Additive Migration from Plastics into Food; A Guide for Analytical Chemists

Additive Migration from Plastics into Foods A Guide for Analytical Chemists

T. R. Crompton

Smithers Rapra Technology Limited

Additive Migration from Plastics into Foods A Guide for Analytical Chemists

T.R. Crompton

Smithers Rapra Technology Limited A wholly owned subsidiary of The Smithers Group Shawbury, Shrewsbury, Shropshire, SY4 4NR, United Kingdom Telephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118 http://www.rapra.net

First Published in 2007 by

Smithers Rapra Technology Limited Shawbury, Shrewsbury, Shropshire, SY4 4NR, UK

©2007, Smithers Rapra Technology Limited

All rights reserved. Except as permitted under current legislation no part of this publication may be photocopied, reproduced or distributed in any form or by any means or stored in a database or retrieval system, without the prior permission from the copyright holder. A catalogue record for this book is available from the British Library. Every effort has been made to contact copyright holders of any material reproduced within the text and the authors and publishers apologise if any have been overlooked.

ISBN: 978-1-84735-055-8 Typeset by Smithers Rapra Technology Limited Cover printed by Livesey Limited, Shrewsbury, UK Printed and bound by Smithers Rapra Technology Limited

Contents

Preface....................................................................................................................1 1

Additive Migration from Plastics into Packaged Commodities.........................5 1.1 Introduction...........................................................................................5 1.2 Principles of Extractability Testing..........................................................5 1.3 Extractability Testing in the European Union.........................................7 1.4 Extractability Testing in the UK..............................................................8 1.5 Extractability Testing in the USA............................................................8

2

Types of Polymers Used in Commodity Packaging.........................................13 2.1 Polyolefins and Copolymers..................................................................13 2.1.1 Polyethylene.............................................................................13 2.1.2 Polyethylene Naphthalate........................................................15 2.1.3 Polyethylene co-methacrylic Acid.............................................15 2.1.4 Polypropylene..........................................................................15 2.2 Polymethyl Pentene...............................................................................16 2.3 Ethylene-vinyl Acetate Copolymers......................................................16 2.4 Ionomers..............................................................................................16 2.5 Vinyl Plastics and Vinyl Copolymers....................................................17 2.5.1 Polyvinyl Acetate.....................................................................18 2.6 Polystyrene and its Copolymers............................................................18 2.7 Polyacrylonitrile...................................................................................19 2.8 Acrylic-based Multi Polymer Compounds............................................19 i

Additive Migration from Plastics into Foods 2.9 Lopac...................................................................................................19 2.10 Borex....................................................................................................20 2.11 Fluorocarbon Polymers.........................................................................20 2.12 Polyamides...........................................................................................20 2.13 Acrylics ...............................................................................................20 2.13.1 Polymethylmethacrylate...........................................................20 2.13.2 Polyethylene Terephthalate (Terylene)......................................20 2.14 Polycarbonates.....................................................................................21 2.15 Polyacetals............................................................................................21 2.16 Polyphenylene Oxide (PPO)..................................................................22 2.17 Polysulfone...........................................................................................22 2.18 Thermosets...........................................................................................22 2.19 Phenol-formaldehyde............................................................................22 2.20 Urea-formaldehyde...............................................................................23 2.21 Melamine-formaldehyde.......................................................................23 2.22 Polyesters..............................................................................................23 2.23 Epoxy Resins........................................................................................23 2.24 Polyurethanes.......................................................................................24 2.25 Silicones................................................................................................24 2.26 Natural Polymers..................................................................................24 2.26.1 Cellulose..................................................................................24 2.26.2 Cellulose Derivatives................................................................25 2.27 p-Hydroxybenzoic Acid and 2-Hydroxy-6-naphthoic Acid...................26 2.28 Polymers Used in Gas Barrier Applications...........................................26 2.28.1 Rubber and Elastomers............................................................26 ii

Contents 3

Non-Polymeric Components of Plastics..........................................................27 3.1 Polymerisation Residues.......................................................................29 3.2 Residual and Unreacted Starting Materials...........................................29 3.3 Polymerisation Medium........................................................................30 3.4 Catalyst Decomposition Agents............................................................31 3.5 Other Impurities Introduced During Polymerisation.............................31 3.5.1 Chemicals Added During Polymerisation.................................31 3.6 Processing Aids.....................................................................................31 3.6.1 Antiblock Agents.....................................................................32 3.6.2 Antioxidants............................................................................32 3.6.3 Antisplit Agents.......................................................................32 3.6.4 Antistatic Agents......................................................................32 3.6.5 Heat Stabilisers........................................................................33 3.6.6 Lubricants................................................................................33 3.6.7 Melt Strength Improvers..........................................................34 3.6.8 Mould Release Agents..............................................................34 3.6.9 Plasticisers................................................................................34 3.6.10 Slip Additives...........................................................................35 3.6.11 Other Stabilisers.......................................................................35 3.7 End-Use Additives................................................................................35 3.7.1 Antiblock Additives.................................................................36 3.7.2 Antifungal agents.....................................................................36 3.7.3 Antioxidants............................................................................36 3.7.4 Antistatic Agents......................................................................36 3.7.5 Bactericidal Agents...................................................................36 3.7.6 Brighteners and Whiteners.......................................................36 3.7.7 Colorants.................................................................................37 3.7.8 Expanding Agents (Blowing Agents, Foaming Agents).............37 3.7.9 Impact Improvers.....................................................................38 3.7.10 Lubricants................................................................................38 iii

Additive Migration from Plastics into Foods 3.7.11 Plasticisers................................................................................38 3.7.12 Ultraviolet Protective Agents....................................................38 3.7.13 Ultraviolet Screens...................................................................39 3.7.14 Ultraviolet Degradation Inhibitors...........................................39 3.7.15 Gas Barrier Packaging Oxygen Scavengers...............................39 4

Determination of Antioxidants.......................................................................43 4.1 Santonox R (4,4′-thiobis-6-tert butyl m-cresol) in Aqueous and Non-aqueous Extractants ....................................................................43 4.1.1 Method....................................................................................43 4.2 DLTDP in Aqueous and Non-aqueous Extractants...............................45 (a) Thin-Layer Chromatography.........................................................46

Method..........................................................................................47



Procedure.......................................................................................49

(b) Infrared Spectroscopy....................................................................51

Method..........................................................................................52



Procedure.......................................................................................55

(c) Oxygen Flash Combustion.............................................................58 Comparison of Infrared and Oxygen Flask Results...............................60 4.3 Irganox 1076 (Octadecyl-3-(3,5-di-t-butyl-4-hydroxy phenyl propionate) in Non-aqueous Extractants..............................................64 4.4 Ionox 330 (1,3,5-trimethyl-2,4,6-tris(3,5 di tert-butyl-4 hydroxybenzyl)benzene) Antioxidant in Fatty Extractants....................65 4.5 Miscellaneous Antioxidants..................................................................68 5

Determination of Ultraviolet Stabilisers in Extractants...................................75 5.1 Uvitex OB (2,5-bis (5′-tert butyl-2-benzoxazolyl)thiophene) in Aqueous Extractants.........................................................................75

6

Determination of Plasticisers in Extractants...................................................81 6.1 Phthalates in Oily and Aqueous Extractants.........................................81 6.1.1 Phthalate Platicisers in Fatty Extractants..................................81 6.1.2 Phthalates in Aqueous Alcoholic Extract..................................91

iv

Contents 6.2 Phthalates and Adipates in Aqueous and Fatty Extractants..................94 7

Determination of Organotin Thermal Stabilisers in Extractants...................105 7.1 Organotin Compounds in Fatty Extractants.......................................105

8

Determination of Organic Sulfur Compounds in Extractants.......................109

9

Determination of Polydimethyl Siloxanes in Extractants..............................119 9.1 Wine and Olive Oil.............................................................................119 9.1.1 PDMS in Olive Oil and Wine.................................................120

10 Determination of Lubricants in Extraction Liquids......................................125 11 Determination of Monomers and Oligomers in Extractants.........................129 11.1 Determination of Acrylonitrile in Aqueous and Fatty (Liquid Paraffin) Extractants..............................................................129 11.1.1 Determination of Acrylonitrile in Aqueous-ethanol Extractants.............................................................................129 11.1.2 Determination of Acrylonitrile Monomer in Acidic and Alkaline Extractants........................................................130 11.1.3 Determination of Acrylonitrile in 6% Hydrochloric Acid Extractant......................................................................131 11.1.4 Determination of Acrylonitrile Monomer in 5% Sodium Carbonate Extractant................................................131 11.1.5 Determination of Acrylonitrile in Light Liquid Paraffin Extractant.................................................................133 11.1.6 Separation of Acrylonitrile from Light Liquid-Paraffin and n-Heptane Extractants....................................................133 11.1.7 Determination of Acrylonitrile in Aqueous, Acidic and Alkaline Extractants........................................................135 11.1.8 Determination of an Acrylonitrile in Aqueous Ethanol Extractant.................................................................136 11.1.9 Calculation of the Acrylonitrile Contents of the Extractants.......................................................................136 11.2 Determination of Styrene Monomer and Aromatic Hydrocarbons in Aqueous and Fatty (Liquid Paraffin) Extractants............................138 v

Additive Migration from Plastics into Foods 11.2.1 Determination of Styrene Monomer and Other Volatiles in Polystyrene.........................................................................138 11.3 Determination of Acrylate Monomers and Oligomers in Aqueous and Fatty Extractants...........................................................141 12 Analysis of Polymer Extraction Liquids Containing More Than One Migrant.................................................................................................... 153 12.1 Preliminary Solvent Extraction of Gross Additives from Aqueous and Alcoholic Extractants....................................................154 12.1.1 Extraction Test (Assuming the Extraction Test is Carried Out on 700/800 ml Scale).........................................154 12.1.2 Ether Extraction.....................................................................155 12.2 Separation of Individual Additives from Solvent Extract of Extractant Liquid...............................................................................155 (a) Mixture of Santonox R, Ionol CP and DLTDP............................155 12.2.1 Indirect Setermination of Additives on the Thin-layer Plate......157 (b) Mixture of Butylated Hydroxy Toluene (BHT) Antioxidant and 2-Hydroxy-4-n-octoxybenzophenone Light Stabiliser in Extractants..............................................................................161 12.2.2 Scaling of Operations.............................................................162 12.3 Ether Extraction of Extraction Liquid..................................................163 13 Determination of Additives and their Breakdown Products in Extractants......171 13.1 Determination of Lauric Diethanolamide and its Degradation Products in Aqueous Fatty Extraction Liquids....................................172 13.1.1 Determination of Lauric Diethanolamide and DEA in the Distilled Water Extractant............................................175 13.1.2 Determination of Lauric Diethanolamide and DEA in the Aqueous 5% Sodium Carbonate Extractant.................176 13.1.3 Determination of Lauric Diethanolamide and DEA in the 50% w/v Aqueous Ethyl Alcohol Extractant................176 13.1.4 Determination of Lauric Diethanolamide and Hydrolysis Products in the 5% w/v Aqueous Citric Acid Extractant......................................................................180 vi

Contents 13.1.5 Determination of Lauric Diethanolamide in the Liquid Paraffin Extractant.................................................................181 13.2 Polymeric Plasticisers..........................................................................195 13.3 Polybutylene Adipate, Poly(1,2-Propylene Adipate) and Polybutylene Succinate.......................................................................196 13.4 Organosulfur Vulcanising Agents.......................................................198 14 Additive Migration Theory...........................................................................201 14.1 Polyethylene Naphthalate ..................................................................218 14.2 The Total Migration Concept.............................................................248 15 Gas Barrier Properties of Food Packaging Plastic Films................................259 16 Legislative Aspects of the Use of Additives in Packaging Plastics..................273 16.1 Regulatory System in the European Union..........................................273 16.2 The Regulatory System Existing in the UK.........................................280 16.3 Regulatory System Existing in the USA...............................................280 FDA compliance for Ultraviolet Electron Beam (EB) Coatings and Adhesives.....................................................................................290 17 Direct Determination of Migrants from Polymers into Foodstuffs................297 17.1 Antioxidants.......................................................................................297 17.2 Plasticisers..........................................................................................301 17.2.1 Olive Oil................................................................................301 17.2.2 Vegetables..............................................................................301 17.2.3 Meat......................................................................................301 17.2.4 Miscellaneous Foods..............................................................302 17.2.5 Hydrocarbons........................................................................302 17.3 Organotin Compounds.......................................................................302 17.4 Polyvinylpyrrolidone..........................................................................303 17.5 Benzoic Acid.......................................................................................303 vii

Additive Migration from Plastics into Foods 17.6 Alkyl Citrates.....................................................................................303 17.7 Metals................................................................................................303 Abbreviations and Acronyms..............................................................................309

viii

Preface

Plastics are now being used on a large scale for the packaging of fatty and aqueous foodstuffs and beverages, both alcoholic and non-alcoholic. This is evident for all to see on the supermarket shelves, namely margarine packed in polystyrene tubs, wine and beer in PVC bottles and meats and bacon in shrink-wrap film. As well as at the point of sale, foods are increasingly being shipped in bulk in plastic containers. Additionally, there is the area of use of plastics utensils, containers and processing equipment in the home and during bulk preparation of food in producing factories, at home and in restaurants and canteens. Contact between plastics packaged commodities also occurs in the products of the pharmaceutical and cosmetics industries and similar considerations apply to these where direct contact occurs between the packed commodity and the container, this is likely that some transfer will occur of polymer additives, adventitious impurities such as monomers, oligomers, catalyst remnants and residual polymerisation solvents and of low molecular weight polymer fractions from the plastic into the packaged material with the consequent risk of a toxic hazard to the consumer. The actual hazard arising to the consumer from any extractable material is a function of two properties, namely, the intrinsic toxicity of the extracted material as evaluated in animal feeding trials (not dealt with in this book) and the amount of extracted from the polymer which enters the packed commodity under service conditions, i.e., during packaging operations and during the shelf life of the packaged commodity at the time of the consumption. The principles covering the design of extraction tests proposed by various authorities are reviewed in Chapter 1. Chapters 2 and 3 review the types of polymers and elastomers, respectively, used in the packaging and the types of deliberately added or adventitious compounds that can occur in plastics and can be extracted into the packaged commodity. The analysis of aqueous or fatty foods, beverages, food simulent liquids, pharmaceuticals and cosmetics which have been contacted with plastics either in extraction tests or during the shelf life of a packaged commodity presents many fascinating and all too difficult analytical problems. Thus, the substance to be determined usually occurs at 1

Additive Migration from Plastics into Foods extremely low concentrations and in complex matrix and several extracted substances may be present in the extraction liquid with mutually interfering effects on the analysis. For example, the extract of a polystyrene may contain an antioxidant, an ultraviolet stabiliser, antistatic agent, mineral oil and plasticiser. Although the analyst may not be required to analyse for all of these substances it would be necessary to be aware of any interference effects that these substances may have on the determination of any particular component. Chapter 4 to 11 discuss methods available for the determination polymer extractants of antioxidants, UV stabilisers, plasticisers, organotin heat stabilisers, organosulfur vulcanising agents, siloxanes and monomers and oligomers, respectively. In order to give an idea of the care with which such methods have to be developed several detailed examples are given of previously unpublished methods developed in the Author’s laboratories for the determination of particular types of extractables in extraction liquids, e.g., dilaurylthiodipropionate, Nonox CI, antioxidants and Ethylon (lauric diethanolamide). The analysis of extraction liquids containing several extracted compounds is reviewed in Chapter 12. Next, there is the question of the analysis of additive breakdown either during polymer manufacture or upon contact with the packaged commodity or simulent liquid. Here there are two considerations, possible interference effects of breakdown products on the determination of the polymer migrant and the necessity to identify such breakdown products, as these too, must be considered from the toxicity point of view. These aspects are discussed in Chapter 13. Chapter 14 reviews modern migration theory as it applies to polymer extraction tests. Currently theory is being developed which given certain basic parameters of the polymer-extractant system would enable extraction data to be predicted theoretically thereby obviating the need for lengthy extraction tests. Eventually it is hoped that such an approach in certain cases might be accepted by the governing authorities. Much more work is needed in this field. The introduction of tailored polymer-based structures as packaging materials for foodstuffs has been increasing over the last decades. The main commercial appeal of the materials lies in their ability to offer a broad variety of tailor-made properties and yet to be cheap and easily processed. A large number of technologies have been put into place, i.e., multilayer structure, modified and equilibrium modified atmosphere packaging, active packaging and so on. The development of resins with high permeability properties for gases e.g., oxygen scavenging, and water and organic vapour is reviewed in Chapter 15. 2

Preface Over the past decade, there have been considerable changes in European and Food and Drug Administration legislation regarding migration of polymer components into packaging commodities. This is discussed in detail in Chapter 16. Finally, for completion, Chapter 17 reviews methodology for the determination of migrants in actual foodstuffs. This book will be of interest to those engaged in the implementation of packaging legislation, including management, analytical chemists and the manufacturers of foods, beverages, pharmaceuticals and cosmetics and also scientific and toxicologists in the packaging industry.

Roy Crompton June 2007

3

Additive Migration from Plastics into Foods

4

1

Additive Migration from Plastics into Packaged Commodities

1.1 Introduction Plastics are now being used on a large scale for the packaging of fatty and aqueous foodstuffs and beverages, both alcoholic and non-alcoholic. This is evident for all to see on the supermarket shelf, namely margarine packed in polystyrene tubs, wine and beer in polyvinylchloride (PVC) bottles, and meats and bacon in shrink wrap film. As well as at the point of sale, foods are increasingly being shipped in bulk in plastic containers. Additionally, there is the area of use of plastics utensils, containers and processing equipment in the home and during bulk preparation of food in processing factories, at home and in restaurants and canteens. Contact between plastics and package commodities also occurs in the products of the pharmaceutical and cosmetics industries and similar considerations apply to these. Where direct contact occurs between the packaged commodity and the container, it is likely that transfer will occur of polymer additives, adventitious impurities such as monomers, oligomers, catalyst remnants and residual polymerisation solvents and of low molecular weight polymer fractions from the plastic into the package material with the consequent risk of a toxic hazard to the consumer. The actual hazard arising to the consumer from any extractable material is a function of two properties, namely the intrinsic toxicity of the extracted material as evaluated in animal feeding trials (not dealt with in this book) and the amount of extracted from the polymer which enters the packed commodity under service conditions, i.e., during packaging operations and during the shelf life of the packaged commodity. The principals governing the design of extraction tests proposed by various authorities are now reviewed.

1.2 Principles of Extractability Testing The extractability of an additive or adventitious substance from a plastic can be determined by contacting the plastic for a specified time and temperature under standard test conditions with either the packed commodity or with a range of oily, 5

Additive Migration from Plastics into Foods alcoholic and aqueous extracts which simulate various types of packed commodities. At the end of the extraction test the extraction liquid or packaged commodity is analysed for extracted substances of interest by approved analytical techniques. However, low molecular weight additives frequently possess a high mobility in plastics materials and, in contrast to macromolecules, are capable of migrating from the packaging material into the packed product. The use of such substances in food packaging is, therefore, subject to strong legal controls. In order to decide whether a plastics packaging material complies with the requirements of the food law, two sets of questions should be considered. Concerning the plastics materials, one must ask whether the type to be used in contact with food is approved for packaging foodstuffs and whether it contains only approved additives in the allowed concentrations. The system, plastics plus foodstuffs to be packed, must also be considered, particularly the extent to which the individual plastics additives or their secondary products and plastics monomers migrate from the packaging material into the food and the extent to which low molecular polymer components similarly migrate. The questions concerning the plastics can in most cases be answered by the manufacturer of the materials. For selecting some packaging materials, it is more important to answer the questions regarding the system - plastics plus foodstuff to be packed. The toxicity of plastics packages, particularly those kept in contact with food for a prolonged period or heated during the pasteurisation, sterilisation or preparation of the foodstuff is first of all determined by the extent to which the additives migrate into the packed foodstuff. It would be ideal if the migration of each additive into the packed material could be determined when the package has been filled and stored under normal conditions of use. This would ensure that no physiologically objectionable plastics material would be admitted and, on the other hand, that no suitable plastics material would be rejected because of a hypercritical assessment. However, quantitative determination of the migrated additives in the heterogeneous foodstuff is extremely difficult. Therefore, natural migration must be simulated in model tests to determine the migrated or extracted additives in food simulants which can more easily be analysed. In this connection, the term ‘migration’ covers the transition of additives under storage conditions, (e.g., at and below 20 °C and 65% relative humidity) from packaging materials into packed foodstuffs or their simulants, while ‘extraction’ is the elimination of additives from a packaging material under extreme experimental conditions, (e.g., at 65 °C or at boiling heat) frequently with low boiling point liquids. Both of the classes of substances mentioned previously, (i.e., additives and impurities) must be considered in polymer extractability investigations and the higher the 6

Additive Migration from Plastics into Packaged Commodities concentration of these substances present in the polymer, then the more of them is likely to extract from the polymer in an extraction test. Obviously, it is advantageous from the point of view of the polymer manufacturer to aim at producing grades of materials intended for food packaging applications with the lowest possible content of contaminants such as unreacted monomer. Efficient control of the manufacturing process will often forestall subsequent difficulties in obtaining acceptance of polymers in food and beverage packaging applications. For this reason it is also desirable to have available, methods for determining the concentration of these impurities in polymers so that the amount of impurity left in the final polymer can be controlled to a suitably low level which is known to produce an acceptable material from the point of view of extractability. The specified conditions of the plastic extraction test and food simulating extraction solvents prescribed in the test by various authorities differ considerably. As an example of the type of test procedure recommended, those of the European Union (EU) and those of the Food and Drug Administration (FDA) of the USA are discussed next in some detail as examples of procedures which receive recognition in countries other than the country of origin.

1.3 Extractability Testing in the European Union The food simulating extraction liquids recommended by the EU include a range of aqueous and fatty extractants. EU Directive 2002/72 EEC [1] and its amendments 2004/19 EC which will be superceded after 2006 by a super directive to cover all aspects of migration testing, for further details, see Chapter 1.6. Aqueous extraction liquids discussed in EU Directives include 10% ethanol, 15% ethanol, 3% aqueous acetic acid, 5% aqueous citric acid and 5% aqueous sodium carbonate. Oily extractants include olive oil triglycerides (tributyrin, tricaprylin), sunflower oil or a synthetic mixture of triglycerides known as HB307. When use of these extractants is not practical due to problems in chemical analysis, the EU decided in 1996 to accept the use of alternate volatile simulents such as isooctane or 95% ethanol. Other fatty simulent liquids that have been considered include isopropanol, ethyl acetate and cyclohexane [2]. Numerous variants exist in the detail of the extraction test, some of these are covered in Chapter 14 (additive migration). A good starting point when organising an extraction 7

Additive Migration from Plastics into Foods test is to refer to original EU Directives [1] (see Chapter 16 on the Legislative Aspects of the use of Additives in Food Grade Plastics). Extraction test durations can vary over a wide range from as low as 2 hours to 60 days. Similarly, the test temperatures used should reflect the conditions that the packaged commodity is subjected to during manufacture and storage life, for example, O’Brien and co-workers [3] in their study of the migration of 2-hydroxy 4-N-octyloxy benzophenone UV stabiliser and octadecyl-3-(3,5-di-tert-4-hydroxyphenyl)propionate antioxidant from high-density polyethylene, polypropylene, high-impact polystyrene and PVC into olive oil used migration test temperatures of 2 hours at 120 °C, 6 hours at 70 °C, 2 hours at 70 °C, 2 hours at 60 °C and 10 days at 40 °C to represent shortterm exposure at high temperatures and room temperature storage.

1.4 Extractability Testing in the UK In the UK, the EU Framework Directive 76/893/EEC, the EU Framework Directive 89/109/EEC and 2002/71/EEC have all been incorporated into UK Statutory Instruments 2006 No.1401, Plastic Materials and Articles in Contact with Food Regulations, 2006 [4]. To assist plastics manufacturers and users in the UK in the interpretation of Statutory Instrument 3145, the British Plastics Federation has published Plastics in Contact with Food – A Guide [5]. There is also the PIRA publication: Food Packaging Migration and Legislation [6].

1.5 Extractability Testing in the USA The US legislation has been adopted outright or in a modified form by other countries. In the US food law is strongly inclined to Statute Law. Although each state can legislate separately, in practice the Federal Law usually prevails which has considerable advantages in interstate commerce. Many states have their own law which is identical to the Federal Law. Details of the US migration test procedures have been published [7-12]. A typical example of the application of a migration test procedure has been described by Lin and co-workers [13]. Migration studies were performed by determining acrylate monomers and oligomers in extractants from electron beam curable acrylate coatings. The cell extraction method used was that proposed by the FDA Centre for Food Safety and Applied Nutrition (www.cfsan.fda.gov). Stainless steel extraction cells with Teflon spacers of 30 ml volume and 50 cm2 opening for the extraction solvent 8

Additive Migration from Plastics into Packaged Commodities

Figure 1.1 Cross section of an extraction cell. (Reproduced with permission from Lin and co-workers, in Proceedings of the Tappi PLACE Conference, Boston, MA, USA, 2002, Session 13, Paper No. 48 [13]. © 2002, Tappi Press)

were used for the extraction test. A test sample, approximately 6.5 cm × 12.5 cm was placed in the extraction cell with the side (direct or indirect) to be extracted facing the cavity where the extracting solvent would be placed (see Figure 1.1). The cell/sample was conditioned with the appropriate extraction solvent (10% or 45% ethanol), temperature schedule (20-100 °C) and dwell time (30 minutes to 10 days) as outlined in FDA Migration Testing Protocols [12]. The cell was allowed to cool to room temperature before handling. The principal United States law is the FDA Food, Drug and Cosmetic Act 1938 and its numerous amendments [14]. This is designed to provide food that is safe and wholesome to the people, honestly labelled and properly packaged. Decisions have to be made as to the appropriate extraction liquids, test temperatures and whether the extraction test is carried out by single or double sided exposure of the test panels to the extractant liquid. Nerin and co-workers [15] in their study of migration of various phthalate plasticisers from paper board into FDA extraction liquids used various test conditions including 10 days at 40 °C, 1 hour at reflux temperature for all aqueous stimulants, 10 days at 40 °C and 1 hour at 175 °C for olive oil and 2 days at 20 °C and 3 hours at 60 °C for the isooctane stimulant. The decision between single or double sided extraction is based upon the construction of the test material, and its thickness. Structures greater than 0.05 cm thick are considered by the FDA to be infinitely thick. Single-sided extractions are most often used for coated materials or structures less than 0.05 cm thick. Two pieces of materials are separated by an inert spacer, thus defining a volume. This layered construction is secured so that the volume can be filled with the food simulating solvent. Double-sided extractions are by far the most common type, i.e., when specimen thickness exceeds 0.05 cm. FDA requirements stipulate that the ratio of food simulating solvent volume to surface area of extracted material must be near to 0.3 ml/cm2. 9

Additive Migration from Plastics into Foods Various extraction liquids are recommended by the FDA including distilled water, 3% aqueous acetic acid, 3% aqueous sodium bicarbonate, 3% aqueous sodium chloride, aqueous ethyl alcohol of the appropriate concentration, 20% sucrose solution containing 1% citric acid adjusted to pH 3.5 (aqueous extractants) and a liquid food fat, e.g., olive oil, vegetable oil, heptane and diethyl ether (fatty extractants). In deciding on the extractant test conditions to be used, the FDA distinguishes between the following conditions of use of the plastic food wrapping or container: 1. High temperature, heat sterilised, 2. Boiling water sterilised, 3. Hot water filled or pasteurised above 65 °C, 4. Hot water filled or pasteurised below 65 °C, 5. Room temperature filled and stored, i.e., no thermal treatment of food in the container, 6. Refrigerated storage, no thermal treatment of food in the container, 7. Frozen storage, no thermal treatment of food in the container, and 8. Frozen or refrigerated storage, a) ready-prepared foods intended to be reheated in container at time of use, b) aqueous, high or low free oil or fat.

References 1.

Commission Directive 2002/72/EC of 6th August 2002, Relating to Plastics Materials and Articles Intended to Come into Contact with Foodstuffs, Official Journal of the European Communities, No.1, 220/18, 8-58, 2002.

2.

I.E. Helmroth, M. Dekker and T. Hankemeier, Food Additives and Contaminants, 2002, 19, 1, 176.

3.

A. O’Brien, A. Goodson and I. Cooper, Food Additives and Contaminants, 1999, 16, 9, 367.

4.

UK Statutory Instrument No.1401, Plastic Materials and Articles in Contact with Food Regulations, HMSO, London, UK, 1992.

5.

Plastics in Contact with Food: A Guide, 2nd Edition, British Plastics Federation, London, UK, 1996.

10

Additive Migration from Plastics into Packaged Commodities 6.

R. Ashby, Food Packaging Migration and Legislation, 2nd Edition, PIRA International, Leatherhead, Surry, UK, 1997.

7.

Food Additives: Food Contact Substance Notification System, US Food and Drug Administration, Federal Register, May 21st 2002, 67, 98 (Rules and Regulations), p.35724-35731.

8.

US Food and Drug Administration, Centre for Food Safety and 38 Applied Nutrition, Office of Premarket Approval, September 1999, Guidance for Industry, Preparation Premarket Notifications for Food Contact Substances Toxicology Recommendations, 2002.

9.

US Food and Drug Administration, Centre for Food Safety and Applied Nutrition, Office of Premarket Approval, May 2002, Guidance for Industry, Preparation of Premarket Notifications for Food Contact Substances Chemistry Recomendations, 1999.

10. US Food and Drug Administration, Centre for Food Safety and Applied Nutrition, Office of Food Additive Safety, August 2002, Inventory of Effective Premarket Notifications for Food Contact Substances, FCN No. 103, 2002. 11. US Food and Drug Administration, Centre for Food Safety and Applied Nutrition, Office of Food Additive Safety, August 2002, Inventory of Premarket Notification Limitations, Specifications and use for Food Contact Substances, FCN No103, 2002. 12. Section 11 D (C) Food Simulents, Appendix 11, Selected Migration Testing Protocols Guidance for Industry Preparation of Pre-Market Notification for Food Contact Substances Chemistry Recommendations, April 2002. 13. A Lin, H. Gao, G. Wind and T. Wornick in Proceedings of the Tappi PLACE Conference, Boston, MA, USA, 2002, Session 13, Paper No.48. 14. Federal Food, Drug and Cosmetics Act, Section 201, USC, 321. 15. C. Nerin, E. Arseno and C. Jiminez, Analytical Chemistry, 2002, 74, 22, 5831.

11

Additive Migration from Plastics into Foods

12

2

Types of Polymers Used in Commodity Packaging

A high proportion of the wide range of polymers manufactured nowadays are used in applications which involve contact with food or beverages particularly in foodstuff packaging applications, utensils, kitchenware and in processing equipment in food factories and other establishments where food is handled in large quantities. In addition to packaging applications, plastics are used extensively in the food and drink vending market and in the manufacture of pre-packed meals, and also in the packaging of pharmaceuticals and cosmetics. In this Chapter, the types of plastics that are being used in these applications are briefly reviewed. The variants of these polymers such as copolymers and blends are discussed together with a brief general discussion of the properties of the polymers which are most relevant from the point of view of their use in applications involving contact with food.

2.1 Polyolefins and Copolymers 2.1.1 Polyethylene There are two distinct forms of polyethylene, namely low-density (high pressure synthesis) and high-density (low pressure synthesis, e.g., the Ziegler route). Also available are a range of copolymers of intermediate density made by either blending or by the copolymerisation of ethylene with other olefins such as propene, butane or hexene. The density range for polymers produced by the low pressure route is about 0.945 to 0.965 g/cm3, whereas the high pressure process produces material with densities between 0.918 and 0.935 g/cm3. Crystallinities are also different for the two types of polyethylene. The high-density polyethylene (HDPE) with its linear structure favours parallel configuration of the chains and hence a high degree of crystallinity (75-90%). The low-density polyethylene (LDPE) has appreciable chain branching which disturbs the regularity of the arrangement of atoms and hence produces a low degree of crystallinity (55-70%). The greater linearity of HDPE also increases the softening point of this grade relative to that of LDPE. This in turn increases the flex resistance of the high-density polymer. Due to its 13

Additive Migration from Plastics into Foods high softening point, HDPE unlike the LDPE can be steam sterilised, a property which is of importance in the food packaging field. Both forms of polyethylene are fairly inert chemically and are not attacked by solvents at temperatures up to 60 °C. However, they do absorb certain hydrocarbon solvents with swelling at room temperature. The HDPE is more resistant in this respect. Both types of polymers are resistant to water and salt solutions and water absorption is negligible at room temperature. Water vapour permeability is fairly low in the HDPE, as would be expected due to its lower permeability. Gas permeabilities are not particularly low, and LDPE, in particular, is not to be recommended as an oxygen barrier, i.e., it would not be used for the packaging of types of foods where oxygen ingress is undesirable. Water and oxygen permeabilities have important implications in the consideration of these polymers as food packaging materials. Both HDPE and LDPE exhibit the phenomenon of environmental stress cracking. This can occur when the material is multiaxially stressed when in contact with certain polar liquids or vapours. These liquids need not be solvents for polyethylene or even be more than slightly absorbed by it, and if the polymer is unstrained, (i.e., no moulding strains) no cracking occurs, no matter how long the contact time with the liquids. Environmental stress cracking has implications in the food packaging field as certain foodstuffs, such as vegetable oils and alcoholic beverages can act as stress cracking agents. Various types of additives can occur in polyethylene used in food packaging. These include, pigments, carbon black (for ultraviolet light resistance), slip additives (e.g., silica) to lower coefficient of friction of film and improve the rate of movement of film through printing and wrapping equipment, antistatic additives (to reduce dust attraction caused by build up of static changes), antiblock additives (used in film grades used for the manufacture of bags in order to overcome difficulties in separating them from a pile) and also to facilitate rapid opening of the bags on automatic filling equipment, and antioxidants (to prevent degradation during processing and during service life of the article). The biggest single application for LDPE is in packaging of food and vegetables. Food contact applications include the household use of polyethylene bags for storing food, especially in the refrigerator, or freezer, polyethylene coated cartons for frozen foods, bags for pre-packed fresh produce and bags for frozen poultry and hams. Blow moulded polyethylene containers are used for the packaging of table salt and sauces. In the injection moulding field, however, LDPE and HDPE are used in many types of food containers, particularly as closures and dispensers. Another very large use is in pots, tubs, beakers and bowls used as food storage containers. There is also an increasing use of extruded LDPE pipe for domestic cold water supplies. 14

Types of Polymers Used in Commodity Packaging

2.1.2 Polyethylene Naphthalate This is a relatively new polymer that has good barrier properties and which has been used in the packaging of olive oil. It is alos ideal for making amber coloured bottles suitable for packaging drinks such as beer. It can also be used for the manaufacture of high-performance fibres that have very high modulus and better dimensional stability than polyester or polyamide.

2.1.3 Polyethylene co-methacrylic Acid This polymer has been used as an additive in long chain branched polyethylene used in food packaging.

2.1.4 Polypropylene Polypropylene manufactured by the low pressure route gives a polymer which is largely isotactic (pendant methyl groups all on same side of carbon backbone). The crystallinity of low pressure polypropylene at 65 to 70% is somewhat lower than that of HDPE. Polypropylene, unlike the polyethylenes, is not subject to environmental stress cracking, which gives it an advantage in the food packaging field. It also has a lower density (0.90 g/cm3) than either LDPE or HDPE. Although the impact strength of polypropylene is lower than that of HDPE, especially at temperatures below 0 °C, this can be improved by incorporating various synthetic rubbers into polypropylene or by copolymerisation with ethylene and propylene. Two important properties of polypropylene are its resiliance and its resistance to fatigue by flexing. This makes it a suitable material for moulding screw cap closures used extensively in food and beverage packaging. A thin section diaphragm or fin is moulded into the inner surface of the closure in such a position that it bears down on the upper surface of the bottle neck. A rigid material would not have enough ‘give’ to take up imperfections in the glass surface and so would not form as good a seal. On the other hand, a more flexible material, such as LDPE, would ‘give’ but would not press back strongly enough to form a seal. The good fatigue resistance of popypropylene is utilised in the manufacture of snap fit lid food containers. Injection moulded polypropylene is used extensively in the manufacture of food storage containers either thick walled or thin walled. Polypropylene, due to its higher softening point when compared to polyethylene, is useful in food packaging applications where 15

Additive Migration from Plastics into Foods hot filling temperatures are involved. Polypropylene film is used extensively in food packaging applications because of its low water vapour and oxygen permeabilities. Examples include: biscuits, potato crisps and snack foods. Animal feeding stuffs are packed in polypropylene woven sacks, as are grains, sugar and vegetables.

2.2 Polymethyl Pentene This is a polymer of 4-methyl-1-pentene and is better known as TPX (Mistsui Chemicals). It is a low specific gravity (0.83 g/cm3) polymer of high clarity and softening point. It is still rather expensive. It has a lower impact strength and a very much higher permeability to gas and water vapour than either polyethylene or polypropylene. It is highly resistant to aqueous salt solutions, acids, alkalies and organic solvents. It is subject to environmental stress cracking. One of its few applications in which it comes into contact with food is as a coating on trays used in bakery ovens. In general, its outstanding properties compared to the other polyolefins are resistance to hot fillings and outstanding clarity.

2.3 Ethylene-vinyl Acetate Copolymers These polymers are very similar to LDPE in many of their characteristics. They are more transparent than polyethylene and have a high flexibility and impact resistance and good resilience. These polymers have a high permeability to water vapour and gases and are rather more susceptible to oil/hydrocarbon attack than is LDPE. When made into film, ethylene-vinyl acetate has a greater tendency towards blocking than LDPE and consequently it is necessary to incorporate a rather high percentage of antiblocking additive. Applications are mainly in the fields where flexibility and resilience are useful, particularly at low temperatures. This makes these copolymers attractive, for example, for the stretch wrapping of deep frozen poultry in which application they require a close contour wrap of the bird.

2.4 Ionomers The only ionomer produced in commercial quantities is produced under the trade name Surlyn A by DuPont. Ionomers are in effect ionic polymers which are solid at normal temperatures but which soften as do thermoplastics, upon heating. Surlyn A is basically a polymer of ethylene containing carboxylate groups and which is similar in many ways to LDPE. The ionic forces, due to the carboxylate groups give it a high melt strength so 16

Types of Polymers Used in Commodity Packaging that it has excellent drawing characteristics. Surlyn A is useful as an extrusion material and very thin coatings with a low ‘neck-in’ can be obtained. Skin packaging is another application with obvious attractions in the food packaging industry. It is resistant to strong and weak alkalies and is slowly attacked by acids. It is resistant to alcohol but swells in contact with oils and hydrocarbons. It is however, more resistant to oils than LDPE and has in fact been used for the packaging of olive oil.

2.5 Vinyl Plastics and Vinyl Copolymers Polyvinylchloride (PVC) is much less crystalline than the polyolefins. The base polymer is very hard and for most applications it has to be plasticised to make it flexible enough for use. There are many plasticisers available for PVC. Food contact applications, of course, impose their own requirements of intrinsic safety and low migration rate of the plasticiser. Due to the nearness of its decomposition and processing temperatures, PVC has to be stabilised against heat by the addition of 1-2% of heat stabiliser and this, in turn, has food contact implications. PVC has a density of about 1.4 g/cm3 which makes it appreciably denser than any of the unchlorinated polyolefins. It is resistant to alcohol and to many fats and this encourages its applications in the packaging of wine, beer, and fatty foods. For food contact uses PVC usually has a low plasticiser content in which form it has good rigidity and can be moulded into sections and film down to 75 mm thick. It has the merit of clarity, which is useful in many food uses. PVC film can be thermoformed into various packages including tubs, trays for foodstuffs, inserts for chocolate boxes and biscuit tins and other confectioneries. Very thin film is used for shrink wrapping of prepackaged meat, fruit and vegetables. PVC bottles are increasingly being used for the packing of foodstuffs and alcoholic beverages, including wine and beer. It is used extensively in the UK for the packaging of fruit squashes and to some extent for edible cooking oils. PVC can be fabricated into jars suitable for the packaging of coffee and chocolate drinks. PVC is also copolymerised with other monomers such as vinyl acetate, vinylidene chloride propylene and acrylonitrile. Copolymerisation with vinyl acetate tends to soften the polymer to the point that plasticiser addition may be unnecessary. For low temperature applications, plasticiser addition may still be desirable and in this form the copolymer is used in the fabrication of refrigerator trays. Vinyl chloride - vinylidene chloride copolymers are used for the manufacture of films which have a very low permeability to water vapour and gases. Such film is used for 17

Additive Migration from Plastics into Foods wrapping cheese and other commodities which require the absence of oxygen for their preservation. Vinyl chloride - vinylidene chloride is also applied as a coating to improve the barrier properties of other food packaging materials such as paper, polypropylene and cellulose film. Propylene - vinyl chloride copolymers are used mainly as bottle blowing materials. Copolymers of vinyl chloride and acrylonitrile have no food packaging applications.

2.5.1 Polyvinyl Acetate This rubbery, synthetic, polymer is used in the food industry, mainly as an adhesive in laminating operations.

2.6 Polystyrene and its Copolymers This is a hard fairly brittle material. Chemically it is unaffected by acids, alkalies, lower alcohols and most paraffinic hydrocarbons. It is attacked by certain foodstuffs, e.g., orange peel oil. It is a reasonably good barrier to gas but a poor one to water vapour. The brittleness of polystyrene can be overcome by incorporation of a synthetic rubber, such as polybutadiene or styrene-butadiene rubber in amounts up to 10%. The increase in impact strength and flexibility thus acquired is accompanied by some loss in clarity, so that only opaque or translucent grades of toughened polystyrene are available. This does not however, limit its uses in food packaging. Polystyrene is also available in an expanded (cellular) form and as such is used extensively in the fabrication of vending machine drinking cups. Polystyrene is used extensively in food packaging and in the home. The toughened polymer is injection moulded into tubs and containers for yogurt, dairy cream, cut peel, ice cream, cottage cheese, honey and syrup. Transparent untoughened polystyrene is used for the fabrication of jam and marmalade containers. Thermoformed thin section toughened polystyrene is used for the fabrication of vending cups, and thicker sections are used to manufacture refrigerator cabinets and door liners where food contact is likely. Biaxially orientated untoughened polystyrene film is also used in the manufacture of transparent food containers. Moulded expanded polystyrene boxes are used extensively for the transport of fruit, vegetables and fish and thermoformed expanded sheet to produce supermarket trays for the prepackaging of meat, fruit and vegetables. 18

Types of Polymers Used in Commodity Packaging The physical properties of polystyrene are improved by copolymerisation with acrylonitrile or acrylonitrile and butadiene. Styrene - acrylonitrile (SAN) is tough and transparent and is used in the manufacture of measuring jugs, orange and lemon squeezers and food homogenisers. Acrylonitrile - butadiene - styrene (ABS) can be manufactured to have a range of properties, depending on the ratio of the monomers present and the order in which the monomers are polymerised. ABS has an improved impact resistance and toughness over polystyrene and also superior chemical resistance. It can be injection moulded, blow moulded and extruded. Applications in food packaging include injection moulded cake and bread trays and margarine tubs. Polystyrene-6-polydimethyl siloxane has been used as an additive to long chain branched polyethylene and narrow molecular weight polystyrene to control barrier properties.

2.7 Polyacrylonitrile This polymer has good chemical and oil and grease resistance. Its only application in which it might come into contact with food is as a filter cloth in food manufacturing equipment.

2.8 Acrylic-based Multi Polymer Compounds This compound is a copolymer of acrylic acid and two other monomers and known as XT Polymer (Cyro Industries). It has excellent oil and grease resistance and has, in fact, been used for the fabrication of containers for peanut butter and medicinal mineral oil.

2.9 Lopac Lopac (Monsanto) is the trade name for a copolymer of methacrylonitrile with small percentages of styrene and methyl styrene. It has been used as a material for the fabrication of carbonated soft drink bottles. It has excellent clarity and barrier properties. 19

Additive Migration from Plastics into Foods

2.10 Borex This is the trade name of a product made by copolymerising acrylonitrile and methyl acrylate in the presence of butadiene - acrylonitrile rubber. It is clear, has good barrier properties and impact strength. This polymer may also, have future applications in the bottle blowing field for carbonated soft drinks bottles.

2.11 Fluorocarbon Polymers Due to its chemical inertness polytetrafluoroethylene is used for pump and pipe lining and gaskets, seals and washers and as such, may be used in applications which involve contact with food. In addition, it is used as a non-stick coating in pans and frying pans and is used in the food industry for the coating of a wide range of mixing equipment. Due to its high price, this material would not be used as a food packaging material.

2.12 Polyamides Various polyamides (Nylon) are manufactured by the condensation of amino acids. Polyamides can be blown into film and here they find many applications in the packaging of oils, fats and greases. The high softening point of polyamides have lead to their extensive use in boil-in-the-bag food packs. The low gas permeability of the film has been utilised in sachets for vacuum packed foods, such as cheese slices and bacon. Moulded articles in polyamide are widely used in food manufacturing processes. The fact that polyamides are self lubricating and consequently do not require the addition of a lubricant additive is a particularly important factor in the food industry where contamination by lubricants cannot be tolerated.

2.13 Acrylics 2.13.1 Polymethylmethacrylate This is a very clear plastic with good light exposure properties. Chemically it is resistant to water, alkalies, many dilute acids and aqueous salt solutions. This polymer is too expensive for packaging. It does however have one application as a food quality plastic, namely the fabrication of artificial dentures.

2.13.2 Polyethylene Terephthalate (Terylene) Polyethylene terephthalate is a condensation product of ethylene glycol and terephthalic acid and one of its trdaenames is Terylene (ICI). It is fabricated as film 20

Types of Polymers Used in Commodity Packaging and fibre and can be injection moulded. Terylene film has excellent strength and transparency but tends to lose strength at the heat seals. Bags are, therefore, often made up using adhesives. It has a high softening point. As a moulded material its important properties are low moisture pick-up, high yield strength and high surface hardness. It has good creep and abrasion resistance. Terylene film is used mainly in the manufacture of food packaging. It is expensive but due to its high strength, it can be used as a very thin film which keeps costs down. When used as a thin film it is sometimes laminated with other lower cost film. Its high softening point renders it attractive in the boil-in-the-bag food market. It is used extensively in drink bottling applications.

2.14 Polycarbonates These polymers are, in effect, linear polyesters of carbonic acid made by reacting aromatic dehydroxy compounds such as bisphenol A with diphenylcarbonate or phosgene. Polycarbonates are hard, rigid and transparent with a very high impact and tensile strength. They have good low temperature properties and are resistant to acids but not alkalies. Polycarbonates are stain resistant and have been found to be satisfactory after long periods in contact with coffee, fruit juices and tea. An additional bonus with this material is its non-flammability. Because of its high strain resistance, high temperature resistance and impact strength, this polymer is used for the fabrication of household items such as plates, cups and saucers and baby’s feeding bottles. It has also been used in the fabrication of food processing equipment.

2.15 Polyacetals A polyacetal homopolymer is essentially a polymer of formaldehyde. Copolymers with cyclic ethers, such as ethylene oxide are also available commercially. These polymers have good abrasion resistance and stand up well to repeated impacts. The range of applicability extends from –40 °C to 80 °C. They have a low water absorption and are creep and fatigue resistant. Chemically they are very resistant to weak and strong alkalies and detergents but are attacked by strong acids. They are affected by ultraviolet light but can be protected with carbon black. The properties of polyacetal have lead to a particular specialised food contact application such as the meat hooks for handling meat carcasses. 21

Additive Migration from Plastics into Foods

2.16 Polyphenylene Oxide (PPO) This is a linear polymer made by the catalytic oxidative coupling of 2,6 dimethylphenol. PPO has a use temperature range of –55 to 175 °C. It is transparent but yellow in colour and is rather expensive being used mainly in engineering applications. Its resistance to repeated steam sterilisation makes it a useful plastic in food engineering applications. It is also used in the fabrication of domestic food mixers. A modified form of PPO is available with the trade name Noryl (General Electric Company). This is cheaper than PPO and has excellent mechanical properties over the range –40 to 120 °C. This polymer is used extensively in food contact applications where an impact strength slightly inferior to that of PPO is acceptable.

2.17 Polysulfone This is another engineering plastic used in food engineering. It is a tough rigid transparent plastic with very high tensile strength, creep resistance and low mould shrinkage and water absorption. It is resistant to acids, alkalies, aqueous salt solutions, alcohols, detergents and oils, even at elevated temperatures and under moderate stress. It has also been used in kitchen hardware and as a replacement for stainless steel in the manufacture of milking machines.

2.18 Thermosets Various thermosets are used in contact with food in significant commercial amounts. Thermosets are distinguished from thermoplastics in that they undergo an irreversible chemical change on heating during moulding. Thermosets are discussed later in Chapter 3.

2.19 Phenol-formaldehyde As manufactured, these polymers usually contain various types of fillers and this has obvious food contact implications. The fillers are used to reduce cost, improve shock resistance and to reduce mould shrinkage. Phenol-formaldehyde resins are resistant to common solvents, acids, aqueous salt solutions and hot oils, although water and alcohols cause slight swelling. As far as food contact applications are concerned the main use of these resins is in the manufacture of closures for containers such as jars and bottles. Direct contact is less likely in these applications because of the interposition of a liner between the closure and the container. The resins are also used for moulding the handles of kitchen implements. 22

Types of Polymers Used in Commodity Packaging

2.20 Urea-formaldehyde This resin produces mouldings which are resistant to common solvents but are attacked by strong acids. This resin is slightly less resistant to water than phenolformaldehyde. The impact strength of urea-formaldehyde resins depend on the type of filler used. Frequently, bleached wood pulp is used as a filler. The main interest to the food industry is again that of container closures and in the domestic field they are used as casings for domestic food mixers.

2.21 Melamine-formaldehyde Colourless white or pastel coloured mouldings can be produced from this resin using bleached wood pulp as a filler. Melamine-formaldehyde mouldings are more resistant to water than urea-formaldehyde resins and are not attacked by weak alkalies and they have a high heat resistance. Due to their superior water resistance, mouldings made in this resin have good staining resistance and have consequently been widely used in tableware. Their high heat resistance has lead to their extensive use as the top layer of decorative laminates for table tops and working surfaces.

2.22 Polyesters A whole range of these resins is available. They are produced by the reaction of polyhydric alcohols and polybasic acids. An important use of these resins is in the manufacture of glass fibre reinforced polyester laminates which, in addition to their great strength, have a high heat and corrosion resistance. Thermosetting polyesters are resistant to most solvents and acids and alkalies. Glass reinforced plastics are widely used in the manufacture of semi-bulk containers for all types of liquid and solids and for storage tanks.

2.23 Epoxy Resins This range of resins, made by the condensation of bisphenol A and epichlorohydrin (or another compound containing the epoxy group) have many similarities with the polyesters. These are extremely tough materials with good heat and chemical resistance. They have low impact strength and when used as coatings are usually modified with flexibilising agents such as low molecular weight polyamides. 23

Additive Migration from Plastics into Foods Epoxy resins, such as polyesters, are used in the manufacture of laminates and reinforced structures. They are used as lacquers in a wide variety of food cans.

2.24 Polyurethanes These resins are formed by the reaction of isocyanates such as toluene diisocyanate and methylene diisocyanate with polyols. This resin can be produced in a rigid form or as a rigid or flexible foam. Applications of polyurethane foam in food packaging are likely to be confined to cushioning, for example, glass jars in outer containers, since polyurethanes must not be used in contact with food because of the great difficulty in removing the last traces of the diisocyanate which is highly toxic.

2.25 Silicones There are three main forms of silicones, namely liquids, solids (resins) and rubbers. The resins are the most important from the food contact point of view. A major application is in the bakery industry where the resin is used to coat bread baking tins.

2.26 Natural Polymers Various natural materials such as cellulose and cellulose acetate are used in applications which involve contact with food. These are discussed in Section 2.26.1.

2.26.1 Cellulose Cellulose film is manufactured by a rather complicated process involving the casting of the film and impregnation with plasticisers and flexibilisers such as glycerol or ethylene glycol to produce uncoated film which is non-moistureproof and non-heat sealable. To make it moisture-proof and heat sealable, a coating of nitrocellulose is applied. To obtain a superior moisture barrier a coating of vinyl chloride-vinylidene copolymer (Saran) is applied. A wide variety of grades of such film are manufactured. There are many possible variations with regenerated cellulose film. These include various degrees of moisture-proofness and single sided coatings. Variants which are in regular production include the following types: nitrocellulose coated on both sides (moisture-proof), heat sealable, non moisture-proof, nitrocellulose coated on one side only and copolymer coated on both sides. Regenerated cellulose film has a light transmission equal to that of glass. Dry cellulose film is practically impermeable to the permanent gases but becomes permeable when wet. Moisture vapour permeability is very high unless the film is coated. 24

Types of Polymers Used in Commodity Packaging Both moisture-proof and non-moistureproof film have a wide variety of uses including important applications in the food packaging field. Non-moistureproof films are used when protection from grease and dirt is required but where mould growth would be a problem if a good moisture barrier were used. Examples of its use in this area include packaging of meat pies, cakes, fondants, chocolate coated sweets and fresh sausages. Moisture-proof films are used for the wrapping of hygroscopic foodstuffs such as biscuits, sugar, confectionary, potato crisps, bread and shelled nuts. Moisture-proof film is also used for the packaging of hygroscopic sweets, pharmaceuticals and dried milk. Single side coated film is used for the packaging of fresh meat with the uncoated side in contact with the meat. The moistening of the film by the meat raises its permeability to gases including oxygen and so gives it the right combination of conditions to preserve the fresh colour of the meat. A single-sided coated film is also used when extrusion coating with polyethylene is required. The polyethylene readily adheres more securely to the base cellulose film than to the nitrocellulose coating. This type of laminate is used for the vacuum packing of bacon, cheese and coffee.

2.26.2 Cellulose Derivatives • Cellulose triacetate and diacetate (secondary cellulose acetate)

This is a hard tough material made by the acetylation of cellulose. It is sensitive to moisture pickup and is dimensionally unstable. It is resistant to weak acids and alkalies but is decomposed by strong ones. Both the triacetate and the secondary cellulose acetate give crystal clear films with high gas and water vapour permeabilities.



Cellulose triacetate has limited use in applications involving contact with food (cutlery handles). The secondary acetate film is widely used in packaging, principally as a laminate. It is not readily heat sealed. It does, however, readily accept printing inks. Cellulose secondary acetate is normally used as the outer layer of a laminate with an inner heat scalable coating. Printing is carried out on the inside of the acetate thus giving a glossy decoration. Secondary acetate film is also used as the window in cartons as it has good adhesion to cardboard. The high permeability of cellulose acetate film to water vapour and oxygen limits its uses in fresh food packaging. It is, however, particularly useful if a breathable film is required allowing the inward passage of carbon dioxide. This renders it particularly useful in certain specific food packaging applications.



Thicker section secondary acetate film is used for the manufacture of rigid food containers and vacuum drawn cellulose acetate containers are used in packaging sweets and chocolates at the high price end of the market. 25

Additive Migration from Plastics into Foods • Ethyl cellulose

This is manufactured by reacting cellulose with ethyl chloride. It is tough and retains flexibility and impact strength down to –40 °C and has a similar moisture pickup to cellulose acetate. It is widely used as a moulding material and in film manufacture.

2.27 p-Hydroxybenzoic Acid and 2-Hydroxy-6-naphthoic Acid These polymers have been used as coating on non-plastic substrates for the packaging of fatty commodities.

2.28 Polymers Used in Gas Barrier Applications These polymers, some of which are listed next are used to modify the atmosphere within packaged foods, beverages, pharmaceuticals and cosmetics and to control the ingress of, for example, oxygen, ethylene-vinyl alcohol, perfectly alternating polyketones, thermoformable ethylene-vinyl, alcohol blends, polylactide, and polyethylene terephthalate.

2.28.1 Rubber and Elastomers These contain additives and their breakdown products such as: benzothiazole, 2-mercaptobenzothiazole, N-cyclo-2-benzothiazole sulfonamide and 2-mercaptobenzothiazyl disulfide. There are two main types of rubbers/elastomers used in packaging and processing: natural rubbers and nitrile rubbers. Both of these are used to produce seals, gaskets, valves, diaphragms and hoses. Natural rubbers are used in latex dipping to form feeding teats and rubber gloves. Filled natural rubbers are used to produce hoses and conveyor belts used in food production. Nitrile rubbers are synthetic copolymers of butadiene and acrylonitrile. These are used more widely than natural rubbers to produce hoses, dairy equipment, seals, gaskets, car seals and bottle closures. Other synthetic rubbers used in packaging include ethylene - propylene - diene terpolymers and vinylidene fluoride - hexafluoropropylene copolymer. Use in food production includes the packaging of milks, yoghurts, infant formulae and beverages. 26

3

Non-Polymeric Components of Plastics

All polymers, in addition to the basic plastic, usually contain several, if not a multiplicity, of non-polymeric components in amounts from less than one part per million to several percent. The presence of these substances obviously have an implication in the suitabilities or otherwise of the plastic for applications involving contact with foodstuffs. Thus, although the plastic itself, due to its very high molecular weight, might not contaminate the foodstuff, it is apparent that certain additives, which are usually of relatively low molecular weight, and therefore of higher solubility, will be transferred from the plastic to the foodstuff during storage. This raises questions regarding the toxicity of the additives, the amounts which transfer and the possible implications of this, from the toxicity point of view, as far as the food consumer is concerned. Non-polymeric components are present in plastics either unavoidably as a result of the process of manufacture, or as the result of deliberate additions to the plastic in order to improve some aspect of ease of manufacture or final polymer properties. Thus, non-polymeric components can be subdivided into three groups: • Polymerisation residues, • Processing aids, and • End-product additives. Polymerisation residues cover substances whose presence is to a large extent unavoidable, such as low molecular weight polymers, oligomers, catalyst remnants, and polymerisation solvents. Raw material non-polymerisable impurities, are impurities picked up from plant materials. Processing aids include such substances as thermal antioxidants and heat stabilisers added to prevent decomposition of the polymer during moulding and slip additives to facilitate moulding. End-products are deliberately added to the polymer either during manufacture or subsequently to improve the properties of the final polymer. As discussed next, a very wide range of non-polymeric substances are used in this category ranging from secondary thermal antioxidant to impact improvers, plasticisers, ultraviolet stabilisers, antistatic agents, and so on. 27

Additive Migration from Plastics into Foods

Table 3.1 Non Polymeric Components Found in a Sample of Polypropylene Determined

Concentration (ppm) Origin

Polymerisation residues Aluminium

40

Titanium

5

Chlorine

230

Sodium

30

Potassium

18

Silicon

<20

Iron

Remnants of Ziegler organoaluminium-titanium halide catalysts Catalyst neutralisation alkalies

5

Manganese

0.1

Nickel

8

Tin

0.3

Vanadium

<0.2

Zinc

<10

Chromium

General contamination from plant

1

C6/C16 hydrocarbons

900

Dissolved propylene

1 cm3/cm3 polymer

Unreacted monomer Residual polymerisation solvent

Processing aids Calcium

40

Possibly calcium stearate stabiliser for protecting polymer during moulding

Ionol

200

Antioxidant (thermally degraded Ionol also present)

1000

Possible barium sulfate filler

End product additives Barium Substituted benzopheno (2-hydroxy 4 noctoxybenzophenone)

Light stabiliser

Dilaurythiodipropionate

Probable secondary antioxidant

Reproduced from Author’s own files

28

Non-Polymeric Components of Plastics The situation regarding the presence of non-polymeric, low molecular weight additives in polymers can be best illustrated by an actual example concerning a batch of polypropylene which, upon detailed but by no means complete examination, was shown to contain the constituents listed in Table 3.1. In many instances the probable origin of these substances can be attributed. Thus, as shown in Table 3.1, a single polymer might contain Ziegler catalyst remnants, neutralising chemicals, residual monomer and polymerisation solvent (polymerisation residues), a calcium salt and an antioxidant (processing aids) and a further antioxidant also a light stabiliser and a filler (end product additives). Further information on the various types of non-polymeric components that can occur in polymers is discussed in the next section.

3.1 Polymerisation Residues Low pressure polyethylene might contain minute traces of oxygen or nitrogen or sulfur chain transfer catalyst residues. These are usually labile and can be ignored from the foodstuff contamination point of view. Polymers such as polyethylene, polypropylene and polystyrene manufactured by a catalysed low pressure route will however, usually contain appreciable catalyst residues usually appearing in the form of aluminium, chromium, magnesium and titanium, also possibly lithium and sodium. As the example quoted in Table 3.1 indicates these impurities can occur in the polymer at levels approaching 100 ppm. Suspension or emulsion processes for the polymerisation of styrene or vinyl chloride can impart to the polymer significant amounts (up to 0.05%) of proton donating processing chemicals such as potassium and ammonium persulfate or benzyl and lauryl peroxide and their decomposition products, (potassium sulfate, ammonium carbonate, urea and benzoic or lauric acid, respectively), and these are of significance in relation to safety in use for food. Catalysts and accelerators used in the manufacture of thermosets include peroxides, organic nitrogen compounds and copper and cobalt salts of naphthenic and other organic acids.

3.2 Residual and Unreacted Starting Materials Many manufactured polymers contain either unreacted monomer, or low molecular weight polymer (oligomers) in amounts varying from a few parts per million to several percent depending on the manufacturing process used and the type of polymer. Low molecular weight products are particularly prevelent in thermosets. However, even 29

Additive Migration from Plastics into Foods polymers such as polyethylene contain a small amount (usually less than 1%) of a waxy low molecular weight tail, whose solubility characteristics and therefore extactability into foods is different from that of the main polymer product. Low pressure polyolefins can in fact contain up to 1% on a volume - volume basis of unreacted monomer, especially when the polymer is newly manufactured. Indeed, simple monomers, such as styrene, ethylene, propylene, hexene, vinyl chloride, acrylonitrile and caprolactam, usually do occur in the corresponding polymers. In addition to unreacted monomer, any non-polymerisable impurities in the original monomer feed to the polymerisation could occur in the final product. Thus, styrene monomers can contain low concentrations of numerous saturated and unsaturated hydrocarbons, ethyl benzene being particularly prevelent and these, particularly the saturated compounds which do not polymerise, will occur in the finished polymer and have implications in the use of the polymer food packaging. It is not unknown for compounds as toxic as benzene to occur at very low concentrations, usually less than 10 parts per million in styrene monomer, and this could, therefore, also occur in the polymer. For foodgrades of polystyrene, the monomer content is usually nowadays limited to 0.2% maximum. Acrylonitrile monomer may be found in amounts up to 0.1% in finished polymer, whilst negligible amounts of monomer are found in polyamide and polymethyl-1-pentene. With thermosets, phenol and formaldelyde are likely to be found even in the most carefully manufactured grades. Regarding oligomers, only those in the relatively very low molecular weight range, and these are usually only greases, are of sufficient potential significance to be regarded as non-polymeric impurities, which may have foodstuff packaging implications. Thus, polypropylene may contain traces of dimer (C6H12) and tetramer (C12H24) hydrocarbons with molecular weight up to approximately 200. The full chemistry of the low molecular weight tail composition, has not yet been fully studied in many polymer systems, especially in the case of copolymers involving two or more monomers.

3.3 Polymerisation Medium Particularly for Ziegler-Natta type, low pressure polymerisations of olefins, the reaction is usually performed in an inert paraffinic solvent medium (from C4 to C18). In such cases, of course, traces of this solvent are found in the final polymer which persist for a long time and, indeed, are difficult to remove even by treatment of the polymer in vacuum. For polystyrene and polyvinylchloride (PVC), made by the suspension or emulsion processes, the reaction medium is an aqueous solution containing wetting agents, detergents, soaps and emulsifiers. Traces of all of these will occur in the final product. 30

Non-Polymeric Components of Plastics

3.4 Catalyst Decomposition Agents Upon completion of the polymerisation of polyolefins by the Ziegler-Natta, low pressure routine, the organo-aluminium-titanium halide catalyst is decomposed and neutralised by the addition of low molecule weight alcohols and possibly aqueous wetting agents and soaps whilst pH control may be effected by the addition of aqueous alkalies such as sodium carbonate. This stage of the process can therefore introduce alcohols and alkali metal salts into the polymer. Similar comments apply in the case of the ferrocenyl catalysts now being used in polyolefin synthesis.

3.5 Other Impurities Introduced During Polymerisation These include general contamination from plant materials such as copper, iron, oil, and silica, and so on. Further sources of contamination include polymerisation inhibitors (styrene and vinyl chloride monomers), emulsifying, suspension and chain transfer agents (polystyrene, PVC).

3.5.1 Chemicals Added During Polymerisation In this category is included the addition of up to 10% mineral oil (to impart flexibility to the product during the manufacture of high impact polystyrene by the copolymerisation of styrene and synthetic rubbers).

3.6 Processing Aids The main types of processing aids are discussed next under separate headings: Antiblock agents Antioxidants Antisplit agents Antistatic agents Heat stabilisers Lubricants Melt strength improvers Mould release agents Plasticisers 31

Additive Migration from Plastics into Foods Slip additives Other stabilisers Gas barrier agents

3.6.1 Antiblock Agents The principal antiblock agent used is silica at the 0.1 to 0.5% level. Its function is to prevent sticking in thin films of the polymer.

3.6.2 Antioxidants Antioxidants which prevent degradation of the polymer by reaction with atmospheric oxygen may be required during moulding operations on the polymer and will be needed to prevent oxygen pickup and oxidation and embrittlement of polymer during long-term usage. The reaction of polymers with oxygen are chain reactions involving hydroperoxy radicals and these reactions can be inhibited or slowed down by compounds known as antioxidants which interrupt the chain reaction at some point. Typical antioxidants are notably hindered phenols and organic sulfides. Some of the more important antioxidants are: Irganox 1076 – Octadecyl-3-(3,5-di-tert-butyl-4-hydroxyphenyl)propionate Irganox 1010 – Pentaerythritol tetrakis(3-(3,5 di-tert butyl-4-hydroxyphenyl) propionate) Ionox 330 – 1,3,5 trimethyl 2,4,6-tris-(3,5 di-tert-butyl-4-hydroxybenzyl)benzene

3.6.3 Antisplit Agents These are used to prevent spontaneous fibrillation of oriented polypropylene film in processing equipment. These additives are usually natural or synthetic rubber added at concentrations up to 10%.

3.6.4 Antistatic Agents Plastics, being good electrical insulators retain electrostatic charges developed by friction between the plastic itself, between plastic and moving machinery, or by 32

Non-Polymeric Components of Plastics electroinisation from dust or radiation. These charges on the plastic lead to end-use problems such as sticking of flowing polymer powders, sticking together or polymer film and dust attraction with the development of a dirty appearance. Also, discharges to earth can lead to the formation of pin-holes in films and electric shocks to operators, fire hazards and explosions of stored petroleum. The requirements of an antistatic agent are a reasonable electrical conductivity and an ability to migrate to the surface of the plastic moulding or film as it is on the surface that the electrostatic charge concentrates. The selection of the types of compounds used as antistatic additives is governed by complex considerations, one of which is that it should have the correct degree of compatability with the base polymer which will enable it to migrate to the polymer surface at a controlled rate during service life. The majority of antistatic agents now in use are either glycol derivatives or quarternary ammonium salt derivatives. Lauric diethanolamide is a typical antistatic additive used in the formulation of polyolefins. The reasons for including antistatic agents in a polymer formulation intended for food packaging are mainly the avoidance of sticking together of films and thin plastic sections such as cups during food packaging operations. Avoidance of dust attraction in film and mouldings is a lesser problem in the case of food packaging due to the relatively low interval of time intervening between packaging and sale.

