Processing Contaminants in Edible Oils: MCPD and Glycidyl Esters [2 ed.] 0128200677, 9780128200674

Citation preview

PROCESSING CONTAMINANTS IN EDIBLE OILS

This page intentionally left blank

PROCESSING CONTAMINANTS IN EDIBLE OILS MCPD and Glycidyl Esters SECOND EDITION

Edited by

SHAUN MACMAHON U.S. Food and Drug Administration, Center for Food Safety and Applied Nutrition (CFSAN), MD, United States

JESSICA K. BEEKMAN U.S. Food and Drug Administration, Center for Food Safety and Applied Nutrition (CFSAN), MD, United States

Academic Press and AOCS Press Academic Press is an imprint of Elsevier 125 London Wall, London EC2Y 5AS, United Kingdom 525 B Street, Suite 1650, San Diego, CA 92101, United States 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom Copyright © 2022 AOCS Press. Published by Elsevier Inc. All rights reserved. Published in cooperation with American Oil Chemists Society. Published in cooperation with American Oil Chemists’ Society www.aocs.org Director, Content Development: Janet Brown No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-820067-4 For information on all Academic Press and AOCS Press publications visit our website at https://www.elsevier.com/books-and-journals Publisher: Nikki P. Levy Acquisitions Editor: Nancy Maragioglio Editorial Project Manager: Sam W. Young Production Project Manager: Bharatwaj Varatharajan Cover Designer: Mark Rogers

Typeset by TNQ Technologies

Contents Contributors

ix

1. Introduction

1

Shaun MacMahon Formation Mitigation Analysis Toxicology Regulations References

2. Formation mechanisms

1 1 2 3 4 5

7

Brian D. Craft and Frédéric Destaillats Introduction MCPD esters Glycidyl esters Conclusions and perspectives References

3. Mitigation of MCPD and glycidyl esters in edible oils

7 8 12 20 20

23

Bertrand Matthäus and Frank Pudel Introduction Influence of precursors on ester formation Mitigation in field Mitigation of the raw material Mitigation in refining Additives Removal from fully refined oils (post-refining) References

4. Indirect detection techniques

23 26 37 39 42 52 56 59

65

Karel Hrncirík Introduction Main steps in the analysis of 2- and 3-MCPD esters Main approaches toward the analysis of glycidyl esters

65 66 75

v

vi

Contents

Officially adopted methods Method comparison Method automation Analysis of oil-based foodstuffs Challenges in foodstuff analysis Conclusions References

5. Direct analytical detection methods for the food-borne toxicants glycidyl esters

83 89 96 97 98 103 104

109

Alice Ewert and Michael Granvogl Introduction Direct methods References

6. Direct detection techniques for MCPD esters

109 111 142

145

Shaun MacMahon Background Direct methods The Institute of Chemical Technology Conclusion References

7. Methods to detect MCPD and glycidyl esters in complex food matrices

145 146 162 191 191

195

Shaun MacMahon, Jessica Beekman and Michael Granvogl Background Discussion Conclusions References

8. Toxicological properties of glycidyl esters

195 196 217 217

221

Gabriele Scholz and Benoît Schilter Introduction Biological properties of glycidyl esters Biological properties of free glycidol Conclusions References

221 222 225 231 232

Contents

9. Toxicological properties of MCPD fatty acid esters

vii

235

Alfonso Lampen Abbreviations Introduction Toxicology of 3-MCPD Hazard potential of 3-MCPD fatty acid esters Toxicology of 2-MCPD esters Risk characterization Conclusions References Index

235 235 236 242 247 248 249 250 253

This page intentionally left blank

Contributors Jessica Beekman FDA Center for Food Safety and Applied Nutrition, College Park, MD, United States Brian D. Craft Nestlé Research Center, Food Science and Technology Department, Lausanne, Switzerland Frédéric Destaillats Nestlé Research Center, Food Science and Technology Department, Lausanne, Switzerland Alice Ewert German Research Center for Food Chemistry, Freising, Germany Michael Granvogl University of Hohenheim, Institute of Food Chemistry, Department of Food Chemistry and Analytical Chemistry (170a), Stuttgart, Germany; Technical University of Munich, Department of Chemistry, Chair for Food Chemistry, Freising, Germany Karel Hrncirík Upfield, Rotterdam, the Netherlands Alfonso Lampen Federal Institute for Risk Assessment (BfR), Berlin, Germany Shaun MacMahon FDA Center for Food Safety and Applied Nutrition, College Park, MD, United States Bertrand Matthäus Max Rubner-Institut, Department for Safety and Quality of Cereals, Working Group for Lipid Research, Detmold, Germany Frank Pudel Pilot Pflanzenöltechnologie Magdeburg e.V., Magdeburg, Germany Benoît Schilter Société des Produits Nestlé S.A., Nestlé Research, Lausanne, Switzerland Gabriele Scholz Société des Produits Nestlé S.A., Nestlé Research, Lausanne, Switzerland

ix

This page intentionally left blank

CHAPTER 1

Introduction Shaun MacMahon

FDA Center for Food Safety and Applied Nutrition, College Park, MD, United States

Formation To improve consumer acceptance, edible oils are industrially processed to remove or modify components that can negatively impact appearance, taste, and shelf stability. However, undesirable chemical changes can take place during the refining process. Fatty acid esters of 3-chloro-1,2-propanediol (3-MCPD), 2-chloro-1,3-propanediol (2-MCPD), and glycidol are heatinduced contaminants that are not present in virgin unrefined oils, but they can be produced by the high temperatures applied during deodorization [1e3]. There is evidence that 3-MCPD esters are formed from iron chloride and/or natural organochlorines present in native oils [4e7]. The predominant precursors and formation pathways for MCPD and glycidyl esters will be thoroughly reviewed in Chapter 2 of this text.

Mitigation The fact that MCPD esters begin forming at 200 C makes mitigation difficult, as deodorizations are generally run at temperatures greater than 200 C [4]. Many factors contribute to the formation of MCPD and glycidyl esters. The growing conditions and harvesting of the palm fruit can have profound effects on an oil’s capacity to form contaminants. The extraction, washing, and processing steps that take place prior to deodorization can influence the formation of these toxicants during deodorization, as can the specifics of the deodorization scheme. It is also possible to remove MCPD and glycidyl esters using appropriate adsorbents or enzymes. There have been a number of recent advances in mitigation of MCPD and glycidyl esters, including industrial scale production of mitigated palm oil for use in infant formula in the United States [8]. Chapter 3 of this text discusses the optimization of all of these steps to reduce and eliminate the presence of these contaminants in refined edible oils.

Copyright © 2022 AOCS Press. Processing Contaminants in Edible Oils Published by Elsevier Inc. All rights reserved. ISBN 978-0-12-820067-4 Published in cooperation with https://doi.org/10.1016/B978-0-12-820067-4.00007-3 American Oil Chemists Society.

1

2

Processing Contaminants in Edible Oils

Analysis Processed edible oils are commonly consumed worldwide and used in the production of many foods including infant formula, highlighting the need for accurate analytical methodology for their detection. Indirect approaches requiring ester hydrolysis followed by derivatization and analysis by GCe MS were the first methods developed to detect MCPD and glycidyl esters [9e11]. While these early methods played a crucial role in bringing attention from industry and regulators to the presence of these contaminants in refined oils, the early conditions for base-catalyzed hydrolyzes were shown to be potentially unreliable, raising questions about the trustworthiness of indirect methodology [12,13]. More recently, however, these issues have been addressed and the quality and reliability of these indirect methods has improved greatly. Several indirect approaches (including those involving automation) have recently completed successful collaborative studies, reaching official compendial method status with organizations such as the American Oil Chemists’ Society (AOCS) and the International Organization for Standardization (ISO). The application of these indirect methods to the analysis of MCPD and glycidyl esters in oils and high fat matrices like dressings/spreads will be covered in Chapter 4. Partly in response to the lack of dependability of early indirect methodology, direct methods were developed for glycidyl esters (GEs) and 3-MCPD esters, with contaminants analyzed intact as they occur in processed oils. Although direct techniques require a large number of standards (some of which are not commercially available) for quantitative analysis, this methodology can provide information about the individual MCPD and glycidyl esters present in a refined oil e indirect methods, on the other hand, provide quantitative information regarding the total ester content in an oil expressed as bound MCPD and glycidol. Detailed information about the structure of the individual esters can be useful in a number of applications. Some studies have indicated potential toxicological differences between different 3-MCPD esters; this information is lost when applying indirect methodology. In addition, as methods have begun to be extended to complex foods, the use of a direct approach can ensure acceptable extraction recoveries of all the different esters, which do have distinct polarity differences in some cases. Information about individual esters can also be useful in evaluating the effectiveness of mitigation approaches. The use of direct methods for the analysis of intact esters of glycidol (Chapter 5) and 3-MCPD (Chapter 6) in edible oils will be discussed.

Introduction

3

Since refined oils are the primary fat source in commercial infant formulas, the presence of MCPD and glycidyl esters is a source of potential concern due to infants’ low body weight and consumption of infant formula as the sole source of nutrition in some children. As a result, research efforts have focused on the development of analytical methods for the analysis of MCPD and glycidyl esters in infant formula, as well as in other complex foods commonly consumed by children. Chapter 7 will review the analysis of MCPD esters and glycidyl esters in infant formula as well as in other complex foods.

Toxicology Free glycidol, 3-MCPD, and 2-MCPD all pose concerns from a food safety perspective. Glycidol is a genotoxic carcinogen that is probably carcinogenic to humans [14]. According to the Federal Institute for Risk Assessment in Berlin, Germany (BfR), concentrations should be kept as low as are reasonably achievable in food [15]. Due to its genotoxicity, the Food Safety Commission of Japan has suggested glycidol intake should be based on the ALARA principle (As Low As Reasonably Achievable) [16]. Negative effects on kidneys and reproductive systems have been seen from 3-MCPD in toxicological studies [17], and it was classified by the European Scientific Committee on Food as a nongenotoxic threshold carcinogen [18]. Free 3-MCPD has been labeled as a group 2B carcinogen (possibly carcinogenic to humans) by the International Agency for Research on Cancer (IARC) [19]. There are toxicological concerns shown in limited studies related to 2-MCPD; one unpublished report showed that high doses affected striated muscles and the heart, as well as the kidneys and the liver in rats [20]. Relative to the free forms of these contaminants, research on the fatty acid esters that are formed in deodorized oils has only been initiated more recently [15,20,21]. In vivo toxicological work has demonstrated that free 3-MCPD is liberated from the diester form in rats [22] as is glycidol from glycidyl esters [23] in relatively high yields. Initial risk assessments conducted by the BfR have concluded that using a worst-case scenario, infants who are fed only commercial infant formulas could potentially ingest amounts of glycidol and 3-MCPD exceeding the Joint Food and Agriculture Organization/World Heath Organization Expert Committee on Food Additives (JECFA) recommended maximum tolerable daily intake levels [21]. Exposures to 3-MCPD esters in the United States from consumption of infant formulas, estimated using data on 3-MCPD ester

4

Processing Contaminants in Edible Oils

concentrations in formula samples collected between 2013 and 2015, are higher than those estimated for European infants. However, estimated glycidyl ester exposures were comparable for US and European infants in that time range [8]. The full results of all toxicological studies on these contaminants will be discussed in Chapter 8 (glycidyl esters) and Chapter 9 (MCPD esters).

Regulations In response to the detection of free 3-MCPD in hydrolyzed vegetable protein, soy sauce, and baked goods, many international organizations addressed the issue in those matrices. The JECFA recommended a maximum tolerable daily intake for 3-MCPD of 2 m[mu]g/kg body weight per day [24]. The European Commission established a maximum level of 20 m[mu]g/kg (ppb) for 3-MCPD in hydrolyzed vegetable protein and soy sauce [25], which was also adopted by Food Standards Australia New Zealand (FSANZ) [26]. The Codex Alimentarius adopted a maximum level of 400 m[mu]g/kg (ppb) in liquid condiments containing acid-hydrolyzed vegetable protein (excluding naturally fermented soy sauce) in 2008 [27]. The US Food and Drug Administration Compliance Policy Guide states that hydrolyzed vegetable protein that contains 3-MCPD at levels greater than 1 m[mu]g/g (ppm) is not generally recognized as safe (GRAS), and therefore is an unsafe food additive [28]. Health Canada also set a maximum contaminant concentration of 1 m[mu]g/g (ppm) in Asian-style sauces [29]. Regulations regarding MCPD or glycidyl ester concentrations in processed oils have begun to appear in recent years. The European Commission (EC) published an initial maximum level for glycidyl esters (expressed as glycidol) in powdered infant formula of 75 mg/kg that was reduced to 50 mg/kg in July 2019; and a maximum level in liquid formula of 10 mg/kg that was reduced to 6 mg/kg in July 2019 [30]. In addition, the maximum level for bound glycidol in vegetable oils and fats is 1000 mg/kg except for those used to produce baby food and processed cereal-based food for infants and young children, where the limit is 500 mg/kg [30]. The EC has proposed levels for the sum of free 3-MCPD and bound 3-MCPD for several products, although these have not been finalized. The proposed limits for the sum of bound and free 3-MCPD are 1250 mg/kg for oils/fats from individual or oil mixes from coconut, corn, canola, sunflower, soybean, and palm kernel; 2500 mg/kg for oils/fats from individual or mixes from other vegetable oils and fish oil, and 2500 mg/kg for mixtures of oils from both categories. Vegetable oils and fats destined for the production of

Introduction

5

baby food and processed cereal-based food for infants and young children have a proposed limit of 750 mg/kg. Infant formula, follow-on formula, and foods for special medical purposes intended for infants and young children have proposed limits of 125 mg/kg for powders and 15 mg/kg for liquids. The EC has also discussed bringing fish/marine oils into the existing published limits for bound glycidol, although no final decisions have been made at press time.

References [1] Hrncirik K, van Duijn G. An initial study on the formation of 3-MCPD esters during oil refining. Eur J Lipid Sci Technol 2011;113:374e9. [2] Matthäus B, Pudel F, Fehling P, Vosmann KL, Freudenstein A. Strategies for the reduction of 3-MCPD esters and related compounds in vegetable oils. Eur J Lipid Sci Technol 2011;113:380e6. [3] Pudel F, Benecke P, Fehling P, Freudenstein A, Matthäus B, Schwaf A. On the necessity of edible oil refining and possible sources of 3-MCPD and glycidyl esters. Eur J Lipid Sci Technol 2011;113:368e73. [4] Destaillats F, Craft BD, Sandoz L, Nagy K. Formation mechanisms of monochloropropanediol (MCPD) fatty acid diesters in refined palm (Elaeis guineensis) oil and related fractions. Food Addit Contam A 2012;29:29e37. [5] Destaillats F, Craft BD, Dubois ML, Nagy K. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part I: formation mechanism. Food Chem 2012;131:1391e8. [6] Nagy K, Sandoz L, Craft B, Destaillats F. Mass-defect filtering of isotope signatures to reveal the source of chlorinated palm oil contaminants. Food Addit Contam 2011;28:1492e500. [7] Smidrkal J, Tesarová M, Hrádková I, Bercíková M, Adamcíková A, Filip V. Mechanism of formation of 3-chloropropan-1,2-diol (3-MCPD) esters under conditions of the vegetable oil refining. Food Chem 2016:124e9. [8] Spungen JH, MacMahon S, Leigh J, Flannery B, Kim G, Chirtel S, Smegal D. Estimated US infant exposures to 3-MCPD and glycidyl esters from consumption of infant formula. Food Addit Contam A 2018:1085e92. [9] Divinová V, Svejkovská B, Dolezal M, Velísek J. Determination of free and bound 3chloropropane-1,2-diol by gas chromatography with mass spectrometric detection using deuterated 3-chloropropane-1,2-diol as internal standard. Czech J Food Sci 2004;22:182e9. [10] Weibhaar R. Determination of total 3-chloropropane-1,2-diol (3-MCPD) in edible oils by cleavage of MCPD esters with sodium methoxide. Eur J Lipid Sci Technol 2008;110:183e6. [11] Zelinková V, Svejkovská B, Dolezal M, Velísek J. Fatty acid esters of 3-chloropropane1,2-diol in edible oils. Food Addit Contam 2006;23:1290e8. [12] Haines TD, Adlaf KJ, Pierceall RM, Lee I, Venkitasubramanian P, Collison M. Direct determination of MCPD fatty acid esters and glycidyl fatty acid esters in vegetable oils by LC-TOFMS. J Am Chem Soc 2011;88:1e14. [13] Kaze N, Sato H, Yamamoto H, Watanabe Y. Bidirectional conversion between 3monochloro-1,2-propanediol and glycidol in course of the procedure of DGF standard methods. J Am Oil Chem Soc 2011;88:1143e51.

6

Processing Contaminants in Edible Oils

[14] IARC (International Agency for Research on Cancer). Some industrial chemicals. In: IARC monographs on the evaluation of carcinogenic risk of chemicals to humans, vol. 77. Lyon, France: International Agency for Research on Cancer; 2000. p. 469e86. [15] Bakhiya N, Abraham K, Gürtler R, Appel KE, Lampen A. Toxicological assessment of 3-chloropropane-1,2-diol and glycidol fatty acid esters in food. Mol Nutr Food Res 2011;55:509e21. [16] Food Safety Commission of Japan (FSCJ). Considerations on glycidol and its fatty acid esters in foods. Risk Assess Rep - Novel Foods & Food Addit 2015. FS/185/2015, Japan. [URL: http://www.fsc.go.jp/english/evaluationreports/others/annex_glycidol_26-52.pdf. [17] Cho WS, Han BS, Nam KT, Park K, Choi M, Kim SH, Jeong J, Jang DD. Carcinogenicity study of 3-monochloropropane-1,2-diol in sprague-dawley rats. Food Chem Toxicol 2008;46:3172e7. [18] European Commission Health and Consumer Protection Directorate. Opinion of the scientific committee on food on 3-monochloro-propane-1,2-diol (3-MCPD). 2001. [19] IARC (International Agency for Research on Cancer). IARC monographs on some chemicals present in industrial and consumer products. Food & Drink Wkly: 3Monochloro-1,2-Propanediol 2016:349e74. [20] Schilter B, Scholz G, Seefelder W. Fatty acid esters of chloropropanols and related compounds: toxicological aspects. Eur J Lipid Sci Technol 2011;113:309e13. [21] Buhrke T, Weißhaar R, Lampen A. Absorption and metabolism of the food contaminant 3-chloro-1,2-propanediol (3-MCPD) and its fatty acid esters by human intestinal caco-2 cells. Arch Toxicol 2011;85:1201e8. [22] Abraham K, Appel KE, Berger-Preiss E, Apel E, Gerling S, Mielke H, Creutzenberg O, Lampen A. Relative oral bioavailability of 3-MCPD from 3-MCPD fatty acid esters in rats. Arch Toxicol 2013;87:649e59. [23] Appel KE, Abraham K, Berger-Preiss E, Hansen T, Apel E, Schuchardt S, Vogt C, Bakhiya N, Creutzenberg O, Lampen A. Relative oral bioavailability of glycidol from glycidyl fatty acid esters in rats. Arch Toxicol 2013;87:1649e59. [24] WHO. Safety evaluation of certain food additives and contaminants, 3-chloro-1,2propanediol. WHO Food Addit Ser 2002;48. http://www.inchem.org/documents/ jecfa/jecmono/v48je18.htm. [25] European Commission Health and Consumer Protection Directorate. Commission regulation (EC) No 1881/2006 of 19 December 2006: setting maximum levels for certain contaminants in foodstuffs. 2006. [26] FSANZ. Chloropropanols in food, an analysis of the public health risk; technical report series No. 15; food standards, Australia/New Zealand. 2003. http://www.foodstandards. gov.au/publications/Documents/Technical Report Chloropropanol Report 11 Sep 03. doc. [27] Codex Alimentarius. Codex general standard for contaminants and toxins in food and feed, codex stan 193e1995; amended. 2012. www.codexalimentarius.org/download/ standards/17/CXS_193e_2012.pdf. [28] U.S. Food and Drug Administration. Guidance levels for 3-MCPD (3-chloro-1,2propanediol) in acid-hydrolyzed protein and asian-style sauces. Compliance Policy Guide Section 500.500. March 14, 2008. http://www.fda.gov/ICECI/Compliance Manuals/CompliancePolicyGuidanceManual/ucm074419.htm. [29] Health Canada. Canadian standards (maximum levels) for various chemical contaminants in foods. Modified. June 28, 2012. http://www.hc-sc.gc.ca/fn-an/securit/chemchim/contaminants-guidelines-directives-eng.php. [30] European Commission. Commission regulation (EU) 2018/290 of 26 February 2018 amending regulation (EC) No 1881/2006 as regards maximum levels of glycidyl fatty acid esters in vegetable oils and fats, infant formula, follow-on formula and foods for special medical purposes intended for infants and young children. 2018.

