Food Standards Agency UK - Benzophenone from Cartonboard (Number 06/00)

• Benzophenone was detected at 0.01 to 7.3 mg/kg (milligrams per kilogram) in 72 per cent (51/71) of food samples packaged in printed cartonboard. 143 samples of such packaging contained 0.05 to 3.3 mg benzophenone/dm² (milligrams per square decimetre).
• No health effects would be expected in an individual's lifetime from the levels found in this survey. Estimated intake of this substance was less than the toxicological Tolerable Daily Intake.
• The Food Standards Agency is taking action to ensure that consumers are protected by pressing for migration of benzophenone into food to be kept to a minimum.

The survey described here was conducted to compare dietary intake of benzophenone with the Tolerable Daily Intake (TDI). A TDI of 0.01 mg/kg bodyweight/day has been set by the European Commission Scientific Committee on Food.¹

Benzophenone is used as a photoinitiator in the curing of inks with ultra-violet (UV) light. UV-cure inks contain typically 5-10 per cent photoinitiator. Only a small portion of the initiator is used up during the curing process. Benzophenone can therefore remain in the printed material. The use of UV-cure inks for printing cartonboard has become widespread because the fast cure permits on-line cutting and folding, enabling rapid production of finished packaging. Research has shown that benzophenone can be present in food contact cartonboard and can migrate to foods packaged in this material even during frozen storage.²

Chemical migration from cartonboard into food is controlled in Great Britain under the general provisions of the Materials and Articles in Contact with Food Regulations 1987 and of the Food Safety Act 1990. Similar controls apply in Northern Ireland. Work continues at the Council of Europe to develop numerical standards of chemical migration from cartonboard.

Sampling
Three hundred and fifty samples of food packaged in printed cartonboard were purchased. The cartonboard from these samples was tested qualitatively for the presence of benzophenone. One hundred and seventy-five of these samples contained benzophenone. Duplicates of these were purchased. The samples, mainly of UK origin, were obtained locally. The packaging from one each of these duplicates was analysed immediately to determine the level of benzophenone in the cartonboard. Benzophenone was found in 143 samples (Table 1). Seventy-one samples were selected at random from the 143 samples that contained benzophenone, and the food contained in these 71 packages was analysed for this substance. (Some of the documents referred to on this Information Sheet can be found through a link in the see also box, top right of this page)

Food samples were stored in their packaging under conditions simulating retail storage. Frozen foods were stored in the freezer at - 20°C, refrigerated foods were stored at 4°C and shelf-stable foods were kept in a cupboard away from sunlight at room temperature, approximately 20^oC. The stored samples were kept until their best-by dates. During this period, they were over-wrapped in aluminium foil to limit any inter-sample migration. Once the best-by date was reached, or after one month from purchase (whichever came first), the food was removed from the packaging. The food and the packaging were then separately wrapped in aluminium foil and stored at - 20°C.

Analysis
A modified version of a published method was used to analyse cartonboard and food samples.² The modifications were that board samples were extracted with dichloromethane in place of chloroform because dichloromethane was more suitable as a solvent for GC-MS analysis of the samples. Size exclusion chromatographic clean-up of food extracts was omitted and a simple selective solvent-solvent extraction was used instead because this streamlined the analysis. Lastly, d₁₀-benzophenone was used as the internal standard in place of d₅-benzophenone, because the d10 form was commercially available whereas the d₅ form had previously been synthesised in-house.² The limit of detection (LoD) for cartonboard analysis was 0.0002 to 0.001 mg/dm2. The limit of quantification (LoQ) was 0.05 mg/dm². For food samples the LoD was 0.01 mg/kg, and the LoQ was 0.05 mg/kg.