3.6.5 Heat Stabilisers These are incorporated into the polymer to protect it from decomposition during the short time that it is held at a relatively high temperature in the moulding machine. Low-density polyethylene and polyamide are generally sufficiently stable not to require heat stabilisers. Polystyrene requires little stabilisation and high-density polyethylene, polypropylene and PVC may require significant additions of such stabilisers. A wide range of heat stabilisers are available and the choice is dictated by considerations of the temperature to be encountered and the time at which the polymer is held at an elevated temperature, also the presence of or absence of atmospheric oxygen and of antioxidants.

3.6.6 Lubricants Internal lubricants are used to reduce the viscosity of the molten polymer in the extruder. These additives may be virtually any compatible stable compound and 33

Additive Migration from Plastics into Foods include compounds such as plasticisers and C12 – C30 hydrocarbons added at the several percent level. External lubricants are used to reduce the friction between the polymer and the surface of the extruding equipment. To function effectively these should be not too soluble in the plastic in order to enhance the concentration of lubricant at the surface. The most commonly used external lubricant is calcium stearate added at the 0.05 to 0.3% level.

3.6.7 Melt Strength Improvers In extrusion - blow moulding operations, molten polymer passes through a stage where it is processed with little external support. In these circumstances the moulding is likely to distort. The incorporation of a melt strength additive reduces the possibility of this occuring. Frequently, the materials used to improve melt strength are the same as those used for internal lubrication or plasticisation.

3.6.8 Mould Release Agents These are used to coat the mould in order to reduce the possibility that the moulded article will stick. Silicones are usually used for this purpose now, although paraffin oil or petroleum jelly are still used to a small extent.

3.6.9 Plasticisers Almost any soluble organic compound can be used to plasticise a polymer including high boiling point alcohol esters of phthalic, adipic, sebacic and phosporic acids and ethylene oxide condensates and polychlorinated hydrocarbons. Polymeric high molecular weight plasticisers are occasionally used. Due to their lower solubility in foods, the general tendency is for these to extract from plastics into food at a lower rate than non-polymeric plasticisers. Good compatability, low volatility, freedom from coloration, toxic hazard or food tainting are factors to be considered in the selection of plasticisers. Plasticisers are seldom added at concentrations exceeding 5% to plastics intended for food contact use. Some important plasticisers include di-2-ethylhexyl phthalate used in PVC formulations, dipentyl phthalate, di-(2-ethyl hexyl) adipate di-octyladipate, diethyl phthalates, diisobutylphthalate and di-n-butyl phthalate used in polyethylene. Polymeric plasticisers include poly(1,2-propylene adipate) and polybutylene succinate. 34

Non-Polymeric Components of Plastics

3.6.10 Slip Additives These are, in effect, external lubricants that operate in the solid state. The most commonly used slip additives are fatty acid amides such as erucamide. These additives are used to prevent blocking in thin films of polymer, e.g., sticking together of thin film bags.

3.6.11 Other Stabilisers To reduce the formation of large crystallites in polypropylene upon heating and slow cooling with consequent loss of impact strength and transparency it is common practise to incorporate in the polymer, nucleating agents such as tertiary butyl benzoic acid or its salts or salts of aromatic sulfonates. To prevent the thermal decomposition of PVC with the consequent evolution of hydrogen chloride, stabilisers such as barium, cadmium, zinc, or calcium salts of fatty acids, and organotin compounds or organophosphites are used to act as acid acceptors. Polydimethylsiloxones have been used as foam suppressors and antifoamimg agents in polymer manufacture.

3.7 End-Use Additives The main types of end-use additives are listed next and are discussed under separate headings: Antiblock additives Antifungal agents Antioxidants Antistatic agents Bactericidal agents Brighteners and whiteners Colorants Expanding agents Impact improvers Lubricants Plasticisers Ultraviolet protective agents and ultraviolet degradation inhibitors Printing ink adhesives Gas barrier packaging oxygen scavengers 35

Additive Migration from Plastics into Foods

3.7.1 Antiblock Additives These are sometimes incorporated into the polymer after moulding as well as in the melt (see also Section 3.6 – Processing Aids).

3.7.2 Antifungal agents These are rarely used in food grade polymers, their use being confined to film or sheet for medical use.

3.7.3 Antioxidants In addition to being used as processing aids, additional antioxidant is sometimes incorporated in the polymer after moulding. This applies to applications where relatively high temperatures are involved as in contact with hot foods, infrared ovens or in tropical areas where substantial exposure to ultraviolet radiation is likely to be encountered.

3.7.4 Antistatic Agents Again, this type of additive might be incorporated after moulding as an alternative to or in addition to inclusion as a processing aid.

3.7.5 Bactericidal Agents Although plastics are usually immune to bacterial attacks there is an exception in the case of plasticised PVC where the plasticiser is susceptible to such attack. The majority of bacteriocidal agents used in polymer formulations are quarternary ammonium compounds.

3.7.6 Brighteners and Whiteners This type of additive (otherwise known as optical bleaching agents), is used to off-set the off-white or pale yellow discoloration of many types of plastics in their moulded state. Brighteners are added to enhance appearance. These substances operate by absorbing incident radiation of suitable wavelength, converting this, and then emitting radiation of a higher frequency in the visible spectrum or in the ultraviolet region. The eye interprets this as a whitening or brightening effect. 36

Non-Polymeric Components of Plastics When used, optical brighteners are usually derivatives of stilbene or thiophen and are incorporated in the polymer at very low concentrations, usually in the 100 parts per million region.

3.7.7 Colorants There are two main ways in which colorants can be used in the manufacture of food grade plastics. Firstly, there is the use of printing inks for decorative or labelling purposes. These inks are applied to the plastic surface which is not in contact with the foodstuff and do not therefore present a health hazard problem. Secondly, there are the types of colorant which are incorporated in the bulk of the plastic, usually at some stage when the plastic is molten. In this method of colouring, the colorant may be present in the form of a fine insoluble dispersion of pigment which, hopefully, is well dispersed or as a solid solution, i.e., a dyestuff. Most pigments are inorganic and include such substances as titanium dioxide (whitening) in the concentration range 0.01 to 1%, cadmium sulfides or sulfoselenides (yellow, red, brown) at the 0.1% level and carbon blacks at the 0.2 to 2% level. In addition, high molecular weight organometallic pigments are occasionally used in food grade plastics such as the anthraquinones (blue, green) and other stable organic pigments. Dyestuffs are usually completely organic and include many of the substances used in textile printing. These are usually incorporated in plastics in the 10 to 1000 parts per million range.

3.7.8 Expanding Agents (Blowing Agents, Foaming Agents) These are used principally in the manufacture of expanded polystyrene foam which is used extensively in the manufacture of vending cups. There are three principal types of expanding agents in use in polymer manufacture, all of which leave residues in the manufactured polymer. Aliphatic hydrocarbons (C4 to C7) are dissolved into polystyrene granules which are then treated with steam to expand the granules into a cellular form. Up to 0.5% residual hydrocarbon can remain in the expanded polymer for a period of several months. Mixtures of sodium carbonate or bicarbonate and citric acid, are blended into the polymer upon heating decompose to produce carbon dioxide to expand the polymer and leave residues of sodium citrate in the polymer. Labile nitrogen compounds, such as azo-dicarbonamide, upon heating, liberate nitrogen to produce a cellular structure and leaving possible traces of the decomposed azo compound in the polymer. 37

Additive Migration from Plastics into Foods

3.7.9 Impact Improvers These additives are used to overcome the inherent brittleness of polymers such as polystyrene, polypropylene, and Terylene. The additives are incorporated into the polymer during manufacture. Pigments, extenders, fillers, nucleating agents, hydrocarbon oils, waxes and rubbers are all used as a means of improving the impact strength of these polymers.

3.7.10 Lubricants External lubricants are used to reduce adhesion between stacks of moulded articles or to reduce friction between moving parts. Calcium stearate at the 0.05 to 1.0% level, and extenders and plasticisers and occasionally antistatic agents all have external lubricating properties.

3.7.11 Plasticisers Plasticisers are incorporated into the more rigid plastics, particularly PVC, to make them more flexible. The types of compounds used as plasticisers also act as impact improvers due to their ability to reduce polymer brittleness. Plasticised PVC pipe is used extensively for transferring beer from the keg to the dispenser. Plasticisers are also used to some extent in cellulose acetate and polyethylene terephthalate, in film, sheet and pipe form.

3.7.12 Ultraviolet Protective Agents Ultraviolet protective agents are used either to protect the plastic itself from strong sources of ultraviolet light such as sunlight and also to protect the packaged foodstuff from such radiation and possibly from the effects of strong lighting when on display in supermarkets. An instance where such protection is required is in the packaging of vitamin C containing cordials where, without protection, ultraviolet light would produce severe degradation of the vitamin. Ultraviolet protective agents fall into two categories, those operating as UV screens, and those which operate by interfering with a degradation chain reaction in the polymer that has been initiated or catalysed by the radiation. Two commonly used UV stabilisers are (2,2-(hydroxyl-3-tert-butyl-5-methylphenyl)5chlorobenzotriazole and 2-hydroxy-4-n-octyloxy-benzophenone. 38

Non-Polymeric Components of Plastics

3.7.13 Ultraviolet Screens Ultraviolet screens act by absorbing or reflecting harmful radiation and converting it to harmless radiation of a different wavelength. All opaque, non-highly coloured additives such as fillers, and carbon black have a beneficial effect as ultraviolet screens when incorporated at the 2 to 3% level. Such compounds protect not only the plastic package from ultraviolet degradation but also the packaged food itself, and in fact, over 90% of all radiation transmission can be eliminated by the incorporation of 0.1 to 0.2% of particular pigments notably carbon black.

3.7.14 Ultraviolet Degradation Inhibitors These are another category of materials to the filler and pigment screens mentioned previously. These materials are generally organic and include derivitives of thiophen, benzotriazole and transition metal dithiocarbamates, e.g., Ferro 101, nickel dithiocarbamate. These substances are effective at concentrations of 0.1% or less. UV curable acrylate coatings and inks have been used in the manufacture of printed polymer containers.

3.7.15 Gas Barrier Packaging Oxygen Scavengers Various materials have been used to avoid damage to packaged commodities and, in some cases, the package by atmospheric oxygen. An obvious example, as mentioned previously, is the use of UV protective agents in the protection of vitamin C containing drinks. Other oxygen scavenger systems include: • Iron oxide sachets, • Cobalt catalysed Nylon MXD6, • Ciba Shelfplus O2 oxygen scavenger, • OSP oxygen scavenging barrier (based on an oxidisable resin ethylene-methylacrylate containing a photoinitiator (Chevron Phillips Chemical Co.), and • A cobalt salt catalyst, Amosorb oxygen scavenger (BP Chemicals). The European Commission provides in Synoptic Document No.7 [1] published in 1990, a provisional list of additives that can be used for the production of food 39

Additive Migration from Plastics into Foods contact plastics. The Synoptic Document, anticipated a Directive on additives used in food contact plastics. In support of this Directive, 100 of the most important additives have been selected to establish a reference collection and a handbook on their spectroscopic and physicochemical data. This new collection of additives used in plastics for food contact [2] can be seen as a logical supplement to the earlier collection of monomers and other starting materials. That earlier collection was described in two handbooks entitled Spectra for the Identification of Additives in Food Packaging [3] and Spectra for the Identification of Monomers in Food Packaging [4]. These books presented information on monomeric substances listed in Directive 90/128/EEC and its amendments which restrict the range of monomers and other starting substances that can be used for the production of plastic materials and articles intended for food contact applications. The preparation of a reference collection and a handbook of analytical data of additives was undertaken with funding from the European Commission under the Standards Measurements and Testing Programme (SMT). Van Lierop and co-workers collected spectra for the identification of the 100 most important additives used in plastic packaging and coatings. Selection of the additives was made after extensive consultation with researchers in the field and with representatives from European industry including the Food Contact Additives Panel (FCA) sector group of the European Chemical Industry Council (CEFIC). As before [3, 4], Fourier Transform infrared and mass spectra were recorded, but Van Lierop and co-workers [5] extended the scope to include proton nuclear magnetic resonance (1H-NMR) spectra and gas chromatographic (GC) data. GC chromatographic retention times were recorded to facilitate identification by retention index data. Chromatographic methods were used to indicate the presence of any impurities in the commercial chemicals. Samples of the reference substances are available on request and the collection of spectra and other information will be made available in printed format and on-line through the Internet. Van Lierop and co-workers [5] give an overview of the work done to establish the reference collection and the spectral atlas, which together will assist enforcement laboratories in the characterisation of plastics and the selection of analytical methods for additives that may migrate.

References 1.

40

Synoptic Document No.7, Draft of provisional list of monomers and additives used in the manufacture of plastics and coatings intended to come into contact with foodstuffs, European Commission Document CS/PM/2356, Brussels, Belgium, 15th May 1994.

Non-Polymeric Components of Plastics 2.

J.B.H. Van Lierop, L. Castle, A. Feigenbaum and A. Boenke, Spectra for the Identification of Additives in Food Packaging, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1998.

3.

J. Bush, J. Gilbert and X. Goenaga, Spectra for the Identification of Monomers in Food Packaging, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1994.

4.

J. Gilbert, J. Bush, A. Lopez de Sa, J.B.H. van Lierop and X. Goenaga, Food Additives and Contaminants, 1994, 11, 1, 71.

5.

B. van Lierop, L. Castle, A. Feigenbaum, K. Ehlert and A. Boenke, Food Additives and Contaminants, 1998, 15, 7, 855.

41

Additive Migration from Plastics into Foods

42

4

Determination of Antioxidants

4.1 Santonox R (4,4′-thiobis-6-tert butyl m-cresol) in Aqueous and Non-aqueous Extractants Ultraviolet spectroscopy is applicable to the determination of phenolic antioxidants in aqueous and alcoholic simulant liquids and also to one of the fatty simulants, liquid paraffin, recommended by the British Plastics Federation (BPF). Full details of the procedure, which is capable of determining down to l-2 ppm of Santonox R in the simulent liquids, is given next. With the exception of the 5% citric acid simulent, errors are generally of the order of less than ±10% of the determined concentration. Typical recoveries of Santonox R in the 5% citric acid extractant were between ±15% of the expected result.

4.1.1 Method Reagents: Citric acid AR Ethanol AR Sodium carbonate AR Liquid paraffin BP (OD in 1 cm cell at 280 nm < 0.2 versus water) Cyclohexane - for spectroscopy 5% w/v citric acid - prepared from citric acid AR and deionised water 50% w/v ethanol solution - prepared from ethanol and deionised water Quartz cells of 1 cm path length are used for all spectroscopic measurements.

4.1.1.1 Procedures • Using Deionised Water as Extractant Shake the vessel containing the test extractant vigorously to obtain a uniform haze of Santonox in suspension in deionised water and immediately dispense 44.1 ml of 43

Additive Migration from Plastics into Foods the test extractant into a standard flask and dilute to 100 ml with ethanol. Wait half an hour before proceeding. Obtain the spectrum of this solution and of the blank (50% w/v aqueous ethanol solution). Then determine the optical density ‘d’ of the Santonox R absorption at 250 nm. Construct a calibration curve for Santonox R in 50% w/v ethanol in the concentration range 2 to 20 ppm. Suitable standard solutions may be prepared by quantitative dilution of a 100 ppm solution of Santonox R in neat ethanol adjusting the ethanol content of the standard solutions to 50% w/v. Note that the concentration of Santonox R in the test extractant is 2.27 times the concentration of Santonox R in the ethanolic solution prepared previously.

• Using 5% w/v Citric Acid as Extractant Shake the vessel containing the test extractant vigorously to obtain a uniform haze of Santonox R in the citric acid solution. Pipette a 25 ml aliquot into a 100 ml separatory funnel, followed by 25 ml of cyclohexane. Shake the separatory funnel vigorously every ten minutes for half an hour. Allow the layers in the funnel to separate. Place an aliquot of the (upper) cyclohexane layer in the sample cell of the spectrophotometer and record a spectrum of the extractant and of a sample of cyclohexane using a 1 cm cell containing cyclohexane in the reference beam. Then obtain the optical density of the Santonox R absorption at 250 nm. Construct a calibration curve for Santonox R in cyclohexane, by obtaining optical densities of 2.5, 5, 10 and 20 ppm solutions at 250 nm. Compare the optical density obtained for the test extractant with the calibration curve and thus obtain the concentration of Santonox R in the test extractant.

• Using 50% w/v Ethanol/Water as Extractant Dispense a 10 ml sample of the extractant liquid into a 1 cm quartz absorption cell. Fill an identical cell with 50% w/v ethanol solution. Measure the optical density at 250 nm. Calculate the Santonox concentration of the solution by reference to a calibration curve for Santonox in 50% w/v ethanol/water. 44

Determination of Antioxidants • Using 5% w/v Sodium Carbonate Solution as Extractant Shake the vessel containing the test extractant vigorously. Obtain a suspension of Santonox in 5% w/v sodium carbonate solution, and immediately measure 44.1 ml of the extractant and transfer to a 100 ml volumetric flask. Slowly add 7 ml of glacial acetic acid and allow the effervescence to subside. Dilute to 100 ml with ethanol. Prepare a blank solution in a similar manner by adding glacial acetic acid and ethanol in the previous proportions to a 5% w/v sodium carbonate solution. Use this solution in the 1 cm reference cell of the spectrophotometer and measure the optical density of the test extractant at 250 nm versus the blank in the reference cell. Prepare a calibration curve for Santonox in the glacial acetic acid - ethanol solution and obtain the Santonox concentration of the test extractant from the calibration curve. The Santonox concentration of the original extractant is obtained by multiplying this figure by 2.27.

• Using Liquid Paraffin BP as Extractant Dispense a 25 ml aliquot of the extractant into a 50 ml standard flask and dilute to the mark with cyclohexane. Record the spectrum of this solution prepared using 50% w/v liquid paraffin in cyclohexane in the reference cell. Then obtain the optical density due to the Santonox R absorption at 250 nm. Construct a calibration curve for Santonox R in 50% v/v liquid paraffin in cyclohexane for concentrations ranging from 0 to 20 ppm. Compare the optical density obtained with the test extractant with the calibration curve and hence obtain the concentration of Santonox R in the test extractant. This concentration is one-half the concentration of Santonox in the extractant. It must be emphasised that methods based on direct spectroscopy of the extractant liquid such as that described previously are only capable of giving a correct analysis if no other substances which absorb at the same or a nearby wavelength are present in the extractant.

4.2 DLTDP in Aqueous and Non-aqueous Extractants The methods discussed next, are based on (a) thin-layer chromatography, (b) infrared (IR) spectroscopy and (c) oxygen flask combustion. These methods are discussed in 45

Additive Migration from Plastics into Foods detail here to illustrate the careful way in which method development must be carried out in extractant analysis.

(a) Thin-Layer Chromatography Preliminary ether extraction of dilaurylthiodipropionate (DLTDP) from the aqueous extractants. In this method the distilled water, sodium carbonate and 5% citric acid extractants (80 ml) are transferred to a separatory funnel. For the distilled water extractant only, solid sodium chloride (10%) is added to assist the extraction of DLTDP. The addition of sodium chloride to the distilled water extractant before ether extraction was essential in order to obtain good recoveries. Omission of the sodium chloride resulted in recoveries in the order of 5%. Addition of sodium chloride to citric acid solutions was deleterious; the citric acid was salted out in the ether, and thus remained on evaporation of the ether. Some of this residue was dissolved along with the DLTDP and was present on the thin-layer chromatograph plate. Neutralisation of the citric acid before extraction resulted in low recoveries. The aqueous phase is extracted with several portions of diethyl ether which are subsequently combined and dried by shaking with anhydrous sodium sulfate. The ether extract is evaporated to dryness and the residue transferred with small volumes of ether to a 2 ml volumetric flask and made up to volume with ether. The ether alcohol:water extractant is evaporated to dryness on a water bath and transferred to a 2 ml volumetric flask as described previously. DLTDP is determined in the ether extracts as described next. The liquid paraffin extractant is diluted with petroleum ether 60 °C to 80 °C in a separatory funnel. The solution is then passed down a column of chromatographic grade silica gel which retains the DLTDP and allows liquid paraffin to percolate through the column. Petroleum ether is then passed through the column to remove the last traces of liquid paraffin. A mixture of chloroform and methanol is then poured down the silica column to desorb DLTDP which is quantitatively recovered in the column effluent. The total effluent is evaporated to dryness in a steam bath and the residue made up to a standard volume with diethyl ether. DLTDP is determined in the extracts as described next. The thin-layer chromatographic separation of DLTDP is carried out on a glass plate coated with a 1 mm thick layer of Merck No.7731 silica gel. Accurately measured volumes of the ether extracts of the various extraction liquids are pipetted onto the plate together with a range of synthetic comparison solutions of known concentrations 46

Determination of Antioxidants of DLTDP dissolved in diethyl ether. The chromatogram is developed with a mixture of acetone and cyclohexane [15:85 (v/v)]. The plates are then sprayed with an alkaline solution of potassium permanganate, where the DLTDP shows as yellow spots on a pink to red background (Rf value of DLTDP = 0.65). The intensity of the sample spots is compared with that of the standard DLTDP spots as soon as possible, and the DLTDP content of the sample solutions deduced accordingly. It is then possible to calculate the DLTDP content of the original extraction liquids. The thin-layer chromatographic method described next is sufficiently sensitive to determine DLTDP in amounts down to 1 ppm in the various extraction liquids. An additional advantage of this procedure, is that because it incorporates a chromatographic separation step, interference effects due to any other polymer additives present in the extraction liquid are minimised, unless these substances migrate on the plate at the same rate as DLTDP itself.

Method • Reagents All reagents should be Analar (AR) quality wherever possible: Acetone Diethyl ether DLTDP Cyclohexane Silica gel (Thin-layer chromatographic grade - particle size 5-20 with calcium sulfate binder, e.g., Merck No.7731) Sodium chloride Sodium sulfate (Anhydrous) Sodium carbonate (Anhydrous) Chloroform Methanol Petroleum spirit (Boiling range 60-80 ºC) Silica gel [Chromatographic grade - particle size 0.05 to 0.20 mm, e.g., Merck No.7734 (for column chromatography of liquid paraffin extractant)] 47

Additive Migration from Plastics into Foods Cyclohexane - acetone solvent mixture (Mix 15 ml of acetone with 85 ml of cyclohexane) Potassium permanganate spray reagent (Dissolve 0.08 g of potassium permanganate and 0.25 g of anhydrous sodium carbonate in 100 ml of distilled water. This solution is stable for about three days and should therefore be prepared as required) Chloroform - methanol solvent mixture (Mix 10 ml methanol with 90 ml chloroform) Standard DLTDP solution

(Dilute 0.1000 g of DLTDP to exactly 100 ml with diethyl ether in a graduated flask. This solution contains 1 mg/ml of DLTDP)

The use of Analar grade solvents throughout is essential to avoid streaking at the thin-layer chromatographic stage of the analysis. Samples of DLTDP in the various extractants were prepared by diluting a known volume of a stock solution of DLTDP in ethanol, containing DLTDP 1 mg/ml to 1 litre, with the appropriate extractant. With distilled water, ethanol - water, and 5% sodium carbonate solution, satisfactory solutions were produced with 5 ml of stock solution. With 5% citric acid solution, 5 ml of stock solution gave a cloudy solution indicating that DLTDP was not soluble to the extent of 5 mg/ml in 5% citric acid. However, 4 ml of stock solution gave a clear solution.

• Apparatus Thin-layer spreading device Thin-layer chromatography plates Thin-layer chromatography tank Spray bottle 2 ml graduated flasks 100 ml separating funnels 48

Determination of Antioxidants • Sampling Aqueous and 50% ethyl alcohol and aqueous extractants – 80 ml of sample is sufficient for one determination. Liquid paraffin extractant - 70 g of sample is sufficient for one determination.

Procedure • Blank Tests These are unnecessary since the materials from which the extractant solutions are prepared are unlikely to contain DLTDP.

• Thin-layer Chromatographic Plates To prepare five plates, each 20 cm × 20 cm, shake 98 g of Kieselgel G (Merck No.7731) with 165 ml of distilled water for about 90 seconds until a creamy slurry of even consistency is obtained. Pour the slurry into the spreading device and coat the plates 1 mm thick without delay. Allow to stand for a few minutes until the layer has set, then transfer to an oven at 105-110 °C for 60-90 minutes. Store the activited plates in a desiccation chamber until required.

• Recovery of DLTDP from the Extractant Samples (a) Extractant - distilled water

Transfer 80 ml of extractant solution to a 100 ml separating funnel and add 8 g of sodium chloride. Extract with 5 × 20 ml portions of diethyl ether, combine the extracts and dry with anhydrous sodium sulfate. Decant a portion of the ether into a 50 ml beaker and evaporate to dryness on a steam bath. Repeat this process until all the ether has been evaporated. Wash the sodium sulfate with several portions of ether, transfer the washings to the 50 ml beaker and again evaporate to dryness. Transfer the residue with small portions of ether to a 2 ml graduated flask and finally dilute to volume with ether.



It is essential that the solutions should be applied to the plate slowly, so that the solvent spreads as little as possible. A gentle current of air blowing on the spot helps with this. 49

Additive Migration from Plastics into Foods (b) Extractants — 5% w/v sodium carbonate solution and 5% w/v citric acid solution

Transfer 80 ml of extractant solution to a 100 ml separating funnel and extract with 5 × 20 ml portions of diethyl ether. Combine the extracts and continue as described for distilled water.

(c) Extractants - 50% w/v ethyl alcohol in water

Evaporate 80 ml of solution, in small portions to dryness, portion-wise, in a 100 ml beaker on a steam bath. Transfer the residue with small portions of ether to a 2 ml graduated flask and finally dilute to volume with ether.

(d) Extractant - Medicinal paraffin BP Levis

Weigh 40 g of sample into a 100 ml separating funnel and dilute with 60-80 °C petroleum spirit to approximately 100 ml. Pass this solution down a column (1 cm × 3.5 cm) of chromatographic grade silica gel which has been activated by heating at 140 °C for 1 hour.



Maintain a flow rate of 2-3 ml per minute by application of gentle suction. Follow with 50 ml of 60-80 °C petroleum spirit, then allow the column to suck dry.



Pass down the column, at the same flow rate, 30 ml of chloroform/methanol solvent mixture, followed by 20 ml of chloroform, collecting the eluent in a 100 ml beaker. Evaporate just to dryness on a steam bath. Transfer the residue with small portions of diethyl ether to a 2 ml graduated flask and finally dilute to volume with ether.

• Thin-layer Chromatographic Separation Extractants - (a), (b) and (c). For a sample containing DLTDP in the range 1 to 5 ppm, apply aliquots of 50 µl of the sample solution in a line 1.5 cm from the bottom of a prepared plate. In addition apply spots of 2, 4, 6, 8 and 10 µl (2, 4, 6, 8, 10 µg) of standard DLTDP solution. For the range 5 to 25 ppm of DLTDP in the sample apply aliquots of 10 µl of the sample solution. Extractant - (d). For a sample containing DLTDP in the range 1 to 5 ppm, apply aliquots of 100 µl of the sample solution and spots of 2, 4, 6, 8, and 10 µl of standard DLTDP solution as for extractants (a), (b) and (c). For the range 5 to 25 ppm DLTDP in the sample apply aliquots of 20 µl of the sample solution. 50

Determination of Antioxidants Having applied the sample spots stand the plate without further delay in a suitable tank starting line downwards, so that the lower edge dips into the solvent mixure. Allow the solvent (acetone:cyclohexane 15:85 v/v) to flow 14 cm to the score line, then remove the plate from the tank and allow the solvent to evaporate. Spray the plate lightly and evenly with alkaline potassium permanganate when the DLTDP traces appear as yellow spots on a pink to red background. The Rf of DLTDP is approximately 0.65. Compare the samples and standard spots without delay. The spot sizes can be compared to an accuracy of ±1.0 µg, representing ±0.5 ppm in the 1 to 5 ppm range. The lower limit of detection of DLTDP on a thin-layer chromatographic plate is 2 µg, it therefore follows that it is not possible to determine the DLTDP content of samples containing less than 1 ppm DLTDP by this method. This procedure when applied to synthetic solutions of up to 5 ppm DLTDP in the various BPF extractants gave recoveries of 80% to 100% DLTDP in the extraction liquid (Table 4.1).

Table 4.1 Recovery of DLTDP from the various BPF extractants by the ether extraction/thin-layer chromatographic procedure DLTDP content of extraction liquid (ppm) Extractant Added Found Distilled water 5.0 4.0 50% Ethyl alcohol:water 5.0 4.0-4.5 5% Sodium carbonate 5.0 4.0-4.5 5% Citric acid 4.0 3.0-4.0 1.0 1.0 2.5 2.0 Liquid paraffin 3.75 3.5 5.0 4.5-5.0 Reproduced from Author’s own files

(b) Infrared Spectroscopy DLTDP does not absorb in the UV region of the spectrum but does absorb strongly in the infrared region. The carbonyl absorption occurring at 572 nm is a useful 51

Additive Migration from Plastics into Foods wavelength for estimating DLTDP. Infrared spectroscopy cannot, of course, be applied directly to aqueous solutions, and this necessitates a preliminary extraction of the DLTDP from the aqueous extractants with a low-boiling immiscible organic solvent. As discussed earlier, infrared spectroscopy can be applied to the direct determination of some types of additives in the liquid paraffin or hexane extractant, although usually with poor sensitivity, e.g., the smallest amount of DLTDP which can be detected in these liquids is about 50 to 100 ppm. The solvent extraction step also provides a concentration factor in the analysis which is essential in order to achieve the necessary analytical sensitivity of down to 10 to 15 ppm of additive in the original extraction liquid. An additional advantage of the infrared technique is that by comparing the infrared spectrum of the solvent extracts of the extraction liquids with that of an authentic specimen of the DLTDP it is possible to detect whether any degradation has occurred, and also to derive information regarding the nature and degree of such additive degradation. In addition to DLTDP, polymers used for extractability tests might contain other additives which could interfere in an infrared spectroscopic method of analysis and these would be present together with DLTDP in the extraction liquids.

Method • Extractability Test (BPF) Contact the plastic film or sheet with 800 ml of each of the four extraction liquids. Use the quantities of plastic and the extraction test conditions as recommended in the second BPF Toxicity Report [1]. It is convenient to carry out the extraction tests in ground-glass stoppered tubes of 1 litre capacity. Include in the test, polymer-free blank tubes containing 800 ml of each extraction liquid. At the end of the extraction test remove the tubes from the heating bath, shake well and wipe clean. While still hot remove the stopper and remove the plastic. Wash any DLTDP containing solution from the surface of the plastic into the extraction tube with a jet of warm distilled water (not exceeding 50 ml) delivered from a glass wash-bottle. Extractants - distilled water, 5% sodium carbonate and 5% citric acid: carefully transfer the hot contents of the extraction tube into a 1 litre capacity liquid-liquid extractor. Wash the interior of the extraction tube with 25 ml hot water and two 25 ml portions of cold ether, and transfer the washings to the extractor. Proceed as described next. 52

Determination of Antioxidants Extractant - ethanol-water (1:1): it is not possible to extract the ethyl alcohol:water extractant directly with diethyl ether as the two phases are miscible upon mixing. The 700 ml ethyl alcohol:water extractant is distilled down to about 200 ml until the distillate no longer has an odour of ethyl alcohol. The alcohol-free extraction liquid is then extracted with diethyl ether as described previously. Carefully transfer the hot contents of the extraction tube into a 1 litre, three-necked round-bottomed flask. Wash the interior of the tube with 25 ml hot water and two 25 ml portions of cold ether and transfer the washings to the 1 litre flask. Connect a separatory funnel containing distilled water, and a horizontally clamped Liebeg condenser with a suitable adaptor to the flask and distill until the distillate has no odour of ethyl alcohol. Add water to the flask when necessary to keep up the volume to a minimum of 200 ml. Ensure that the flask contains between 700 and 800 ml of liquid at the end of the distillation. Transfer the hot distillation residue into a 1 litre capacity liquid-liquid extractor. Wash the interior of the flask and the condenser with 25 ml hot water and two 25 ml portions of cold ether and transfer the washings to the extractor. Proceed as described next. Extraction of DLTDP from the liquid paraffin extractant: DLTDP could not be determined in the liquid paraffin extractant in amounts of less than 100 ppm by direct infrared spectroscopy, nor was it possible to devise a preliminary solvent extraction procedure to obtain a DLTDP concentrate as was achieved in the case of the aqueous extractants. The results obtained in trial extractions of liquid paraffin with various solvents (Table 4.2) indicate that neat or anhydrous methanol (10%), ethanol or

Table 4.2 Extraction of DLTDP from liquid paraffin extractant with low-boiling solvents Volume of liquid paraffin used (ml)

700

Volume of cyclohexane used to dilute liquid paraffin (ml)

300

Original DLTDP content of liquid paraffin (ppm)

Solvent extraction cycle used

Recovery of DLTDP in solvent extract (%)

% of original liquid paraffin present in solvent extract

100

20 hour extraction with anhydrous methanol in liquid-liquid extractor

30

10

53

Additive Migration from Plastics into Foods

Table 4.2 Continued

700

100-200

100-200

100

100

100

100-200

300

0-100

0-100

0

0

0*

0-100

100

10-35 hours extraction with anhydrous ethanol in liquid-liquid extractor

30-35

15-30

200-400

Four 20 ml extractions with anhydrous methanol

<10

2

200-400

Four 20 ml extractions with 9:1 (v/v) methanol: water

<10

2

400

Four 20 ml extractions with anhydrous ethanol

20

2

400

Four 20 ml extractions with 9:1 (v/v) ethanol:water

<10

2

400

Four 20 ml extractions with anhydrous acetone*

20

10

200-400

Four 20 ml extractions with 9:1 (v/v) acetone: water*

<10

2

*Liquid paraffin and acetone are completely miscible in presence of cyclohexane.

Reproduced from Author’s own files

54

Determination of Antioxidants acetone do not extract more than 35% of the DLTDP present in the liquid paraffin extractant, even when the extraction is carried out for extended periods of up to 35 hours using a liquid-liquid extractor. Moreover, from the results in Table 4.2 it can be seen that due to the slight miscibility of liquid paraffin with these various solvents, the extract would contain appreciable amounts of liquid paraffin which would reduce the concentration factor achieved in the concentration stage and might interfere in a subsequent analytical procedure. Consequently a solvent extraction procedure was not available for obtaining a DLTDP concentrate from the liquid paraffin extractant.

• Reagents Diethyl ether - shake 4 litres diethyl ether (Analar) with 300 ml of 30% aqueous sodium hydroxide in a 5 litre separating funnel. Run off the lower aqueous phase and reject. Wash the ether phase four times with 300 ml distilled water. Finally, distil the ether from 20 g solid sodium hydroxide into an amber glass bottle. Sodium sulfate, anhydrous

• Apparatus Liquid-liquid extractor, upward displacement type, 1 litre capacity, comprising distributor, extractor tube, boiling flask and condenser. Round-bottomed, 3-necked, 1 litre flask, with separatory funnel and connected to horizontally mounted Liebig condenser.

Procedure To the contents of the extractor containing 800 to 900 ml aqueous polymer extractant add 100 g solid sodium chloride. Stir with a glass rod to dissolve the salt as completely as possible. Charge the extractor with ether and extract for a total period of fifteen hours. To the flask containing the ether extract, add 3 g of powdered anhydrous sodium sulfate powder, and shake to dry the solvent phase. Filter this solution, in portions, through several layers of Whatman No. 3 filter paper into a 100 ml beaker standing in a warm water bath. Wash the interior of the flask containing the sodium sulfate with several 25 ml portions of fresh ether and transfer these via the filter paper to the 100 ml beaker, ensuring that the whole surface of the filter paper is washed with ether. Immerse the beaker in a water bath until the volume of ether is reduced to approximately 10 ml. Ensure that no droplets of water or undissolved organic matter remain in the ether at this stage. Carefully transfer this solution to a 55

Additive Migration from Plastics into Foods 25 ml volumetric flask. Wash the whole interior of the beaker with several portions of fresh ether and transfer these to the volumetric flask to dilute the volume up to 25 ml (at 25 °C). As soon as possible transfer to three 25 ml beakers, 6 ml and two 8 ml portions, respectively, of the ether extract by means of a dry 10 ml graduated pipette. Apply a nitrogen line to remove ether completely from the beakers. Use the two beakers containing 8 ml of ether extract for a duplicate determination of sulfur by the Schoniger method described in Method (c). Use the beaker containing 6 ml of ether extract for the determination of DLTDP by the infrared spectroscopic method described in Method (b). If there is any delay in continuing the analysis, cover the beakers with aluminium foil. Carry out IR analysis as follows:

• Reagents Carbon tetrachloride, spectroscopic grade.

• Apparatus Infrared spectrometer, double beam. Cells, rock salt, 1 mm path length. Volumetric glassware, 10 ml flask. Into the beaker containing the ether free extract of the extraction liquid, (see previously), pour approximately 5 ml spectroscopic grade carbon tetrachloride and swirl to completely dissolve the solid. Transfer this solution quantitatively to a dry 10 ml volumetric flask. Wash the walls of the beaker with further small portions of carbon tetrachloride, transfer to the volumetric flask to dilute the volume to 10 ml and shake the flask well. Transfer a portion of the carbon tetrachloride solution to a 1 mm path length, rock salt cell, and record its infrared spectrum in the range 500 to 600 nm. Construct a base line to the DLTDP peak at 575 nm by drawing a straight line between the absorption minima at 560 and 595 nm and measure the distances Io and I in millimetres as shown in Figure 4.1. Calculate the DLTDP content of the original extraction liquid as described next. Calibrate the procedure as described next. Calculate the absorbance (A575) of the 575 nm DLTDP peak as follows (see Figure 4.1): A575 = Log10 Io /I

56

Determination of Antioxidants

Figure 4.1 Infrared spectrum of a solution of DLTDP in carbon tetrachloride showing the measurements required for determining the absorbance at the 575 nm carbonyl peak. (Reproduced from Author’s own files)

The DLTDP content of the original 800 ml of extraction liquid in ppm w/v is then obtained by reference to the calibration graph of the absorbance of the 575 nm peak versus ppm DLTDP in 800 ml extraction liquid, prepared as described next.

• Calibration of the infrared procedure Use 6 ml/25 ml portions of the twenty ether extracts for calibration on the infrared procedure. Dilute each of the extracts (ether removed) up to 10 ml with spectroscopic grade carbon tetrachloride, and measure the absorbance (in a 1 mm path length rock salt cell) of the carbonyl peak occurring at 575 nm, as described previously. Prepare a calibration graph for each of the four extraction liquids by plotting determined absorbance (A) versus ppm DLTDP added to 800 ml of original extraction liquid (i.e., between 0 and 94 ppm w/v). Use this calibration graph to determine the DLTDP content of plastic extraction liquids of unknown composition. 57

Additive Migration from Plastics into Foods

(c) Oxygen Flash Combustion This technique analyses the diethyl ether extracts of the extraction liquids prepared as described under method (b).

• Reagents Oxygen cylinder Sodium hydroxide 0.01 N - aqueous. Prepare daily by dilution of a standard 0.1 N stock solution. Ensure that the 0.01 N sodium hydroxide solution is protected from contact with atmospheric carbon dioxide.

• Tashiro Mixed Indicator Dissolve 0.125 g methyl red in 100 ml absolute alcohol. Dissolve 0.083 g methylene blue (BP grade) in 100 ml absolute alcohol. Use equal volumes of each indicator for each titration. Hydrogen peroxide - 100 volume. Microanalytical reagent grade quality. Percolated water is prepared by percolating distilled water through a mixed resin bed containing Amberlite IR-120(H) and Amberlite IRA-400(OH). This water must be boiled immediately before use and should be used throughout during the Schoniger combustion stage of the analysis.

Apparatus Schoniger flasks – 500 ml with B24 neck. Platinum wire supports 10 cm long and fused into B24 stoppers. The free end of the platinum wire is either shaped into an acute S bend or fitted with a 2.5 cm square of platinum gauze. Absorbent ash-free filter paper cut into 2.5 cm squares. Graduated burettes – 5 ml (Grade A - volumetric glassware) Pipette 1 ml of a freshly prepared 1% solution of DLTDP (i.e., 0.01 g) into a 100 ml beaker, apply this solution to the paper and carry out the combustion as described next. Add to a 500 ml Schoniger combustion flask, 1 ml of 100 volume hydrogen peroxide and 10 ml of recently boiled-out distilled water. Add 0.1 ml each of methyl red and 58

Determination of Antioxidants of methylene blue indicators. Titrate the solution with 0.01 N sodium hydroxide solution until it becomes a clear green colour. Fill the flask with pure oxygen and close immediately with a B24 stopper. To the beaker containing DLTDP add 1 to 2 ml of ether. Into this solution dip a 2.5 cm square of filter paper held by a pair of tweezers. Allow the paper to soak up all the ether solution. Gentle air blowing will evaporate ether from the paper once it has become saturated. Rinse down the beaker walls with 1 to 2 ml of fresh ether and similarly apply this solution by means of a small glass dropping pipette to the filter paper. Wash the beaker again with 1 to 2 ml ether, and apply the solution to the filter paper. Allow the filter paper to dry, and fold twice by means of two pairs of tweezers. Insert the paper between the platinum gauze attached to the stopper of the Schoniger flask. Into a fold of the paper insert a 5 cm × 0.3 cm strip of filter paper to serve as a wick. Light the top of the wick and quickly insert into the oxygen-filled combustion flask. Invert the flask during the combustion (during combustion the operator should be screened from the flask). Set the flask aside until the mist has cleared (10 to 15 minutes). Shake the flask occasionally during this period. Open the combustion flask and wash down the stopper, wire and neck with recently boiled-out water. Boil the solution for one minute. Quickly connect a Carbosorb-filled B24 trap into the mouth of the flask and cool rapidly. Titrate-immediately with 0.01 N sodium hydroxide to the clear green end-point. Carry out duplicate blank combustions exactly as described above, except that no sample is placed on the filter paper. Calculate the DLTDP content of the original extraction liquid as described next. Calibrate the procedure as described next. Calculate the DLTDP content of the extraction liquid as follows: DLTP (ppm; w /v) extraction liquid =

(T

A

− TB ) × f × 514 × 106 2000 × V

Where: ta = Titration (ml) of sodium hydroxide after Schoniger combustion of sample extract.

tb = Titration (ml) of sodium hydroxide after Schoniger combustion of reagent blank.



V = Volume (ml) of original extraction liquid represented by portion of ether extract taken for Schoniger combustion.



f = Normality of sodium hydroxide solution.

The method is calibrated against synthetic solutions of DLTDP in the various extraction liquids which have been heated for 10 days at 60 °C and then analysed in exactly the 59

Additive Migration from Plastics into Foods same manner as that used for the extraction liquids obtained in extractability tests carried out on plastics. Weigh out accurately 0.75 g pure DLTDP into a 100 ml beaker and dissolve in warm absolute ethyl alcohol. Transfer to a 100 ml volumetric flask together with beaker washings and make up to the 100 ml mark with alcohol and mix well. Into five, 1 litre ground-glass stoppered extraction tubes, measure 800 ml of distilled water. Into three further sets of five tubes measure 800 ml of the ethanol:water extractant (1:1), 800 ml of the 5% sodium carbonate extractant and 800 ml of the 5% citric acid extractant. Into each of the four sets of five tubes, accurately pipette 0.0, 1.5, 3.5, 7.0 and 10.0 ml of 0.75% DLTDP in ethyl alcohol, stopper and mix well. Leave the tubes at 60 °C for 10 days. For the five ethyl alcohol:water extractants (1:1) remove alcohol by distillation and transfer the hot alcohol-free distillation residue to the liquid-liquid extractor as described previously. For the five distilled water, 5% sodium carbonate and 5% citric acid extractants, transfer the hot extractants to the liquid-liquid extractor as described previously. Carry out the ether extraction of the four sets of extractants and split each of the twenty 25 ml ether extracts into separate portions as described next. Use these solutions to calibrate the Schoniger procedure and the infrared procedures as described next. Determine total sulfur (in duplicate) in 8 ml/25 ml portions of the twenty ether extracts produced by the Schoniger procedure as described previously. Calculate the weight of the DLTDP equivalent to the determined sulfur content. Prepare a calibration graph for each of the four extraction liquids by plotting ppm DLTDP added to the 800 ml of original extraction liquids (i.e., between 0 and 94 ppm w/v) versus ppm DLTDP determined in the 800 ml of the original extraction liquid by Schoniger combustion. Use this calibration graph to determine the sulfur content (i.e., the DLTDP content) of the plastic extraction liquids of unknown composition.

Comparison of Infrared and Oxygen Flask Results Using the previous procedures, extractability tests have been carried out using the distilled water and the 50% ethyl alcohol:water extractants on 0.03 mm thick polypropylene film containing 0.25% w/v, DLTDP. In these tests the film (31 to 32 g) was contacted (0.23% w/v) with 700 ml of each extractant for 10 days at 60 °C. Each extraction experiment was carried out in duplicate. At the end of the extraction test, DLTDP was determined in the extractants by the oxygen combustion and the infrared procedures. The results (see Table 4.3) show that duplicate extraction tests carried out by either method and under identical conditions, gave results which differed by factors of up to 10. The differences obtained in duplicate extractability tests (Table 4.4) can be ascribed to a fault in the technique used. In these experiments, the extraction liquid was decanted 60

Determination of Antioxidants

Table 4.3 Extractability of DLTDP from polypropylene - duplicate extraction tests DLTDP extracted (g/100 ml film)*

Extractant

By oxygen combustion method

By infrared method

Distilled water

0.01

0.01

Extraction 1

<0.01

<0.01

<0.01

Extraction 2

0.02

0.03

0.03

Extraction 1

<0.01

<0.01

0.01

Extraction 2

0.07

0.07

0.07

50% ethyl alcohol:water

*Calculated according to the method prescribed by the British Plastics Federation [1] for plastic samples less than 0.5 mm thick. Reproduced from Author’s own files

Table 4.4 Recovery of DLTDP from 700 ml aqueous BPF extractants by ether extraction Recovery of DLTDP by ether extraction (%) DLTDP content of extractant (ppm)

Determined as sulfur

Determined by infrared spectroscopy at 572 nm

16

80, 83

88

75

75, 76

82

50% ethyl alcohol water

16

56, 65

69

75

87

91

5% sodium carbonate

16

73, 82

85

75

89, 90

92

16

83, 65

66

75

39, 49

69

Extractant

Distilled water

5% citric acid

Reproduced from Author’s own files

61

Additive Migration from Plastics into Foods from the tube after the 10 day heating period, leaving the film in the tube. The tube and its contents was then washed several times with hot water and then diethyl ether to ensure quantitative recovery of DLTDP. However, separate experiments showed that cold ether can extract considerable quantities of DLTDP from thin films of polypropylene after a few minutes contact, and it was this which was responsible for the variable extraction test results reported in Table 4.4. The cold ether polymer washing stage was therefore omitted in the further experiments described next. The polypropylene used in these determinations contained 0.25% w/v DLTDP, i.e., approximately 0.23% w/v. This may be compared with the maximum extractability figure obtained for DLTDP from this film, namely 0.01 g DLTDP extracted from 100 ml film, i.e., 0.01% w/v. It is concluded that a maximum of only 5% of the original additive content of film was extracted into the distilled water extractant during 10 days at 60 °C. DLTDP does not therefore tend to migrate to any appreciable extent from polypropylene into this extraction liquid. The results in Table 4.3 show that the migration of the additive into the 50% ethyl alcohol:water extractant is of a similar low order of magnitude to its migration into distilled water. These concentrations indicate the care with which etractability studies must be conducted. Synthetic solutions of DLTDP were prepared in the four aqueous and the alcoholic extractants. The DLTDP was added to 700 ml of each of the extractants at the 16 and the 75 ppm level as a solution of 5 ml of methyl alcohol. These solutions were then heated for 10 days at 60 °C in order to simulate the conditions occurring in an actual polymer extraction test. At the end of this period, DLTDP was extracted from the extractants with diethyl ether prior to analysis by the infrared procedure. Accurately pipetted portions of the ether extract were transferred to small, dry beakers and evaporated to dryness with a nitrogen stream. The residue was diluted to volume with carbon tetrachloride in a 10 ml volumetric flask. Figures 4.2 (a) and (b) show the infrared spectra in the 2 to 15 µm region of: (a) Synthetic solution of DLTDP in carbon tetrachloride (b) DLTDP found in an ether extract of distilled water, 50% w/v ethyl alcohol:water, 5% sodium carbonate and 5% citric acid extractants after 10 days at 60 °C. Comparison of the infrared spectrum of DLTDP (Figure 4.2(a) with those of the ether extracts of the BPF extractants (Figure 4.1(b)) shows that the infrared spectrum is virtually unchanged by heating DLTDP for 10 days at 60 °C in the various BPF extractants. In Table 4.3 are shown DLTDP recoveries obtained by the infrared and the oxygen flask methods for synthetic solutions of 16 ppm and 75 ppm DLTDP in the four 62

Determination of Antioxidants

Figure 4.2 Recoveries of DLTDP in carbon tetrachloride extracts of (a) distilled water, (b) ethanol: water (1:1), British Plastics Federation extraction liquids after heating extractants for 10 d at 60 °C. (Reproduced from Author’s own files)

aqueous extractants. These results show that by either method of analysis with the exception of the citric acid extractant, DLTDP recoveries usually exceed 70% to 80% of the amount added to the original extraction liquid before the heating period for 10 days at 60 °C. As seen in Table 4.4, DLTDP recoveries of up to 40% lower than expected are obtained by the ether-extraction/infrared method when the following four extractants have been heated for 10 days at 60 °C: distilled water, ethanol:water (1:1), 5% sodium carbonate, and 5% citric acid. These low recoveries could be due either to (a) incomplete recovery of the DLTDP during ether extraction, or (b) to the occurrence of partial degradation of the DLTDP to another compound brought about by contact with the extraction liquid during the 10 day heating period at 60 °C. It was noticed in the case of these three extractants that a white solid was present which was insoluble in the aqueous phase and was not removed by prolonged extraction with ether. Evidently, this solid was not DLTDP, which is very ether-soluble, but degraded DLTDP produced during the heating period (the original DLTDP used in experiments is completely soluble in ether). 63

Additive Migration from Plastics into Foods In view of the evidence concerning the possible degradation of DLTDP during the analysis, it is desirable to apply the same treatment to the standard DLTDP calibration solutions as applied during the analysis of plastic extraction liquids of unknown DLTDP. This would cancel out errors in the analysis caused by partial degradation of DLTDP. To calibrate the procedure, various standard concentrations of DLTDP were made up in each of the extractants, which were then heated for 10 days at 60 °C. The aqueous extractants were then ether-extracted and the extracts used to calibrate the Schoniger combustion and the infrared procedures.

4.3 Irganox 1076 (Octadecyl-3-(3,5-di-t-butyl-4-hydroxy phenyl propionate) in Non-aqueous Extractants Helmroth and co-workers [2] in the course of studying the effect of solvent adsorption on additive migration from low-density polyethylene described a gas chromatographic (GC) method for the determination of this antioxidant in the solvents used in the extraction studies. These include cyclohexane, ethanol, ethyl acetate, isooctane, isopropanol, olive oil, tributyrin and tricaprylin. The concentration of Irganox 1076 in all low molecular weight solvents was analysed by GC with flame ionisation detection (FID) equipped with an on-column injector. The column used was a 15 m × 0.25 mm id DB5-MS with a film thickness of 0.1 µm (J & W Scientific). The analytical column was connected to a 0.5 m × 0.53 mm id retention gap, which was deactivated with a thin film of OV-1701-OH (BGB Analytik). Carrier gas was helium at a constant flow rate of 1.8 ml/min. Samples of 1 µl were injected on-column into the retention gap by an autosampler. The temperature programme of the GC oven was 1 °C under the boiling point of the solvent during injection and held for 1 minute after injection. The temperature was then increased to 150 °C at 15 °C/min followed by an increase at 10 °C/min to 310 °C, at which it was held for 1 minute. The FID detector temperature was kept at 315 °C. The determination of Irganox 1076 in triglyceride samples was based on a method using high-performance liquid chromatography (HPLC) developed by O’Brien and co-workers [3] in which olive oil is diluted with acetone (1:4) before analysis. Analyses were performed by reversed-phase HPLC using an Waters Alliance 2690 system with LC pump, injector and thermostatted LC oven and a Waters 474 fluorescence detector (Waters, Milford, USA). Separation was achieved on a 150 × 4.6 mm id XTerra RP18 column (particle size: 5 µm; Waters). The LC column was kept at 40 °C and the samples in the autosampler at room temperature. Detection was achieved by fluorescence detection (excitation wavelength 282 nm; emission wavelength 308 nm). A 10 µl aliquot of the triglyceride - acetone mixture was injected. The mobile phase was initially a linear concentration gradient of water:acetonitrile (2:8 v/v) 64

Determination of Antioxidants to water:acetonitrile (1:19 v/v). After 5 minutes, the mobile phase was changed to acetonitrile (100%) and held for 5 minutes. Next, the mobile phase was changed to tetrahydrofuran (100%) to clean the column. Finally, the column was conditioned by water:acetonitrile (2:8 v/v) for 5 minutes before the next injection.

4.4 Ionox 330 (1,3,5-trimethyl-2,4,6-tris(3,5 di tert-butyl-4hydroxybenzyl)benzene) Antioxidant in Fatty Extractants The direct quantitative determination of additives migrating from plastics into edible oils or fatty foodstuffs (Waggon and Uhde [4]) is extremely difficult even with the most sophisticated analytical techniques, and in some cases it may be impossible. Analytical procedures in which the migrated additives are concentrated or isolated from the fatty foodstuffs before their determination are tedious. Moreover they are inaccurate because of losses of material and they may even fail completely. Therefore, in order to determine the migrated additives quantitatively in an analytically suitable simulant, it is essential to simulate the natural migration into foodstuffs in appropriate model tests. The fat simulents that have been studied range from simple organic solvents such as n-heptane [5], diethyl ether [6], isopropanol [1], ethanol [1], isooctane [1], ethyl acetate [1], liquid paraffin [6] and ethyl acetate [1]. Alternatively, edible oils such as olive oil [2], coconut oil [7], tributyrin [2] and tricaprylin [2] have been used. Detailed data on the extraction capacities of these fat simulents in comparison with actual migration of additives from plastics into fatty food have been elaborated by Figge and others [8-22]. Figge proposed the use of radioactive 14C-labelled Ionox 330 [8, 22-24] which he incorporated in known amounts (0.5-2%) in polyvinylchloride (PVC), polystyrene (PS) and high-density polyethylene (HDPE) and low-density polyethylene (LDPE). Migration tests were carried out for 5 hours at 60 °C or 30-60 days at 20 °C. As was to be expected, the amounts of Ionox 330 migrating from different plastics into any one edible oil differed widely. The amounts of Ionox 330 migrating into olive oil from PVC, LDPE, HDPE and PS, for example, were in the proportions 1:14:160:2000. The migration of different additives from the same plastics into a given edible oil also varied widely. Even under extraction conditions (i.e., at a higher test temperature), the interactions between different edible oils and a given plastic material were similar. Further investigations showed this conformity to apply also to fats with a higher amount of medium-chain fatty acids, such as butter fat or coconut oil. Thus edible oils behave so similarly that any one of them could serve as a standard fat simulant for all other oils and fats. However, for laboratory work, substances that can more easily be analysed are preferred, provided they are also appropriate fat simulants under migration and extraction conditions. 65

Additive Migration from Plastics into Foods Curves obtained for the migration of Ionox 330 from HDPE and PS into various oily extractants are illustrated in Figures 4.3 and 4.4. The synthetic triglyceride tricaprylin is an excellent fat simulent [25] for the assessment of HDPE packages (Figure 4.5). The course of the migration with time of Ionox 330 from this polymer into tricaprylin was in good agreement with that into sunflower seed oil. This conformity was also observed in tests conducted at 65 °C. The migration and extraction values of Ionox 330 into PS into tricaprylin was unrealistically high (Figure 4.4) showing that tricaprylin cannot be recommended as a standard fat simulent for all types of plastics. These and other considerations lead Figge to investigate the possibility of developing a single standard fatty simulent liquid for use in extraction tests [2]. Figge [8] is of the opinion that a fat simulent which he designated as HB307 presents an appropriate generally applicable simulent for pure edible oils and fatty foodstuffs. Figge and co-workers [13, 15, 17] have described the extraction test apparatus in detail. For scintillation counting of the concentration of extracted labelled additive in the fatty foodstuff simulant Figge used a scintillation spectrometer employing a mixture of 4 g of 2,5-diphenyloxazol and 0.3 g of 1,4 bis-2-(4-methyl-5-phenyloxazolyl)-benzol in 1 litre of Merck 8325 toluene as the scintillation liquid. The contact area between plastic film and simulant liquid was varied between 28.25 cm2 and 67.93 cm2. In any particular series of tests the contact area of film and the volume of the extraction cell was kept constant. Figge and Piater [14, 20] developed formulae for the calculation of the corrected time dependent migration rates (%) from their radioactivity measurements on the extractant after specified time intervals. A typical extraction curve obtained for olive oil extractant and polystyrene is illustrated in Figure 4.6. It can be seen from Table 4.5 that the amount migrating from different test films into HB 307, exceed the values for reference fats by a maximum factor of 1.8. Thus, in contrast to the organic solvents, this synthetic standard triglyceride mixture represents a good and generally acceptable simulant for pure edible fats and fatty foodstuffs. Figge and co-workers also carried out detailed studies on the influence of temperature and time on additive migration [27-31]. A typical temperature – migration curve is shown in Figure 4.6. Although much of the published work in the determination of extracted polymer additives in the synthetic triglyceride fat simulant HB 307 has been based on radiochemical methods, using labelled additives, non-radiochemical methods have been described. Koch and other workers [32] have described methods, which are based on alternative techniques such as visible and UV spectroscopy and GC for the determination of antioxidants in the fat simulant HB 307. 66

Determination of Antioxidants

Figure 4.3 Migration of antioxidant Ionox 330 from high-density polyethylene into different contact liquids as a function of time. Tests were carried out at 20 °C for 60 days (- - - -) and at 65 °C for 5 hours (——) using sunflower seed oil (SF), tricaprylin (TC) and paraffin oil (Par) and n-heptane. (Reproduced with permission from K. Figge, Food and Cosmetics Toxicology, 1972, 10, 6, 815 [8])

Figure 4.4 Migration of antioxidant Ionox 330 from polystyrene into different contact liquids as a function of time. Tests were carried out at 20 °C for 60 days (- - - -) and at 65 °C for 5 hours (——) using sunflower seed oil (SF), tricaprylin (TC) and n-heptane. (Reproduced with permission from K. Figge, Food and Cosmetics Toxicology, 1972, 10, 6, 815 [8]) 67

Additive Migration from Plastics into Foods

Figure 4.5 Migration of labelled additives (Ionox 330 o-o) from high-density polyethylene film, Ionox 330 ◆ from polystyrene film. (Reproduced with permission from Figge and Koch, Food and Cosmetic Technology, 1973, 11, 4, 975 [27])

4.5 Miscellaneous Antioxidants Koch [33] has also described visible colorimetric methods for the determination of phenolic antioxidants in HB 307 in the concentration range around 1 ppm. A method (Koch [33] utilising diphenyl-picrylhydrazyl as the chromogenic reagent has been applied to the determination of compounds such as 4,4′-butylidene-bis(3-methyl-6-tert-butylphenol-1) and 2,2′-methylene-bis-(4-ethylene-6-tert-butyl phenol-1). In this method, 2 ml of an ethanolic solution (0.05 to 1 mg per 100 ml) of an antioxidant was added to 5 ml of a 50% solution of HB 307 in chloroform or 1,2-dichloroethane in a stoppered test-tube. Then 3 ml of 0.0075% ethanolic 2,2-diphenyl-1-picrylhydrazyl was added, and the stoppered tube was heated for 30 or 60 minutes (according to the antioxidant) in a water bath at 60 °C. 68

Determination of Antioxidants

Figure 4.6 Extraction curves, Ionox 330, from polyethylene and polystyrene into: (a) olive oil, (b) olive oil containing 2% free fatty acids at 65 °C. (Reproduced with permission from Figge and Piater, Empfehlungen der Kunststoffe-Kommission des Bundesgusundheitsamtes (BGA) I Mitteilung, Bundesgesundheitsblatt, 1967, 189 and 10 [26])

Table 4.5 Comparison of the amounts of Ionox 330 migrating from different test films into edible fats and the fat simulent HB 307 during one-sided contact for 60 days at 20 °C Identity and concentration (% w/v) of labelled additive

Test film

Proportion (%) of radioactivity or additive migrating into

Ratio R† for

Biskin

Coconut oil

Butter

HB 307

Biskin

Coconut oil

Butter

HDPE

Ionox 330 – [14C]‡ (1.0)

0.090

0.098

0.120

0.140

1.6

1.4

1.2

PS

Ionox 330 – [14C]‡ (2.0)

2.08

2.53

3.07

3.05

1.5

1.2

1.0

R: =

†

amount of additive migrating into fat simuleent HB 307 nto Biskin, coconut oil or butter amounts for additive migrating in

: 1,3,5-trimethyl-2,4,6-tris-(3,5-tert-butyl-4-hydroxybenzyl) [14C] benzene

‡

Reproduced from Author’s own files

69

Additive Migration from Plastics into Foods The tube was cooled in water at 20 °C, and, as soon as possible, the extinction of the solution was measured at 517 nm against a similarly prepared solution of a sample without antioxidant. The order of extinction measurement was the same as that of sample preparation, to give a standardised reaction time. Air was excluded in all operations, and all solvents were saturated with nitrogen. The extinctions obeyed Beer’s law for up to 4 to 16 ppm of antioxidant in the HB 307, depending on the compound. Down to 1 ppm of many compounds could be determined, but the applicability of this method is limited by inadequate activation by substituents or by high steric hindrance of the phenolic group. Uhde and co-workers [34, 35] have studied the migration of 4,4′-thiobis-6-tert-butyl-mcresol (Santonex R) from plastics utensils into sunflower seed oil. Sunflower seed oil that had been stored in vessels of polyethylene containing this antioxidant was diluted (3:5) with pentane and extracted with acetonitrile containing 5% of water. The concentrated acetonitrile extract (or an ethanol solution of the residue on evaporation) was subjected to thin-layer chromatography on Kieselgel with hexane-ethyl acetate (10:3) as solvent. To detect the antioxidant (down to 0.1 ppm) the plate was sprayed with 3,5-dichlorop-benzoquinonechlorimine solution. To determine the antioxidant, the zone at Rf = 0.44 (located by means of iodine vapour) was removed and treated with fuming nitric acid:sulfuric acid (1:1). The nitro-derivative of the antioxidant was determined in the product by polarography after the addition of urea and sodium acetate [35]. Alternatively, an aliquot of the original concentrated acetonitrile extract was treated with ethanol and diazotised sulfanilic acid, and the extinction of the resulting dye, measured at 480 nm. This was compared with that for a solution containing 10 µg of antioxidant in acetonitrile, treated similarly. Uhde and Woggon [30-38] also studied the migration of 2,6-di-tert-butyl-p-cresol from PS, impact resistant polystyrene and polypropylene utensils. Samples of the utensils (10 cm × 10 cm × 1 mm) containing less than 0.5% of antioxidant were immersed for 10 days at 45 °C in water, 3% acetic acid, 15% or 50% aqueous ethanol, heptane, or sunflower seed oil. The migration of the antioxidants from the plastics into the liquids was followed by the use of spectrophotometric or polarographic methods. The tests showed that there was little tendency for migration of the various antioxidants from the PS into the aqueous alcoholic or fatty liquids but there was considerable migration of 2,6-di-t-butyl-p-cresol from polypropylene into sunflower seed oil, and migration values were high whenever heptane was used. Alternatively, the oil sample was dissolved in isopropyl ether and the extinction was measured at 340, 343 or 353 nm against a similar solution of the original untreated oil and referred to a calibration graph prepared with 1 to 50 µg of the UV absorber added to 1 g of oil [38]. 70

Determination of Antioxidants Sampaolo and co-workers [39] determined the specific migration of 2,2′ methylene bis-(6-tert-butyl-4-methylphenol) from rubber goods into ailimentary fats. They determined the antioxidant by the methods of Hilton [40] and Wadelin [41] after extraction from styrene-butadiene, natural, butyl and nitrile rubbers into triolein or coconut oil by immersing discs of the rubber into the fat for 10 days at 40 °C. The results obtained from the two methods agreed well. The concentration of antioxidant in the extracts ranged from approximately 18 to 63 ppm, the largest amounts being extracted from natural rubber and the smallest amounts from nitrile rubber. Similar results were obtained for the migration of the antioxidant into heptane.

References 1.

Plastics in Contact with Food: A Guide, 2nd Edition, British Plastics Federation, London, UK, 1996.

2.

I.E. Helmroth, M. Dekker and T. Hankemeir, Food Additives and Contaminants, 2002, 19, 2, 176.

3.

A.R. O’Brien, I. Cooper and P.A. Tice, Food Additives and Contaminants, 1997, 14, 6-7, 705.

4.

H. Woggon and W-J. Uhde, Ernahrungforschung, 1971, 16, 227.

5.

US Food and Drug Administration, Code of Federal Regulations, Chapter 1, Part 121, US Government Printing Office, Washington, DC, USA, 1967.

6.

Italian Ministry of Health, Health Regulations for Packages, Wrappings, Containers and Utensils Intended for Contact with Foodstuffs or with Substances for Personal Use, Gazetta Ufficiale, 1963, 64, 18.

7.

R. Franck, Kunststoffe im Lebensmittelverkehr, Empfehlungen der Kunststoffe-Kommission des Bundesgesundheitsamtes, 1967, Carl Heymanns Verlag, Cologne, Germany, Teil B, 7th issue, p.6.

8.

K. Figge, Food and Cosmetics Toxicology, 1972, 10, 6, 815.

9.