CHAPTER 2

Formation mechanisms Brian D. Craft and Frédéric Destaillats

Nestlé Research Center, Food Science and Technology Department, Lausanne, Switzerland

Introduction Since the publication of [1] and the heightened awareness of fatty esters of monochloropropanediol (MCPD-FE) in refined edible oils, the circumstances surrounding their formation have been subject to large amounts of speculation. For instance, some researchers speculated that precursors for MCPD-FE formation (e.g., chlorine and diacylglycerols) are present in partially refined oils [2]. Other researchers suspected that the refining process results in the uncontrolled introduction of certain compounds to the oils (e.g., inorganic chlorides in the stripping stream), so it should be the first place to explore mitigation strategies [3]. Further, amid analytical developments in MCPD-FE quantification, another family of compounds was discovered in refined edible oils, namely the fatty esters of glycidol (G-FE). G-FE were found to be partially responsible for inflation of the results of MCPD-FE quantifications due to the generation of artifacts during sample preparation before analysis using indirect methods [4]. Despite these early hurdles, some recent breakthroughs were made by Refs. [5e7] on the formation mechanisms of both MCPD-FE and G-FE during palm oil refining. Within this chapter we will take a focused look at the status of the literature to date as it pertains to the formation pathways of MCPD-FE and G-FE in refined edible oils. Critical topics will be covered, including the most prevalent precursor compounds and detailed formation mechanisms responsible for the generation of these process contaminants during oil production and refining. Because both MCPD-FE and G-FE have been found in the highest average abundance in palm oil, the majority of the research reviewed herein involves crude palm oil production and refining.

Copyright © 2022 AOCS Press. Processing Contaminants in Edible Oils Published by Elsevier Inc. All rights reserved. ISBN 978-0-12-820067-4 Published in cooperation with https://doi.org/10.1016/B978-0-12-820067-4.00004-8 American Oil Chemists Society.

7

8

Processing Contaminants in Edible Oils

MCPD esters Precursors Chlorine is ubiquitous in nature. Thus, one can speculate about a wide variety of chlorine sources, whether organic or inorganic, as potential precursor compounds to MCPD-FE formation during edible oil production. Further, a host of lipid types and compositions (e.g., acylglycerols, phospholipids, glycolipids) are available in the raw materials used to produce edible oils. Many of these lipids could theoretically interact with chlorine sources and result in the formation of MCPD-FE during oil refining. The critical precursors responsible, however, are mostly dependent on the oil type, quality, and, to a lesser degree, the circumstances of manufacture, as will be described below. Given that refined palm oil specifically has been shown to contain significant levels of MCPD-FE (2.7 mg/kg) [8], it has been exclusively used as a model matrix in the literature. The first question often raised regarding MCPD-FE precursors is the origin of chlorine involved in the MCPD-FE reaction during oil refining and why it is potentially more abundant in crude palm oil (CPO) in comparison to other crude vegetable oils [9]. Recently [5], demonstrated that many sources of covalently bound inorganic chlorine exist at ppm (mg/kg) levels in crude palm oil, including FeCl3, FeCl2, MgCl2, and CaCl2. Further, a “pool” (n ¼ 300) of organic monochlorinated compounds was also found and it appears to undergo a transformation throughout the stages of oil refining with certain compounds being formed while others decompose over time. In order to elucidate the composition of the more predominant chlorine “donor” compounds [5], used LC-MSn in the framework of model experiments. Therein the authors identified a specific family of chlorinated compounds present in both the lipids extracted from hand-picked Malaysian palm fruits and commercially procured CPO samples. Fig. 2.1, taken from Ref. [5]; shows the proposed structure and chemical formulas of this monochlorinated family of compounds. The authors suggest that given their structural similarities to phytosphingosines, it is perhaps more likely that the chlorine donors identified are endogenous plant metabolites as opposed to chlorinated contaminants introduced to the oil palm’s direct environment during growth and maturation. Extrapolating from this hypothesis, one might tend to the logic that the raw materials intended for production of each edible vegetable oil have their own reactive-chlorine pool capable of donating chlorine during oil refining and ultimately resulting in MCPD-FE generation.

Formation mechanisms

9

OH O

HN

OH

HO

Cl

OH Elemental composition

Exact m/z

C42H83O4NCl

700.60280

C42H85O4NCl

702.61807

C42H83O5NCl

716.59723

C42H85O5NCl

718.61357

C42H85O6NCl

734.60809

Figure 2.1 Proposed structure and chemical formulas of an organochlorine family of compounds found in crude palm oil. (Reprinted with permission from Nagy K, Sandoz L, Craft BD, Destaillats F. Mass-defect filtering of isotope signatures to reveal the source of chlorinated palm oil contaminants. Food Addit Contam 2011;28:1492e1500).

In terms of the most predominant lipid precursors of MCPD-FE in edible oils [1], proposed that there may be a link between the content of diacylglycerols (DAGs) in refined edible oils and their MCPD-FE levels. This assertion was likely due to the fact that the highest MCPD-FE levels were observed within the fruit pulp oils analyzed. Fruit pulp oils, such as olive and palm, are known for having high DAG contents compared to seed oils, due to the greater prevalence of lipolytic reactions during harvest [10]. This correlation, however, has been disproved in recent literature [11,12]. Although DAGs could potentially react with chlorine donors during oil refining and result in the formation of MCPD-FE, they are not the most critical lipid precursors of these process contaminants. Further, lipids such as monoacylglycerols (MAGs), phospholipids, and glycolipids are largely removed during oil degumming and are not present during the later stages of oil refining [10]. Because the bulk of MCPD-FE have been shown to be generated during oil deodorization [2,13], the entirety of the aforementioned lipid classes is not expected to be greatly involved in MCPD-FE formation reactions. This of course leaves the triacylglycerols (TAGs) up for consideration. TAGs can represent more than 90%e95% (v/v) of refined vegetable oils, whether pressed from nuts, seeds, or fruit pulps. TAGs are, therefore, the most logical critical lipid precursor available for MCPD-FE formation during oil deodorization. The results of in vitro thermal reaction experiments carried out by Ref. [7] appear to confirm this hypothesis [7]. demonstrated in controlled conditions that TAGs, not DAGs, are preferentially reacting with chlorine donors to form MCPD-FE.

10

Processing Contaminants in Edible Oils

Formation pathways As previously mentioned, MCPD-FE are formed almost completely during the deodorization unit operation of edible oil refining. As such, the majority of scientific research conducted on MCPD-FE formation pathways has been carried out either in conditions mimicking oil deodorization or within bench-top, pilot scale, or commercial deodorization units [7]. showed through in vitro experiments that 3-MCPD diesters, which are the most predominant form of MCPD-FE in refined oils [14], can be generated at temperatures as low as 180e200 C. It follows from this observation that within either type of edible oil refining (chemical or physical), the typical deodorization conditions employed strongly favor MCPD-FE formation. For example, typical chemical refining operations for palm oil involve a deodorization step at around 240 C, whereas physical refining operations can involve a deodorization step at even higher temperatures (260e270 C) in order to remove excess free fatty acids (FFAs) [10]. In order to determine the origin of chlorine involved in MCPD-FE formation during oil refining [2,11], examined the content of chloride ions present in oils pre- and post-deodorization. They then attempted to correlate these levels with the ultimate levels of MCPD-FE observed in the fully refined oils. Unfortunately, little or no correlation was observed [3]. attempted to determine if the chlorine responsible for MCPD-FE formation was originating from the stripping steam applied during oil deodorization, but with a similar negative result. Only recently [5], demonstrated that both inorganic and organic chlorinated compounds are present at ppm (mg/kg) levels in partially refined edible oils. Further, the authors demonstrated that the thermal decomposition of organic chloride-containing compounds in CPO was found to coincide strongly with the evolution of 3-MCPD diesters during the thermal treatment of CPO (see Fig. 2.2). Given that 3-MCPD diesters are the predominant class of MCPD-FE in refined oils, the study of [5] has proven causality behind the greatest portion of MCPD-FE formed during the deodorization of edible oils. A given organic (or inorganic) chlorinated compound may decompose at a certain temperature, above which the released reactive chlorine can then interact with TAG and result in the formation of MCPD-FE. The fact that the graphic in Fig. 2.2 crosses at 180 C, whereas the formation temperature of MCPD-FE in the standardized in vitro experiments of [7] was 180e200 C, is quite a coincidence. In order to uncover whether a certain reactive chlorine intermediate was most responsible for MCPD-FE

Formation mechanisms

11

120 Sum of organochlorines Sum of MCPDs

Relative abundance [%]

100 80 60 40 20 0 100

120

140

160

180

200

220

240

260

Temperature [°C]

Figure 2.2 The simultaneous decomposition of some key organochorines (n ¼ 8) monitored and the formation of 3-MCPD diesters during the thermal treatment of crude palm oil. (Reprinted with permission from Nagy K, Sandoz L, Craft BD, Destaillats F. Mass-defect filtering of isotope signatures to reveal the source of chlorinated palm oil contaminants. Food Addit Contam 2011;28:1492e1500).

formation [5], monitored the decomposition of the organochlorines found in crude palm oils via LC-MSn experiments. The authors reported that hydrogen chloride (HCl) is a typical thermal decomposition product of the organochlorine pool monitored. Thus, HCl could prove to be the predominant form of reactive chlorine responsible for MCPD-FE formation during oil deodorization. Several potential formation mechanisms of MCPD-FE have been recently reviewed in the literature [15,16]. Given that many of the past mechanistic studies of MCPD-FE formation were carried out in hydrophilic media, it has long been suggested that TAG underwent hydrolysis to DAG as a first step in the MCPD-FE formation reaction. DAGs then reacted with chlorine donor compounds resulting in the formation of acyloxonium ion intermediates and eventual nucleophilic substitution of chloride ion on the glycerol backbone [17]. More recently, however, it was proven that TAGs can act directly as a substrate for MCPD-FE formation [7]. Further [18], proved with infrared (IR) spectroscopy that heating TAG in the presence of Lewis acids can lead to cyclic acyloxonium ion formation. Thus, is it theoretically plausible to assume that the two pathways could be favored in the case of hydrophobic systems like oil deodorization.

12

Processing Contaminants in Edible Oils

One pathway involves the reaction of TAG directly with HCl formed by thermal degradation of chlorine donors, nucleophilic substitution of chloride ion on the glycerol backbone to form an MCPD-FE, and finally the release of a fatty acid. The second mechanism first involves the formation of an acyloxonium ion intermediate compound. These two MCPD-FE formation mechanisms were summarized by Ref. [7] and appear in Fig. 2.3 [7]. also proved through in vitro experiments that the MCPD-FE formation reaction is regioselective and preferential to the sn-1 (3) positions on the glycerol backbone. Because the critical precursors and predominant formation mechanisms have been elucidated in the case of palm oil production and refining, it was then possible to speculate about the potential predominant root causes responsible for the manifestation of MCPD-FE therein [5]. demonstrated that chlorinated compounds (n ¼ 300 found) can be monitored in crude palm oils utilizing mass-defect filtering of isotope signatures. Further, the authors then discovered that these same compounds can be utilized to segregate commercial palm oil samples (n ¼ 26) based on their processing stage using multivariate statistical analysis (see Fig. 2.4). The grouping of commercial samples from crude to refined-bleached to refined-bleacheddeodorized, as shown in Fig. 2.4, suggests that chlorinated compounds undergo a transformation throughout palm oil production and refining. This truth might lend to the logic that chlorine is similarly transformed throughout the agricultural process involved in palm oil growth, maturation, and harvest. Fig. 2.5 is a schematic taken from Ref. [19] and serves as a root-cause analysis of MCPD-FE formation during refined palm oil production. It summarizes potential locations/sources for the influx of chlorine from the environment, accumulation of the chlorine in the palm plant and fruits, and transformation of chlorine into more liposoluble forms during CPO production, followed by the resultant formation of MCPD-FE during oil refining. A similar type of root-cause analysis could prove beneficial in the case of other refined vegetable oil crops and the assessment of their potential for production of MCPD-FE after harvest and refining.

Glycidyl esters Precursors Given the history behind the discovery of G-FE in edible oils, researchers often pooled G-FE with MCPD-FE (often termed MCPD esters and related compounds) [3]. This fact led to the assumption that these

Formation mechanisms

13

Figure 2.3 Proposed mechanisms for the formation of MCPD-FE from TAG at high temperatures in the presence of HCl evolved from the thermal decomposition of trace organochlorines. Two putative pathways including the formation of reactive cyclo acyloxonium ion intermediate (upper panel) or a direct nucleophilic substitution reaction (lower panel) are displayed. Both pathways result in the formation of an MCPD-FE molecule and the release of a fatty acid. (Reprinted with permission from Destaillats F, Craft BD, Sandoz L, Nagy K. Formation mechanism of monochloropropanediol (MCPD) fatty acid diesters in refined palm (Elaeis guineensis) oil and related fractions. Food Addit Contam 2012;29:29e37).

14

Processing Contaminants in Edible Oils

Figure 2.4 Principal component analysis of chlorine-containing compounds (n ¼ 300) present in crude, partially, and fully refined palm oil samples (n ¼ 26). The grouping of oil samples based on refining stage (i.e., from crude to refined-bleached [RB], to refined-bleached-deodorized [RBD]) suggests that these compounds undergo a transformation during oil processing. (Reprinted with permission from Nagy K, Sandoz L, Craft BD, Destaillats F. Mass-defect filtering of isotope signatures to reveal the source of chlorinated palm oil contaminants. Food Addit Contam 2011;28:1492e1500).

compound families were very closely related and potentially subject to interconversion. As such, both MCPD-FE and G-FE were thought to share the same precursors. This assumption has been disproved recently [20]. reported a strong positive correlation between DAG levels alone and the amount of G-FE contained in refined palm oils. Because fruit pulp oils like palm oil naturally contain higher levels of DAG (3%e4%) [21], this may suggest possible causality as to why their deodorized counterparts also contain higher levels of G-FE in comparison to seed oil crops, as demonstrated by Ref. [4]. The strong correlation between DAG contents and G-FE formation has essentially been validated on an industrial scale [22]. recently reported that high-DAG oils marketed for health and wellness had to be taken off the market purportedly due to “high levels” of G-FE. Further [23], demonstrated that commercially refined oils rich in DAG (87%) can contain more than 10-fold greater G-FE levels relative to oils with lower DAG (3.9%e6.8%) contents. The only other proven lipid precursor to G-FE is MAG [6]. showed that both MAG and DAG can result in formation of GFE upon thermal treatment, although formation from DAG is most

Formation mechanisms

15

Palm Oil Production Formation of lipophilic organochlorines during fruit bunch sterilization

Palm Oil Refining

(4)

Oil Palm Growth & Maturation

Biosynthesis of hydrophilic organochlorines in palm fruits

(3)

Reaction of liposoluble organochlorines with palm oil triacylglycerols at high temperatures

(5) Accumulation of inorganic chloride in the oil palm plant

MCPD diester formation

(2) (1)

Influx of inorganic chloride from the environment (e.g., KCl, NH4Cl, MgCl2, FeCl3, FeCl2)

Soil, Fertilizer, & Irrigation Figure 2.5 Root-cause analysis of the factors involved in the formation of MCPD-FE during refined palm oil production, including (1) chlorine influx from the environment, (2) accumulation of inorganic chloride in the plant, (3) bioconversion of inorganic chlorides to organochlorines in palm fruits, (4) formation of liposoluble organochlorines during fruit bunch sterilization, and (5) reaction of liposoluble organochlorines with TAG in palm oil during oil deodorization. (Reprinted with permission from Craft BD, Nagy K, Sandoz L, Destaillats F. Factors impacting the formation of monochloropropanediol (MCPD) fatty acid diesters during palm (Elaeis guineensis) oil production. Food Addit Contam 2012;29:354e361).

favored. The fact that MAG levels in refined-bleached oils are often quite low (mean < 0.1%) [24], however, renders this reaction route less significant in the case of edible vegetable oils. Formation pathways The formation of G-FE from DAG during oil deodorization was shown to be significant at temperatures at or above 230 C [11]. Ref. [20] confirmed this temperature of G-FE formation by deodorizing refined-bleached palm oil from 180 to 240 C at increments of 20 C. G-FE values were reported as the sum of the glycidyl-palmitate (16:0 G-FE), stearate (18:0 G-FE), oleate (18:1 G-FE), linoleate (18:2 G-FE), and linolenate (18:3 G-FE) species

16

Processing Contaminants in Edible Oils

2.0 1.8

Glycidyl esters content (ppm)

1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 Nondeodorized

180°C

200°C

220°C

240°C

Figure 2.6 Formation of G-FE during the thermal treatment of refined-bleached palm oil heated at different temperatures for 2 h in a bench-top deodorization unit. These GFE totals are a sum of the glycidyl-palmitate (16:0 G-FE), stearate (18:0 G-FE), oleate (18:1 G-FE), linoleate (18:2 G-FE), and linolenate (18:3 G-FE) species. (Reprinted with permission from Craft BD, Nagy K, Seefelder W, Dubois M, Destaillats F. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part II: practical recommendations for effective mitigation. Food Chem. 2012;132:73e79).

(see Fig. 2.6). As seen in Fig. 2.6, the formation of G-FE proceeds at an exponential rate above 220 C. To further demonstrate the link between DAG and G-FE formation during edible oil refining [20], spiked refined cottonseed oil (w100% TAG) with a DAG standard in concentrations ranging from 1% to 5% and submitted it to thermal treatment at 235 C for 2 h within both an ampoule system and a bench-top deodorization unit (see Fig. 2.7). As seen in Fig. 2.7, G-FE formation appears to increase exponentially when DAG contents exceed 3%e4% of total lipids. Further, the marked reduction in the levels of G-FE found in the deodorization trials, when compared to the ampoule system, suggest that a large amount of G-FE are being stripped into the deodorizer distillate. This is logical given the typical range of molecular weights of G-FE species found in palm oil (i.e., 284.4 g/mol for glycidyl myristate to 340.5 g/mol for glycidyl stearate). This could mean that the relative concentration of G-FE manifested in refined edible oils is dependent on the efficiency of the deodorizer units in the oil factories (e.g., vacuum pressure, stripping medium).

Formation mechanisms

17

Relative abundance to internal standard

5 Heated in ampoules

Deodorized

4

3

2

1

0 0

1

2

3

4

5

Level of diacylglycerol in oil (%)

Figure 2.7 Influence of DAG concentration (standardized at 1%e5%) on the formation of G-FE during the thermal treatment of refined cottonseed oil at 235 C for 2 h. The greater presence of G-FE in the ampoule experiments (dark bars) when compared to the bench-top deodorizer trials (light bars) suggests that a significant portion of G-FE are stripped off in the deodorizer. (Reprinted with permission Craft BD, Nagy K, Sandoz L, Destaillats F. Factors impacting the formation of monochloropropanediol (MCPD) fatty acid diesters during palm (Elaeis guineensis) oil production. Food Addit Contam 2012;29:354e361).

Ref. [6] proposed that G-FE can be formed from the thermal treatment of DAG proceeding through an intramolecular rearrangement mechanism followed by the elimination of a fatty acid (see Fig. 2.8). The authors also proposed a similar mechanism for the formation of G-FE from the thermal treatment of sn-1 (3)-MAG resulting in the elimination of a water molecule. Unexpectedly, while conducting G-FE formation experiments and subsequent quantifications [6], discovered some very prominent isomers of G-FE, namely oxopropyl esters. The formation of oxopropyl esters was found to occur at temperatures as low as 150 C and plateaus around 200 C, at which point G-FE proceed to be formed at an increasing rate (exponentially at temperatures > 230 C). Fig. 2.9, taken from Ref. [6]; summarizes this phenomenon of simultaneous oxopropyl ester and G-FE production with increased thermal treatment. The authors proposed that oxopropyl esters are formed from DAG through a similar intramolecular rearrangement reaction as G-FE, but this reaction proceeds through fatty acid loss followed by the formation of an enol intermediate that then undergoes tautomerization.

18

Processing Contaminants in Edible Oils

Figure 2.8 Proposed mechanism for the formation of G-FE from DAG at high temperatures. The DAG molecule undergoes an intermolecular rearrangement followed by the loss of a fatty acid. (Reprinted with permission from Destaillats F, Craft BD, Dubois M, Nagy K. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part I: formation mechanism. Food Chem. 2012;131:1391e1398).

Figure 2.9 Relative formation of oxopropyl esters and G-FE from DAG at different temperatures. Note that the ordinate axis is log scale and both individual and sum abundance of isomers are shown. (Reprinted with permission from Destaillats F, Craft BD, Dubois M, Nagy K. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part I: formation mechanism. Food Chem. 2012;131:1391e1398).