Benzophenone was detected at very low levels in a blank card control sample for four of the six analytical batches. The levels were in the range 0.0002 to 0.002 mg/dm². In all batches the result for the blank control sample was more than a factor of 20-times lower than the LoQ. This very low background may have been carry-over of benzophenone on shredding the packaging. A quantity of blank cardboard was put through the shredder between samples, to minimise any carry-over, but given the design of paper document shredders and the large cutting surfaces involved a small level of carry-over would not be surprising. It was not feasible to wash the shredder with solvent between every sample because the materials used in construction are not resistant to the organic solvents in which benzophenone is soluble. This background level was insignificant compared to the levels of benzophenone found in cartonboard samples.

Analytical quality assurance
For benzophenone to be considered present in a sample of packaging or food, all of the following criteria had to be met for the analytical batch:

• not more than 0.02 minutes difference between the retention time for m/z 105 and 182 (benzophenone) and between m/z 110 and 192 (internal standard);
• not more than 0.1 minutes difference in the retention time for m/z 182 between standards and samples;
• not more than 0.1 minutes difference in the retention time for m/z 192 between standards and samples; and
• ion ratios m/z 182/105 and 192/110 for samples to be within plus or minus 20 per cent of the mean of these ratios for standards.

Packaging samples were analysed in batches. Each batch of cartonboard samples contained at least one sample of blank card, one solvent blank and two 'spiked' samples. If the recovery of the 'spike' did not fall in the range 70-110 per cent then the results for that batch were rejected. It was not necessary to reject the results for any batch. The analytical recovery for packaging samples 'spiked' at the LoQ of 0.05 mg/dm² was 95 per cent. The between-batch relative standard deviation (rsd) was 6 per cent for the 12 'spiked' samples analysed across the 6 analytical batches. Results were corrected for the average recovery of the 'spiked' samples in that batch.

Each batch of food samples contained two food blanks (oil and cornflakes), a solvent blank, and a 'spiked' sample of corn oil used as the in-house reference material. If the recovery of the 'spike' did not fall in the range 70-110 per cent then the results for that batch were rejected. Similarly, if either or both of the two batch blanks was found to contain benzophenone then the batch results were rejected. As before it was not necessary to reject the results of any batch. Method repeatability was good with an rsd of 12 per cent for five replicates of a pizza sample containing approximately 0.2 mg/kg benzophenone. The analytical recovery of benzophenone, 'spiked' into corn oil at 1 mg/kg, was good with an average of 99 per cent.

To check that there was no significant loss of label from the deuterated d10-benzophenone internal standard during sample extraction and work-up, which might bias any subsequent results, a determining experiment was performed. The d₁₀, d₉, d₈, d₇, d₆, d₅, d₄, d₃, and d₂ channels from the internal standard were monitored for three sample extracts (selected at random) and the results were compared with the ion abundance for the calibration standards. The deuterated internal standard as purchased contained a small amount of d9-labelled species, at 17 per cent, but there was no increase in this and there was little or no (less than 0.3 per cent) formation of the lower-labelled members of d6 and below. This is evidence that, under the conditions of use, the deuterium label of the internal standard was not lost.

In a 'blind' check sample exercise that was externally refereed, six samples of a breakfast cereal were 'spiked' with benzophenone by an independent laboratory and sent to the surveying laboratory for analysis. The samples were analysed and the results were reported to the independent laboratory. Two blank samples that had no benzophenone added were correctly identified. The results reported for the remaining four 'spiked' samples were satisfactory with results of 104 per cent, 101 per cent, 103 per cent and 100 per cent of the values expected, for the check samples 'spiked' at 1.16 and at 3.28 mg/kg.

In addition to analysing 'spiked' corn oil in-house reference material to check for consistency of analytical results between batches, a set of three foods was analysed on two occasions as a further guard against long-term 'drift' of analytical results. The first analysis was made in February 1999, when the food samples were recently purchased and unwrapped for analysis. The second analysis was made 14 months later in April 2000 at which time the food samples were removed from frozen storage and a second portion taken for extraction and analysis. The agreement between results was very good and was as follows: (sample ID, first analysis in mg/kg, second analysis in mg/kg): no. 15, 1.8, 2.0; no. 17, 5.7, 5.4; no. 21, 0.84, 0.88. This demonstrates not only that the analysis was repeatable in the long-term but also that the storage of the food samples, individually wrapped in aluminium foil and stored at -18^oC, was suitable with no rise or fall in the level of benzophenone in the food.