K. Figge in Proceedings of Aus den Arbeit von Chemischen Forschungslaborotorien, Hamburg, Germany, 1972.

10. K. Figge, Angewandte Chemie, 1971, 83, 22, 901. 11. K. Figge, Kunststoffe, 1971, 61, 11, 832. 71

Additive Migration from Plastics into Foods 12. K. Figge, S.R. Eder and H. Piater, Deutsche Lebensmittel-Rundschau, 1972, 68, 359. 13. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1971, 67, 9. 14. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1971, 67, 47. 15. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1971, 67, 110. 16. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1971, 67, 154. 17. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1971, 67, 235. 18. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1972, 68, 313. 19. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1970, 66, 281. 20. H. Piater and K. Figge, Migration von Hilfsstoffen der Kunststoffverarbeitung aus Folien in Flüssige und feste Fette bzw. Simulantien. VIII. Mitteilung: Vergleich der Gravimetrisch bestimmen Rückstände der Extraktionslosungen mit den Tatsächlich Extrahierten Additivmengen, 1971. 21. C.G. vom Bruck, K. Figge, H. Piater and V. Wolf, Deutsche LebensmittelRundschau, 1971, 67, 444. 22. C.G. vom Bruck, K. Figge, H. Piater and V. Wolf, Deutsche LebensmittelRundschau, 1970, 66, 253. 23. K. Figge in Proceedings of Praparative Radiochemie, Gesellschaft Deutscher Chemiker (GDCh), Lindau/Bodensee, 1968. 24. K. Figge, Journal of Labelled Compounds, 1969, 5, 2, 122. 25. Ministre de la Santé Publique. Circulation du 12 September 1963 relative aux demandes d’autorisation d’emploi de substance chimiques destinées à être introduites dans les aliments ou utilisées dans les matériaux mis aux contact des ailments, Journal Officiel, 26 September 1963. 26. Empfehlungen der Kunststoffe-Kommission des Bundesgusundheitsamtes (BGA) I Mitteilung, Bundesgesundheitsblatt, 1967, 189 and 10. 27. K. Figge and J. Koch, Food and Cosmetics Toxicology, 1973, 11, 4, 975. 28. R.F van der Heide, The Safety for Health of Plastics Food Packaging Materials Principles and Chemical Methods, Kemink En Zoon NV, Utrecht, 1964. 72

Determination of Antioxidants 29. H. Woggon, D. Jehle and W-J. Uhde, Nahrung, 1969, 13, 343. 30. H. Woggon, W-J. Uhde and G. Zydek, Zeitschrift für Lebensmittel untersuchung und -Forschung, 1968, 138, 169. 31. K. Figge and A. Zeman, Kunststoffe, 1973, 63, 8, 543. 32. J. Koch, Deutsche Lebensmittel-Rundschau, 1972, 68, 401. 33. J. Koch, Deutsche Lebensmittel-Rundschau, 1972, 68, 404. 34. W-J. Uhde, H. Woggon and U. Köhler, Nahrung, 1968, 12, 813. 35. U. Köhler, H. Woggon and W-J. Uhde, Plaste und Kautscuh, 1968, 15, 9, 630. 36. W-J. Uhde and H. Woggon, Deutsche Lebensmittel-Rundschau, 1971, 67, 257. 37. W-J. Uhde and H. Woggon, Nahrung, 1968, 12, 825. 38. W-J. Uhde and H. Woggon, Analytical Abstracts, 1970, 18, 2024. 39. A. Sampaolo, L. Rossi, R. Binetti, C. Cesolari and G. Fava, Raoss Chimo, 1972, 24, 3. 40. C.L. Hilton, Analytical Abstracts, 1960, 7, 4893. 41. O. Wadelin, Analytical Abstracts, 1957, 4, 982.

73

Additive Migration from Plastics into Foods

74

5

Determination of Ultraviolet Stabilisers in Extractants

5.1 Uvitex OB (2,5-bis (5′-tert butyl-2-benzoxazolyl)thiophene) in Aqueous Extractants The determination of Uvitex OB illustrates an example of the application of ultraviolet spectroscopy to the determination of additives in foodstuff simulent extraction liquids. Figures 5.1 and 5.2 show an ultraviolet spectrogram and a calibration graph, respectively, which demonstrates that the optical brightener Uvitex OB can be estimated in amounts down to 3 ppm in these extractants by direct spectroscopy at its absorbtion maximum which occurs at 378 nm. Table 5.1 gives the formulation of a polyethylene and a polystyrene used in some Uvitex OB extractability studies. In addition to the ultraviolet stabiliser, liquids obtained in extractability tests carried out on these plastics would contain various other substances, some of which are ultraviolet absorbers and which may be present in the extractant at higher concentrations than Uvitex OB. In this case the polymer also contained Santonox R, Wingstay T, phenolic antioxidant and styrene monomer. The presence of such ultraviolet absorbing substances in the extractant will interfere in the determination of Uvitex OB at 378 nm. In applying spectroscopic methods of analysis to extractants, consideration must always be given to the possibility of interference by any polymer additives present other than that which one is required to determine. Ultraviolet spectra were run on solutions of the various additives in the region of the Uvitex OB absorption maximum (378 nm) to ascertain to what extent these would interfere in the determination of Uvitex OB. Table 5.1 and Figure 5.3(a) show that for polyethylene, the level of Santonox R present would not seriously interfere in the determination of Uvitex OB, i.e., if the extractant contained twice as much Santonox R as Uvitex OB then the reported Uvitex OB analysis would be only approximately 10% higher than theoretical. A rather different picture emerged, however, in the case of the polystyrene sample which contains a wider range of additives than the polyethylene sample; these are present in the polymer at appreciably higher concentrations than the level of Uvitex OB. In fact, if all the additives present migrate from the polymer at the same rate as Uvitex OB, the latter could not be determined in the extractant by ultratiolet spectroscopy because at 378 nm the absorption due to other polymer additives would be about 30 times greater than that of the Uvitex OB alone. 75

Additive Migration from Plastics into Foods

Figure 5.1 Ultraviolet absorption and visible fluorescence spectrograms of Uvitex OB. (a) Ultraviolet absorption spectrum(240-400 nm) of Uvitex OB (53 ppm) in 50% of ethyl alcohol:distilled water extractant (1 cm cell) showing absorption maximum at 378 nm using tungsten lamp (slit width 0.04 mm). (b) Visible fluorescence spectrogram (380-500 nm) of Uvitex OB (28 ppm) in 50% ethyl alcohol:distilled water extractant (1 cm cell) showing maxima at 415 and 435 nm using light filter with transmission 380-540 nm and high pressure mercury vapour lamp (slit width 0.1 mm). (Reproduced from Author’s own files)

Extraction liquids can contain a mixture of extracted substances, and one method of applying spectroscopy in these cases is to apply a correction procedure. This can be demonstrated by considering the extractant from a polystyrene formulation containing, in addition to Uvitex OB, Wingstay, butyl stearate and mineral oil. Figures 5.3(b) to (d) are ultraviolet spectra in the 250 to 418 nm region of synthetic solutions in the 5% sodium carbonate extractant of Uvitex OB (53 ppm), and up to three times this concentration of the other three polymer additives. Each of these additives would seriously interfere in the determination of Uvitex OB by evaluation of its absorption maximum at 378 nm. In the correction procedures, measurements are made at not only at this maximum but also at the two Uvitex OB minima at 275 and 418 nm (Figure 5.3(a)) Using suitable calibration and calculation procedures it is then possible to calculate the corrected optical density at 378 nm due to Uvitex OB alone, provided that in the region 275-418 nm, background absorption due to any other substances present is low, fairly linear, and not too steep. 76

Determination of Ultraviolet Stabilisers in Extractants

Figure 5.2 Calibration Curve. Ultraviolet spectroscopic determination of Uvitex OB at 378 nm in aqueous extractants (British Plastics Federation). (Reproduced from Author’s own files)

The ultraviolet spectra in Figures 5.3 (b) to (d) show that, due to the presence of the Wingstay T maximum at 275 nm, the correction procedure would not be applicable to the determination of Uvitex OB in solutions which also contain a similar level of Wingstay T. However, it is quite feasible to determine Uvitex OB accurately in the presence of up to twice its concentration of butyl stearate or mineral oil, as the background due to these substances occurring between 275 and 418 nm is low and is sufficiently linear to permit application of the correction procedure. Thus, in some, but not all cases, this interference can be overcome by the use of a correction procedure. For additives which absorb weakly in the ultraviolet region, or for strongly absorbing additives which have to be determined at concentrations of much less than 5 ppm, it is necessary to prepare a concentrate from the extraction liquid by extracting up to one litre of it containing about 10% added sodium chloride with a suitable low boiling solvent such as diethyl ether, hexane, methylene dichloride or carbon disulfide. The extraction can be achieved in about 24 hours in an upward displacement or downward displacement liquid-liquid extractor. The 50% w/v ethyl alcohol:water extractant is distilled to about 20% of its original volume to remove ethyl alcohol, and the residue extracted with a low-boiling solvent as before. The hexane extract used in Food and 77

Additive Migration from Plastics into Foods

Table 5.1 Interference by other Polymer Additives in the Ultraviolet Spectroscopic Determination of Uvitex OB in Foodstuff Extractants (a) Additives present in polyethylene Santonox R Uvitex OB - 50 ppm Zinc sulfide - 300 ppm 100 ppm Concentration Optical density in extraction at 378 nm liquid(a), ppm (1 cm cell) Uvitex OB 5 0.052 The presence of Santonex R at this level would not cause more than 10% error in the Santonox R 10 0.005 determination of Uvitex OB at 378 nm (b) Additives present in polystyrene Uvitex OB - 200 Styrene monomer Butyl stearate 30,000 ppm ppm 2,000 ppm Wingstay T ‘RISELLA’ oil 40,000 ppm 6,000 ppm Concentration Optical density in extraction at 378 nm Comments liquid(b), ppm (1 cm cell) Uvitex OB 5 0.05 Optical density at 378 nm due to Uvitex OB Wingstay T 150 0.12 is less than 3% of the Butyl stearate 750 0.55 optical density due to Shell ‘RISELLA’ interference at 378 nm 1,000 1.00 oil 33 by all the other migrated polymer additives Styrene present. Uvitex cannot 50 0.06 monomer be determined under these conditions. 3 (a) Assuming that in the extraction test, 1 cm of plastic is contacted with 10 ml of extraction liquid and that both additives completely migrate from the plastic into the extractant. (b) Assuming that in the extraction test, 1 cm3 of plastic is contacted with 40 ml of extraction liquid and that all the additives completely migrate from the plastic into the extractant. (Reproduced from Author’s own files)

78

Determination of Ultraviolet Stabilisers in Extractants

Figure 5.3 Ultraviolet spectrograms in the 250-418 nm region of Uvitex OB and other polystyrene additives in the 5% sodium carbonate extractant. (Reproduced from Author’s own files)

Drug Administration extraction procedures can be concentrated by evaporation in a water bath. No satisfactory extraction procedure has been found for concentrating the liquid paraffin extractant. Finally, the extracts are dried with anhydrous sodium sulfate, evaporated to dryness and made up to 2 ml with distilled water or another appropriate spectroscopic solvent, preparatory to ultraviolet spectroscopy. Such 79

Additive Migration from Plastics into Foods extraction procedures should be checked against solutions of known concentrations of the additives in the extraction liquids in order to confirm that quantitative recovery is being obtained through the analytical procedure. Figure 5.1 shows ultraviolet and visible fluorescence spectrograms of a solution of Uvitex OB in the 50% w/v ethyl alcohol:water extract. This substance absorbs ultraviolet energy at 360 and 378 nm and re-emits a strong fluorescent light at 415 and 435 nm (i.e., about 55 nm higher in each case). Figure 5.4 shows a calibration curve for the direct spectrofluorimetric determination of up to 25 ppm Uvitex OB in the 5% sodium carbonate extractant at 415 and 435 nm. In many instances, visible fluorescence methods are less subject to interference by other polymer additives present in the extractant than are ultraviolet methods. Thus, Uvitex OB has an intense ultraviolet absorption at a wavelength high enough (378 nm) to be outside the region where many interfering substances in the extractant would be excited to fluoresce. Therefore, in some instances visible fluorimetry offers a method of determining an extractant constituent without interference from other constituents when this would not be possible by ultraviolet spectroscopy.

Figure 5.4 Calibration curves: direct determination of Uvitex OB in the 5% sodium carbonate extractant. (Reproduced from Author’s own files) 80

6

Determination of Plasticisers in Extractants

6.1 Phthalates in Oily and Aqueous Extractants Polyvinylchloride (PVC) films have found wide applications in the packaging of a large variety of foodstuffs, such as fresh meat, fruits and vegetables, and cheeses [1]. These commercial films, in addition to the polymeric component, contain a number of additives. The one used in the highest proportion is the plasticiser. Plasticisers are low-melting solids or high-boiling organic liquids added to plastic polymers in order to aid flow and processing, to extend and modify the natural properties of the resin and to develop new, technologically important properties not present in the resin itself [2]. Among the most commonly used plasticisers are di-2-ethylhexyl phthalate (DEHP) and di-2-ethyl hexyl adipate (DEHA) [3]. The amount of packaging components that may be leached by food or food simulating solvents depends on the original concentration of the particular component or migrant in the polymer, its solubility in the solvent and/or the partition coefficient between the polymer and solvent as well as temperature and time [4]. Various workers have discussed the occurence of phthalate plasticisers in food packaging materials [5, 6].

6.1.1 Phthalate Platicisers in Fatty Extractants Gas chromatography (GC) has been employed for the direct detection of the migration of phthalate plasticisers from plastics materials into fat and fatty foods [7]. In this method a 7% solution of the sample, e.g., a solution of butter fat containing a plasticiser such as dibutyl phthalate in acetone is cooled in a centrifuge tube in a bath of ice-salt (3:1) to freeze out the greater part of the fat. The tube is centrifuged for 1 minute and the supernatant solution is decanted. This solution (2 or 5 µl) is injected on to a pre-column (17 cm × 3 mm) that is packed with Chromosorb G AW-DMCS 81

Additive Migration from Plastics into Foods (80 to 100 mesh) coated with a mixture of SE-30 and dicyclohexyl phthalate and has been conditioned for 15 hours at 370 °C in the presence of nitrogen. This precolumn adsorbs the remaining fat from the test solution and the plasticiser passes to the analytical gas chromatographic column (2 m × 3 mm) packed with 4% of silicone oil DC on Chromosorb G AW-DMCS (80 to 100 mesh) and temperature programmed from 140 °C to 250 °C at 22.5 °C per minute with nitrogen as carrier gas (28 ml per minute) and flame ionisation detection. Down to 30 ppm of plasticisers of high boiling point can be determined in fat with an error within ±6%. Rohleder and Bruchhousen [7] applied this method to cheese with a plastic outer coating. Bergner and Berg [8] also carried out early work on the migration of citric acid and phthalate plasticisers from PVC and cellulose acetate. Figge and Koch [9] also studied the influence of the depth of the HB 307 (fat simulent) layer on the migration of additives from polymer films. To determine the extent to which the ratio between the contact area and the volume of simulant (i.e., the depth of the layer of HB 307 on the test material) affects the amount of migration they designed migration cells of differing width but otherwise identical construction. The basic design of these cells has been described by Figge and Piater [10] and is shown in Figure 6.1. With a constant area of contact, the layers of simulant between pairs of test films fixed in the cells parallel to each other were exactly 5, 10 or 20 mm thick. These studies were carried out with pressed films of PVC containing levels of 14C-DEHP between 4 and 25% (w/v). Separate determinations of plasticiser migration were carried out with each depth of simulant studied and the mean amount of plasticiser was calculated in each

Figure 6.1 Migration cell. (Reproduced with permission from Figge and Piater, Deutsche Lebensmittel-Rundschau, 1971, 67, 9 [10]) 82

70.92

34.90

34.90

34.90

34.90

34.90

6.0

9.0

12.0

16.0

20.0

25.0

1.06

1.06

1.06

1.06

1.06

1.11

1.17

SD (%) n = 10

8.617

6.878

5.823

4.155

3.162

4.225

4.139

µCi/g

1.58

2.24

2.07

1.57

1.96

1.24

2.71

SD (%) n = 40

Film

213

216

210

220

218

224

222

µm

4.8

5.6

5.5

7.0

11.8

11.2

7.3

SD (%) n = 50

Thickness

2849.1

28.9.5

2758.5

2960.3

3003.2

2994.6

2990.3

mg/dm2

2.7

2.8

2.7

3.0

2.8

3.0

3.0

SD (%) n=5

Weight

0.043 (2.7) 0.58 (2.4)

0.017 (2.1) 0.039 (2.7) 0.51 (1.9) 2.13 (1.4) 8.73 (0.8) 22.51 (0.9) 79.76 (0.4)

0.019 (3.0) 0.039 (1.1) 0.58 (0) 2.11 (1.0) 9.07 (0.4) 24.63 (0.7) 82.47 (0.6)

72.48 (0.9)

23.91 (1.6)

8.30 (0.5)

2.04 (0.8)

0.009 (7.4)

10 mm

5 mm

20 mm

Migration of additive (%)† with a depth of simulant between two films of

Reproduced with permission from Figge and Koch, Food and Cosmetics Toxicology, 1973, 11, 4, 975 [9].

† Values in parentheses are the coefficient of variation (SD) of the single results about the mean of five determinations. Migration was determined after one-sided contact for 10 days at 40 °C.

SD = standard deviation/mean value of n single values × 100

*Composition: 95.0-74% (w/v) Vinflex 503 (BASF, Ludwigshafen), 4.0-25% di-(2-ethyl-n-hexyl) [7,8-14C] phthalate, 0.5% Wachs E (Farbwerke Hoechst, Frankfurt/Main), 0.5% Stabiliser C (Farbenfabriken Bayer, Leverkusen).

104.21

µCi/g

Specific radioactivity Di(2-ethyl-n-hexyl) 7,8 14 C phthalate)

4.0

Content of di-(2ethyl-nhexyl) phthalate (w/v)

PVC test film*

Table 6.1 Migration of Plasticiser from PVC test films into varying volumes of fat simulent HB 307 during 10 days at 40 °C

Determination of Plasticisers in Extractants

83

Additive Migration from Plastics into Foods case. In addition, for each test film, the mean amounts of transferred plasticiser and the coefficient of variation of the single results about the mean were calculated thus giving a mean migration value independent of the depth of the simulant layer. The results of the study on the migration of plasticiser from PVC films into different depths of HB 307 layers are given in Table 6.1. The extent of migration of DEHP phthalate in 10 days at 40 °C increased markedly with increasing plasticiser content and the concentration of DEHP in the simulant layers reached relatively high values. Nevertheless, the amounts of plasticiser migrating, for example, from the PVC film containing 25% (w/v) of plasticiser into the different depths of fat simulant layers were approximately equal, and this similarity was even more marked with films containing lower levels of plasticiser. The curves in Figure 6.2 are the means of the 15 individual determinations of migration from each test film plotted as a function of the plasticiser content and the coefficient of variation of the single results about the

Figure 6.2 Mean migration of plasticiser di-(-2-ethyl-n-hexyl)phthalate from PVC test films as a function of the plasticiser content. Broken lines indicate the coefficient of variation of the single results about the mean. (Reproduced with permission from Figge and Koch, Food and Cosmetics Toxicology, 1973, 11, 4, 975 [9]) 84

Determination of Plasticisers in Extractants mean. The narrow range of error confirms that the amounts of plasticiser migrating from the PVC test films into HB 307 are practically independent of the depth of the simulant layer. The amounts of the antioxidant, Ionox 330, that migrated from the high-density polyethylene film into HB 307 were 0.084, 0.082 and 0.069% for simulant layers of 5, 10 and 20 mm, respectively. Figge and Koch [9] consider, therefore, that a simulant layer of 5 mm between two test films is sufficient for a migration test. This makes the quantitative determination of migrated additives easier by increasing the ratio of contact area to the amount of simulant. In comparative migration tests, the contact areas between the test film and test fat should be kept constant throughout the study, since the amount of additive migrating in each case is proportional to this area. The design of the migration cell reflects normal conditions of use, in that the test film is in one-sided contact with the simulant. The final version of this migration cell provides a contact area between the test film and the fat simulant of 2 × 50 cm and a simulant layer 5 mm in depth between the two circular films. Figge and Koch [9] conclude that for test films used for packaging edible oils and foodstuffs containing fats, it is appropriate to test migration into fat simulant HB 307 (mp 29.3 °C) in a migration cell providing a contact area of 2 × 50 cm, and a simulant layer 5 mm in depth. After the required amount of simulant has been placed in the cell, migration is tested either for 10 days at 40 °C or under conditions actually encountered in practice. For carrying out migration tests on plastic containers, the same fat simulant and test conditions may be used, and when its volume and internal surface have been determined, the container is filled with glass beads and the required amount of simulant is placed in it. Analysis of the simulant follows either type of contact and the resulting level of migrate in the simulant (µg/g) can be converted to µg migrate/dm2 contact area and thence to µg migrate/g foodstuff. These workers stress that high temperatures encountered for limited periods in practice, (e.g. during pasteurisation, sterilisation or preparation of foodstuffs in plastics packs) must be simulated by use of appropriate test temperatures and times. Moreoever, experimental films made only for test purposes should be of the same thicknesses as the article that will actually be used, since the thickness of the film has a definite influence on the amount of additive migrating. The extent of migration also depends on the physico-chemical properties of the test material, including its density, surface quality and crystallinity. Test specimens, both of films and plastics containers should therefore be made under standard production conditions. Bergner and Berg [8] studied the migration of citrate and phthalate plasticisers from PVC and cellulose triacetate. For the determination of the extractability of phthalate esters from soft PVC, Wildbrett and co-workers [11] acidify the extract with hydrochloric acid, extract with light 85

Additive Migration from Plastics into Foods petroleum and dry the extract over anhydrous sodium sulfate, then evaporate in a rotary vacuum evaporator. The residue is dissolved in ethanol and the extinction measured at 274 nm. A calibration graph was prepared with standard solution of DEHP. More recently Belhaneche-Bensemra and co-workers [12] studied the interactions between plasticised PVC packaging and food. The additives, including DEHP plasticiser, tin-based heat stabiliser, processing aid, and internal and external lubricants, were first characterised and kinetic studies of their specific migrations were then carried out using various analytical methods such as GC, Fourier transform infrared (FTIR) spectroscopy, atomic absorption spectrometry and dynamic scanning calorimetry (DSC) analysis. The influence of various parameters such as temperature, stirring, nature of food simulant and initial concentration of plasticiser was investigated. Migration tests were conducted using four food simulants (sunflower oil, distilled water, 3% (w/v) aqueous acetic acid, and 15% (v/v) aqueous ethanol). These food simulants represent all fatty, liquid and moist foods and beverages, except for beverages with a high alcoholic strength. The test conditions used were 10 days at 25 °C and 45 °C with and without agitation. Ten circular samples were immersed in 100 ml of each food simulant. A circular sample and 10 ml of food simulant were taken out every day to be analysed. Each sample was wiped and weighed. The rate of variation of the mass (τ) was determined as a function of time following the relationship: τ (%) = mt – m0/m0 × 100 where: m0 = initial mass before immersion, and mt = mass of the sample at the time t. The rate of mass variation (τ) as a function of time gives information about the phenomenon which occurred between the samples and the food simulants used. An increase in weight of the plastic means that the food simulant penetrated the sample while a decrease means that some additives migrated in the food simulant. Hence τ gives information about the overall migration that occurred. Figures 6.3 and 6.4 illustrate the variation of τ as a function of time for the four food simulants used at 45 °C with agitation. It can be noted that the highest rates of mass variations, were obtained in sunflower oil indicating that migration of additives occurred. For aqueous simulants the rates of mass variations are practically the same and a very small increase of τ was observed indicating penetration of the food simulant in the PVC discs. On the other hand, the highest values of τ were obtained for the formulations containing the highest amounts of plasticiser; slightly lower values [14] were obtained when agitation was not used in the extraction test. This is expected, since the solubility of DEHP is better in sunflower oil than in aqueous simulants. Furthermore, the migration of DEHP can involve the migration of the other additives which are present in the formulation like stabiliser and lubricant. Details of the gas chromatography, infrared and DSC methods are given next. 86

Determination of Plasticisers in Extractants

Figure 6.3 Effect of the nature of the food simulant on the rate of mass variation τ at 45 °C with agitation: (a) 60% DEHP; (b) 50% DEHP; (c) 40% DEHP; (d) 30% DEHP. ▲: Sunflower, l: 15% v/v aqueous ethanol, ■: 3% w/v aqueous acetic acid, ✴: distilled water. (Reproduced with permission from Belaneche-Bensemra and coworkers, Macromolecular Symposia, 2002, 180, 1, 191 [12])

• Application of Gas Chromatography [12] A Philips PYE Unicam 304 GC with a flame ionisation detector was used. Chromatographic conditions were a column of OV-1 (1.8 m × 4 mm od), column temperature: 280 °C; injection temperature: 300 °C; detector temperature: 300 °C; and nitrogen flow: 20.5 ml/min. The effect of the initial concentration of DEHP on its specific migration in sunflower oil is illustrated by Figure 6.4. It is observed that the amount of migrating DEHP is related to its initial concentration in the PVC discs, to the temperature of migration testing and to the presence or absence of agitation. It is obvious that the mobility of the plasticiser molecules increased with increasing temperature and that the migration is favoured by agitation. 87

Additive Migration from Plastics into Foods

Figure 6.4 Effect of the initial concentration of DEHP on its specific migration in sunflower oil: (a) 45 °C with agitation; (b) 45 °C without agitation; (c) 25 °C with agitation; (d) 25 °C without agitation. ▲: 60% DEHP, l: 50% DEHP, ✴: 40% DEHP, n: 30% DEHP. (Reproduced with permission from Belaneche-Bensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [12])

Figures 6.5 and 6.6 illustrate the effect of the nature of food simulant on the migration of DEHP at 45 °C and 25 °C with agitation, respectively. In all the cases, it is observed that the amounts of migrating DEHP are more pronounced in sunflower oil than in aqueous ethanol. This is due to the better solubility of DEHP in sunflower oil. Furthermore, for the same initial concentration of plasticiser, the amounts of migrating DEHP are higher at 45 °C in the two food simulants. This is due to the fact that the mobility of the plasticiser molecules increases with increasing temperature.

Application of Fourier Transform Infrared spectroscopy [12] A Philips type PU 9800 FTIR spectrophotometer was used. The food simulants were placed between two KBr pellets and analysed directly. On the other hand, the PVC circular samples were dissolved in tetrahydrofuran. After evaporation of the solvent, a polymeric film was recovered and analysed. 88

Determination of Plasticisers in Extractants

Figure 6.5 Effect of the nature of food simulant on the migration of DEHP at 45 °C with agitation: (a) 60% DEHP; (b) 50% DEHP; (c) 40% DEHP. ▲: sunflower oil, n: ethanol. (Reproduced with permission from BelanecheBensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [12])

The direct analysis of the spectra of the food simulents did not show the migration of the additives due to overlapping of characteristic bands. However, direct analysis of the PVC did show that concentration of plasticiser decreased with time for bands located at 1585 cm-1, 1128 cm-1 and 742 cm-1.

DSC Analysis Samples were heated from 5 to 100 °C under nitrogen at 10 °C/min. It is known that the presence of a plasticiser decreases the glass transition temperature (Tg) of PVC. For that purpose, the Tg of virgin PVC and of four formulations containing 0, 10, 20 and 30% of DEHP in addition to the other additives, were determined. The results are given in Table 6.2 which show that the values of Tg decreased with increasing the amount of the plasticiser. 89

Additive Migration from Plastics into Foods

Figure 6.6 Effect of the nature of food simulant on the migration of DEHP at 25 °C with agitation: (a) 60% DEHP; (b) 50% DEHP; (c) 40% DEHP. ▲: sunflower oil, n: ethanol. (Reproduced with permission from BelanecheBensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [12])

Table 6.2 Glass transition temperatures for various formulations of PVC Sample Tg (°C) Virgin PVC 81.00 0% DEHP 69.40 10% DEHP 53.50 20% DEHP

33.30

30% DEHP

15.40

Reproduced with permission from N. Belhaneche-Bensemra, C. Zeddam and S. Ouahmed, Macromolecular Symposia, 2002, 180, 1, 191 [12]

As would be expected the Tg of PVC exposed to sunflower oil decreased with an increase in exposure time to sunflower oil. 90

Determination of Plasticisers in Extractants

6.1.2 Phthalates in Aqueous Alcoholic Extract Recycled paper and board is being used more and more in a wide variety of applications, and food packaging is one of the ongoing uses that could be explored. Recently, the Council of Europe proposed limitations on the concentration of several potential migrants in recycled paper and board intended for use as food packaging materials, and the proposed limits depend on the type of food. Without doubt, the use of any recycled paper or board materials in contact with food involves previous analysis of the materials, to know the potential migrants that could be transferred to the food [13]. However, one of the critical points in the analysis of migrants is the extraction of the paper or board samples. Most of the published papers about the analysis of components in paper and board employ solvent extraction with ethanol [14] or toluene, both by direct immersion or using the accelerated solvent extraction procedure [15]. One attractive approach in this context is the use of supercritical fluid extraction (SFE) with carbon dioxide as the supercritical fluid, which can provide a fast and efficient extraction of a wide range of compounds. SFE has been shown as an excellent extraction system of different matrixes [16-26]. Nerin and co-workers [26] examined the applicability of SFE for the extraction of several paper and board samples of different composition, including virgin and recycled fibre. They optimised the SFE using carbon dioxide for the extraction of contaminants in 15 samples of recycled paper and board. An experimental design was used for simultaneous optimisation of the variables involved in both the extraction step and the collection of the extract. Methanol was used as modifier. Several plasticisers such as diethyl phthalate, diisobutyl phthalate, di-n-butyl phthalate, dioctyl adipate, and DEHP (from 2 to 100 µg/g of paper) were found in the recycled samples. A discriminate analysis applied to all results obtained, allowed them to classify the samples in three different groups according to the content of recycled pulp (0, 10-30, and >80% of recycled pulp), the sample thickness (from <300 to >600 µm), and the surface treatment of the paper. The optimum conditions for collection and extraction procedures are listed in Table 6.3. SFE analysis does not require previous digestion, dissolution, or special treatment of the samples. Each paper sample was extracted using the optimum conditions. The SFE extracts were directly analysed by GC-mass spectrometry (MS). Samples of paper and board were analysed by SFE-GC-MS using the optimised procedure described previously. Figure 6.7 shows the chromatograms of one virgin sample (R11) and one 100% recycled sample (R1), which can be considered as representative of those virgin and recycled samples, respectively, analysed by SFE and liquid extraction with ethanol. 91

92

St

Dy

Dy

Eq

1

2

5

5

Time (min)

10

0

0 43

43

43 120

120

120

Methanol modifier P (MPa) T (°C) (%) 1.2

Solvent vol (ml) 1

Trap flow (ml/min) 85

T restrictor (°C)

Collection Step

0

Tads (°C)

Reproduced with permission from C. Nerin, E. Arsenio and C. Jiminéz, Analytical Chemistry, 2002, 74, 22, 5831 [26]

Tdes: Desorption temperature

Tads: Adsorption temperature

Dy: Dynamic step

St: Static step

1

0

0

CO2 flow time (ml/ min)

Eq: Equilibrium step

Mode

Step

Extraction Step

Table 6.3 Optimum values for the global process SFE

50

Tdes (°C)

Additive Migration from Plastics into Foods

Determination of Plasticisers in Extractants

Figure 6.7 Chromatograms of (a) virgin paper (R11), recycled paper (R1) analysed by supercritical fluid chromatography and by liquid extraction. (Reproduced with permission from Nerin and co-workers, Analytical Chemistry, 2002, 74, 22, 5831 [26]) 93

Additive Migration from Plastics into Foods As can be seen, the compounds extracted at higher concentration were the plastisicers (phthalates). Comparing the SFE extraction with the liquid extraction using ethanol applied to the same sample (Figure 6.7), similar qualitative analysis was obtained. Looking through the chromatograms of all the samples extracted, it can be seen that the behaviour of the recycled samples was very similar within the group, while the concentration of all the analysed compounds was much lower in the virgin samples. This result agrees with the idea of a stronger penetration of the supercritical fluid into the recycled pulp. Similar values were obtained on board samples using the ethanol extraction GC-MS procedure with the exception of some of the recycled board samples. Several interesting conclusions can be emphasised from this study: (1) SFE is a good procedure to extract the contaminants present in paper samples. This extraction technique, easy to use with minimum handling and time, has been shown to be very efficient for paper analysis. (2) The lower amount of organic solvents (<2 ml) required in SFE for an efficient extraction of contaminants from paper samples, compared with the liquid extraction, confirms SFE as a safer and environmental friendly technique. (3) The SFE procedure should be optimised following an experimental design, so that all the variables involved in the process, as well as the interaction between the different variables, were optimised at the same time. (4) The collection system of the final extract from SFE is very critical, and a polar solvent such as methanol is required to elute the more polar compounds. This way, the SFE extraction provides results similar to those obtained by liquid extraction using ethanol. (5) The characteristics of the paper samples affect the content of contaminants in the samples, and such contaminants are potential migrants. SFE provides a good way to evaluate the potential migration of chemicals to food, and assuming 100% migration, additional migration tests could be avoided.

6.2 Phthalates and Adipates in Aqueous and Fatty Extractants Simoneau and Hannaert [27] applied GC to study the stability of DEHA, DEHP and octadecyl 3-(3,5-di-tert-butyl-4-hydroxy phenyl) propionate (Irganox 1076) in three EU aqueous and fatty food simulents including 15% aqueous ethanol, 95% aqueous ethanol, 3% aqueous acetic acid, olive oil and iso-octane. This work was carried out within the framework of the AIR3-CT94-2360 EU project. 94

Determination of Plasticisers in Extractants Such studies on the stability of additives used in food packaging are designed for regulatory purposes as an aid to decide whether the legislation should regulate limits for plasticisers based on a quantity in the food packaging itself (quantity in packaging material, QM) or based on an ingested dose by the consumer (specific migration limit, SML). The stability data will be integrated and used by the EU legislative body (DG-III-E-1) to decide whether the regulation limits for plasticisers should be based on a quantity in the food packaging itself (QM) or based on an ingested dose by the consumer, i.e., a dose migrating to the simulants (SML). In the case of these additives it was shown that their stability allows the use of an SML rather than a QM [27]. The various test conditions included exposures of 10 days at 40 °C, 1 hour at reflux temperature for all aqueous simulants, 10 days at 40 °C and 1 hour at 175 °C for the olive oil and two days at 20 °C and 3 hours at 60 °C for the isooctane simulant. Following the exposure, the additive samples are extracted from aqueous simulants with hexane. A sonication step is necessary to ensure maximum extraction of control samples. In the case of the isooctane simulant, the samples are analysed directly from the simulant. The oil samples are extracted by acetonitrile. The extracts of samples exposed to various heat conditions as well as unexposed spiked controls and blanks were analysed by GC on a non-polar 5% phenylmethylpolysiloxane capillary column with high temperature capabilities. The results show that DEHA, DEHP and Irganox 1076 are stable at 40 °C and at reflux temperature in ethanolic or acidic aqueous simulants. The various additives are also stable in the organic isooctane simulant as well as in the fatty simulant olive oil. Spiked, unspiked (blank) and control samples were extracted using the same protocol. Extraction from 95% ethanol was best achieved by adding hexane in proportions of 1:1 with respect to the volume of simulant and 10% distilled water to provoke the instantaneous separation of the aqueous ethanol/hexane layers. Extraction from 15% ethanol or 3% acetic acid was also performed with hexane as previously described. The extraction from olive oil was performed by adding a volume of acetonitrile equal to that of the simulant. After addition of the solvent, the mixtures were shaken manually for 1 minute and subsequently sonicated for 15 minutes. They were then left to separate (approximately 1 hour). Aliquots of the organic phase were taken and transferred into small labelled GC vials for analysis. Samples exposed in isooctane were analysed directly. All samples were analysed by GC. The stationary phase was 5% phenylmethylpolysiloxane. Injection volume was 1 µl and carrier gas was helium at 33 cm/s. When analysing the aqueous stimulant extracts, a column of HP 5 (25 m length, 0.25 mm internal diameter) and 0.25 µm film thickness was used (Hewlett Packard, Milan, Italy). The temperature programme was from 100 °C to 300 °C at 30 °C/min, with a holding 95

Additive Migration from Plastics into Foods time of 5 minutes for DEHA and DEHP and 10 minutes for Irganox 1076. The injector and detector temperatures were 250 °C and 300 °C, respectively. The split mode was (1:1) for DEHA and Irganox, and splitless for DEHP. When analysing the oil simulant extracts, a column of DB-5 HT of 25 m length, 0.32 mm internal diameter and 0.15 µm film thickness was used. The analysis of the acetonitrile extracts of oil simulant was performed with the injector at 300 °C, detector at 365 °C, and the column was programmed from 100 °C to 365 °C at 30 °C/min and held for 25 minutes. In aqueous simulants, the extraction yields for controls were lower than that of the samples in the same experimental scheme. A sonication treatment of the controls during the extraction led to extraction yields similar to that of the samples. Both samples and controls (as well as blanks were therefore subsequently sonicated in the extraction step, both in the cases of 15% ethanol and of 3% acetic acid, and for all time-temperature exposures. The yields from the extraction for the samples in aqueous solutions ranged from 70 to 100%. Yields of extraction were generally lower for the exposure of 1 hour at reflux temperature compared with 10 days at 40 °C. The analysis of the three additives from aqueous simulants is represented in Figure 6.8. The SML concentrations for the different compounds analysed was sufficient to yield reproducible peaks and quantitation. Both methanol and acetonitrile were assessed as extracting solvent for the olive oil. The results showed that the extracts obtained with acetonitrile and methanol gave the same yield for the recovery of additive, while the acetonitrile extract was cleaner of interfering compounds from the olive oil. The acetonitrile extract in the peak region of DEHA, DEHP and Irganox are shown in Figures 6.9-6.11, respectively. The data show that each additive peak appears in a clear region and can be quantified without ambiguity. The results obtained by this method for DEHA and DEHP were reproducible, which made the method acceptable for fast screening purposes. However, the yields were relatively low around 45-50%. Acetonitrile gave slightly better recoveries but they never exceeded 60%. The extraction of Irganox from olive oil could not be achieved to an acceptable level of recovery, and thus a more suitable method would be high-performance liquid chromatography (HPLC) rather than GC even for screening purposes. The stability of additives was expressed as percentage remaining in the exposed samples compared with controls (Table 6.4). Values for pH measurements before and after the time-temperature exposure were similar. The results obtained suggested that the additives tested were fairly stable in all simulants. Only DEHP exhibited a slight 10% instability when exposed for 10 days at 40 °C in acetic acid. Aignasse and co-workers [28] have described a HPLC method for the determination of DEHP in PVC packaging and studied its release into polymer extracts. 96

Determination of Plasticisers in Extractants

Figure 6.8 Chromatograms of the three additives analysed. These typical chromatograms show the peak for each spiked at its SML in the simulant, following their time-temperature exposure and extraction. (Reproduced with permission from Simonceau and Hannaert, Food Additives and Contaminants, 1999, 16, 5, 197 [27]) 97

Additive Migration from Plastics into Foods

Figure 6.9 Chromatogram of DEHA in the olive oil extract. The portion of chromatogram presented corresponds to the region of elution of DEHA within the acetonitrile extract from the oil. The upper part represents the blank specimen and shows the chromatographic baseline at the retention time of the additive, whereas the bottom part shows the presence of the additive DEHA. (Reproduced with permission from Simonceau and Hannaert, Food Additives and Contaminants, 1999, 16, 5, 197 [27]) 98

Determination of Plasticisers in Extractants

Figure 6.10 Chromatogram of DEHP in the olive oil extract. The portion of chromatogram corresponds to the region of elution of DEHP within the acetonitrile extract from the oil. The upper part represents the blank specimen and shows the chromatographic baseline at the retention time of the additive, whereas the bottompart shows the presence of the additive DEHP. (Reproduced with permission from Simonceau and Hannaert, Food Additives and Contaminants, 1999, 16, 5, 197 [27]) 99

Additive Migration from Plastics into Foods

Figure 6.11 Chromatogram of Irganox 1076 in the olive oil extract. The portion of chromatogram presented corresponds to the region of elution of Irganox 1076 within the acetonitrile extract from the oil. The upper part represents the blank specimen and shows the chromatographic baseline at the retention time of the additive, whereas the bottom part shows the presence of the additive Irganox 1076. (Reproduced with permission from Simonceau and Hannaert, Food Additives and Contaminants, 1999, 16, 5, 197 [27]) 100

Determination of Plasticisers in Extractants

Table 6.4 Summary of the results of stability testing of the additives. The results are expressed as percentage remaining in the exposed samples compared with the controls, after extraction. Exposure Simulant DEHA DEHP Irganox 1076 15% ethanol 101 101 104 1 h/100 °C 95% ethanol 104 103 105 3% acetic acid 10 104 105 15% ethanol 104 107 97 95% ethanol 102 102 100 10 d/40 °C 3% acetic acid 98 90 103 Olive oil 101 101 2 d/20 °C Iso-octane 101 102 100 3 h/60 °C Iso-octane 103 99 99 1 h/175 °C Olive oil 100 100 Reproduced with permission from Simonceau and Hannaert, Food Additives and Contaminants, 1999, 16, 5, 197 [27]

References 1.

L. Castle, A.J. Mercer, J.R. Startin and J. Gilbert, Food Additives and Contaminants, 1987, 4, 4, 399.

2.

M. Pizzoli and M. Scandola in Polymeric Materials Encyclopedia, Ed., J.C. Salamone, CRC Press, Inc, FL, USA, 1996, CD-ROM.

3.

E. Kondyli, P.G. Demertzis and M.G. Kontominas, Food Chemistry, 1992, 45, 2, 163.

4.

S. Chong, C.M. Guttman, I.C. Sanchez and L.E. Smith in Food Packaging Interactions, Ed., J.H. Hotchkiss, ACS Symposium Series 365, 1988, ACS, Washington, DC, USA, p.106.

5.

L. Castle, J. Gilbert and T. Eklund, Food Additives and Contaminants, 1990, 7, 5, 591.

6.

Y-S. Fung and A.S-K. Tang, Fresenius’ Journal of Analytical Chemistry, 1994, 350, 12, 721. 101

Additive Migration from Plastics into Foods 7.

K. Rohleder and V.B. Bruchhousen, Deutsche Lebensmittel-Rundschau, 1972, 68, 180.

8.

K.G. Bergner and H. Berg, Deutsche Lebensmittel-Rundschau, 1972, 68, 282.

9.

K. Figge and J. Koch, Food and Cosmetics Toxicology, 1973, 11, 4, 975.

10. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1971, 67, 9. 11. G. Wildbrett, K.W. Evers and F. Kiermeier, Fette Seifen Anstrichmittel, 1969, 71, 330. 12. N. Belhaneche-Bensemra, C. Zeddam and S. Ouahmed, Macromolecular Symposia, 2002, 180, 1, 191. 13. A. Damant and L. Castle in Proceedings of the Pira International Symposium, Paper in Contact with Foodstuffs, Edinburgh, UK, 1999, Paper No.9. 14. BS EN 1186-13, Materials and Articles in Contact with Foodstuffs – Plastics – Part 13: Test Methods for Overall Migration at High Temperatures, 1998. 15. C. Jiminéz, Internal Report, CPS, University of Zaragoza, 2000. 16. J.L. Hedrick, L.J. Mulcahey and L.T. Taylor, Mikrochimica Acta, 1992, 108, 3-6, 115. 17. I.J. Barnabas, J.R. Dean and S.P. Owen, Analyst, 1994, 119, 11, 2381. 18. S. Bøwardt and S.B. Hawthorne, Journal of Chromatography, 1995, 703, 1-2, 549. 19. T.L. Chester, J.D. Pinkston and D.E. Raynie, Analytical Chemistry, 1996, 68, 12, 487R. 20. I.A. Stuart, J. MacLachan and A. McNaughtan, Analyst, 1996, 121, 3, R11. 21. J.L. Snyder, R.L. Grob, M.E. McNally and T.S. Oostdyk, Analytical Chemistry, 1992, 64, 17, 1940. 22. K. Schäfer and W. Baumann, Fresenius’ Zeitschrift für Analytische Chemie, 1989, 332, 8, 884. 23. EPA Method 3560, Supercritical Fluid Extraction of Total Recoverable Petroleum Hydrocarbons, 1.0 Scope and Applications, 1996. 102

Determination of Plasticisers in Extractants 24. EPA Method 3561, Supercritical Fluid Extraction of Polynuclear Aromatic Hydrocarbons, 1.0 Scope and Applications, 1995. 25. S.B. Hawthorne, Analytical Chemistry, 1990, 62, 11, 633A. 26. C. Nerin, E. Arsenio and C. Jiminéz, Analytical Chemistry, 2002, 74, 22, 5831. 27. C. Simoneau and P. Hannaert, Food Additives and Contaminants, 1999, 16, 5, 197. 28. M.F. Aignasse, P. Prognon, M. Stachowicz, R. Gheyouche and D. Pradeau, International Journal of Pharmaceutics, 1995, 113, 2, 241.

103

Additive Migration from Plastics into Foods

104

7

Determination of Organotin Thermal Stabilisers in Extractants

7.1 Organotin Compounds in Fatty Extractants Koch [1] has described a direct quantitative determination of di-n-octyltin compounds in fat simulant HB 307 using dithizone reagent. This method has a lower detection limit of 0.75 ppm of the organotin compound. If the migration tests are carried out at a ratio of plastics surface (cm2) to amount of fat simulant (g) of 5:1, then the detection limit for organotin stabilisers is 15 µg/dm2 (Figure 7.1). The extractability of labelled organotin compounds (di-n-octyl [14C] tin-dithioglycollic acid-2-ethyl-n-hexyl ester) from rigid polyvinylchloride (PVC), low and high-density

Figure 7.1 Extraction curve, migration of labelled Advastab TM-181 from PVC into HB 307 (1.5% additive in PVC film), TMSI: methyl tin thioglycollic acid-2-ethyln-hexyl ester. (Reproduced with permission from J. Koch, Deutsche LebensmittelRundschau, 1974, 70, 209. [1] ©1974 Wissenschaftliche Verlagsgesellschaft mbH) 105

Additive Migration from Plastics into Foods polyethylene and polystyrene into mono, di and tri-glycerides has also been determined radiometrically [2]. Figge and Zeman [3] have also studied the extractability of organotin compounds from rigid PVC using radiometric procedures. The organotin compounds studied included, di-n-octyltin-dithioglycollic acid-2-ethyl hexyl ester, di-n-octyltin oxide, din-butyl tin dichloride, n-octylstannone, n-butylthiostannone, tri-n-octyltin fluoride, di-n-butyl-tin dilaurate, di-n-butyl tin dipropionate and di-n-octyltin maleate. Figge and Koch [4] have studied the migration of Advastab TM-181 FS (methyltin thioglycollic acid-2-ethyl-n-hexyl ester) from PVC into HB 307 fat simulant. Two preparations of this stabiliser which were 14C labelled, the methyltin and the thioglycollic acid-2-ethyl-n-hexyl-ester ,were incorporated into powdered PVC in amounts of 0.5, 1.0 and 1.5% by weight. From the resulting rigid PVC mixtures compression-moulded films were made having specific radioactivities between 1.7 and 8.8 µCi/g. The even distribution of the radioactive quantities of stabiliser in the different samples was checked by direct continuous radioanalytical measuring methods as well as by autoradiography and liquid scintillation counter. In model tests the migration of methyl tin stabiliser and of its methyl-tin and thioglycollic acid ester residues from rigid PVC into edible fats was checked radiometrically during 30 days at 20 °C/65% relative humidity and 10 days at 40 °C/65% relative humidity in one-sided contact as well as for 30 minutes at 70 °C in all-round contact between the film samples and the fat simulant HB 307. It was found that under the same test conditions, despite an increasing stabiliser content of the film samples, the percentage migration of stabiliser as related to the total content of the films remains constant. However, the migration and extraction values depend on the different test conditions (Figure 7.1). Assuming that the intact methyl tin stabiliser migrates from the rigid PVC into the fat simulant and that 0.1 m2 of PVC are in contact with 1 kg of each simulating agent, for films with a stabiliser content of 1.5% by weight, migrate concentrations found in the simulant were: 0.18 and 0.35 ppm after 30 days at 20 °C, 0.14 and 0.15 ppm after 10 days at 40 °C and 0.044 and 0.11 ppm after 30 minutes at 70 °C. The migration/time curves show that with a prolonged testing time these concentrations do not increase (Figure 7.2). Figge and Koch [4] also discuss the migration of hydrolysis products of the stabiliser. In further work on the migration of Avastab TM181FS stabiliser from PVC, Figge and Bieber [5] and Figge [6] study the migration rate from PVC into HB 307 of four methyl tin stabilisers of identical chemical composition, containing either monomethyl [14C]-tintrithioglycollic acid-2-ethyl hexyl ester, di-methyl [14C]-tin-dithioglycollic acid-2-ethyl 106

Determination of Organotin Thermal Stabilisers in Extractants

Figure 7.2 Extraction curve, migration of Advastab TM-181 MS (– – o – –), thioglycollic acid-2-ethylhexyl ester (....o....), methyl tin chloride (—o—) from PVC film into HB 307. Extraction temperature, 40 °C; relative humidity, 65%; area of PVC film, 0.1 m2. (Reproduced with permission from J. Koch, Deutsche LebensmittelRundschau, 1974, 70, 209. [1] ©1974 Wissenschaftliche Verlagsgesellschaft mbH)

hexyl ester, trimethyl [14C]-tin-mono-thioglycollic acid-2-ethyl hexyl ester, or all three 14 C-labelled methyl tin compounds as radioactive indicators. These were incorporated into four identical powdered PVC mixtures in amounts of 1.5% by weight each. The resulting rigid PVC mixtures were processed into compression-moulded films, the specific radioacitivities of which varied between 0.6 and 4.5 µCi/g. In model tests the migration of radioactive stabiliser components from the rigid PVC films into the fat simulant HB 307 was determined radiometrically over a period of 10 days during which the films were stored at 40 °C/65% relative humidity, one side of the film being in contact with the simulant. Assuming that the intact stabiliser components migrate, the results show that fat-containing food products packed and stored in rigid PVC materials stabilised with Avastab TM 181 FS will contain not more than 1 ppb trimethyl-tin-mono-, 73 ppb dimethyl-tin-di- and 39 ppb monomethyl-tintri-thioglycolate. Assuming that the methyl tin thioglycolates are converted into the methyl tin chlorides before or during migration, the corresponding concentrations are 0.5 ppb trimethyl-tin-mono-, 29 ppb dimethyl-tin-di- and 13 ppb monomethyl-tintrichloride. The migration/time curves show that these concentrations hardly increase 107

Additive Migration from Plastics into Foods on prolonging the test period. Figge and Bieber [5] also found that irrespective of the duration of any heat treatment given to the PVC the amounts of trimethyl tin compounds present in stabilised rigid PVC materials never exceed those originally added with the methyl tin stabilisers. Adcock and Hope [7] have described methods for the determination of the extractability of dioctyl-S,S-bis-(iso-octylmercapto) acetate stabilisers from PVC into sunflower seed oil. Belhaneche-Bensemra and co-workers [8] used atomic absorption spectrometry to determine the levels of tin and other metals in sunflower oil containing an organotin stabiliser during a migration test carried out at 45 °C with agitation. Tin levels in the oil ranged from 395 ppm (3 days exposure) to 650 ppm (12 days exposure).

References

1.

J. Koch, Deutsche Lebensmittel-Rundschau, 1974, 70, 209.

2.

K. Figge, Kunststoffe, 1971, 61, 11, 832.

3.

K. Figge and A. Zeman, Kunststoffe, 1973, 63, 8, 543.

4.

K. Figge and J. Koch, private communication.

5.

K. Figge and W.D. Bieber, private communication.

6.

K. Figge, Verpackungs-Rundschau, 1975, 8, 59.

7.

L.N. Adcock and W.G. Hope, Analyst, 1970, 95, 1135, 868.

8.

N. Belhaneche-Bensemra, C. Zeddam and S. Ouahmed, Macromolecular Symposia, 2002, 180, 1, 191.

108

8

Determination of Organic Sulfur Compounds in Extractants

The main uses of rubber components in contact with foodstuffs include natural rubber (NR) or nitrile rubber (NBR) seals and gaskets used in food-processing equipment, valve diaphragms and rubber hoses. Natural rubber is used in latex dipping to form feeding bottle teats and rubber gloves. With addition of fillers it is made into hosing and conveyor belting. Nitrile rubber (a synthetic acrylonitrile - butadiene copolymer) is more widely used than natural rubber and in particular for hosing, dairy equipment, and many types of seals and gaskets including can sealants and bottle closures. The service conditions in these applications can be characterised into two sets. The first is typified by tubing, hosing and belting where a large contact area of rubber is used but which makes only transient contact with a large volume of food or drink. The second type is typified by can sealants where only a small surface area is exposed to the food or drink but for a prolonged period. Rubber can contain rather high concentrations of additives and vulcanisation residues. A typical formulation for NR may contain 1% by weight of N-cyclohexyl-2benzothiazole sulfonamide (CBS). A nitrile rubber could use 2-mercaptobenzothiazyl disulfide (MBTS) at typically 0.5% by weight. The benzothiazoles are quite soluble in water and can migrate extensively into aqueous foods and food simulents. Sidwell [1] reported that 2-mercaptobenzothiazole (MBT) was the major water-soluble migrant from rubber. For example, he took a NR, vulcanised using 0.53% sulfur and 1.3% w/v CBS and tested it for migration at 40 °C into beer and 10% ethanol solution. After 8 hours of contact at a ratio of 0.5 dm2:100 ml, the migration of benzothiazole (BT) was 2.3 and 2.8 mg/kg and the migration of MBT was 16 and 24 mg/kg into the beer and simulant, respectively. This research highlighted the need to test retail foods that may come into contact with natural and nitrile rubber for possible migration of MBT, the MBT precursors MTBS and CBS, and related substances such as BT. There have also been recommendations from the Nordic countries [2, 3] that the migration of MBT merits attention. Sidwell of Rapra Technology [4] carried out a study of the types of organosulfur compounds present in a range of rubber materials and other elastomers and the effect of ageing of various sanitisers and cleaning agents on the migration of these compounds from rubbers in contact with food. 109

Additive Migration from Plastics into Foods Sidwell [4] used combined gas chromatography – mass spectrometry and, in some cases, high-performance liquid chromatography (HPLC) to obtain data. Elastomers examined included NR, nitrile rubber, ethylene-propylene dimer and a vinylidene fluoride – hexafluoropropylene copolymer. Examination of migrating species was carried out by methods discussed in the EC Framework Directive 89/109/ EC (see Chapter 16). Distilled water and diethyl ether extractants were included in this study. From the studies undertaken it was evident that: • All of the rubber compounds tested were found to contain chemical species that have the potential to migrate into food. Whether these migrating species are of concern will depend on the time, temperature, surface area of rubber that is in contact with them, and the fat content of the food (Table 8.1).

Table 8.1 Migration of mercaptobenzothiazole and benzothiazole (mg/kg) into distilled water at 40 °C 3rd 24 hour 1st 24 hour 2nd 24 hour Migrant extraction after extraction extraction autoclaving Sample NR Mercaptobenzothiazole 17.6 6.5 13.9 Benzothiazole 1.2 1.0 Not detected Sample EPDM Mercaptobenzothiazole 1.0 1.4 2.8 Benzothiazole 0.5 0.3 Not detected NR: Natural rubber EPDM: Ethylene-propylene-diene terpolymer Reproduced with permission from J. Sidwell in Proceedings of a Rapra Conference RubberChem 2002, Munich, Germany, 2002, Paper 16, p.117-130, Table 1 [4]. © 2002, Rapra Technology

• The use of CBS curative in the NR sample led to relatively high levels of water soluble migrating species (Figure 8.1). Compounds detected mainly relate to the cure reaction products of the CBS cure accelerator i.e., MBT, BT, and related species. Levels decreased on repeat exposure and high temperature ageing.

110

Determination of Organic Sulfur Compounds in Extractants

Figure 8.1 Water extractables - 24 hours at 40 °C from compound NR2 (partitioned into dichloromethane and examined by GC-MS). (Reproduced with permission from J. Sidwell in Proceedings of a Rapra Conference RubberChem 2002, Munich, Germany, 2002, Paper 16, p.117-130, Figure 1 [4]. © 2002, Rapra Technology)

Similar species and levels of migrants would be expected from related sulfenamide accelerators. This finding probably confirms why sulfenamide curatives are not approved for food contact applications in the German BgVV recommendations despite their approval in rubbers compounded to the US FDA requirements. Diethyl ether extraction data indicated that fatty foods could potentially extract organic additives such as plasticisers, extenders and antidegradants (unless of high molecular weight or chemically reacted with the polymer chain) and also cure system reaction products (Figure 8.2). The polycyclic aromatic hydrocarbon (PAH), pyrene has been extracted with diethyl ether from two rubbers containing SRF N762 carbon black (ethylene-propylene-diene terpolymer (EPDM) and NBR). Pyrene and fluoranthene were present in the volatiles from EPDM on heating at 150 °C. Another PAH extract from the NR sample which contained HAF N330. These findings confirm the requirement for using high purity carbon blacks in food contact rubber materials. 111

Additive Migration from Plastics into Foods

Figure 8.2 Diethyl ether extractables from NBR3. (Reproduced with permission from J. Sidwell in Proceedings of a Rapra Conference RubberChem 2002, Munich, Germany, 2002, Paper 16, p.117-130, Figure 4 [4]. © 2002, Rapra Technology)

The toxic monomer, acrylonitrile was released from the NBR compound on heating at 150 °C. Dicyclopentadiene monomer was also released from the EPDM sample under similar conditions. It is unclear whether or not these monomers are a product of the heating process. The temperature of 150 °C is higher than would be expected as a normal use condition. Migration testing for these monomers is advisable. Chemical sanitation with sodium hypochlorite can lead to the formation of chlorinated organic species from the rubber compound. For example, chlorinated BT, and various chloroanilines have been detected in the NR and NBR compounds studied (from the phenylene diamine antiozonants present). Over the last few years there have been major developments in ionisation systems for examining the composition of components separated or present in liquid media. When examining extractable or migrating components from rubber materials, it is important to remember that certain types of extractable components may either breakdown or react when being examined by gas chromatography mass spectrometry (GC-MS). Alternatively components may be too high in molecular weight terms or be too involatile to be detected. 112

Determination of Organic Sulfur Compounds in Extractants

Table 8.2 Species extracted from EPDM sample and detected by GC-MS Peak time Assignment Comments 5.11 Ethane, isothiocyanatoFrom cure system 7.52 Dicyclopentadiene Monomer 8.50 N-Formylpiperidine From cure system 9.19 Tetramethylthiourea From cure system 9.27 Benzothiazole From cure system 11.92 Benzothiazole, 2-(methylthio)From cure system 12.16 2-Benzothiazolamine, N-ethylFrom cure system 14.48 Dodecanamide, N-(2-hydroxyethyl)From process aid 14.88 Pyrene From carbon black 15.09 N,N-dimethylpalmitamide From process aid 15.42 Dodecanamide, N-(2-hydroxyethyl)From process aid 16.57 Di(2-ethylhexyl)phthalate Contaminant 17.18 Thiazole, 4-ethyl-2-propylFrom cure system Reproduced with permission from J. Sidwell in Proceedings of a Rapra Conference RubberChem 2002, Munich, Germany, 2002, Paper 16, p.117-130, Table 4 [4]. © 2002, Rapra Technology

Table 8.2 shows the total ion current GC-MS trace for the examination of diethyl ether extractable species from the sulfur cured EPDM compound. Note these are all of relatively low molecular weight. Examination of the diethyl ether extract by liquid chromatography (LC) using a C18 reverse phase gradient elution separation with atmospheric pressure chemical ionisation (APCI) in the positive mode gave the total ion current trace shown in Figure 8.3. The last three peaks showed ions of masses 538, 566 and 594 and are believed to relate to the presence of tellurium dithiocarbamates in the extract (ions +2H+ from ionised protonic solvent). These species were not detected by GC-MS. Koch [5, 6] has described an ultraviolet (UV) spectroscopic method for the determination of 2-(2′-hydroxyphenyl-5′-methylphenyl)benzotriazole, thioethyleneglycol-bis-(βaminocrotonate) and 2-phenylindole in HB 307 glyceride fat simulent. Barnes and co-workers [7] developed an LC-S method for the analysis of food and drink for residues of specific vulcanisation accelerators used to crosslink rubber. The method was applied to the analysis of 236 samples of selected retail foodstuffs 113

Additive Migration from Plastics into Foods

Figure 8.3 LCMS examination of the diethyl ether extractables from EPDM2 using APCI ionisation (positive total ion chromatogram). (Reproduced with permission from K.A. Barnes, L. Castle, A.P. Damant, W.A. Read and D.R. Speck, Food Additives and Contaminants, 2003, 20, 2, 196 [7]. © 2003, Taylor & Francis)

(Table 8.3) that may have been in contact with rubber during their manufacture, transport and storage. The method of analysis involved extraction of the food using acidified solvent and analysis by LC-APCI-MS. The detection limit depended on the sample type and was in the range 0.005-0.043 mg/kg for MBT and BT. The average analytical recovery rate was 82% for MBT and 87% for BT. No trace of MBT or BT was found in any of the retail samples. It is also concluded that no sample contained significant MBTS or CBS. Both MBTS and CBS are important accelerators used to vulcanise rubber and they break down in foodstuffs to form MBT and BT. The absence of MBT and BT in the foodstuffs therefore also provides proof of the absence of MBTS and CBS. The method described by Barnes and co-workers [7] focused on MBT and BT because tests confirmed that both MBTS and CBS were unstable and broke down rapidly during food analysis to give MBT and BT as the main products. Two extraction procedures were used: one for milks, yoghurts and infant formulae; the other for all other foodstuffs. Sample extracts were then analysed by LC-APCI-MS using 114

Determination of Organic Sulfur Compounds in Extractants

Table 8.3 Limit of detection for MBT and BT in the various food types for the average recovery for that sample type Sample type

MBT (mg/kg)

BT (mg/kg)

Canned beers

0.012

0.005

Vending machine water

0.024

0.010

Baby food

0.026

0.017

Long-life fruit juice

0.027

0.027

Tinned fruit-juice

0.043

0.039

Canned soft drink

0.042

0.014

Tinned fruit syrup

0.009

0.005

Yoghurt

0.034

0.012

Infant and evaporated milk

0.022

0.029

Bottled drinks and draught beer

0.019

0.010

Milk

0.036

0.010

Vending machine coffee

0.008

0.007

Reproduced with permission from K.A. Barnes, L. Castle, A.P. Damant, W.A. Read and D.R. Speck, Food Additives and Contaminants, 2003, 20, 2, 196 [7]. © 2003, Taylor & Francis

hydroxybenzothiazole as a marker. Milks, yoghurts, beverages and infant formulations were extracted with acetonitrile. HPLC was performed using 1% acetic acid in water then 1% acetic acid in acetonitrile as mobile phases. Mass spectra were obtained using a benchtop mass spectrometer using APCI operated in the positive-ionisation mode. The instrument was tuned and calibrated on a mixture of polyethylene glycol (PEG) 300, PEG 600 and PEG 1000. The source temperature was 140 °C and the APCI probe temperature 500 °C. The ions were monitored at m/z 168 (MBT), 132 (hydroxybenzothiazole marker) and 136 (BT). The ion at m/z 109 was included as a confirmation ion for all three compounds. The cone voltage was set at 20 V for all three parent ions and at 60 V for the confirmation ion. The dwell time was 0.08 second, the interchannel delay 0.02 second and the mass span 0.4 daltons. Spectra were acquired from 15 to 25 minutes for each injection. Figure 8.4 gives a typical LC-APCI-MS trace for a mixture of five benzothiazoles. Analytical recoveries through the procedure ranged from 166% (MBT in beer) to 45% (MBT in soft drinks). 115

Additive Migration from Plastics into Foods

Figure 8.4 LC-APcI-MS trace (reconstructed total-ion current) for some vulcanisation agents and residues (each 200 μg/kg). (Reproduced with permission from K.A. Barnes, L. Castle, A.P. Damant, W.A. Read and D.R. Speck, Food Additives and Contaminants, 2003, 20, 2, 196 [7]. © 2003, Taylor & Francis)

Niessen and co-workers [8] showed that MBT and BT can be analysed by GC. The presence of sulfur conveys good sensitivity and selectivity using flame photometric detection (FPD) or, alternatively, more selective detection is possible using MS. Special precautions are required during sample extraction, clean up and GC analysis since the thiol is prone to oxidation. GC analysis is not suitable for analysis of the parent accelerators since they break down on heating to yield MBT. Therefore, while with care GC-FPD or GC-MS analysis can give the total quantity of MBT and related species that migrate, any information on individual substances is lost. Blosczyk [9] described an LC method with UV detection for the determination of MBT and zincMBT in migration tests of teats. An LC-MS method has also been reported for the determination of thermolabile alkyl bis (2-benzothiazolylsulfenamides [10]. This method uses N-cyclohexyl-2-benzothiazole sulfenamide as an internal standard. 116

Determination of Organic Sulfur Compounds in Extractants The use of LC-MS for the identification of minor components in BT derivatives was investigated by Niessen and co-workers [8] who evaluated the merits of GCMS, moving belt LC-MS, particle-beam LC-MS and LC-MS-MS with thermospray ionisation (TSP). They found that LC-TSP-MS, LC-TSP-MS-MS and LC-PB-MS were the best techniques. Recent advances in LC–MS applied to food analysis [11] meant that LC-MS was the method of choice for the investigations reported here as it offered the potential to analyse with minimal sample clean up. A level of interest of 0.05 mg/kg food was set at the commencement of this work and this set the level of detection that should be achieved using LC-MS.

References 1.

J.A. Sidwell, Migration Studies – Food Contact of Elastomeric Materials, Final Project Report, Food Standards Agency, London, UK, 1997.

2.

M.L. Binderup, L. Lillemark, A. Petersen and J.H. Petersen, Screening Analysis of Soothers and Feeding Bottles, Report No.13, National Food Agency of Denmark, Søborg, Denmark, 1997.

3.

T. Hallas-Moller, Proposals for Official Control of Materials Intended to Come into Contact with Foodstuffs, Temanord 510, Nordic Council of Ministers, Copehagen, Denamrk, 1997.

4.

J.A. Sidwell in Proceedings of a Rapra Conference: RubberChem 2002, Munich, Germany, 2002, Paper 16, p.117-130.

5.

J. Koch, Deutsche Lebensmittel-Rundschau, 1972, 68, 401.

6.

J. Koch, Deutsche Lebensmittel-Rundschau, 1972, 68, 404.

7.

K.A. Barnes, L. Castle, A.P. Damant, W.A. Read and D.R. Speck, Food Additives and Contaminants, 2003, 20, 2, 196.

8.

W.M.A. Niessen, C.C. McCarney, P.E.G. Moult, U.R. Tjaden and J. van der Greef, Journal of Chromatography, 1993, 647, 1, 107.

9.