Formation mechanisms

19

Ref. [20] compiled analytical data on the occurrence of G-FE in a variety of commercially refined-bleached-deodorized (RBD) palm oils used in their experiments to determine if relationships were present. Fig. 2.10, taken from Ref. [20] shows the relationship between G-FE and DAG levels in palm oil and palm olein samples (n ¼ 15; panel A). Analysis of a subgroup of samples obtained from the same refinery (n ¼ 6; Fig. 2.10, panel B) show a strong positive correlation (R2 w0.8) [20]. also developed a predictive model that directly correlates the level of FFA in CPO to the DAG contents of fully refined palm oil. For example, a DAG concentration of 3% in refined palm oil was found to be equivalent to 1.2%e1.3% FFA in the 20.0

A

Glycidyl esters (ppm)

18.0 16.0 14.0 12.0 10.0 8.0 6.0 4.0 2.0 0.0 3.8

4.3

4.8

5.3

5.8

6.3

6.8

7.3

7.8

DAG level (g/100g of oil) 18.5

B

Glycidyl esters (ppm)

17.5 16.5 15.5 14.5 13.5 12.5 11.5 10.5

R² = 0.79

9.5 3.8

4.3

4.8

5.3

5.8

DAG level (g/100g of oil)

Figure 2.10 Relationship between G-FE and DAG levels in palm oil and palm olein samples (n ¼ 15, panel A). Analysis of a subgroup of samples obtained from the same refinery (n ¼ 6, panel B) show a strong positive correlation (R2 w0.8). (Reprinted with permission from Craft BD, Nagy K, Seefelder W, Dubois M, Destaillats F. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part II: practical recommendations for effective mitigation. Food Chem. 2012;132:73e79).

20

Processing Contaminants in Edible Oils

initial CPO. Making such a correlation is practical given that CPO is often purchased based on the level of FFA as a general indicator of quality (i.e., the lower the FFA, the higher the quality). From this information the answer seems clear: The higher quality the CPO used to produce refined palm oil, the lower the G-FE levels will be in their refined-bleached counterparts. This relationship would likely also prove true for other raw materials used to create refined edible oils.

Conclusions and perspectives In summary, the most predominant formation mechanisms of MCPD-FE and G-FE in refined palm oil were revealed based on quantifications of known matrix components and the execution of well-defined experimental investigations. Once the formation mechanisms were proven, critical precursors and detailed formation pathways could be elucidated. These factors will prove essential in allowing for targeted and efficient mitigation procedures by oil producers.

References [1] Zelinkova Z, Svejkovska B, Velisek J, Dolezal M. Fatty acid esters of 3-chloropropane1,2-diol in edible oils. Food Addit Contam 2006;23:1290e8. [2] Franke K, Strijowski U, Fleck G, Pudel F. Influence of chemical refining process and oil type on bound 3-chloro-1,2-propanediol contents in palm oil and rapeseed oil. LWT - Food Sci Technol (Lebensmittel-Wissenschaft -Technol) 2009;42:1751e4. [3] Pudel F, Benecke P, Fehling P, Freudenstein A, Matthaus B, Schwaf A. On the necessity of edible oil refining and possible sources of 3-MCPD and glycidyl esters. Eur J Lipid Sci Technol 2011;113:368e73. [4] Weisshaar R, Perz R. Fatty acid esters of glycidol in refined fats and oils. Eur J Lipid Sci Technol 2010;112:158e65. [5] Nagy K, Sandoz L, Craft BD, Destaillats F. Mass-defect filtering of isotope signatures to reveal the source of chlorinated palm oil contaminants. Food Addit Contam 2011;28:1492e500. [6] Destaillats F, Craft BD, Dubois M, Nagy K. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part I: formation mechanism. Food Chem 2012;131:1391e8. [7] Destaillats F, Craft BD, Sandoz L, Nagy K. Formation mechanism of monochloropropanediol (MCPD) fatty acid diesters in refined palm (Elaeis guineensis) oil and related fractions. Food Addit Contam 2012;29:29e37. [8] Weisshaar R. Fatty acid esters of 3-MCPD: overview of occurrence and exposure estimates. Eur J Lipid Sci Technol 2011;113:304e8. [9] Matthaus B. Organic or not organicdthat is the question: how the knowledge about the origin of chlorinated compounds can help to reduce formation of 3-MCPD esters. Eur J Lipid Sci Technol 2012;114:1333e4.

Formation mechanisms

21

[10] Dijkstra AJ, Segers JC. The lipid handbook: with CD-ROM. Boca Raton, FL: CRC Press; 2007. p. 143e262. [11] Hrncirik K, van Duijn G. An initial study on the formation of 3-MCPD esters during oil refining. Eur J Lipid Sci Technol 2011;113:374e9. [12] Matthaus B, Pudel F, Fehling P, Vosmann K, Freudenstein A. Strategies for the reduction of 3-MCPD esters and related compounds in vegetable oils. Eur J Lipid Sci Technol 2011;113:380e6. [13] Ramli MR, Siew WL, Ibrahim NA, Hussein R, Kuntom A, Razak RAA, Nesaretnam K. Effects of degumming and bleaching on 3-MCPD esters formation during physical refining. J Am Oil Chem Soc 2011;88:1839e44. [14] Seefelder W, Varga N, Studer A, Williamson G, Scanlan FP, Stadler RH. Esters of 3chloro-1,2-propanediol (3-MCPD) in vegetable oils: significance in the formation of 3-MCPD. Food Addit Contam 2008;25:391e400. [15] Hamlet CG, Asuncion L, Velisek J, Dolezal M, Zelinkova Z, Crews C. Formation and occurrence of esters of 3-chloropropane-1,2-diol (3-CPD) in foods: what we know and what we assume. Eur J Lipid Sci Technol 2011;113:279e303. [16] Rahn AKK, Yaylayan VA. What do we know about the molecular mechanism of 3MCPD ester formation? Eur J Lipid Sci Technol 2011;113:323e9. [17] Velisek J, Dolezal M, Crews C, Dvorak T. Optical isomers of chloropropanediols: mechanisms of their formation and decomposition in protein hydrolysates. Czech J Food Sci 2002;20:161e70. [18] Rahn AKK, Yaylayan VA. Monitoring cyclic acyloxonium ion formation in palmitin systems using infrared spectroscopy and isotope labelling technique. Eur J Lipid Sci Technol 2011;113:330e4. [19] Craft BD, Nagy K, Sandoz L, Destaillats F. Factors impacting the formation of monochloropropanediol (MCPD) fatty acid diesters during palm (Elaeis guineensis) oil production. Food Addit Contam 2012;29:354e61. [20] Craft BD, Nagy K, Seefelder W, Dubois M, Destaillats F. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part II: practical recommendations for effective mitigation. Food Chem 2012;132:73e9. [21] D’Alonzo RP, Kozarek WJ, Wade RL. Glyceride composition of processed fats and oils as determined by glass capillary gas chromatography. J Am Oil Chem Soc 1982;59:292e5. [22] Watkins C. Kao suspends DAG oil shipments. Inform 2009;20:689e90. [23] Masukawa Y, Shiro H, Nakamura S, Kondo N, Jin N, Suzuki N, Ooi N, Kudo N. A new analytical method for the quantification of glycidol fatty acid esters in edible oils. J Oleo Sci 2010;59:81e8. [24] Goh EM, Timms RE. Determination of mono- and diglycerides in palm oil, olein and stearin. J Am Oil Chem Soc 1985;62:730e4.

This page intentionally left blank

CHAPTER 3

Mitigation of MCPD and glycidyl esters in edible oils Bertrand Matthäusa and Frank Pudelb a

Max Rubner-Institut, Department for Safety and Quality of Cereals, Working Group for Lipid Research, Detmold, Germany; bPilot Pflanzenöltechnologie Magdeburg e. V., Magdeburg, Germany

Introduction In different stages of food processing, so-called heat-induced contaminants can be formed depending on the processing conditions. These contaminants are often produced with desired aroma compounds that strongly affect the quality of food and the taste experience, as with acrylamide, or they can form during purification of the raw material, as with trans fatty acids. Thus, it becomes difficult in practice to avoid their formation completely without changing food to an extent that it no longer fits the expectation of the consumer or it decreases the safety of the food, respectively. In a first publication Zelinkova et al. [94], described the discovery of fatty acid esters of 3-monochloro-1,2-propanediol (3-MCPD esters) in a series of 25 virgin and refined edible oils. Although in virgin oils the levels were low, they found higher amounts of 3-MCPD esters in refined oils and oils obtained from roasted raw materials. In December 2007, 3-MCPD esters became the focus of further attention when the official laboratory of Stuttgart and the Max Rubner-Institut in Germany announced findings of 3-MCPD esters in different edible fats and oils. 3-MCPD esters are compounds in which the basic structure of the 3monochloro-1,2-propanediol can be connected to different fatty acids. Monoesters bearing one fatty acid at the glycerol base body, as well as diesters bearing two linked fatty acids, are possible. Free 3-MCPD, not esterified with fatty acids, has long been known to form as a reaction product in the processing of acid hydrolyzed vegetable protein [21,84]. The finding of 3-MCPD esters in edible oils is important because an assessment of the International Agency for Research on Cancer (IARC) identified free 3-MCPD as “possibly carcinogenic to humans” (group 2B) [37]. The German Federal Institute for Risk Assessment (BfR) assumed Copyright © 2022 AOCS Press. Processing Contaminants in Edible Oils Published by Elsevier Inc. All rights reserved. ISBN 978-0-12-820067-4 Published in cooperation with https://doi.org/10.1016/B978-0-12-820067-4.00003-6 American Oil Chemists Society.

23

24

Processing Contaminants in Edible Oils

already in 2007 that 3-MCPD esters are 100% degraded by the enzymes in the human body, releasing free 3-MCPD [6], which is justified by toxicological studies [2,9,19]. In addition, 2-MCPD esters and glycidyl esters were found in edible oils. These compounds are also considered to be potentially hazardous to human health. Herein, the different classes of compounds including bound 2-MCPD, 3-MCPD, and glycidol are summarized as “3-MCPD esters and related compounds” if the total amount of these compounds has been determined. One conclusion of the assessment from EFSA and BfR was that industry should search for alternative techniques to reduce the formation of 3MCPD and glycidyl esters during oil processing. Zelinkova et al. [94] assumed that the heat treatment during roasting or refining could be responsible for the formation of the esters. Matthäus [55] showed that the deodorization step is most critical, a finding later confirmed by Franke et al. [28]. Thus, at the beginning of the 3-MCPD ester story, minimization strategies focused on the refining process, but later research showed that efforts to minimize the contents of the esters in edible oils should begin with the selection of the raw material or even with the cultivation of the seeds or fruits [56]. Looking at the entire production chain, three starting points for the mitigation and reduction of 3-MCPD and glycidyl esters in edible vegetable oils become evident: (1) reduction or avoidance of precursors in the raw material before processing and selection of suitable raw materials for the production of oils and fats, respectively; (2) changing of the conditions of the oil extraction and particularly of the refining process and introduction of new steps into the process; and (3) reduction of the esters in the refined oil by suitable absorbent materials or enzymatic treatment. In the last fifteen years since the first publication about the occurrence of 3-MCPD esters in edible oils [94], many papers have been published showing that mitigation is possible. Several of the approaches, successful in laboratory or pilot plant scale, have meanwhile been adopted by the oil refining industry. As a result of these studies, several documents summarizing mitigation strategies for industry have been published. The first document was released in 2015 by the EU Vegetable Oil and Protein Meal Industry Association (FEDIOL) giving information about the refining process, identifying critical areas with regard to 3-MCPD and glycidyl esters formation and providing information on credible mitigation techniques known in 2014 and available in published literature for the prevention and reduction of these compounds [27]. Later the “Toolbox for the Mitigation

Mitigation of MCPD and glycidyl esters in edible oils

25

of 3-MCPD Esters and Glycidyl Esters in Food” was published in 2016 and organized by the leadership of the Food Federation Germany (formerly German Federation for Food Law and Food Science) together with representatives from the German food sector, research institutes and private laboratories [7]. This toolbox details tested “tools” or strategies along the edible oils processing chain to aid industry in reducing the levels of 3MCPD esters and glycidyl esters in their products. In 2019, the 42nd Codex Alimentarius Commission adopted a “Code of Practice (COP) for the reduction of 3-Monochloropropane-1,2- diol esters (3-MCPDE) and glycidyl esters in refined oils and food products made with refined oils” (CXC 79e2019) [26]. The aim of this COP is to provide authorities, producers, and manufactures with information on how to mitigate 3MCPD and glycidyl esters in edible oils and food products. In 2018, the European Union (EU) defined limits for glycidyl esters in edible oils and infant formulas in regulation (EU) 2018/290, and, more recently, limits were also issued for 3-MCPD esters (regulation (EU) 2020/ 1322) effective January of 2021. Most of the work on the mitigation of 3MCPD and glycidyl esters has been done with palm oil because it has a high capacity for the formation of the esters in comparison to most other edible oils and it is an important edible oil in human nutrition. This resulted in more research interest in palm oil in comparison to other edible oils, but most of the results presented here can also be applied for the mitigation of the esters in other oils. In the following discussion, a distinction is made between the capability of crude oils to form 3-MCPD esters and related compounds and the formation in oils that undergo the entire refining process. The measurement of the capability of crude oils to form the esters is a simulated deodorization by heating the crude oil at a higher temperature and measuring the content of formed esters under these standardized conditions. The refining process includes not only the deodorization but also the other steps necessary to obtain high quality oils, such as degumming and bleaching. In that case, the content of 3-MCPD esters and related compounds produced in the final oil provides evidence about the effectiveness of the refining process as a whole, but not about the capability of the crude oil to form the esters. Matthäus et al. [56] investigated the capability of vegetable oils to form 3-MCPD esters and related compounds by heating 10 g of the crude oil samples at 240
C for 2 h. The authors mentioned that a lab scale deodorization under similar temperature and time conditions resulted in comparable formation of 3-MCPD esters and related

26

Processing Contaminants in Edible Oils

compounds. A similar approach was used by Craft et al. [17]; who heated 0.5 mL of oil in sealed glass ampoules under nitrogen for 2 h at 270
C before analyzing the content of formed glycidyl esters. Shimizu et al. [73] also used airtight, closed tubes flushed with nitrogen to remove the air in the headspace and heated at 240
C for 2 h. Since the first publication of this book in 2013, a series of new publications have appeared dealing with various aspects of minimizing 3-MCPD and glycidyl fatty acid esters in edible oils. Much of the mitigation research has continued to focus on reaction pathways and the influence of precursors on contaminant formation, but other studies have also detailed the contribution of the refining steps to the mitigation process. The aim of this chapter is to present an overview of the effects of the individual steps of the entire value-added chain of palm oil (and other edible oils) refining on the mitigation of 3-MCPD esters and related compounds, as well as the influence of precursors on ester formation. Publications from the last 14 years (since 2007) on this topic are considered and discussed.

Influence of precursors on ester formation Formation of 3-MCPD and glycidyl esters in fats and oils is a complex problem, in which the presence of chlorine donating ions; appropriate precursors such as triacylglycerols (TAGs), mono- (MAGs) and diacylglycerols (DAGs), phospholipids, or glycerol; as well as processing conditions (e.g., temperature, time, pH) all may play an important role. All these precursors exist in vegetable fats and oils, and during processing, sufficient thermal treatment occurs. In addition, during processing, other precursors may be introduced into the process by addition of water, citric or phosphoric acid, sodium hydroxide solution or bleaching earths, and the reaction conditions in the oil may be changed to favor the formation of the esters. Before looking at the effect of different precursors, it is interesting to know more about the capability of different types of vegetable oils to form the esters. This can give an indication of the relationship between composition and ester formation. Matthäus et al. [56] investigated the capability of vegetable oils to form 3-MCPD esters and related compounds. This screening revealed that avocado oil, extra virgin olive oil, rapeseed oil, soybean oil, and palm kernel fat formed only about 1 mg of 3-MCPD esters and related compounds per kilogram of oil under the applied conditions, while the levels in palm oil, corn oil, and coconut oil were markedly higher at up to 14 mg/kg

Mitigation of MCPD and glycidyl esters in edible oils

27

Figure 3.1 Capability of different vegetable oils to form 3-MCPD esters and related compounds.

(Fig. 3.1). Later, they found that hazel nut oil also seems to have a very high potential to form the esters but the reasons for this are not yet clear. Particularly for palm oil, a wide range in levels of formation of the esters was found, which implies that different factors may influence formation of the esters. Similar results were published by Weisshaar [89] showing higher amounts of 3-MCPD and glycidyl esters for palm oil, while corn oil, rapeseed oil, soybean oil, and sunflower oil only formed moderate amounts of the esters. A higher capability for the formation of 3-MCPD esters and related compounds was not seen in other fruit flesh oils, as heat treatment of both avocado oil and extra virgin olive oil resulted in relatively low amounts of 3-MCPD esters and related compounds. However, olive pomace oils of poor quality (lampante oils), not meeting the categories of extra virgin or virgin, showed a significantly higher capability for the formation of the esters. This finding was confirmed by Yan et al. [92] who found 0e6 mg 3-MCPD esters (3-MCPDE)/kg, 0e1.5 mg 2-MCPD esters (2-MCPDE)/kg, and 0e1 mg glycidyl esters (GE)/kg oil, with much higher amounts observed in lower grade olive oils after refining than in extra virgin olive oil. It was suggested that the detection of esters in highquality extra virgin olive oil (EVOO) could be used as a parameter for the identification of an extra virgin olive oil combined with refined one. The authors defined a detection limit of 2% when detecting lower grade oils in

28

Processing Contaminants in Edible Oils

EVOO when using 3-MCPD esters, 5% when using 2-MCPD esters, and 13%e14% when using GE. Similar findings were published by Kamikata et al. [39] who observed ester levels ranging from 0.28 to 3.77 mg/kg for 3MCPDE, 0.17e1.91 mg/kg for 2-MCPDE and not detected to 1.88 mg/ kg for GE in olive oils, while lower amounts were detected in EVOO. The authors also suggested analysis of the esters could be used as an indicator of adulteration of EVOO. Razak et al. [69] investigated 105 crude palm oils, finding that 80% of them contained no 3-MCPD esters, 10% contained esters below the detection limit, and the remaining 10% had concentrations below 0.9 mg/ kg. In 50% of 56 bleached oil samples, the authors found 3-MCPD esters below the detection limit, with the remaining samples ranging from 0.25 to 1.8 mg/kg. In refined bleached and deodorized (RBD) palm oils, the authors found that only 1% of the samples contained 3-MCPD esters below the detection limit, while 56% of the samples showed values between 0.25 and 2.0 mg/kg. In the remaining 43%, the content of 3-MCPD esters was higher, with a maximum content of 5.8 mg/kg. The 3-MCPD ester content of almost 2% of RBD palm olein samples was below the detection limit, and most of the samples contained 3-MCPD ester concentrations in the range of 1.4 to 3.2 mg/kg, with a maximum of 4.0 mg/kg. In RBD palm stearin, 1.5 mg/kg 3-MCPD esters were found in 20% of the samples, but 40% of the samples contained less than the mean value (1.25 mg/kg). The partitioned coefficient of the esters in palm olein to palm stearin was found to be about 2.0, because [69] found 3.0 to 3.8 mg/kg 3-MCPD esters in RBD palm olein and only 1.6 to 1.8 mg/kg in palm stearin, indicating that the esters prefer to partition into the more liquid or unsaturated phase during fractionation. Kyselka et al. [45] has also described reduced amounts of 3-MCPD and glycidyl esters in hard and super stearin fractions of palm oil, whereas higher amounts were found in the olein fractions, with highest levels in top olein. Under hydrogenation conditions with commercial and tailor-made catalysts, the esters were not stable and rapid degradation occurred (3-MCPD esters 51%e70%; glycidyl esters 1%e42%, depending on the type of nickel catalyst) into non-carcinogenic and non-genotoxic products by hydrogenolysis and deoxygenation reactions. Bleaching and deodorizing of hydrogenated palm stearins resulted in the repeated formation of glycidyl esters up to 1.01e1.76 mg/kg. Kuhlmann [41] showed that, in addition to palm oil products, walnut oil, hazelnut oil, fish oil (salmon), and grape kernel oil contained higher relative amounts of both 3-MCPD and glycidyl esters when compared to

Mitigation of MCPD and glycidyl esters in edible oils

29

typical seed-oil crops. He also investigated a wide range of different oils not only for the content of 3-MCPD and glycidyl esters, but also for 2-MCPD esters. Higher amounts of 3-MCPD esters were found in fish oils (0.7e13 mg/kg), evening primrose oil (0.8e5.2 mg/kg), hazelnut oil (19 mg/kg), grapeseed oil (0.8e4.2 mg/kg), olive oil (0.3e1.2 mg/kg), and walnut oil (1.2e19 mg/kg). Higher amounts of 3-MCPD esters were not necessarily accompanied by higher amounts of glycidyl esters. In coconut oil (0.5e3.0 mg/kg), fish oil (0.1e1.2 mg/kg), grapeseed oil (0.3e2.5 mg/ kg), palm kernel oil (0.3e2.5 mg/kg), shea butter (3.1 mg/kg), and walnut oil (0.7e1.4 mg/kg), glycidyl esters were detected. In all refined oils, 2MCPD esters were found with levels between 20% (evening primrose oil) and 60% (grapeseed oil and walnut oil) of the amount of 3-MCPD esters. Higher amounts of 3-MCPD esters and related compounds in palm oil, corn oil, coconut oil, and lower quality olive oil may indicate that oil quality, characterized by the lipid composition, can have a strong impact on the capability for formation of the esters, as higher amounts of free fatty acids and DAGs are often characteristic to these types of oils due to enzymatic degradation of TAG during storage. For palm oils from different locations [56], found remarkable differences in the capability to form 3-MCPD esters and related compounds under standardized heating conditions. While palm oil from Ghana formed only 1.5 mg/kg, palm oil from Malaysia produced up to 14 mg/kg, with a variation in oils from different Malaysian regions ranging from 3 to 14 mg/ kg (Fig. 3.2). This great variation might not only be explained by differences in climate, soil, and growth conditions because genotype, harvest technique, and processing approach all have a strong influence on the capability of the crude oil to form 3-MCPD esters and related compounds. These conditions influence the amount of chlorine-containing compounds (input of inorganic or organic chlorine-containing compounds via salts in the soil or pesticides) and partial acylglycerols in the oil, which can act as precursors. The discussion of influences of agronomic factors on the formation of 3-MCPD and glycidyl esters will be revisited later in this chapter. Chloride From the structure of 3-MCPD esters it becomes clear that chloride or other chlorine-containing compounds have to act as chlorine donors for the formation of 3-MCPD esters. For the development of mitigation strategies, knowledge about the source of the chlorine donor was considered to be a key point. At the beginning of the discussion about the formation of

30

Processing Contaminants in Edible Oils

Figure 3.2 Capability of palm oils from different origin to form 3-MCPD esters and related compounds.

