Applied Spectroscopy, 43, 1252-1257.
GILBERT, J., STARTIN, J. R., and MCGUINNESS, J. D., 1986, Compositional analysis of commercial PVC bottles and studies of aspects of specific and overall migration into foods and simulants. Food Additives and Contaminants, 3, 133-144.
JARVIS, K. E., GRAY, A. L., and HOUK, R. S., Sample preparation for ICP-MS. Handbook of Inductively Coupled Plasma Mass Spectroscopy (Glasgow and London: Blackie), 179 pp.
LONGERICH, H. P., 1989, Effect of nitric acid, acetic acid and ethanol on inductively coupled plasma mass spectrometric ion signais as a function of nebuliser gas flow, with implications on matrix suppression and enhancements. Journal of Analytical Atomic Spectrometry, 5, 346R-359R.
LORD, 111, C.1., 1991, Determination of trace metals in crude oil by ICP-MS with micro emulsion sample introduction. Analytical Chemistry, 63, 1594-1599.
MARSHALL, J., and FRANKS, 1., 1990, Applications of advanced spectrometry to the analysis of advanced materials. Analytical Proceedings, 27, 240-241.
MARSHALL, J., CARROLL, J., and CRIGHTON, J. S., 1990, Atomic spectrometry update-industrial analysis: metals, chemicals and advanced materials, 3) Advanced materials. Journal of Analytical Atomic Spectrometry, 4, 665-667.
TAN, S. H., and HORLICK, G., 1986, Background spectral features in inductively coupled plasma-mass spectroscopy. Applied Spectroscopy, 40, 445-449.
If you have come to this page from an outside location click here to get back to mindfully.org