G. Blosczyk, Deutsche Lebensmittel-Rundschau, 1992, 88, 392.

10. M. Stolcova, A. Kaszonyi, M. Hronec and T. Liptaj, Journal of Chromatography, 1995, 710, 2, 351. 11. K.A. Barnes, Applications of High-Performance Liquid Chromatography/ Atmospheric Pressure Ionisation/Mass Spectrometric Strategies in Problems of Food Safety, University of East Anglia, Norwich, UK, 1999. [PhD Thesis] 117

Additive Migration from Plastics into Foods

118

9

Determination of Polydimethyl Siloxanes in Extractants

9.1 Wine and Olive Oil In the food-processing industry, polydimethylsiloxanes (PDMS) categorised as E900 are used as structure-shaping food additives [1-3]. They are used in various stages of food processing as foam-suppressing and anti-foaming agents. Silicones and PDMS may also be present in foodstuffs as contaminants. The source of contamination of foodstuffs with silicone polymers may also be the packaging material, which may be manufactured by impregnation in order to obtain silicon paper or the food processing operation involving the use of silicone grease [4-6]. PDMS in foodstuffs may also result from silicone plant protection agents used during production of edible plants [7, 8]. When silicones started to be used in food and pharmaceutical industries, it became necessary to identify and determine trace amounts of these substances. Food and medicine agencies, such as the US Food and Drug Administration or the European Union require that adding silicones in a production process should be controlled, i.e., that the amount in the finished product should be identifiable [9]. According to the existing regulations [10], amounts of PDMS in foodstuffs should not exceed 10 mg/kg of the product. There is also a recommended acceptable daily intake of 1.5 mg/kg body weight given by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) in 1974. Atomic absorption spectrometry (AAS), atomic emission spectrometry (AES) [11], infrared (IR), Fourier transform infrared (FTIR) and Raman spectroscopy have all been studied at various times for the determination of silicon compounds. AAS has been used to determine silicon in methylisobutylketone, chloroform or petroleum ether extracts of packaging materials and foodstuffs [12-17]. However, these methods suffer from the disadvantage that they do not distinguish between organic and inorganic silicon compounds, similarly inductively coupled plasma AES measures total silicon [11]. In speciation analysis, spectroscopic methods such as IR absorption spectroscopy, FTIR and Raman spectroscopy are very useful. However, high detection limits of these methods render their application impossible for trace analysis of PDMS, in particular for biological samples or foodstuffs [11]. 119

Additive Migration from Plastics into Foods Various workers have reported on the application of 1H nuclear magnetic resonance spectroscopy (1H-NMR) to the determination of PDMS in biological materials [18-23]. Mojsiewicz-Pienkowska and co-workers [23] were the first to apply this technique to the determination of PDMS in foodstuff extraction liquids, particularly olives and wine.

9.1.1 PDMS in Olive Oil and Wine For wine samples, approximately 500 g were weighed into a conical flask with a ground-glass stopper and shaken for 2 hours with 5 ml of carbon tetrachloride. After separation of the layers, the organic layer was removed. For edible oil samples, approximately 5 g were weighed into a conical flask connected to a condenser and 50 ml of 10 M aqueous sodium hydroxide solution were added in a 10:1 v/v ratio. Hydrolysis was carried out by heating the sample at the temperature for gentle boiling for 90 minutes. After cooling the contents of the flask, the solid layer was separated by filtration. PDMS was extracted from the residue by stirring with 30 ml of carbon tetrachloride. The contents were transferred to a test tube and the carbon tetrachloride was isolated by centrifugal separation. Having obtained a clear extract of carbon tetrachloride, the sample was concentrated five times by evaporating the solvent. Extracts were analysed by 1H-NMR. For measurement of 1H-NMR spectra, 0.3 ml extract were taken and 0.3 ml hydroxyl dimethyl siloxane (HMDS) internal standard containing 12 µg/ml C6D6 were added. The chemical shift of the proton signal in PDMS methyl groups was 0.18 ppm (14 Hz) and 0.10 ppm (8 Hz) for the HMDS used as an intensity standard. The position of resonance lines for protons in Si-CH3 for PDMS and HDMS is shown in Figure 9.1. Owing to the considerable chemical similarity between both compounds, their signals have similar 1H chemical shifts, however, the distance between the HMDS signal in relation to PDMS is sufficient to allow HMDS to be used as an internal standard. Limits of detection achieved are reported in Table 9.1. Recoveries were in the range 95 to 97%. Reported values of PDMS contents in a range of 28 imported wines range between <0.06 to 0.35 mg/kg and for edible oils range between <0.6 to 11.9 mg/kg (Table 9.2) [21]. The latter value being obtained for Italian corn oil. With the exception of Italian corn oil none of the wine or oil samples exceeded the permissible standards of 10 mg/kg laid down by the Codex Alimentarius Commission. 120

Determination of Polydimethyl Siloxanes in Extractants

Figure 9.1 Resonance line positions for protons in Si–CH3 groups for PDMS and HMDS (spectrometer 1H-NMR 80 MHz). (Reproduced with permission from Mojsiewicz-Pienkowska, Z. Jamrogiewicz and J. Lukasiak, Food Additives and Contaminants, 2003, 20, 5, 438 [23]. © 2003, Taylor and Francis)

Table 9.1 Limits of quantification for wine and edible oils after a silicone (PDMS) spiking process Sample type Limits of quantification (mg/kg) 1 1 H-NMR @ 80 MHz H-NMR @ 500 MHz Wine 0.06 0.006 Edible oil 6.0 0.6 Reproduced with permission from Mojsiewicz-Pienkowska, Z. Jamrogiewicz and J. Lukasiak, Food Additives and Contaminants, 2003, 20, 5, 438. (Table 2, p.442) [23]. © 2003, Taylor and Francis

121

Additive Migration from Plastics into Foods

Table 9.2 Results of PDMS determination in edible oil samples with 80 MHz and 500 MHz spectrometers. PDMS content (mg/kg) Edible oil sample

80 MHz spectrometer

500 MHz spectrometer

Olio di Semi di Mais – Olitalia Italian corn oil

11.9

11.8

Olio di Semi di Arachide – Olitalia Italian peanut oil

<6.0

<0.6

Goya Spanish olive oil

<6.0

<0.6

Olio di Semi di Soia – Olitalia Italian soybean oil

8.3

8.4

Costa d’Oro – Spoleto

<6.0

<0.6

Frideal Mix – Olitalia Italian frying oil blend

<6.0

5.4

Panensky Slovakian sunflower seed oil

<6.0

<0.6

Mazola Best Foods Division CPT-American corn oil

6.5

6.5

Italian oil blend

6.3

6.2

Olio di Semi di Soia – Costanza — Salvadori Italian soybean oil

<6.0

4.3

Oleificio – Costa d’Oro Italian soybean oil

9.8

9.7

Bartek – ZPT Warszawa Polish soybean oil

<6.0

<0.6

Kujawski – ZPT Kruszwica Polish soybean oil

<6.0

<6.0

Olio di Semi Vari – Olitalia

Reproduced with permission from Mojsiewicz-Pienkowska, Z. Jamrogiewicz and J. Lukasiak, Food Additives and Contaminants, 2003, 20, 5, 438, Table 5 [23]. © 2003, Taylor and Francis

122

Determination of Polydimethyl Siloxanes in Extractants

References 1.

Eighteenth Report of Joint FAO/WHO Expert Committee on Food Additives (JECFA), World Health Organisation, Geneva, Switzerland, 1974, p.24.

2.

Joint FAO/WHO Expert Committee on Food Additives (JECFA), Evaluation, World Health Organisation, Geneva, Switzerland, 1996, p.9.

3.

A. Rutkowski, St. Gwiazda and K. Dabrowski, Dodatki Funkcjonalne do Zywnosci, Agro and Food Technology, Katowice, Poland, 1993.

4.

M. Kazo, New Food Industry, 1992, 34, 17.

5.

Food and Drug Administration Federal Register, Title 21 Food and Drugs, Part178, Indirect Food Additives: Adjuvants, Production Aids and Sanitizers, FDA, Washington, DC, USA, 59, 123, 33194-33195.

6.

S. Masa Lero, A. Makoto, K. Shiro and E. Mitsuru, inventors; Kokai Tokkya Koho, JP 1994,158,586, 1994.

7.

P. Steven, Pesticide Science, 1993, 38, 103.

8.

K. Klein and S. Wilkowski in Proceedings of the 4th International Symposium on Adjuvants for Agrochemicals, 1995, p.27-31.

9.

H.J. Horner, J.E. Weiler and N.C. Angelotti, Analytical Chemistry, 1960, 32, 7, 858.

10. Food and Drug Administration, Code of Federal Regulation, 21 Section 173340, FDA, Washington, DC, USA, 1981. 11. B.A. Cavic-Vlasak, M. Thompson and D.C. Smith, Analyst, 1996, 121, 6, 53R. 12. I.P. Freeman, F.B. Padley and W.L. Sheppard, Journal of the American Oil Chemists’ Society, 1973, 50, 101. 13. W.G. Doeden, E.M. Kushibab and A.C. Ingala, Journal of the AOAC International, 1980, 57, 73. 14. J.L. Kacprzak, Journal of AOAC International, 1982, 65, 1, 148. 15. D.A. McCamey, D.P. Ianelli, L.J. Bryson and T.M. Thorpe, Analytica Chimica Acta, 1986, 188, 1, 119. 16. R.D. Parker, Journal of AOAC International, 1990, 73, 721. 123

Additive Migration from Plastics into Foods 17. E.G. Gooch, Journal of AOAC International, 1993, 76, 3, 581. 18. L. Garrido, B. Pfleiderer, M.Papisov and J.L. Ackerman, Magnetic Resonance in Medicine, 1993, 29, 6, 839. 19. B. Pfleiderer, J.L. Ackerman and L. Garrido, Magnetic Resonance in Medicine, 1993, 29, 656. 20. B. Pfleiderer, J.L. Ackerman and L. Garrido, Magnetic Resonance in Medicine, 1995, 30, 5, 534. 21. B. Pfleiderer and L. Garrido, Magnetic Resonance in Medicine, 1995, 33, 1, 8. 22. M.W. Lieberman, E.D. Lykissa, R. Barrios, C.N. Ou, G. Kala and S.V. Kala, Environmental Health Perspectives, 1999, 107, 2, 161. 23. K. Mojsiewicz-Pienkowska, Z. Jamrogiewicz and J. Lukasiak, Food Additives and Contaminants, 2003, 20, 5, 438.

124

10

Determination of Lubricants in Extraction Liquids

To study the influence of temperature and time on additive migration, Figge and Koch [1] measured the amount of labelled additive migrating into HB 307 at a range of storage temperatures between 30 and 80 °C and times between 1 and 30 days. Consider, for example, the migration of labelled n-butyl stearate from polystyrene film into HB 307. It can be seen in Figure 10.1 that at 40 °C, in contrast to the situation at higher temperatures, there is no further migration of the additive after 10 days. In fact, in this case, the final migration value, amounting to 5.68% of the amount of additive in the original film, was reached after only four days. With all the polymer/additive combinations studied by Figge and Koch there was a marked increase in migration values above 50 °C. The extent of this increase and the point at which it began depended both on the type of plastics material and the physicochemical properties of the additive. Thus, while the migration of Ionox 330 from either polystyrene or high-density polyethylene (HDPE) showed a marked increase only above 65 °C, an increase in the migration of n-butyl stearate from polystyrene and of stearic acid amide from HDPE was clearly observed at 50 °C. Consequently, in order to simulate migration under storage conditions, it is necessary to know the extent of this increased migration and the temperature at which it begins to occur, since the use of too high a test temperature could indicate an unrealistically high level of additive transfer. Comparison of the amounts that had migrated from the test films into HB 307after 30 days at 20 °C and after 10 days at 40 °C (Table 10.1) indicates that the migration of additives from plastics packs into fatty foods stored at about 20 °C can be simulated satisfactorily in tests with HB 307 for 10 days at 40 °C. Figge [2] also investigated use of radiochemical methods for measuring the migration of butyl stearate from PVC and HDPE into fatty extractants. Percentage migrations of n-butyl stearate into edible oils ranged from 46% (sunflower oil) to 57% (Biskin) from HDPE and from 1.2% (olive oil) to 1.8% (sunflower seed oil) from polystyrene. Thus the migration of stearic acid from HDPE was some 40 times greater than the rate of migration from PS in extraction tests conducted at 20 °C for 60 days. Piacentini [3] developed an indirect gas chromatographic method for the determination of the migration of stearic acid from rubber articles into olive oil and coconut oil. Test 125

Additive Migration from Plastics into Foods

Figure 10.1 Migration of n-butyl stearate [14C] from polystyrene film into fat simulant HB 307 during periods of 1 to 30 days at temperatures between 30 and 80 °C. (Reproduced with permission from K. Figge and J. Koch, Food and Cosmetics Toxicology, 1973, 11, 4, 975 [1]. © 1973, Elsevier)

pieces (25 mm × 100 mm × 0.5 mm) of nitrile rubber were weighed and suspended in olive or coconut oil and control pieces were similarly suspended in air. After the desired time, the pieces were removed, cleaned, dried and weighed. The total weight of material absorbed by the oil from the rubber was derived by deducting the final weight from the sum of the initial weight plus the weight of oil absorbed. To determine the weight of oil absorbed, the pieces of rubber were heated under reflux with N-ethanolic potassium hydroxide for eight hours, and after acidification with hydrochloric acid, the liberated fatty acids were extracted with ethyl ether. The methyl esters or the silyl derivatives were prepared and subjected to gas chromatography on a stainlesssteel column (250 cm × 4 mm) packed with 15% of LAC 72S on Diatoport S (60 to 126

Determination of Lubricants in Extraction Liquids

Table 10.1 Comparison of the amounts of additive migrating from different test films into fat simulent HB 307 and coconut oil Migration of additive (%) After 30 days at 20 °C into Coconut Oil

HB 307

HB 307

Avastab 17 MOK- [14C]

0.010

0.014

0.012

Ionox 330 - [14C]

0.080

0.120

0.106

Stearic amide - [ C]

3.86

3.48

3.66

Ionox 330 - [14C]

4.89

4.76

5.58

n-Butyl stearate - [14C]

5.27

5.42

5.76

Test film

Labelled additive

PVC HDPE Polystyrene

After 10 days at 40 °C into

14

Reproduced with permission from K. Figge and J. Koch, Food and Cosmetics Toxicology, 1973, 11, 4, 975 [1]. © 1973, Elsevier

80 mesh) or on a similar column (180 cm × 2 mm) packed with 10% of W 9S on Diatoport S (80 to 100 mesh). The columns were operated at 200 °C, with nitrogen as carrier gas (50 ml/min) and flame ionisation detection. The chromatography of the methyl esters was more accurate, but less rapid, than that of the silyl derivatives. The chromatograms were evaluated by reference to the peak area for a fatty acid characteristic of the oil, it being important that any peak due to a constituent of the rubber, e.g., stearic acid, did not interfere. For this reason, Piacentini [3] proposed that the ‘fat’ used in such migration studies should be a synthetic triglyceride prepared from a saturated fatty acid having an odd number of carbon atoms.

References 1.

K. Figge and J. Koch, Food and Cosmetics Toxicology, 1973, 11, 4, 975.

2.

K. Figge, Food and Cosmetics Toxicology, 1972, 10, 8, 15.

3.

R. Piacentini, Industria della Gomma, 1972, 16, 4, 46.

127

Additive Migration from Plastics into Foods

128

11

Determination of Monomers and Oligomers in Extractants

11.1 Determination of Acrylonitrile in Aqueous and Fatty (Liquid Paraffin) Extractants The procedure described next is based on cathode-ray polarography of a solution of the extractant in 0.02 M aqueous tetramethyl ammonium iodide (TMAI). Polarographic cell solutions were prepared by mixing 1 ml of 0.2 M aqueous TMAI base electrolyte base to the distilled water extractant. The results obtained by polarography of 0.02 M aqueous base-electolyte solutions (start potential of –1.8 volts) showed that it is possible with either the cathodic direct or the cathodic derivative circuits to determine down to 1 ppm of acrylonitrile monomer in the distilled water extractant. Confirming the conclusions of Bird and Hale [1] and Daues and Hamner [2] it was found that the presence in the cell solution of dissolved oxygen did not interfere in the polarographic determination of acrylonitrile in the range –1.8 to –2.3 volts. Liquids obtained in extractability tests on styrene-acrylonitrile copolymers also usually contain a small amount of styrene monomer besides acrylonitrile monomer. It has been shown that the presence of up to 500 ppm of styrene monomer in the test solution does not interfere in the determination of acrylonitrile in aqueous solutions.

11.1.1 Determination of Acrylonitrile in Aqueous-ethanol Extractants To 9 ml of each ethanol-water mixture containing 50 ppm of added acrylonitrile was added 1 ml of 0.2 M TMAI base-electrolyte solution. Reagent blank solutions were also prepared by mixing 9 ml of the appropriate acrylonitrile free ethanol-water mixture with 1 ml of 0.2 M base electrolyte. These samples and blank solutions were examined polarographically at a start potential of –1.8 volts and the peak currents occurring at the acrylonitrile maximum were noted. The influence of the ethanol content of the extraction liquid on the peak current obtained with 50 ppm of acrylonitrile (corrected for the peak current of the reagent blank solution) is shown in Figure 11.1. It is seen that lower peak currents are obtained as the alcohol content of the test solution increases from zero to 50%, i.e., the procedure for determining acrylonitrile becomes rather less sensitive as the alcohol content of the extraction liquid is increased. Acrylonitrile could be reproducibly determined, however, in amounts down to 1 ppm in all the alcohol solutions. 129

Additive Migration from Plastics into Foods

Figure 11.1 Graph of peak current versus ethanol concentration of test solution for 50 ppm of acrylonitrile in aqueous ethanol solutions containing 0.02 M tetramethylammonium iodide base electrolyte. (Reproduced from Author’s own files)

11.1.2 Determination of Acrylonitrile Monomer in Acidic and Alkaline Extractants Direct polarography of synthetic solutions of acrylonitrile in the 6% hydrochloric acid and in the 5% sodium carbonate extractants is not possible as both extractants produce interfering waves in the acrylonitrile polarogram. Also, when TMAI was added to the 6% hydrochloric acid extractant, the quaternary salt decomposed and free iodine was liberated. The applicability of the azeotropic-distillation procedure described by Daues and Hamner [2] for separating acrylonitrile monomer from interfering impurities before polarography was examined. In this procedure the aqueous sample is distilled in the presence of a mixture of methanol and aqueous sulfuric acid. The methanolacrylonitrile azeotrope, boiling at 61.4 °C, distils first from this mixture and thus the acrylonitrile is recovered in the initial distillate. The azeotropic-distillation procedure was tested by using a synthetic solution of acrylonitrile in water that had been shown by direct polarographic analysis to contain 0.75 ppm of acrylonitrile. A measured volume (500 ml) of this solution, i.e., 0.375 mg of acrylonitrile, together with 5 ml of concentrated sulfuric acid, 25 ml of methanol and 0.1 g of 2,4-dinitrophenyl-hydrazine (to destroy carbonyl compounds) 130

Determination of Monomers and Oligomers in Extractants was transferred to a round-bottomed flask fitted with a lagged glass column packed with 3.175 mm helices and a reflux head. This mixture was distilled for 1 hour under total reflux and then three 4 ml portions of the methanol distillate were collected in separate 10 ml calibrated flasks. Then 1 ml of 0.2 M TMAI and 5 ml of distilled water were added to each flask. The weight of acrylonitrile recovered from the three fractions was determined polarographically by making ‘standard additions’ of a synthetic solution of acrylonitrile in a methanol:water (40:60) mixture to the cell solution. It can be seen from Table 11.1 that approximately 90% of the 0.375 mg of acrylonitrile monomer present in the original 500 ml of water was recovered in the first two 4 ml methanol distillates. The somewhat low recovery of acrylonitrile might be due to a slight hydrolysis of this substance to ammonium acrylate under the acidic conditions used during azeotropic distillation.

11.1.3 Determination of Acrylonitrile in 6% Hydrochloric Acid Extractant The methanol azeotropic-distillation procedure was also applied to a synthetic solution of acrylonitrile in 6% hydrochloric acid extractant. Polarographic analysis of the methanol-acrylonitrile azeotrope was not possible, however, owing to the presence of an appreciable amount of free acid originating from the hydrochloric acid extractant, in the distillates, which interfered in the polarography of acrylonitrile. In a further experiment, a 6% hydrochloric acid solution of 47.3 ppm of acrylonitrile was neutralised by the addition of a small excess of solid calcium oxide. Methanol and sulfuric acid were added and the azeotropic distillation continued as before. It can be seen from Table 11.1 (sample B) that under these conditions more than 90% of the added amount of acrylonitrile was recovered in the first 8 ml of methanol distillate. A preliminary neutralisation with lime was incorporated, therefore, into the procedure for determining acrylonitrile in 6% hydrochloric acid extraction liquids. This procedure should also be applicable to the determination of acrylonitrile in the 3% aqueous acetic acid extractant recommended by the Food and Drug Administration (FDA) [3].

11.1.4 Determination of Acrylonitrile Monomer in 5% Sodium Carbonate Extractant The results in Table 11.1 (samples C and D) show that above 80% of the added amount of acrylonitrile was recovered in the methanol distillate when the azeotropicdistillation procedure was applied to 5% sodium carbonate extractants containing up to 72.4 ppm of acrylonitrile monomer. The recoveries of acrylonitrile in these experiments are lower than those obtained for the distilled water and 6% hydrochloric acid extractants. This may be due to an increased amount of hydrolysis of acrylonitrile 131

132

47.3

14.4

72.4

B

C

D

25 25

5% Na2CO3

5% Na2CO3 1810

360

2370

50 (+ 3.5 g of calcium oxide)

6% HCl

Acrylonitrile added, µg (X)

376

Volume of test solution used for azeotropic distillation, ml

500

Distilled water

Solvent

Reproduced from Author’s own files

0.75

Acrylonitrile, ppm, w/v

A

Sample

Composition of synthetic test solution used for azeotropic distillation

-

-

-

376

Found by direct polarography of sample

1390

248

2250

315

1st 4 ml

1

83

54

93

21

2nd 4 ml

2

Nil

Nil

Nil

Nil

3rd 4 ml

3

1473

302

2343

336

Y

Total

Found in distillates obtained by azeotropic distillation with methanol

Weight of acrylonitrile in test solution, µg

82

84

99

90

Recovery of acrylonitrile in first two fractions obtained by azeotropic distillation Y × 100/X, % w/v

Table 11.1 Determination of Acrylonitrile in Distilled Water, 6% Hydrochloric Acid and 5% Sodium Carbonate Extractants after Azeotropic Distillation with Methanol

Additive Migration from Plastics into Foods

Determination of Monomers and Oligomers in Extractants to ammonium acrylate occurring during reflux in the presence of sodium sulfate, as this salt will elevate the temperature at which the mixture boils during the preliminary one hour reflux period. The azeotropic-distillation procedure is also applicable to the 3% sodium hydrogen carbonate extractant recommended by the FDA (USA) [2].

11.1.5 Determination of Acrylonitrile in Light Liquid-Paraffin Extractant The possibility of aqueous extraction of acrylonitrile from the liquid paraffin was examined. Various synthetic solutions of acrylonitrile in liquid paraffin were extracted with two 250 ml portions of distilled water. The distilled water extracts were filtered into a one litre flask and an azeotropic distillation with methanol made as described previously. The recoveries of acrylonitrile obtained by this procedure for liquid-paraffin extractants containing up to 538 ppm of acrylonitrile are shown in Table 11.2. In all these mixtures, the recovery of acrylonitrile in the first 8 ml of the methanol azeotrope was within 5% of the amount known to be present. Duplicate recoveries obtained in the azeotropic-distillation procedure are reasonably reproducible. Procedures, based on these principles, involving extraction with water and azeotropic distillation with methanol before polarography should also be applicable to the n-heptane extractant recommended by the FDA [3] and also might be useful for the determination of acrylonitrile in the FDA vegetable oil extractants, provided that these can be successfully extracted with water. Trials have not been made on these particular extractants. Detailed procedures for determining acrylonitrile monomer in the various extractants are described next.

11.1.6 Separation of Acrylonitrile from Light Liquid-Paraffin and n-Heptane Extractants Weigh an amount of extractant containing between 0.001 and 0.01 g of acrylonitrile into a clean one litre separating funnel. Into a further separating funnel, weigh the same amount of the blank extractant that has not been brought into contact with the plastic under test. Into each funnel pour 250 ml of water. Stopper the funnels and shake the contents thoroughly. When the two phases have separated, filter the lower aqueous phase through two layers of Whatman No.40 filter paper into a one litre round-bottomed flask. Extract the organic phase with a further 250 ml of water and combine the aqueous extracts. As soon as possible after this extraction procedure, carry out an azeotropic distillation of the aqueous extracts with methanol as described next. 133

134

10

20

200

9

538

297

34

11.1

100

6800

5940

5380

µg (X)

Weight of acrylonitrile present in light liquid paraffin test solution

Reproduced from Author’s own files

g

ppm, w/v

Acrylonitrile Weight of content of the light liquid synthetic light paraffin liquid paraffin sample sample extracted with solution for 2 × 250 ml extraction of water with water

2 2

(ii) 96

Nil

(ii) 6840 (i) 96

Nil

(i) 7220

460

370

(ii) 4980 5160

350

2nd 4 ml

2

(i) 4940

1st 4 ml

1

Nil

Nil

Nil

Nil

Nil

Nil

Nil

3rd 4 ml

3

98

98

6840

7220

5620

5350

5290

Y

Total

Weight of acrylonitrile recovered in fractions of azeotropic distillation, µg

Table 11.2 Determination of acrylontrile in light liquid paraffin extractant

98

103

95

99

%, w/v

Y × 100/X

Mean recovery of acrylonitrile in first two fractions obtained by azeotropic distillation

Additive Migration from Plastics into Foods

Determination of Monomers and Oligomers in Extractants

11.1.7 Determination of Acrylonitrile in Aqueous, Acidic and Alkaline Extractants Into a one litre round-bottomed flask transfer a volume of the plastic-extraction liquid containing between 0.001 and 0.01 g of acrylonitrile. Into a further one litre flask transfer the same volume of the blank extraction liquid. Into each flask introduce distilled water to make the volume up to 500 ml. To the 6% w/v hydrochloric acid extractant only, add a slight excess of calcium oxide (3.5g of calcium oxide per 50 ml of 6% w/v hydrochloric acid is sufficient). To the 5% sodium carbonate and the 3% sodium hydrogen carbonate extractants only, add sufficient concentrated sulfuric acid to neutralise the alkali present. To all extractants add 5 ml of concentrated sulfuric acid, 25 ml of redistilled carbonyl-free methanol, 0.1 g of 2,4-dinitrophenylhydrazine and a few glass beads. Place the two flasks in one litre electric mantles and connect to each flask a 60 mm column packed with 3.175 mm glass helices and a reflux head. Turn the stopcock on the reflux head to total reflux and turn on the water supply to the condenser on the reflux head. Set the voltage of the heating tape on the column to bring the column to about 10 °C above room temperature. Commence heating the flasks and leave them to equilibrate for 1 hour after methanol starts to condense at the reflux head. After the 1 hour reflux period, transfer by pipette, 2 ml of 0.2 M TMAI base electrolyte and 10 ml of water into each of two dry 25 ml stoppered graduated cylinders. Place one of these cylinders at the outlet of the reflux head and open the stopcock to provide a reflux ratio of approximately 1:1. Allow methanol to distil into the cylinder at a rate of approximately 1 ml per minute, until the volume of solution reaches the 20 ml mark. Immediately examine the distillates polarographically as described next. It has been shown that up to 0.02 g of acrylonitrile in the distillation flask charge is recovered in the first 8 ml of methanol-acrylonitrile azeotrope. Completeness of recovery of acrylonitrile in this distillate can be checked by collecting a further 4 ml of distillate in a 10 ml graduated cylinder (containing 1 ml of 0.2 M TMAI and 5 ml of distilled water). Polarography of the solution shows whether any acrylonitrile is present in the second distillate. If acrylonitrile is found in this distillate then it should be included in the reported analytical result. When the K1000 polarograph is used for the analysis, use a dropping mercury electrode and a mercury-pool reference electrode on the cathodic direct circuit. Transfer by pipette 5 ml of sample solution from the 25 ml cylinder into a polarograph cell and immerse the cell in the constant temperature tank of the cathode-ray polarograph. Lower the dropping mercury electrode system over this cell and insert the anode connection into the side-arm of the polarographic cell. If the approximate concentration of monomer in the polarographic cell solution is known, set the instrument to the appropriate sensitivity setting at a start potential 135

Additive Migration from Plastics into Foods of –1.8 volts. If the composition of the solution is unknown, then adjust the polarograph to its maximum sensitivity setting and move the ‘X-shift’ control and the ‘Y-shift’ control until the light spot on the graticule of the cathode-ray tube commences its horizontal sweep at the origin of axes at the left of the graticule. Repeat this operation at different sensitivity settings until the polarographic wave is visible on the graticule. Read off from the graticule the maximum height of the peak, and note the voltage, V, at which this maximum polarographic reading occurs. Transfer by pipette a further 5 ml of solution from the 25 ml graduated cylinder into a dry 25 ml beaker and into this solution deliver a suitable ‘standard addition’ of a solution of acrylonitrile in a methanol-water (40 + 60 v/v) mixture (by using a horizontally held micrometer syringe for delivery). To avoid dilution errors, limit the volume of the ‘standard addition’ added to less than 0.05 ml. Mix the contents of the beaker and pour them into a dry polarographic cell. Note the height of the acrylonitrile wave at V volts. Adjust the instrument to obtain the azeotropic-distillation blank wave on the graticule. Measure the blank peak height corresponding to V volts.

11.1.8 Determination of an Acrylonitrile in Aqueous Ethanol Extractant The azeotropic-distillation procedure cannot be applied to this alcoholic extraction liquid. Transfer 16 ml of the aqueous alcohol plastic-extraction liquid and 16 ml of the blank alcoholic extraction liquid into two 25 ml stoppered graduated cylinders. To each cylinder add 2 ml of 0.2 M TMAI and 2 ml of distilled water and mix. Examine these solutions polarographically as described previously.

11.1.9 Calculation of the Acrylonitrile Contents of the Extractants The amount of acrylonitrile in the plastic-extraction liquid, ppm w/v, is given by: 20 × A × 106  h1S1 − h3S3    5× S  h2S2 − h1S1  (assuming that the methanolic azeotropic distillate is made up to 20 ml and that 5 ml of this solution is used for polarography), where: S = volume, in ml, of plastic-extraction liquid used in azeotropic distillation, 136

Determination of Monomers and Oligomers in Extractants h1 = peak height, in cm, of sample solution before the ‘standard addition’, h2 = peak height, in cm, of sample solution after the ‘standard addition’, h3 = peak height, in cm, obtained in the azeotropic-distillation blank determination, S1, S2, S3 = the corresponding instrument sensitivity settings (the products of h and S are known as peak currents, in µA), and, A = weight, in g, of acrylonitrile present in volume of ‘standard addition’, solution added to the cell solution. Provided that a suitable sample size is taken for analysis, the azeotropic distillationpolarographic procedure can be used for determining acrylonitrile in extractants in concentrations down to 1 ppm or a little lower. Thus, it is seen from Table 11.1 that approximately 90% of the added amounts of acrylonitrile is recovered when the procedure is applied to 500 ml of a 0.75 ppm solution of acrylonitrile in the distilled-water extractant. The method can be used for achieving a similar level of sensitivity in the determination of acrylonitrile in the other aqueous alcoholic or oily extractants for plastics recommended by the British Plastics Federation [4] and the FDA [3]. This level of sensitivity is quite adequate for the examination of extractants that have been brought into contact with styrene-acrylonitrile copolymers under the British Plastics Federation and FDA extractability-test conditions. Actual extraction liquids obtained in extractability tests made on various styreneacrylonitrile copolymers have been found by the previously described procedures to contain from less than 10 ppm up to 200 ppm of acrylonitrile. The amount of acrylonitrile monomer found in the extraction liquids depends, of course, on the extent to which this monomer is removed from the plastic during the manufacturing process. It is advisable to analyse an extraction liquid for acrylonitrile as soon as possible after the completion of the extractability tests. This is because a slow hydrolysis of acrylonitrile to acrylamide or ammonium acrylate occurs on standing, especially in acidic or alkaline extractants. Consequently, low acrylonitrile contents are obtained by the polarographic method. Thus, on analysis immediately after the extraction test, a 6% hydrochloric acid extractant was found to contain 110 ppm of acrylonitrile. Analysis of the same sample two months later showed that the acrylonitrile content had decreased to 75 ppm because of the hydrolysis. 137

Additive Migration from Plastics into Foods

11.2 Determination of Styrene Monomer and Aromatic Hydrocarbons in Aqueous and Fatty (Liquid Paraffin) Extractants 11.2.1 Determination of Styrene Monomer and Other Volatiles in Polystyrene Conventional and modified grades of polystyrene contain low concentrations of residual styrene monomer left in them from the manufacturing process. In addition to the monomer, the polymer might contain other aromatic volatiles such as ethyl benzene, cumene, and so on, which originate as impurities in the styrene charge-stock. For grades of polystyrene used for foodstuff or beverage packaging applications, some transfer of these volatiles will occur from the container to its contents with consequent risk of contamination and tainting of the packaged commodity. In order to measure the extent to which the aromatic polystyrene volatiles migrate from a polymer, gas chromatography can be used to determine low concentrations of these substances in the aqueous and alcoholic British Plastics Federation extractants [5]. The gas chromatographic separation column comprised polyethylene glycol 20,000 (Carbowax 15-20M) supported on Celite (60-72 mesh), and detection was achieved using a flame ionisation detector which has the advantage of high sensitivity and freedom from interference by water in the sample. The instrumental conditions are those of Shapras and Claver [5] in which a 5 µl sample is injected into a separation system comprising two columns in series consisting of a 2 m × 6.4 mm OD column of Tween 81 (20%), on Chromosorb W (30-60 mesh) and Resoflex 446 (10%) on Chromosorb W (30-60 mesh) using a hydrogen flame detector to achieve the desired sensitivity and nitrogen (25 ml/mm) as carrier gas. Figure 11.2 shows the chromatogram obtained with synthetic solutions in the aqueous and an alcoholic British Plastics Federation extractant of various aromatic volatiles likely to occur in the polystyrene. The extractants are injected directly into the gas chromatograph (the olive oil or liquid paraffin extractants cannot be injected directly into the gas chromatographic column). Mixtures of benzene, toluene, ethyl benzene, n-propyl benzene, o-xylene, cumene, m/p-xylene and styrene monomer were resolved from each other and could each be detected in amounts down to approximately 10 ppm. To obtain quantitative determinations of aromatic volatiles in the extraction liquids it is necessary to adopt the standard gas chromatographic technique of incorporating into the extractant a known concentration of a miscible internal standard substance which has a different retention time to any of the sample components. Under the extractability test conditions prescribed by the British Plastics Federation for plastic test pieces of less than 0.5 mm thickness, the plastic and the extractant are contacted under standard test conditions at a ratio of 1 cm3 plastic volume to 138

Determination of Monomers and Oligomers in Extractants

Figure 11.2 Gas chromatographic determination of aromatic volatiles in polystyrene extraction liquids. (Reproduced with permission from P. Shapras and G.C. Claver, Analytical Chemistry, 1964, 36, 12, 2282 [5]. © 1964, ACS)

20 ml extractant. Thus, if a polystyrene test piece contained 0.1% styrene monomer and this completely migrated into the extractant during the test, then at the end of the extraction test the extractant would contain approximately 50 ppm monomer. Styrene monomer can be determined in amounts down to 10 ppm in extraction liquids by gas chromatography. Thus, the gas chromatographic method would be sufficiently sensitive to ascertain whether 20% or more of the original 0.1% styrene monomer in the polystyrene sample had migrated into the extraction liquid during the extraction test. 139

Additive Migration from Plastics into Foods

Table 11.3 Migration of aromatic volatiles from 0.5 mm thick polystyrene sections in the British Plastics Federation extractants Aromatics determined in polystyrene*, % w/v Extraction Styrene Cumene Ethyl benzene liquid Before After Before After Before After extraction extraction extraction extraction extraction extraction Distilled 0.14 0.13 0.012 0.012 0.016 0.015 water 5% Sodium 0.13 0.12 0.012 0.013 0.014 0.015 carbonate 5% 0.13 0.12 0.012 0.011 0.13 0.015 Citric acid 50% Ethyl 0.12 0.09 0.012 0.011 0.015 0.013 alcohol: water Liquid 0.10 0.02 0.011 0.004 0.014 0.003 paraffin *2 g polystyrene contacted with 100 ml extraction liquid Reproduced from Author’s own files

An alternative, more sensitive approach to the measurement of the migration of aromatic volatiles from polystyrene into extraction liquids is to determine them in the polymer both at the beginning and at the end of the extraction tests. The decrease is a measure of the degree of migration of volatiles. Table 11.3 shows some results of analyses for styrene, cumene and ethyl benzene present in 0.5 mm thick sections of polystyrene before and after contact with the five extractants for 10 days at 60 °C showing that migration from the polystyrene occurs to a much smaller extent for the aqueous extractants than for the alcoholic or oily extractants. Fifty to 100 ppm extractable benzene in the polymer would render the polymer unacceptable for food packaging applications. Small concentrations of benzene have been detected occasionally in polystyrene. Considerably higher concentrations of extractable aromatics other than benzene can be tolerated in the polymer. Usually, however, in the case of food or beverage packaging grades of polystyrene, it is necessary to reduce the level of all aromatics to a very low level as, in addition to toxicity considerations, food tainting by migrated polymer volatiles is an additional factor governing the acceptability of a polymer in this field. 140

Determination of Monomers and Oligomers in Extractants

11.3 Determination of Acrylate Monomers and Oligomers in Aqueous and Fatty Extractants Lin and co-workers [6] have discussed recent advances in testing protocols for FDA compliant ultraviolet (UV) electron beam (EB) curable acrylate coatings, adhesives and inks used for food packaging applications. These workers review a comprehensive testing protocol utilising cell extraction method and liquid chromatography (LC) with a mass selective spectrometer for testing the above types of sample, in order to determine the suitability of the chemistry as part of the food packaging materials. The testing protocol helps to determine FDA compliance of the UV/EB curable chemistry by supporting the ‘no migration/no food additive’ statutory exemptions under FDA regulations (see Chapter 16). Examples are described, which demonstrate the suitability, advantages and selectivity of LC-mass spectroscopy (MS)/LC-MS-MS detection methods over the gas chromatography (GC)-MS method. Also demonstrated is the possibility of achieving FDA compliance with EB curable, acrylated chemistry, and to be able to maintain the same compliance with normal process variation, e.g., coat weight, curing voltage and curing dosage. Migration studies were performed by determining the amount of extractables from coated and cured samples through the cell extraction method using FDA migration testing protocols provided by the ‘FDA Center for Food Safety and Applied Nutrition’ [7]. Stainless steel extraction cells with Teflon spacers of 30 ml volume and 48 cm2 opening for the extracting solvent were used for the extraction test. A test sample, approximately 6.5 cm × 12.5 cm, was placed in the extraction cell with the side (direct or indirect) to be extracted facing the cavity where the extracting solvent would be placed (Chapter 1). The cell/sample was conditioned with the appropriate extraction solvent, temperature schedule, and dwell time as recommended in Tables 11.4 and 11.5 and as outlined in [8]. The cell was allowed to cool to room temperature before handling. In this study both extracting solvents were used and condition of use ‘E’ (Table 11.5) was selected.

Table 11.4 Simulating Solvents Food Types Simulating Solvent Aqueous, acidic, and low alcohol foods 10% Ethanol (in water) Fatty foods 95% Ethanol (in water) Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press

141

Additive Migration from Plastics into Foods

Table 11.5 Extraction Conditions Condition of Use Description

C Hot filled >65 °C

D Hot filled <65 °C

Extraction 100 °C for 66 °C for Temperature 30 min then 30 min then and 40 °C for 40 °C for Duration 10 days 10 days

E

F

Room temperature Refrigerated filled and storage stored 40 °C for 10 days

20 °C for 10 days

G Frozen storage

20 °C for 5 days

Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press

Identification and concentrations of migrants were determimed by both gas chromatography with a mass selective detector (GC-MS) (Agilent 6890 equipped with mass selective detector model 5973, column cooling accessory and a 30 m × 0.025 mm proprietary capillary column and conditions for adequate separation) and liquid chromatography with a mass selective detector (LC-MS-MS) (HPLC: Agilent 1090 system equipped with Micromass Quattro II triple Quad mass spectrometer with atmospheric pressure ionisation source; HPLC was operated using a solvent gradient mobile phase at 3 ml/min, 150 × 2.0 mm proprietary column and conditions; for MS-MS experiments, argon was used as collision gas at cell pressure of 50 Pa). The extracted solution for GC analysis has to be concentrated for detection sensitivity reasons either by extraction with another solvent or by room temperature evaporation through nitrogen purging. Internal standard naphthalene-d8 was added before the concentrating step for better, more consistent results. The extracted solution for LC-MS-MS analysis can be used directly after the internal standard, dipropylphthalate, is added with further concentrating. EB coatings were applied at a coating weight range of 0.7-1. 0 kg per 300 m2 (0.7-1.0 kg/ream) on generic oriented polypropylene (OPP) film with a ‘Little Joe’ ink proofer for screening study and cured with a lab EB curing unit from Energy Sciences, Inc. (ESI) (model: Electrocurtain CB175) at various kV and Mrad or with an offset gravure coater for the process variable study using a pilot coater from ESI at various kV and Mrad. EB adhesive was applied at a coating weight of 0.4 to 0.6 kg/ream for film to film laminates (48 gauge Melinex 813 to proprietary heat-seal film) using ESI’s pilot coater and laminator. 142

Determination of Monomers and Oligomers in Extractants Hartman [9] has been involved in the development of protocols using GC-MS by using the Short-Path ‘Purge & Trap’ thermal desorption technique. The focus of this initial method for UV/EB curable acrylate formulations was purely from off-odour and off-taste standpoints to determine the chemical components (from the acrylate type formulation) that contribute to the undesirable side effects. However, this thermal desorption method did not bring in the migration or extraction aspects of the FDA regulation which requires the use of food simulants to perform the necessary extraction. At a later date, cell extraction was used by Hartman [9] to develop a specific concentrating method for the extracted solution to allow for the analysis by the GC-MS method. However, detection by the GC method can only detect volatile compounds and not non-volatile compounds. Lin and co-workers [6] have elected to use HPLC with a triple Quad mass spectrometer as their primary analytical detection technique for the following advantages over the GC-MS method: • Detection of both volatile and non-volatile compounds. • Non-destructive separation method (no extreme heat involved in the injection and separation steps). • Extracted solution can be used ‘as is’ without concentrating steps. • Reduction of detection variation due to potential loss of ingredients in the extra concentrating steps. Comparison of the workflow for GC-MS and LC-MS-MS methods is shown in Figure 11.3. Lin and co-workers [6] analysed five acrylic monomers and four acrylic oligomers by both the GC-MS and LC-MS method. Figures 11.4 and 11.5 are spectra of five such monomers by GC-MS and LC-MS methods, respectively. One can see that for monomer #5 and #6 where volatility is very low, the GC-MS method could not effectively analyse them by showing a much lower abundance level and/or very high noise leyel (no response) while LC-MS had no problem identifying and analysing all five monomers. Figures 11.6 and 11.7 are spectra of four oligomers obtained from GC-MS and LC-MS methods, respectively. For oligomer #1 (standard bisphenol A epoxy diacrylate or ‘epoxy acrylate’) and oligomer #6, only the LC-MS method can detect them due to their low volatility. For oligomer #2 where there are lower molecular weight (MW) volatile portions and high MW portions, the GC-MS method can only detect the lower MW portions while the LC-MS method can identify both. For oligomer #7 (a polyurethane diacrylate), an isocyanate moiety was identified in the GC-MS analysis. This species was generated by thermal decomposition at the injection port, meanwhile, LC-MS analysis showed the spectrum of the oligomer. 143

Additive Migration from Plastics into Foods

Figure 11.3 Workflow comparison for GC-MS versus LC-MS methods. (Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press)

It can be seen that the GC-MS method would be likely to fail to properly detect potential migratable species such as low volatility monomers, higher MW oligomers and possibly miss out on certain compounds that decomposed during the analysis. These considerations explain why Lin and co-workers [6] utilised the ‘non-destructive’ LC-MS method for detection and identification of the potential migrants from UV/EB curable, acrylate type formulations. Multiple reaction monitoring (MRM) mode LC-MS analysis of a model EB coating formula (Formula #1: consisting of epoxy oligomer, trimethylol propane triacetate (TMPTA), tripropylene glycol diacrylate (TPGDA), and other necessary additives such as wetting agent and wax) is shown in Figure 11.8. Using the MRM detection mode, a migration study was performed with Formula #1 coated on a 50 gauge polypropylene film. The extraction results are listed in Table 11.6. Direct extraction means that the EB coated side is in ‘direct’ contact with the extracting solvent (food simulant) to test for the suitability for the direct food contact applications. Indirect extraction means that extracting solvent and EB coating are separated by the substrate (50 gauge OPP film), and thus the substrate is being tested to see if it can be claimed as a ‘functional barrier’ for the EB coatings for indirect food applications. 144

Determination of Monomers and Oligomers in Extractants

Figure 11.4 GC-MS spectra – full scans for five acrylic monomers. (Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press)

145

Additive Migration from Plastics into Foods

Figure 11.5 LC-MS spectra – total ion full full scan of five acrylic monomers. (Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press)

146

Determination of Monomers and Oligomers in Extractants

Figure 11.6 GC-MS spectra – full scan of four acrylic monomers. (Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press)

147

Additive Migration from Plastics into Foods

Figure 11.7 LC-MS spectra – total ion full scan of four oligomers. (Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press)

148

Determination of Monomers and Oligomers in Extractants

Figure 11.8 TPGDA in model coating Formula #1 with retention time of 6.6 minutes. Peak is expanded to show the major parent ion, m/z = 301, and a fragment ion (model EB coating consisting of epoxy oligomer, TMPTA, TDGDA, wetting agents and wax), m/z = l 13 as circled. (Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press)

The results again confirm that only the LC method could detect this non-volatile epoxy oligomer and that its migration through 50 gauge OPP film was hindered. The results also indicate that this model EB coating formulation, Formula #1, cured at the conditions specified is not suitable for food packaging materials (both direct and indirect with this specific OPP substrate) because of the high level of total migrants. A better formulation has to be developed and/or a better barrier material (for indirect application) has to be used. With the help of LC-MS-MS (operating at MRM mode) technique, Lin and co-workers were able to screen through various raw materials and optimise formulations to satisfy the ‘no migration/no food additive’ statutory exemption of FDA regulation. Extractable results using two of such EB curable formulas (Formula #2 and Formula #3) are listed in Table 11.7 along with those of model Formula #1 using both GC-MS and LC-MS detection methods. 149

Additive Migration from Plastics into Foods

Table 11.6 Ethanol (95%) cell extraction result of model EB coating formula #1 (epoxy oligomer TMPTA or TPGDA) Extraction type LC-MS-MS MRM mode GC-MS 806 ppb (No oligomer 98 ppb (No oligomer Indirect detected) detected) 5010 ppb (oligomer 300 ppb (No oligomer Direct detected) detected) Note: Coating was cured at 165 kV, 3 Mrad on OPP films at about 0.7-0.9 kg per 280 m2. Extraction conditions: 95% ethanol, 40 °C for 10 days though cell extraction (condition of use ‘E’: room temperature filled and stored) Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No. 48 [6]. © 2000, Tappi Press

Table 11.7 Comparison of 95% ethanol cell extractable results based on detection methods (GC method versus LC method) Formula Type Model Formula #1 Formula #2 Formula #3

Extractable (ppb) GC-MS

GC Result Pass/Fail

Indirect

300

Direct

98

Extraction Type

Extractable (ppb) LC-MS

LC Result Pass/Fail

SIR

MRM

Fail

720

806

Fail

Fail

4000

5010

Fail

Indirect

ND

Pass

ND

ND

Pass

Direct

ND

Pass

110

104

Fail

Indirect

ND

Pass

ND

ND

Pass

Direct

ND

Pass

ND

ND

Pass

Note: All formulas cured at 150 kV, 3 Mrad on OPP films at about 0.7-0.9 kg per 280 m2 Extraction condition: 95% ethanol, 40 °C for 10 days through cell extraction (Condition of use E: room temperature filled and stored). ND: not detected. Detection limit = 50 ppb (ppb: parts per billion based on 1.5 g of food contact per cm2 area). SIR: Selective ion recording Reproduced with permission from A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No.48 [6]. © 2000, Tappi Press

150

Determination of Monomers and Oligomers in Extractants Samples made from model formula #1 were determined to be non-FDA compliant with both GC-MS and LC-MS methods in both direct and indirect extraction tests. For samples made from formula #2, the results are quite different with the different methods. With results from the GC-MS method, the sample made from formula #2 could easily be determined as FDA compliant (through the ‘no migration’ exemption) for both direct and indirect applications, while results from the LC-MS method only allows its use for indirect food packaging applications under the ‘Functional Barrier Doctrine’ exemption. Samples made from Formula #3, on the other hand, are determined to be FDA compliant (through the ‘no migration’ statutory exemption) even with the more stringent LC-MS method. It can be seen that only with the LC-MS detection technique can one clearly determine the actual level of migration especially when there are non-volatile species involved.

References 1.

W.L. Bird and C.H. Hale, Analytical Chemistry, 1952, 24, 3, 586.

2.

G.W. Daues and W.F. Hamner, Analytical Chemistry, 1957, 29, 7, 1035.

3.

Federal Register, 2002, Volume 67, No.98, Rules and Regulations, US Food and Drug Administration, Washington, DC, USA, p.25, 724-25, 731.

5.

Plastics in Contact with Food: A Guide, 2nd Edition, British Plastics Federation, London, UK, 1996.

5.

P. Shapras and G.C. Claver, Analytical Chemistry, 1964, 36, 12, 2282.

6.

A. Lin, H. Gao, G. Wind and F. Wornick in Proceedings of the 2002 PLACE Conference, Boston, MA, USA, Tappi Press, 2000, Session 13, Paper No.48.

7.

FDA Center for Food Safety and Applied Nutrition, www.cfsan.fda.gov

8.

Section 11 D (c) Food Simulents, Appendix 11, Selected Migration Testing Protocols of Guidance for Industry. Preparation of Pre-Market Notification for Food Contact Substances, Chemistry Recommendations, US Food and Drug Administration, Washington, DC, USA, April 2002, http:/vm.cfsan.

9.

J.R. Hartman, unpublished work.

151

Additive Migration from Plastics into Foods

152

12

Analysis of Polymer Extraction Liquids Containing More Than One Migrant

Polymer formulations usually include one or more compounds such as antioxidants, secondary antioxidants, antistatic additives, light stabilisers, lubricants, plasticisers, stabilisers, slip and antiblock agents. In addition, the polymer and hence the extractant liquid might contain other substances not deliberately added such as unreacted monomers, residual polymerisation solvents and catalysts. The result is that practical extractants from such plastics can contain very low concentrations of several very different types of substances which may or may not mutually interfere with each other during these subsequent analyses and one or more of which it may be necessary to determine. The problem resolves itself into three stages. Firstly, the additives must be extracted from the extraction liquid in the form of an extract which is suitable for subsequent analysis. Frequently, extraction with diethyl ether or another low boiling organic solvent will achieve the required separation, certainly in the case of the aqueous and simple hydrocarbon polymer extractants. In addition to isolating the additives in the form of a suitable extract, this process will achieve a useful concentration factor of up to 100 fold in the level of additives present in the extract. Secondly, it is usually necessary to separate in this extract the additive or additives which it is required to determine from those for which analysis is not required, in order to avoid any analytical interference effects. Techniques such as thin-layer or column chromatography are particularly useful in this respect and are discussed in further detail next. Chromatographic techniques are of value in problems other than that of determining one or more additives in the presence of interfering substances. Thus, many polymer additives breakdown either during polymer processing or due to hydrolysis by aqueous extraction liquids. This facet of extractability testing is discussed further in Chapter 13. Thus, many phenolic antioxidants are partially oxidised in the polymer during extrusion at elevated temperatures and as the toxicity of both the original antioxidant and its oxidation product may be of interest from the points of view of their extractability from the polymer and their toxicity it will be necessary to analyse for both. In a further example polymers additives such as alkyl dialkanolamides are hydrolysed by aqueous extractions to a fatty acid and an dialkanolamine immediately they migrate from the polymer: RCON(CH2 CH2OH)2 + H2O = RCOOH + HN(CH2 CH2OH)2 153

Additive Migration from Plastics into Foods If it is required to ascertain the extent to which such hydrolysis of the additive has occured then, again, chromatographic techniques are amenable to the determination in the extractant of both the unchanged additive (RCON(CH2CH2OH)2) and its hydrolysis product (RCOOH). Finally, having extracted total additives from the extraction liquid and, if necessary, separated these into individual fractions by chromatography it is necessary to apply appropriate analytical techniques of sufficient sensitivity to the determination of the individual additives. Procedures for the solvent extraction of total polymer additives and their breakdown products from aqueous and simple hydrocarbon polymer extraction liquids and for the chromatography of these extracts are discussed next.

12.1 Preliminary Solvent Extraction of Gross Additives from Aqueous and Alcoholic Extractants In addition to using diethyl ether as the extraction solvent, depending on the partition coefficient of the additives between the extraction liquid and the solvent, other low boiling solvents might be suitable in certain instances, e.g., petroleum ether, methylene dichloride and benzene. The ether used for extraction can be purified by shaking 2 litres of ether with 300 ml of 30% aqueous sodium hydroxide in a separating funnel. The lower phase is rejected and the ether phase washed with water. Finally the ether is distilled from 20 g solid sodium hydroxide and stored in an amber bottle.

12.1.1 Extraction Test (Assuming the Extraction Test is Carried Out on 700/800 ml Scale) • Extractants: distilled water, 5% sodium carbonate and 5% citric acid. Transfer the hot contents of the polymer extraction tube into a 1 litre liquid extractor (upward displacement type). Wash the interior of the extraction tube with 25 ml of hot water and ether and transfer to extractor as discussed next. • Extractant: 50% aqueous ethanol. Transfer the contents of the extraction tube to a round bottomed flask. Wash the tube with 25 ml of hot water and 25 ml of ether and combine. Distill until the distillate has no odour of alcohol. Dilute the distillation residue up to 700 ml with water and transfer to a liquid-liquid extractor with ether washings as discussed next. 154

Analysis of Polymer Extraction Liquids Containing More Than One Migrant

12.1.2 Ether Extraction To the contents of the extractor (700/800 ml) add 100 g solid sodium chloride and stir to dissolve. Charge with ether and extract for 10-15 hours. To the ether extract add 3 g of anhydrous sodium sulfate. Shake, filter off the ether and wash the sodium sulfate with fresh ether into the receiving flask. Remove ether on a warm water bath, working the residue in stages into a 10 ml beaker with appropriate ether washings at each stage to avoid losses of residue. Finally, quantitatively transfer the residue to a 2 ml volumetric flask. This procedure was used to check the recoveries of dilauryl thiodipropionate (DLTDP) by ether extraction. The method was checked at the 15 and 75 ppm DLTDP level from 700 ml of the aqueous and alcoholic extraction liquids. DLTDP was determined in the extracts by infrared spectroscopy and by elemental analysis for sulfur as discussed in Chapter 4). Table 12.1 shows that the recoveries are in excess of 80%.

Table 12.1 Recoveries of DLTDP using different extratcants Extractant Recovery (%) Water 80-85 Ethyl alcohol, water 90 5% sodium carbonate 85-95 5% citric acid 70-85 Reproduced from Author’s own files

12.2 Separation of Individual Additives from Solvent Extract of Extractant Liquid Some examples of the application of these techniques to the determination of additives in extractants are discussed next.

(a) Mixture of Santonox R, Ionol CP and DLTDP An example of the application of silica-gel column chromatographic techniques to the separation of polymer additives will be described first. An extraction liquid containing the following three additives, Santonox R [4,4′ thio-bis(6-tert-butyl-meta-cresol)], Ionol CP (butylated hydroxytoluene), and DLTDP was used and it was necessary to determine the concentrations of two of them (Santonox R and Ionol CP). 155

Additive Migration from Plastics into Foods

Figure 12.1 Ultraviolet spectra of polypropylene additives; 200-400 nm. (Reproduced from Author’s own files)

Inspection of the ultraviolet (UV) spectra of synthetic cyclohexane solutions of each of these additives (Figure 12.1) shows that the Santonox R could be determined by direct UV spectroscopy with little or no interference from either Ionol CP or DLTDP by evaluation of the maximum occurring at 250 nm. However, Ionol CP could not be determined by direct spectoscopy of the extractant by evaluation of its peak at 276 nm as Santonox R also has a maximum at this wavelength. Interference by Santonox R in the determination of Ionol CP at 276 nm was overcome by means of a preliminary separation on a column of 100-200 mesh silica gel (containing 4% water). To isolate total additives from the aqueous extraction liquid it was saturated with sodium chloride and then extracted with diethyl ether, followed by extraction with chloroform. The chloroform was dried with sodium sulfate and 156

Analysis of Polymer Extraction Liquids Containing More Than One Migrant evaporated to dryness and the residue was dissolved in spectroscopic grade carbon tetrachloride. Percolation of this solution down a column of silica gel gave an effluent containing Ionol CP only, which could be determined by evaluation of its maximum occurring at 276 nm. Many useful separations of extractant constituents can be achieved by column chromatographic techniques. Generally, however, such separations are best achieved by thin-layer chromatography (TLC) or, in the case of volatile materials, gas chromatography (GC). Two different applications of the TLC techniques direct and indirect: are discussed next for the analysis of additive mixtures. It is first necessary to extract total polymer additives from the extraction liquid with a low boiling organic solvent and concentrate the extract to 2 ml as described earlier in this Chapter. TLC plate adsorbents (e.g., silica gel, alumina and so on) usually contain small amounts of substances which migrate with the development solvent along the plate towards the solvent front. Solvents such as methanol cause migration of adsorbent impurities almost completely to the solvent front, whereas non-polar solvents such as n-hexane do so to a lesser extent. These impurities adsorbed in the UV region of the spectrum but do not appear to absorb much in the infrared region; the impurities also produce a char when the plate was sprayed with sulfuric acid and heated to about 170 °C, indicating that they contain organic matter. The presence of these impurities in the adsorbent might interfere in procedures for the quantitative analysis of mixtures carried out by TLC. Such interference can usually be overcome by the pre-migration procedure in which the plate is migrated to the top with methyl alcohol which removes impurities out of the region where the analysis chromatogram will be subsequently developed (a second pre-migration with methanol might be desirable sometimes to complete removal of impurities). The plate is air-dried and then conditioned in the usual way prior to carrying out the analysis procedure. Plates which have been pre-migrated are quite adequate for carrying out quantitative TLC analysis of mixtures by either of the two procedures described next.

12.2.1 Indirect Setermination of Additives on the Thin-layer Plate Suitable spray reagents which produce coloured derivatives upon reaction with separated substances on a chromatoplate are not available for all the types of substances likely to occur in extractants. If an extractant contains a mixture of additives which have to be separated, and for some or all of which spray reagents are not available, then the following approach is recommended. 157

Additive Migration from Plastics into Foods

Figure 12.2 Chromatoplate of mixture of 1 mg of Santonox R, Ionol and dilaurylthiodipropionate. (Reproduced from Author’s own files)

To obtain sufficient material for subsequent analysis a portion of the extract of the extraction liquid (containing up to 1 mg of each component) is applied in a straight line along one side of the plate with a suitable applicator. After solvent development, the bands on the plate containing each additive are marked off. The position of these bands on the chromatograph can be calculated from the known retardation factor (Rf) values of the various sample components, as determined in a separate run carried out under identical conditions using a sulfuric acid reagent as detector. Figure 12.2 is a chromatoplate obtained under these conditions for a mixture of Santonox R, Ionol CP and DLTDP. The silica gel corresponding to each of these separate bands is then carefully scraped off from the plate and each transferred to a filter stick and eluted with a low boiling polar solvent such as ethyl alcohol (Figure 12.3) which quantitatively desorbs the additive from the gel. Solvent is then removed from each extract, which is diluted to 2 ml in a volumetric flask with a suitable spectroscopic solvent. These solutions are suitable for analysis by appropriate techniques as discussed elsewhere. It is advisable to check on the recovery obtained for each additive in these procedures by carrying 158

Analysis of Polymer Extraction Liquids Containing More Than One Migrant

Figure 12.3 Filtration apparatus for extracting separated additives from adsorbent. (Reproduced from Author’s own files)

out suitable control experiments covering all stages of the analytical procedure on known weights of each of the additives concerned. Polymer additives are usually colourless compounds and cannot, therefore, be determined by comparing the intensity of colour of the separated constituents with standard control solutions of known concentration, a technique which works excellently when determining organic dye stuffs. However, many polymer additives react to produce coloured reaction products when the plate is sprayed with a suitable reagent. Comparison of the intensity of the coloured spots obtained with those obtained for the standards enables an estimate to be obtained of the concentrations of various substances present. Figure 12.4, for example shows the thin-layer plates obtained upon development of 20 µl of an ether solution of a mixture of Santonox R, Ionol CP and DLTDP. It is seen that 2,6-dibromo-p-benzoquinone-4-chlorimine is an excellent spray reagent for showing up Santonox R and Ionol CP but that DLTDP does not show up distinctly using this reagent, however, it produces a yellow colour when sprayed with alkaline potassium permanganate solution (Figure 12.5). The unknown and standard spots can be quantified visually either by the bracketing technique or by densitometric scanning of the plate. These procedures are sufficiently sensitive to detect one microgram of 159

Additive Migration from Plastics into Foods

Figure 12.4 Chromatoplate of unknown mixture of Santonox R, Ionol and dilaurylthiodipropionate with standard calibration mixtures. (Reproduced from Author’s own files)

substance which, with suitable scaling-up of the analytical operations, is equivalent to determining additives at the 1 to 5 ppm level in polymer extraction liquids. Pre-migration of the adsorbent with methanol is not usually necessary when determining additives by the direct technique. In some cases, however, a suitable spray reagent is not available for a particular additive. In these cases an approximate estimate of the amount of the compound present can be obtained by spraying the plate with 20% aqueous sulfuric acid (followed by heating at 160/200 °C) and comparing the intensity of the charred spots with that obtained for a range of standard calibration solutions of the substance. In such instances it is, of course, necessary to pre-migrate the plate with methanol before it is used for the analysis, to remove organic impurities from the adsorbent layer. 160

Analysis of Polymer Extraction Liquids Containing More Than One Migrant

Figure 12.5 Chromatoplate of unknown mixture of Santonox R, Ionol and dilaurylthiodipropionate with standard calibration mixtures. (Reproduced from Author’s own files)

(b) Mixture of Butylated Hydroxy Toluene (BHT) Antioxidant and 2-Hydroxy-4-n-octoxybenzophenone Light Stabiliser in Extractants As an example of the thin-layer technique, a method is described next for the determination of both BHT and 2-hydroxy 4-n-octoxybenzophenone in the aqueous and alcoholic foodstuff simulent extractants of the British Plastics Federation. The additives are first separated from the extractant liquids and thereby concentrated by ether extraction as described earlier. As both of these substances have similar absorption maxima at 280 nm and 290 nm in the UV region, it is first necessary to separate them and this is achieved by TLC. The separated bands of additives are 161

Additive Migration from Plastics into Foods then removed from the plate and determined independently by UV spectroscopy. The overall recovery of additives by this procedure is in the order of 90-100%. It is first necessary, to work out the scaling of operations so that the desired analytical sensitivity and accuracy is achieved.

12.2.2 Scaling of Operations If it is assumed that in an extraction test 10-20 g of plastics is contacted with 700 ml of extraction liquid which is subsequently extracted with a low boiling solvent to remove additives and this extract is concentrated to 2 ml, i.e., all the additives which migrate from the original polymer are concentrated into 2 ml of solvent, then if each additive is originally present in the polymer at 0.1%, then if complete additive migration occurs, the extract will contain 10-20 mg of each additive per 2 ml. If however, only 10% additive migration occurs, then the extract will contain 1-2 mg of each additive per 2 ml. Knowing the extinction coefficients of BHT and 2-hydroxy-4-n-octoxybenzophenone it is possible to calculate the weight of each of these substances that must be present in the volume of test solution applied to the plate (Table 12.2).

Table 12.2 1% BHT

E = 86.3 (in ethanol) 1 cm 280 nm 1%

Substituted benzophenone

E = 527 (in ethanol) 1 cm 290 nm

Reproduced from Author’s own files

The minimum detectable concentrations are: BHT = 0.1/86.3 × 3 = 386 µg/100 ml i.e., 39 µg in 10 ml (3 cm cell) 162

Analysis of Polymer Extraction Liquids Containing More Than One Migrant Substituted benzophenone = 0.1/527 × 3 = 60 µg/100 ml i.e., 6 µg in 10 ml (3 cm cell) Thus a suitable volume of the 2 ml extracts for application to the plate is 0.1 ml (which contains 500-1000 µg of each additive if complete additive migration from the polymer has occurred and 50-100 µg of each if only 10% additive migration has occurred). The TLC procedure described next for separating the additives and recovering each component for UV spectroscopy is based on the assumption that only 10% additive migration occurs from the polymer. Quantities should be adjusted if additive migration differs appreciably from 10%.

12.3 Ether Extraction of Extraction Liquid As discussed earlier, the volume of final extract is adjusted to 2 ml. • Thin-layer chromatography - Calibration Prepare a calibration solution as follows: Accurately weigh 0.2 g BHT and 0.2 g of 2-hydroxy-4-n-octoxy benzophenone into a 100 ml graduated flask and dilute up to the mark with diethyl ether. Spot 0, 10, 20, 40, 60 and 80 µl portions of this calibration solution on to the base line of a 20 × 20 cm plate (coated with a 0.25 mm layer of silica gel GF254) in duplicate and treating these as ‘sample’, proceed as described under chromatography next (Section 1). Include blanks as described. Measure the UV absorptions of the calibration samples and blanks as described in Section 3 and calculate the net optical density for each sample. Plot calibration curves of net optical density against concentration of additive in the ethanol solution used for spectroscopy for both BHT and 2-hydroxy-4-n-octoxy benzophenone. • Procedure Preparation of standard solutions Solution I. Accurately weigh 50 mg of BHT into a 100 ml graduated flask and dilute up to the mark with ether. 163

Additive Migration from Plastics into Foods Solution II. Accurately weigh 50 mg of 2-hydroxy-4-n-octoxy benzophenone into a 100 ml graduated flask and dilute up to the mark with ether. 1. Chromatography a) Apply the following solutions side by side at the base line of a 20 × 20 cm TLC plate in the form of a spot or short band 1 to 2 cm long: i) 1 × 0.1 ml portion of solution I (BHT reference) ii) 1 × 0.1 ml portion of solution II (light stabiliser reference) iii) 2 × 0.1 ml portions of polymer extract. b) Develop an ascending chromatogram to a distance of 10 cm in a petroleum ether (100-120°):ethyl acetate (9:1) eluent in a tank saturated with solvent vapour. Remove the plate from the tank and allow the solvent to evaporate. 2. Removal of the additives from the plate a) Inspect the plate under a 254 nm UV light source and by means of a sharp pointer mark out the dark blue zones in the two sample chromatograms which correspond in Rf to the major constituents of the two reference chromatograms.

Approximate Rf values should be 0.75 for BHT and 0.60 for 2-hydroxy-4-noctoxy-benzophenone.