3-MCPD esters in edible oils, the chloride content of water used for the generation of strip steam during deodorization was suspected to have an effect as a source for chloride, but [63] showed that the type of water and the content of chloride have no influence. Matthäus et al. [56] assumed that the chlorine donor must be available in an oil-soluble form to enable the reaction with other precursors. Craft et al. [17] and Nagy et al. [59] described the role of chlorine donors in 3-MCPD ester formation, discussing their origin and their changes during plantation, harvest, and oil mill processing (see Chapter 1). Later, in 2018 Tiong et al. [82] showed the presence of organochlorine compounds in vegetable oils such as crude palm, soybean, rapeseed, sunflower, corn, coconut and olive oil by measuring chlorine isotope mass patterns with high-resolution mass spectrometry. In crude palm oil they identified the organochlorine compounds as constituents of wax esters, fatty acids, diacylglycerols and sphingolipids, with sphingolipid-based organochlorines as the most active precursors for the formation of 3-MCPD esters. For the development of mitigation strategies this finding is important as the levels of these compounds can indicate the potential of the crude oils to form 3-MCPD esters, thereby suggesting that the refining process should be adjusted accordingly. There are many potential sources for inorganic chloride in palm plantations. Beginning with the soils, it is clear that saline soils, found by the sea or near river mouths that are constantly inundated by sea or brackish water,

Mitigation of MCPD and glycidyl esters in edible oils

31

contain higher amounts of chloride in the form of salts. Because soil moisture must be optimal for good growth and yield, problematic soils like peat, saline, acid sulfate, or podzol have to be drained or irrigated [90]. In the case of irrigation with treated wastewater, ferric chloride is often used as coagulant in water treatment in Malaysia [1]. In particular, the use of FeCl2þ and FeCl3þ seems to support the formation of 3-MCPD esters directly from TAG, as shown by Zhang et al. [95] in model experiments. They explained this observation by the potential of Fe ions to promote free radical generation. Cheng et al. [11] showed that Fe3þ is involved in glycidyl ester formation by catalyzing free radical generation and thereby promoting the free radical mediated mechanisms for the formation of cyclic acyloxonium free radical intermediates. Moreover, fertilizers are needed to meet the requirements of oil palm growth. Mechanical fertilizer spreaders and erial application techniques are adopted and among potassium fertilizers, potassium chloride is the cheapest and widely used in oil palm plantations [81]. In addition [17], mentioned the use of ammonium chloride in fertilizers to support oil palm growth and bunch yield. Due to sustainable plantation management, empty fruit bunches (EFBs) and palm oil mill effluent (POME) are used as fertilizers as well [90], which could enhance the risk of an enrichment of chloride ions within the plantation. In addition to the environmental sources of inorganic chloride, many different chlorine-containing substances are used in palm oil plantations for weed and pest control, such as the herbicides paraquat, diuron 2,4-D amine, dicamba, or fluroxypyr [42,70], as well as the insecticides trichlorphon (against bagworm and nettle caterpillars), endosulfan (nettle caterpillars and bunch moth), phosmamidon and leptophos (nettle caterpillars), lambda-cyhalothrin and cypermethrin (rhinoceros beetle), and diflubenzuron and cyfluthrin (bunch moth) [71,87]. Similar results were found by Freudenstein et al. [29] when adding sodium chloride or tetra-butylammonium chloride (TBAC) to a model system consisting of palm oil after the removal of polar compounds. The authors showed that mainly 3-MCPD esters formed after addition of sodium chloride, although glycidyl esters were also produced. For TBAC, the content of 3-MCPD esters increased linearly with increasing chloride content, but the slope was significantly weaker than after sodium chloride addition. The percentage of glycidyl esters on the total value was significantly greater after addition of TBAC than after addition of sodium chloride. A possible reason for this finding might be the influence of the added compounds on the pH value. Shimizu et al. [75] showed that

32

Processing Contaminants in Edible Oils

3-MCPD ester levels were strongly influenced by the chloride content with a conversion ratio of chloride to 2-MCPD esters of 2%e8%. However, the author’s stated that most of the chloride remained inactive. Nevertheless, they assumed that removal of the chloride source from the crude oil before heating during deodorization would be an important approach to mitigate 3-MCPD ester formation. In model systems consisting of crude or refined oil and NaCl solution Li et al. [47], detected a high positive correlation between the amount of 3-MCPD esters and the concentration of sodium chloride when simulating the thermal processing of edible oils by heating of the well-blended mixtures at 160
C for 1 h in a silicone oil bath. They observed increasing formation of the esters until an amount of 10 g/100 g Cl, at which point decreasing formation of 3-MCPD esters was observed with higher amounts of sodium chloride. The authors also showed that Fe3þ promoted the formation of the esters. This was confirmed by Zhang et al. [96] who showed that only FeCl2 and FeCl3 were able to form 3-MCPD esters from tristeraroylglycerol. They explained this by the potential of Fe ions to promote free radical generation under experimental conditions using electron spin resonance (ESR) analysis. Davidek et al. [20] and Velisek et al. [85] described the finding of monoesters and diesters of 3-chloropropanediol after hydrolysis of triacylglycerols with hydrochloric acid at 110
C. Gardner et al. [30] discussed, in connection with toxic oil syndrome in Spain, the formation of C16 and C18 fatty acid monoesters and diesters of chloropropanediol in rapeseed oil samples with high levels of chlorine. In this investigation, the authors found chloropropanediol esters in the TLC fraction of the toxic oils and they explained the finding with the reaction of hydrochloric acid during refining with the aim to remove aniline, which was used as a denaturant of rapeseed oil [91]. Karahadian and Lindsay [40] also mentioned the potential production of chloropropanediol in hydrochloric acidedeodorized fish oils, if cod liver oil was deodorized with 0.01 N hydrochloric acid for 2 h. At low deodorization temperatures (near 100
C), hydrochloric acideacidified steam did not cause the formation of the esters. The knowledge about the formation mechanism of 3-MCPD esters as result of the decomposition of chlorine-containing compounds during deodorization is of great importance for the development of mitigation strategies and is discussed fully in the previous chapter. This can explain the reduction of 3-MCPD esters in palm oils deodorized after washing with polar solvents like water or waterealcohol mixtures, as suggested by Matthäus et al. [56]. The washing process removes polar chlorine-containing

Mitigation of MCPD and glycidyl esters in edible oils

33

compounds from the oil, resulting in a remarkably lower capability to form 3-MCPD esters. Because of the conversion of polar chlorinated compounds into more nonpolar compounds during processing [17], suggested that it would be easier to remove the reactive chlorine species from the pulp during oil extraction in the oil mill rather than from the crude palm oil. The earlier the removal of the chlorinated species takes place, the more effective it is for the mitigation of 3-MCPD esters. Knowledge about the compounds involved and insight into the formation pathways can help to explain the mitigation effect of processing steps like washing palm oil. It also helps to improve the development of strategies to minimize the formation of 3-MCPD esters. Thus, the introduction of a washing step by Matthäus et al. [56] was an effective approach, later corroborated by the work of [23,59]. Acylglycerols Vegetable fats and oils are comprised of 88%e96% triacylglycerols (TAGs). For most fats and oils, the content of diacylglycerols (DAGs) is low, with amounts ranging from 1% to 2%, but for palm oil the amount of DAGs ranges from 4% to 12%, with a mean value of about 6.5% [34,52,76]. Craft et al. [18] found DAG levels in commercial samples of palm oil and palm olein ranging from 3.91 g/100 ge7.55 g/100 g, with a median at 5.37 g/ 100 g and an average of 5.34 g/100 g oil. Olive pomace oils, which are not classified as edible olive oil for quality reasons, are often characterized by more than 3% DAGs [14]. Fig. 3.3 shows the lipid composition of different vegetable oils, with higher amounts of DAG for palm oil, palm kernel oil, rice bran oil, and grapeseed oil. So-called DAG-rich oils, which are mainly intended for particular nutritional or dietetic uses, contain between 80% and more than 90% DAGs. DAGs are in general very low in most commonly used vegetable oils. Rahn and Yaylayan [65] stated that the formation of 3-MCPD esters is the result of a nucleophilic attack of chloride ions on lipids via acyloxonium ion formation. In addition, the formation of the acyloxonium ion from an appropriate precursor is influenced by the steric hindrance effect and the leaving group involved. As precursors, partial acylglycerols, phospholipids, carotenoids, and other lipid-related compounds are suspected [16,66]. A similar pathway for the formation of 3-MCPD esters via acyloxonium ion or cyclic acyloxonium ion was postulated by Destaillats et al. [23]. They heated a mixture of TAG and DAG with a proportion of TAGeDAG adjusted to 2%, which is comparable to the average composition of

34

Processing Contaminants in Edible Oils

Figure 3.3 Lipid composition of different vegetable oils.

vegetable oils. After addition of a chlorine-containing donor, the production of 3-MCPD esters occurred from TAG (triolein) as well as DAG (diheptadecanoate) upon thermal treatment, with the extent of formation higher for TAGs. When correcting the results with the weight percentage of TAG and DAG, TAG was a bit more than two times more likely to form the MCPD diesters. In contrast [73], found that 3-MCPD esters were generated in higher levels from diolein and especially monoolein in the presence of a chloride source, while only small amounts were found for triolein after heating the samples at 240
C for 2 h. One reason for these different results could be that the efficacy of the chlorine donors used in different experiments also affects the kinetics of the formation of the esters from different lipid sources. Zhang et al. [96] showed that free radicals were suitable to mediate the formation of 3-MCPD diesters from DAGs via a cyclic acyloxonium free radical and assumed that this finding might lead to the improvement of oil and food processing conditions. Fig. 3.4 shows a poor correlation between the amount of DAGs and the formation of 3-MCPD esters and related compounds for industrial samples (R ¼ 0.4) [56]. A similarly poor correlation was described by Craft et al. [18] between DAG and the formation of glycidyl esters, where the authors explained that finding by the fact that the tested oil samples were procured from different suppliers who likely used different refining conditions. The authors also described that when focusing the correlation only on samples

Mitigation of MCPD and glycidyl esters in edible oils

35

Figure 3.4 Influence of the content of diacylglycerols on the capability of palm oil to form 3-MCPD esters and related compounds. ( ) Pure diacylglycerols added to virgin rapeseed oil; ( ) different crude oils from the industry.

36

Processing Contaminants in Edible Oils

that originate from one given refinery, a positive linear correlation of R2 ¼ 0.8 was found. The explanation was that these samples likely have been deodorized under similar conditions, resulting in a good correlation between concentrations of DAG and the formation of glycidyl esters. The addition of DAGs to a virgin rapeseed oil with a low capability to form the esters resulted in a very good correlation (R > 0.8) between the content of DAGs and the formation of 3-MCPD esters and related compounds (Fig. 3.4) [56]. This linear increase is attributed mainly to the increase of glycidyl esters, since the increase of glycidyl esters was likely the driving factor for the increase in the total amount of the esters during this experiment. Usually, rapeseed oil from sound seeds shows a low concentration of DAGs ( Lascrobyl palmitate (1.51 mg/kg) > a-tocopherol (1.76 mg/kg). Rosemary extract, however, can have a strong odor, which may add unwanted fragrance to the oil.

Mitigation of MCPD and glycidyl esters in edible oils

55

Cheng et al. [11] showed that the artificial antioxidant tert-butyl hydroquinone (TBHQ) is suitable for decreasing the content of glycidyl esters in bleached oils after heating in comparison to the control. This was explained due the fact that the antioxidant directly affects the formation of the cyclic acyloxonium intermediate. Li et al. [46] investigated the effects of six antioxidants, butylated hydroxytoluene (BHT), butylated hydroxy anisole (BHA), tert-butyl hydroquinone (TBHQ), propyl gallate (PG), Lascorbyl palmitate (AP), and alpha-tocopherol (VE) on the formation of 3-MCPD esters in chemical models and in oil models. They found strong inhibiting capacities for all antioxidants, with the strongest effect (44% decrease in 3-MCPD esters) observed for TBHQ (66 mg/kg oil), followed by AP (40% for 166 mg/kg oil) and PG (37% for 84 mg/kg oil), after heating of rapeseed oil at 230
C for 30 min. For BHA (31% for 72 mg/kg oil), BHT (29% for 88 mg/kg oil), and VE (22% for 172 mg/kg oil) the inhibiting capacity was weaker. The authors also showed that the deodorization temperature strongly affected the inhibiting capacity of PG and VE, with increasing temperatures resulting in decreased 3-MCPD ester formation. The effect of chelating agents like EDTA-2-Na on the reduction of 3MCPD ester formation was shown by Zhang et al. [96] in a model system consisting of tristearylglycerol and FeCl3. The authors showed that controlling the availability of radical promoting ions like Fe2þ or Fe3þ with chelating agents may reduce the amount of 3-MCPD and glycidyl esters during refining. Today it is generally accepted that the formation of 3-MCPD and related compounds can be prevented if the acidity of the oil is changed toward a neutral pH value. Smidrkal et al. [78] suggested the addition of bicarbonates or carbonates to the oil during deodorization in order to neutralize free hydrogen chloride. During laboratory scale experiments, the addition of potassium hydrogen carbonate or sodium bicarbonates to sunflower oil spiked with sodium chloride and heated to 240
C led to a remarkable reduction of the capability to form 3-MCPD esters. It was shown that the most effective bases were bicarbonates or carbonates of alkali metals that acted even at concentrations equivalent to the potassium hydroxide used for neutralization of free fatty acids present in the oil. These results are in good agreement with investigations from Freudenstein et al. [29], showing that the addition of sodium carbonate and sodium hydrogen carbonate to a model oil with a low capability for the formation of the

56

Processing Contaminants in Edible Oils

esters and spiked with sodium chloride resulted in a reduction of the formation of the esters from about 6 mg 3-MCPD esters/kg to about 2 mg/kg (with sodium carbonate) and 1 mg/kg (with sodium hydrogen carbonate), respectively. While the addition of sodium carbonate decreased the formation of 3-MCPD esters before reaching a steady state after adding 5 mmol Na2CO3/kg oil, the addition of NaHCO3 resulted in a decrease of 3-MCPD esters beginning at a concentration of 1 mmol/kg.

Removal from fully refined oils (post-refining) Besides avoiding or removing precursors for the formation of the esters or changing the purification process of crude oil, another promising approach to produce edible vegetable oils low in 3-MCPD and glycidyl esters is the removal of the esters from the fully refined oil. Based on the different polarities of 3-MCPD and related compounds in comparison to TAGs, the removal of the compounds should be possible. Examples for the treatment of oils by adsorption materials to remove polar compounds are known from the treatment of used frying oils [31,50]. Strijowski et al. [80] suggested the removal of the esters from palm oil by using different adsorption materials after the deodorization step. They found that calcined zeolite and a synthetic magnesium silicate were able to reduce the amount of 3-MCPD esters and related compounds by about 40%, with the reduction based largely on the removal of glycidyl esters; 3-MCPD ester contents were not influenced by the treatment (Fig. 3.13). They also found materials, such as amorphous magnesium silicate, that actually increased the content of the esters after treatment of the oil. Calcinated zeolite and synthetic magnesium silicate showed remarkable differences in the transposition rate. While calcined zeolite reached the total reduction effect at 60
C, the activation of magnesium silicate required temperatures of 80
C. The application of calcined zeolite enabled a spontaneous reduction of 3-MCPD esters and related compounds after addition, whereas the reduction after addition of magnesium silicate required at least 60 min to reach similar results. The authors found a slight improvement of the oil quality with calcined zeolite, but an adverse effect of the treatment with synthetic magnesium. In contrast, the oxidation stability of the oil was improved by the treatment with synthetic magnesium, and no significant effect was found for calcined zeolite. Shimizu et al. [74] described the removal of glycidyl esters from both TAG and DAG oils by the use of activated bleaching earth (ABE). They

Mitigation of MCPD and glycidyl esters in edible oils

57

Figure 3.13 Content of 3-MCPD esters and related compounds in refined palm oil after treatment with different adsorbents (total column length: 3-MCPD esters and related compounds; solid lines: amount of 3-MCPD esters and related compounds in refined palm oil before treatment; dark columns: 3-MCPD esters; dotted lines: amount of 3-MCPD esters in refined palm oil before treatment). (With kind permission of Dr. K. Franke Leibnitz University, Hannover, Germany).

found that glycidyl palmitate was completely removed from glycerol dioleate, not by absorption of glycidyl esters on ABE, but rather by transformation of glycidyl esters to glycerol monopalmitate, glycerol palmitate oleate, and glycerol dipalmitate. The authors proposed a ring-opening reaction of glycidyl esters with water contained in the ABE and in the bulk oil, followed by an interesterification reaction among the resultant monopalmitate and the glycerol dioleate of the bulk oil. A material for the removal of 3-MCPD esters from refined vegetable oils was described by Bertoli et al. [4], demonstrating the effect of a treatment with carboxymethylcellulose or a cation-exchange resin on the content of the esters. Cheng et al. [12] used acid-washed oil palm wood-based activated carbon (OPAC) to remove glycidyl esters from palm oil. At a concentration of 30 mg OPAC/100 mL palm oil, a 95% reduction of glycidyl esters was observed (from 3.75 to 0.2 mg/kg). The process showed a strong positive correlation with the initial concentration of glycidyl esters and the amount of added acid-washed OPAC, while the amount of absorbed glycidyl esters decreased with increasing temperature. From these findings, the authors

58

Processing Contaminants in Edible Oils

concluded an exothermic adsorption process on acid-washed OPAC must be taking place. In addition, the results revealed that removal of glycidyl esters from palm oil was not only a result of adsorption at OPAC but, in the case of acid washed OPAC, also a degradation at activated sites with acidic character. The quality of the palm oil based on oxidation stability, Gardner value, and acid value was not affected by treatment with acid-washed OPAC. The degradation of free 3-chloro-1,2-propanediol with Saccharomyces cerevisiae was first described by Bel-Rhlid et al. [3]. Later Bornscheuer and Hesseler [8], published their study on the removal of 3-MCPD esters from refined oils by an enzymatic treatment. In laboratory scale experiments, they showed the removal of 3-MCPD esters from aqueous and biphasic systems by converting the contaminants via an enzyme cascade into glycerol. In detail, 3-MCPD esters were converted first in a biphasic system into free 3-MCPD by Candida antarcitca lipase A, and afterwards, the free 3-MCPD was converted into glycerol via treatment with halohydrin dehalogenase from Arthrobacter sp. AD2 and an epoxide hydrolase from Agrobacterium radiobacter AD1. Recommendations Summarizing all the findings, the following measures of mitigation can significantly decrease the content of 3-MCPD esters and glycidyl esters in the resulting oils. Because processed oils are consumed worldwide and commonly used in products for which it is desirable to have low amounts of contaminants like 3-MCPD and glycidyl esters, such as infant formula, these measures should prove to be extremely useful: Optimization of the palm fruit growing, harvest, and preprocessing in the oil mill and selection of crude oil with low contents of precursors. Optimization of oil extraction processes, avoiding the formation of lipophilic organochlorines from hydrophilic ones. Introduction of an additional washing step prior to refining. Neutralization prior to deodorization to increase pH values. Use of chemical refining, which appears to be the better option for producing oils with lower contents of 3-MCPD esters and related compounds, in comparison to the commonly established physical refining in the case of palm oil. Use of natural bleaching earths and acid-activated bleaching earths with more neutral pH values. Addition of alcohols, diacetin, carbonates, or bicarbonates before deodorization.