As an additional check for any cross-contamination of foods during extraction and analysis, the analytical batch that contained the long-term stability samples described above also contained 16 food products that had been found to be packaged in cartonboard not employing benzophenone as a UV-photoinitiator. The 16 packaging samples each contained less than 0.005 mg/dm2 of benzophenone which is more than 10-times lower than the LoQ for cartonboard analysis. None of the 16 food samples contained benzophenone above the LoD for food analysis of 0.01 mg/kg. This is further evidence that there was no cross-contamination.

Reporting
Brand names are reported as this survey was carried out in accordance with guidelines for reporting survey results published in the Food Safety Information Bulletin in September 1997. The absence of a particular brand from Table 1 means only that the brand was not included in the survey.

Benzophenone was detected in 51 out of 71 food samples (Table 2). Twenty-nine samples contained benzophenone at 0.05 to 0.5 mg/kg, seventeen samples were in the range of 0.5 to 5 mg/kg and three exceeded 5 mg/kg. The highest level was 7.3 mg/kg. Two samples contained levels that were detectable but not quantifiable. Benzophenone was found in 143 cartonboard samples at 0.05 to 3.3 mg/dm2 board (Table 2).

The results confirm that benzophenone can migrate from printed cartonboard to food. However no health effects would be expected in an individual's lifetime from the levels found in this survey. Estimated intake of this substance was less than the toxicological Tolerable Daily Intake of 0.01 mg/kg bodyweight/day set by the European Commission Scientific Committee on Food (SCF).¹ Intake was estimated at 0.001 mg/kg bodyweight/day. This estimate was made using the data in Table 2, a 97.5 percentile value for daily consumption for the foods that were tested³ (102 g), a mean value of bodyweight³ (60 kg) and a mean contaminant level. Upper and lower bound values were calculated for the mean contaminant level to take account of results that were less than the limits of quantification or detection, but the estimates of intake using these were the same.

The Food Standards Agency is taking action to ensure that consumers are protected. Industry has been directed to keep migration of benzophenone from cartonboard into food to a minimum. This is because the Tolerable Daily Intake set by the SCF was based on toxicological work that was reported some years ago, in 1970, and research on this substance continues (e.g. it is included in the US National Toxicology Program). The Food Standards Agency is contributing to such research. For example a current research project is looking into chemical migration into frozen foods. Research work on functional barriers was reported in April.⁴ Both of these aspects of chemical migration may have led to the observation apparent from this survey that there was not a clear correlation between levels of benzophenone in cartonboard and food. Differences in composition of the foods tested may have been important in this respect. There may also have been some contribution to benzophenone in some of the food samples from its use as a flavouring. However this substance was not detected in some of the food samples but it was found in all of the packaging samples (Table 2).

1. Commission of The European Communities, Directorate General III Synoptic Document: provisional lists of monomers and additives notified to the European Commission as substances which may be used in the manufacture of plastics intended to come into contact with foodstuffs.Updated to 10 January 2000.
2. Johns, S. M., Gramshaw, J. W., Castle, L. and Jickells, S. M. Studies on functional barriers to migration. 1. Transfer of benzophenone from printed paperboard to microwaved food. Deutsche Lebensmittel-Rundschau, 1995, 91, 69-73.
3. Gregory, J., Foster, K., Tyler, H. and Wiseman, M. The dietary and nutritional survey of British adults. London: HMSO, 1990.
4. Food Standards Agency. Food Safety Information Bulletin 119, 2000, p. 17.

Mr. Patrice Mongelard
Food Standards Agency
Chemical Safety and Toxicology Division

Email: patrice.mongelard@foodstandards.gsi.gov.uk

A copy of the full report of this survey has been placed in the Food Standards Agency library. If you wish to consult a copy please contact the library for an appointment giving at least 24 hours notice or alternatively copies can be obtained from the library; a charge will be made to cover photocopying and postage.