Also mark off duplicate blank areas equal in size and Rf to the spots in the sample chromatograms.

b) Carefully scrape off the plate the areas of silica gel marked and transfer each portion of gel separately to a sintered glass filter column. Extract the additives from the silica gel with absolute ethanol and make up in graduated flasks as follows:

Sample and blank of Rf - 0.75 dilute each up to 10 ml with absolute ethanol



Sample and blank of Rf - 0.60 dilute each up to 50 ml with absolute ethanol.

3. Ultraviolet Spectroscopy Measure the optical density of the duplicate sample and blank solutions in 3 cm silica cells against absolute ethanol in the reference beam, at the following wavelengths: BHT - 280 nm, 2-hydroxy-4-n-octoxy benzophenone - 290 nm. 164

Analysis of Polymer Extraction Liquids Containing More Than One Migrant • Calculations Calculate the mean value of the sample and blank optical densities and obtain the net optical density as follows: Net OD = (mean sample OD - mean blank OD) Determine the concentration of each additive in the ethanol by reference to the appropriate calibration graph and relate this to the weight of additive in the extraction liquid. Fichtner and Giese [1] have pointed out that a high-performance liquid chromatography (HPLC)/mass spectroscopy (MS) method is an effective method for achieving a comprehensive analysis of elutable components (extractables) from polymeric materials. When applied correctly, it was superior to the classical HPLC/ UV method and complemented the GC/MS method for the identification of thermally unstable, reactive and high molecular weight analytes (with molecular weights of about 100-4000 g/mol). The HPLC-MS method required fundamental and wideranging work to optimise the liquid chromatographic separation system and the ionisation conditions for the mass selective detection for an effective use of the ion trap method. The ion trap method does not need such detailed work and offered the possibility of forming MS/MS spectra with relatively high fragmentation from certain analytes. The MS/MS technology required the development of parameterspecific libraries of spectra. There is a similar problem with the coupling of HPLC and MS as for the GC-MS coupling. The solvent has to be selectively separated from the analytes, as only those ions that are in the gas phase in the mass analyser’s high-vacuum zone are detected. For routine equipment mainly two techniques have established themselves on the basis of reliability, sturdiness and ease of use: atmospheric pressure chemical ionisation (APCI) and electrospray ionisation (ESI) [2-5], (i.e., (a) and (b) next). a) Atmospheric pressure chemical ionisation (APCI)

In this process the analyte solution is evaporated with the help of nitrogen and charged on a corona needle. Protected by the surrounding solvent, a gentle positive or negative chemical ionisation of the analyte molecules takes place in the resulting plasma in the gas phase with little noticeable fragmentation. The aerosol then passes through a heated capillary into the mass analyser’s highvacuum zone. The APCI source can be used with the usual HPLC flow rates for chromatographic separation, but low eluate flow rate does have a favourable effect on the quality of the total ion current (TIC) chromatogram and the mass spectra. 165

Additive Migration from Plastics into Foods b) Electrospray ionisation

With the electrospray ionisation method a low eluate flow rate does have a favorable effect on the quality of the TIC and the mass spectra.

After these two processes have been run, the analyte is focused by means of octapoles and lenses and fed into the mass filter, such as, for example, an ion trap. In this segment, the solvent has been completely separated, leaving the ions isolated in the ion trap. After separation on the basis of their mass to charge ratio, the ions are accelerated to the detector (electron multiplier). Figure 12.6 provides a schematic representation of a mass spectrometer with ESI sprayhead and ion trap and liquid chromatography apparatus.

Figure 12.6 HPLC-MS (schematic view). (Reproduced with permission from Fichtner and Guise, Kautschuk und Gummi Kunststoffe, 2004, 57, 3, 116 [1])

Because of the lack of fragmentation in this gentle ionisation process, structure determination and identification of unknown analytes is not reliable. This disadvantage can be offset with the of so-called MS-MS technology. Figure 12.7 shows a chromatogram of a standard solution recorded with a Nucleosil C18 column (250 mm length × 4.6 mm id, particle size of 5 µm), and a gradient of water/ acetonitrile (1 ml/min flow rate). On this chromatogram, the diverse ‘response’ (ratio of analyte concentration and signal intensity) of the individual substances can be read in the UV and in the MS. The peak of the dimethyl phthalate, for example, is the most intensive peak in UV, whereas it can hardly be detected at all in the TIC chromatogram. 166

Analysis of Polymer Extraction Liquids Containing More Than One Migrant

Figure 12.7 Chromatogram (TIC and UV) of standard solution 1, measured with Nucleosil column and flow rate 1 ml/min. (Reproduced with permission from Fichtner and Guise, Kautschuk und Gummi Kunststoffe, 2004, 57, 3, 116 [1])

The reduction in the flow rate to 0.4 ml/min, in combination with a column of smaller diameter (Reprosil-Pur ODS-3 column, 5 µm particle size, 250 mm length and 3 mm id), brings about a marked reduction in background noise in the TIC chromatogram, allowing for the likelihood of an improved detection limit and clearer mass spectra. A corresponding chromatogram with adapted gradient elution is shown in Figure 12.8. With an excitation amplitude in the ion trap of approximately 40-50% (equipmentspecific value) the fragments 219 and 149 of dipentylphthalate, for example, could be detected with high sensitivity, as the spectrum in Figure 12.9 shows. 167

Additive Migration from Plastics into Foods

Figure 12.8 Chromatograms (TIC and UV) of standard solution 1, reduced flow rate 0.4 ml/min (Reprosil-Pur column). (Reproduced with permission from Fichtner and Guise, Kautschuk und Gummi Kunststoffe, 2004, 57, 3, 116 [1]) To check the purity of registered peaks in the HPLC chromatogram, it is possible to subtract spectra of individual peaks or individual peak segments from a given chromatogram. To this end the mass spectrum is subtracted from the complete chromatogram at the maximum of a peak. In the case of pure peaks, only the base line should remain after subtraction. This process is carried out using the example of standard solution 1. The TIC chromatograms before and after subtraction are shown in Figure 12.10.

References 1.

S. Fichtner and U. Giese, Kautschuk und Gummi Kunststoffe, 2004, 57, 3, 116.

2.

R. Willoughby, E. Sheehan and S. Mitrovich, A Global View of LC/MS: How to Solve your Most Challenging Analytical Problems, Global View Publishing, Pittsburgh. PA, USA, 1998, p.66.

3.

W.M.A. Niessen, Journal of Chromatography A, 1999, 856, 1-2, 179.

168

Analysis of Polymer Extraction Liquids Containing More Than One Migrant 4.

Electrospray Ionisation Mass Spectrometry: Fundamentals, Instrumentation and Applications, Ed., R.B. Cole, John Wiley & Sons, New York, NY, USA, 1997, p.3.

5.

W.M.A. Niessen, Liquid Chromatography - Mass Spectrometry, 2nd Edition, Marcel Dekker, New York, NY, USA, 1999, p.337.

Figure 12.9 MS, MS-MS, MS3 of dipentylphthalate (excitation amplitude: 40%). (Reproduced with permission from Fichtner and Guise, Kautschuk und Gummi Kunststoffe, 2004, 57, 3, 116 [1]) 169

Additive Migration from Plastics into Foods

Figure 12.10 Chromatogram (TIC) of standard solution 2 with positive and negative ionisation with and without subtraction. (Reproduced with permission from Fichtner and Guise, Kautschuk und Gummi Kunststoffe, 2004, 57, 3, 116 [1])

170

13

Determination of Additives and their Breakdown Products in Extractants

It is a fact that the analyst concerned with the analysis of additives in either synthetic extractants or foodstuffs may occasionally find that his problem is simple in that he has only to determine one substance without interferences and at an analysible concentration in a range of extractants. In practical circumstances, however, the problem is not generally as simple as this. Other additives or adventitious polymer impurities which have also extracted from the polymer into the extractant may be present and this frequently causes complications which the analyst has to circumvent. Also, of course, in many cases, analysis is required for more than one additive and this may necessitate the adoption of a separation technique prior to carrying out the final analysis as discussed in Chapter 12. A particular circumstance where such problems manifest themselves is in the analysis of extractants of polymers in which complete or partial degradation has occured either during polymer processing or during its service life, (e.g., exposure to sunlight) or during the extraction test itself, (e.g., hydrolysis of extracted additives by aqueous extractants). As such degradation products might themselves have toxicological properties it might be deemed to be necessary to carry out analysis for them as well as for the undegraded additive. Even if this is not considered necessary the analyst must be aware of any such degradation processes so that due allowance can be made for any interfering effects they may exert on the determination of the undegraded form of the additive or on the determination of any other substance which is to be determined in the extractant. In addition to these considerations, there is also the rather more estoeric possibility of reaction between the extracted additive or its degradation products with a component of the extractant liquid to form compounds of different toxicology to the original additive. This factor should particularly be taken into account when dealing with edible oil types of extractants where distinct possibilities exist for reaction between migrated additives and components of the oil. Examples of the possibility of a degradation of an additive and its effect on the design of the analytical procedure are supported next by means of an example, i.e., the antistatic additive lauric diethanolamide which has been used in foodgrade polyolefin and polystyrene formulations. Lauric diethanolamide upon migration 171

Additive Migration from Plastics into Foods from the plastic into aqueous extractant liquids hydrolyses quite readily to lauric acid and diethanolamine (DEA): C11H23CON(CH2CH2OH)2 + H2O = C11H23COOH + HN(CH2CH2OH)2 In fact any of this additive which extracts from a polymer into distilled water at 60 °C is hydrolysed fairly completely within a few days. The analytical problem, resolves itself, therefore, into the determination of traces of DEA degradation product in the presence of relatively small concentrations of lauric diethanolamide. A method for the determination of lauric diethanolamide and DEA in the aqueous and alcoholic extraction liquids of the British Plastics Federation (BPF) test conditions and in liquid paraffin is described next. These methods are considered here in some detail as they illustrate the amount of method development that is sometimes required to achieve the required analysis.

13.1 Determination of Lauric Diethanolamide and its Degradation Products in Aqueous Fatty Extraction Liquids Lauric diethanolamide is a fairly unreactive substance. A likely approach to the problem of determining low concentrations of this substance involved hydrolysing it to DEA and fatty acid and determining the former by a conventional iodometric method. The lauric diethanolamide content of the extractant could then be calculated from the amount of DEA produced upon complete hydrolysis. It is more appropriate to determine DEA rather than the fatty acid, as it was possible that the extractant could contain fatty acids originating from sources other than hydrolysis of the dialkanolamide, (e.g., free fatty acids or their metal salts or fatty acid esters, all of which might be present in the original polymer). The required sensitivity of the analytical procedure is indicated, for example, by the following considerations. Under the extractability test conditions proposed by the British Plastics Federation on thin films (<0.5 mm) of polymer containing 0.03% w/v of additive, each cubic centimetre volume of plastic is contacted with 20 cm3 of the extraction liquid. Thus, if all the additive present in the original polymer film migrates during the extraction test into the extraction liquid, at the end of the test this liquid will contain only approximately 15 ppm of additive. Consequently it was necessary to devise a procedure for determining extracted polymer additive in each of the extractants in amounts down to 3 ppm. To estimate lauric diethanolamide it is first necessary to hydrolyse it to DEA, which can then be determined by the periodic acid method. Tests to determine the conditions necessary to completely hydrolyse lauric diethanolamide (Table 13.1) showed that it hydrolyses quite slowly. In fact, a reflux period of approximately 20 hours with 0.03 N acid is needed to hydrolyse the dialkanolamide completely. The effect of increasing 172

14

14

30

60

13

13

13

Found

93

93

93

Recovery (%)

4.9

4.9

4.9

Titre difference (sampleblank) (0.01 N iodine ml)

31

31

31

Added

27

27

29

Found

87

87

94

Recovery (%)

-

10.3

-

Titre difference (sampleblank) (0.01 N iodine ml)

34

36

-

34

36 -

34

-

Found

36

-

Added

-

94

94

94

-

Recovery (%)

-

12.9

12.9

12.9

-

Titre difference (sampleblank) (0.01 N iodine ml)

Reproduced from Author’s own files

In this procedure the sample (100 ml) containing between 14 and 36 ppm DEA (1.4 to 3.6 mg) was reacted for various periods of time, between 20 and 60 minutes, with excess 0.01 M periodic acid in 0.03 N sulfuric acid medium. To this solution was added excess 0.02 N sodium arsenite and sodium bicarbonate. Unused arsenite was then back titrated with 0.01 N iodine solution using potentiometric titration using a platinum foil indicating electrode – calomel reference electrode. These results show that it is possible to determine DEA amounts down to about 3 ppm in the 100 ml of test solution, with a recovery of about 90% or higher of the amount added.

14

Added

20

Reaction time of DEA with 0.02 N periodic acid (min)

Table 13.1 Experimental conditions in the determination of diethanolamine (DEA) by the periodic acid method DEA (ppm) in 100 ml test solution – 0.03 N with respect to sulfuric acid

Determination of Additives and their Breakdown Products in Extractants

173

Additive Migration from Plastics into Foods

Table 13.2 Investigation of hydrolysis conditions required for the conversion of lauric diethanolamide* to DEA Recovery (% of added amount of lauric diethanolamide)** Sulfuric acid, Concentration of sulfuric acid present during hydrolysis of reflux time (h) lauric diethanolamide 0.03 N 0.05 N 0.10 N 1 29 32 42 2 38 49 64 3 52 60 80 4 93 5 79 100 6 70 100 9 82 16 96 * Using commercial lauric diethanolamide which had been twice recrystallised from methanol (assumed to be 100% lauric diethanolamide for the purpose of calculating results). ** 12 mg recrystallised lauric diethanolamide used in experiment Composition of original lauric diethanolamide: Component

% w/v

C11H23CON (CH2CH2OH)2

80.6

HN

CH2CH2OOCC11H23



7.1

CH2CH2OH C11H23CON

CH2CH2OOCC11H23



4.7

CH2CH2OOCC11H23 C11H23CON (CH2CH2OOCC11H23)2

0.0

Free diethanolamine

0.0

Free fatty acid

0.0

Water

7.6

Reproduced from Author’s own files

174

Determination of Additives and their Breakdown Products in Extractants the strength of sulfuric acid up to 0.1 N during reflux was investigated. At the end of the reflux the acid strength was adjusted back to 0.03 N by the addition of sodium hydroxide. Excess periodic acid was then added and the estimation continued as described previously. Because a 5 to 6 hour reflux with 0.1 N sulfuric acid is needed to hyrolyse lauric diethanolamide completely to DEA (Table 13.2), lauric diethanolamide was estimated in the distilled water extractants by adjusting the sample to 0.1 N with respect to sulfuric acid, refluxing for at least 6 hours, and then adjusting back to 0.03 N acid strength before determination of DEA by the periodic acid method.

13.1.1 Determination of Lauric Diethanolamide and DEA in the Distilled Water Extractant The results in Table 13.3 confirm that lauric diethanolamide is indeed hydrolysed by contact with distilled water for 10 days at 60 °C. To estimate free DEA, one portion of the extractant was analysed directly by the periodic acid procedure. The other portion of extractant was refluxed with 0.1 N sulfuric acid for 6 hours to hydrolyse

Table 13.3 Hydrolysis of lauric diethanolamide in distilled water during 10 days at 60 °C % of original lauric Lauric diethanolamide (g/100 ml test solution) diethanolamide addition Not Hydrolysed hydrolysed during during Added Found extraction extraction test test Degraded (B - C) × (C), i.e., Undegraded C × 100/A Total A Total B 100/A after acid (B – C) reflux 0.0008 0.0010 0.0017 0.0015 0.0017 0.0020 0.0022 0.0022 0.0040 0.0040 0.0017 0.0023 42.5 47.5 Reproduced from Author’s own files

175

Additive Migration from Plastics into Foods the dialkanolamide completely to DEA, which was then analysed by the periodic acid method, (i.e., total undegraded lauric diethanolamide plus degraded dialkanolamide. About 50% of the amount of lauric diethanolamide originally added was hydrolysed by heating in water for 10 days at 60 °C.

13.1.2 Determination of Lauric Diethanolamide and DEA in the Aqueous 5% Sodium Carbonate Extractant As lauric diethanolamide is extensively hydrolysed to DEA by distilled water, (i.e., heating for 10 days at 60 ºC) it would be expected that even more extensive hydrolysis of the additive might occur under the alkaline conditions prevailing in the case of the 5% w/v sodium carbonate extractant. The results in Table 13.4 show that lauric diethanolamide is almost completely hydrolysed to DEA and fatty acid in aqueous sodium carbonate when heated to 60 °C for 10 days.

Table 13.4 Hydrolysis of lauric diethanolamide occurring in 5% sodium carbonate during 10 days at 60 °C Lauric diethanolamide added to 200 ml 5% sodium % of original lauric carbonate (g) diethanolamide addition Not Hydrolysed hydrolysed during during Added Found extraction extraction test test Degraded (B - C) × (C), i.e., Undegraded C × 100/A Total A Total B 100/A after acid (B – C) reflux 0.0042 0.0042 Nil 105 Nil 0.0040 0.0042 0.0042 Nil 105 Nil Reproduced from Author’s own files

13.1.3 Determination of Lauric Diethanolamide and DEA in the 50% w/v Aqueous Ethyl Alcohol Extractant In the 50% w/v aqueous ethyl alcohol extractant, lauric diethanolamide is not completely hydrolysed to DEA when refluxed in the presence of 0.1 N sulfuric acid for periods of up to 8 hours, (i.e., low total recoveries obtained). 176

Determination of Additives and their Breakdown Products in Extractants

Table 13.5 Determination of DEA in 50% w/v ethyl alcohol:water extractant Volume Volume of Volume Reaction Concentration of 0.03 N water added of water time with of DEA in 100 sulfuric acid immediately added before periodic acid ml test solution added before prior to periodic acid (min) (ppm) periodic acid iodine back addition (ml) addition (ml) titration (ml)

DEA recovery (%)

30

15-40

0

0

0

94-104

15-60

15

0

0

100

96-104

30

40

100

0

0

91-99

30

40

0

100

0

48*-95

Method of Analysis: 100 ml of 50% w/v ethanol:water extractant adjusted to 0.03 N with respect to sulfuric acid, is added to 15 ml 0.01 M periodic acid and then left to react for 30 minutes. Saturated sodium bicarbonate (30 ml) and 25 ml 0.02 N sodium arsenite are added and then the mixture is left for 10 minutes. Potassium iodide (2 ml of 15%) and 10 g of solid sodium bicarbonate are added to the mixture and the arsenite is back titrated with standard 0.01 N iodine. *Low results were due to the presence of too much acid in the reaction mixture during reaction of excess periodic acid with sodium arsenite. Reproduced from Author’s own files

The results in Table 13.5 show that the low lauric diethanolamide recoveries must be due to a slowing down in the rate of hydrolysis of lauric diethanolamide to DEA, brought about by the presence of ethyl alcohol during the preliminary reflux with 0.1 N sulfuric acid. Experiments on synthetic solutions of lauric diethanolamide in the 50% w/v ethyl alcohol aqueous extractant showed (Table 13.6) (rows 1-5) that in the absence of ethyl alcohol a recovery of between 100% and 120% of the added amount of lauric diethanolamide is obtained by the periodic acid method following hydrolysis with either 0.1 N or 1 N sulfuric acid for 6 to 8 hours. However, in the presence of ethyl alcohol, the lauric diethanolamide recoveries obtained following hydrolysis with 0.1 N sulfuric acid (row 6) are only about two-thirds of the expected value, but when the acid strength during hydrolysis is increased to 1 N, a quantitative lauric diethanolamide recovery is obtained (row 7-8). Thus, the low lauric diethanolamide recoveries observed earlier can be overcome by increasing the sulfuric acid strength during hydrolysis from 0.1 to 1.0 N. 177

178

Concentration of lauric diethanolamide in original 200 ml extraction liquid (ppm)

8

20

17

69

69

Dilution of extraction liquid with distilled water

200 ml of distilled water extractant diluted to 250 ml

200 ml of distilled water extractant diluted to 250 ml

200 ml of distilled water extractant diluted to 250 ml

200 ml of distilled water extractant diluted to 250 ml

200 ml of distilled water extractant diluted to 500 ml

No.

1

2

3

4

5 200b

100a

100a

100a

100a

Volume of diluted extraction liquid used for determination of lauric diethanolamide (ml)

0

0

0

0

0

1.0

1.0

0.1

0.1

0.1

Concentration of ethyl alcohol Concentration in diluted of sulfuric extraction liquid acid in diluted during reflux extraction liquid with sulfuric acid during reflux (N) (% w/v)

200

100

100

0

0

Volume of distilled water added to extraction liquid immediately prior to iodine back titration (ml)

100

101

99

110

120

Recovery of added lauric diethanol-amide (%)

Table 13.6 Hydrolysis of lauryl diethanolamide by sulfuric acid in distilled water and in ethyl alcohol:water media (refluxed for 6 to 8 hours). The requirement for a higher concentration of acid for complete hydrolysis in the presence of alcohol (Experiment No.6)

Additive Migration from Plastics into Foods

69

69

200 ml of ethyl alcohol:water extractant diluted to 250 mlb

200 ml of ethyl alcohol:water extractant diluted to 500 ml

6

7

8 200c

100a

100a

20

40

40

1.0

1.0

0.1

200

100

100

101

102, 99

62, 66

As above but double quantities were used of periodic acid and all other reagents added subsequently.

Reproduced from Author’s own files

c

b

When 200 ml of 50% w/v ethyl alcohol:water extractant is diluted to 250 ml with water, then the final mixture contains 40% w/v ethyl alcohol (see row 7 in this table). A 100 ml portion of this solution is adjusted to 1 N with respect to sulfuric acid, refluxed to hydrolyse Ethylan MLD, and then adjusted back to 0.03 N acid strength by addition of sodium hydroxide prior to determination of DEA. Owing to its high alcohol content, this solution precipitates out some sodium sulfate which causes difficulties in the subsequent analysis for DEA. To overcome this, the original 200 ml of 50% w/v ethyl alcohol:water extractant is diluted to 500 ml with water (i.e., the final solution contains 20% w/v ethyl alcohol, see row 8 in this table). A 200 ml portion of this is adjusted to 1 N with respect to sulfuric acid and treated as previously described. This solution, upon addition to sodium hydroxide, does not precipitate out sodium sulfate and can be satisfactorily analysed for its DEA content.

a

100 ml extraction liquid adjusted to 0.1 N or 1.0 N with sulfuric acid and refluxed for 6 to 8 hours to hydrolyse lauric diethanolamide to DEA. Solution adjusted back to 0.03 N with respect to sulfuric acid by addition of sodium hydroxide. 15 ml of 0.01 M periodic acid added and left for 30 minutes. 30 ml of saturated sodium bicarbonate added and 25 ml of 0.02 N sodium arsenite and left for 10 minutes. 2 ml of 18% potassium iodide, 10 g solid sodium bicarbonate added, and back titrated excess arsenite and 0.01 N iodine.

40

200 ml of ethyl alcohol:water extractant diluted to 250 ml

Determination of Additives and their Breakdown Products in Extractants

179

Additive Migration from Plastics into Foods

13.1.4 Determination of Lauric Diethanolamide and Hydrolysis Products in the 5% w/v Aqueous Citric Acid Extractant The periodic acid method cannot be applied directly to the determination of DEA in the 5% citric extractant due to the fact that citric acid itself reacts with periodic acid and interferes in the analysis. It is, therefore, necessary to devise a method for removing citric acid from this extractant prior to the determination of DEA. By adding a slight excess of barium carbonate to the extractant, citrate ions are precipitated as barium citrate, and the insoluble salt then removed from the solution by centrifuging. However, it was found that the clear aqueous phase recovered by this procedure still contained sufficient citrate ions to interfere in the periodic acid method (due, presumably, to the presence of dissolved barium citrate, which is slightly soluble in water). To remove the last trace of barium citrate from the extractant, it was passed down a column of Amberlite IRA-400 ion exchange resin (in the chloride form). Citrate ions remained on the ion-exchange column and DEA passed through with the column effluent: Resin Cl + Ba citrate → Resin citrate + BaCl Blank estimations by the periodic acid method carried out on the 5% citric acid extractant following the barium carbonate precipitation and ion-exchange chromatographic separations described previously showed that these procedures had completely removed citrate ions from the extractant liquid. Table 13.7 shows the results obtained in applying the citrate removal procedure described previously to a synthetic solution of DEA (70 ppm w/v) in the 5% citric acid extractant. A recovery of 97% of the added amount of DEA was obtained in the first 500 ml of the ion-exchange column effluent, proving that the method is quantitative. In a further experiment a synthetic solution of lauric diethanolamide (40 ppm w/v) in the 5% citric acid extractant was heated for 10 days at 60 °C) during which time it was considered likely that citric acid would completely hydrolyse lauric diethanolamide to DEA. The citric acid removal procedure was then applied to the extractant and DEA determined in the column effluent by the periodic acid method. The amount of DEA obtained was that which was expected, assuming lauric diethanolamide had completely hydrolysed to DEA during the extraction test. Thus, in determining total lauric diethanolamide plus hydrolysed lauric diethanolamide in the 5% citric acid extractant, it is unnecessary to apply the preliminary reflux with 0.1 N sulfuric acid prior to analysis by the periodic acid method. 180

Determination of Additives and their Breakdown Products in Extractants

Table 13.7 Determination of lauric diethanolamide and DEA in 5% w/v aqueous citric acid extractant Weight of DEA found in various ion exchange column fractions (g) Fraction 1 (250 ml)

Fraction 2 (100 ml)

Fraction 3 (100 ml)

Fraction 4 (100 ml)

Total weight DEA (g) Found

Added

DEA recovery (%)

Synthetic solution of DEA (0.0140 g) in 5% citric acid (200 ml) 0.0134

0.0001

<0.0001

<0.0001

0.0135

0.0140

97

Synthetic solution of lauric diethanolamide (0.0080 g)* in 5% citric acid, (200 ml heated under BPF test conditions, i.e., 10 days at 60 °C) before application of barium carbonate precipitation and ion-exchange procedures. 0.0025

0.0004

<0.0001

-

0.0029 0.0029*

100

*The weight of lauric diethanolamide (0.0080 g) taken would, upon complete hydrolysis yield 0.0029 g diethanolamine. The fact that this weight of DEA was obtained in the column effluent indicated that lauric diethanolamide is completely hydrolysed by 5% citric acid during 10 days at 60 ºC, i.e., the sulfuric acid hydrolysis step can be omitted in the analysis. Reproduced from Author’s own files

13.1.5 Determination of Lauric Diethanolamide in the Liquid Paraffin Extractant Although complete or partial hydrolysis of lauric diethanolamide to DEA occurs with the four aqueous extractants, no such hydrolysis was expected in the case of a solution of lauric diethanolamide in the liquid paraffin extractant. Table 13.8 gives the results obtained in checks for the presence of DEA in synthetic liquid paraffin solutions of lauric diethanolamide after it had been heated for 10 days at 60 °C. At the end of the extraction test, the solution (20 ppm w/v lauric diethanolamide) was diluted with cyclohexane and divided into two portions for determination of DEA and undegraded lauric diethanolamide. DEA is water-soluble and can be determined by applying the periodic acid method to a water extract of the liquid paraffin/cyclohexane mixture. A small amount of DEA was found in the liquid paraffin extractant (less than 20% of the 20 ppm w/v of lauric diethanolamide present in the liquid paraffin before the extraction test, see Table 13.8, some of which was present as an impurity in the original batch of lauric diethanolamide used in this work. 181

Additive Migration from Plastics into Foods

Table 13.8 Determination of lauric diethanolamide* and DEA in the liquid paraffin extractant after 10 days at 60 °C Lauric diethanolamide Percentage of original lauric diethanolamide Added Found addition As lauric Undegraded As DEA C diethanolTotal A Total B DEA** C (B – C) × 100/A amide (B-C) × 100/A 0.0080 0.0096 0.0015 0.0081 19 101 0.0080 0.0097 0.0013 0.0084 16 105 *Using commercial lauric diethanolamide (which had not been recrystallised from methanol). Lauric diethanolamide assumed for the purpose of calculating results to be 100% pure lauric diethanolamide, although it is evident from these results that this assumption is not quite correct (total lauric diethanolamide recovery some 20% higher than expected). **Determined as diethanolamine, calculated as lauric diethanolamide. Reproduced from Author’s own files

To determine lauric diethanolamide, a further portion of the cyclohexane solution of liquid paraffin was refluxed with 0.5 N sulfuric acid until complete hydrolysis to DEA had occurred. Under these conditions, a twenty-four hour reflux period was needed to hydrolyse lauric diethanolamide completely. Determination of total DEA in the water extract by the periodic acid method (Table 13.8) showed that the lauric diethanolamide recovery was reasonably near to the theoretical value. Detailed procedures are given next for the determination of lauric diethanolamide and DEA in amounts down to 3 ppm in aqueous extractants and in liquid paraffin recommended by the British Plastics Federation.

13.1.5.1 Method Reagents Periodic acid (0.01 M): weigh out 2.28 g of periodic acid (HIO4.2H2O) to four decimal places, and dissolve in approximately 100 ml of distilled water. Transfer quantitatively to a 1 litre standard volumetric flask, dilute to the mark and mix thoroughly. 182

Determination of Additives and their Breakdown Products in Extractants Sodium arsenite (0.02 N): Dissolve 0.8 g of sodium hydroxide (Analar) and 1.0 g of arsenious oxide (Analar) in a minimum quantity of distilled water, warm in the beaker to obtain complete solution. Transfer quantitatively to a 1 litre standard volumetric flask and add 2.0 g of solid bicarbonate (Analar), swirl the flask until the solid is completely dissolved, dilute to the mark and mix thoroughly. Iodine solution (0.01 N): Dilute 100 ml of 0.1 N iodine solution to 1 litre with distilled water. Potassium iodide (15% aqueous): Dissolve 15 g of potassium iodide (Analar) in 100 ml of water. Sodium bicarbonate: saturated aqueous. Sodium bicarbonate (Analar): solid. Sodium hydroxide (5 N aqueous): Dissolve 200 g of sodium hydroxide (Analar) in water and dilute to 1 litre. Accurately standardise to three places of decimals. Sodium hydroxide (1 N aqueous): Accurately standardised. Sulfuric acid (10 N aqueous): Dilute 70 ml concentrated sulfuric acid (Analar; SG 1.84) to 250 ml with distilled water. Accurately standardise to three places of decimals. Sulfuric acid (5 N aqueous): Accurately standardised. Sulfuric acid (0.03 N aqueous): Accurately standardised. Hydrochloric acid (1 N): Accurately standardised. Ethanol: absolute. Cyclohexane: re-distil cyclohexane and discard the first and last 10% of the distillate. Methyl orange indicator (0.02%): aqueous. Barium carbonate (Analar): solid. Amberlite IRA 400-Cl ion exchange resin Starch (1%): aqueous. 183

Additive Migration from Plastics into Foods

Procedure • Transfer of liquids from extraction tubes (a) Take the tube used in the extractability test containing 200 ml extraction liquid, and warm to 60 °C in a water bath for a few minutes. Shake the tube to mix thoroughly and disperse any insolubles. Transfer the warm liquid from the extraction tube into a 250 ml standard volumetric flask. Rinse out the extraction tube with three separate 10 ml portions of hot ethanol (to dissolve any deposited solid adhering to the walls of the tube) and transfer these washings to the volumetric flasks. Allow the solution to cool to room temperature, make up to 250 ml with distilled water and mix thoroughly. • Analysis of distilled water extractant (b) Determination of the free DEA content of the distilled water extractant. Pipette 100 ml of the diluted extraction liquid referred to in Section (a) into a 250 ml conical flask. Add 6.5 ml of 0.5 N standardised sulfuric acid solution (i.e., sufficient acid to adjust the solution to approximately 0.03 N with respect to sulfuric acid). Into a second flask (blank) pipette 100 ml of 0.03 N aqueous sulfuric acid. (c) Into the sample and blank flasks pipette 15 ml of 0.01 M periodic acid solution and leave to react for 30 ± 1 minutes. Immediately add 30 ml of saturated sodium bicarbonate solution and 25 ml of 0.02 N sodium arsenite solution. Leave to react for 10 ± 1 minutes. Add 2 ml of 15% potassium iodide solution and 10 g of solid sodium bicarbonate, swirl to dissolve the sodium bicarbonate and add a few drops of starch indicator solution. Titrate the solution with 0.01 N iodine to the blue end-point (blue colour should persist for at least two minutes). Record the sample and blank iodine titrations. An alternative method of end-point detection is by means of a potentiometric titration procedure using either manual methods or an automatic titrimeter (see Figure 13.1). (d) Calculate the weight of free DEA present in the whole 250 ml of test solution. (e) Determination of lauric diethanolamide in the distilled water extractant. Pipette a further 100 ml portion of the 250 ml of extractant referred to in Section (a) into a 250 ml conical flask. Into a 250 ml flask (blank) pipette 100 ml of 0.03 N sulfuric acid. Pipette accurately 1 ml of standardised 10 N sulfuric acid solution into the sample flask and attach a vertical condenser. Reflux the sample solution for 8 hours on a hot plate to hydrolyse diethanolamide to DEA. At the end of this period allow to cool, and rinse down the condenser with sufficient distilled water to make the volume up to 120 ml. To this flask add accurately 6.2 ml standardised 1 N sodium hydroxide solution, (i.e., sufficient alkali to make the solution about 0.03 N with respect to sulfuric acid). Proceed as described in Section (c). 184

Determination of Additives and their Breakdown Products in Extractants

Figure 13.1 Potentiometric titration of sodium arsenite with standard iodine solution using platinum/calomel electrode and automatic titrator. (Reproduced from Author’s own files)

(f) Calculate the weight of lauric diethanolamide present in the whole 250 ml of test solution. • Analysis of 5% sodium carbonate extractant (g) Determination of free DEA content of the sodium carbonate extractant. Transfer the 5% aqueous sodium carbonate extractant from the extraction tube to a 250 ml volumetric flask with distilled water, as described in Section (a). Pipette 25 ml of this solution into a 100 ml conical flask, add a few drops of methyl orange indicator solution, and titrate with standardised normal hydrochloric acid solution to the pink coloured end-point. Calculate the normality of the sodium carbonate solution (1 litre of 1 N hydrochloric acid is equivalent to 1 litre of 1 N sodium carbonate solution: 53 g per litre, at the methyl orange end-point). (h) Pipette 100 ml of the extraction liquid into a 250 ml conical flask and add, from a graduated pipette, a calculated volume of accurately standardised 10 N sulfuric acid (sufficient to neutralise the sodium carbonate present and make the final volume of solution 0.03 N with respect to sulfuric acid - see Section (g)). Into a second (blank) flask, pipette 100 ml 0.03 N sulfuric acid. Proceed as described in Section (c). Calculate the weight of DEA present in the whole 250 ml of test solution. 185

Additive Migration from Plastics into Foods (i) Determination of lauric diethanolamide in 5% sodium carbonate extractant. Pipette 100 ml of the extraction liquid into a 250 ml conical flask and add, from a graduated pipette, a calculated volume of accurately standardised 10 N sulfuric acid sufficient to neutralise the sodium carbonate present and make the final volume of solution 0.1 N with respect to sulfuric acid (see Section (g)). Into a second (blank) flask pipette 100 ml 0.03 N sulfuric acid. Attach a condenser to the sample flask and reflux for 8 hours to hydrolyse lauric diethanolamide to DEA. Allow the flask to cool, and rinse down the condenser with sufficient distilled water to make the volume up to 120 ml. To this flask add accurately 6.2 ml standardised 1 N sodium hydroxide solution, (i.e., sufficient alkali to make the solution about 0.03 N with respect to sulfuric acid). Proceed as described in Section (c). Calculate the weight of lauric diethanolamide in the whole 250 ml of test solution. • Analysis of liquid paraffin extractant (j) Transference of liquid paraffin from extraction tube. At the end of the BPF extraction test, remove the extraction tube from the constant temperature bath, mix the contents well, and clean the outside of the tube thoroughly with a cloth. With a pair of clean tongs remove the sample from the tube and allow the liquid paraffin adhering to the sample to drain back into the tube. Transfer the extractant to a 500 ml standard volumetric flask, rinsing out the extraction tube with a 5 × 60 ml portion of warm cyclohexane, and add these washings to the contents of the 500 ml flask. Allow the solution to cool to room temperature, dilute to the mark with cyclohexane and mix thoroughly. (k) Determination of the free DEA content of the liquid paraffin extractant. Measure 200 ml of the 500 ml of cyclohexane solution of liquid paraffin into a 500 ml separating funnel, wash this solution with a 4 × 50 ml portion of 0.03 N sulfuric acid and run these washings into a 500 ml conical flask. Into a second (blank) flask run 200 ml of 0.03 N sulfuric acid. (l) Into the sample and blank flasks run 30 ml of 0.01 M periodic acid and leave to react for 30 ± 1 minute. Immediately add 60 ml of saturated sodium bicarbonate solution and 50 ml of 0.02 N sodium arsenite solution. Leave to react for 10 ± 1 minute. Add 4 ml of 15% potassium iodide solution and 20 g solid sodium bicarbonate, swirl to dissolve and add a few drops of starch indicator solution. Titrate the solution with 0.01 N iodine to the blue end-point (blue colour should persist for at least two minutes). Record the sample and blank iodine titrations. (m) Calculate the weight of free DEA present in the whole 500 ml of cyclohexane test solution. 186

Determination of Additives and their Breakdown Products in Extractants (n) Determination of lauric diethanolamide in the liquid paraffin extractant. Measure a further 200 ml portion of the 500 ml of cyclohexane solution of liquid paraffin into a 500 ml conical flask and accurately pipette it into 12 ml of 0.5 N sulfuric acid solution. Place the flask on a magnetic stirrer/hot plate and agitate the solution by means of a stirrer bar. Attach a condenser to the flask and reflux for 24 hours. Allow the solution to cool, wash down the condenser with a few millilitres of distilled water, and run the cyclohexane and aqueous phase into a 500 ml separating funnel. Carefully rinse the reflux flask with 50 ml of distilled water and 20 ml of cyclohexane, and transfer these washings to the separator funnel. Wash the organic layer with several 50 ml portions of distilled water and run the water washings into a 250 ml conical flask using a total of 200 ml of water for the extractions, i.e., the final extract is 0.03 N with respect to sulfuric acid. Into a second (blank) conical flask run 200 ml of 0.03 N sulfuric acid. Proceed as described in Section (l). Calculate the weight of lauric diethanolamide present in the whole 500 ml of cyclohexane test solution. • Analysis of 50% w/v ethanol:water extractant (o) Take the tube used in the extractability test and warm to 60 °C in a water bath for a few minutes. Shake the tube thoroughly, mix and disperse any insolubles. Transfer the warm liquid from the extraction tube into a 500 ml volumetric flask. Rinse out the tube with three separate 10 ml portions of hot ethanol (to dissolve any deposited DEA adhering to the walls of the tube) and transfer these washings to the volumetric flask. Allow to cool to room temperature, make up to 500 ml with distilled water and mix thoroughly. (p) Determination of free DEA in 50% w/v ethanol:water extractant. Measure 200 ml of the diluted extractant into a conical flask and accurately pipette in 13 ml of standardised 0.5 N sulfuric acid from a graduated pipette, (i.e., sufficient acid to adjust the solution to 0.03 N with respect to sulfuric acid). Into a second (blank) flask pipette 200 ml of 0.03 N sulfuric acid. Proceed as described in Section (l).

It will be observed that, under the previous conditions, the 20 g solid sodium bicarbonate added immediately prior to the iodine titration does not always completely dissolve (due to the presence of a high concentration of ethyl alcohol). Solubilise the sodium bicarbonate at this stage of the analysis by adding 100 ml of distilled water to the sample and the blank solutions immediately prior to the iodine titration. Calculate the weight of free DEA present in the whole 500 ml of test solution.

(q) Determination of lauric diethanolamide in 50% w/v ethanol:water extractant. Accurately measure 200 ml of the 500 ml extractant referred to in Section (o) 187

Additive Migration from Plastics into Foods into a 500 ml conical flask. Into another (blank) flask measure 200 ml of 0.03 N sulfuric acid. Into the sample flask accurately pipette 22.2 ml of standardised 10 N sulfuric acid (i.e., diluted solution is 1 N with respect to sulfuric acid). Attach a condenser to the flask and reflux for 8 hours to hydrolyse lauric diethanolamide to DEA. Allow the flask to cool, and rinse down the condenser with a few ml of distilled water. To the sample solution add a few drops of methyl orange indicator solution, and titrate with standardised 5 N sodium hydroxide solution to the yellow-coloured end-point (the theoretical titration of 5 N sodium hydroxide is 44.4 ml). Then add accurately, by pipette, sufficient standardised 10 N sulfuric acid to convert the solution to 0.03 N with respect to sulfuric acid (the theoretical addition of 10 N sulfuric acid is 0.8 ml). Proceed as described in Section (l). If the 20 g addition of sodium bicarbonate made immediately before the iodine titration of the sample does not completely dissolve (due to the presence of a high concentration of ethanol), add 100 ml distilled water to solubilise the sodium bicarbonate before carrying out the iodine titration. Calculate the weight of lauric diethanolamide present in the whole 500 ml of test solution. • Analysis of 5% w/v citric acid extractant (r) Transfer the 5% aqueous citric acid extractant from the extraction tube into a 250 ml volumetric flask as described in Section (a). Only free DEA need be determined in this extractant as it has been shown that lauric diethanolamide is completely hydrolysed in this medium to DEA during the course of 10 day extractability test at 60 °C. (s) Removal of a major proportion of citric acid from the extractant. Into a 250 ml centrifuge bottle, accurately pipette a 120 ml portion of the 250 ml citric acid extractant referred to in Section (r). Into a second (blank) 250 ml centrifuge bottle, measure 100 ml 5% citric acid extractant which has not been in contact with the plastic sample. To each bottle add 8 ± 0.l g (i.e., 20% to 30% excess) of solid barium carbonate. Stir for one hour to enable the barium carbonate to convert citric acid into insoluble barium citrate.

Centrifuge the bottles for one hour at 2,500 rpm to separate an upper clear phase from the settled barium carbonate/barium citrate layer. Remove the bottles from the centrifuge and carefully pipette off as much as possible of the clear upper phases without disturbing the settled solid phase. Measure the volume of clear liquid recovered (it should be possible to recover 90 to 100 ml of the undiluted citric acid extractant liquid at this stage of the analysis) and filter the clear phase into two 250 ml separatory funnels. Wash through the filter papers into the separatory funnels with 50 ml distilled water.

188

Determination of Additives and their Breakdown Products in Extractants (t) Removal of the last trace of citric acid from extractant by ion exchange chromatography. Prepare an ion exchange column of Amberlite IRA 400 (Cl) as follows: slurry 130 g Amberlite IRA-400(Cl) resin in 200 ml distilled water and completely transfer it into an 45 cm chromatographic column (10 mm diameter) fitted at the lower end with a No. 2 porosity sintered disc and stopcock. Allow the column to drain until the water level is just above the level of the resin.

Swirl to dissolve the sodium bicarbonate and add a few drops of starch indicator. Titrate the solutions with 0.01 N iodine solution to the blue end-point (blue colour should persist for at least two minutes).



The iodine titration obtained with flask (B) should be identical to the blank titration obtained with flask (C). This confirms that citric acid has been completely removed from the citric acid blank solution by barium carbonate treatment and ion-exchange chromatography. From the iodine titrations obtained with the sample flask (A) and the reagent blank flask (C) calculate the weight of DEA present in the whole 250 ml of original test solution (Section (r)).

13.1.5.2 Calculation Determination of free DEA in extraction liquid Weight (g) of DEA present in volume of extraction liquid in contact with plastic is calculated by:

(T − T ) × t × 105 × V S

B

2

103 × 4 × V1

g diethanolamine

where V2 = volume (ml) of extraction liquid in contact with plastic during extractability test

V1 = actual volume (ml) of above extraction liquid represented by portion of sample used in periodic acid analysis



TS = back titration (ml) of iodine obtained with sample



TB = back titration (ml) of iodine obtained in reagent blank determination, and



f

= normality of iodine solution. 189

Additive Migration from Plastics into Foods Weight (g) of lauric diethanolamide present in volume of extraction liquid in contact with plastic is calculated by:

(T − T ) × f × 105.1 × V S

B

3

10 × 4 × V4 3

=

A×M 105.1

where V3 = volume (ml) of extraction liquid in contact with plastic during extractability test

V4 = actual volume (ml) of above extraction liquid represented by portion of sample used in periodic acid analysis



TS = back titration (ml) of iodine obtained with sample



TB = back titration (ml) of iodine obtained in reagent blank determination



f



A = weight (g) of DEA present in volume of extraction liquid in contact with plastic, (see previously), and



M = molecular weight of lauric diethanolamide.

= normality of iodine solution

The data in Table 13.9 indicate that, after it had migrated from the polyethylene film, lauric diethanolamide extensively hydrolysed to DEA, especially in the case of the 5% sodium carbonate and 5% citric acid extractants. Also, between 60% and 100% of the original lauric diethanolamide content of the film migrated from the film into the extraction liquid during ten days exposure at 60 °C. The effect of carrying out the extraction test at room temperature instead of at 60 °C is illustrated in Table 13.10. The results show that reduction of the temperature during the extraction test reduces the rate of migration of lauric diethanolamide from the film into the extraction liquid. The extracted additive, however, is still completely hydrolysed to DEA , even when the test is carried out at 20 °C. Extractability tests were carried out on injection moulded polystyrene cups with a nominal wall thickness of 0.02 inch containing 1.5% lauric diethanolamide as an antistatic additive. Tests were run for 10 days at 60 °C and total undegraded plus degraded lauric diethanolamide (i.e., DEA) was determined in each extractant and calculated as grams lauric diethanolamide extracted per 4000 cm2 of plastic surface. 190

6.3, 6.3

Distilled water

5% Aqueous sodium carbonate

6.2, 6.2

Liquid paraffin

Calculated as lauric diethanolamide (b) Calculated as DEA

0.027, 0.028 0.010, 0.010 0.019, 0.020

0.032, 0.035

0.00, 0.00 0.00, 0.00

0.00, 0.00

0.00, 0.00

0.026, 0.023 0.010. 0.008 0.019, 0.017

0.029 0.006 0.002 0.004 0.009, 0.006 0.013, 0.017 0.005, 0.006 0.009, 0.012

0.00, 0.00

0.033

0.025

0.035 0.023

0.028

0.018

0.028

0.028

0.028 0.028

0.028

0.028

Nil

100

Approx. 70

100

Approx. 30

Total added % of lauric Lauric acid, g Total concentration diethanolextracted per determination of lauric amide of 100 cm3 film (calculated as diethanolextraction lauric amide in liquid diethanoloriginal hydrolysed to amide), g unextracted DEA at end Calculated extracted per film, g present of 10 days at 100 cm3 film per 100 cm3 60 °C, (b) × (a) + (b) 100/(a) + (b) film

0.012, 0.012 0.005, 0.007 0.002, 0.002 0.004, 0.005

Lauric diethanolamide, g extracted per 100 cm3 film (a)

Reproduced from Author’s own files

6.3, 7.8

5% Aqueous citric acid

7.7 6.5

6.8, 6.8

Extractant

50% Aqueous ethanol

Volume of plastic film contacted with 200 ml extractant (ml)

DEA, g extracted per 100 cm3 film

Extractability data

Table 13.9 Extractability of lauric diethanolamide from low-density polyethylene by BPF extractants Determination of Additives and their Breakdown Products in Extractants

191

192

6.2

90 days at 20 °C 0.000, 0.000

0.000, 0.000

Lauric diethanolamide, g extracted per 100 cm3 film (a)

Reproduced from Author’s own files

6.3

10 days at 60 °C

Extractability test conditions

Volume of plastic film contacted with 200 ml of 5% sodium carbonate (ml)

0.020, 0.016

0.027, 0.028

Calculated as lauric diethanolamide (b)

0.007, 0.006

0.010, 0.010

Calculated as DEA

DEA, g extracted per 100 cm3 film

0.018

0.028

Total determination (calculated as lauric diethanolamide), g extracted per 100 cm3 film

Extractability data

0.028

0.028

65

100

Total % of total concentration original lauric of lauric diethanoldiethanolamide amide in content of original un film extracted extracted during film, g extraction present per test 100 cm3 film

Table 13.10 Extractability of lauric diethanolamide from polyethylene into 5% aqueous sodium carbonate at 20 °C and 60 °C.

Additive Migration from Plastics into Foods

Determination of Additives and their Breakdown Products in Extractants The results of these tests are given in Table 13.11. The total lauric diethanolamide content of the various extraction liquids at the end of the test was between 3 and 13.5 ppm w/v, and maximum extraction occurred in the case of the liquid paraffin extractant, 5% sodium carbonate and 50% ethanol:water. Comparison under Set 1 and Set 2 of results obtained for the 5% sodium carbonate and liquid paraffin extractants in tests on polystyrene containing 1.5% lauric diethanolamide shows that two to three times as much DEA and lauric diethanolamide, respectively, was present in extracts obtained in the second test (Set 2), compared with those obtained in the first test (Set 1). As expected, a perceptibly higher concentration of lauric diethanolamide (or its hydrolysis product) was present in the extracts obtained for polystyrene containing 2.0% lauric diethanolamide compared with polystyrene containing 1.5% lauric diethanolamide. Finally, the extractability tests in 5% sodium carbonate and liquid paraffin were repeated several months later, on different samples of the polystyrene containing 1.5% and 2.0% lauric diethanolamide (Set 3), and the results obtained were very similar to those shown under Set 2. The extractability test procedure can be quite reproducible when carried out on different occasions and on different batches of the plastic with a particular additive formulation. It is inadvisable to rely on the results obtained in a single extractability test on a particular polymer formulation. Extractability tests should be repeated over a period of time on different batches of the polymer made in typical manufacturing conditions. The results quoted in Tables 13.9 and 13.10 show that lauric diethanolamide is almost completely extracted from 0.05 mm thick low-density polyethylene film into each of the five BPF extractants during 10 days exposure at 60 °C. Complete additive extraction would not necessarily occur in the case of thicker polyethylene films. Also, it has been shown that in the aqueous extractants most, or all, of the lauric diethanolamide extracted from polyethylene (Tables 13.9 and 13.10) or polystyrene (Table 13.11) is then hydrolysed to DEA and fatty acid, although, of course, this does not occur in the liquid paraffin extractant. The BPF liquids are intended to simulate various types of foodstuffs and beverages and it seems likely that the hydrolysis of lauric diethanolamide to DEA, referred to above, could also occur in some foodstuffs. Such hydrolytic decomposition of additives might also occur in the case of other types of additives. It has also been shown that lauric diethanolamide is hydrolysed fairly rapidly by human gastric juice at body temperature. Thus, it seems that even if lauric diethanolamide which has been extracted from a plastic into a foodstuff is not hydrolysed during storage in the container, it could still be hydrolysed following ingestion. 193

194

200

200

200

200

200

200

200

200

200

200

200

(b)

5% Sodium carbonate (a)

(b)

50% Ethanol:water (a)

(b)

(c)

(d)

5% Citric acid (a)

(b)

Liquid paraffin (a)

(b)

100

100

97

109

100

104

100

100

100

100

100

100

Total surface area of plastic sample (both sides) (cm2)

Reproduced from Author’s own files

*Based on results in Table 13.9

(a) – set 1, (b) – set 2, (c) – set 3, (d) – set 4

200

Distilled water (a)

Extractant

Volume of extractant (ml)

2.46

2.58

2.42

3.73

2.65

2.59

2.61

2.44

2.67

2.68

2.64

2.65

Weight of plastic sample (g)

13.5 ± 0.5

13.5 ± 0.5

2.9 ± 0.5

2.9 ± 0.5

10.1 ± 0.5

5.1 ± 0.5

6.0 ± 0.5

4.5 ± 0.5

6.0 ± 0.5

7.5 ± 0.5

5.5 ± 0.5

3.0 ± 0.5

Lauric diethanolamide content of 200 ml extraction liquid (ppm, w/v)

0.108 ± 0.004

0.108 ± 0.004

0.024 ± 0.004

0.022 ± 0.004

0.081 ± 0.004

0.040 ± 0.004

0.048 ± 0.004

0.036 ± 0.004

0.048 ± 0.004

0.060 ± 0.004

0.044 ± 0.004

0.024 ± 0.004

Average 0.108 ± 0.004

Average 0.023 ± 0.004

Average 0.051 ± 0.004

Average 0.054 ± 0.004

Average 0.034 ± 0.004

Weight of lauric diethanolamide extracted per 4000 cm2 of plastic surface (g)

Nil

100

Approximately 70

100

Approximately 30

% of DEA content of extraction liquid hydrolysed to DEA at end of 10 d period at 60 °C*

Table 13.11 Extractability of lauric diethanolamide from injection moulded polystyrene cups containing 1.5% DEA

Additive Migration from Plastics into Foods

Determination of Additives and their Breakdown Products in Extractants

13.2 Polymeric Plasticisers Further studies of the effect of in vitro hydrolysis by digestive fluid simulents on the hydrolysis of polymer additives migrating into food were conducted by Hamdani and co-workers [1]. These workers assessed the safety of polymeric plasticisers capable of migrating into food by an in vitro study of the hydrolysis of poly(1,2-propylene adipate) by such fluids. A high extent of hydrolysis was obtained with freshly prepared intestinal fluid solutions. High-performance size exclusion chromatography (HP-SEC) analysis indicated that the bulk plasticiser, i.e., poly(1,2-propylene adipate) completely disappeared and that low molecular weight oligomers were formed within four hours. Hydrolysis, as was shown by gas chromatography (GC), did not result in a significant conversion to the free monomers, like adipic acid. Measurements by nuclear magnetic resonance (NMR) indicated that the cleavage selectively occurred at primary ester linkages. Fractionation of the hydrolysis products on silica gel gave six oligomeric fractions, which were recovered and analysed. The absolute molecular weight of the plasticisers and the average molecular weight of the hydrolysis products were evaluated using proton NMR (1H-NMR). Hydrolysis did not take place (<2%) under simulated gastric and saliva conditions. The evaluation made by the EU Scientific Committee for Food (SCF) uses a risk assessment approach. The potential danger represented by consumer ingestion of these food contaminants is considered, and the amount of toxicity data that is required depends on the level of exposure. The following criteria are taken into account by the SCF [2]: 1. If the migrant has a molecular weight higher than 1000 g/mol, SCF considers that there is little absorption in the gastrointestinal tract, thus, in principle, no toxicological data are required for the substance itself. Therefore, for polymeric additives and oligomers, only the fraction with a molecular weight below 1000 g/mol is relevant for safety evaluation. 2. If migration is shown to be lower than 5 mg/kg of food in worst case situations and if the substance is not mutagenic, a reduced compendium of toxicological information is required. 3. If the substance hydrolyses in digestive fluids, the toxicological information available on hydrolysis products may be taken into account. In favorable situations, when hydrolysis is complete and the hydrolysis products are already authorised, no other information is needed. The European Commission recently defined three simulants to be used for digestive fluids [2], which are saliva, gastric, and intestinal fluid simulants. 195

Additive Migration from Plastics into Foods Saliva fluid stimulant [2] was NaHCO3 (4.2 g), sodium chloride (0.5 g), and potassium carbonate (0.2 g) dissolved in distilled water with the volume adjusted to 1 litre. The pH of the solution was 8.9 ± 0.1. The gastric fluid stimulant [2] was a 0.07 M hydrochloric acid solution (pH 1.2 ± 0.1) prepared from a 0.1 M standard solution. The intestinal fluid stimulant [2] consisted of a solution of pancreatin. First potassium dihydrogen phosphate (KH2PO4) (6.8 g) was dissolved in distilled water (250 ml) in a 1 litre volumetric flask. Then 0.2 M sodium hydroxide (190 ml), sodium taurocholate (0.5 g), and distilled water (around 400 ml) were added. A solution of pancreatin (10 g, USP) was prepared separately by gradually adding the enzyme into distilled water (150 ml) to avoid the formation of lumps, and it was then transferred into the volumetric flask with gentle shaking. The pH was adjusted to 7.5 ± 0.1 with sodium hydroxide (0.2 M), and the resulting solution was diluted to 1 litre with distilled water. The plasticiser was allowed to react under exposure to simulated gastric juice. The analysis of the HP-SEC chromatograms showed that at low concentration the bulk plasticiser poly(1,2-propylene adipate) was recovered intact while the low molecular weight constituents were hidden by impurities from the blank. At high concentrations there were no significant differences between the chromatographic profiles of acetylated poly(1,2-propylene adipate) before and after hydrolysis. The same results were observed with the addition of pepsin and mucin as catalysts [3]. These results were confirmed by 1NMR, because no free hydroxyl signal of the endchain groups could be detected, even after prolonged exposure (3 days). The GC analysis showed that there was no detectable formation of free adipic acid. Hydrolysis under saliva conditions gave identical results - the polymeric plasticiser was recovered unchanged. It could therefore be deduced that under gastric and saliva conditions the hydrolysis of the polymeric plasticiser, acetylated poly(1,2-propylene adipate) was insignificant.

13.3 Polybutylene Adipate, Poly(1,2-Propylene Adipate) and Polybutylene Succinate Lindström and co-workers [4] carried out quantitative determinations of aqueous hydrolytic degradation products produced in the early stages of degradation in linear and branched polybutylene adipate (PBA) and polybutylene succinate (PBS). 196

Determination of Additives and their Breakdown Products in Extractants The low molecular weight products migrating from linear and branched PBA and PBS during hydrolysis were extracted, identified and quantified by solid-phase extraction (SPE) and GC-mass spectroscopy (MS). The combination of SPE and GCMS proved to be a sensitive tool, able to detect small differences in the degradation rate during early stages of hydrolysis before any significant differences were observed by weight loss and molecular weight measurements. The detected low molecular weight products included monomers, i.e., adipic acid and 1,4-butanediol for the PBA polymers and succinic acid and 1,4-butanediol for PBS. Several dimers and trimers, i.e., hydroxybutyl adipate (BA), hydroxybutyl succinate (BS), dihydroxybutyl adipate (BAB), dihydroxybutyl succinate (BSB) and hydroxybutyl disuccinate (SBS) were also detected. Comparison of measured weight loss and the amount of monomeric products showed that weight loss during the early stages of hydrolysis was mainly caused by the release of water-soluble oligomers. On prolonged ageing these oligomers continued to hydrolyse to monomeric species.

Figure 13.2 Gas chromatogram of linear polybutylene succinate after 12 weeks in water at 70 °C. The identity of the numbered peaks is (1) succinic acid, dimethyl ester, (2) glutaric acid, dimethyl ester (internal standard), (3) 1,4-butanediol, (4) 1,5-pentanediol (internal standard), (5) succinic acid, monomethyl ester, (6) glutaric acid, monomethyl ester (internal standard). (Reproduced with permission from Lindström and co-workers, Polymer Degradation and Stability, 2004, 83, 3 487 [4]) 197

Additive Migration from Plastics into Foods Figure 13.2 shows the gas chromatogram of the low molecular weight compounds extracted after hydrolysis of linear PBS for 12 weeks in water at 70 °C. The most abundant products in the gas chromatograms were the original monomers, i.e., 1,4butanediol and succinic acid in the case of PBS, and 1,4-butanediol and adipic acid in the case of PBA. In addition several dimers and trimers such as BA, BS, BAB, BSB and SBS were detected throughout the whole ageing period. The most common oligomers were BAB and BSB, i.e., the oligomers with alcohol end-groups. The relative areas of the monomeric species were much higher than the relative areas of the oligomeric species. However, since the relative areas are heavily affected by the volatility of the analytes, this tells little of the real amounts of the different compounds. Figure 13.2 shows the amount of succinic acid and 1,4-butanediol that migrated from linear PBS after different hydrolysis times in water at 37 °C and 70 °C. During ageing at 37 °C the amount of succinic acid and 1,4-butanediol increased slowly as a function of hydrolysis time. Extraction tests carried out at 70 °C, not unexpectedly, increase the amount of monomers formed over the amount produced at 37 °C.

13.4 Organosulfur Vulcanising Agents The chemical structure of some commonly employed vulcanising agents and their breakdown products are given next:

2-Mercaptobenzothiazyl disulfide (MBTS): C14H8N2S4. MW = 332, also known as 2,2′-dithiobis(benzothiazole)

N-Cyclohexyl-2-benzothiazole sulfenamide (CBS): C13H16N2S2. MW = 264

2-Mercaptobenzothiazole (MBT): C7H5NS2. MW = 167 198

Determination of Additives and their Breakdown Products in Extractants

Benzothiazole (BT): C7H5NS. MW = 135 Two important accelerators are MBTS and CBS. MBT is a cure accelerator in its own right and is also a breakdown product of MBTS and CBS. Homolytic cleavage of the S-S bond in MBTS followed by the hydrogen abstraction by the S-centred free radical that results leads to the formation of MBT. Similarly, the formation of BT is formally a result of cleavage of the C-S bond in MBTS or CBS, either homolytically or heterolytically, followed by abstraction of a hydrogen atom or a proton. Furthermore, the double bonds in rubbers can participate in the decomposition of sulfenamidetype accelerators whereby the high electron density in the π-orbital of the double bond polarises the S-N bond and makes it susceptible to the decomposition reaction, which may involve heterolytic cleavage as well as homolytic cleavage to form MBT and BT. The analysis of these organosulfur compounds and their breakdown products is discussed in Chapter 8.

References 1.

M. Hamdani, L. Thil, G. Gans and A.E. Feigenbaum, Journal of Applied Polymer Science, 2002, 83, 5, 956.

2.

Commission of European Communities Materials and Articles in Contact with Foodstuffs, Guidelines for Measurement of Hydrolysis of Plastics Monomers and Additives in Digestive Food Simulents, Commission of European Communities, Brussels, Belgium, 1999.

3.

W. Rotard, W. Christmann, W. Knoth and W. Mailahn, Zeitschrift für Umweltchemie und Ökotoxikologie, 1995, 7, 3.

4.