Mitigation of MCPD and glycidyl esters in edible oils

59

Application of short-path distillation instead of deodorization. Reduction of the temperature load during deodorization, for example, by dual deodorization with different temperature profiles. Post-treatment of fully refined oils with appropriate adsorbents or enzymes.

References [1] Abdollah NAB. Some problem in chemical and physical treatment of water supply. Ph.D. Thesis. Universiti Teknologi Malaysia; 2010. [2] Barocelli E, Corradi A, Mutti A, Petronini PG. Comparison between 3-MCPD and its palmitic esters in a 90-day toxicological study. Scientific report submitted to EFSA, CFP/EFSA/CONTAM/2009/01. 2011 [Accessed 5 April 2021]. [3] Bel-Rhlid R, Talmon JP, Fay LB, Juillerat MA. Biodegradation of 3-chloro-1,2propanediol with Saccaromyces cerevisiae. J Agric Food Chem 2004;52:6165e9. [4] Bertoli C, Cauville F, Schoonman AJH. A deodorized edible oil or fat with low levels of bound MCPD and process of making by carboxmethyl cellulose and/or resin purification. US patent 20120122983 A1. Switzerland: Nestec S.A.; May 17, 2012. [5] Sim BI, Muhamad H, Lai OM, Abas F, Yeoh CB, Imededdine AN, Yih PK, Chin PT. New insights on degumming and bleaching process parameters on the formation of 3monochloropropane-1,2-diol esters and glycidyl esters in refined, bleached, deodorized palm oil. J Oleo Sci 2018;67:397e406. [6] BfR Stellungnahme N. Säuglingsanfangs- und Folgenahrung kann gesundheitlich bedenkliche 3-MCPD estersttsäureester enthalten. 2007. 047/, http://www.bfr.bund. de/cm/343/saeuglingsanfangs_und_folgenahrung_kann_gesundheitlich_bedenkliche_ 3_mcpd_fettsaeureester_enthalten.pdf. [Accessed 5 April 2021]. [7] BLL. Toolbox for the mitigation of 3-MCPD esters and glycidyl esters in food. 2016. https://www.lebensmittelverband.de/download/toolbox-for-the-migration-of-3mcpd-esters-and-glycidyl-ester. [Accessed 2 April 2021]. [8] Bornscheuer UT, Hesseler M. Enzymatic removal of 3-monochloro-1,2-propandiol (3-MCPD) and its esters from oils. Eur J Lipid Sci Technol 2010;112:552e6. [9] Buhrke T, Weisshaar R, Lampen A. Absorption and metabolism of the food contaminant 3-chloro-1,2-propanediol (3-MCPD) and its fatty acid esters by human intestinal Caco-2 cells. Arch Toxicol 2011;85:1201e8. [10] Calta P, Velisek J, Dolezal M, Hasnip S, Crews C, Reblova M. formation of 3chloropropane-1,2-diol in systems simulating processed foods. Eur Food Res Technol 2004;218:501e6. [11] Cheng W, Liu G, Liu X. Effects of Fe3þ and antioxidants on glycidyl ester formation in plant oil at high temperature and their influencing mechanisms. J Agric Food Chem 2017;65:4167e76. [12] Cheng W, Liu G, Wang X, Han L. Adsorption removal of glycidyl esters from palm oil and oil model solution by using acid-washed oil palm wood-based activated carbon: kinetic and mechanism study. J Agric Food Chem 2017;65:9753e62. [13] Cheng W, Liu G, Guo Z, Chen F, Cheng K-W. Kinetic study and degradation mechanism of glycidyl esters in both palm oil and chemical models during hightemperature heating. J Agric Food Chem 2020;68:15319e26. [14] Catalano M, Deleonardis T, Comes S. Diacylglycerols in the evaluation of virgin olive oil quality. Grasas y Aceites 1994;45:380e4. [15] Chong CL. An overview of the effect of milling practice and storage on the quality of crude palm oil. In: Proceedings of the seminar on developments in palm oil milling technology and environmental management. Malaysia: Genting Highlands; May 16e17, 1991.

60

Processing Contaminants in Edible Oils

[16] Collier PD, Cromie DDO, Davies AP. Mechanism of formation of chloropropanols present in protein hydrolysates. J Am Oil Chem Soc 1991;68:785e90. [17] Craft BD, Nagy K, Sandoz L, Destaillat F. Factors impacting the formation of monochloropropanediol (MCPD) fatty acid diesters during palm (Elaeis guineensis) oil production. Food Addit Contam 2012;29:354e61. [18] Craft BD, Nagy K, Seefelder W, Dubois M, Destaillats F. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part II: practical recommendations for effective mitigation. Food Chem 2012;132:73e9. [19] Creutzenberg O, Berger-Preiß E. 3-MCPD- und glycidol fettsäureester e stand zur toxikologie e untersuchungen zur bioverfügbarkeit und metabolisierung. Berlin, Germany: BLL-OVID-Informationsveranstaltung; January 18, 2011. [20] Davidek J, Velisek J, Kubelka V, Janicek G, Simicova Z. Glyceral chlorohydrins and their esters as products of the hydrolysis of tripalmitin, tristearin and triolein with hydrochloric acid. Z Lebensm Unters Forsch 1980;171:14e7. [21] Davidek J, Velisek J, Kubelka V, Janicek G. New chlorine containing organic compounds in protein hydrolysates. In: Baltes W, Czedik-Eysenberg PB, Pfannhauser W, editors. Recent developments in food analysis, proceedings of EuroFood Chem I, Vienna, Austria, February 17e20. Weinheim: Deerfield Beach, FL; 1981. p. 322e5. [22] Destaillats F, Nagy KL, Sandoz L, Craft B. Plant oil refinement in the presence of alcohol. European patent application EP 2 502 501 A1. 2011. [23] Destaillats F, Craft BD, Sandoz L, Nagy K. Formation mechanisms of monochloropropanediol (MCPD) fatty acid diesters in refined palm (Elaeis guineensis) oil and related fractions. Food Addit Contam 2012;29:29e37. [24] Destaillats F, Craft BD, Dubois M, Nagy K. Glycidyl esters in refined palm (Elaeis guineensis) oil and related fractions. Part I: formation mechanism. Food Chem 2012b;131:1391e8. [25] Diks R. The effect of oil refining on contaminant removal and secondary product formation. In: Presented at 3rd Leipzig symposium, processing: quality and safety, Leipzig, Germany, March 17e18; 2010. [26] FAO/WHO. Code of practice for the reduction of 3-monochloropropane-1,2- diol esters (3-MCPDEs) and glycidyl esters (GEs) in refined oils and food products made with refined oils. CXC 79-2019. 2019. [27] FEDIOL. MCPD esters and glycidyl esters - review of mitigation measures, Revision 2015. June 24 , 2015. Ref. 15SAF108. [28] Franke K, Strijowski U, Fleck G, Pudel F. Influence of chemical refining process and oil type on bound 3-chloro-1,2-propanediol contents in palm oil and rapeseed oil. LWT - Food Sci Technol (Lebensmittel-Wissenschaft -Technol) 2009;42:175 1e1754. [29] Freudenstein A, Weking J, Matthäus B. Influence of precursors on the formation of 3MCPD and glycidyl esters in a model oil under simulated deodorization conditions. Eur J Lipid Sci Technol 2013;115:286e94. [30] Gardner AM, Yurawecz MP, Cunningham WC, Diachenko GW, Mazzola EP, Brumley WC. Isolation and identification of C16 and C18 fatty acid esters of chloropropanediol in adulterated Spanish cooking oils. Bull Environ Contam Toxicol 1983;31:625e30. [31] Gertz C. Optimising the baking and frying process using oil-improving agents. Eur J Lipid Sci Technol 2004;106:736e45. [32] Gibon V, DeGreyt W, Kellens M. Palm oil refining. Eur J Lipid Sci Technol 2007;109:315e35. [33] Gibon V, Vila Ayala J, Dijckmans P, Maes J, de Greyt W. Future prospects for palm oil refining and modifications. Oleagineux 2009;16:193e200. [34] Goh EM, Timms RE. Determination of mono- and diglycerides in palm oil, olein and stearin. J Am Oil Chem Soc 1985;62:730e4.

Mitigation of MCPD and glycidyl esters in edible oils

61

[35] Hamlet CG, Sadd PA, Gray DA. Generation of monochlorophropanediols (MCPDs) in model dough systems. 2. Unleavened doughs. J Agric Food Chem 2004;52:2067e72. [36] Hrncirik K, van Duijn G. An initial study on the formation of 3-MCPD esters during oil refining. Eur J Lipid Sci Technol 2011;113:374e9. [37] IARC. Some chemicals in industrial and consumer products, some food contaminants and flavourings, and water chlorination by-products. In: IARC monographs on the evaluation of carcinogenic risks to humans, vol. 101. Lyon, France: International Agency for Research on Cancer; 2012. [38] Incontech; n.d. http://www.incontech.com/equip_syst.htm. [Accessed 4 November 2013]. [39] Kamikata K, Vicente E, Arisseto-Bragotto AP, Rauen de Oliveira Miguel AM, Milani RF, Verdiani Tfouni SA. Occurrence of 3-MCPD, 2-MCPD and glycidyl esters in extra virgin olive oils, olive oils and oil blends and correlation with identity and quality parameters. Food Contr 2019;95:135e41. [40] Karahadian C, Lindsay RC. Low temperature deodorizations of fish oils with volatile acidic and basic steam sources. J Am Oil Chem Soc 1990;67:85e91. [41] Kuhlmann J. Determination of bound 2,3-Epoxy-1-propanol (glycidol) and bound monochloropropanediol (MCPD) in refined oils. Eur J Lipid Sci Technol 2011;113:335e44. [42] Kuntom A, Tan YA, Kamaruddin N, Yeoh CB. Pesticide application in oil palm plantation. Oil Palm Bull 2007;54:52e67. [43] Krisdiarto AW, Sutiarso L. Study on oil palm fresh fruit bunch bruise in harvesting and transportation to quality. Makara J Technol. 2016;20:67e72. [44] Krisdiarto AW. Mapping of bruise of oil palm fresh fruit bunch during loading and transportation from field to mill. Makara J Technol 2018;22:84e7. [45] Kyselka J, Matejkova K, Smidrkal J, Bercikova M, Pesek E, Belkova B, Ilko V, Dolezal M, Filip V. Elimination of 3-MCPD fatty acid esters and glycidyl esters during palm oil hydrogenation and wet fractionation. Eur Food Res Technol 2018;244:1887e95. [46] Li C, Jia HB, Shen MY, Wang YT, Nie SP, Chen Y, Zhou YQ, Wang YX, Xie MY. Antioxidants inhibit formation of 3-monochloropropane-1,2-diol esters in model reactions. J Agric Food Chem 2015;63:9850e4. [47] Li C, Zhou Y, Zhu J, Wang S, Nie S, Xie M. Formation of 3-chloropropane-1,2-diol esters in model systems simulating thermal processing of edible oil. LWT - Food Sci Technol (Lebensmittel-Wissenschaft -Technol) 2016;69:586e92. [48] Li DM, Qin XL, Sun BG, Wang WF, Wang YH. A feasible industrialized process for producing high purity diacylglycerols with no contaminants. Eur J Lipid Sci Technol 2019;121:1900039. [49] Lin SW. Enhancement of oil quality. In: Basiron Y, Jalani BS, Chan KW, editors. Advances in oil palm research. Malaysian palm oil board, vol. 2; 2001. p. 935e67. Malaysia. [50] Lin S, Akoh CC, Estes-Reynolds A. Recovery of used frying oils with adsorbent combination: refrying and frequent oil replenishment. Food Res Int 2001;34:159e66. [51] Lin SW, Kuntom A, Ibrahim NA, Ramli MR, Razak RAA. The possible mitigation procedures for the reduction of formation of chloropropanol esters and related compounds. Palm Oil Develop 2012;57:21e7. [52] Long K, Jamari MA, Ishak A, Yeok LJ, Latif RA, Ahmadilfitri A, Lai OM. Physicochemical properties of palm olein fractions as a function of diglyeride content in the starting material. Eur J Lipid Sci Technol 2005;107:754e61. [53] Martins PF, ItoM VM, Batistella CB, Maciel MRW. Free fatty acids separation from vegetable oil deodorizer distillate using molecular distillation process. Separ Purif Technol 2006;48:78e84. [54] Masukawa Y, Shiro H, Nakamura S, Kondo N, Jin N, Suzuki N, Ooi N, Kudo N. A new analytical method for the quantification of glycidol fatty acid esters in edible oils. J Oleo Sci 2010;59:81e8.

62

Processing Contaminants in Edible Oils

[55] Matthäus B. What are 3-MCPD esters? Presented at the ad-hoc seminar. In: What we know today about 3-MCPD fatty acid esters. Germany: Frankfurt; June 24, 2008. [56] Matthäus B, Pudel F, Fehling P, Vosmann K, Freudenstein A. Strategies for the reduction of 3-MCPD esters and related compounds in vegetable oils. Eur J Lipid Sci Technol 2011a;113:380e6. [57] Matthäus B, Freudenstein A, Pudel F, Rudolph T. Final results of the German FEI research project concerning 3-MCPD esters and related compoundsdmitigation strategies. In: Presented at 9th Eurofed lipid congress, Rotterdam, The Netherlands, Sept 18e21; 2011. [58] Matthäus B, Freudenstein K, Brühl L, Pudel F, Fehling P, Rudolph T, Granvogl M, Schieberle P, Franke K, Strijowski U. Final report: investigations on the formation of 3-monochloropropane-1,2-diol fatty acid esters (3-MCPD esters) in vegetable oils and development of minimization strategies. Bonn, Germany: Research Association of the German Food Industry; 2011. [59] Nagy K, Sandoz L, Craft BD, Destaillats F. Mass-defect filtering of isotope signatures to reveal the source of chlorinated palm oil contaminants. Food Addit Contam 2011;28:1492e500. [60] Nagy K, Redeuil K, Lahrichi S, Nicolas M. Removal of organochlorines from vegetable oils and its benefits in preventing formation of monochloropropanediol diesters. Food Addit Contam 2019;36:712e21. [61] Ooi CK, Choo YM, Yap CM, Ma AN. Refining of red palm oil. Elaeis 1996;8:20e8. [62] Poku K. Small-scale palm oil processing in Africa; FAO agricultural services bulletin 148. Rome, Italy: Food and Agriculture Organization of the United Nations; 2002. [63] Pudel F, Benecke P, Fehling P, Freudenstein A, Matthäus B, Schwaf A. On the necessity of edible oil refining and possible sources of 3-MCPD and glycidyl esters. Eur J Lipid Sci Technol 2011;113:368e73. [64] Pudel F, Benecke P, Vosmann K, Matthäus B. 3-MCPD- and glycidyl esters can be mitigated in vegetable oils by use of short path distillation. Eur J Lipid Sci Technol 2016;118:396e405. [65] Rahn AKK, Yaylayan VA. Monitoring cyclic acyloxonium ion formation in palmitin systems using infrared spectroscopy and isotope labelling technique. Eur J Lipid Sci Technol 2011;113:330e4. [66] Rahn AKK, Yaylayan VA. What do we know about the molecular mechanism of 3MCPD ester formation? Eur J Lipid Sci Technol 2011;113:323e9. [67] Ramli MR, Siew WL, Ibrahim NA, Hussein R, Kuntom A, Razak RAA, Nesaretnam K. Effects of degumming and bleaching on 3-MCPD esters formation during physical refining. J Am Oil Chem Soc 2011;88:1839e44. [68] Ramli MR, Siew WL, Ibrahim NA, Kuntom A, Abd Razak RA. Other factors to consider in the formation of chloropropandiol fatty esters in oil processes. Food Addit Contam 2015;32:817e24. [69] Razak RAA, Kuntom A, Siew WL, Ibrqahim NA, Ramli RM, Hussein R, Nesaretnam K. Detection and monitoring of 3-Monochloropropane-1,2-diol (3MCPD) esters in cooking oils. Food Chem 2012;25:355e60. [70] Rutherford M, Flood J, Sastroutomo S. Roundtable for sustainable palm oil (RSPO): research project on integrated weed management strategies for oil palm. Final Rep 2011;209:205. [71] Sahid I, Weng CK. Integrated ground cover management in plantations. In: Basiron Y, Jalani BS, Chan KW, editors. Advances in oil palm research. Malaysian palm oil board, vol. I; 2001. p. 623e52. Malaysia. [72] Schurz K. Method for reducing the 3-MCPD content in refined vegetable oils. Patent cooperation treaty application WO10063450. 2010.

Mitigation of MCPD and glycidyl esters in edible oils

63

[73] Shimizu M, Vosmann K, Matthäus B. Generation of 3-monochloro-1,2-propanediol and related materials from tri-, di, and monoolein at deodorization temperature. Eur J Lipid Sci Technol 2012;114:1268e73. [74] Shimizu M, Moriwaki J, Shiiba D, Nohara N, Kudo N, Katsuragi Y. Elimination of glycidyl palmitate in diolein by treatment with activated bleaching earth. J Oleo Sci 2012;61:23e8. [75] Shimizu M, Weitkamp P, Vosmann K, Matthaus B. Influence of chloride and glycidylester on the generation of 3-MCPD- and glycidyl-esters. Eur J Lipid Sci Technol 2013;115:735e9. [76] Siew WL, Ng W-L. Diglyceride content and composition as indicators of palm oil quality. J Sci Food Agric 1995;69:73e9. [77] Silva WC, Santiago JK, Capristo MF, Ferrari RA, Vicente E, Sampaio KA, Arisseto AP. Washing bleached palm oil to reduce monochloropropanediols and glycidyl esters. Food Addit Contam 2019;36:244e53. [78] Smidrkal J, Ilko V, Filip V, Dolezal M, Zelinkova Z, Kyselka J, Hradkova I, Velisek J. Formation of acylglycerol chloro derivatives in vegetable oils and mitigation strategy. Czech J Food Sci 2011;29:449e56. [79] Smidrkal J, Tesarová M, Hrádková I, Bercíková M, Adamcíková A, Filip V. Mechanism of formation of 3-chloropropan-1,2-diol (3-MCPD) esters under conditions of vegetable oil refining. Food Chem 2016;211:124e9. [80] Strijowski U, Heinz V, Franke K. Removal of 3-MCPD esters and related substances after refining by adsorbent material. Eur J Lipid Sci Technol 2011;113:387e92. [81] Tarmizi AM. Nutritional requirements and efficiency of fertilizer use in Malaysian oil palm cultivation. In: Basiron Y, Jalani BS, Chan KW, editors. Advances in oil palm research. Malaysian palm oil board, vol. I; 2001. p. 411e40. Malaysia. [82] Tiong SH, Saparin N, The HF, Ng TLM, bin Md Zain MZ, Neoh BK, Md Noor A, Tan CP, Lai OM, Appleton DR. Natural organochlorines as precursors of 3monochloropropanediol esters in vegetable oils. J Agric Food Chem 2018;66:999e1007. [83] Unnithan UR. Refining of edible oil rich in natural carotenes and vitamin E. U.S. patent 5932261. 1999. [84] Velisek J, Davidek J, Kubelka V, Bartosova J, Tuckova A, Hajslova J, Janicek G. Formation of volatile chlorohydrins from glycerol (triacetin, tributyrin) and hydrochloric acid. Z Lebensm- Wiss Technol 1979;12:234e6. [85] Velisek J, Davidek J, Kubelka V, Janicek G, Svobodova Z, Simicova Z. New chlorinecontaining organic compounds in protein hydrolysates. J Agric Food Chem 1980;28:1142e4. [86] Velisek J, Calta P, Crews C, Hasnip S, Dolezal M. 3-chloropropane-1,2-diol in models simulating processed foods: precursors and agents causing its decomposition. Czech J Food Sci 2003;21:153e61. [87] Wahid MB, Kamarudin N. Insect pests, pollinators and barn owl. In: Basiron Y, Jalani BS, Chan KW, editors. Advances in oil palm research. Malaysian palm oil board, vol. I; 2001. p. 466e541. Malaysia. [88] Wang QY, Ji Z, Han B. Density functional theory study of the mechanism for the formation of glycidyl esters from triglyceride. J Mol Model 2017;23:83. [89] Weisshaar R. Fatty acid esters of 3-MCPD: overview of occurrence and exposure estimates. Eur J Lipid Sci Technol 2011;113:304e8. [90] Weng CK. Soils management for sustainable oil palm cultivation. In: Basiron Y, Jalani BS, Chan KW, editors. Advances in oil palm research, Vol. I. MPOB; 2001. p. 371e410. [91] World Health Organization (WHO). Analysis of contaminated rapeseed oil. European regional office report No. ICP/RCE 903 8920 B. Copenhagen, Denmark: World Health Organization; 1981.