A. Lindström, A-C. Albertsson and M. Hakkarainen, Polymer Degradation and Stability, 2004, 83, 3, 487.

199

Additive Migration from Plastics into Foods

200

14

Additive Migration Theory

In earlier work on the theory governing the rate of migration of extractable additives and other components from plastics into liquids, Garlanda and Masoero [1, 2] considered the extraction from the point of view of the Fickian Laws of diffusion and they mathematically examined the way in which this migration varies with respect to the principal parameters of the system such as time and temperature, polymer thickness, choice of extractant liquid and so on. Garlanda and Masoero [1, 2] start off their treatment by considering the simple case of the migration of additives from a flat sheet of plastic. For a plastic sheet which contains a substance capable of diffusing along the x axis, perpendicular to the surface of the sheet: Fick’s first law is: F = −D

δC δx

(14.1)

Fick’s second law is: δC δ2C =D 2 δt δx

(14.2)

Where: C is the concentration of migrant in plastic at time t and position x, F is the quantity of diffusing substance moved in units of time and area, t is the time, and, D is the coefficient of diffusion, which is a constant since the initial concentration of additive within the sheet (Co) varies in the range of concentrations considered. Assuming that at the start of the migration test, the concentrations of additive within and at the surface of the sheet are Co and C, respectively, and that the sheet is immersed 201

Additive Migration from Plastics into Foods in a large volume of extraction liquid, then for short time intervals or great polymer sheet thicknesses (denoted by the quantity ΔS) of diffusing substance which has moved from the surface unit of one single side of the sheet from the commencement to time (t) is expressed by: Δs =

2Co Dt π

(14.3)

Where: Co is the initial concentration of additive within the sheet, D is the coefficient of diffusion, and, t is the time of diffusion. DS is therefore proportional to the initial concentration of additive within the sheet (Co) and to √Dt whereas as a first approximation it does not depend upon the sheet thickness. The initial phase of desorption can therefore be represented by a linear relationship with regard to the square root of the time, both for a thin film and a thick one, other conditions being equal, but in the case of the thin film the interval of time within which this can be verified is much shorter than for the thick film. To generalise further - solving Equations (14.1) and (14.2) - it is observed that, at time t, and with diffusion along the x axis, the local concentration within the sheet (C) is: C=

4Co π

∞

∑ n =0

(2n + 1)πx −(2 n+1)2 π2 h−2Dt 1 e sin h 2n + 1



(14.4)

Where: Co is the initial concentration of additive within sheet, C is the initial concentration of additive at the surface of sheet, h is the thickness of sheet, t is the time interval, and D is the coefficient of diffusion. Garlanda and Masoero [1, 2] applied the above considerations to the practical aspects of extractability testing. Regarding the effect of extraction time (t), they state that the quantity of plastic additive that migrates varies and is a function of √t and tends towards a limit value of M∞, which is a function of the initial content of the diffusing substance in the plastic and also of the coefficient of distribution, K. The relative graphs represent functions which are never convex towards the axis of the 202

Additive Migration Theory abscissae. These workers quote as an example the extraction of styrene monomer from polystyrene into vegetable and paraffin oils. The quantity of styrene which migrates (mg/dm2) is linearly proportional to √t during the course of a 30 day test and at the end of this time only 1.3% of the monomer present in the sample had migrated into the oil. With aqueous solvents, after 10 days at 40 °C, the graph had already begun to curve towards the axis √t but the percentage of monomer which had migrated remained small with respect to the total amount of monomer originally in the polymer even after a 30 days extraction test. When a solvent which was capable of reacting strongly with the polystyrene was used, e.g., heptane at 20 °C or at 40 °C, the extraction was obviously rapid: even from the very first hours of testing, graphs which are convex towards the axis of the abscissae are obtained, which show the great difference in performance between n-heptane and vegetable oils. With regard to the effect of extraction test temperatures, it is concluded that the coefficient of diffusion (D) is related to the absolute temperature (T) by a relationship which, in its simplest form, can be expressed as: D = Doe− ED / RT

(14.5)

Where: R is the gas constant, ED is the energy of diffusion, Do is the unit of diffusion coefficient (1 m2/s), and T is the temperature, degrees absolute. The approach of the EC in controlling the safety of food packaging and other food contact plastics has been to assign specific migration limits (SML) to those substances with adverse toxicological properties. The SML is a limit on the quantity of the substance allowed to migrate into foods or food simulants. Plastics monomers have already been covered by this approach (EEC 1990) and it is likely that many other additives will be similarly restricted in the near future. In order to demonstrate compliance with the limits, specific migration tests will need to be carried out on the finished packaging or article. Testing for migration of plastics additives into foods or food simulants in many cases will be costly and difficult, particularly with fatty food and the fatty food simulants owing to analytical difficulties in isolating and determining lipophilic, low volatility substances at low levels in the fat matrix. Moreover, it is recognised that, in the main, 203

Additive Migration from Plastics into Foods fats represent the worst case scenario compared with aqueous foods, giving much higher migration values. In cases where the analytical difficulties in determining substances in fat simulant are insurmountable, provisions have been made for using simple solvent ‘substitutes’ such as 95% ethanol and isooctane (EEC 1997) and taking the highest of the two results. Experimental studies investigating the use of alternative simulants to determine specific migration have been reported elsewhere [3-5]. However, even when using these substitutes a large number of migration tests will still be required. One way of overcoming this potential compliance testing burden with the European Food Regulations is to make use of a validated migration model. These models have been discussed elsewhere [6-12] and are based upon Fickian diffusion principles. Recently a commercial software package has become available based upon the ‘Piringer model’ [12]. In this connection, O’Brien and co-workers [13, 14] carried out detailed studies of the use of the mathematical models for the prediction of migration of additives from food contact polymers and compared these predictions with experimental data. In their first paper [13] experimental migration data was obtained and used to evaluate two Fickian-based models for predicting the migration of four additives from high-density polyethylene (HDPE) into olive oil. Additives were a benzophenone UV stabiliser (2hydroxy-4-N-octyloxy benzophenone), an antioxidant and a benzothiazole stabiliser, a colorant carrier (adipic acid, bis(2-ethylhexyl) ester, a propionate antioxidant (octadecyl-3-(3,5-di-tert-butyl-4-hydroxyphenyl)propionate) and a benzotriazole UV stabiliser (2-(2-hydroxy-3-tert-butyl-5-methylphenyl)-5-chlorobenzotriazole). Predicted migration values were calculated by inserting the measured initial concentration of additive in the HDPE into the equations along with known variables, such as additive molecular weight, temperature and exposure time. This study was undertaken to evaluate independently the models proposed by Piringer and co-workers [11] and Limm and Hollifield [10] in a systematic fashion with well defined HDPE materials and at the same time, provide ‘benchmark data’ for underpinning the use of this approach as an alternative to migration testing for demonstrating future regulatory compliance. It is important that the models should not significantly underestimate migration levels, but, on the other hand not overestimate migration by such an excessive degree that modelling would offer no advantage. This paper reports the practical migration experiments and details the results obtained with HDPE containing selected additives and compares experimental values against predicted values. 204

Additive Migration Theory All tests were conducted using olive oil, representing the most severe case for fatty foods with test conditions of 2 hours at 70 °C, 6 hours at 70 °C, and 10 days at 40 °C, representing short-term exposures at high temperatures and room temperature storage. Predicted migration values were calculated by inserting the measured initial concentration of additive in the polymers (Cp,0) into the equations together with known variables such as additive molecular weight, temperature and exposure time. The results indicate that both models predict migration values into olive oil close to, or in excess of, the experimental results. The Piringer migration model, using the ‘exact’ calculations of the Migratest Lite programme, gave an overestimation for 83% of the migration values generated in this study. The highest overestimation was 3.7 times the measured value. For all measurements, the predicted migration from the Migratest Lite programme was greater than 50% of the observed value. The Food and Drug Administration (FDA) model was found to predict more accurately migration in most situations but underestimated migration more frequently. Differences in the polymer specification had little effect on specific migration of the additives investigated. All migration experiments were conducted using sealed single side cells exposing only one surface of the polymer plaque to the fatty food simulant olive oil. Cell design has been reported [15, 16]. To ensure a good seal, the cell was preheated to 105 °C and clamped down onto the test specimen to slightly soften the polymer at the cell edge and then cooled. As all four additives are reasonably soluble in olive oil, no partitioning effects were expected. Both assembled cells and olive oil were preheated to the test temperature prior to starting the exposure. High temperature gas chromatography (GC) or high performance liquid chromatography were used to determine the additives in olive oil [17, 18]. All of the measured migration data together with predicted values are given in Figures 14.1 to 14.4. These values have been corrected to the EC conventional surface area to food ratio of 6 dm2/kg. Analysis of the spiked simulant from the migration test gave additive recoveries of 83-105%, validating the analytical methods used to measure migration. All calibration graphs were linear with correlation coefficients better than 0.997. For each polymer/blend, the proportion of each additive found to migrate into the food simulant was calculated. In no case was more than 60% of the available additive found to migrate and for the vast majority of samples this value was less than 40%. In general, the test conditions of 10 days at 40 °C were found to be the most severe, giving higher migration results than 2 hours or 6 hours at 70 °C. The exception to this was for migration of the antioxidant, octadecyl-3-(3,5-di-tert-butyl-4-hydroxy-phenyl) 205

Additive Migration from Plastics into Foods

Figure 14.1 Migration of octadecyl-3-(3,5-di-tert-butyl-4-hydroxy phenyl) propionate from different polymer blends P/1 to B/5, into olive oil. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 1999, 16, 9, 367 [13]. © 1999, Taylor & Francis)

Figure 14.2 Migration of 2-hydroxyl-4-N-octoxybenzophenone into olive oil. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 1999, 16, 9, 367 [13]. © 1999, Taylor & Francis) 206

Additive Migration Theory

Figure 14.3 Migration of additive – 2-(2-hydroxy-3-tert-butyl-5-methyl phenyl)-5chlorobenzotriazole into olive oil. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 1999, 16, 9, 367 [13]. © 1999, Taylor & Francis)

Figure 14.4 Migration of adipic acid, bis(2-ethyl hexyl)ester co-colorant into olive oil. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 1999, 16, 9, 367 [13]. © 1999, Taylor & Francis) 207

Additive Migration from Plastics into Foods propionate where 6 hours at 70 °C gave the highest values. By comparing the values at 70 °C it can be seen that migration increases with time as expected in cases where migration is diffusion controlled. It should be emphasised that all data in this study have been obtained using olive oil food simulant. It is recognised that olive oil gives higher migration values than most fatty foods except for pure fats. Reduction factors are given in EC Directive 85/572/ EEC [19] to correct for this when demonstrating compliance with safety limits. O’Brien and co-workers [17] also studied the effect of polymer specification on migration. The polymers covered a range of material that could be found commercially with melt flow indices ranging from 10 g/10 minutes (2.16 kg at 190 °C) to 0.3 g/10 minutes (5 kg at 190 °C). The manufacturers’ recommended conversion process for the polymers varied from injection moulding to blow moulding and film production. To account for the small variations of additive concentrations in the polymer blends the migration value, Cf, was divided by Cp,0 to give migration per mg/kg of additive in the polymer. No consistent trend of specific migration with polymer specification can be noted for any of the additives evaluated. From these figures it can also be noted that migration values for all four additives show similar variations across the sample range - % relative standard deviations (SD) were good (3.7-89%) and % recoveries of additives through the whole process ranged from 83 to 106%. The findings are in agreement with those from studies on high impact polystyrene (HIPS) reported previously [20]. O’Brien and co-workers [13] also compared the effect of the molecular weight of the additive on migration test results. Higher molecular weight additives such as (octadecyl-3-(3,5-di-tert-butyl-4-hydroxyphenyl)propionate show no difference between injection moulded or blow moulded samples. However, for the lower molecular weight additives, 2-hydroxy-4-n-octyloxybenzophenone, adipic acid, bis(2ethylhexyl) ester and 2-(2-hydroxy-3-tert-butyl-5-methylphenyl) 5-chlorobenzotriazole migration from the injection moulded plaques was higher than the blow-moulded bottles manufactured from the same material. The cause for these differences was not investigated in this study but may be due to loss of the more volatile lower molecular weight additives from the surface of the bottles during blow moulding. In Figures 14.1-14.4 experimental migration results obtained in this study are compared with predicted migration values calculated from the measured additive levels in the polymer, Cp,0. Two similar empirical approaches were used to estimate the diffusion coefficient, D(cm2/s), which in turn was used to calculate migration by Fickian diffusion principles. For migration predictions based upon the ‘FDA model’, the Equations described by Limm and Hollifield [10] and Begley [23] were used. Equation 14.6 contains empirical constants derived from Irganox 1010 experimental migration measurements. 208

Additive Migration Theory FDA model [10]: ln D = lnD0 + α(MW)½ - K(MW)⅓/T

(14.6)

Mt = 2Cp,0(Dt/π)½

(14.7)

This equation is identical to that used by Garlanda and Masoero [1, 2] used in earlier works, see Equation (14.3). Where for HDPE: ln D0 (unit of diffusion coefficient, cm2/s) is 0.9 a

is 0.819

K

is 1760.6

Where: D

is the diffusion coefficient (cm2/s),

D0

is the unit if diffusion coefficient (cm2/s),

Cp,0 is the initial concentration of additive in polymer (g/cm2), MW is the molecular weight of additive, α

is the mass ratio of migrant in food to that in packaging film at equilibrium (where α = MF∞/(Mp,0 - MF∞) = KFPVF/VP, and

MF∞ is the migrant amount of migrant in food at equilibrium, Mp,0 is the initial migrant amount in packaging film, KFP is the partition coefficient of migrant between food and packaging film, where KFP = CF∞/CP∞, CF∞ is the migrant concentration in food at equilibrium, CP∞ is the migrant concentration in packaging film at equilibrium, K

is the partition coefficient,

T

is the temperature, K,

Mt is the migrant concentration in food (mg/kg), t

is the migration time, s. 209

Additive Migration from Plastics into Foods Piringer Model [11, 12]: The ‘Migratest Lite’ programme version 1997 (concept Mercea and Piringer) was supplied by Piringer (FABES GmbH, Munich, Germany). All calculations were based upon use of the ‘exact’ calculation based upon Equation (14.9), which corresponds closely to the full Crank Equation. The principles and equations have been described in detail [6-8, 11, 12, 22]. D = 104 exp (Ap – 0.01 MW - 10450/T)

(14.8)

Where, for HDPE: D

is the diffusion coefficient (cm2/s),

Ap is the A diffusion conductance, 22 - 5190/T for fatty foods and fatty food simulants, MW is the additive molecular weight, and T

is the temperature, K.

D = 104 exp(A p − 0.1351(MW)2 / 3 + 0.003MW − 10450 / T)

(14.9)

Where (in Equations 14.6 to 14.9): KFP is the partition coefficient of migration from polymer (P) into extraction liquid or food (F), VF/VP is the volume of extraction liquid or food (VF) divided by volume of polymer (Vp), A

is the area of interface (cm2),

dp

is the thickness of polymer (cm),

Mt

is the migrant concentration in food (mg/kg),

Do

is the unit of diffusion coefficient (m2/s),

D

is the diffusion coefficient (cm2/s),

t

is the time (s),

Cp,0 is the initial concentration of the additive in the polymer (mg/kg), ρp

is the density of polymer (g/cm3),

VF

is the volume of simulant (cm3),

Vp

is the volume of polymer (cm3),

210

Additive Migration Theory qn

is the non zero positive root of tan qn = -α qn,

α

is the mass ratio of migrant in food to that in packaging film at equilibrium where α = MF∞ (MP, 0 - MF∞) = KFP VF/VP,

F∞

is the migrant concentration in food at equilibrium (mg/kg), and

PP∞

is the migrant concentration in packaging film at equilibrium (mg/kg).

An important objective of this study was to evaluate the effectiveness of the models when used as a tool for compliance testing. It must be emphasised that the FDA and Piringer models both use factors to calculate the diffusion coefficient that has been derived empirically from experimental data. For the Piringer model, the data used to generate the factors were obtained from the scientific literature. It is possible that some of these data [11, 12] are unsuitable, and may give misleading information. An example of this is if the data used include substances which are not 100% stable in the food simulant, give a low result. Subsequent prediction of migration of other stable substances based upon this could lead to significant underestimates being made. In addition, information on the polymer specification and formulation may not be available to establish if data are available across the whole range of HDPE compositions and properties commonly used in contact with food. This study, in general, indicates that the Piringer model [11, 12] overestimates migration significantly for the 70 °C tests so there is an additional margin of safety in these predictions. The FDA model (Equations 14.6 and 14.7) predicts lower migration values and underestimates a greater proportion of results than the Piringer model (Equation 14.9), however, the FDA model was not designed to be used as a control measure but was intended to confirm that migration data from petitioners was valid. As a consequence it was designed to give as close as possible to the correct value and not overestimate migration. The data presented in this report confirm that the FDA models (Equations 14.6 and 14.7) perform their intended application with some success for HDPE. It may also be noted that both the FDA and the Piringer models [11, 12] overestimate the migration of the UV stabiliser [2-(2-hydroxy-3-tert-butyl-5-methyl-phenyl)-5chlorobenzotriazole] to a greater degree than any of the other three additives. The differences arising from variations in the polymer specification do not have a significant effect on the observed migration of additives and consequently do not compromise the accuracy of the models. The data from this study could be used to further refine the models. To obtain a better indication of the degree of correlation between the observed migration values and those obtained using the two models the ratio of the predicted migration (Cf) value to the measured migration (Cf) value, was calculated for each result. 211

Additive Migration from Plastics into Foods

Table 14.1 Comparison of predicted data with observed migration Ratio Cfcalc/Cfmeasured % of Piringer results % of FDA results <0.5

0

1.4

0.5-0.9

16.7

44.4

1.0-1.4

30.6

37.5

1.5-1.9

19.4

5.6

2.0-2.4

18.1

11.1

2.5-2.9

6.9

3.0-3.4

5.6

>3.5

2.8

Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis.

To ensure consumer safety, the predictions ideally should be equal to or greater than the experimental values. The results obtained from the experimental data previously are given in Table 14.1. In no case did the Piringer specific migration model give a Cfcalc/Cfmeasured ratio of 0.5 or less (i.e., the model did not underestimate migration by 50% or more). The model gave predictions of 0.5 to 0.8 for 16.7% of the samples, equivalent to 12 measurements. For the majority of these the ratio was 0.8 (i.e., 80% of measured value) or greater. For 92% of the measurements predicted migration was less than three times the measured migration. Using the FDA model, all predictions are less than twice the measured value, but for nearly 50% of the results, predictions are less than the measured value. These findings suggest that it would be valid to use the Piringer mathematical model [11, 12] to predict migration of additives from HDPE. In more recent work, O’Brien and Cooper [14] extended their study to cover a range of different polymers which, in addition to HDPE, included HIPS, polypropylene, polyvinyl chloride (PVC) and polyethylene naphthalate (PEN). The measured migrations were compared with predictions obtained using mathematical models. A large amount of data was obtained and used to evaluate a Fickian-based migration model in the prediction of specific migration of additives into olive oil. All tests were conducted using olive oil, representing the most severe case for fatty foods with test conditions including 2 hours at 121 °C, 6 hours at 70 °C, 2 hours at 70 °C, 2 hours at 60 °C and 10 days at 40 °C, representing shortterm exposures at high temperatures, and room temperature storage. Predicted 212

Additive Migration Theory migrations were calculated by inputting the measured initial concentration of additive in the polymers into the Equations together with known variables such as additive molecular weight, temperature, and exposure time. The results indicated that the Piringer migration model, using the ‘exact’ calculations of the Migratest Lite programme, predicted migrations into olive oil that were close to or in excess of the experimental results and gave an overestimation for more than 95% of the migrations generated in this study. The model used to predict migrations was the ‘Migratest Lite’ programme updated in February 2000 (concept Mercea and Piringer), which was supplied by Piringer (FABES, Munich, Germany). The model uses an empirical approach to estimate the diffusion coefficient, D, which in turn allows calculation of migration following Fickian diffusion principles. The measured additive levels in the polymer, Cp0, were used together with migration conditions, sample size and thickness to calculate migration. The constant in Equation 14.10 which accounts for the polymer type, is Ap. This constant may increase with temperature to account for the fact that the polymer will soften on heating. All calculations used here were based upon use of the ‘exact’ calculation based upon Equation 14.11 which corresponds to the solution given in Crank’s Mathematics of Diffusion for the case of a finite polymer in contact with a finite liquid and is the same as the Piringer model used by O’Brien and Cooper in their earlier work [13] (see Equation 14.9). D = 104exp(Ap - 0.1351(MW)2/3 + 0.003MW - 10450/T)

(14.10)

Where the diffusion conductance Ap = AP′ - C/T ∞   M t / A = C p,0 ρp ⋅ d p (α / 1 + α) 1 − ∑ 2α (1 + α) 1 + α + α2q2n exp {−Dp ⋅ t ⋅ q2n / d2p }  n =1 

(

a = 1 / K P, F ⋅ VF / VP

tan q n = −αq n

)

(14.11)

Where (in Equations 14.10 and 14.11): A

is the area of interface (cm2),

dp

is the thickness of polymer (cm),

Mt is the migrant amount in food (mg/kg), D

is the diffusion coefficient (cm2/s),

MW is the additive molecular weight, T

is the temperature (K), 213

Additive Migration from Plastics into Foods t

is the time (s),

Cp,0 is the initial concentration of the additive in the polymer (mg/kg), pp

is the density of polymer (g/cm3),

VF

is the volume of stimulant (cm3),

VP

is the volume of polymer (cm3),

K

is the partition coefficient,

Ap

is 22–5190/T

Ap′ is the A ‘diffusion conductance’ parameter specific for the polymeric matrix, KPF is the partition coefficient (D) of migration of additive from polymer into extraction liquid or food (F), and C

is the An ‘activation energy adjustment’ parameter specific for the polymeric matrix qn α equal the quantities listed under Equation 14.9.

The performance of the model and its ability to predict migration of additives are discussed next. The aim of this work [14] was to evaluate the effectiveness of the models when used as a tool for compliance testing. The model evaluated here uses empirically derived Ap to estimate a diffusion coefficient. As mentioned previously, these factors have been derived from experimental migration data reported in the scientific literature. It is possible that some of these data are unsuitable, not validated and may give misleading information. An example of this is if the data used include substances that are not 100% stable in the food simulant giving a low result. Subsequent prediction of migration of other stable substances based upon this could lead to a significant underestimation of migration. Figure 14.5 illustrates a scheme for evaluating SML compliance using the model. Table 14.2 gives the migration of two additives (i.e., Cf, which is the concentration of substances found in the food simulant after the migration experiments) from high impact polystyrene (HIPS) and compares the values with the predicted values with a diffusion conductance parameter of Ap = 0 and it can be seen that in all cases the model overestimates migration. Ap could be reduced significantly to get closer agreement. In all cases, the prediction is less than the proposed SML for the additive evaluated and would appear to be an effective control measure for these polymer/additive combinations. Studies by other workers in an EU-funded project, ‘Evaluation of Migration Models’, have found that assuming an Ap value of -4 is effective in achieving agreement between observed and predicted migrations. 214

Additive Migration Theory

Figure 14.5 How to test compliance with model. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis)

There was little effect from polymer specification on migration. However, it was noted that the sample with the lowest mineral oil content and melt flow index did give the lowest migration of additives. It must be emphasised, however, that migration from all grades of HIPS was low and this effect was small. 215

Additive Migration from Plastics into Foods

Table 14.2 Migration of additives from HIPS, 6 hours at 70 °C Polymer 1 Polymer 2 Polymer 3 Polymer 4 Polymer 5 % Mineral oil 6 3 3.7 0 1.7 2-Hydroxy-4-n-octyloxy benzophenone (UV stabiliser) 0.34 0.29 0.42 0.12 0.27 CfPredicted (mg/kg) 2.57 3.03 2.91 2.82 2.11 Octadecyl-3-(3,5-di-tert-butyl-4-hydroxyphenyl) propionate (antioxidant) < 0.6 <0.6 <0.6 <0.6 <0.6 CfPredicted (mg/kg) 0.18 0.26 0.24 0.21 0.18 (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis)

Figure 14.6 Migration of octadecyl-3-(3,5-di-tert-butyl-4hydroxyphenyl)propionate from high-density polyethylene into olive oil. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis) Figure 14.6 compares observed migration, predicted using the migration model for the migration of octadecyl-3-(3,5, di-tert-butyl-4-hydroxyphenyl) propionate from HDPE into olive oil and the results obtained from this investigation. Here, the data set labelled ‘measured’, the experimentally obtained data, are compared with the data set labelled ‘Model’, which have been derived using Equation 14.10. 216

Additive Migration Theory From Figure 14.6, it can be seen that in all cases the predicted migrations are higher than the experimental data giving a high confidence that safety margins will not be compromised by the use of the model. To obtain a better indication of the degree of correlation between observed data and predicted migration, the ratio of the migration value Cfpredicted to the migration value of Cfmeasured was calculated and the distribution of these is given in Figure 14.7 for all migration experiments. Cfpredicted should be > Cfmeasured and this ratio should be ≥ 1 which was the case for all the data points of the measurements. C is the ‘activation energy adjustment’ parameter specific for the polymer matrix. In the case of the migration of 2-hydroxy-4-n-octyloxy benzophenone (UV stabiliser) from eight polypropylene random copolymers the degree of correlation between the observed migrations and those obtained using mathematical models the ratio of Cfpredicted to Cfmeasured was calculated for each result. The predictions ideally should be equal to or greater than the experimental values and this ratio should be ≥ 1. The comparison of observed migration to predicted data are given in Figure 14.8. The model underestimated migration for 4% of the samples evaluated with a Cfpredicted / Cfmeasured ratio of 0.7-1.0. In no case was this ratio < 0.7. For 90% of the measurements predicted, migration was < 4.5 times the measured migration with the highest overestimate being by a factor of 7. Studies on the migration of methyl and octyl tin stabilisers from PVC show that comparing the predicted migration to the determined migration for the additives, the model overestimates migration in all cases for 10 days at 40 °C. However, for the higher temperature test there is a significant underestimation (Table 14.3).

Figure 14.7 Predicted versus observed data for HDPE. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis) 217

Additive Migration from Plastics into Foods

Figure 14.8 Predicted migration versus observed migration. (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis)

Table 14.3 Migration of tin stabilisers from PVC Specific migration (ppb) 2 h/60 °C 10 d/40 °C Sn Additive Predicted Sn Additive Predicted 1 14 100 59 24 175 236 2 16 116 59 17 125 236 3 20 137 117 21 141 471 4 5.5 30 130 5.4 29 523 5 18 144 52 31 187 210 6 26 157 57 30 172 231 (Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis)

14.1 Polyethylene Naphthalate A range of PEN samples were investigated. Into each of these samples, 1% benzyl butyl phthalate (BBP) and 0.1% of a benzotriazole were compounded and injection moulded plaques suitable for testing were manufactured from the blends. Migration of the additives into olive oil food simulant has been measured using test condition of 10 days at 40 °C and 2 hours at 121 °C. The results are listed in Table 14.4 together with the measured Cp,0. This polymer is relatively new to the market place and insufficient migration data are available to generate a realistic Ap value. Consequently, 218

Additive Migration Theory

Table 14.4 Migration of additives from PEN Polymer Polymer Polymer Polymer Polymer 1 2 3 4 5

Polymer 6

BBP Cf (mg/kg) 10 d at 40 °C

0.1

0.1

0.1

0.1

0.1

0.1

2 h at 121 °C

0.2

0.2

5.3

0.2

0.1

0.2

CP,0, (mg/kg)

8700

8700

7100

8600

8600

8200

(Reproduced with permission from O’Brien and co-workers, Food Additives and Contaminants, 2002, 19, Supplement 1, 63 [14]. © 2002, Taylor & Francis)

predictions are not possible for this polymer. The initial concentration of additive in the polymer Cp,0 and the predicted migration value Cf for these polymers are given Table 14.4 and can be used to derive a suitable Ap. The conclusions reached in this study are as follows. A mathematical model has been used to predict migration of additives from food contact grade polymers. The FABES migration model used in these studies proposed by Piringer (Equation 14.11) is the ‘exact calculation’ of the Migratest Lite programme. The data generated have been used by the developers of the programme to validate the empirically derived factors used in the model. This has significantly improved the performance of the model that now overestimates migration for > 95% of the formulations evaluated. In more recent work on diffusion, Chung and co-workers [22] point out that two migration models based on Fick’s second law of diffusion have been extensively used for the assessment of migration of additives and contaminants from food packaging films [23-31] (compare with Equations 14.1 and 14.2). M F, t M P ,0 M F, t M F ,∞

0.5

=

2  Dt    LP  π 



(4.12)

=

2  Dt    LP  π 

(4.13)

0.5

Where: MF,t is the amount of migrant in the food at time t, MP,0 is the initial amount of migrant in the packaging film, 219

Additive Migration from Plastics into Foods MF, ∞ is the amount of migrant in the food at equilibrium, D

is the diffusion coefficient of migrant in the packaging film,

Lp

is the thickness of the packaging film, and

t

is time (s).

The two models are the same and known to provide accurate estimates of diffusioncoefficients for complete migration: MP,0 = MF, ∞ (initial migrant amounts in packaging film – migrant amount in food at equilibrium), where partitioning and resistance to mass transfer are negligible. The diffusion coefficient can be determined easily from the plot of MF,t / Mp,0 versus t0.5 or the plot of MF,t / MF, ∞ versus t0.5 using initial migration data and linear regression analysis. Theoretically, it is possible to determine the diffusion coefficient from Equations (l4.12 or 14.13) with a single datum point, however, the use of multiple data points and linear regression analysis can provide a far more reliable estimate of diffusion coefficient. The two models are less accurate in determining diffusion coefficients for partitioned migration (Mp,0 > MF, ∞). When partitioning occurs, the models are no longer the same, and they yield different diffusion coefficients for the same migration [28]. In addition, the models could result in considerable errors in the determination of diffusion coefficients when partitioning is significant [23, 26]. This is not surprising because Equation (14.12) does not include any parameter for partition behaviour. Equation (14.13) includes an equilibrium quantity (MF, ∞), but it is doubtful that simply replacing MP,0 with MF, ∞ can greatly improve the estimation of diffusion coefficients. Nevertheless, the two models are often used to estimate diffusion coefficients from data that involve partitioned migration. One reason is that the models are simple and convenient to use. Another reason is to maximise the use of available data. For example, it is difficult to determine if partitioning will occur before an experiment. If partitioning is observed after the experiment, it is possible to redesign and repeat the experiment to avoid partitioning, however, doing so also requires considerable additional time and effort. When repeating the experiment is not affordable, the models are often used to provide ‘best estimates’ of diffusion coefficients based on the available data, knowing that some errors must be accepted. Chung and co-workers [22] attempted to clarify the mathematical implications of the two migration models and examine their errors in estimating diffusion coefficients for partitioned migration. Another objective was to propose a simple migration model that can provide better estimates of diffusion coefficients for partitioned migration than the two models. Chung and co-workers [22] examined the assumptions and deviations of general migration models, for the purpose of defining the applications and limitations of 220

Additive Migration Theory Equations (14.12) and (14.13). Migration of an additive or contaminant from an amorphous polymeric packaging film above the glass transition temperature is often controlled by the molecular diffusion of the migrant in the film, which can be described by Fick’s Second Law [9] (see also Equation 14.2): δC P δ2C = D 2P δt δx

(14.4)

Where: CP is the concentration of the migrant in the packaging-film at time t and position x, and D

is the diffusion coefficient of migrant within the packaging film.

To describe the migration process, the following assumptions are often used [24, 32-35]: (1) initially, the migrant is distributed uniformly in the packaging film; (2) the migration occurs from one side of the packaging film to a liquid food; (3) the liquid food is well mixed so that there is no migrant concentration gradient in the food, implying very large surface mass transfer coefficient of migrant (km); (4) the surface mass transfer coefficient is much larger than the diffusion coefficient (Biot number [Bi] = KFPkmLP/D ≈ ∞, (where KFP is the partition coefficient of migrant between food and packaging film, where KFP = CF∞/CP∞) implying the migration is controlled by Fickian diffusion in the packaging film and the effect of mixing is negligible; (5) diffusion coefficient, and partition coefficient KFP = CF∞/ CP∞) (where CF∞ = migrant concentration in food at equilibrium, and CP∞ = migrant concentration in packaging film at equilibrium) are constant during migration and depend only on temperature; (6) equilibrium exists all the time during migration at the interface of packaging film and food; and (7) edge effects and interactions between packaging film and food are negligible. Furthermore, the following two cases may be considered to obtain migration models depending on whether the volume of food is infinite or limited [16]. 221

Additive Migration from Plastics into Foods A summary of the nomenclature in Chung’s paper [22] is given next: A

exposed surface area of packaging film

Bi

Biot number, where Bi = KFPkmLP/D

CF,0 initial migrant concentration in food CF∞ migrant concentration in food at equilibrium CP

migrant concentration in packaging film

Cp,0 initial migrant concentration in packaging film CP,L migrant concentration in packaging film at the interface CP,∞ migrant concentration in packaging film at equilibrium D

diffusion coefficient of migrant within packaging film

KFP partition coefficient of migrant between food and packaging film, where KFP = CF∞/CP∞ km

surface mass transfer coefficient of migrant

L P

thickness of packaging film

MF,L migrant amount in food at the end of migration for Case 1 MF, ∞ migrant amount in food at equilibrium MF,t migrant amount in food at time t MP,0 initial migrant amount in packaging film qn

non-zero positive roots of tan qn = –αqn.

t

migration time

VF

volume of food

VP

volume of packaging film

x

position in packaging film

α

mass ratio of migrant in food to that in packaging film at equilibrium, where α = MF, ∞/(MP,0 - MF,∞) = KFP VF/VP

ε

diffusion coefficient prediction error for the linear migration models.

n

index variable

222

Additive Migration Theory Special functions: erf(ω) is the error function, where: erf (ω) =

2

∫

π

ω − n2 0

e δn

erfc(ω) is the complementary error function, where: erfc(ω) = 1 - erf(ω) ierfc(ω) is the integral of the error function, where: ierfc(ω) =

∫

∞ 0

erfc(n)δn =

1 π

2

e−ω − ω ⋅ erfc(ω)

Case 1: Limited Packaging, Infinite Food This is the case where migration occurs from a limited-volume packaging film into a well-mixed infinite volume of food. The infinite volume of food means that the concentration of migrant in the food remains constant during migration and equals its initial value (CF,0). If KFP is constant, then the migrant concentration in the packaging film at the interface between packaging film and food (CP,L) also remains constant and equals CF,0 / KFP. Migration proceeds until the concentration of migrant in the film decreases from its initial value (CP,0) to the interface value (CP,L). Based on these assumptions, the initial condition for CP,L (migrant concentration in packaging film) is: CP = CP,0 at 0 < x < Lp, t = 0

(14.15)

Where x = position in packaging film and the boundary conditions are: C P = C P,L =

C F ,0 K FP

at x = L P, t > 0

δC P = 0 at x = 0, t > 0 δx

(14.16)

(14.17)

223

Additive Migration from Plastics into Foods A solution of Equation (14.4) useful for short migration times is given by Crank [32]: C P − C P ,0 C P , L − C P ,0

∞  (2n + 1)L P − x  ∞  (2n + 1)L P − x  n = ∑ (−1)n erfc   + ∑ (−1) erfc   0.5 0.5  2(Dt)  2(Dt)  n =0  n =0

(14.18)

The amount of migrant transferred from the packaging film to the food from time zero to time t (MF,t) can be obtained using the migrant flux at packaging film/food interface (J| × = LP): M F, t =

∫

t 0

AJ

x = LP

dt, where J

x = LP

= −D

δC P δx



(14.19)

x = LP

The amount of migrant transferred from the packaging film to the food for Case 1 during the entire migration period (MF,L) is:  C  VC M F,L = AL P (C P,0 − C P,L ) = VP C P,0 − F,0  = M P,0 − P F,0 K FP  K FP 

(14.20)

Where: A is the exposed area of the packaging film, and VP is the volume of packaging film. From Equation (14.21), MF,t is given by Crank [32]: M F, t M F,L

0.5

 Dt  = 2 2   LP 

∞  nL P   1  0.5 + 2∑ (−1)ierfc   0.5   π  (Dt)  n =1

(14.21)

An equivalent form of Equation 14.21 useful for large migration times is given by Crank [32]: M F, t M F,L

 −D(2n + 1)2 π2t  8 exp   2 2 4L2P   n =0 (2n + 1) π ∞

= 1∑

(14.22)

• Case 2: Limited Packaging, Limited Food This is the case where migration occurs from a limited-volume packaging film into a well-mixed limited volume of food. The food initially does not contain any migrant, 224

Additive Migration Theory and as migration occurs, the concentration of migrant in the food increases from zero (CF,0 = 0) to its equilibrium value (CF,∞). This case describes actual migration better than Case 1. The boundary condition is expressed as the mass balance of migrant by Gandek and co-workers [33] and Chung and co-workers [34]:  V  δC δC K FP  F  P = −D P at x = L P , t > 0 δx  A  δt

(14.23)

Where: VF is the volume of food. The initial condition and the boundary condition at x = 0 are the same as Equations (14.15) and (14.17), respectively. A solution of Equation (14.14) for the above conditions is proposed by Crank [32]:

C P − C P ,0 C P ,∞

 −Dq2 t  q x 2 1 α + ( )exp  2 n  cos  n  ∞  LP   LL  × = 1+ ∑ 2 2 cos (q n ) 1 + α + α qn n =1

(14.24)

Where: CP,∞ is the migrant concentration in the packaging film at equilibrium, α is the mass ratio of migrant in the food to that in the packaging film at equilibrium (α = KFP VF /Vp), and qn is the non-zero positive roots of tan qn = – α qn. The total amount of migrant transferred from the packaging film to the food until equilibrium (MF, ∞) is: C  VC M F,∞ = AL P (C P,0 − C P,∞ ) = VP  F,∞  = M P,0 − P F,∞ K FP  K FP 

(14.25)

Where CF,∞ is the migrant concentration in the food at equilibrium: C F ,∞ =

K FP C P ,0 1+ α

(14.26)

From Equations (14.19), (14.24) and (14.25), MF,t is given by Crank [32]: M F, t M F ,∞

2α (1 + α)

 −Dq2 t  exp  2 n   LP  n =1 1 + α + α q ∞

= 1− ∑

2

2 n

(14.27)

225

Additive Migration from Plastics into Foods Equation (14.27) is the most rigorous general model for describing the migration controlled by Fickian diffusion in a packaging film. A simpler form of Equation (14.27), which is applicable for most values of α, is given by Crank [32]: M F, t M F ,∞

  = (1 + α) 1 − eωerfc ω0.5   

( )

(14.28)

• Two Migration Models The two models (Equations 14.12 and 14.13) are simplified forms of Equation (14.21), which belongs to the type of ‘limited packaging, infinite food’. For short migration times, ierfc[nLP/(Dt)0.5} → 0 and Equation (14.21) can be simplified as: M F, t M F,L

0.5

2  Dt  =   LP  π 



(14.29)

i.e., Equation 14.12 Equation (14.29) can also be-obtained when LP is large, implying it belongs to the type of ‘infinite packaging, infinite food’. Equation (14.12) can be obtained from Equation (14.29) by assuming that the food initially does not contain any migrant (CF,0 = 0), since the assumption leads to MF,L = MP,0 in Equation (14.20). The assumption CF,0 = 0 prescribes that CP,L remains constant at zero during migration, since the ‘infinite food’ and constant KFP are also assumed in Case 1. Furthermore, Equation (14.12) also prescribes complete migration, since migration proceeds until CP decreases from CP,0 to zero. If Equation (14.12) is used for partitioned migration, some errors are expected due the violation of the prescribed conditions. For partitioned migration, CP,L no longer remains zero at all time, but it increases from zero to its equilibrium value (CP,∞). Therefore, the use of Equation (14.12) for partitioned migration will result in a lower CP,L, a larger (δCP/δx)|x = LP, and consequently a lower estimate of diffusion coefficient, compared with the corresponding actual values. Similarly, Equation (14.13) can be obtained from Equation (14.29) by assuming the initial migrant concentration in the food (CF,0) is the same as CF,∞ since the assumption leads to MF,L = MF, ∞ in Equation (14.25). The assumption CF,0 = CF,∞ prescribes that CP,L 226

Additive Migration Theory remains constant at CP,∞ during migration, since the ‘infinite food’ and constant KFP are also assumed in Case 1. Furthermore, Equation (14.13) also prescribes partitioned migration, since migration proceeds until CP decreases from CP,0 to CP,∞. If Equation (14.13) is used for partitioned migration, some errors are also expected because the prescribed conditions (CF,0 = CF,∞ and CP,L = CP,∞ at all times) are not strictly satisfied. In real situations, CP,L is lower than CP,∞ until it reaches CP,∞ at the end of migration. Therefore, the use of Equation (14.13) for partitioned migration will result in a higher CP,L, a smaller (δCP/δx)|x = LP, and consequently a higher estimate of diffusion coefficient, compared with the corresponding actual values.

• Estimation Errors for Partitioned Migration The estimation error (εi) involved in using the two models to obtain diffusion coefficients for partitioned migration is defined here as: εi (%) ≡ 100 ×

Di − Do Do

(14.30)

Where: Di is the diffusion coefficient estimated using the models i = 1 and 2 for Equations (14.12) and (4.13), respectively, and Do is the actual diffusion coefficient. The estimation error is examined using simulated data generated from Equation (14.27), which is the most rigorous model for partitioned migration. To generate the simulated data, Do = 10-11 cm2/s and LP = 10 μm are assumed. DO is within the range of reported values for packaging migration (10-9 - 10-12 cm2/s; [35]). LP is smaller than the normal packaging film thickness, but this was done on purpose, because the accuracy of Equation (14.29) increases with LP. The simulated data of MF,t / MF, ∞ versus t are then generated at five alphas (α = 0.5, 1, 2, 5, 20) to represent various partitioned migration conditions, using Equation (14.27) and Mathcad 14 software (PTC, Needham, MA, USA). For each set of simulated data, diffusion coefficients (Di) are estimated based on the two models using linear regression (R2 > 0.99). The estimated diffusion coefficients (D1, D2) are shown in Table 14.5. The results show that Equation (14.12) underestimates while Equation (14.13) overestimates diffusion coefficients, which is consistent with the earlier discussion. The estimation errors (ε1, ε2; Figure 14.9) are negligible when α is very large (complete migration) and increase with decreasing α (partitioned migration). Considerable estimation errors (ε1 > 14% and ε2 > 25%) occur for α < 5 under the given migration conditions (Figure 14.10). 227

Additive Migration from Plastics into Foods

Table 14.5 Diffusion coefficients estimated from the two migration models (D1 from Equation 14.2 and D2 from Equation 14.3) and the proposed model (D3 from Equation 14.35) under different partitioning conditions. Simulated data (MF,t/ MF, ∞ versus time) are generated from Equation 14.27 assuming D0 = 10-11 cm2/s and LP = 10 μm. Diffusion coefficient cm2/s × 10-11 0.5

1.0

2.0

5.0

20.0

D1

0.42

0.57

0.71

0.86

0.96

D2

3.85

2.29

1.59

1.26

1.08

D3

1.18

1.10

1.03

1.01

1.00

Reproduced with permission from Chung and co-workers, Food Additives and Contaminants, 2002, 19, 6, 611 [22]. © 2002, Taylor & Francis

Figure 14.9 Estimation errors of diffusion coefficients from the two models (ε1 and ε2) and the proposed model (ε3) as a function of α. Simulated data (MF,t/MF,∞ versus time) are generated from equation (14.27) assuming D0 = 10-11 cm2/s and LP = 10 µm. (Reproduced with permission from Chung and co-workers, Food Additives and Contaminants, 2002, 19, 6, 611 [22]. © 2002, Taylor & Francis)

228

Additive Migration Theory

Figure 14.10 Simulated migration profiles generated from the proposed model (equation 23) and the rigorous model (Equation 14.27] for α = 1 and 5, D0 = D3 = 10-11 cm2/s and LP = 10 µm. (Reproduced with permission from Chung and coworkers, Food Additives and Contaminants, 2002, 19, 6, 611 [22]. © 2002, Taylor & Francis)

Equation (14.12) yields a more accurate diffusion coefficient than Equation (14.13) at any partition condition ε1 < ε2), although partition is not considered in Equation (14.12). Since the two models are derived for short migration times, the assumption of CF,0 = CP,L = 0 seems to be better than the assumptions of CF,0 = CF,∞ and CP,L = CP,∞. Chung and co-workers [22] proposed a sample migration model for Case 2 quoted previously. While there are simple models [Equations (14.12) and (14.13)] available for Case 1 (infinite food), there is no simple model available for Case 2 (limited food), although Case 2 is based on more realistic assumptions. Equation (14.27) is seldom used for packaging migration assessment due to its high non-linearity, even though it is the most rigorous migration model (Begley [23] and Hamdani [28]). This is because more than three or four terms of the infinite summation are often required to achieve sufficient prediction accuracy, especially for the initial period of migration. When migration time is large, it is sufficient to consider only the first term of the infinite summation, and the diffusion coefficient can be estimated using linear regression analysis after calculating q1, the first non-zero positive root of tan qn = – α qn (Chung and co-workers [34]). Nevertheless, detailed migration data and elaborate calculations are still needed to ensure safe approximation of the infinite summation by the first term. Equation 229

Additive Migration from Plastics into Foods (14.28), a simpler form of Equation (14.27) has appeared in many papers (Miltz [36], Limm and Hollifield [27], Hamdani and co-workers [28] and Lickly and co-workers [29]), but it is also rarely used because of its non-linear nature. Chung and co-workers [34] propose a simple migration model for Case 2, which is derived from Equation (14.28). Since MF, ∞ / MP,0 = α /(l + α) and erfc(ω0.5) = 1 erf(ω0.5), Equation (14.28) can be rearranged to: M F, t M P ,0

( )

= α 1 − eω + eω erf ω0.5 

(14.31)

From the Maclaurin series of eω and erf(ω0.5) (Kreyszig [37]), the following relationships are obtained: 1 − eω = −

ω ω 2 ω3 − − −… 1! 2 ! 3!

( )

eωerf ω0.5 =

(14.32)

 2  0.5  21  1.5  22  2.5 ω +   ω +… ω +  0.5 π   1⋅ 3 ⋅ 5  1⋅ 3 

(14.33)

For short migration times (ω « 1), we may retain only the first terms of Equations (14.32) and (14.33). Substituting these first terms into Equation (14.31), we obtain: M F, t M P ,0

0.5

0.5

ω 2  Dt  Dt = 2α   − αω =   − 2 LP  π  αL P π

(14.34)

Compared with Equation (14.12), Equation (14.34) has an additional term describing partition behaviour on its right-hand side expression. This term is negligible when α→ ∞ (complete migration) or LP → ∞ (infinite packaging), which implies that Equation (14.34) retains the ‘limited packaging, limited food’ assumption of Case 2 with this term. An equivalent form of Equation (14.34) useful for linear regression analysis is: 0.5

1 1 M  D0.5 0.5 1  − ⋅ F, t  = − ⋅ t + 0.5 αL P π  π α M P,0 

(14.35)

Thus, diffusion coefficient can be estimated from the slope of the plot of [1/π - (1/ α)(MF,t / MP,0)]0.5 versus t0.5. The simple models described here (Equations 14.12, 14.34 and 14.35), are valid only for initial short migration times, however, migration 230

Additive Migration Theory data needs to be collected until equilibrium to determine MF, ∞ in Equation (14.13) and α in Equation (14.35). Table 14.5 compares the diffusion coefficient (D3) estimated from the proposed model (Equation 14.35) to those estimated from the two models. Figure 14.9 shows the estimation errors as a function of α. The results indicate that the proposed model yields more accurate diffusion coefficients (ε3 < ε1 < ε2) than the two models, especially when α is small. The estimation errors for the proposed model are less than 3% for most α range (α > 2.0) under the given migration conditions. Figure 14.10 compares the predictions of MF,t / MP,0 versus t0.5 from the proposed model (Equation 14.34) and the rigorous model (Equation 14.27). The two models agree well for short migration times, and the length of time, for the agreement increases with α. The proposed model yields a parabolic curve with a maximum of α/π at t0.5 = αLP/π(D)0.5. The maximum is equal to the equilibrium fractional migration (MF, ∞ / MP,0) when α = π - 1, since MF, ∞/MP,0 = α/(l + α). The predictions from the two models deviate at large migration times: the proposed model overestimates the fractional migration when α > π -1 (e.g., α = 5), but it underestimates the fractional migration when α < π -1 (e.g., α = 1). By examination of the assumptions and derivations, Chung and co-workers [22] were able to reveal subtle differences between the different migration models, which are important in determining their application for partitioned migration. The two models belong to the case of ‘infinite packaging, infinite food’. More specifically, Equation (14.12) assumes ‘well-mixed infinite food without any migrant’, while Equation (14.13) assumes ‘well-mixed infinite food, containing equilibrium amount of migrant’. When applied to partitioned migration, both Equations yield significant errors in the estimation of diffusion coefficients. On the other hand, the proposed model belongs to the more realistic case of ‘limited packaging, limited food’. It has the same form as Equation (14.12), except a partition parameter is also included. Its advantages are simplicity and the ability to provide far better estimates of diffusion coefficient than the two models for partitioned migration. Chung and co-workers [22] considered only the theoretical aspects of migration models. Further work is needed to test the proposed model with real data. Helmroth and co-workers [38] point out that given a diffusion coefficient D and a partition coefficient K, migration of a component from a polymer to a solvent can be calculated (according to Fick) as a function of time. Both D and K are macroscopic parameters that describe the combined properties of migrant, polymer and contacting solvent at a certain temperature. Empirical relationships of D with the molecular weight of the additive, polymer type and temperature have been established [10, 39]. However, the properties of the contacting solvent can also influence the migration rate. 231

Additive Migration from Plastics into Foods For example, Koszinowski [40] and Piergiovanni and co-workers [41] showed that components permeating through different polyolefins move faster when the polymer film is in contact with a more apolar than with a more polar solvent. Piergiovanni and co-workers [41] also found that the diffusion rate of (the more apolar) isooctane is significantly higher than that of (the more polar) ethanol in low-density polyethylene (LDPE), implying that isooctane migration enhanced the migration rate of the other components. Similarly, the migration rate of small molecules in PVC film was shown to be governed by penetration of the contacting solvent, tertiary butyl acetate [31]. Diffusion depends not only on polarity, but also on molecular size and shape. For example, the polarity of larger triglycerides is almost equivalent to that of polyolefins, but results in the literature do not underpin any hypothesis of an increase in migration rate initiated by triglycerides. For example, as pointed out earlier, Goydan and co-workers [25] found that the migration rates of both Irganox 1076 (octadecyl-3-(3,5-di-tert-butyl4-hydroxyphenyl propionate)) and Irganox 1010 (tetrakis (methylene-3-(3,5-di-tertbutyl-4-hydroxyphenyl) propionate)) from polyolefins into 95% ethanol were similar to the migration rates to corn oil. Also, O’Brien and co-workers [13] showed that specific migrations obtained for Irganox 1076 and di(2-ethylhexyl)adipate migrating from polyolefins into 95% ethanol were in good agreement with those migrations to olive oil. In contrast with this, Reynier and co-workers [42] suggested using a D that depends on the local solvent concentration to compensate for the effect on additive migration caused by penetrating solvents, including olive oil and other triglycerides. The exact role of triglycerides on additive migration is thus not yet clear. According to the free-volume theory, diffusion in polymers occurs by low molecular weight components ‘jumping’through holes available between the twined polymer chains. The presence of the components in the holes will result in a larger total free volume, since the total free volume of polymer and components together is considerably larger than that of the polymer alone. This extra free volume allows easier segmental motion of the polymer chains. Consequently, the diffusion of the components (and other components present) becomes faster as their local concentration increases. This effect is called plasticisation and is similar to the effect caused by increasing the temperature [43-45]. A rule of thumb reported for elastomers is that the solvent diffusion coefficient increases by a factor of ~10 for an increase in solvent concentration of ~15 vol% [45]. Helmroth and co-workers [38] focussed on the effect that solvent absorption by the polymer has on migration of additives. Specifically, the D of Irganox 1076 is related to the maximum solvent absorption in LDPE film. Solvents were selected to cover a broad range of polarities and molecular sizes, including fatty food simulants as used in EC legislation. This resulted in five low-molecular organic solvents (ethanol, isopropanol, isooctane, ethyl acetate and cyclohexane) with varying polarity, two triglycerides (tributyrin and tricaprylin) and olive oil. 232

Additive Migration Theory Diffusion and partition coefficients were determined by fitting the migration curves, i.e., the concentration of Irganox 1076 in solvent as a function of time, with Fick’s diffusion Equation. The results for the low molecular weight solvents showed that with increasing maximum solvent absorption, D (diffusion coefficient of migrant within the packaging film) of Irganox 1076 also increased. This trend was not observed for the two triglycerides and olive oil. Despite the absorption, no increase in D was observed. The result obtained was the basis of an extended predictive migration model that, apart from migrant and polymer properties, was also based on the maximum solvent absorption in the polymer. An analytical solution of Fick’s second diffusion Equation for one-dimensional diffusion and limited volumes of packaging and solvent is given by Crank [32]: M S, t M P ,0

∞   α   2α (1 + α) q2n  = exp t − D    × 1 − ∑ eff L2   1 + α   n =1 1 + α + α2q2n 

(14.36)

Where: MS,t is the mass of additive in the solvent after time, t (s), MP,0 is the initial mass of additive in the polymer (kg), Deff is the effective diffusion coefficient (m2/s), L

is polymer thickness (m),

qn

is the non-zero, positive roots of Equation: tan (α) = –qn α, and

n

is the index variable,

α

is Vs/(VpK)

Equation (14.36) further assumes that: (1) the additive is homogeneously distributed in the polymer, (2) there is no mass transfer resistance at the interface between polymer and solvent, (3) there is no diffusion from the polymer surface that is not in contact with the solvent, and (4) the polymer matrix is not changing throughout the migration process. A migration cell for single-sided contact was designed to follow migration, i.e., the concentration of Irganox 1076 in the solvent, as a function of time. The cell was made 233

Additive Migration from Plastics into Foods of stainless steel and contained a screw cap with a septum for sampling with an injection syringe. Preheated cells were filled with 20 ml solvent, a pre-cut polymer film (contact area 38.5 cm2) was inserted and the cell tightened firmly by hand. Turning the cell upside down immediately started the contact period. The cells were incubated under gentle shaking (55 rpm) in a water-bath of 40 °C (± 0.3 °C). Samples of 40 μl were transferred to an autosampler vial with a 100 μl insert and 5 μl internal standard solution was added. The samples were analysed by GC with flame ionisation detection (FID). Estimation values of diffusion coefficient D and partition coefficient K are given in Table 14.6. K varied in all cases between 0.3 and 11 and should, due to the observed variation, only be considered as an indicative value. Low partition coefficients were expected as Irganox 1076 has a good solubility in all selected solvents. Diffusion coefficients, found in the literature, of Irganox 1076 in LDPE at 40 °C were measured without solvent contact

Table 14.6 Effective diffusion coefficients (D) and partition coefficients (K) of Irganox 1076 migrating from LDPE to the test solvents at 40 °C (duplicate or triplicate values) Solvent D (cm2/s) × 10-9 K (m3/m3) 0.8 11 Tributyrin 0.7 11 0.9 5 Tricaprylin 0.9 6 0.8 nd Olive oil 0.9 nd 1.0 nd 1.0 0.3 Ethanol 1.2 3 1.2 3 Isopropanol 1.4 2 7.4 4 Ethylacetate 9.3 11 24 4 Isooctane 27 7 64 4 Cyclohexane 65 9 nd: no data Reproduced with permission from Helmroth and co-workers, Food Additives and Contaminants, 2002, 19, 2, 176 [38]. © 2002, Taylor & Francis

234

Additive Migration Theory by a stack of polymer films pressed against a source containing the additive, were 6.9 × 10-10 and 1.0 × 10-9 cm2/s [46, 47]. The values found here for ethanol, isopropanol and the triglycerides (0.7-1.4 × 10-9 cm2/s) are in good agreement with these values. Figure 14.11 shows the measured migration curves as the amount of migrated Irganox 1076 in the solvent relative to the initial amount in the polymer as a function of time at 40 °C. A close look at the data points and best-fitting curves in Figure 14.12 shows that for ethanol, isopropanol, tributyrin and tricaprylin the applied Fick Equation fits very well. For isooctane, ethylacetate and cyclohexane, the fit is less perfect. For olive oil, the deviation from the applied Fick Equation was very large. This was most likely due to the well-known difficulty of analysing Irganox 1076 in olive oil, which is a complex mixture of different triglycerides. A more quantitative way of determining how well the applied Fick Equation fits, is calculating the root of mean-square error (RMSE) using Equation (14.37). The smaller the RMSE, the better the experimental data fit the Equation. The calculated RMSE in Table 14.7 confirm the results observed by eye. Notice the large RMSE of the triplicate measurements of olive oil.

RMSE =

1 M P ,0

(

1 N ∑ (MS,t )experiment, i − (MS,t )prediccted, i N i =1

)

2



(14.37)

Effective diffusion coefficients of Irganox 1076 are shown in Figure 14.12 as a function of the maximum solvent absorption. For the low molecular solvents, a clear trend exists: the more a solvent is absorbed in the polymer, the higher is the migration rate of the additive into that solvent. The migration behaviour of an additive to a solvent can thus, in principle, in accordance with the free-volume theory, be related to the maximum absorption of that solvent in the polymer. When little absorption occurred, the migration behaviour was pure Fickian and D was independent of the type of solvent. In contact with isooctane, ethylacetate and cyclohexane, LDPE was clearly plasticised since considerable solvent absorption occurred that significantly increased the value of the effective diffusion coefficient (D). For example, compared with (almost) no absorption, the diffusion coefficient of Irganox 1076 increased by a factor of ~25 at a solvent absorption of ~15 vol% (of isooctane). This increase is in the same order of magnitude as that reported for elastomers. For the triglycerides, the results do not clearly follow the observed relationship between maximum solvent absorption and effective diffusion coefficient (D). Although the maximum solvent absorption of the triglycerides was slightly higher than that of ethanol and isopropanol, the D were more or less similar. The relatively low maximum absorption of the three different types of triglyceride confirms the theory that solvent absorption depends on both polarity and molecular size and shape. 235

Additive Migration from Plastics into Foods

Figure 14.11 Amount of Irganox 1076 migrated from LDPE in the test solvents relative to the initial amount in the polymer as a function of time at 40 °C. Experimental data are shown as points and best-fitting curves as solid lines. (Reproduced with permission from Helmroth and co-workers, Food Additives and Contaminants, 2002, 19, 2, 176 [38]. © 2002, Taylor & Francis) 236

Additive Migration Theory

Figure 14.12 Effective diffusion coefficient of Irganox 1076 as a function of maximum solvent absorption. Error bars are the SD for the determination of solvent absorption obtained for at least five measurements at equilibrium. (Reproduced with permission from Helmroth and co-workers, Food Additives and Contaminants, 2002, 19, 2, 176 [38]. © 2002, Taylor & Francis)

Table 14.7 Root of mean-square error (RMSE) as a measure of fit of the experimental data to the applied diffusion equation (duplicate or triplicate values) Solvent RMSE Tributyrin 0.011, 0.025 Tricaprylin 0.016, 0.020 Olive oil 0.087, 0.087, 0.15 Ethanol 0.02, 0.012 Isopropanol 0.010, 0.015 Ethylacetate 0.055, 0.047 Isooctane 0.033, 0.028 Cyclohexane 0.045, 0.048 Reproduced with permission from Helmroth and co-workers, Food Additives and Contaminants, 2002, 19, 2, 176 [38]. © 2002, Taylor & Francis

237

Additive Migration from Plastics into Foods Belhaneche-Bensemra and co-workers [48] studied the interactions between plasticised PVC packaging and food. The additives, including di-2-ethylhexyl phthalate (DEHP) plasticiser, tin-based heat stabiliser, processing aid, and internal and external lubricants, were first characterised and kinetic studies of their specific migrations were then carried out using various analytical methods such as GC, Fourier-transform infrared (FTIR) spectroscopy, atomic absorption spectrometry and dynamic scanning calorimetry analysis. The influence of various parameters such as temperature, stirring, nature of food simulant and initial concentration of plasticiser was investigated. Migration tests were conducted using four food simulants (sunflower oil, distilled water, 3% (w/v) aqueous acetic acid, and 15% (v/v) aqueous ethanol). These food simulants represent all fatty, liquid and moist foods and beverages, except for beverages with a high alcoholic strength. The test conditions used were 10 days at 25 °C and 45 °C with and without agitation. Ten circular samples were immersed in 100 ml of each food simulant. A circular sample and 10 ml of food simulant were taken off every day to be analysed. Each sample was wiped and weighed. The rate of variation of the mass (τ) was determined as a function of time following the relationship: τ (%) =

m t − m0 × 100 m0

(14.38)

Where: m0 is the initial mass before immersion, and mt is the mass of the sample at the time t. The rates of mass variation (τ) as a function of time gives information about the phenomenon which occurred between the samples and the food simulants used. An increase means that the food simulant penetrated the sample while a decrease means that some additives migrated into the food simulant. Hence, τ gives information about the overall migration that occurred. Figures 14.13 and 14.14 illustrate the variation of τ as a function of time for the four food simulants used at 45 °C with and without agitation, respectively. It can be seen that the highest rates of mass variations were obtained in sunflower oil indicating that migration of additives occurred. For aqueous simulants the rates of mass variations are practically the same and a very weak increase of τ was observed indicating penetration of the food simulant into the PVC discs. On the other hand, the highest values of τ were obtained during the migration tests with agitation and for the formulations containing the highest amounts of plasticiser. This is expected, since the solubility of DEHP is better in sunflower oil than in aqueous simulants. Furthermore, the migration of DEHP can involve the migration of the other additives which are present in the formulation like stabiliser and lubricant. 238

Additive Migration Theory

Figure 14.13 Effect of the nature of the food simulant on the rate of mass variation τ at 45 °C with agitation: (a) 60% DEHP; (b) 50% DEHP; (c) 40% DEHP; (d) 30% DEHP: ▲: Sunflower, l: 15% w/v aqueous ethanol, n: 3% w/v aqueous acetic acid, ✴ distilled water. (Reproduced from Belhaneche-Bensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [48]. © 2002, Wiley)

The effect of the initial concentration of DEHP on its specific migration in sunflower oil is illustrated by Figure 14.15. It is observed that the amount of migrating DEHP is related to its initial concentration in the PVC discs, to the temperature of migration testing and to the presence or absence of agitation. It is obvious that the mobility of the plasticiser molecules increased with increasing temperature and that the migration is favoured by agitation. Figures 14.16 and 14.17 illustrate the effect of the nature of food simulant on the migration of DEHP at 45 °C and 25 °C with agitation, respectively. In all the cases, it is observed that the amounts of migrating DEHP are more pronounced in sunflower oil than in aqueous ethanol. This is due to the better solubility of DEHP in sunflower oil. Furthermore, for the same initial concentration of plasticiser, the amounts of migrating DEHP are higher at 45 °C in the two food simulants. This is due to the fact that the mobility of the plasticiser molecules increased with increasing temperature. 239

Additive Migration from Plastics into Foods

Figure 14.14 Effect of the nature of the food simulant on the rate of mass variation τ at 45 °C without agitation: (a) 60% DEHP; (b) 50% DEHP; (c) 40% DEHP; (d) 30% DEHP: ▲: Sunflower, l: ethanol, n: acetic acid, ✴: water. (Reproduced from Belhaneche-Bensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [48]. © 2002, Wiley)

Flodberg and co-workers [49, 50] have demonstrated that liquid-crystalline co-polyesters (e.g., a random copolyester based on p-hydroxybenzoic acid (HBA) and 2-hydroxy-6naphthoic acid (HNA) known as Vectra A950) offer good barrier properties. For foodcontact use they require overall and specific migration testing and this is described in detail. For Vectra A950 films, the highest overall migration level obtained was 2.3 mg/dm2 (13.8 mg/kg) in olive oil (10 days at 40 °C), well below the EC limit of 10 mg/dm2 (60 mg/kg). The highest specific migration of HBA was 15.2 μg/dm2 (91.2 μg/kg) in olive oil (2 hours at 175 °C). In this case, the migration level was well below the EC limit of 10 mg/dm2 (60 mg/kg). For HNA, the highest value obtained was 4.3 μg/dm2 (26 μg/kg) in 10% ethanol (4 hours at 100 °C), well below the SML of 50 μg/kg. The results obtained show that even at these severe conditions, the migration values comply with the European Union Directive 2002/72/EC, which regulates plastic materials and articles for food contact use. In addition, the polymer Vectra A950 complies with Food Contact Notification No.103 of the United States Food and Drug Administration. Vectra A950 is therefore permitted for food-contact use both in the European Union and the USA. 240

Additive Migration Theory

Figure 14.15 Effect of the initial concentration of DEHP on its specific migration in sunflower oil: (a) 45 °C with agitation; (b) 45 °C without agitation; (c) 25 °C with agitation; (d) 25 °C without agitation: ▲: 60% DEHP, l 50% DEHP, ✴: 40% DEHP, n: 30% DEHP. (Reproduced from Belhaneche-Bensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [48]. © 2002, Wiley)

Ashby [51] has reviewed European Committee for Standardisation (CEN) test methods and technical specifications relating to the testing polymers in contact with foodstuffs. CEN TC194 Subcommittee 1 has been granted a mandate by the European Commission with respect to preparation of test methods for overall migration, preparation of test methods for SML (dose of an additive to food migrating simulents) and quantity of additive in food packaging (QM), preparation of test methods for measuring temperature at the food/plastics interface, preparation of test methods for the determination of molecular weight. CEN methods for overall migration are listed, together with details of CEN standards for monomers. Also included are details of ongoing drafting of standards for monomers, and details of future mandates for CEN methods and future work in the fields of additives and new monomers, estimation of exposure and application of factors, modelling, and threshold regulations. This review includes a discussion on test methods for determining overall migration of additives into olive oil, triglycerides, isooctane, 95% aqueous ethanol, fatty acid simulents and the measurement of SML and QM. Standardised methods for the determination of a wide range of monomers in food simulent liquids are also discussed. 241

Additive Migration from Plastics into Foods

Figure 14.16 Effect of the nature of food simulant on the migration of DEHP at 45 °C with agitation: (a) 60% DEHP; (b) 50% DEHP; (c) 40% DEHP: ▲: sunflower oil, n: ethanol. (Reproduced from Belhaneche-Bensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [48]. © 2002, Wiley)

Lee and Archer [52] report on a study on flow-induced migration of polyethyleneco-methacrylic acid (PE-co-MA) and polystyrene-b-polydimethylsiloxane (PS-bPDMS) copolymer additives in commercial long-chain branch PE and narrowmolecular distribution PS hosts in a capillary flow device. Attenuated total reflection Fourier transform infrared (ATR-FTIR) spectroscopy and dynamic contact angle measurements were used to characterise surface composition of polymer specimens following extrusion through metallic dies with various length-to-diameter (L/D) ratios, (1100 < L/D < 3000). Results from experiments covering a broad range of shear rates and polymer residence times in the dies are reported. Provided that the polymer residence time in the die is sufficiently long, shear is found to increase the concentrations of low molecular weight copolymer additives on the host polymer’s surface. The surface composition of copolymer additive is found to vary strongly with the wall shear rate and die L/D ratio. Decreasing the die diameter at fixed flow rate is found, for example, to be a more effective method for enhancing transport of additive to a polymer’s surface than increasing shear rate at fixed diameter. A mechanism based on shear-induced diffusion is proposed to explain the observed migration. ATR-FTIR 242

Additive Migration Theory

Figure 14.17 Effect of the nature of food simulant on the migration of DEHP at 25 °C with agitation: (a) 60% DEHP; (b) 50% DEHP; (c) 40% DEHP: ▲: sunflower oil, n: ethanol. (Reproduced from Belhaneche-Bensemra and co-workers, Macromolecular Symposia, 2002, 180, 1, 191 [48]. © 2002, Wiley)

absorbance spectra for PE/PE-co-MA are provided in Figure 14.18. The absorption peak at ~1750-1650 cm-1 is easily assigned to the asymmetric stretching vibration of C=O bonds in the methacrylic acid groups, while the absorption band at ~1465 cm-1 has been assigned to vibrations of C-H bonds present in the host polymer and copolymer additive. Surface concentrations of the PE-co-MA additive were therefore calculated using the ratio of the two peak heights, i.e., the C-H vibration was used as an internal reference for quantitative analysis. The migration of additives from plastics to food products or pharmaceuticals is an important problem in packaging technology. In spite of extensive research in this area, there is still a lack of understanding about the migration of large molecules (>200 g/mol) from a polymer matrix to a contacting solvent. Current migration research focuses mainly on the measurement of the additive concentration in the contacting solvent. Measurements in the solvent give a direct indication of contamination risk for food products or pharmaceuticals. However, they provide limited information 243

Additive Migration from Plastics into Foods

Figure 14.18 ATR-FTIR spectra for the surface and bulk of a PE-01/PE-co-MA blend extruded at a nominal wall shear rate γw of 0.1 s-1 using a capillary die (L/D = 102). Bulk samples were sectioned from the centreline of the extrudate at the same axial location where the surface chemical composition was evaluated using ATR-FTIR. (Reproduced with permission from Lee and co-workers, Polymer Engineering and Science, 2002, 42, 7, 1568 [52]. © 2002, Wiley)

about the migration process, which takes place inside the polymer. The equations that are used to describe the transport of molecules through polymers, are, in general, partial differential equations in terms of both time and space and are therefore best studied as such when insight in the transport process is required. However, very little work has dealt with the direct measurement of the local additive concentration in polymers, despite the advantage of obtaining information about the migration process as a function of both time and space. Slicing the polymer with a microtome has been shown to give some promising results [53-56], but the method still lacks satisfactory validation. Besides the purpose of studying complex migration processes, a proper evaluation of the method is also important as microtoming may be used to validate new and promising concentration profiling techniques, such as confocal microscopy [57, 58], Raman microscopy [59] and nuclear magnetic resonance [60-62]. Helmroth and co-workers [63] have described a method based on microtoming and GC analysis for studying transport processes of additives in polymers. As a test case, they used a combination of the polymer antioxidant Irganox 1076, LDPE, and the contacting solvent, ethanol [25, 46, 47, 64]. Validation of the method was performed in two steps: 244

Additive Migration Theory 1. By checking the mass balance of Irganox 1076 in both the polymer and solvent. The total amount of Irganox 1076 in the polymer and solvent should, at all time, be equal to the initial amount of Irganox 1076 in the polymer. 2. By comparing the experimental data with the diffusion equations of Fick [43, 65] that are known to describe the transport of Irganox 1076 from LDPE to ethanol [38]. In this work the mass balance is written as: MP,initial = MP,t + MS,t

(14.39)

Where: M is the amount (kg) of an additive in a polymer (MP,t) or solvent (MS,t), initially (t = 0) or after contact time t. The migration process in a polymer slab may be described by the second diffusion equation of Fick for unidirectional transport [65], see Equation (14.14). Where: C is the additive concentration (kg/m3), t

is the contact time (s),

x

is the position in the slab (m), and

D is the diffusion coefficient (m2/s). This equation was solved numerically for C(x,t) using the following initial and boundary conditions [66]: Initial conditions: CP(x,0) = C0

(14.40)

Cs(0) = 0

(14.41)

Boundary condition: CP(0,t) = CP(L,t) = CS(t)

(14.42)

Where: C0 is the initial additive concentration in the polymer, CP is the migrant concentration in packaging film, 245

Additive Migration from Plastics into Foods L

the polymer thickness; and

x

the position which ranges from 0 to 1.

The solution of Equation (14.14) using Equations (14.40-14.42) is based on the following assumptions: 1. Initially, the additive is homogeneously distributed throughout the polymer. 2. For the used polymer/additive/solvent combination and temperature, the diffusion coefficient (D) is constant. Effects due to penetration of the solvent into the polymer may be ignored due to the low interaction between LDPE (δ = 16.9 MPa½) and ethanol (δ = 26.0 MPa½) [67]. As the Irganox 1076 concentration in the polymer is low (<0.4%), concentration effects of Irganox 1076 itself may also be ignored. 3. There is no concentration gradient in the solvent, as diffusion through the polymer is much slower than is diffusion through ethanol. For comparison, the diffusion coefficient of Irganox 1076 in LDPE to ethanol at 40 °C is 1 × 10-13 m2/s [38] whereas the diffusion coefficient of molecules in liquids under ambient conditions is in the order of magnitude of 10-9 m2/s [45]. 4. The additive concentrations at both sides of the interface between the polymer and the solvent are equal, according to a partition coefficient (ratio of additive concentrations in polymer and solvent) of 1. This is justified as the solubility of Irganox 1076 in ethanol is good. In practice, as the solvent volume is large with respect to the polymer volume, the solvent concentration will always be almost equal to zero. Mathematically, the concentration in the polymer at the interface is assumed to be equal to the concentration in the solvent at one timestep earlier. This assumption is justified as long as the step size in time is very small. The concentration in the solvent was calculated using the mass balance (Equation 14.39). The ethanol extracts were analysed by GC with FID. The mass balance given by Equation (14.39) for Irganox 1076 in the polymer and solvent after different incubation times is presented in Table 14.8. The total amount of Irganox 1076 in each polymer slab piece was obtained by summing up the amount in all slices of half the piece and, assuming a symmetrical profile, multiplying this value by two (as it was technically not possible to slice the whole polymer slab piece). The amounts of migrated Irganox 1076 in ethanol corresponding in time with the measurements in the polymer were obtained by interpolation of the curve of Irganox 1076 concentrations in ethanol as a function of time. Ideally, the deviation, shown in the last column, should be zero. 246

Additive Migration Theory

Table 14.8 Mass balance of Irganox 1076 in polymer and solvent standardised to a polymer mass of 1 g Deviation a b t (min) MP (mg) MS (mg) MP + MS (mg) from Minitial (%)c 0

3.58 (Minitial)

0

3.58

0

235

3.37

0.40

3.77

5.2

975

3.23

0.59

3.82

6.7

1130

3.37

0.63

4.00

11.6

1455

3.08

0.70

3.78

5.6

14,400

2.13

1.63

3.76

5.0

25,920

1.56

2.13

3.69

3.1

Symbols explained in Equation (14.2) a

Sum of the amount in all slices of half the polymer slab piece multiplied by two.

Interpolated values of curve of Irganox 1076 concentration in ethanol as a function of time.

b

c

Deviation calculated as [(MP + Ms) - Minitial] Minitial × 100.

Reproduced with permission from Helmroth and co-workers, Journal of Applied Polymer Science, 2002, 86, 12, 3185 [63]. © 2002, Wiley

Figure 14.19(a-g) shows the experimental data of seven incubation times at 40 °C together with the best-fitting physical model given by Equation (14.40) using the least-square error criterion [66] for all observations simultaneously. Figure 14.19 confirms that initially the additive was homogeneously distributed in the polymer. If one considers each point as a separate estimate of the initial concentration, the average initial concentration was 3.58 mg/g (SD = 0.11 mg/g, n = 14). In general, the model fits the experimental data quite well throughout the whole process. The diffusion coefficient obtained was 1.1 × 10–13 m2/s. Helmroth and co-workers [63] conclude that their method using microtoming and GC analysis for the determination of additive migration in polymers as a function of position and time was valid for the combination LDPE/Irganox 1076/ethanol. The consistency of the mass balance was good and the concentration profiles inside the polymer corresponded to Fick diffusion equations. The diffusion coefficient obtained corresponded to that obtained from measurements of the Irganox 1076 concentration in ethanol as a function of time. Overwhelmingly, the method is well-suited, for studying additive migration inside polymers as a function of both space and time. 247

Additive Migration from Plastics into Foods

Figure 14.19 Concentration profiles of Irganox 1076 in LDPE for seven different incubation times at 40 °C expressed as the amount in the polymer at time t relative to the initial amount in the polymer per unit polymer weight. Experimental data are shown by crosses and the best-fitting curve by solid lines. (Reproduced with permission from Helmroth and co-workers, Journal of Applied Polymer Science, 2002, 86, 12, 3185 [63]. © 2002, Wiley)

14.2 The Total Migration Concept Various workers [68-72] have carried out extensive studies of the determination of total migration of additives from plastics packaging materials into edible fats using a 14 C-labelled fat simulent (HB 307). Methods based on measurement of total extract into volatile solvents such as n-heptone and diethylether were rejected at an early stage [73-86]. Unlike tests for specific migration, determination of the total migrate does not involve any direct toxicological considerations because both toxic substances and physiologically harmless packaging components are included. This test is therefore only used for checking new batches of already approved packaging materials, for which the total migration limits are known to the manufacturers. However, determining 248

Additive Migration Theory the total migrate from new packaging materials awaiting approval could give an important indication of total contamination and thus of the possible adulteration of the packed foodstuffs. Nevertheless, it has generally been considered that the determination of total migration is inadequate for a final physiological assessment of a new packaging material. Figge and others [70, 71, 86-92] described a radiochemical method in which the total amount of material that has migrated into the fat simulant HB 307 can be estimated indirectly from the decrease in weight of the packaging material being tested. A representative sample of the packaging material is weighed before and after storage in the radioactively-labelled fat simulant (HB 307-14C) and the amount of HB 307-14C retained in and on the sample in spite of careful cleaning is determined by radioanalysis. The suitability of the synthetic standard triglyceride mixture, HB 307, as a fat simulant is indicated by Table 14.9 which shows that the amounts of additives migrating from various test films into HB 307 are never more than 1.8 times as great as the values for the reference fats, Biskin, coconut oil and butter. The principles of the method for determining total migration into HB 307-14C have been described by Figge [68]. The total migrate (GM) is calculated from the radioanalytically determined amount (mg) of fat absorbed (FP) and the weight (mg) of the packaging-material sample before storage (Gv) and after storage (Gn) in HB 307-14C [Equation (14.43)]: GM = Gv - (Gn - FP) (mg)

(14.43)

The total migration during all-sided contact and the specific migration of plasticiser from one-sided contact of unplasticised PVC films with the fat stimulant – HB 307 after a contact time of 10 days at 40 °C amounted to 2.06 mg/dm2 film of 1.03 mg/dm2 contact area. Increases of 5% in the plasticiser content of the film to provide concentrations of 5, 10, 15, 20 and 25% led to increases in the amounts of total migrate of 7.2, 10.7, 23.1, 112.6 and 382.6 mg/dm2 film, respectively. The increase in total migration thus becomes larger the higher the plasticiser content of the film, as is also the case with the specific migration of plasticiser from PVC films into HB 307. This is clearly demonstrated in Figure 14.20 in which the specific migration of the plasticiser and the total migration into HB 307 are plotted as a function of the plasticiser content of the PVC test film. In addition to the increasing steepness of the migration curves for PVC films containing more than 15% plasticiser, this graph also shows the close similarity between the values for total migration from two-sided contact and for specific migration of the plasticiser from one-sided contact of the test film with the fat simulant over this range of plasticiser concentrations. This indicates that in a 10 day period at 40 °C, a PVC 249

Additive Migration from Plastics into Foods

Figure 14.20 Influence of plasticiser content of PVC test films on the specific migration (SM) of di-(2-ethyl-n-hexyl)phthalate calculated on contact area total migration (GM) calculated on the size of film (l), and on the contact area ( ), into the fat simulant HB 307-14C. (Reproduced with permission from K. Figge, S.R. Eder and H. Piater, Deutsche Lebensmittel-Rundschau, 1972, 68, 359 [91])

film containing more than 15% plasticiser imparts approximately the same amount of plasticiser to HB 307 whether the contact is one-sided or two-sided. Consequently, the total migration from a 1 dm2 piece of a highly plasticised PVC film should not be halved to obtain the value for a contact area of 1 dm2. It can in fact be seen from Figure 14.20 that the curve obtained by halving the total migration values in this way and plotting them against the plasticiser content of the films does not conform to the actual migration as indicated by the curve of the specific migration of plasticiser. Halving of the total migration values is only appropriate if the test films are not swollen by the test fat, or are only swollen at the surface.

250

Additive Migration Theory

Table 14.9 Comparison of the amounts of additive migrating from different test films into edible fats and the fat simulant HB 307 during one-sided contact for 60 days at 20 °C. Proportion of radioactivity or additive migrating into Test film*

Identity and concentration Coconut (%, w/v) Biskin Butter Oil of labelled additive

Rigid PVC

Irgastab 17 MOK-[14C]# (1.5)

0.009

0.014

Ionox 330[14C]§ (1.0)

0.090

Stearic acid [1 - 14C] amide (0.2)

HDPE

PS

Ratio R† for

HB 307

Biskin

Coconut Butter Oil

0.017

0.016

1.8

1.1

1.0

0.098

0.120

0.0

1.6

1.4

1.2

0.80

0.96

1.05

1.36

1.7

1.4

1.3

Ionox 330[14C] (2.0)

2.08

2.53

3.07

3.05

1.5

1.2

1.0

n-Butyl stearate [1 - 14C] (0.5)

5.20

5.61

7.11

7.57

1.5

1.4

1.1

* Figge and co-workers [68] R = amount of additive migrating into fat simulant HB 307/amount of additive migrating into Biskin, coconut oil or butter

†

#

Di-n-octyl[1-14C]-tin-bis-(2-ethylhexylthioglycollate)

§

1,3,5-Trimethyl-2,4-tris-(3,5-di-tert-butyl-4-hydroxybenzyl[14C]) benzene

Reproduced with permission from K. Figge, S.R. Eder and H. Piater, Deutsche Lebensmittel-Rundschau, 1972, 68, 359 [91]

251

Additive Migration from Plastics into Foods

References 1.

T. Garlanda and H. Masoero, La Chimica e l’Industria, 1966, 48, 9, 936.

2.

H. Masoero and T. Garlanda, La Chimica e l’Industria, 1965, 47, 9, 973.

3.

N. de Kruijf and M.A.H. Rijk, Food Additives and Contaminants, 1988, 5, 467.

4.

A.L. Baner, W. Bieber, K. Figge, R. Franz and O. Piringer, Food Additives and Contaminants, 1992, 9, 137.

5.

I. Cooper, Food Additives and Contaminants, 1998, 15, 72.

6.

A.L Baner, R. Franz and O. Piringer, Deutsche Lebensmittel-Rundschau, 1994, 90, 181.

7.

A.L Baner, J. Brandsch, R. Franz and O. Piringer in the Proceedings of the 8th ICI/PIRA International Symposium on Plastics for Packaging Food, Washington, DC, USA, 1995.

8.

A.L Baner, J. Brandsch, R. Franz and O. Piringer, Food Additives and Contaminants, 1996, 13, 5, 587.

9.

T.H. Begley in Proceedings of the PIRA Conference on Plastics for Packaging Food, Prague, Czechoslovakia, 1997.

10. W. Limm and H.C. Hollifield, Food Additives and Contaminants, 1996, 13, 8, 949. 11. O. Piringer in Proceedings of the PIRA Conference on Plastics for Packaging Food, Prague, Czechoslovakia, 1997. 12. O. Piringer in Proceedings of the PIRA Conference on Plastics in Contact with Foodstuffs, London, UK, 1997. 13. A. O’Brien, A. Goodson and I. Cooper, Food Additives and Contaminants, 1999, 16, 9, 367. 14. A. O’Brien and I. Cooper, Food Additives and Contaminants, 2002, 19, Supplement 1, 63. 15. P.A. Tice and I. Cooper in Food Science Reviews, Volume 2, Chemical Migration from Food Packaging, Eds., D.H. Watson and M.N. Meah, Ellis Horwood, Chichester, UK, 1994, p.3-33. 252

Additive Migration Theory 16. I. Cooper and P.A. Tice, Food Additives and Contaminants, 1995, 12, 2, 235. 17. A.P. O.Brien, I. Cooper and P.A. Tice, Food Additives and Contaminants, 1997, 14, 6-7, 705. 18. A.P. O’Brien, I. Cooper, A. Goodson and J. Simal-Gándara in Proceedings of the PIRA Conference on Plastics for Packaging Food, Prague, Czechoslovakia, 1997. 19. European Council Directive 85/572, Laying down the list of simulents to be used for testing migration of of constituent plastic materials and articles intended to come into contact with foodstuffs, EEC, Brussels, Belgium, 19th December 1985. 20. A.P. O’Brian and I. Cooper in Proceedings of the PIRA Conference on Plastics in Contact with Foodstuffs, London, UK, 1997. 21. J. Brandsch, P. Mercea, V. Tosa and O. Piringer in Proceedings of the PIRA Conference on Plastics in Contact with Foodstuffs, Edinburgh, UK, 1999. 22. D. Chung, S.E. Papadakis and K.L. Yam, Food Additives and Contaminants, 2002, 19, 6, 611. 23. T.H. Begley, Food Additives and Contaminants, 1997, 14, 6-7, 545. 24. R.C. Reid, K.R. Sidman, A.D. Schwope and D.E. Till, Industrial and Engineering Chemistry Product Research and Development, 1980, 19, 4, 580. 25. R. Goydan, A.D. Schwope, R.C. Reid and G. Cramer, Food Additives and Contaminants, 1990, 7, 3, 323. 26. T.H. Begley and H.C. Hollifield, Food Technology, 1993, 47, 11, 109. 27. W. Limm and H.C. Hollifield, Food Additives and Contaminants, 1995, 12, 609. 28. M. Hamdani, M. Feigenbaum and J-M. Vergnaud, Food Additives and Contaminants, 1997, 14, 5, 499. 29. T.D. Lickly, M.L. Rainey, L.C. Burgert, C.V. Breder and L. Borodinsky, Food Additives and Contaminants, 1997, 14, 1, 65. 30. J. Miltz, A. Ram and M.M. Nir, Food Additives and Contaminants, 1997, 14, 649. 253

Additive Migration from Plastics into Foods

Additive Migration Theory

31. A.M. Riquet and A. Feigenbaum, Food Additives and Contaminants, 1997, 14, 1, 53. 32. J. Crank, The Mathematics of Diffusion, 2nd Edition, Clarendon Press, Oxford, UK, 1975, p.44-68. 33. T.P. Gandek, T.A. Hatton and R.C. Reid, Industrial and Engineering Chemistry, 1989, 28, 7, 1030. 34. D.H. Chung, S.E. Papadakis and K.L. Yam, Journal of Food Processing and Preservation, 2001, 25, 71. 35. D. Chung, Antimicrobial activities and release phenomena in packaging materials containing propylparaben or triclosan, Rutgers University, New Brunswick, NJ, USA, 2000. [PhD Thesis]. 36. J. Miltz in Handbook of Food Engineering, Eds., D.R. Heldman and D.B. Lund, Marcel Dekker, New York, NY, USA, 1992, p.678-718. 37. E. Kreyszig, Advanced Engineering Mathematics, 7th Edition, Wiley, New York, NY, USA, 1993, p.805. 38. I.E. Helmroth, M. Dekker and T. Hankemeier, Food Additives and Contaminants, 2002, 19, 2, 176. 39. O.G. Piringer, Food Additives and Contaminants, 1994, 11, 2, 221. 40. J. Koszinowski, Journal of Applied Polymer Science, 1986, 31, 6, 1805. 41. L. Piergiovanni, P. Fava and A. Schiraldi, Food Additives and Contaminants, 1999, 16, 8, 353. 42. A. Reynier, P. Dole and A. Feigenbaum, Food Additives and Contaminants, 1999, 16, Supplement 1, 42. 43. Diffusion in Polymers, Eds., J. Crank and G.S. Park, Academic Press, London, UK, 1968. 44. C.M. Hansen, Hansen Solubility Parameters: A Users Handbook, CRC Press, Boca Raton, FL, USA, 2000. 45. J. Wesselingh and R. Krishna, Mass Transfer in Multicomponent Mixtures, Delft University Press, Delft, The Netherlands, 2000. 46. E. Földes, Journal of Applied Polymer Science, 1993, 48, 11, 1905. 254

Additive Migration Theory 47. J.Y. Moisan, European Polymer Journal, 1980, 16, 10, 979. 48. N. Belhaneche-Bensemra, C. Zeddam and S. Ouahmed, Macromolecular Symposia, 2002, 180, 1, 191. 49. G. Flodberg, L. Hojväll, M.S. Hedenquist and U.W. Gedde, Food Additives and Contaminants, 2003, 20, 3, 313. 50. G. Flodberg, L. Hojväll, M.S. Hedenquist and U.W. Gedde, Food Additives and Contaminants, 2002, 19, 5, 492. 51. B. Ashby in Proceedings of a PIRA International Conference on Completion of a CEN Mandate and Future Programmes, Coventry, UK, 2001. 52. H. Lee and L.A. Archer, Polymer Engineering and Science, 2002, 42, 7, 1568. 53. V. Dudler and C. Muiños in Polymer Durability: Degradation, Stabilisations and Lifetime Prediction, Eds., R.L. Clough, N.C. Billingham and K.T. Gillen, Advances in Chemistry Series No.249, ACS, Washington, DC, USA, 1996, p.441-453. 54. D. Messadi, J-M. Vergnaud and M. Hivert, Journal of Applied Polymer Science, 1981, 26, 2, 667. 55. K. Figge and F. Rudolph, Die Angewandte Makromolekulare Chemie, 1979, 78, 1, 157. 56. D. Richman and F.A. Long, Journal of the American Chemical Society, 1960, 82, 3, 509. 57. Y. Song, M. Srinivasarao, A. Tonelli, C.M. Balik and R. McGregor, Macromolecules, 2000, 33, 12, 4478. 58. D. de Beer, P. Stoodley and Z. Lewandowski, Biotechnology and Bioengineering, 1997, 53, 2, 151. 59. C. Sammon, S. Hajatdoost, P. Eaton, C. Mura and J. Yarwood, Macromolecular Symposia, 1999, 141, 247. 60. A.A. Goodwin, A.K. Whittaker, K.S. Jack, J.N. Hay and J. Forsythe, Polymer, 2000, 41, 19, 7263. 61. M. Knörgen, U. Heuert, H. Schneider, D. Kuckling, S. Richter and K.F. Arndt, Macromolecular Symposia, 1999, 145, 83. 255

Additive Migration from Plastics into Foods 62. J.L. Koenig in Magnetic Resonance Microscopy – Methods and Applications in Materials Science, Agriculture and Biomedicine, Eds., B. Blümich and W. Kuhn, VCH, Weinheim, 1992, p.168. 63. I.E. Helmroth, H.A.M. Bekhuis, J.P.H. Linssen and M. Dekker, Journal of Applied Polymer Science, 2002, 86, 12, 3185. 64. T.P.Gandek, Migration of Phenolic Antioxidants from Polyolefins to Aqueous Media with Application to Indirect Food Additive Migration, Massachusetts Institue of Technology, Cambridge, MA, USA, 1986. [PhD Thesis] 65. J. Crank, The Mathematics of Diffusion, 2nd Edition, Clarendon, Oxford, UK, 1975. 66. W.H. Press, B.P. Flannery, S.A. Teukolsky and W.T. Vetterling, Numerical Recipes in C: The Art of Scientific Computing, 2nd Edition, Cambridge University Press, Cambridge, UK, 1994. 67. Polymer Handbook, 4th Edition, Eds., J. Brandrup, E.H. Immergut and E.A. Grulke, Wiley, New York, NY, USA, 1999. 68. K. Figge, Food Cosmetics and Toxicology, 1973, 11, 6, 963. 69. K. Figge, Bundesgesundheitsblatt, 1975, 24, 18, 27. 70. K. Figge, Deutsche Lebensmittel-Rundschau, 1975, 71, 129. 71. K. Figge and J. Koch, Fette Seifen Anstrichmittel, 1975, 77, 184. 72. W.G. Aldershoff, Annali Dell’Istituto Superiore di Sanita, 1972, 8, 550. 73. E. Baumgartner, Kunststoffe-Plastics, 1968, 15, 3. 74. Federal Health Office, Berne Schweizerisches Lebensmittelbuch, Vorentwurf, V. Auflage, Band III, Kap 48 - Kunststoffe, 1966. 75. K. Figge in Proceedings of the Gesellschaft Deutscher Chemiker (GDCh) meeting on Preparative Radiochemie, Lindau/Bodensee, 1968. 76. United States Code of Federal Regulations, Section 121, 2501. 77. United States Code of Federal Regulations, Section 121, 2514. 78. K. Figge and H. Piater, Deutsche Lebensmittel-Rundschau, 1972, 68, 313. 256

Additive Migration Theory 79. Deutsche Lebensmittel-Rundschau, 1972, 68, 37. 80. R. Brugger, University of Bern, Switzerland, 1971. [PhD Thesis] 81. J.H. de Wilde, Deutsche Lebensmittel-Rundschau, 1966, 61, 369. 82. E. Flückiger and C. Rentsch, Alimenta, 1968, 7, 41. 83. L. Robinson-Görnhardt, Kunststoffe, 1957, 47, 265. 84. L. Robinson-Görnhardt, Kunststoffe, 1958, 48, 463. 85. Determination de la migration des constituents des materiaux destinés a être mis au contact des denrées alimentaires ayant un contact gras, BITMP (Bureaux Internationaux Techniques des Matières Plastiques), Brussels, Belgium, 1971. 86. W. Pfab, Annali Dell’Istituto Superiore di Sanita, 1972, 8, 385. 87. D. van Battum and M.A.H. Rijk, Annali Dell’Istituto Superiore di Sanita, 1972, 8, 421. 88. G. Wilbrett, K-W. Evers and F. Kiermeier, Zeitschrift für LebensmittelUntersuchung und -Forschung, 1970, 142, 205. 89. J. Koch, Deutsche Lebensmittel-Rundschau, 1972, 66, 216. 90. K. Figge, Food Cosmetics and Toxicology, 1972, 10, 6, 815. 91. K. Figge, S.R. Eder and H. Piater, Deutsche Lebensmittel-Rundschau, 1972, 68, 359. 92. K. Figge, Kunststoffe, 1971, 61, 11, 832.

257

Additive Migration from Plastics into Foods

258

15

Gas Barrier Properties of Food Packaging Plastic Films

The introduction of polymer-based structures as packaging materials for foodstuffs has been increasing over the last few decades. The main commercial appeal of these materials lies on their ability to offer a broad variety of tailor-made properties and yet be cheap, and easily processed and conformed into a myriad of shapes and sizes. Given the diversity of food products and the various packaging requirements of those, a large number of packaging technologies have also been put into place, i.e., multilayer structures, modified and equilibrium modified atmosphere packaging, active packaging, and so on [1]. Nevertheless, one of the limiting properties of polymeric materials in the food packaging field is their inherent permeability to low molecular weight substances, including permanent gases, water and organic vapours. This has boosted the interest for developing new resins with higher barrier properties and to carry out research aimed at the understanding of the structure/barrier properties relationship. The most efficient and widely used high barrier polymeric materials are the ethylenevinyl alcohol copolymers (EVOH), although there are also other new high barrier material developments and specific improvements like perfectly alternating aliphatic polyketones [2] and thermoformable EVOH-based blends [3]. Barrier properties in polymers are necessarily associated to their inherent ability to permit the exchange, to a higher or lower extent, of low molecular weight substances through mass transport processes like permeation. The permeation of low-molecular weight chemical species usually takes place through the polymer amorphous phase and is generally envisaged as a combination of two processes, i.e., sorption and diffusion. A permeate gas is first sorbed into the upstream face of the polymer film, and then, undergoes a molecular diffusion to the downstream face of the film where it desorbs into the external phase again. A sorption-diffusion mechanism is thus applied, which can be formally expressed in terms of permeability (P), being this the product of solubility (S) and diffusion (D) coefficients, as defined by Henry’s and Fick’s laws, respectively. Transport of gases and vapours in polymers is an important subject both from the technological and scientific point of view. Applications include protective coatings, 259

Additive Migration from Plastics into Foods packing materials for food, and selective barriers for gas or liquid mixtures. In the case of food packages, polymeric materials should exhibit an adequate CO2/O2 ratio (generally lower than 7) [4]. The process of permeation involves dissolution of the gas in one side of the membrane, diffusion of the gas through it and release of the gas from the other side of the membrane. When dealing with glassy polymers, it is important to note that permeability characteristics depend on thermal history of the polymer [5]. A conventional method for determining permeability and diffusion coefficients in polymers involves the measurement of membrane weight gain versus time until the final mass of equilibrium mass is reached. As consumers demand improved product quality and longer shelf life, there is a continuing need for improved barrier properties and extended shelf life in packaging for food and beverages, cosmetics, and pharmaceuticals. Additives that help create a protective package environment, polymers with good barrier properties, processes for creating modified atmosphere packaging and sensors to measure the package environment can be combined. It is expected that within the next decade, active and intelligent packaging options will become key elements in how food processors and manufacturers protect the longevity and nutrient value of their products. Active packaging responds to changes in the package environment. For example, when the relative humidity in a package reaches a certain level, a desiccant in the package will begin to absorb moisture. Active packaging systems may absorb molecules such as oxygen, ethylene, or moisture, or release agents such as antimicrobials or flavours. While many active packaging technologies such as desiccants, odour scavengers and ethylene absorbers are commonly used in sachets that are inserted into a package, there is a drive to find ways to incorporate active packaging technologies directly into the package walls. Consumers are looking for foods without any preservatives, but they still want foods to be long-lasting. Putting additives into the package itself can fulfill this desire. Some of the additives being considered or used for compounding into plastic packaging are oxygen scavengers and absorbers, flavour and odour controllers and antimicrobials. Antifogging agents and UV absorbers also play an important role. Active packaging has been used for years in Japan, and in recent years active packaging has been used in more limited areas in Europe and the US. Europe does not have many commercial products on the market due to legislation restrictions. However, in 2003, the Commission of the European Community accepted amendments to a framework directive for food contact materials that will allow the use of active and intelligent packaging concepts. It is expected that they will become law in 2005 and 2006. Market studies predict significant growth once the legislation changes. Growth is likely to occur first in proven applications such as oxygen scavengers and later in more 260

Gas Barrier Properties of Food Packaging Plastic Films untried areas such as preservative releasing packaging systems. A directive specific to active and intelligent packaging is also being discussed. As more examples of active packaging become commercial, they will give direction to the specific directive. In the US, regulatory issues limit antimicrobial use, but other areas in which active packaging can add value are growing. Active packaging has good growth potential in single-serve and fresh cut produce packaging where higher value, higher margin products support the use of more sophisticated packaging systems that extend shelf life. Photoacoustic spectroscopy and related photothermal products [7] are well-established spectroscopic techniques. The photoacoustic technique, apart from providing direct optical absorption spectra, can also be used to perform depth profile analysis, and characterisation of thermal properties. In addition, there has been a substantial development of new versatile and competitive instrumentation and experimental methodologies suitable for use in daily practice [6, 7]. Further details on the photothermal wave phenomenon and its applications can be found in the books by Rosencwaig [6] and Almond [7] and in some of many published reviews on the subject [8-10]. Recently, photothermal techniques have been applied to the determination of diffusion coefficients in biopolymers [11]. However, the complete validation of this new methodology demands the characterisation of diffusion coefficients of plastic films used commercially, which has been achieved successfully using traditional gravimetric techniques. Regarding this, the aim of work by Poley and co-workers [12] was to demonstrate the potentiality of photothermal techniques in the plastic packages quality control studying the diffusion properties of commercial low-density polyethylene (LDPE) packages, a widespread used material for wrapping food and vegetables in day-to-day life. In order to support the gas diffusion results, thermophysical properties and crystallinity degree of samples were also determined. This work involved the use of photothermal techniques for determining diffusion coefficients of oxygen and carbon dioxide of commercial LDPE. The methodology involved the monitoring of diffused gas by a photoacoustic analyser. Diffusion coefficients measured for carbon dioxide and oxygen were 2.77 × 10-8 cm2/s and 16.8 × 10-8 cm2/s, respectively. To support the gas diffusion results, thermal properties were studied using photoacoustic spectroscopy and crystallinity was determined using X-ray diffraction. Values obtained for thermal diffusivity and specific heat capacity were 0.00165 cm2 and 2.33 J/cm3/K, respectively, which were in good agreement with values available in the literature for pure LDPE and thus assured the reliability of the diffusion coefficient values. The room temperature characterisation of the LDPE sample’s thermal properties was based upon the measurements of the thermal diffusivity, α, and of the heat capacity, 261

Additive Migration from Plastics into Foods ρcp, where ρ is the material density and cp the specific heat at constant pressure (heat capacity per unit volume). To complete the determination of the sample thermal properties, knowing α and ρcp, the sample thermal conductivity, k, is readily obtained from Equation 15.1: K = α ρcp

(15.1)

The thermal diffusivity can be accurately measured by the photoacoustic technique. This technique looks directly at the heat generated in a sample, due to a thermal relaxation process, following the absorption of light. Among several experimental set-ups the open photoacoustic cell method was used by Poley and co-workers [12]. It consists of mounting the samples directly onto a cylindrical electret microphone and using the front air chamber of the microphone itself as the usual gas chamber of conventional photoacoustic cell. As a result of a periodic heating of the sample following the absorption of modulated light, the pressure in the microphone chamber oscillates at the chopping frequency yielding the photoacoustic signal. The experimental arrangement (Figure 15.1) consisted of a 100 mW Ar laser whose beam was modulated with a mechanical chopper. The sample was placed directly above the opening of the microphone covering it. The signal from the microphone was connected to a lock-in amplifier used to register both signal amplitude and phase. The contribution to the photoacoustic signal from the thermoelastic bending is the dominant. This effect is essentially due to the temperature gradient created inside the sample along an axis perpendicular to the surface exposed to incident radiation [13]. Permeation studies of carbon dioxide and oxygen were performed using a sample holder having its sides sealed off one from another. The permeate gas being studied was introduced on one side and kept at a constant pressure of 0.10 MPa. On the other side the concentration of the gas was measured using a photoacoustic gas analyser as it was being stored (Figure 15.2). The gas concentration rise, in these conditions, is given by [11]:

(

)

C = C0 1 − e−1/ τD Where: C0 is the gas concentration at saturation, τD = ls2/2D is the gas diffusion time, D

is the gas diffusion coefficient, and,

ls

is the sample thickness (40 μm) in this case.

262

(15.2)

Gas Barrier Properties of Food Packaging Plastic Films

Figure 15.1 Experimental arrangement for photoacoustic thermal diffusivity (a) and specific heat capacity (b) measurements. (Reprinted with permission from Poley and co-workers, Polimeros: Ciencia e Tecnologia, 2004, 205, 225 [12]. © 2004, Associacao Brasileira de Polimeros)

Thus, by monitoring the time evolution of the gas concentration, the gas diffusion time, τD, can be determined and thus, the diffusion coefficient D. All measurements were performed at 27 °C. Figure 15.3 shows a typical photoacoustic signal frequency dependence for LDPE. It can be seen that for modulation frequencies higher than 100 Hz, the signal amplitude scales essentially as f-0.94 a value close to f-1. This frequency dependence of the photoacoustic signal of a thermally thick sample confirms the thermoelastic bending as the dominant mechanism responsible for the acoustic signal [14]. Accordingly, the thermal diffusivity was evaluated. 263

Additive Migration from Plastics into Foods

Figure 15.2 Experimental arrangement used for gas diffusion studies. (Reprinted with permission from Poley and co-workers, Polimeros: Ciencia e Tecnologia, 2004, 205, 225 [12]. © 2004, Associacao Brasileira de Polimeros)

Figure 15.3 Modulation frequency dependence of photoacoustic signal for LDPE. (Reprinted with permission from Poley and co-workers, Polimeros: Ciencia e Tecnologia, 2004, 205, 225 [12]. © 2004, Associacao Brasileira de Polimeros) 264

Gas Barrier Properties of Food Packaging Plastic Films

Figure 15.4 Oxygen (a) and carbon dioxide (b) concentration evolution for LDPE. The solid line represents the best fit of the experimental data to Equation 15.2 using τD as an adjustable parameter. (Reprinted with permission from Poley and co-workers, Polimeros: Ciencia e Tecnologia, 2004, 205, 225 [12]. © 2004, Associacao Brasileira de Polimeros)

Concentrations of carbon dioxide and oxygen in the analyser chamber as a function of time are presented in Figure 15.4. The fitting of experimental data to Equation (15.2) allows the determination of diffusion coefficients. The values obtained for diffusion coefficients of LDPE to oxygen and carbon dioxide were 1.68 × 10-7 cm2/s and 2.77 × 10-8 cm2/s, respectively. These values agree with those presented in literature for LDPE [15] (6.9 × 10-8 cm2/s for carbon dioxide and 4.6 × 10-7 cm2/s for oxygen). High barrier polymers are of increasing interest and use nowadays in food packaging applications. The reason for this is that packaging [15] is being more often regarded as a very efficient and convenient means of preserving foodstuffs from deterioration during products handling, transport and shelf-life. Vibrational spectroscopy is one of the most widely used techniques for the morphological characterisation of high gas and aroma barrier materials in commercial food packaging, i.e., the EVOH copolymers. Nevertheless, and as a result of the highly hydrophobic character of the EVOH resins, it is usually LDPE the high water barrier material that is put in direct contact with foods, i.e., multi-layer EVOH - LDPE film. The excellent properties of EVOH resins in terms of high gas, hydrocarbon and aroma barrier, and transparency have allowed them to become widely implemented in many 265

Additive Migration from Plastics into Foods commercial applications where high barrier properties are needed to minimise product losses or deterioration. Despite the excellent performance of these materials in high barrier food packaging applications, the materials are easily plasticised by moisture and, consequently, in most packaging applications are commonly encapsulated in multilayer structures between hydrophobic polymers such as polypropylene or polyethylene. Various workers have studied the application of Fourier-transform infrared spectroscopy (FTIR) to determine diffusion coefficients of mostly water vapour, in various polymers [17-20]. In particular, the work of Lagaron [16] is discussed next. Lagaron and co-workers [16] used vibrational spectroscopy to characterise the morphology and barrier properties of polymers used in food packaging applications. Raman and infrared (IR) spectroscopies were used for morphological characterisation of EVOH copolymers (Soarnol from Nippon Synthetic Chemical Industry), one of the families of high gas and aroma barrier materials most widely used in commercial food packaging. The effects of thermal history, annealing and thermal degradation of the copolymers on their crystallinity, morphological and chemical structure were examined. FTIR spectroscopy was used to determine the diffusion coefficients of citrus fruit aroma components in LDPE. Figure 15.5 shows typical substraction FTIR spectra of a chosen band characteristic of the citrus fruit aroma limonene as a function of time during desorption from LDPE.

Figure 15.5 Subtraction spectra from a completely desorbed LDPE sample of a CH deformation mode of limonene as a function of time. (Reprinted with permission from Lagaron and co-workers, Macromolecular Symposia, 2004, 205, 225 [16]. © 2004, Wiley) 266

Gas Barrier Properties of Food Packaging Plastic Films By fitting the relative area of the chosen band (At/Ae) as a function of time to the modified solution form of the second law of Fick for a desorption case in an FTIR experiment [Equation (15.3)] it is easy to derive the corresponding diffusion coefficient:  −D(2n + 1)2 π2t  At 8 ∞ 1 = 2∑ exp   A e π n =0 (2n + 1)2 L2  

(15.3)

In Equation (15.3) (only strictly valid if we assume Fickian behaviour, no changes in film thickness and constant absorption coefficient), At is the absorbance of the chosen band at a given time, t, and Ae is the absorbance at saturation or equilibrium sorption, L is the thickness of the film and D the diffusion coefficient. Lagaron and co-workers [16] carried out measurements for a number of aroma components, namely, limonene, α-pinene and citral and results are given in Table 15.1.

Table 15.1 Diffusion coefficients (m2/s) of limonene, α-pinene and citral Aroma component D (gravimetry) D (FTIR) (18.5 ± 0.6) × 10-13 Limonene (20.0 ± 0.3) × 10-13 α-Pinene (9.7 ± 0.2) × 10-13 (9.6 ± 0.6) × 10-13 Citral (3.5 ± 0.2) × 10-13 (5.5 ± 0.1) × 10-13 Reprinted with permission from Lagaron and co-workers, Macromolecular Symposia, 2004, 205, 225 [16]. © 2004, Wiley In order to compare results obtained in the determination of diffusion coefficients by FTIR they also measured desorption of these penetrants by gravimetry, i.e., by weight loss measurements. Figure 15.6 shows the corresponding fits of the desorption experiments of α-pinene by FTIR and gravimetry showing an excellent correspondence. In the case of limonene and citrate small differences between FTIR and gravimetric results occur (Figure 15.6). The general criteria for selecting a good IR band in terms of evaluating transport properties is that it should not undergo significant changes in shape and position during desorption, because this is indicative of vibrational modes highly sensitive to the different levels of interactions that can be established, depending on the sorption level, between the penetrant and the polymer matrix. The reason why strong frequency shifts and changes in band shape, often the case for N-H or O-H stretching bands and CO stretching bands, may not give good correlations with for example gravimetry, lies in the fact that absorption coefficients are changing during the desorption process, and therefore, the relation between absorbance and concentration may be significantly different during the experiment. 267

Additive Migration from Plastics into Foods

Figure 15.6 Fickian-like desorption experiments of α-pinene in LDPE by FTIR and gravimetry. (Reprinted with permission from Lagaron and co-workers, Macromolecular Symposia, 2004, 205, 225 [16]. © 2004, Wiley)

Auras and co-workers [21] studied the variations in the oxygen diffusion, solubility, and permeability coefficients of polylactide (PLA) films at different temperatures (5, 23, and 40 °C) and water activities (0-0.9). The results were compared with the oxygen diffusion, solubility, and permeability coefficients obtained for polyethylene terephthalate (PET) films under the same experimental conditions. The water sorption isotherm for PLA films was also determined. Diffusion coefficients were determined with the half-sorption time method. Also, a consistency test for continuous-flow permeability experimental data was run to obtain the diffusion coefficient with the lowest experimental error and to confirm that oxygen underwent Fickian diffusion in the PLA films. The permeability coefficients were obtained from steady-state permeability experiments. The results indicated that the PLA films absorbed very low amounts of water, and no significant variation of the absorbed water with the temperature was found. The oxygen permeability coefficients obtained for PLA films (2-12 × 10-18 kg.m/m2.s.Pa) were higher than those obtained for films (1-6 × 10-19 kg.m/m2.s.Pa) at different temperatures and water activities. Moreover, the permeability coefficients for PLA and PET films did not change significantly with changes in the water activity at temperatures lower than 23 °C. A modified Oxtran 100-Twin with a coulometric sensor (Mocon, Inc., Minneapolis, MN) was used by Auras and co-workers [21] for measuring the permeability of oxygen through the polymer films as a function of Aw (Aw is the ratio of the partial pressure 268

Gas Barrier Properties of Food Packaging Plastic Films

Figure 15.7 Schematic diagram of the test apparatus for permeability experiments. (Reprinted with permission from Auras and co-workers, Journal of Applied Polymer Science, 2004, 92, 3, 1790 [21]. © 2004, Wiley)

of water to the saturation vapour pressure of water at a specified temperature). The experiments were carried out at 0.021 MPa of pressure for PLA films and at 0.1 MPa pressure for PET films. Figure 15.7 shows a schematic diagram of the test apparatus. Before a polymer film was mounted in a cell, each film was vacuum-dried at 60 °C for at least 10 hours. After the films were mounted to cells A and B (Figure 15.7), a stream of dry nitrogen was circulated through the system. For each Aw value at which the permeability of oxygen was measured, a nitrogen stream adjusted to the selected Aw value flowed throughout a cell containing a film for at least 24 hours. During this period, each film acquired a moisture content in equilibrium with that in the nitrogen stream. The initial time was marked when an oxygen gas stream at the selected Aw value, equal to Aw of the film in the cell, was introduced. The amount of oxygen that permeated per unit of time was continuously monitored with a Mocon 269

Additive Migration from Plastics into Foods

Figure 15.8 Permeation flow fraction versus time for oxygen permeation experiments for PET films at 5, 23, and 40 °C (Aw = 0.6). (Reprinted with permission from Auras and co-workers, Journal of Applied Polymer Science, 2004, 92, 3, 1790 [21]. © 2004, Wiley)

D-200 integrator until a steady state was reached (i.e., the oxygen flux changed by less than 1% for 30 minutes). The instrument was calibrated at each temperature at which the samples were tested (i.e., 5, 23, and 40 °C) through the measurement of the oxygen flow rate through a standard reference material provided by Mocon. Aw values were determined with a Hygrodynamics model 1820A wide-range humidity sensor (Newport Scientific, Inc., Jessup, MD, USA). The experimental error in the determination of the oxygen flow rate on the Oxtran 100-Twin apparatus was estimated to be 2-5%, with highly reproducible and consistent readings. The experimental error in the measurement of Aw was estimated to be 2%. A typical plot of oxygen permeation as a function of time is presented in Figure 15.8 for PET films at 5, 23 and 40 °C at Aw = 0.9. The oxygen barrier properties of PET and PLA films at three different temperatures (5, 23, and 40 °C) were measured. The permeability coefficients of PET and PLA decreased as the water content increased. An increase in the permeability coefficient of PET and PLA was observed as the temperature increased. Oxgen diffusion in PET and PLA showed an exponential increase as Aw increased at each temperature. The effect was more pronounced at higher temperatures. The oxygen solubility coefficient decreased linearly as Aw increased because of the reduction in the free volume due to its occupation by water molecules. 270

Gas Barrier Properties of Food Packaging Plastic Films Markarian [23] has reviewed the latest development of types of additives used to improve gas barrier and other properties of food packaging, cosmetics and pharmaceutical products. These include oxygen scavengers, flavour and odour controllers, antimicrobials, antifogs and UV absorbers. Barrier packaging oxygen scavengers discussed in this review include ascorbic acid, iron oxide (in sachets), cobalt catalysed Nylon MXD6, Ciba Shelfplus O2 oxygen scavenger, Chevron Phillips Chemical C6OSP oxygen scavenger (based on an oxidisable resin, ethylene methyl acrylate cyclohexane, methylacrylate containing a photoiniator and a cobalt salt catalyst) and BP Chemicals Anosorb oxygen scavenger. Mazzocca [24] stuidied the interactions between foodstuffs, polymers and environments (vapours, micro- and macro-organisms and radiation) leading to migration of additives from plastics packaging films into foods and diffusion of materials through the films. Toxicological implications of migration are considered. The gas and water vapour permeability, chemical resistance and resistance to biological attack of plastics commonly used in packaging applications are also discussed.

References 1.

J.M. Lagaron, R. Catala and R. Gavara, Materials Science and Technology, 2004, 20, 1, 1.

2.

J.G. Bonner and A.K. Powell in Proceedings of CSIR New Plastics ’98 Conference, London, UK, 1998.

3.

R. Gavara, R. Catala, J.M. Gimenez, J.M. Lagaron and C. Sanz in Proceedings of WorldPak2002: Improving the Quality of Life Through Packaging Innovation, Volume 1, CRC Press LLC, Boca Raton, FL, USA, 2002, p.400.

4.

S. Marais, J.M. Saiter, C. Devallencourt, Q.T. Nguyen and M. Metayer, Polymer Testing, 2002, 21, 4, 425.

5.

V. Compari, A. Ribes, R. Diaz-Callega and E. Rianda, Polymer, 1996, 37, 11, 2243.

6.

A. Rosencwaig, Photoacoustic and Photothermal Spectroscopy, 2nd Edition, RE Krieger Publishing Company, Malabar, FL, USA, 1990.

7.

D. Almond and P. Patel, Photacoustic and Photothermal Sciences and Techniques, Chapman and Hall, London, UK, 1996. 271

Additive Migration from Plastics into Foods 8.

H. Vargas and L.C.M. Miranda, Physics Reports, 1988, 161, 2, 43.

9.

J.B. Kinney and R.H. Staley, Annual Review of Materials Science, 1982, 12, 295.

10. A Torres Filho, N.F. Leite, L.C.M. Miranda, N. Cella and H. Vargas, Journal of Applied Physics, 1989, 66, 1, 97. 11. M.G. da Silva, S.S. Goncalves, M.S. Sthel, D.U. Schramm, R.R. Sanchez, J.B. Rieumont and H. Vargas, Review Scientific Instruments, 2003, 74, 1, 831. 12. L.H. Poley, A.D.L. Siqueiro, M.G. da Silva, H. Vargas and R. Sanchez, Polimeros: Ciencia e Tecnologia, 2004, 14, 1, 8. 13. G. Rousset, F. Lepoutre and L. Bertrand, Journal of Applied Physics, 1983, 54, 5, 2383. 14. N.F. Leite, N. Cella, H. Vargas and L.C.M. Miranda, Journal of Applied Physics, 1987, 61, 8, 3025. 15. Polymer Handbook, 3rd Edition, Eds., J. Brandup and E.H. Immergut, John Wiley and Sons, New York, NY, USA, 1989. 16. J.M. Lagaron, D. Cava, E. Gimenez, P. Hernandez-Munoz, R. Catala and R. Gavara, Macromolecular Symposia, 2004, 205, 225. 17. G.T. Fieldson and T.A. Barbari, Polymer, 1993, 34, 6, 1146. 18. C. Sammon, N. Everall and J. Yarwood, Macromolecular Symposia, 1997, 119, 189. 19. C.M. Balik and W.H. Simendinger III, Polymer, 1998, 39, 20, 4723. 20. S. Cotugno, G. Larobina, G. Mensitiere, P. Musto and G. Ragosta, Polymer, 2001, 42, 15, 6431. 21. R. Auras, B. Harte and S. Selke, Journal of Applied Polymer Science, 2004, 92, 3, 1790. 22. D.A.S. Ravens and I.M. Ward, Transactions of the Faraday Society, 1961, 57, 150. 23. J. Markarian, Plastics, Additives and Compounding, 2004, 6, 1, 22. 24. I. Mazzocca, Materie Plastiche ed Elastomeri, 1982, 12, 716.

272

16

Legislative Aspects of the Use of Additives in Packaging Plastics

This chapter is reviews the legislation pertaining to the use of additives in foodgrade plastics in various countries. The legislation is reviewed and also the method of its implementation and the procedure that must be followed by anyone who wishes to seek approval for the use of any polymer or additive in applications involving contact with foodstuffs and in the packaging of other materials involving contact with humans, pharmaceuticals and cosmetics. The situation in member states of the European Union (EU) and the USA is reviewed in some detail. The situation in the following countries is reviewed more briefly: Austria and East Europe, Canada, Australia, New Zealand, South Africa, India, Japan and Latin America. The harmonisation of legislation in the EU is proceeding steadily. The Council of Europe has tended to concentrate its efforts on intrinsic toxicity of additives The EU has been studying the legal aspects of the harmonisation of legislation since 1963 and has also been studying migration of additives.

16.1 Regulatory System in the European Union The Commission of European Communities (EEC) has compiled a number of Directives relating to plastics used in food contact applications. The Framework Directive 76/893/EEC was issued in 1976 and a new Framework Directive 89/109/ EEC which governs all food contact materials was issued in 1989. In 1990 a Plastics Directive 90/128/EEC [1] was issued to be followed by several amendments, the fourth amendment being issued in March 1996 and the fifth in December 1996. In 2002, Directive 2002/72/EEC [2] was issued. This involves amendments to the Framework directive 89/109/EEC to include traceability and includes active packaging systems, e.g., packaging which deliberately interacts with the packaged food, for example as permeable gas barriers used in packaged meat applications. 273

Additive Migration from Plastics into Foods Directive 2002/72/EEC [2] and its amendments include items which govern the use of plastics food contact materials, and the creation of a positive list of raw materials such as the polymer and the additive system used. This Directive also requires petitioners to prove the safety of these substances. From 2004, 2002/72/EEC [2] is the only positive list for use in all EU States. In 2005 to 2006 a Superdirective is expected which will pull together several pieces of earlier legislation including in a single Directive, plastics, additive migration testing and vinyl chloride and also the material in Directive 2002/72/EEC [2]. Also, presumably matter in the following Directives will be included in the Superdirective: The Dangerous Substances Directive, 67/548/EEC, The Dangerous Preparations Directive, 88/379/EEC, the existing Substances Regulation, 793/93/EEC and the Marketing Directive 76/769/EEC. Published information on these Directives is discussed next in date order. In 1996, Coupard and co-workers [3] reviewed the European (90/128/CEE) and French (82/711/CEE, and modifications 93181 CEE and 85/572/CEE) regulations which are applicable to the use of polymers in contact with foods. These two sets of regulations discuss the mechanisms of the migration of additives and other chemicals (e.g., monomers and oligomers) from rubbers into foodstuffs and the migration of food constituents into rubbers. Chromatographic techniques for the detection of trace elements in rubbers susceptible to migration are described and applied to the determination of nitrosamines and polycyclic aromatic hydrocarbons. Prior to the introduction of the EU Directives, France had its own regulations (Degree 73-138, issued in February 1973) covering all materials in contact with food. Leadbitter [4] discusses the current and future regulations controlling the safe use of polyvinylchloride (PVC) in food contact applications and the implication of this for PVC manufacturers. The European Union in 1985 published a list of official fatty food simulents recommended for use in extractability tests. This has been discussed by Riquet and Feigenbaum [5] in their study of the tailoring of fatty food simulents for use in extraction tests. The fatty simulents quoted in the European list include: triglycerides, olive oil, sunflower oil or a synthetic mixture of triglycerides known as HB307. When the use of triglycerides as extraction liquids is not practicable, due to problems in chemical analysis, the European Commission has since 1996 accepted the use of alternate volatile fatty simulents such as iso-octane or aqueous ethanol, reporting the highest extraction results obtained with these liquids. Riquet and Feigenbaum [5] found that drawbacks still exist with these media. They suggest the use of a mixture of t-butyl acetate and iso-octane as a non-aggressive simulent that interacts with the polymer in the same fashion as the European Union’s official fatty food simulents. 274

Legislative Aspects of the Use of Additives in Packaging Plastics In 1994 the European Commission published a Synoptic Document No.7 [6] which provides a provisional list of additives suitable for the production of food contact plastics. This synoptic document [6] published in Belgium anticipates an EU Directive on additives used in food contact plastics. In 1998, van Lierop and co-workers [7] in support of what was at the time a future Directive, selected 100 of the most important polymer additives to establish a reference collection. Van Lierop and co-workers in 1998 produced a handbook [8] on their infrared (IR) spectroscopic, mass spectrometric, gas chromatographic, proton nuclear magnetic resonance data also physicochemical data. This new collection of additives used in plastics for food contact [8] can be seen as a logical supplement to the earlier collection of monomers and other starting materials described in a handbook entitled ‘Spectra for the Identification of Monomers in Food Packaging’ [9, 10]. This book provides information on monomeric substances listed in the Directive 90/128/EEC and amendments which restricts the range of monomers and other starting materials that can be used for the production of plastics materials and articles intended for food contact applications. An example of a typical page of Van Lierop’s handbook [8] is given in Figure 16.1. This page provides basic information, in this case for adipic acid, bis (2-ethylhexyl) ester. For each substance, this page is followed by an infrared spectrum, mass spectrum and proton magnetic resonance spectrum. This information is now available through the Internet and this will make both access and updating much easier. Van Lierop and co-workers [7] point out that a systematic approach to control materials has been elaborated in the Netherlands [9-13] to meet Dutch Regulations [14] that were in existence before the introduction of Regulation 90/128/EEC. Certainly in 1998 at least, these Dutch Regulations were in use in Government and industrial laboratories in Denmark, Greece, Norway, Sweden and Switzerland. In an attempt to harmonise the legislation, the EU has initiated global control by means of positive lists of substances which can be used, where substances with toxic potential are restricted. The first efforts have focussed on plastics, monomers and starting substances and more recently on plastics additives, as discussed previously in Directive 90/128/EEC. Such a restriction can be implemented in either of two ways: 1. By the quantity of a deliberately added or adventitious material which may be present in the packaging material, defined as quantity of material (QM). 275

Additive Migration from Plastics into Foods

Figure 16.1 Specimen entry from the handbook. (Reprinted with permission from Gilbert and co-workers, Food Additives and Contaminants, 1994, 11, 71 [10]. © 1994, Taylor & Francis)

2. By the quantity of material which could migrate from the polymeric packaging material or additives therein to the packaged commodity, defined as the specific migration limit (SML). Obviously, if the QM is sufficiently low to constitute no hazard to public health even if 100% of the substances in the polymer migrate into the packaged commodities, then the determination of SML may be deemed unnecessary. Plastics monomers are covered by Directive 90/128/EEC. The 5th Amendment to 90/128/EEC [1], i.e., Directive 99/91/EEC includes about 50 additives with restrictions and more will be added in future amendments. A QM should be used rather than a SML when the compound is shown to degrade. 276

Legislative Aspects of the Use of Additives in Packaging Plastics Although the additives are listed in Synoptic document No.7 [6], this document is provisional, and it is still undecided exactly how control on their use would be best exercised. The two most likely forms of control are a restriction on the migration of the specific substances to food or food simulants (SML) or controls on the composition of the plastic (QM). Several studies have been carried out to provide data on this topic, to relate the maximum level of residual substance in the finished plastic (QM limit) with the amount of substance migrating to foods or food simulants under defined conditions (SML limit). Feigenbaum [15] identified the substances in the polymer with the aid of a reference collection of additives. In subsequent parts of that work, a mathematical model was then developed to relate SM values to QM values. It will be necessary to validate the migration model by generating the necessary experimental data. These developments notwithstanding, in fact whichever form of control, SML or QM, is finally adopted for additives, the first step in any investigation must be to identify the substances that may migrate and therefore need to be quantified. Van Lierop in his handbook [8] gives information to assist in this identification. When considering how these reference substances and spectra may be used in practice, it is instructive to consider how enforcement might be conducted effectively. Although the principles intended to govern the control of materials and articles are clear from the relevant Directives, the practical problems of implementation and the development of an approach that should be adopted by enforcement authorities in real situations, have not yet been addressed. The most systematic approach to control has been elaborated in the Netherlands [11-13] to meet Dutch Regulations that were in existence before Directive 90/128/EEC [1]. The approach used has been initially to identify polymeric materials by IR spectra and then to identify the potential migrants by solvent extraction followed by gas chromatography - mass spectrometry (GC-MS) and by liquid chromatography (LC). Practical application of this approach over more than 10 years has shown that considerable experience is required in knowing what type of polymer is used in what food contact situation, as well as the likely additives and other constituents that might be present. The Dutch test method has been discussed in a CEN working group [16]. Several studies have been conducted on the stability of selected plastics additives for food contact in EU aqueous, fatty and alternate stimulants. Thus, Simoneau and Hannaert [17] showed that bis(2-ethylhxyl)phthalate plasticiser and octadecyl 3-(3,5di-tert-butyl 4-hydroxy phenyl) propionate antioxidant are quite stable under all heat exposure conditions tested (20-40 °C, 1 h to 10 days) and all food simulents used (15% ethanol, 95% ethanol, 3% acetic acid, olive oil and iso-octanol). In these tests, the amounts of additives remaining in the plastic at the end of the test were between 90% and 107% of the amounts present in the plastic before extraction. 277

Additive Migration from Plastics into Foods Mountford [18] discusses current EU chemical legislation and examines the shortcomings of some of the regulations in place with respect to dangerous chemicals. The Commission white paper is discussed and in particular, the REACH system, which involves the registration, evaluation and authorisation of chemicals used in food-contact applications. The impact of the REACH system on food contact plastic manufacturers is examined with respect to supplies of monomer and additives, converters and packagers. The following four legal instruments are discussed: 1. The Dangerous Substances Directive, 67/348/EEC 2. The Dangerous Preparations Directive, 88/379/EEC 3. The Existing Substances Regulation, 793/93/EEC 4. The Marketing and Use Directive, 76/769/EEC Key dates for the introduction of the new EEC legislation are reviewed next, i.e., concerning registration, evaluation of polymer contaminants, and authorisation of use of such chemicals: April 1998

EU Council: Instigation of the review

February 2001

White paper on a strategy for a future EU Chemicals Policy

June 2001

Council conclusions

November 2001 Parliament’s opinion 2002

Commission legislative proposals

2003-2004

Legislative process

2005-2006

New legislation in place

Gueris [19] of DuPont International discusses some of the pending issues with regard to plastics and polymers in contact with foodstuffs which were at the time not fully harmonised by Directive 2002/72/EC [2]. Particular attention is paid to the reduction factors mentioned in directive 85/572/EEC for migration of additives in Annex 111 Section B, to polymeric additives and attention is also given to polymerisation production aids, including antifoam agents, antiskinning agents, buffering agents, emulsifiers, surfactants, solvents and thickening agents - particular attention being paid to solvents. Gueris [19] has proposed that Annex 111 of Directive 2002/72/EC should be amended to include a Regulatory Proposal concerning substances which could be present in finished products, such as impurities in materials used, reaction intermediates, decomposition products and solvents used in the manufacture of plastics materials. 278

Legislative Aspects of the Use of Additives in Packaging Plastics Reade [20] reports that the European Commission is planning changes to its food contact plastics directives. The aims of the Superdirective referred to earlier are to be applauded: to pull together several pieces of legislation, including those on plastics, migration testing and vinyl chloride, into a single Directive. However, there is widespread feeling that the Superdirective expected in 2006 will make compliance more difficult and expensive. There are two existing pieces of legislation that cover the use of food contact plastics: the Framework Directive (89/109/EC) and the existing Plastics Directive (2002/72/EEC). There are two main additions to the Framework Directive that will affect converters: one is to make traceability compulsory, the second is the inclusion of active packaging systems. The main change to the existing Plastics Directive (2002/72/EEC) is the creation of a new ‘positive list’ of raw materials. Producers of raw materials, whether they are resins or additives, must ‘petition’ the EU to have these substances included on the list. Substances that do not appear on the list are effectively banned from food contact use. Tschech, Head of International and Regulatory Affairs at the RCC Institute [21] warns that European harmonisation of food contacting plastics will affect the whole supply chain. Currently the EU Directive EEC2002/72 [2] establishes ‘positive lists’ of monomers and additives, which can be used in the manufacture of food contact plastics. But the Commission is now proposing regulations relating to labelling, traceability and certification. Tscech states that: ‘This means that national positive list will disappear and then we will have EU-wide harmonised regulations for food contacting plastics. Suppliers of additives will no longer be able to supply their product for food contact material.’ Materials containing these additives can no longer be sold, and will no longer be available for packaging materials. So the whole supply chain will be affected.’ Complete petitions for inclusion of additives had to be submitted by 31 December 2004. Materials not complying with the Directive and not on an interim list of national approval cannot be sold legally on the market. This will be the first time that a harmonised list has existed at European level. To date, each country had its own list of approved substances. From 2004, there is only one list for EU member states. Flodberg and co-workers [22] report that the two monomers, p-hydroxy benzoic acid (HBA) and 2-hydroxy-6-naphthoic acid (HNA) present in poly(p-hydroxybenzoic acid-co-2-hydroxy-6-naphthoic acid are now included in the positive list of the new Commission Directive 2002/72/EC [2]. In the former Directive 91/128/EC [1], these two monomers were not included in the positive list and, therefore, could not be used for applications involving direct contact with food. 279

Additive Migration from Plastics into Foods In an amendment to Directive 2002/72/EC [2], namely 2002/17/EC [23] issued on 21st February 2002 an SML of 0.05 mg/kg for HNA was included for the first time. The new directive 2002/72/EC [2] of 6th of August 2002 now supersedes the former Directive 90/128/EC [1] and all it’s amendments and confirms the specific migration limit for HNA for the copolymer films. The highest overall migration level obtained was 2.3 mg/dm2 (13.8 mg/kg) in olive oil (10 days at 40 °C), well below the EU limit of 10 mg/dm2 (60 mg/kg). The SML of HBA was 15.2 µg/dm2 (91.2 µg/kg) in olive oil (2 h at 175 °C). In this case, the migration level was well below the EU limit of 10 mg/dm2 (60 mg/kg). For HNA, the highest value obtained was 4.3 µg/dm2 (26 µg/kg) in 10% ethanol (4 h at 100 °C), well below the SML of 50 µg/kg). The results obtained show that even at these severe conditions, the migration values comply with the new EU Directive 2002/72/EC [2], which regulates plastic materials and articles for food contact use. In addition, the copolymer complies with Food Contact Notification (FCN) No.103 of the United States Food and Drug Administration (FDA). Vectra A950 is therefore permitted for food contact use both in the European Union and the USA. Other relevant EU Directives are covered in further references [24-27].

16.2 The Regulatory System Existing in the UK In the UK, the EU Framework Directive 76/893/EEC, the new Framework Directive 89/109/EEC and the Plastics Directive 90/128/EC [1] and 2002/72/EC [2] have been incorporated into UK Statutory Instrument 1992 No.3145 – The Plastics Materials and Articles in Contact with Food Regulation, 1992 and subsequent amendments [28, 29]. To assist plastics manufacturers and users in the UK in the interpretation of Statutory Instrument 1992 No.3145, the British Plastics Federation has published a guide [30]. In line with this publication and a similar PIRA publication [31], Guise [32] has looked at migration testing and simulents also users of migration test data in the supply chain and he also discusses some permitted additives and existing regulations. This very useful discussion covers the current situation in the UK regarding these items. Other useful reviews have been published by PIRA [33-39] and HMSO [40].