64

Processing Contaminants in Edible Oils

[92] Yan J, Oey SB, van Leeuwen SPJ, van Ruth SM. Discrimination of processing grades of olive oil and other vegetable oils by monochloropropanediol esters and glycidyl esters. Food Chem 2018;248:93e100. [93] Yang T, Zang H, Mu H, Sinclair AJ, Xu X. Diacylglycerols from butterfat: production by glycerolysis and short path distillations and physical properties. J Am Oil Chem Soc 2004;81:979e87. [94] Zelinkova Z, Svejovska B, Velisek J, Dolezal M. Fatty acid esters of 3-chloropropane1,2-diol in edible oils. Food Addit Contam 2006;23:1290e8. [95] Zhang H, Jin P, Zhang M, Cheong L, Hu P, Zhao Y, Yu L, Wang Y, Jiang Y, Xu X. Mitigation of 3-monochloro-1,2-propanediol ester formation by radical scavengers. J Agric Food Chem 2016;64:5887e92. [96] Zhang Z, Gao B, Zhang X, Jiang Y, Xu X, Yu L. formation of 3-monochloro-1,2propanediol (3-MCPD) di- and monoesters from tristearoylglycerol (TSG) and the potential catalytic effect of Fe2þ and Fe3þ. J Agric Food Chem 2015;63:1839e48. [97] Zieverink MMP, Berg I. Oil processing development. In: Presented at 8th EuroFed lipid congress, Munich, Germany, Nov 21e24; 2010. [98] Zulkurnain M, Lai OM, Latip RA, Nehdi IA, Ling TC, Tan CP. The effects of physical refining on the formation of 3-monochloropropane-1,2-diol esters in relation to palm oil minor components. Food Chem 2012;135:799e805. [99] Zschau W. Bleaching of edible fats and oils. Cooperative work of the German Society for Fat Science (DGF). Eur J Lipid Sci Technol 2001;103:505e51.

CHAPTER 4

Indirect detection techniques Karel Hrncirík

Upfield, Rotterdam, the Netherlands

Introduction Fatty acid esters of 2- and 3-monochloropropanediol (MCPD) and glycidol are process contaminants occurring in refined vegetable oils and oil-based food products. Their widespread occurrence in such foodstuffs has been a cause of concern for food safety authorities. The European Food Safety Agency (EFSA) published their scientific opinion regarding these contaminants in 2016 [1]. Soon after, the European Commission established regulatory limits for glycidyl esters [2] and 3-MCPD esters [3] in March 2018 and January 2021, respectively. That implies the need for ongoing monitoring of these contaminants in edible oils and fats on the European market. Retrospectively, it is fair to mention that early research around MCPD esters and glycidyl esters was hampered by considerable knowledge gaps, particularly with regard to the availability of accurate analytical methodologies [4]. Reliable quantification of these contaminants in oils and fats has been essential for mapping their occurrence, performing exposure assessments, studying the mechanisms of formation, developing mitigation technologies, and since recently, monitoring their levels and ensuring regulatory compliance. Intensive work on the development of an analytical methodology began in 2008. The early methods were based on various chemical principles differing in scope and performance. The discovery of glycidyl esters one year later complicated the issue further because MCPD and glycidol were shown to be easily interconvertible under the analytical conditions of some methods. With an increasing number of new methods appearing in the public domain the situation became confusing, and it was soon obvious that harmonization of analytical methodologies was urgently needed. MCPD esters and glycidyl esters include a large number of compounds (more than 100 individual species occur in the most common oils and fats) that vary in type and position of the esterified fatty acids. Analytical Copyright © 2022 AOCS Press. Processing Contaminants in Edible Oils Published by Elsevier Inc. All rights reserved. ISBN 978-0-12-820067-4 Published in cooperation with https://doi.org/10.1016/B978-0-12-820067-4.00009-7 American Oil Chemists Society.

65

66

Processing Contaminants in Edible Oils

methods used for the quantification of MCPD esters and glycidyl esters can be divided into two main categories: direct methods (see Chapters 5 and 6) and indirect methods (featured in this chapter). Indirect methods of analysis are based on the conversion of MCPD esters and glycidyl esters into the corresponding free forms: 2-MCPD, 3-MCPD, and glycidol. These three compounds are further isolated, derivatized, chromatographically separated, and quantified. The results are then expressed as the amount of 2-MCPD, 3-MCPD, or glycidol that can be released from their esterified forms (for this reason the terms bound 2-MCPD, bound 3-MCPD, and bound glycidol are also commonly used). This chapter chronicles the development of methods for the analysis of MCPD esters and glycidyl esters within the past decade, classifies indirect analytical methods, describes chemical principles of different analytical approaches and provides an overview of state-of-the-art analytical methods, which have been adopted as official standards. Major features of each presented method, including their performance and limitations, are discussed in detail. The last part of the chapter provides an overview of the most important collaborative studies and ring-tests, and touches on method automation and other trends in the analysis of MCPD and glycidyl esters.

Main steps in the analysis of 2- and 3-MCPD esters An interest in developing analytical methodologies for the determination of MCPD esters was sparked by findings of the working group of Prof. Velísek, who reported the presence of 3-MCPD esters in fried food [5] and refined oils [6]. Over the following decade, several indirect methods for the analysis of MCPD esters, predominantly 3-MCPD esters, were developed and applied. The individual analytical protocols may vary, but nevertheless, they feature a common sequence of steps: addition of an internal standard to the sample, chemical or enzymatic cleavage of MCPD esters, neutralization of the reaction mixture and salting out with various salts, derivatization of the cleaved 2- and 3-MCPD, and as a final step, chromatographic separation and quantification by GCeMS (Fig. 4.1). Several studies demonstrated that even minor variations of the individual steps in the analytical protocol may significantly affect the performance of the overall analysis, implying that each of these steps needs to be meticulously designed and optimized. In the next section the entire analytical protocol of indirect determination is discussed in detail.

Indirect detection techniques

67

Figure 4.1 Sequence of steps applied to the indirect determination of MCPD esters.

Cleavage of the MCPD esters The initial step of indirect analysis of bound MCPD in oils and fats is the cleavage of the fatty acids from MCPD esters by either chemical transesterification in the presence of methanol (methanolysis) or by enzymatic hydrolysis. In this manner, free 2-MCPD and 3-MCPD are released from their parent esters. Concurrently, triacylglycerols and partial acylglycerols are converted either into fatty acid methyl esters (FAMEs) and glycerol (methanolysis) or to free fatty acids and glycerol (enzymic hydrolysis), as depicted in Fig. 4.2. The transesterification reaction can be catalyzed either by acid or alkali, as first proposed by Divinová et al. [7] and Weißhaar [8]; respectively. Both of these basic approaches became the subject of further assessment and improvements, resulting in development of modern analytical methods for reliable determination of 2-MCPD esters, 3-MCPD esters, and glycidyl esters.

68

Processing Contaminants in Edible Oils

Figure 4.2 Main chemical reactions occurring during the transesterification step.

The advantage of the original alkali-catalyzed transesterification [8] is its short duration (only up to 15 min). Unfortunately, 3-MCPD (released from its fatty acid esters) is unstable in alkaline media and it is quickly dechlorinated and converted to glycidol. This critical phenomenon was independently observed by several authors [9e12], who reported rapid degradation of 3-MCPD under conditions of alkali-catalyzed transesterification over time (Fig. 4.3). The situation is further complicated by the presence of 2-MCPD, which is converted to glycidol in a similar manner (Fig. 4.2) but was reported to have substantially higher stability in comparison to 3-MCPD [11e13]. Since 3-MCPD undergoes a faster conversion, the ratio between 2- and 3-MCPD is gradually changing during alkaline transesterification in favor of 2-MCPD. Depending on the transesterification conditions applied (namely temperature, time, sodium methoxide concentration), this phenomenon can negatively affect the reliability of the 2-MCPD quantification, as was the case in the very first method based on alkaline transesterification [14]. These side reactions require close attention and several strategies of how to deal with these reactions have been proposed, as described later. Next, the degradation of the primary analytes (2-MCPD and 3-MCPD) and their reduced recovery in alkali-catalyzed transesterification inherently affects the sensitivity of methods based on this principle. In a systematic study focused on this subject Hrncirík et al. [9] observed that the transesterification time had an impact on the trueness of the method as well: the levels of bound 3-MCPD found at short transesterification times (1e2 min) were 10%e20% higher than those obtained at longer times (5e10 min). This phenomenon was attributed to the difference among conversion

Indirect detection techniques

69

100 R² = 0.9379 90

Recovery of 3-MCPD [%]

80 70 60 50 40 30 20 10 0 0

2

4

6

8

10

Transesterifica on me (min)

Figure 4.3 Degradation of 3-MCPD during alkali-catalyzed transesterification. (Data originate from studies of Hrncir ík K, Zelinková Z, Ermacora A. Critical factors of indirect determination of 3-chloropropane-1,2-diol esters. Eur J Lipid Sci Technol 2011;113:361e367. (diamonds) Kuhlmann J. Determination of bound 2,3-epoxy-1propanol (glycidol) and bound monochloropropanediol (MCPD) in refined oils. Eur J Lipid Sci Technol 2011;113:335e344; (squares) Sato H, Kaze N, Yamamoto H, Watanabe Y. 2-monochloro-1,3-propanediol (2-MCPD) dynamics in DGF standard methods and quantification of 2-MCPD. J Am Oil Chem Soc 2013;90:1121e1130; (triangles) and Zwagerman R, Overman P. A novel method for the automatic sample preparation and analysis of 3-MCPD-, 2-MCPD-, and glycidylesters in edible oils and fats. Eur J Lipid Sci Technol 2016;118:997-1006 (circle)).

rates for various forms of 3-MCPD (e.g., monoesters being converted to 3-MCPD faster than diesters). To increase the reproducibility of this approach by eliminating the impact of different conversion rates, the authors suggested that a pentadeuterated diester of 3-MCPD should be used as an internal standard in indirect methods, as the majority of the native 3-MCPD esters occurs in the form of diesters. The application of milder conditions for alkaline transesterification is another strategy to deal with the undesirable degradation of 3-MCPD. Kuhlmann [10] applied the combination of a diluted solution of transesterification reagent (sodium hydroxide) and reduced temperature (below
22 C) and achieved a very high recovery of 3-MCPD at the expense of a prolonged transesterification time (16 h).

70

Processing Contaminants in Edible Oils

More recently Zwagerman and Overman [12,15], experimented with conditions of alkali-catalyzed transesterification in detail. They reported the conversion of 3-MCPD and 2-MCPD to glycidol under the optimized conditions (reaction time 12 min at 10 C) to be approx. 7% and 0.4%, respectively. In acid-catalyzed transesterification, no degradation of 2-MCPD or 3-MCPD [9] was observed, which is a favorable prerequisite for the simultaneous analysis of both isomers. Acid media, however, also facilitate the occurrence of nucleophilic substitution reactions, which pose a risk of ex-novo formation of MCPD in the presence of chlorides [8,16]. This, however, does not pose a problem for the analysis of bound MCPD, as the amount of inorganic chlorides occurring in oils and fats is insignificant [17]. In addition, inorganic chlorides can be easily eliminated from the sample by a simple liquideliquid extraction [16,18]. Another option for the cleavage of MCPD from its fatty acid esters is enzymatic hydrolysis. Hamlet and Sadd [19] optimized a protocol for the determination of 2- and 3-MCPD esters in cereal products, based on the enzymatic hydrolysis of the esters with a lipase from Aspergillus oryzae. The method showed good sensitivity and precision, but it did not find large application, probably due to the long incubation time required for the hydrolysis (24 h). Chung and Chan [20] proposed an alternative method for the analysis of 2- and 3-MCPD esters based on the enzymatic hydrolysis of the esters with Candida antarctica lipase A and shortened the incubation time to 16 h. A further substantial reduction of the hydrolysis time to just 30 min was achieved by applying a lipase from Candida cylindracea (previously referred to as Candida rugosa) [21]. In this method the internal standards were proposed to be added only after the enzymatic reaction, which means that a complete ester cleavage is critical for the performance of this method. There are several challenges, very different from the chemical transesterification, that must be overcome in order to accomplish the full ester cleavage. First, while the lipophilic esters are dissolved in an organic solvent they must be processed by the lipase in an aqueous solution, which requires an intensive shaking (under reproducible conditions) to maximize the contact between the lipase and the substrate, both dissolved in immiscible phases [22]. Second, the enzymatic reaction takes place at room temperature, which poses the risk of incomplete dissolution (or recrystallization) of fats with a higher melting point [23]. Third, some reagents, e.g., sodium bromide salt, used for the conversion of glycidol to 3-MBPD (described later in the

Indirect detection techniques

71

Conversion of glycidol to 3-MBPD after alkaline or enzymatic ester cleavage section), may exert an inhibitory effect on the enzyme activity [24]. The fourth challenge is related to the substrate specificity (with respect to both the position and the length of the esterified fatty acid). In fact Miyazaki and Koyama [24] demonstrated that C. cylindracea lipase, previously established in the official protocol for the analysis of MCPD and glycidyl esters [25], exhibited only a very low substrate specificity for esters of long-chain polyunsaturated fatty acids (EPA, DHA) and was therefore not suitable for the analysis of fish oil. The same authors proposed the use of a lipase from Burkholderia cepacia, alongside a modification of the analytical protocol, to enable the analysis of MCPD and glycidyl esters in these matrices, which demonstrates the critical aspect of the substrate specificity in the analysis of a very wide variety of individual species (esters). Neutralization and salting out In the next step, the transesterification reaction is stopped by the addition of a neutralizing agent. Sodium hydrogen carbonate is used in the case of acidcatalyzed transesterification, while various acids are used to stop alkalicatalyzed transesterification. After the transesterification reaction is stopped, the next step is a salt out. The purpose of the salt out is to purify the sample by facilitating the extraction of lipophilic compounds formed during transesterification, mainly fatty acid methyl esters (FAMEs), from the aqueous phase containing the water-soluble 2- and 3-MCPD. Several common salts have been proposed in various procedures for the analysis of 3-MCPD, namely sodium chloride, sulfate salts, and sodium bromide [7e10]. These salts are added to the reaction mixture either after the neutralization or together with the neutralizing agent. It was clearly demonstrated that the salt out is critical with respect to the specificity (and subsequently the trueness) of methods employing alkalicatalyzed transesterification. Shortly after the publication of the method based on alkaline transesterification and salting out by sodium chloride [8], it became clear that the level of bound 3-MCPD measured by this method was substantially overestimated in some samples. This overestimation was caused by the ex-novo formation of 3-MCPD during sample preparation. The precursor of the additional 3-MCPD formed was confirmed to be bound glycidol, which can be present in refined vegetable oils at levels comparable with bound MCPD [26,27] and liberates free glycidol during alkaline transesterification. During neutralization and salting out, glycidol

72

Processing Contaminants in Edible Oils

subsequently reacts with chlorides (when sodium chloride is used as a salting out reagent) to form 3-MCPD and, to a lesser extent, 2-MCPD. The reaction between glycidol and sodium chloride became the basis of the first method for the indirect determination of bound glycidol in oils and fats, adopted by the German Society for Fat Science (DGF) in 2009 [14]. The reactivity of glycidol (or glycidyl esters) with halides is now utilized by several established procedures for the determination of bound glycidol, as described in detail later. In the early methods based on alkali transesterification, the co-present glycidyl esters in the oil sample interfered with the correct determination of bound MCPD, as they could be easily converted into 2- and 3-MCPD during sample preparation. In order to avoid glycidol conversion, two alternative strategies can be applied: either the elimination of glycidyl esters prior to transesterification, or the substitution of sodium chloride by other salts. Concerning the former option, acid pretreatment of the sample was implemented into the official DGF method [14], but the elimination of glycidyl esters was proven to be incomplete and the method was withdrawn. The usage of a non-chloride salt, such as sodium bromide [28], proved to be more effective, although this approach does not avoid the complex chemistry and loss of 2- and 3-MCPD (affecting both the genuine analytes and their pentadeuterated analogues used as internal standards) during the alkali-catalyzed transesterification performed at room temperature [29]. Methods based on acid-catalyzed transesterification are not affected by the choice of salt or by the conditions applied during salting out. Glycidyl esters are irreversibly degraded during the acid-catalyzed transesterification (i.e., prior to salting out), and hence do not interfere with the determination of MCPD esters [9]. Similarly, the enzymatic methods do not seem to suffer from the presence of glycidyl esters, as the released glycidol reacts with bromide salt present without any apparent side-reactions [23]. Derivatization and GCeMS analysis Gas chromatographic (GC) analysis poses several major advantages: high speed, good resolution, high sensitivity, and the usage of a simple/affordable GCeMS interface. However, the application range of GC is restricted by the characteristics of the analyzed compounds, particularly their molecular weight, polarity, and thermal stability.

Indirect detection techniques

73

Both 2- and 3-MCPD possess relatively low volatility (b.p. 248.5 C and 213 C, respectively) and high polarity (readily soluble in water and ethanol). These physicalechemical properties complicate the direct analysis by GC, as 3-MCPD gives rise to undesirable interactions with components of the GC systems, resulting in poor peak shape and low sensitivity. Next to that, the low molecular weight of 3-MCPD (110.5 g/mol) makes mass detection difficult because diagnostic ions cannot be reliably distinguished from background noise [30]. These limitations have been overcome by converting 3-MCPD into volatile derivatives prior to GC analysis. The GC methodology for the analysis of free 3-MCPD has been evolving since the 1980s, and several derivatization techniques have been developed since then [30]. By far the most common derivatization agent is phenylboronic acid (PBA), followed by N-(heptafluorobutyryl)imidazole (HFBI) (Fig. 4.4). Both derivatization agents represent good alternatives, although HFBI is sensitive to moisture and the derivatization reaction must be carried out under strict anhydrous conditions. The formation of a stable volatile derivative of 3-MCPD, either dioxaborolane derivative (MW 196 g/mol) or heptafluorobutyryl ester derivative (MW 502 g/mol), prior to GC analysis greatly improves the sensitivity of the analysis. In addition, the method specificity is significantly enhanced by greater possibilities for monitoring specific fragment ions of the 3-MCPD derivatives. Importantly, both of these derivatization reagents can also react with 2-MCPD to form corresponding derivatives (Fig. 4.4) that can be easily separated from the 3-MCPD analogs by GC (Fig. 4.5). PBA is nowadays the more common derivatization reagent that is used in all standard methods currently available (see Section Officially adopted methods). The 3-MCPD dioxaborolane derivative showed high stability during storage at different temperatures (ambient and e20 C) over the period of two months [16]. The derivatization with PBA can be performed in two manners. The first option is to carry out the reaction of the analytes with PBA in an aqueous solution, typically in a saturated solution of PBA in acetone/water (19/1, v/v) that is added in variable amounts to the sample mixture [7,8,16]. The derivatization reaction takes place under ultrasonic treatment at room temperature. The other option is to carry out the derivatization in an organic solvent (e.g. diethyl ether). This requires the extraction of the

74

Processing Contaminants in Edible Oils

Figure 4.4 Structure of 2-MCPD and 3-MCPD derivatives.

analytes from an aqueous solution with organic solvents prior to the addition of the derivatization agent in order to reduce the solubility of (unreacted) PBA in the derivatized sample [10,28]. The negative impact of an excess of PBA on the performance of the GC system was reported by Weißhaar [8], who suggested the use of a PTV injector or a back-flush unit [31] to overcome this issue. Another option is to dissolve the sample ready for GC injection in a non-polar solvent (n-heptane); the application of such an additional step of sample concentration is aimed to both facilitate the precipitation of the excess of PBA and improve the sensitivity [32].

Indirect detection techniques

75

Figure 4.5 Selected ion chromatograms (m/z 196 and 201) of a standard solution of 1,2-dipalmitoyl-3-MCPD and 1,3-dipalmitoyl- 2-MCPD subjected to the analytical procedure (bromination, transesterification and derivatization) and analyzed by GCeMS.

Main approaches toward the analysis of glycidyl esters The occurrence of glycidyl esters in refined oils and fats was confirmed when Weißhaar and Perz [27] identified glycidyl palmitate and glycidyl stearate in various samples of palm-based fats. As discussed in the previous section, this breakthrough originated from the observation that the level of 3-MCPD esters measured in samples of refined oils and fats was dependent on the analytical protocol applied. In particular, the application of protocols based on alkaline transesterification and salting out by sodium chloride resulted in the detection of consistently higher levels of bound 3-MCPD. During alkaline transesterification, the fatty acid esters of glycidol are cleaved to generate free glycidol, which is stable under these conditions and accumulates throughout the reaction. During the next step of neutralization and salting out, glycidol can react with the salts used as salting out reagents,

76

Processing Contaminants in Edible Oils

Figure 4.6 Pathways of chemical reactions used for designing strategies for the analysis of glycidyl esters.

undergoing nucleophilic addition to the epoxide ring. When sodium chloride is used as salting out reagent, 2- and 3-MCPD are formed, causing the overestimation of the true level of these compounds in the sample. The properties and reactivity of the epoxide function of glycidol and glycidyl esters were later advantageously employed for the development of analytical methods for their determination in samples of oils and fats. In four separate sections below, major approaches proposed for the analysis of glycidyl esters are discussed individually. These are based on either chemical (alkaline or acid catalyzed) transesterification or enzymatic hydrolysis, and all involve a conversion step of glycidyl esters to a more stable halogenated derivative during sample preparation, as depicted in Fig. 4.6. In each case, the method was developed starting from a protocol previously optimized for the determination of 3-MCPD esters; therefore, all methods included the determination of both classes of compounds. This is particularly advantageous considering that 2- and 3-MCPD esters and glycidyl esters are formed under similar conditions of oil refining and typically co-occur in refined vegetable oils and fats. Elimination of glycidyl esters by acid treatment The first strategy proposed for the indirect determination of glycidyl esters in oils and fats involved a differential analysis. According to this approach, the oil sample is divided into two aliquots, one of which is used for the

Indirect detection techniques

77

determination of the sum of 3-MCPD esters and glycidyl esters (part A), while the other is analyzed to determine solely the content of 3-MCPD esters (part B) [27]. The amount of glycidyl esters (called “3-MCPD related substances” in the method scope) in the oil sample is calculated from the difference between the two determinations: c ¼ ðA
½minus signBÞ  ½multiplication signt where c is the concentration of glycidyl esters (expressed as mg of bound glycidol per kg of oil), A is the concentration of 3-MCPD esters plus glycidyl esters (expressed as mg of bound 3-MCPD per kg of oil) determined in part A, B is the concentration of 3-MCPD esters determined in part B, t is a transformation factor, which takes into account a stoichiometric conversion between 3-MCPD and glycidol, and the experimentally obtained conversion of glycidol to 3-MCPD. The analytical protocol used in part A involves a first step of alkaline transesterification in which both 3-MCPD esters and glycidyl esters are converted into the corresponding free forms, followed by neutralization of the reaction mixture and salting out using sodium chloride as salting out reagent. As a result, glycidol is also converted into 3-MCPD during this step. In part B, the same analytical procedure is preceded by a step aimed to eliminate glycidyl esters, thus allowing the quantification of 3-MCPD esters only. The elimination of glycidyl esters is achieved by reaction in acid alcoholic medium (propanol/sulfuric acid 100/0.5, v/v). During this step, glycidyl esters are predominantly converted into 2- and 3-propyl ether derivatives, which do not interfere with the following steps of the analysis, while 3-MCPD esters remain unchanged. The trueness of methods based on this approach relies on the assumption that complete elimination of glycidyl esters is achieved during the first step of sample preparation (part B). In fact, an incomplete reaction would result in the overestimation of the levels of 3-MCPD esters and the complementary underestimation of the levels of glycidyl esters, due to the differential analysis applied. Several research groups questioned and tested this hypothesis by performing comparative studies with either direct [29,33,34] or indirect methods based on a different approach [35] to verify the completeness of the reaction. All research groups concluded that complete elimination of glycidyl esters was not accomplished during this step of acid treatment. Shimizu et al. [34] observed a dependency of the residual level of glycidol/glycidyl esters after the treatment on their initial concentration in the oil sample. Similar results were also reported by Kaze

78

Processing Contaminants in Edible Oils

et al. [33], who estimated that approximately 10% of the glycidyl esters initially present in the oil sample are converted into 3-MCPD during the next steps of sample preparation. Prolongation of the time applied during the acid treatment (from 15 to 30 min) resulted in just a limited increase of the conversion yield of glycidyl esters into propyl ether derivatives [34]. Further extension of the reaction time did not affect the results since under the conditions specified by the method [27] the reaction reaches equilibrium within 30 min [34]. Although the protocol proposed by Weißhaar and Perz [27] was established by the German Society for Fat Science (DGF) as the official DGF method C-III 18 (09), it was withdrawn in 2011 due to the poor accuracy resulting from drawbacks captured in the text above. Conversion of glycidol to 3-MCPD after alkaline transesterification In an attempt to solve the issues related to the method accuracy, a second approach based on alkaline transesterification and a differential analysis was proposed. Analogous to the previous case, the oil sample is divided into two aliquots, which are analyzed separately to quantify either the sum of 3-MCPD esters and glycidyl esters (part A), or 3-MCPD esters only (part B). The difference between the two determinations is used to calculate the level of glycidyl esters in the oil sample. Because the main reason for inaccurate results in methods using the elimination approach is attributed to the quantification of 3-MCPD esters in part B, this assay was modified, while part A remained practically unchanged. The improved protocol for part B involves a first step of alkaline transesterification (carried out under the same conditions applied in part A), followed by neutralization of the sample mixture and salting out by a chloride-free salt. As a result, glycidol, which is formed during alkaline transesterification, is converted into a derivative different than 2- or 3-MCPD (depending on the salts used, bromide or sulfate derivatives can be formed) during the step of salting out. The newly formed derivative is stable during sample preparation and does not interfere with the 3-MCPD analysis, thus avoiding any negative impact on its quantification. Due to the differential approach applied, an accurate determination of 3-MCPD esters is necessary for an accurate quantification of glycidyl esters as well. In addition, the complete conversion of glycidyl esters into 3-MCPD must be achieved in assay A, since an incomplete reaction would result in the underestimation of their level in the oil sample. Because only 3-MCPD is

Indirect detection techniques

79

quantified by the assay, it is also important that the nucleophilic addition to the epoxide ring is specific to position sn-3. A nucleophilic addition to position sn-2 would generate 2-MCPD, which is not quantified by this procedure. The conversion reaction of glycidol into 3-MCPD has been studied by different research groups that found a clear dependency of the conversion yield on the pH of the reaction medium [9,33]. Glycidol conversion was found to be limited (up to 54%) in neutral media [9], but considerably higher (100% and up to 94%) at pH 1.9 and 2, respectively [9,33]. Hrncirík et al. [9] also observed that the pH of the neutralized mixture varies substantially among the different methods based on alkaline transesterification (1.5e7.0, depending on the protocol applied), thus, it is expected that the accuracy of the results obtained by such methods will vary accordingly. Sato et al. [11], who studied regioselectivity of the glycidol conversion reaction in samples spiked with glycidyl esters, found that approximately 95% of the spiked compound was converted to 3-MCPD during sample preparation, while the remaining 5% generated 2-MCPD. The precision of these differential methods is inherently affected by the fact that two independent analyses are required for the determination of bound glycidol; therefore, the error is expected to be higher than in a single analysis. In their infancy, analytical protocols based on this approach were suspected to be less reliable because of the possible formation of artifacts during sample preparation. It was hypothesized that substances other than glycidyl esters could generate 3-MCPD given the conditions applied. Nevertheless, the reaction conditions were optimized and following the positive outcome of interlaboratory comparison studies the method was adopted as the DGF method C-VI 18 (10), and later as AOCS Official Method Cd 29c-13 [36] and ISO standard 18363-1:2015 [37] (see Section Officially adopted methods). Conversion of glycidol to 3-MBPD after alkaline or enzymatic ester cleavage During alkaline transesterification glycidol is generated by two pathways: first, by the cleavage of the fatty acids of glycidyl esters, and second, by partial degradation of 2- and 3-MCPD released from the corresponding esters. That means that a reliable quantification of both MCPD and glycidol requires a differential approach based on two analyses. In order to mitigate this issue Kuhlmann [10] proposed an alternative protocol based on alkaline transesterification under mild conditions that represents a third strategy to quantify bound glycidol. The basic principle is

80

Processing Contaminants in Edible Oils

that upon reduced concentration of sodium hydroxide (approximately tenfold lower when compared to the DGF method C-VI 18 (10)) and reduced temperature (below
22 C), the degradation of 2- and 3-MCPD is considerably suppressed. Therefore, the formation of glycidol during this step can be almost entirely attributed to the cleavage of glycidyl esters. During the next step of neutralization and salting out with sodium bromide, glycidol is converted to 3-monobromopropandiol (3-MBPD), which is then quantified. The structural similarity of 3-MBPD and 3-MCPD allows the simultaneous analysis of the two compounds (see Fig. 4.7). The level of glycidyl esters in the oil sample is calculated based on the 3-MBPD/3-MBPD-d5 ratio and by applying a correction factor aimed to compensate for the partial degradation of 2- and 3-MCPD to glycidol during alkaline transesterification. In this manner the simultaneous analysis of MCPD esters and glycidyl esters is achieved, although two separate analyses of each sample are still needed to calculate the correction factor, which is matrix dependent. In addition, the conversion of glycidol into a derivative other than 3-MCPD allows for a more accurate quantification of glycidyl esters. In fact, the use of a deuterated glycidyl ester as an internal standard compensates for the regioselectivity of the conversion reaction (deuterated and nondeuterated

Figure 4.7 Selected ion chromatograms (m/z 147 and 150) of a standard solution of 1,2-dipalmitoyl-3-MCPD, 1,3-dipalmitoyl- 2-MCPD and glycidyl palmitate, subjected to the analytical procedure (bromination, transesterification and derivatization) and analyzed by GCeMS.

Indirect detection techniques

81

compounds are expected to have almost identical reactivity), thus leaving the accuracy of the method unaffected. On the other hand, the regioselectivity of the conversion reaction has an impact on the sensitivity, since only the sn-3 isomer (i.e., 3-MBPD) is quantified. The 3-MBPD/2-MBPD signal ratio was calculated to range between 5 and 7, depending on the conditions applied, indicating a sufficient specificity for the sn-3 position [10]. The application of mild conditions during transesterification, which is key for an effective suppression of the formation of glycidol from 2- and 3-MCPD, also increases the complexity of the analytical protocol. The need for a constant very low temperature (
22 C/
25 C) increases the chances of errors during the analysis and slows down the transesterification reaction, which reaches completion after 16e18 h. This principle became the basis for the SGS method, which was later adopted as AOCS Official Method Cd 29b-13 [38] and ISO standard 18363-2:2018 [39] (see Section Officially adopted methods). An alternative approach to tackle the issue of undesirable conversion of 3-MCPD in alkaline environment was presented by Zwagerman and Overman [12,15]. Similar to the DGF and SGS methods, they applied alkaline transesterification (followed by bromination), but chose to reduce the reaction temperature to 5 C. That resulted in the reduced conversion of 3-MCPD to glycidol from 27% room temperature to 2.5% (5 C) during the reaction time 5.5 min [12]. By further optimization of the reaction conditions a compromise between the reaction time and the undesirable conversion of 2- and 3-MCPD to glycidol was found. Using the optimized method conditions (10 C, 12 min), they reported that 0.4% and 7% of 2-MCPD and 3-MCPD, respectively, was converted to glycidol [15]. Although low, the conversion of 3-MCPD to glycidol is not negligible, therefore, 3-MCPD-13C3 is added as an internal standard. The conversion of 3-MCPD-13C3 (yielding glycidol-13C) reflects the conversion of genuinely present 3-MCPD, and the amount of glycidol-13C detected after stopping the reaction enables quantification of the overall conversion of 3-MCPD, thereby correcting the overestimation of glycidyl esters. To accomplish the transesterification step, the reaction time was prolonged from 5.5 min (DGF method) to 12 min, which is considerably shorter than the transesterification reaction time (16 h) required in the SGS method. The method, denoted as the BLC (Bunge Loders Croklaan) method was recently adopted as ISO standard 18363-4:2021 [40] and is currently undergoing validation by AOCS (see Section Officially adopted methods).

82

Processing Contaminants in Edible Oils

An alternative to chemical ester cleavage (transesterification) is an enzymatic hydrolysis (see Section Cleavage of the MCPD esters). This approach is relatively straightforward in terms of chemical conversions: glycidol is released from its parent esters after enzymatic hydrolysis and subsequently converted to 3-MBPD by reaction with bromide under mild acidic conditions (80 C, 10 min) in a similar fashion as in the SGS and BLC methods [22]. More details about the so-called JOCS enzymatic method, which was adopted as AOCS Cd 29d-19 method [25], are provided in the Officially adopted methods section. Conversion of glycidyl esters to 3-MBPD esters prior to acid transesterification Different from the previous strategies, the approach based on acid transesterification is based on the reactivity of intact glycidyl esters. As discussed earlier in this chapter, during acid transesterification glycidyl esters are irreversibly converted into propyl- (or methyl-, when methanol is used) ether derivatives due to the opening of the epoxide ring in acid alcoholic media. In order to prevent the irreversible breakdown of glycidyl esters, an additional step converting glycidyl esters into a more stable derivative prior to transesterification was proposed by Ermacora and Hrncirík [41]. The conversion reaction is carried out under mild acid conditions and in the presence of a strong nucleophile (bromide), which undergoes nucleophilic addition to the epoxide ring of glycidyl esters, forming 3-MBPD monoesters. The newly formed derivatives have very similar structural and physical-chemical properties to 3-MCPD esters; thus, they can be advantageously analyzed with the same analytical protocol. Carrying out the conversion reaction on the esterified (glycidyl esters) rather than the free (glycidol) form results in the higher regioselectivity of the reaction due to the effect of steric hindrance exploited by the esterified fatty acid chain. The 3-MBPD/2-MBPD signal ratio of 13e15 [41] was approximately twice as high as reported for the conversion in alkaline environment (3-MBPD/2-MBPD signal ratio of 5e7) [10], increasing the sensitivity of this method. The accuracy of methods based on this approach is strongly dependent on the conditions applied during the conversion reaction of glycidyl esters. From an extensive study on the underlying chemistry of the method [41] it was discovered that the reaction time, temperature, strength of the acid solution, and concentration of bromide ions in the reaction mixture are the main factors affecting both the conversion yield and the potential occurrence of undesired side reactions.

Indirect detection techniques

83

In fact, the presence of bromide ions (strong nucleophiles) in acid media could induce the ex-novo formation of 3-MBPD esters by nucleophilic substitution of bromide ions on partial acylglycerols, which are naturally present in oils and fats. This undesired reaction, however, occurs less readily than the nucleophilic addition to the epoxide ring of glycidyl esters and, thus, it can be easily suppressed by careful tuning of the reaction conditions. In theory, the concentration of partial acylglycerols in the oil sample may also affect the measured level of 3-MBPD. Nevertheless, under the optimized conditions, this side reaction was found to have no significant impact on the trueness of the results, even for oil samples with a content of partial acylglycerols up to 25% (w/w), which is unrealistically high for any type of oil or fat [41]. Further, it was observed that during the next step (acid transesterification) of the analytical procedure no interconversion between 3-MBPD and 2-/3-MCPD occurs [32], which confirms the high stability of 3-MBPD esters under the conditions of the method. In a similar manner to the previous approaches in which glycidyl esters were ultimately converted into 3-MBPD upon sample preparation, a simultaneous GC analysis of all three classes of compounds can be performed with this method due to their structural similarity and good chromatographic resolution. A further advantage of this approach is that it requires only a single sample preparation and GC run for the simultaneous quantification of bound 2-MCPD, 3-MCPD, and glycidol. This is due to the fact that upon acid transesterification (similar to enzymatic cleavage) there is no risk of 2- and 3-MCPD conversion to glycidol, hence there is no need for an additional analysis to correct for the occurrence of such a reaction. This procedure is generally referred to as the Unilever method and following an official interlaboratory evaluation it was adopted as AOCS Official Method Cd 29a-13 [42] and ISO standard 18363-3:2017 [43] (see Section Officially adopted methods).

Officially adopted methods Many attempts to develop the ultimate analytical procedure have been published in the past 15 years. Researchers have utilized various combinations of the analytical steps previously described in this chapter and experimented with reaction conditions. These attempts resulted in the development of a number of different methods, which diverge with regard to the compounds analyzed (typically 2- and/or 3-MCPD esters, or both

84

Processing Contaminants in Edible Oils

MCPD esters and glycidyl esters) and the main method parameters, such as accuracy, selectivity, precision, sample throughput, etc. These have been reviewed and listed by Ermacora and Hrncirík [44]. Significant efforts were spent in the last decade toward the harmonization of such analytical protocols, with the intent of identifying the most reliable procedures. Three indirect methods (DGF, SGS and Unilever) widely used for the routine analysis of MCPD and glycidyl esters were largely accepted and adopted by AOCS and ISO as official methods. A few years later, two other methods (JOCS and BLC) were added to the list of the official methods. In principle, they are all based on either acid transesterification, alkaline transesterification (under standard or mild conditions) or enzymatic hydrolysis, followed by derivatization of the analytes with phenylboronic acid, and GCeMS analysis. The five officially adopted indirect methods are described further in more details and their main features are summarized in Table 4.1. Unilever method This method [41] is applicable for the determination of 2- and 3-MCPD esters and glycidyl esters in edible oils and fats. It is certified as AOCS Official Method Cd 29a-13 [42] and ISO standard 18363-3:2017 [43]. Procedure: The oil sample is dissolved in tetrahydrofuran and the internal standards (pentadeuterated 3-MCPD diester and pentadeuterated glycidyl ester) are added. During the first step of sample preparation, glycidyl esters are converted into 3-MBPD monoesters by the addition of an acidified solution of sodium bromide (0.1%, w/v). In order to maximize the conversion yield, the reaction is carried out at 50 C for 15 min. Upon completion of the reaction, the organic phase, containing 2- and 3-MCPD esters and 3-MBPD esters, is separated and evaporated to dryness. The residue is dissolved in tetrahydrofuran and the acid transesterification is initiated by the addition of an acid alcoholic solution (1.8%, v/v, sulfuric acid in methanol). After 16 h at 40 C, the sample mixture is neutralized and a sodium sulfate aqueous solution (20%, w/v) is added to facilitate the removal of the fatty acid methyl esters generated during the transesterification. The purified sample is derivatized with phenylboronic acid prior to GCeMS analysis. The quantification of 2- and 3-MCPD esters (expressed as bound 2- and 3-MCPD) is based on the 2-MCPD/3-MCPDd5 and 3-MCPD/3-MCPD-d5 signal ratios, respectively. The quantification of glycidyl esters (expressed as bound glycidol) is based on the 3-MBPD/3-MBPD-d5 signal ratio.

Table 4.1 Overview of indirect methods for the analysis of MCPD esters and/or glycidyl esters officially adopted by AOCS and ISO. Method denomination

Unilever method

SGS method

References Code of the official method

[41] AOCS Cd 29a-13 ISO 18363e3:2017

Compounds analyzed

2- MCPD-E 3-MCPD-E glycidyl-E Conversion of glycidyl-E to 3-MBPD-E, acid-catalyzed ester cleavage, derivatization, GCeMS analysis

[10] AOCS Cd 29b-13 ISO 18363e2:2018 2- MCPD-E 3-MCPD-E glycidyl-E Alkaline-catalyzed ester cleavage, conversion of glycidol to 3-MBPD, derivatization, GCeMS analysis

Method principle

Internal standards

3-MCPD-diE-d5 glycidyl-E-d5

Transesterification/ hydrolysis Salting out

Acid, 40 C, 16 h Na2SO4 (20%, w/v)

2-MCPD-diE-d5 3-MCPD-diE-d5 2-MCPD-d5 3-MCPD-d5 glycidyl-E-d5 Alkaline, -22/25 C,16 h NaBr (60%, w/v)

Comment

Single assay

Two assays

DGF method C-VI 18 (10)

BLC method

JOCS method 2.4.14e2016

[35] AOCS Cd 29c-13 ISO 18363e1:2015

[15] ISO 18363e4:2021

[22] AOCS Cd 29d-19

3-MCPD-E

2- MCPD-E 3-MCPD-E glycidyl-E Alkaline-catalyzedester cleavage, conversion of glycidol to 3-MBPD, derivatization (PBA), GCeMS analysis

2- MCPD-E 3-MCPD-E glycidyl-E Enzyme assisted ester cleavage, conversion of glycidol to 3-MBPD, derivatization, GCeMS analysis

3-MCPD-13C3 glycidol-d5

2-MCPD-d5 3-MCPD-d5 3-MBPD-d5

glycidyl-E Alkaline-catalyzed ester cleavage, conversion of glycidol to 3-MCPD (as. A) or 3-MBPD (as. B), derivatization, GCeMS analysis 3-MCPD-diE-d5

Alkaline, RT, 3.5e5.5 min A: NaCl (20% w/v) B: NaBr (60%, w/v) Two assays; glycidyl-E calculated

List of abbreviations in the table. -E, ester; -diE, diester; PBA, phenylboronic acid; RT, room temperature.