16.3 Regulatory System Existing in the USA The US legislation has been adopted outright or in a modified form by other countries. In the USA food law is strongly inclined to statute law. Although, each State can 280

Legislative Aspects of the Use of Additives in Packaging Plastics legislate separately, in practice the federal law usually prevails which has considerable advantages in inter-state commerce. Many states have their own laws which are identical to the federal law. The principal US law is the Food, Drug and Cosmetic Act (1938) and its amendments. This law is designed to ‘provide food safe and wholesome to the people, honestly labelled and properly packaged’. This Amendment states: 1. The burden of proof of safety ‘to the health of man or animal’ of a food additive is placed on the person causing the addition. 2. The Secretary of the Health, Education and Welfare Department (HEW), acting on the advice of the FDA is authorised to ‘prescribe the conditions under which an additive may be safely used’. 3. The Delaney Amendment overrides the HEW Department and absolutely prohibits use at any concentration of an additive which induces cancer in man or animal. 4. Regarding definition of a food additive: these are any substances the intended use of which may be reasonably be expected to result, directly or indirectly, in its becoming a component or otherwise affecting the characteristics of any food including any substance intended for use in producing, manufacturing, packaging, processing, preparing, treating, transporting or holding food. 5. A substance may be exempted from such control if it is: a) ‘generally recognised among experts, qualified by scientific training or experience to evaluate its safety, as having been adequately shown, through scientific procedures (or in the case of a substance used in food prior to January 1st 1958 through scientific procedures or experience based on common use in food), to be safe under the conditions of its intended use’, or b) ‘used in accordance with a sanction or approval granted prior to the enactment of the Amendment, or under the similar clause in the Poultry Products Inspection or Meat Inspection Acts as amended’. In the USA, therefore, the use of plastics in contact with food is under the direct jurisdiction of the FDA who issue detailed positive lists of permitted plastics and additives. The whole cost of test and evaluation is put on the plastics and additive manufacturers. The Delaney Amendment has caused great difficulty but has not yet affected plastics. 281

Additive Migration from Plastics into Foods The positive list of plastics and additives has three categories. Generally recognised as safe (some additives approved under this category), prior sanction (i.e., prior to January 1st, 1958) and all others (this includes approval of most base plastics). To market in or to export to the USA it is essential to conform to FDA regulations and in the case of packaged meat and poultry, United States Department of Agriculture Regulations (USDA). The merchant who intends to use the plastic should seek a written guarantee from the supplier of plastics raw material that it conforms with the regulations for its intended use. The plastics manufacturer must consider his product from the points of view of basic polymer, additives/adjuvants and migration of anything from the plastic into the packaged commodity. The base polymer must conform to the FDA or USDA specifications. All additives must similarly conform and any restrictions with regard to foods and conditions of use complied with. In this connection it is usually necessary for the plastics producer to seek a guarantee from the additive manufacturer or supplier that his material (i.e., the neat additive) is of the quality specified in the FDA (or USDA) regulations. The polymer manufacturer must be satisfied that the additives present in his polymer meet any migration limits specified in the regulations. The FDA considers that materials which come into contact with food, such as processing machinery and packaging, to be indirect food additives, and the potential migration of materials into the food requires determination [41]. Ethanol/water solutions and food oils are commonly used as food simulating solvents when determining migration. The subsequent determination of the migrating species (which may include residual monomers, oligomers, additives and modifiers) may be by LC and GC, and MS. The FDA regulates substances that may come into contact with food, food packaging and processing machinery for example, as indirect food additives. These regulations are found in 21 CFR Parts 174-178. New substances or substances used in a new manner must receive approval from the FDA for use as an indirect food additive. There are several approaches that may be used under given sets of circumstances. A substance may be shown not to migrate from the food contact material into the food under the intended conditions of use. A substance may be shown to migrate, but in quantities that are generally recognised as safe, or a substance may migrate and its safety must be demonstrated. Each of these three scenarios require the generation of migration data, that is the quantitative description of the substances that transfer from the food contact material into the food. The third scenario requires the generation of 282

Legislative Aspects of the Use of Additives in Packaging Plastics toxicological data in addition to the migration data. Eberhard and McCort-Tipton [41] focus on the generation of migration data that is acceptable to the FDA. To assist those wishing to obtain FDA approval for the use of a food contact substance, the FDA’s Chemistry Review Branch of the Office of Premarket Approval at the Center for Food Safety and Applied Nutrition has published a guidance document, ‘Recommendations for Chemistry Data for Indirect Food Additive Petitions’. The first version of the document was published in 1995, however, recent legislative developments have lead to a new version in 2000. This guidance document describes experimental protocols to be employed for simulating migration. It addresses classes of migrating substances, types of food, and duration and temperature at which the contact occurs. It further describes the data quality requirements for migration data, and the relationship between migration and human dietary exposure to a migrating substance. In the design of a migration experiment, several key decisions are made at the outset. What substances are expected to migrate from the food contact material? This question is answered with knowledge of the synthesis and or manufacturing of the food contact material. What food simulating solvent is appropriate? This question is answered by understanding the intended use of the food contact material, and the raw materials used in its construction. When these decisions have been made, the appropriate migration, or extraction experiments can be conducted. These extraction experiments yield samples of food simulating solvent (matrix) that contain some quantity of the migrating substance (analyte). Assay for these analytes in food simulating matrices is often quite challenging. Extraction experiments may be broadly classified as single-sided, double-sided or exhaustive. The decision between single- or double-sided extraction is based upon the construction of the test material, and its thickness. Structures greater than 0.5 mm thick are considered by the FDA to be infinitely thick. Single sided extractions are most often used for coated materials, or structures less than 0.5 mm thick. Two pieces of the material are separated by an inert spacer, thus defining a volume. This layered construction is secured so that the volume can be filled with food simulating solvent. The mechanism by which this layered construction is secured must be able to withstand the pressures generated by heating food-simulating solvents past their boiling points. Double-sided extractions are by far the most common type. The test material is simply immersed in the food simulating solvent in a container that is capable of withstanding 283

Additive Migration from Plastics into Foods the pressures generated by heating food simulating solvents past their boiling points. Alternatively, an autoclave may be employed. When considering either single or double sided extraction experiments, the volume of food simulating solvent used and the surface area of food contact material to be extracted are important. Surface area of the food contact material to be extracted may be constrained by manufacturing considerations. The vessel in which the food simulating solvent and the material to be extracted are contained may impose further constraints on the surface are of the test material, as well as on the volume of food simulating solvent. In any event, common practice and FDA requirements dictate that the ratio of food simulating solvent volume to surface area of extracted material must be in the range 0.3-1.6 ml/cm2. The most common, though by no means the only application for exhaustive extractions is the generation of oligomer standards used in the quantitation of oligomers which migrate from polymeric food contact materials. The test material is simply refluxed in a solvent capable of freeing low molecular weight oligomers from a matrix of higher molecular weight polymeric material. This solution of oligomers is evaporated to dryness, weighed, taken up in a known volume of appropriate solvent, and used as an analytical standard for quantitation of oligomers which migrate from polymeric packaging materials. Method detection limits dictated by the FDA, particularly in the cases where migrants are carcinogens, are generally lower than any instrument detection limit. Methods for concentrating the analyte, or otherwise enhancing sensitivity are the rule rather than the exception. Concentration over steam, or evaporative concentration at reduced pressure are very commonly used for non-volatile analytes. This is true for aqueous/ethanolic food simulating solvents. When food oil simulants are used, aqueous buffers or acetonitrile are used to partition the analyte out of the oil and into a more volatile matrix for subsequent evaporative concentration. Volatile migrants pose a different sort of challenge. The usual evaporative concentration techniques will result in the loss of analyte. The analyte may be partitioned into a small volume of non-volatile solvent for subsequent analysis, but identification of a useful combination of solvent miscibility and analyte solubility is often impossible. Solid phase extraction and solid phase micro-extraction are often very useful for aqueous/ethanolic matrices. Here, a large volume of matrix is brought into contact with a small amount of solid material for which the analyte has a high affinity. The analyte is then desorbed from the solid phase with an appropriate solvent for subsequent analysis. 284

Legislative Aspects of the Use of Additives in Packaging Plastics

Figure 16.2 Headspace GC-MSD chromatogram of benzene at 4 ppb. (Reproduced from Author’s own files)

Headspace analysis is useful for assay of volatile analytes in aqueous/ethanolic or food oil simulents. The most important consideration here is that the boiling point of the matrix must be greater than the boiling of the analyte. This technique is used in conjunction with GC, and extremely low detection limits (sub-ppb) are attainable, especially when mass selective detectors (MSD) are employed. A sample is enclosed in a sealed container so that there is a headspace above the liquid phase. The sample is heated to a temperature greater than the boiling point of the analyte, but less than that of the solvent matrix. The analyte is forced out of the liquid phase and into the gas phase (headspace). The headspace of the sealed container is then analysed by GC. Figure 16.2 shows the headspace GC - MSD chromatogram of benzene at 1 ppb. Another advantage of headspace analysis is that there is limited interference from non-volatile migrants that may be present in the food simulating solvent. Co-migration of materials other than the migrant of interest merits discussion. In general, the reason chromatography must be used in an assay for a migrant is 285

Additive Migration from Plastics into Foods

Figure 16.3 HPLC Chromatogram of a migrant with UV detection. (Reprinted with permission from Eberhard and McCort-Tipton in Proceedings of ANTEC 2000, Orlando, Florida, 2000, Paper No. 438 [41]. © 2000, SPE)

to resolve the migrant of interest from other substances that migrate from a test material. Control materials are often used to correct for migration resulting from substances other than the migrant of interest. A control material is usually the same as the test material, but without the addition of the ingredient that results in the migrant of interest. Figure 16.3 shows the chromatographic resolution between the migrant of interest, and other substances that migrate from the test material. Figure 16.4 is an overlay chromatogram of a test material extract and a control material extract. The area under the control material peak is subtracted from the area under the test material peak to yield the signal that results from the migrant of interest alone. Selection of the appropriate chromatographic technique is specific to the analyte and matrix under consideration. GC is appropriate for volatile analytes and many organic soluble analytes. Separation is based upon polarity of the analyte molecule, and changes in its volatility with increasing temperature. LC is appropriate for nonvolatile analytes and many water soluble analytes. Separation is based on polarity of the analyte molecule and changes in its solubility in different solvents. 286

Legislative Aspects of the Use of Additives in Packaging Plastics

Figure 16.4 HPLC chromatograms with migration in both the control and the test material extracts. (Reprinted with permission from Eberhard and McCort-Tipton in Proceedings of ANTEC 2000, Orlando, Florida, 2000, Paper No. 438 [41]. © 2000, SPE)

287

Additive Migration from Plastics into Foods

Figure 16.5 GPC-ELSD chromatogram with oligomer migration and migration of other substances. (Reprinted with permission from Eberhard and McCort-Tipton in Proceedings of ANTEC 2000, Orlando, Florida, 2000, Paper No. 438 [41]. © 2000, SPE)

In the view of Eberhard and McCortney-Tipton [41], the quantitation of migrating oligomers is generally the most challenging analytical task faced in a migration study. A type of LC - gel permeation chromatography (GPC), is usually used in conjunction with evaporative light scattering detection (ELSD). In GPC, separation is based on size of the oligomer molecule. The FDA is generally concerned only with those oligomers having a molecular weight less than 1000 daltons. There are GPC techniques available for organic and aqueous oligomers. In ELSD, the eluent from a GPC column is forced through a heated orifice and aerosolised. The solvent evaporates and is swept away, leaving only solute molecules. Quantitation occurs as the solute molecules pass through and deflect a laser beam. GPC-ELSD chromatograms are quite different in appearance than other LC chromatograms. Figure 16.5 is the GPC-ELSD chromatogram of a migrant in a food simulating solvent concentrate. Another topic that merits discussion is the consideration of residual monomer in polymer formulations. Whole polymer samples, rather than food simulating solvent extracts of the polymeric material are sometimes assayed for residual monomer. The thought here is that if the amount of residual monomer in a formulation is low enough, and if one assumes that all of the residual monomer migrates into food, the human health impact 288

Legislative Aspects of the Use of Additives in Packaging Plastics of that residual monomer can be assessed. This is an alternative in some situations to assay of food simulating solvents for the presence of residual monomer. The FDA prescribes the manner in which data quality is to be assured in its ‘Recommendations for Chemistry Data for Indirect Food Additive Petitions’ document [42]. This quality assurance requirement is implemented in two ways: analyses of replicate samples, and analyses of fortified samples. The replicate analyses serve primarily to assess the inter-sample variability in migration, rather than as an indication of precision of analysis. The analyses of fortified samples assures the accuracy of the measurement system. These analyses are usually conducted on composited extracts of control material. In cases where no analyte was detected at the required method detection limit, the composited extracts are fortified with the analyte of interest at a level that corresponds to the required method detection limit. These fortified extracts are then assayed as samples. The purpose here is to insure that if the analyte were present in the extract, it would be able to be detected by the measurement system. Results of these limit of detection fortification experiments are expressed simply as ‘detected’ or ‘not detected’. When an analyte is detected in test material extracts, the fortification experiments are conducted in a slightly different fashion. Composited control material extracts are fortified at levels corresponding to one-half of, equal to, and twice the amount of the analyte detected. These fortified extracts are then assayed as samples. Results of these fortification experiments are expressed as percentage recovery of the amount added. In order to be useful in the assessment of human exposure to a migrant, the FDA express migration data as mass of migrant per unit surface area of test material extracted. This value is then used to calculate the potential human dietary exposure to a migrant. The raw migration data are expressed as mass of migrant per unit volume of food simulating solvent. The conversion to mass of migrant per unit surface area of test material extracted is as follows: μg migrant in

2 test material

=

μg migrant mL solvent

 mL  ×  2 solvent  × C f  in test material 

Where: Cf is the concentration factor. 289

Additive Migration from Plastics into Foods The mass of migrant per unit surface area of test material extracted is determined for each applicable food simulating solvent, then human dietary exposure is calculated as follows: EDI = < M > × CF

(16.2)

Where: EDI is the estimated daily intake, CF is the consumption factor, or fraction of the diet expected to be in contact with the packaging material, and < M > is the concentration of the analyte in the food contacting material. < M > is calculated as follows: < M > = ∑ ffood typeM food simulant Where: M is the mass of migrant per unit surface area of test material extracted into each applicable food simulating solvent, and f is a constant for each food type.

FDA compliance for Ultraviolet Electron Beam (EB) Coatings and Adhesives The Food and Drug Administration [43] has authority over food packaging or processing equipment materials only to the extent that they are encompassed by the definition of a ‘food additive’ under the Federal Food, Drug, and Cosmetic Act (FDCA) [44]. Section 201(s) of the act defines a food additive as ‘any substance the intended of which results or may reasonably be expected to result, directly or indirectly, in its becoming a component or otherwise affecting the characteristics of any food (including any substance intended for use in ... packing, ... packaging, or holding food.)’. The definition also specifically excludes substance that, when used as intended, are: 1) not reasonably expected to become a component of food. 2) generally recognised as safe (GRAS). 3) Prior sanctioned substances (prior to 1958, issued by either the FDA or the USDA. 290

Legislative Aspects of the Use of Additives in Packaging Plastics Any food additive that is not within these exclusions is required to be the subject of a FCN according to section 409 of the FDCA. However, there are other evaluation options by which food contact substances can be determined to be exempt from or compliant with FDA regulations (Statutory Exemption). Food contact substance is defined as ‘any substance intended for use as a component of materials used in manufacturing, packaging, packing, transportation, or holding food if such use is not intended to have a technical effect in food’. Provided next is a list of mechanisms by which a statutory exemption can be established (i.e., the substance is not reasonably expected to become a component of food): 1) Threshold of Regulation Letter: Prior to the implementation of the FCN, FDA would issue opinion letters as to whether or not a particular substance represented a significant dietary exposure. Previously issued letters will still be in effect. 2) ‘No Migration’/’No Food Additive’ Determination: This determination, often referred to as the ‘no migration’ exemption, is the most significant escape clause, for it provides the most-used premise for self-determination that a food packaging material need not be subject to FDA’s review.

This option is based on the 1979 Monsanto v. Kennedy decision, and the 1969 FDA Ramsey proposal which established that the FDA must determine that a substance migrates into food in more than insignificant amounts in order for that substance to be subject to the FDA regulations. To establish ‘No-Migration’ the evaluator must first determine an appropriate level of concern (level of detection). For most substances, an analytical sensitivity of 50 ppb, as a finding of ‘nondetected’ is reasonable. However, for certain applications where packaged foods are consumed in large amounts so that dietary exposure is expected to be high, or if the substance is considered to be highly toxic, a level of 10 ppb is recommended. Carcinogens must be evaluated on a case-by-case basis. As an alternative to extraction studies, calculations may be performed in which it is assumed that 100% of the substance migrates to food. In such a ‘worst case’ scenario, standard assumptions are used regarding the food contact area. In addition, factors such as application rate, film thickness, and so on, must be incorporated. These factors are case specific. Therefore, the calculation-based risk assessment is case specific and application to similar scenarios must be carefully considered.

3) Functional Barrier Doctrine: A subset of the No-Migration exclusion is the Functional Barrier doctrine. This concept dictates that if a substance is not part of the food contact surface of a package and is separated from the food by a functional barrier that prevents migration of the substance to food, then the substance is not reasonably be expected to become a component of food and therefore not subject to the FDA regulations. Whether or not a functional barrier exists can be 291

Additive Migration from Plastics into Foods determined based on knowledge of package structure or by conducting migration calculations or migration testing. It is important to recognise that such evaluations cannot be extrapolated to all situations. For example, migration data generated for a specific package configuration does not suggest that a particular coating is suitable for other package configurations or food types. Such functional barrier determinations may be case specific and application of such data must be carefully considered. (Also consider ‘back to front’ contamination - e.g., stacked sheets - which would bypass a functional barrier.) 4) GRAS Substances: Substances properly deemed generally recognised as safe are excluded for the definition of food additive and thus, are exempt from the premarket clearance requirement that applies to food additives. 5) Prior-Sanctioned Substances: The prior-sanctioned substances exclusion is found directly in the FDCA. Prior to the passage of the Act in 1958 a number of approvals had been issued by FDA and USDA. These pre 1958 ‘prior-sanctions’ are still valid. 6) Basic Polymer Doctrine: Substances such as catalysts, chain regulators, chain transfer agents and other material used at low levels (generally used at 1% or below) and which are required to produce the resin are considered part of the basic resin and are not subject to independent regulatory consideration. Therefore, a clearance which applies to a basic resin extends to the substances that are necessarily used during the polymerisation stage to produce it. 7) Mixture Doctrine: Manufacturers may physically blend substances which have been separately cleared or excluded. No further clearance is required provided that each substance in the mixture complies with any limitations applicable to the substance in its respective regulation. 8) Houseware Exemption: The FDCA statute does not authorise FDA to regulate containers such as paper cups, baby bottles, cooking utensils, plastic tableware, plastic or paper plates and eating utensils intended for picnic use, which are used in the home, the restaurant or beverage dispensers. However, if such containers are used as a packaging for food being merchandised in a retail market, they would be regulated. The adulteration provisions of the FDCA still apply and FDA will take action against housewares that may adulterate food. Therefore, it is prudent to apply other evaluation options, such as no-migration, to establish that adulteration will not occur. The Consumer Product Safety Commission (CPSC) regulations may also apply to specific situations.) Among these statutory exemption mechanisms, the ‘No Migration’/’No Food Additive’ exemption and the Functional Barrier Doctrine exemptions are potentially applicable to UV/EB curable chemistry. This application assumes that the packaging material 292

Legislative Aspects of the Use of Additives in Packaging Plastics is manufactured in accordance with good manufacturing practice and that the user can support, either through ‘worse case calculations’ (that is assumption of 100% migration), or through analytical determinations that ‘no-migration’ will occur. This application focuses on the use of LC-MS methods to support the ‘No Migration’/’No Food Additive’ statutory exemptions. David and co-workers [45] have given a comprehensive coverage of the Code of Federal Regulations and the specific concerns of the FDA. Other relevant US regulations are covered in [46-49].

References 1.

European Commission Directive, 90/128/EEC, 23/2/90, Relating to plastics materials and articles intended to come into contact with foodstuffs, Official Journal of the European Communities, 1990, L349, 26-47.

2.

European Commission Directive, 2002/72/EC, 06/08/2002, Relating to plastic materials and articles intended to come into contact with foodstuffs, Official Journal of the European Communities, 1990, L220/18, 18-58.

3.

A. Coupard, M. Le Huy and H. Khalfoune, Revue Generale des Caoutchoucs et Plastiques, 1996, 73, 750, 97.

4.

J. Leadbitter in Proceeedings of PVC 96 Conference, Brighton, UK, 1996, p.315.

5.

A.M. Riquet and A. Feigenbaum, Food Additives and Contaminants, 1997, 14, 1, 53.

6.

Draft of provisional list of monomers and additives used in the manufacture of plastics and coatings intended to come into contact with foodstuffs, Synoptic Document No.7, CS/PM 2356, European Commission, Brussels, Belgium, 1994.

7.

B. van Lierop, L. Castle, A. Feigenbaum, K. Ehlert and A. Boenke, Food Additives and Contaminants, 1998, 15, 7, 855.

8.

J.B.H. van Lierop, L. Castle, A. Feigenbaum and A. Boenke, Spectra for the Identification of Additives in Food Packaging, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1998.

9.

J. Bush, J. Gilbert and X. Goenaga, Spectra for the Identification of Monomers in Food Packaging, Kluwer Academic Publishers, Dordrecht, The Netherlands, 1993. 293

Additive Migration from Plastics into Foods 10. J. Gilbert, J. Bush, A. Lopez de Sa, J.B.H. Van Lierop and X. Goenaga, Food Additives and Contaminants, 1994, 11, 71. 11. D. von Battum and J.B.H. Van Lierop, Food Additives and Contaminants, 1988, 5, 381. 12. J.B.H. Van Lierop, Food Additives and Contaminants, 1994, 11, 131. 13. J.B.H. Van Lierop, Food Additives and Contaminants, 1997, 14, 555. 14. Warenwet, Dutch Packaging and Food Utensils Regulation, 1980. 15. A.F. Feigenbaum, Safety and Quality Control of Plastics Materials for Food Content, EU DG XII Research Programme, AIR 941025 1998. 16. D. Van Battum and J.B.H. van Lierop, Materials and Articles in Contact with Foodstuffs, Guide for Examination of Plastic Food Contact Materials, CEN TC 194/SCI/WG2 document N118, 1997. 17. C. Simoneau and P. Hannaert, Food Additives and Contaminants, 1999, 16, 5, 197. 18. J.P. Mountford in Proceedings of a PIRA Conference – Plastics and Polymers in Contact with Foodstuffs, Coventry, UK, 2001, Paper No.15. 19. C. Gueris in Proceedings of a PIRA Conference – Plastics and Polymers in Contact with Foodstuffs, Edinburgh, UK, 2002, Paper No.1. 20. L. Reade, European Plastics News, 2003, 30, 6, 33. 21. A. Tschech, European Plastics News, 2003, 30, 6. 22. G. Flodberg, L. Hojvall, M.S. Hedenqvist and U.W. Gedde, Food Additives and Contaminants, 2003, 20, 3, 313. 23. Commision Directive 2002/17/EC of 21st February 2002 amending Directive 90/128/EEC relating to plastics materials and articles intended to come into contact with foodstuffs, Official Journal of the European Communities, 2002, L58, 45, 19. 24. Commission Directive 97/48/EC of 29th July 1997 amending for the second time Council Directive 82/711/EEC 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, 1997, L222, 10. 294

Legislative Aspects of the Use of Additives in Packaging Plastics 25. Commisson Directive 2001/62/EC of 9th August 2001 amending Directive 90/128/EEC relating to plastics materials and articles intended to come into contact with foodstuffs, Official Journal of the European Communities, 2002, L221, 18. 26. Food Contact Materials, A Practical Guide for the Users of European Directives, European Communities, Directorate General III, Industry, Industrial Affairs III: Consumer Goods and Industries, Foodstuffs-Legislation and Scientific and Technical Aspects, 1998. 27. Food Contact Materials, A Practical Guide for the Users of European Directives, European Commission, Health and Consumer protection Directorate General, Directorate D-Food Safety: production and distribution chain, D3-Chemical and Physical Risks; Surveillance, 2002. 28. UK Statutory Instrument No. 1376, The Plastic Materials and Articles in Contact with Food Regulations, HMSO Publications, London, UK, 1998. 29. UK Statutory Instrument No.360, The Plastic Materials and Articles in Contact with Food Regulations, HMSO Publications, London, UK, 1995. 30. Plastics in Contact with Food – A Guide, 2nd Edition, British Plastics Federation, London, UK, 1996. 31. R. Ashby, Food Packaging, Migration and Legislation, 2nd Edition, PIRA International, Leatherhead, Surry, UK, 1997. 32. B. Guise, Packaging, 1997, 4, 2. 33. M. Knowles in Proceedings of a PIRA International - Packaging Materials in Contact with Foodstuffs Conference, London, UK, 1995. 34. N. De Kruijf in Proceedings of a PIRA International - Packaging Materials in Contact with Foodstuffs Conference, London, UK, 1995. 35. I. Renvoize in Proceedings of a PIRA International - Packaging Materials in Contact with Foodstuffs Conference, London, UK, 1995. 36. L. Rossi in Proceedings of a PIRA International - Packaging Materials in Contact with Foodstuffs Conference, London, UK, 1995. 37. D. Thomas in Proceedings of a PIRA International - Packaging Materials in Contact with Foodstuffs Conference, London, UK, 1995. 295

Additive Migration from Plastics into Foods 38. R. Ashby in Proceedings of a PIRA International - Packaging Materials in Contact with Foodstuffs Conference, London, UK, 1993. 39. M. Biggs in Proceedings of a PIRA International - Packaging Materials in Contact with Foodstuffs Conference, London, UK, 1994. 40. UK Statutory Instrument No.1523, Materials and Articles in Contact with Food Regulations, HMSO Publications, London, UK, 1987. 41. J.S. Eberhard and M.M. McCort-Tipton in Proceedings of the SPE ANTEC 2000 Conference, Orlando, FL, USA, 2000, Paper No.438. 42. Guidance for Industy – Preparation of Premarket notifications for Food Contact Notifications and Food additive Petitions for Food Contact Substances: Chemistry Recommendations, US Food and Drug Administration, Center for Food Safety and Applied Nutrition, Office of Premarket Approval, April 2002. [http://.cfsan.fda.gov/~dms/opa2pmnc.html] 43. J.H. Heckman and D.W. Ziffer, Fathoming Food Packaging Regulation Revisited, 2001. www.packaginglaw.com/index_fcn.cfm?id=27 44. Federal Food, Drug and Cosmetics Act, Section 201(s), [21 USC § 321 (s)], US FDA, Rockville MD, USA. 45. J.R.D. David, R.H. Graves and V.R. Carlson, Aseptic Processing and Packaging of Food: a Food Industry Perspective, CRC Press Inc., Boca Raton, FL, USA, 1996. 46. Inventory of effective premarket notifications for food contact substances, FCN No.103, US Food and Drug Administration, Center for Food Safety and Applied Nutrition, Office of Food Additive Safety, August 2002. [http://www.cfsan.fda. gov/~dms/opa-fcn.html] 47. Inventory of effective food contact substance notifications, limitations, specifications and use, FCN No.103 US Food and Drug Administration, Center for Food Safety and Applied Nutrition, Office of Food Additive Safety, August 2002. [http://www.cfsan.fda.gov/~dms/opa-fcn2.html] 48. Rules and Regulations, Federal Register, May 21st 2002, 67, 98, 3574235731, Food and Drug Administration, USA. [http://www.cfsan.fda.gov/ ~lrd/fr020521.html] 49. Guidance for Industry – Preparation of Premarket Notifications for Food Contact Substances: Toxicology Recommendations, US Food and Drug Administration, Center for Food Safety and Applied Nutrition, Office of Premarket Approval, April 2002. [http://www.cfsan.fda.gov/~dms/opa-pmnt.html] 296

17

Direct Determination of Migrants from Polymers into Foodstuffs

This approach involves a direct determination of substances in the packaged food that have migrated from the polymer. This is perhaps not a very appealing approach as it involves determination of migrants in what are often very complex foodstuff matrices presenting great analytical difficulties. Neither is this approach discussed to any great extent by the various regulatory bodies such as the European Union (EU) or the Food and Drug Administration (FDA). Therefore, the subject is not discussed extensively in this book and what follows is not a complete coverage of the subject. Nevertheless it is included here as it may appeal to some commodity packagers as a means of quality control of their product. This, of course, does not exempt the supplier of additives or the food package from attaining the necessary official approvals,

17.1 Antioxidants Hartman and Rose [1] described a rapid gas chromatographic (GC) method for the determination of butylated hydroxy anisole (t-butyl-4-methoxyphenol; BHA) and butylated hydroxy toluene (2,6-di-t-butyl-p-cresol; BHT) in vegetable oils. To 2.5 g of oil was added 150 µg of methyl undecanoate (as internal standard) in 5 ml of carbon disulfide, this was then diluted to 25 ml with carbon disulfide. A 3 to 7 µl aliquot was analysed by GC on an aluminium column (1.8 m × 4.5 mm) or a glass column (1.8 m × 4 mm) containing 10% of DC200 on gas chrom Q (80 to 100 mesh), operated at 160 °C with helium as carrier gas (50 ml/min) and flame ionisation detection, and measure the peak areas. To prevent contamination of the columns with non-volatile matter, the aluminium column is fitted with a stainless-steel sleeve containing a 6.4 mm plug of siliconised glass wool, and the first 10 cm of the glass column is packed with siliconised glass wool; the glass wool is replaced at the end of each day’s analyses. Recoveries of 97 to 104% are obtained for concentrations of 20 to 100 ppm of both BHT and BHA. Phillips [2] described a liquid-liquid extraction of BHT and BHA from vegetable oils. He studied the partition of BHT and BHA between heptane and several common 297

Additive Migration from Plastics into Foods polar solvents. The sample (4 g) was dissolved in heptane (20 ml) and extracted with dimethyl sulfoxide (4 × 25 ml). The extracts were combined, add water (100 ml) and saturated sodium chloride solution (100 ml) was added and the mixture was extracted with light petroleum (2 × 75 ml). The combined extracts were filtered and the filtrate was concentrated by evaporation. The residue was examined by thinlayer chromatography (TLC) on silica gel by development with hexane and spraying with Folin-Ciocalteu reagent, or by high-speed liquid chromatography on a 1 metre column of Corasil II, by development with heptane. The latter technique gives 97% recovery of 0.01 to 0.02% of BHT in ground nut oil, and 90 ± 10% recovery at the 0.005% level. Lemieszek-Chodorowska and Snycerski [3] have described a procedure based on TLC for the detection of phenolic antioxidants in edible fats and in oils. In this procedure the sample (10 g) was dissolved in light petroleum (25 ml) at 50% and extracted with 72% ethanol (3 × 5 ml) - shaking for 5 minutes each time. The combined ethanolic extracts were evaporated at 20 °C dissolve the residue in 96% ethanol (0.5 ml) and this solution was applied to 0.25 mm silical gel layers that had been activated at 105 °C for 1 hour. The chromatograms were developed with chloroform:anhydrous acetic acid (17:3) for separating propyl gallate, octyl gallate, dodecyl gallate and nordihydroguaiaretic acid (NDGA), or with chloroform for separating BHA and BHT. The chromatogram was dried and the spots were reveaed by spraying with 1% silver nitrate solution in 25% aqueous ammonia and heating at 70 °C to 80 °C for 20 minutes. This method permits the detection of as little as 0.5 µg of propyl gallate, 1 µg of octyl and dodecyl gallates and NDGA and BHA and 3 µg BHT. BHT, BHA, propyl gallate and tocopherol can be determined in fats extracted from mayonnaise, biscuits or margarine by TLC on layers (0.5 mm) of silica gel activated at 105 °C for 1 hour. Development is with chloroform, spots are located by spraying the air-dried chromatogram with 10% ethanolic molybdophosphoric acid, and the coloured zones are compared with those of antioxidant standards [4, 5]. TLC on polyamide-kieselguhr has been used to separate fat antioxidants [6]. The separation of propyl, isoamyl, lauryl, hexadecyl and stearyl gallates, BHA, BHT and ethyl protocatechuate is described by these workers. Using a thin layer of Nylon 6 - Kieselguhr G (2:1), they applied the antioxidants as a 0.5% solution in ethanol, and developed the chromatogram with one of the following solvents: isoamyl alcohol for 9 hours; isoamyl acetate - acetone (5:1) for 2 hours; isoamyl acetate – xylene - ethanol (20:1:1) for 2.5 hours; acetone - water (5:3) for 2.5 hours; or dioxan - water - ethanol (10:7:5) for 3 hours. The spots were revealed by spraying the chromatogram with a 0.07% ethanolic solution of Rhodamine B (CI Basic Violet 10) and then examination under ultraviolet light, and by exposure to iodine vapour. 298

Direct Determination of Migrants from Polymers into Foodstuffs Lehman and Moran [7] used a micro-column of polyamide powder in the analysis of antioxidants in fats. The antioxidants were extracted from a light petroleum solution of the fat by acetonitrile, or were adsorbed from the fat on to Celite (light petroleum as solvent) and extracted with methanol. They were then separated into two groups on a micro-column of polyamide powder, on which gallate esters, dihydrocaffeic acid and NDGA were adsorbed from a 70% methanol medium. The column was washed with water and 70% methanol, and ascorbyl acetate, t-butyl-4-methoxyphenol, 2,6-di-t-butyl-p-cresol, guaiacol and tocopherols in the percolate and washings were separated by TLC and identified. The adsorbed antioxidants were then eluted with methanolic sodium hydroxide and, after acidifying the eluate, were separated by TLC and identified. To determine antioxidants in food oils, Vigneron and Spicht [8] extracted an hexane solution of the oil with 1% ammonium acetate, 32% aqueous acetonitrile and 48% aqueous acetonitrile to isolate propyl, octyl and dodecyl gallates, respectively. As traces of t-butyl-4-methoxyphenol were also extracted by the last solvent, a hexane solution of another portion of the sample was extracted with 30% water (to remove the gallates) and washed with aqueous ammonia, then the BHA was extracted with 72% aqueous ethanol and the BHT remains in the hexane solution. Pino and co-workers [9] carried out a comparative study of chromatographic and colorimetric methods for the identification of phenolic antioxidants in edible fats. They found that the best solvent for extracting the antioxidants from the oil is acetonitrile saturated with light petroleum (bp 40 to 60 oC). The best separation was obtained on layers of polyamide powder with light petroleum - benzene - acetic acid - dimethylformamide (40:40:20:1) as solvent. For locating the spots, 0.5% ethanolic 2,5-dichloro-p-benzoquinonechlorimine was used. McBride and Evans [10] developed a rapid voltametric method for the estimation of antioxidants and tocopherols in oils and fats. The sample solutions were prepared by dissolving the oil or lard sample in an appropriate solvent, e.g., in most cases 0.12 M sulfuric acid in ethanol - benzene (2:1). The solutions were analysed with use of a linearly varying potential and a stationary, planar vitreous-carbon electrode, with a standard calomel electrode and a platinum-wire counter-electrode. Separate peaks were obtained for α-, γ- and δ-tocopherol: the peak for the β-isomer, was superimposed on that for the γ-tocopherol. BHA (>10 ppm) can be determined in vegetable oil under the same conditions, provided that δ-tocopherol is absent. Kohler and co-workers [11] described a polarographic method for the determination of 4-4′ thiobis(BHT) in food. The antioxidant was first nitrated, preferably with fuming nitric acid - concentrated sulfuric acid (1:1) at 20 ºC for 1 hour. The polarography was carried out on the resulting solution after dilution and addition of urea and sodium acetate buffer solution. The E½ for the nitrated compound was –0.54 V versus the 299

Additive Migration from Plastics into Foods mercury pool. The limit of the determination was 0.02 µg per ml with a cathode-ray polarograph or l µg per ml with a conventional DC polarograph. Takahashi [12, 13] used gas liquid chromatography to determine BHA and BHT in breakfast cereals. He obtained quantitative recoveries of 20 to 30 ppm of the antioxidants added to cereals. To determine antioxidants and preservatives in soya sauce and other foods, Takemura [14] converted these compounds to trimethylsilyl derivatives and analysed the product by GC. The sample was shaken with the silation reagent (1 ml) for 30 seconds, then set aside for 5 minutes and 2 µl of the solution was analysed by GC on, for example 20% of SE-31 silicone on Celite 545 at 200 °C with helium as carrier gas (22 ml/min) and hexane as internal standard. The method was applied to 15 preservatives and 7 antioxidants; well-separated peaks were obtained for 4-hydroxybenzoate estars, phenol, xylenol derivatives and salicylic acid. The calibration graphs for methyl, ethyl, propyl and butyl-4-hydroxybenzoates were rectilinear for the range 0.5 to 4 mg. The method was used to determine 4-hydroxybenzoates in soya sauce. Halot [15] reviewed methods of separating, detecting and determining six commercially used antioxidants in foods at concentrations down to 0.01%. Lee [16, 17] has described procedures for the detection of additives in foodstuffs utilising TLC. Ethyl protocatechuate, propyl isoamyl, lauryl cetyl and stearyl gallates, t-butyl methoxyphenol and 2,6-di-t-butyl-p-cresol are separated on layers (0.25 mm) of polyamide - silica gel (8:15) by development with a mixture of anhydrous acetic acid with chloroform, benzene or carbon tetrachloride (1:4.) The dried chromatograms were sprayed with ammoniacal to silver nitrate solution to locate the antioxidants as light-brown to black spots. The limits of detection ranged from 9.05 to 2 µg. The method was applied to butter and lard. Brown and Baxter [18] and ter Heide [19] applied different solvent extraction systems to the separation of antioxidants from fats. Stahl [20] was one of the first to apply TLC to the separation of antioxidants from fats. Seher [21, 22] carried out an extensive study of the applicability of TLC on 250-300 µm thick layers of silica gel to the identification of naturally occuring and added synthetic antioxidants in fats. Phosphomolybdic acid proved to be the most sensitive chromogenic reagent. To identify the separated compounds Seher [21, 22] sprayed the air dried plates with a 20% solution of phosphomolybdic acid in ethanol or ethylene glycol momomethyl ether (methyl ‘Cellosolve’) until they become a uniform yellow. The first antioxidants appeared as blue spots within the first one or two minutes. The plate was then treated with ammonia vapour which caused the substrate to turn pure white while the substances stood out very clearly as dark blue, in some cases violet or greenish tinged spots. 300

Direct Determination of Migrants from Polymers into Foodstuffs

17.2 Plasticisers Methods have been described for determining plasticisers in a variety of foodstuffs, milk, cheese and butter. Kiermeier and co-workers [23] and Wildbrett and co-workers [24] have developed schemes for the measurement of the extractability of monomeric phthalate ester and C10/C20 alkane phenyl sulfonates from polyvinyl chloride (PVC) containers into milk. After leaving raw milk at 38 °C for 4 hours in contact with PVC tubing containing bis(2-ethylhexyl) phthalate or phenol or cresol alkanesulfonate, the respective plasticiser was determined. Schemes are described for separation and hydrolysis of the phthalic acid ester, with spectrophotometric determination of the liberated phthalic acid at 284 nm, and for the separation and saponification of the alkane sulfonic acid ester and for spectrophotometric determination at 470 nm, of the liberated phenol or cresol after coupling with diazotised nitroaniline and addition of sodium carbonate solution. These methods are applicable to the determination of dibutyl and di-isononyl phthalates, and of the C10 to C20 alkane sulfonates of phenol or cresol. Several workers have discussed the determination of di(2-diethylhexyl)phthalate [25] in milk using HPLC [25] and in milk and cheese [26, 27] and butter [26, 28]. Page and Lacroix studied the transfer and migration of phthalate esters from aluminium foil paper to butter and margarine [29].

17.2.1 Olive Oil Badeka and Kontominas [30] studied the effect of microwave heating on the migration of dodecyl adipate and acetyl tributyl citrate plasticisers from food grade PVC into olive oil.

17.2.2 Vegetables Cohen and co-workers [31] and Coccieri [32] identified di-(2-ethylhexyl) phthalate in corn packed in a plastic bag.

17.2.3 Meat Konoyli and co-workers [33] determined dioctylphthalate and dioctyl adipate plasticisers in ground meat products and Goulas and Kontominas [34] studied the effect of γ radiation on the transfer of dioctyl adipate from PVC into chicken meat products. 301

Additive Migration from Plastics into Foods

17.2.4 Miscellaneous Foods Di(2-ethylhexyl)adipate has been determined in food stuffs [35-38] using isotope dilution GC -mass spectrometry [36] and other techniques.

17.2.5 Hydrocarbons Uhler and Miller [39] developed a GC multiple headspace extraction technique for the determination of volatile hydrocarbons in butter [39]. In a related procedure to determine styrene and its dimer in milk, acetone was used to precipitate proteins and extract fat and the residues from the packaging material. Using direct injection of these extracts for GC analysis, detection limits were 0.16 mg/kg for styrene and 0.28 mg/kg for the dimer [40].

17.3 Organotin Compounds The isolation and determination of organotin compounds in vinegar and salad oil after contact with stabilised hard PVC foil has also been studied by Wieczorek [41]. Foil samples were extracted under reflux, with pentane or ethyl ether; other samples were left in contact with oil or vinegar for periods up to 11 months. The organotin compounds in the solvent extracts or in the oil (diluted with pentane) or a chloroform extract of the vinegar were then adsorbed on to a column of Florisil (which was then washed with hexane and ether, if necessary, to remove the oil), eluted with chloroform - acetic acid - ether (5:4:3) and separated by TLC [42]. In comparison with the amount of organotin compounds that migrates from the foil into salad oil during storage for 11 months at room temperature, ether extraction for 5 hours removes ≈ 830 times as much, but pentane extraction for 1 hour removes only about 20% of the amount. From 11.5 to 27 µg of organically bound tin per dm2 of foil was found to migrate into the oil over 11 months. Adcock and Hope [43] developed a spectrometric method based on catechol violet for the determination of dioctyltin 5,5′ bis(isooctylmercaptoacetate) as tin in vinegar and olive oil. A method has been described for determining the migration of organotin stabilisers such as bis(2-ethylhexyloxycarbonyl)methylthiodioctyltin from, PVC bottles into beer. After storage of beer for 8 weeks at 20 °C in bottles stabilised with 1.13% of bis(2-ethylhexyloxycarbonyl)methylthiodioctyltin, it was treated with sulfuric acid - nitric acid and tin was determined with catechol violet by the method of Newman and Jones [44]. 302

Direct Determination of Migrants from Polymers into Foodstuffs The tin was also determined by measurement of radioactivity after storage of beer in bottles stabilised with (C-labelled bis [(2-ethylhexyloxycarbonyl) methylthio]dioctyltin. Only 1.7 µg of tin per litre was found - the FDA limit being 1 mg per litre.

17.4 Polyvinylpyrrolidone Cheng amd Malawer [45] applied pyrolysis – GC to the determination of 140 ppm of residual polyvinylpyrrolidone in beers that had been treated with a compound to remove haze.

17.5 Benzoic Acid De La Riva Reyero [46] applied HPLC to the determination of benzoic acid in milk.

17.6 Alkyl Citrates Tsuji [47] determined mono, di and triisopropyl citrates in amounts down to 1 µg/g in butter and milk by GC.

17.7 Metals Methods for the determination of metals in milk are reviewed in Table 17.1. Whilst all of these methods were not specifically developed for the determination of metals in milk which had been contacted with metal containing plastic materials these studies would, nevertheless, be applicable to metals that had migrated into milk in such tests.

303

Additive Migration from Plastics into Foods

Table 17.1 Determination of Metals in Milk Method

Detection limit

Reference

Lead

Chelation, then atomic absorption spectrometry (AAS)

0.2 µg/ml

[48]

Zinc

Kinetic fluorimetry of 2-hydroxybenzaldehyde thiosemicarbazone derivative

50-400 ppb

[49]

Tin

Polarography

0.5 mg/kg

[50]

Selenium

Specrofluorimetry of 2,3diamino derivative

-

[51]

Cadmium

AAS

0.06 mg/kg

[52]

Calcium, magnesium, potassium, sodium

AAS

mg/100 kcal

[53]

Continuous flow analysis

mg/dm2

[54]

Calcium, magnesium

-

-

[55]

Calcium, magnesium

Complexiometric titration

-

[56]

Calcium, magnesium, potassium

-

-

[57]

Miscellaneous metals

Analysis of butter

-

[58]

Antimony, bromine, chlorine, iron, lead, manganese, sodium, zinc

-

-

[59]

Cadmium, copper, iron, lead, zinc

Graphite furnace AAS

-

[60]

Cadmium, copper, lead

Flow potentiometric stripping analysis

-

[61]

Metals

Calcium

From Author’s own files

304

Direct Determination of Migrants from Polymers into Foodstuffs

References 1.

K.T. Hartman and L.C. Rose, Journal of the American Oil Chemists’ Society, 1970, 47, 7.

2.

A.M. Phillips, Journal of the American Oil Chemists’ Society, 1973, 50, 21.

3.

K. Lemieszek-Chodorowska and A. Snycerski, Roczniki Panstwowego Zakladu Higieny, 1969, 20, 261.

4.

M.V. de Vincente and T. Vincente, Anales de Bromatologia, 1971, 23, 107.

5.

O. Schwein and O.J. Conroy, Journal of the Association of Official Agricultural Chemists, 1965, 48, 489.

6.

H-C Chiang and R-G. Tseng, Journal of Pharmaceutical Sciences, 1969, 58, 12, 1552.

7.

G. Lehman and M. Moran, Zeitschrift für Lebensmittel-Untersuchung undForschung, 1971, 145, 344.

8.

P.Y. Vigneron and P. Spicht, Revue Francaise des Corps Gras, 1970, 17, 295.

9.

A.M.I. Pino, J.V. Leiro and H. Schmidt-Hebbel, Grasas y Aceites, 1969, 20, 129.

10. H.D. McBride and D.H. Evans, Analytical Chemistry, 1973, 45, 3, 446. 11. U. Köhler, H. Woggon and W-J. Uhde, Plaste und Kautsch, 1968, 15, 9, 630. 12. D.M. Takahashi, Journal of the Association of Official Analytical Chemists, 1970, 53, 39. 13. D.M. Takahashi, Analytical Abstracts, 1969, 17, 3063. 14. I. Takemura, Bunseki Kagaku (Japan Analyst), 1971, 20, 61. 15. D. Halot, Chimie Analytique (Paris), 1971, 53, 776. 16. S.C. Lee, Chemistry (Taipei), 1968, 43, 155. 17. S.C. Lee, Analytical Abstracts, 1967, 14, 7864. 18. F. Brown, Biochemical Journal, 1952, 51, 2, 237. 19. R. ter Heide, Fette Seifen Anstrichmittel, 1958, 60, 360. 305

Additive Migration from Plastics into Foods 20. E. Stahl, Chemiker-Zeitung, 1958, 82, 323. 21. A. Seher, Fette Seifen Anstrichmittel, 1959, 61, 345. 22. A. Seher, Fette Seifen Anstrichmittel, 1958, 60, 1144. 23. F. Kiermeier, K-W. Evers and G. Wildbrett, Zeitschrift für LebensmittelUntersuchung und- Forschung, 1979, 142, 205. 24. G. Wildbrett, K-W. Evers and F. Kiermeier, Zeitschrift für LebensmittelUntersuchung und- Forschung, 1968, 137, 365. 25. J.A. Giust, T. Seipelt, B.K. Anderson, D.A. Deis and J.D. Hinders, Journal of Agricultural and Food Chemistry, 1990, 38, 2, 415. 26. J.H. Petersen, Food Additives and Contaminants, 1991, 8, 701. 27. J.H. Petersen, E.T. Naamansen and P.A. Nielsen, Food Additives and Contaminants, 1993, 12, 245. 28. M. Sharman, W.A. Read, L. Castle and J. Gilbert, Food Additives and Contaminants, 1994, 11, 375. 29. B.D. Page and G.M. Lacroix, Food Additives and Contaminants, 1992, 9, 197. 30. A.B. Badeka and M.G. Kontominas, Zeitschrift für LebensmittelUntersuchung und- Forschung, 1996, 202, 313. 31. H. Cohen, C. Charrier and J. Sarfaty, Archives of Environmental Contamination and Toxicology, 1991, 20, 3, 437. 32. R.A. Cocchieri, Journal of Food Protection, 1986, 49, 4, 265. 33. E. Konoyli, P.G. Demertzis and M.G. Kontominas, Food Chemistry, 1992, 45, 3, 163. 34. A.E. Goulas and M. Kontominas, Zeitschrift für Lebensmittel-Untersuchung und- Forschung, 1996, 202, 250. 35. A. Mercer, L. Castle, J. Comyn and J. Gilbert, Food Additives and Contaminants, 1990, 7, 497. 36. J.R. Startin, I. Parker, M. Sharman and J. Gilbert, Journal of Chromatography, 1987, 387, 1, 509. 37. B.D. Page and G.M. Lacroix, Food Additives and Contaminants, 1992, 12, 129. 306

Direct Determination of Migrants from Polymers into Foodstuffs 38. C. Nerin, P. Gancedo and J. Cacho, Journal of Agricultural and Food Chemistry, 1992, 40, 10, 1833. 39. A.D. Uhler and L.J. Miller, Journal of Agricultural and Food Chemistry, 1988, 36, 4, 772. 40. S. Abrantes, Journal of High Resolution Chromatography, 1993, 16, 2, 113. 41. H. Wieczorek and D.E. Deutsch, Lebensmittel-Industrie, 1970, 62, 92. 42. H. Wieczorek, Analytical Abstracts, 1970, 19, 348. 43. L.N. Adcock and W.G. Hope, Analyst, 1970, 95, 1135, 868. 44. E.J. Newman and P.D. Jones, Analyst, 1966, 91, 1084, 406. 45. T.M.H. Cheng and E.G. Malawer, Journal of the American Society of Brewing Chemistry, 1996, 54, 2, 85. 46. C. De La Riva Reyero, Alimentaria (Madrid), 1987, 24, 57. 47. S. Tsuji, Y. Tonogai and Y. Ito, Journal of Food Protection, 1986, 49, 914. 48. D.C. Manning, American Laboratory, 1973, 5, 37. 49. A. Moreno, M. Silva, D. Perez Bendito and M. Valcarcel, Analyst, 1983, 108, 1282, 85. 50. E.M. Godar amd O.R. Alexander, Industrial Engineering Chemistry, Analytical Edition, 1946, 18, 11, 681. 51. M. Inhat, Journal of the Association of Official Analytical Chemists, 1974, 57, 368. 52. B. Bibr and J. Lener, Journal of Agricultural and Food Chemistry, 1971, 19, 5, 1011. 53. M. de la Guardia, A. Salvador, P. Bayarri and R. Farre, Analyst, 1986, 111, 12, 1375. 54. W.D. Basson and J.F. van Staden, Analyst, 1979, 104, 1238, 419. 55. F. Salinas, M.C. Mahedero and M. Jiminez-Arrabal, Quimica Analitica (Barcelona), 1992, 11, 11. 307

Additive Migration from Plastics into Foods 56. M.D. Alvarez Jiminez, M.I. Serrano Gil, M.A. Palacios Corvillo and L.M. Polo Diez, Analyst, 1988, 113, 4, 633. 57. S. Zuccheti and G. Contarina, Atomic Spectroscopy, 1993, 14, 60. 58. F. Corradini, L. Marcheselli, A. Marchetti, C. Preti and C.J. Biancardi, Journal of AOAC International, 1994, 77, 714. 59. V.A. Maihara and M.B.A. Vasconcellos, Journal of Radioanalytical and Nuclear Chemistry, 1988, 122, 1, 161. 60. N. Oikawa, K. Taguchi, N. Suzuki, K. Kojima, Shokuhin Eiseigaku Zasshi, 1987, 28, 180. 61. R. Alamestrand, D. Jagner and L. Renman, Talanta, 1986, 33, 12, 991.

308

Abbreviations and Acronyms

1

H-NMR

Proton nuclear magnetic resonance

AAS

Atomic absorption spectoscopy

ABS

Acrylonitrile – butadiene - styrene

AES

Atomic emission spectrometry

APCI

Atmospheric pressure chemical ionisation

ATR-FTIR

Attenuated total reflection Fourier-transform infrared

BA

Hydroxybutyl adipate

BAB

Dihydroxybutyl adipate

BBP

Benzyl butyl phthalate

BHA

Butylated hydroxy anisole (t-butyl-4-methoxyphenol)

BHT

Butylated hydroxy toluene (2,6-di-t-butyl-p-cresol)

bp

Boiling point

BPF

British Plastics Federation

BS

Hydroxybutyl succinate

BSB

Dihydroxybutyl succinate

BT

Benzothiazole

CBS

N-Cyclohexyl-2-benzothiazole sulfenamide

CEFIC

European Chemical Industry Council

CEN

European Committee for Standardisation

CPSC

Consumer Product Safety Commission

DC

Direct current

DEA

Diethanolamine

DEHA

Di-2-ethylhexyl adipate

DEHP

Di-2-ethylhexyl phthalate

DLTDP

Dilauryl thiodipropionate 309

Additive Migration from Plastics into Foods DSC

Dynamic scanning calorimetry

EB

Electron beam

EC

European Community

EEC

The Commission of European Communities

EI

Electrospray ionisation

ELSD

Evaporative light scattering detection

EPDM

Ethylene-propylene-diene terpolymer

ESI

Energy Sciences, Inc.

EU

European Union

EVOH

Ethylene - vinyl alcohol copolymers

FCA

The Food Contact Additives Panel sector group of the CEFIC

FCN

Food Contact Note

FDA

Food and Drug Administration

FDCA

Federal Food, Drug, and Cosmetic Act

FID

Flame ionisation detection

FPD

Flame photometric detection

FTIR

Fourier-Transform infrared spectroscopy

GC

Gas chromatography/chromatographic

GC-MS

Gas chromatography–mass spectrometry

GM

Total migrate

GPC

Gel permeation chromatography

GRAS

Generally recognised as safe

HBA

p-Hydroxybenzoic acid

HDPE

High-density polyethylene

HEW

Health, Education and Welfare Department

HIPS

High impact polystyrene

HMDS

Hydroxyl dimethyl siloxane

HMSO

Her Majesty’s Stationary Office

HNA

2-Hydroxy-6-naphthoic acid

HPLC

High-performance liquid chromatography

HP-SEC

High-performance size exclusion chromatography

310

Abbreviations and Acronyms id

Internal diameter

IR

Infrared

L/D

Length-to-diameter ratios

LC

Liquid chromatography

LC-MS

Liquid chromatography - mass spectrometry

LDPE

Low-density polyethylene

MBT

2-Mercaptobenzothiazole

MBTS

2-Mercaptobenzothiazyl disulfide

mp

Melting point

MRM

Multiple reaction monitoring

MS

Mass spectroscopy

MSD

Mass selective detectors

Mw

Molecular weight

NBR

Nitrile rubber

ND

Not detected

NDGA

Nordihydroguaiaretic acid

NMR

Nuclear magnetic resonance

NR

Natural rubber

OD

Outer diameter

OPP

Oriented polypropylene

PAH

Polycyclic aromatic hydrocarbon

PBA

Polybutylene adipate

PBS

Polybutylene succinate

PDMS

Polydimethylsiloxanes

PE

Polyethylene

PE-co-MA

Polyethylene-co-methacrylic acid

PEG

Polyethylene glycol

PEN

Polyethylene naphthalate

PET

Polyethylene terephthalate

PLA

Polylactide

ppb

Parts per billion

311

Additive Migration from Plastics into Foods ppm

Parts per million

PPO

Polypropylene oxide

PS

Polystyrene(s)

PS-b-PDMS

Polystyrene-b-polydimethylsiloxane

PVC

Polyvinylchloride

QM

Quantity of material

Rf

Retardation factor

RMSE

Root of mean-square error

rpm

Revolutions per minute

SAN

Styrene – acrylonitrile

SBS

Hydroxybutyl disuccinate

SCF

EU Scientific Committee for Food

SD

Standard deviation(s)

SFE

Supercritical fluid extraction

SIR

Selective ion recording

SML

Specific migration limit

SMT

Standards Measurements and Testing Programme

SPE

Solid-phase extraction

Tg

Glass transition temperature(s)

TIC

Total ion current

TLC

Thin-layer chromatography

TMAI

Tetramethyl ammonium iodide

TMSI

Methyl tin thioglycollic acid-2-ethyl-n-hexyl ester

TSP

Thermospray ionisation

USDA

United States Department of Agriculture

USP

United States Pharmacoepia

UV

Ultraviolet

312

Subject Index

A acrylates determinations 141–151 acrylic-based multipolymer compounds 19 acrylonitrile determinations acidic and alkaline extractants 130–131 aqueous, acidic and alkaline extractants 135–136 hydrochloric acid extractant 131, 132 liquid paraffin and n-heptane extractants 133 liquid paraffin extractant 133, 134 sodium carbonate extractant 131–133 water/ethanol extractants 129, 136 acrylonitrile–butadiene–styrene (ABS) copolymers 19 active packaging 260–261 additives used in polyethylene 14 adipates determinations 94–101 alkyl citrates 303 anthraquinones 37 antiblock additives 36 antiblock agents 32 antifungal agents 36 antioxidants 32, 36 determinations dilaurylthiodipropionate (DLTDP) 45–64 Ionox-330 65–68 Irganox-1076 64–65 miscellaneous 65–71 Santonox R 43–45, 70 direct determination of migration into foods 297–300 antislip agents 32 antistatic agents 32–33, 36 aqueous extraction liquids 7, 10 acidic and alkaline extractants: acrylonitrile 130–131 citric acid solution 44

313

Additive Migration from Plastics into Foods lauric diethanolamide 180–181 ethanol/water mixtures 44 acrylonitrile 129, 136 lauric diethanolamide 176–179 phthalates 91–94 hydrochloric acid acrylonitrile 131, 132 sodium carbonate solution 45 acrylonitrile 131–133 lauric diethanolamide 176 water 43–44 lauric diethanolamide 175–176 atmospheric pressure chemical ionisation (APCI) 165 azeotropic distillation acrylonitrile 130–137 azo-dicarbonamide 37

B bactericidal agents 36 benzoic acid 303 benzothiazole (BT) 26, 109–117, 198–199 N-cyclo-2-benzothiazole sulfonamide 26 2,5-bis(5′-tert-butyl-2-benzoxazolyl)thiophene (Uvitex OB) 75–80 Borex 20 breakdown products see extraction liquids with additives and breakdown products brighteners/whiteners 36–37 butylated hydroxyanisole (BHA) 297–300 butylated hydroxytoluene (BHT) 297–300 determination in multiple antioxidants 155–161 determination with light stabiliser 161–168 butylbenzoic acid 35

C cadmium sulfide 37 calcium stearate 34, 38 carbon black 39 carbon dioxide permeation studies 262–265 cellulose 24–25 cellulose acetates 25 citric acid 37 colorants 37 Crank’s equations 213–214, 225, 226, 233 N-cyclohexyl-2-benzothiazole sulfonamide (CBS) 109–117, 198–199

314

Subject Index

D degradation products see extraction liquids with additives and breakdown products Delaney Amendment 281 determinations antioxidants dilaurylthiodipropionate (DLTDP) 45–64, 155 Ionox-330 65–68 Irganox-1076 64–65 miscellaneous 65–71 Santonox R 43–45, 70 direct determinations of migrants into foods 297 alkyl citrates 303 antioxidants 297–300 benzoic acid 303 hydrocarbons 302 metals 303 organotin compounds 302–303 plasticisers 301–302 polyvinylpyrrolidone 303 lubricants 125–127 monomers and oligomers acrylates 141–151 acrylonitrile 129–137 styrene 138–140 organic sulfur compounds 109–117 plasticisers adipates 94–101 phthalates 81–101 polydimethyl siloxanes wine and olive oil 119–122 thermal stabilisers organotin compounds 105–108 ultraviolet stabilisers Uvitex OB 75–80 di-(2-ethylhexyl) adipate di-octyladipate 34 di-n-butylphthalate 34 2,6-di-tert-butyl-p-cresol 70 di-2-ethylhexylphthalate (DEHP) 238–240 diethanolamine (DEA) 172–175 diethylphthalate 34 differential scanning calorimetry (DSC) phthalates 89–90 diffusion, coefficient of 203 digestive fluids 195–196 diisobutylphthalate 34

315

Additive Migration from Plastics into Foods dilaurylthiodipropionate (DLTDP) 45–46, 155 determination in multiple migrants 155–161 infrared (IR) spectroscopy 51–57 oxygen flash combustion 58–64 thin-layer chromatography (TLC) 46–51 dipentylphthalate 34

E elastomers 26 electrospray ionisation 166–168 end-product additives 27 end-use additives antiblock additives 36 antifungal agents 36 antioxidants 36 antistatic agents 36 bactericidal agents 36 brighteners/whiteners 36–37 colorants 37 expanding agents 37 impact improvers 38 lubricants 38 oxygen scavengers 39–40 plasticisers 38 ultraviolet degradation inhibitors 39 ultraviolet protective agents 38 ultraviolet screens 39 environmental stress cracking polyethylene 14 polymethyl pentene 16 polypropylene 15 epoxy resins 23–24 erucamide 35 estimated daily intake (EDI) 290 ethyl cellulose 26 ethylene–vinyl acetate copolymers 16 ethylene–vinyl alcohol copolymers (EVOH) 259, 265–266 di-2-ethylhexylphthalate 34 European Committee for Standardisation (CEN) methods 241 expanding agents 37 extractability testing EU practice 7–8 principles 5–7 UK practice 8 US practice 8–10

316

Subject Index extraction cell 9 extraction liquids with additives and breakdown products 171–172 lauric ethanolamide and degradation products 172–175 analytical methods 182–194 citric acid solution extractant 180–181, 188–189 ethanol/water extractant 176–179, 187–188 liquid paraffin extractant 181–182, 186–187 sodium carbonate solution extractant 176, 185–186 water extractant 175–176, 184–185 organosulfur vulcanising agents 198–199 PBA and PBS 196–198 polymeric plasticisers 195–196 extraction liquids with multiple migrants 153–154 separations antioxidant and light stabiliser 161–168 multiple antioxidants 155–161 solvent extraction 154 ether extraction 155 extraction test 154 extraction of additives 6

F FABES migration model 219 FDA import regulations 282–283 FDA model of migration 208–209, 211–212 Fick’s first law 201 Fick’s second law 201, 219, 221 fluorocarbon polymers 20 Fourier-transform infrared (FTIR) spectroscopy diffusion studies 266–267 phthalates 88–89

G gas barrier polymers for packaging 26, 259–271 rubber and elastomers 26 gas chromatography phthalates 87–88 gas permeability 14 gastric fluid stimulant 196

H headspace analysis 285 heat stabilisers 33

317

Additive Migration from Plastics into Foods high-density polyethylene (HDPE) 13–14 2-hydroxy-4-n-octyloxy-benzophenone 38 determination with antioxidant 161–168 2-hydroxy-6-naphthoic acid (HNA) 26, 240 SML 279–280 p-hydroxybenzoic acid (HBA) 26, 240 SML 279–280 (2,2-(hydroxyl-3-tert-butyl-5-methylphenyl)-5-chlorobenzotriazole 38

I impact improvers 38 infrared (IR) spectroscopy dilaurylthiodipropionate (DLTDP) 51–57 intelligent packaging 260–261 intestinal fluid stimulant 196 Ionol CP (butylated hydroxytoluene) determination in multiple migrants 155–161 ionomers 16–17 Ionox-330 (1,3,5-trimethyl-2,4,6-tris(3,5-di-tert-butyl-4-hydroxybenzyl)benzene) 32, 65–68 Irganox-1010 32 migration rates 232 Irganox-1076 (octadecyl-3-(3,5-di-tert-butyl-4-hydroxyphenyl propionate) 32, 64–65 coefficients of diffusion and partition 234 migration rates 232–233, 235–237

L lauric diethanolamide 33, 171–172 determination with degradation products 172–175 analytical methods 182–194 citric acid solution extractant 180–181, 188–189 ethanol/water extractant 176–179, 187–188 liquid paraffin extractant 181–182, 186–187 sodium carbonate solution extractant 176, 185–186 water extractant 175–176, 184–185 legislation and regulation 273 EU 273–280 UK 280 USA 280–290 ultraviolet electron beam (EB) coatings and adhesives 290–293 Lopac 19 low-density polyethylene (LDPE) 13–14 lubricants 33–34, 38 determinations 125–127

318

Subject Index

M meat plasticiser migration 301 melamine–formaldehyde 23 melt strength improvers 34 2-mercaptobenzothiazole (MBT) 26, 109–117 2-mercaptobenzothiazyl disulfide (MBTS) 26, 109–117, 198–199 metals 303 in milk 304 2,2′-methylene-bis-(6-tert-butyl-4-methylphenol) 71 microtoming of samples 244–247 Migratest Lite program 210–211, 213–214 migration experiment design 283 migration of additives 6 migration per unit surface area 289 migration theory 201–218 polyethylene naphthalate 218–248 case 1 – limited packaging, infinite food 223–224 case 2 – limited packaging, limited food 224–226 estimation errors for partitioned migration 227–228 migration models 226–227 total migration concept 248–251 milk determination of metals 304 molecular weights of additives, effects on migration 208 monomers residual/unreacted 30 mould release agents 34

N natural rubber (NR) 26, 109 nitrile rubber (NBR) 26, 109 non-polymeric components 27–29 catalyst decomposition agents 31 end-use additives 35 antiblock additives 36 antifungal agents 36 antioxidants 36 antistatic agents 36 bactericidal agents 36 brighteners/whiteners 36–37 colorants 37 expanding agents 37 impact improvers 38

319

Additive Migration from Plastics into Foods lubricants 38 oxygen scavengers 39–40 plasticisers 38 ultraviolet degradation inhibitors 39 ultraviolet protective agents 38 ultraviolet screens 39 other introduced impurities 31 polymerisation medium 30 polymerisation residues 29 processing aids 31–32 antiblock agents 32 antioxidants 32 antislip agents 32 antistatic agents 32–33 heat stabilisers 33 lubricants 33–34 melt strength improvers 34 mould release agents 34 plasticisers 34 slip additives 35 stabilisers, general 35 residual/unreacted starting materials 29–30 Noryl 22 Nylon see polyamides 20

O octadecyl-3-(3,5-di-tert-butyl-4-hydroxyphenyl propionate (Irganox-1076) 32, 64–65 coefficients of diffusion and partition 234 migration rates 232–233, 235–237 oily extraction liquids 7, 10 liquid paraffin 45 acrylonitrile 133, 134 lauric diethanolamide 181–182 phthalates 81–90 oligomers 30 olive oil 205–208 plasticiser migration 301 polydimethyl siloxanes 119–122 optical bleaching agants 36–37 organic extraction liquids 7, 10 organic sulfur compounds determinations 109–117 organosulfur vulcanising agents 198–199 organotin thermal stabilisers 105–108, 302–303 oxygen flash combustion

320

Subject Index dilaurylthiodipropionate (DLTDP) 58–64 oxygen permeation studies 265, 268–271 oxygen scavengers 39–40, 271

P packaging polymers 13 acrylic-based multipolymer compounds 19 acrylics polyethylene terephthalate (PET; Terylene) 21–21 polymethylmethacrylate 20 Borex 20 epoxy resins 23–24 ethylene–vinyl acetate copolymers 16 fluorocarbon polymers 20 gas barrier applications 26, 259–271 rubber and elastomers 26 ionomers 16–17 Lopac 19 melamine–formaldehyde 23 natural polymers 24 cellulose 24–25 cellulose acetates 25 ethyl cellulose 26 p-hydroxybenzoic acid and 2-hydroxy-6-naphthoic acid 26 phenol formaldehyde 22 polyacetals 21 polyacrylonitrile 19 polyamides (Nylon) 20 polycarbonates 21 polyesters 23 polymethyl pentene 16 polyolefins and copolymers polyethylene 13–14 polyethylene-co-methacrylic acid 15 polyethylene naphthalate 15 polypropylene 15–16 polyphenylene oxide (PPO) 22 polystyrene and its copolymers 18–19 polysulfone 22 polyurethanes 24 silicones 24 thermosets 22 urea formaldehyde 23 vinyl plastics and vinyl copolymers 17–18 polyvinyl acetate (PVA) 18

321

Additive Migration from Plastics into Foods pancreatin 196 phenol formaldehyde 22 phenolic antioxidants 68–70 photoacoustic spectroscopy 261, 262–263 photothermal techniques 261 phthalates aqueous alcoholic extractant 91–94 oily extractants 81–90 differential scanning calorimetry (DSC) 89–90 Fourier transform infrared (FTIR) spectroscopy 88–89 gas chromatography 87–88 Piringer model of migration 204–205, 210–212 plasticisers 34, 38 determinations adipates 94–101 phthalates 81–101 direct determination of migration into foods 301 meat 301 miscellaneous foods 302 olive oil 301 vegetables 301 polarography acrylonitrile 129–137 poly(1,2-propylene adipate) 34, 196–198 polyacetals 21 polyacrylonitrile 19 polyamides (Nylon) 20 polybutylene adipate (PBA) 196–198 polybutylene succinate (PBS) 34, 196–198 polycarbonates 21 polydimethylsiloxanes 35 determinations wine and olive oil 119–122 polyesters 23 polyethylene 13–14 additives 14 polyethylene-co-methacrylic acid 15 polyethylene naphthalate (PEN) 15 migration theory 218–248 estimation errors for partitioned migration 227–228 limited packaging, infinite food 223–224 limited packaging, limited food 224–226 migration models 226–227 polyethylene terephthalate (PET; Terylene) 21–21 polylactide (PLA) 268–270 polymerisation residues 27, 29

322

Subject Index polymethyl pentene 16 polymethylmethacrylate 20 polyphenylene oxide (PPO) 22 polypropylene 15–16 non-polymeric components 28 polystyrene 18–19 determination of styrene monomer and other volatiles 138–140 polystyrene-6-polydimethyl siloxane 19 polysulfone 22 polyurethanes 24 polyvinyl acetate (PVA) 18 polyvinyl chloride (PVC) 17 polyvinylpyrrolidone 303 processing aids 27, 31–32 antiblock agents 32 antioxidants 32 antislip agents 32 antistatic agents 32–33 heat stabilisers 33 lubricants 33–34 melt strength improvers 34 mould release agents 34 plasticisers 34 slip additives 35 stabilisers, general 35 prolonging food life with antimicrobial additives to packaging 260–261

Q quantity of material (QM) 275–277

R REACH system 278 regulation see legislation and regulation root mean square error (RMSE) 235 rubbers 26, 109 organosulfur vulcanising agents 198–199

S saliva fluid stimulant 196 Santonox R (4,4′-thiobis-6-tert-butyl-m-cresol) 43–45, 70 determination in multiple migrants 155–161 silica 32 silicones 24

323

Additive Migration from Plastics into Foods slip additives 35 sodium bicarbonate 37 sodium carbonate 37 sodium citrate 37 specific migration limits (SML) 203, 275–277 compliance modelling 215 HDPE 216–218 high-impact polystyrene (HIPS) 214–216 stilbene 37 styrene determinations 138–140 migration 203 styrene–acrylonitrile (SAN) copolymers 19 sulfoselenides 37 synthetic rubbers 26

T Terylene see polyethylene terephthalate (PET) theoretical background see migration theory thermal conductivity 262 thermal stabilisers determinations organotin compounds 105–108 thermosets 22 thin-layer chromatography (TLC) dilaurylthiodipropionate (DLTDP) 46–51 multiple antioxidants 157–161 4,4′-thiobis-6-tert-butyl-m-cresol (Santonox R) 43–45, 70 thiophen 37 titanium dioxide 37 total migration 248–251 1,3,5-trimethyl-2,4,6-tris(3,5-di-tert-butyl-4-hydroxybenzyl)benzene (Ionox-330) 65–68 trimethylol propane triacetate (TMPTA) 144–151 tripropylene glycol diacrylate (TPGDA) 144–151

U ultraviolet degradation inhibitors 39 ultraviolet protective agents 38 ultraviolet screens 39 ultraviolet stabilisers determinations Uvitex OB 75–80 urea formaldehyde 23

324

Subject Index USDA import regulations 282 Uvitex OB (2,5-bis(5′-tert-butyl-2-benzoxazolyl)thiophene) 75–80

V vegetables plasticiser migration 301 vinyl chloride–vinylidene chloride copolymers 17–18 volatile migrants 284–285

W water vapour permeability 14 whiteners/brighteners 36–37 wine polydimethyl siloxanes 119–122

Z Ziegler catalyst 28, 29

325

Additive Migration from Plastics into Foods

326

ISBN: 978-1-84735-055-8

Smithers Rapra Technology is a leading international organisation with over 85 years of experience providing technology, information and consultancy on all aspects of rubber and plastics and is part of the Smithers Group of Companies. Rapra has extensive processing, analytical and testing laboratory facilities and expertise and provides a wide range of services for processors, additive suppliers, product manufacturers and end users from all industry sectors. The Rapra Information Group publishes books, technical journals, reports, conference proceedings and trade directories. They organise several key conferences each year and hold regular on- and off-site training courses. Also, the Information Group maintains and develops the world’s most comprehensive database of commercial and technical information on rubber and plastics – the Polymer Library.

Smithers Rapra Technology Limited Shawbury, Shrewsbury, Shropshire SY4 4NR, UK Telephone: +44 (0)1939 250383 Fax: +44 (0)1939 251118 http://www.rapra.net