Alkaline, 10 C, 12 min Enzymatic, RT, 30 min NaBr (60%, w/v) NaBr (30%, w/v) Single assay; GCeMS/MS needed

Single assay

86

Processing Contaminants in Edible Oils

The Unilever method is a well-established procedure, it is rugged and resources-efficient as it allows for the simultaneous quantification of all three analytes in a single assay. A disadvantage is a long (overnight) transesterification step, which affects the speed of reporting the results. SGS method This method [10] is applicable for the determination of 2- and 3-MCPD esters and glycidyl esters in edible oils and fats. It is certified as AOCS Official Method Cd 29b-13 [38] and ISO standard 18363-2:2018 [39]. Procedure: Two assays, A and B, are applied for the determination of glycidyl esters and MCPD esters, respectively. Assay A: An aliquot of the oil sample is dissolved in diethyl ether, the internal standards (pentadeuterated free 2- and 3-MCPD and a pentadeuterated glycidyl ester) are added. The alkaline transesterification reaction is carried out at
22 C/
25 C for 16 h with a diluted solution of sodium hydroxide in methanol (0.25%, w/v). The reaction is stopped by the addition of an acid solution of sodium bromide (60%, w/v), which enables the conversion of glycidol into 2- and 3-MBPD. The sample mixture is then purified by liquideliquid extraction with iso-hexane and the analytes are extracted with diethyl ether/ethyl acetate (3/2, v/v), derivatized with phenylboronic acid and injected into a GCeMS system. The quantification of glycidyl esters (expressed as bound glycidol) is based on the 3-MBPD/3MBPD-d5 signal ratio. A correction factor (calculated in assay B) is applied to compensate for the partial conversion of MCPD into glycidol (and subsequently to MBPD) during transesterification and salting out. Assay B: The second aliquot of the oil sample is dissolved in diethyl ether, a different set of internal standards (pentadeuterated 2- and 3-MCPD diester) is added to the sample, and the resulting mixture is treated according to the same protocol used in assay A. The quantification of 2and 3-MCPD esters (expressed as bound 2- and 3-MCPD) is based on the 2-MCPD/2-MCPD-d5 and 3-MCPD/3-MCPD-d5 signal ratios, respectively. The SGS method represents another well-established approach widely used over past years. Similar to the Unilever method, it suffers from a prolonged 16 h transesterification routine. Although the method is not based on a differential analysis (like the DGF method), two separate assays are required to account for the partial conversion of 2- and 3-MCPD to glycidol in alkaline media (which requires an application of the appropriate correction factor) and to monitor the extent of the ester cleavage during transesterification.

Indirect detection techniques

87

DGF method C-VI 18 (10) This method [28] is applicable for the quantification of 3-MCPD esters and glycidyl esters in edible oils and fats. It is certified as AOCS Official Method Cd 29c-13 [36] and ISO standard 18363-1:2015 [37]. Procedure: The method consists of two assays. In assay A, the sum of 3-MCPD esters and glycidyl esters (expressed as bound 3-MCPD) is determined, whereas assay B is used for the quantification of solely 3-MCPD esters. The level of glycidyl esters is calculated from the difference between the two assays by applying a nonstoichiometric factor accounting for the conversion of glycidol to 3-MCPD under the conditions applied during sample preparation. Assay A: An aliquot of the oil sample is dissolved in tert-butyl methyl ether and the internal standard (pentadeuterated 3-MCPD diester) is added. The alkaline transesterification is carried out at room temperature for 3.5e5.5 min by reaction with a methanolic solution of sodium hydroxide (2%, w/v) or sodium methoxide (2.5%, w/v). The reaction mixture is neutralized by the addition of an acidified solution of sodium chloride (20%, w/v) that induces the simultaneous conversion of glycidol into 2- and 3-MCPD. The fatty acid methyl esters generated during alkaline transesterification are separated from the sample mixture by liquideliquid extraction with iso-hexane. Both 2- and 3-MCPD are extracted with a mixture of diethyl ether and ethyl acetate (3/2, v/v) and derivatized with phenylboronic acid prior to GCeMS analysis. The quantification of 3-MCPD esters and glycidyl esters (expressed as bound 3-MCPD) is based on the 3-MCPD/3-MCPD-d5 signal ratio. Assay B: A second aliquot of the oil sample is dissolved in tert-butyl methyl ether, the internal standard (pentadeuterated 3-MCPD diester) is added. Alkaline transesterification is carried out according to the same protocol used in assay A. Upon completion of the reaction, the sample mixture is neutralized by the addition of an acid solution containing chloride-free salts (i.e., sodium bromide) as salting out reagent and glycidol is converted into non-chlorinated derivatives that do not interfere with the determination of 3-MCPD esters. The next steps of sample purification and derivatization of the analytes prior to GCeMS analysis follow the same procedure used in assay A. The DGF method attained its popularity due to a considerably faster procedure (compared to the two aforementioned methods), which makes the method suitable for circumstances when results are required on a short

88

Processing Contaminants in Edible Oils

notice (e.g., in a positive release situation). Nonetheless, the quantification of glycidyl esters by this method is based on a differential analysis performed in two independent assays; such a procedure inherently increases the potential for analytical error and the generated results for glycidyl esters are subject to a higher uncertainty, particularly when present at lower levels. BLC method This method [15] is applicable for the quantification of 2- and 3-MCPD esters and glycidyl esters in edible oils and fats. It is certified as ISO standard 18363-4:2021 [40] and is currently undergoing validation by AOCS. Procedure: An aliquot of the oil sample is dissolved in a toluene/tertbutyl methyl ether mixture and the internal standards (3-MCPD-13C3 and pentadeuterated glycidol) are added. The transesterification is carried out at 10 C for 12 min by reaction with a methanolic solution of sodium methoxide (0.35 M). The reaction is stopped by adding an acidified solution of sodium bromide (60%, w/v), which further reacts (5 min at RT) with glycidol to yield 3-MBPD. Subsequently, fatty acid methyl esters that were formed during the transesterification are eliminated by a double extraction with iso-octane. Finally, the target analytes are derivatized by phenylboronic acid and analyzed by a GCeMS/MS system. The quantification of 2- and 3-MCPD esters (expressed as bound 2- and 3-MCPD) is based on the 2-MCPD/3-MCPD-13C3 and 3-MCPD/3-MCPD-13C3 signal ratios, respectively. The quantification of glycidyl esters (expressed as bound glycidol) is based on the 3-MBPD/3-MBPD-d5 signal ratio, with an additional correction formula to compensate for glycidol formed from 3-MCPD during alkaline transesterification. The BLC method was presented relatively recently. It is derived from the DGF Method C-VI 18 (10) [28]and benefits from a short time required for sample preparation. Opposite to the DGF method, only a single analysis is required by application of the 13C-labeled internal standard, which corrects the overestimation of glycidyl esters caused by undue conversion of 3-MCPD to glycidol during alkaline transesterification. In theory, this should have a positive impact on the method precision. JOCS method 2.4.14e2016 This method [22] is applicable for the determination of 2- and 3-MCPD esters and glycidyl esters in vegetable oils and fats. It is certified as Joint JOCS/AOCS Official Method AOCS Cd 29d-19 [25].

Indirect detection techniques

89

Procedure: The oil sample is dissolved in isooctane and then sodium bromide acidic (pH 5) aqueous solution (30%, w/v) containing 90 U/mL lipase (C. cylindracea) is added. The enzymatic cleavage of the analytes from their esters, facilitated by intensive agitation of the reaction mixture by a high-speed shaker, is carried out at room temperature for 30 min. The mixture is then incubated at 80 C for 10 min in order to accelerate the bromination reaction of glycidol. Only after cooling to room temperature the internal standards (pentadeuterated analogues of 3-MCPD, 2-MCPD and 3-MBPD) are added. Afterward, the sample mixture is purified by liquideliquid extraction with hexane and the analytes present in the water phase are derivatized with phenylboronic acid, followed by GCeMS analysis. Similar to the previously described methods, the quantification of 2- and 3-MCPD esters (expressed as bound 2- and 3-MCPD) is based on the 2-MCPD/2-MCPD-d5 and 3-MCPD/3-MCPD-d5 signal ratios, while the quantification of glycidyl esters (expressed as bound glycidol) is based on the 3-MBPD/3-MBPD-d5 signal ratio. The JOCS method has been evaluated predominantly by laboratories in Japan, while it found only very limited application for routine analysis of oils and fats in the rest of the world. The method conveniently combines the enzymatic hydrolysis of 2- and 3-MCPD esters and glycidyl esters with the additional conversion of glycidol into 3-MBPD, which makes the working procedure relatively short. Similar to the Unilever and BLC methods, this method allows for the quantification all three analytes in a single analysis. On the other hand, several challenges have been encountered during the optimization and collaborative validation of the method [22,23] as described in the Cleavage of the MCPD esters section in this chapter. Moreover, the choice of the internal standards and their application after (and not before) the enzymatic cleavage, which is assumed e rather than checked e to be complete, raise some additional concerns about the accuracy and the applicability of this method to the wide range of edible oils and fats. It should be noted that the JOCS method 2.4.14e2016 failed in the analysis of fish oil [24] and a modified procedure (using a different lipase) was proposed and later certified as Joint JOCS/AOCS Recommended Practice Cd 29e-19 [45].

Method comparison Soon after confirming the occurrence of 3-MCPD esters in refined vegetable oils it became apparent that the level of contamination reported

90

Processing Contaminants in Edible Oils

by different laboratories was dependent on the analytical procedure used for quantification. The pioneering method of Weißhaar [8] suffered from poor selectivity because glycidyl esters (which at that time were not known to occur in oils and fats) were quantified together with 3-MCPD esters as a single entity. Later, the method was modified to allow for the separate quantification of 3-MCPD esters and glycidyl esters, and officially adopted as the DGF method C-III 18 (09). Unfortunately, this modification was proven unsuccessful and the method was withdrawn (see Section Elimination of glycidyl esters by acid treatment). Nonetheless, this episode triggered a debate on the reliability of indirect methods in general since they all consist of a series of chemical reactions that could potentially affect selectivity and, thereby, accuracy. With a number of emerging procedures, it became clear their main parameters, namely accuracy, selectivity, and precision, required a thorough comparison by interlaboratory comparison studies, ring tests, and ultimately, official collaborative studies. In the section below selected studies that had an impact on the overall method harmonization are mentioned. Studies comparing indirect methods The first thorough study [9] focused on the evaluation of the conditions applied during transesterification and neutralization/salting out and their impact on the correct quantification of 3-MCPD esters. The results of this study showed that indirect methods based on both acid and alkaline transesterification can provide an accurate determination of 3-MCPD esters under the analytical conditions (namely chemical conversions) strictly prescribed. The methods based on alkaline transesterification were found to be prone to condition variation, mainly due to the degradation of 3-MCPD in alkaline media, which explained some discrepancies in reported data. Later, two indirect methods for the analysis of both 2- and 3-MCPD esters and glycidyl esters, based on acid transesterification (Unilever method) and alkaline transesterification (SGS method) were extensively studied by interlaboratory comparison [46]. A set of 65 samples of different botanical origin, degrees of processing, and wide-ranging contamination were analyzed by both laboratories using their own respective methods. A remarkably good match between the two sets of results was found (R2 values ranged from 0.997 to 0.999) for all three analytes (see Fig. 4.8). Such a demonstration of method agreement was an important prerequisite for a collaborative validation study of both methods, which led to the adoption of these methods as official methods (see Table 4.1).

Indirect detection techniques

91

Figure 4.8 Level of bound 3-MCPD (A), 2-MCPD (B) and glycidol (C) in a set of 65 samples of edible oils and fats of different origin and composition, analyzed by the Unilever method and the SGS method [46].

92

Processing Contaminants in Edible Oils

More recently, the newly optimized BLC method was compared against the SGS method by analyzing a set of 11 samples of various oils and fats [15]. An extensive statistical evaluation of the obtained data sets of 2-MCPD, 3-MCPD and glycidyl esters confirmed good agreement between both methods and the authors concluded that the new method was characterized by a precision that was far better than expected in an interlaboratory comparison. Soon after, the method was subjected to a collaborative study, which commenced its path to the official adoption as the BLC method (Table 4.1). Studies comparing indirect with direct methods Direct methods are characterized by a simple sample preparation that avoids any chemical modification of the analytes, with a potential advantage for accuracy. They require a chromatographic separation and quantification of each individual fatty acid ester of MCPD and glycidol. Comparative studies including both direct and indirect methods are helpful in assessing the method performance. Being based on completely different principles, direct and indirect methods are subjected to different analytical errors; therefore, matching results represents a clear indication of good accuracy. Amid the doubt about the accuracy of the DGF method C-III 18 (09) ten years ago, it was compared with a direct method based on LC-TOF-MS analysis for the quantification of 3-MCPD esters [29] and a direct method based on purification of the oil sample by solid phase extraction, followed by LC-MS analysis for quantitation of the glycidyl esters [34]. Both research groups concluded that the results obtained by the DGF method and the direct approaches were not comparable due to the incomplete elimination of glycidyl esters during the acid pretreatment applied in the DGF method C-III 18 (09), option B. Dubois et al. [47] compared the level of MCPD esters determined by an indirect method based on acid transesterification with a direct method based on sample purification by solid phase extraction and silica gel separation, coupled with LC-TOF-MS analysis. Data obtained by the analysis of a set of 29 samples (Fig. 4.9) showed a good match between the two datasets (R2 > 0.96), confirming the reliability of indirect methods based on acid transesterification. Direct methods are less complex for glycidyl esters than for MCPD esters since the number of different glycidyl esters present in oils and fats is typically limited to five to seven (see Chapter 5 for more details).

Indirect detection techniques

93

Figure 4.9 Level of bound MCPDs (sum of 2- and 3-MCPD mono- and diesters) in a set of 22 samples of palm oil (black diamonds) and 7 samples of palm olein (white diamonds), analyzed by a direct method based on LC-TOFMS analysis and an indirect method based on acid transesterification [47].

Granvogl and Schieberle [48] proposed a direct method for glycidyl esters based on the purification of the oil sample on a silica gel column, followed by LC-MS analysis of the intact glycidyl esters. Comparison of their method with the SGS method through analysis of seven samples showed a decent correlation between both data sets within a wide range of contamination. A similar comparison was carried out between a newly developed direct method based on GCeMS analysis [49] and the Unilever method. The direct method included a first step of acetonitrile extraction of the oil sample, followed by a liquid chromatography purification and GCeMS analysis of the intact glycidyl esters. Analogous to previous studies, the set of samples analyzed (77 in total) covered a wide range of contamination levels (0.1e10 mg/kg of bound glycidol). The good match of the datasets (R2 ¼ 0.993, Fig. 4.10) confirmed the accuracy of both methods. Official collaborative studies and proficiency tests A number of ring tests and proficiency studies for the comparison and/or validation of existing methods have been held in the past decade. The major ones are listed chronologically below.

94

Processing Contaminants in Edible Oils

Figure 4.10 Level of bound glycidol in a set of 77 samples of oils and fats analyzed by an indirect method based on acid transesterification and a direct method based on GCeMS analysis [49].

The first proficiency test for the analysis of 3-MCPD esters in oils was organized by the Institute for Reference Materials and Measurements (IRMMs) of the European Commission’s Joint Research Center in 2009 [50]. Two samplesda virgin olive oil spiked with 3-MCPD dioleate (bound 3-MCPD level: 4.6 mg/kg) and a fully refined palm oil (assigned value of bound 3-MCPD: 8.8 mg/kg)dwere analyzed by 34 participants, which included commercial, industrial, and official control laboratories, and without any restrictions in terms of the method used. The test reflected the situation regarding the analysis of 3-MCPD esters at that time: 85% of participants provided results that were considered satisfactory (absolute z-score below 2) for the spiked sample, but only 56% of participants succeeded with the analysis of refined palm oil. The impact of the analytical method on the outcome was obvious: All five datasets generated by methods based on acid transesterification were found compliant (the reported level ranged between 6.7 and 9.1 mg/kg). Methods based on alkaline transesterification, either those using acid pretreatment or those that were chloride-free, showed much larger variation (3.0e16.5 mg/kg), presumably due to the presence of glycidyl esters, which interfered with the analysis of bound 3-MCPD.

Indirect detection techniques

95

The Food Analysis Performance Assessment Scheme (FAPAS) conducted another proficiency test two years later [51]. From the analysis of a single sample of refined palm oil (an assigned value for bound 3-MCPD was 4.7 mg/kg), 26 datasets were obtained. The results did not show substantial improvement of the situation: Only 62% of participants (16 datasets) complied with the minimum performance criteria (absolute z-score below 2), with reported values ranging from 3.5 to 6.1 kg/mg. Analogous with the previous study, 10 laboratories failed, mainly because of the use of methods based on alkaline transesterification and salting out by sodium chloride (seven datasets). In 2012, the American Oil Chemists’ Society (AOCS) organized a collaborative study focused on the comparison of indirect methods. Complete protocols of three indirect methodsdUnilever method [41], SGS method [10], and DGF method C-VI 18 (10) [28]dwere provided to the participants of the study. Seven samples of spiked or refined oils were analyzed by 20 participating laboratories, resulting in 31 datasets obtained. The three methods tested showed satisfactory performance at the contamination level above 1 mg/kg [52]. As a result, all three methods were adopted as AOCS Official Methods 29-13 [36,38,42]. The same year, the IRMM conducted another interlaboratory study in order to compare current analytical methods for both MCPD esters and glycidyl esters [53]. Only laboratories experienced with this type of analysis were invited to the study. Twenty-two participants had a free choice of the method of analysis to apply; nevertheless, the laboratories showed a strong preference for the established indirect methods e Unilever method, SGS method, and DGF method. Seven samples, comprising both refined and spiked oils, were submitted for the analysis. Satisfactory results were reported by the vast majority of participating laboratories. The obtained results showed good agreement with spiked/assigned values for both bound 3-MCPD and glycidol. Importantly, all three indirect methods applied in this study did not show significant differences in their performance, which supported the outcome of the AOCS collaborative study. Another collaborative study was organized by the Japan Oil Chemists’ Society (JOCS) with the objective to evaluate the performance of an indirect enzymatic method [23]. There were a total number of thirteen participating laboratories from Japan, which were instructed to follow the prescribed method [22] and analyze seven samples of vegetable oils spiked with 0.6e4.6 mg/kg of 3-MCPD dioleate, 2-MCPD dipalmitate, and glycidyl oleate. Despite certain difficulties in the analysis of a high-melting

96

Processing Contaminants in Edible Oils

point solid fat sample, at least 11 laboratories provided acceptable results. The repeatability (RSDr < 8%) and reproducibility (RSDR < 18%) were considered satisfactory and the method was registered as the official JOCS Standard Method 2.4.14e2016 and later as Joint JOCS/AOCS Official Method AOCS Cd 29d-19 [25]. In 2019 an interlaboratory validation study was designed as part of the ISO standardization process of the BLC method. Seventeen participants committed to follow the prescribed protocol [15]. Eight samples of vegetable oils with the concentration range of 0.1e1.4 mg/kg per each contaminant were analyzed. The performance criteria of the method were found satisfactory [54] and the method was certified as ISO standard 183634:2021 [40]. From the very recent studies, two proficiency tests organized by FAPAS are worth mentioning. The test from 2019 [55] entailed the analysis of one sample (assigned values 0.31, 0.15 and 0.20 mg/kg for bound 3-MCPD, 2MCPD and glycidol, respectively). The results showed a compliance with the minimum performance criteria (absolute z-score below 2) in 86% of datasets for bound 3-MCPD (98 datasets), in 92% of datasets for bound 2MCPD (65 datasets) and in 88% of datasets for bound glycidol (94 datasets). One year later [56] another sample (assigned values 0.57, 0.28 and 0.46 mg/ kg for bound 3-MCPD, 2-MCPD and glycidol, respectively) was submitted to participants of the test. An impressive 97% compliance was recorded for bound 3-MCPD (89 datasets), 90% for bound 2-MCPD (60 datasets) and 90% for bound glycidol (89 datasets). Somewhat higher compliance from FAPAS 2662 test can be partly attributed to the higher (about double) contamination level. It should be noted that the vast majority of results were obtained by the officially adopted methods. In addition, about one-third of the participating laboratories did not report bound 2-MCPD, which is not regulated by EU legislation. In any case, these recent positive findings demonstrate that the current methods for MCPD and glycidyl esters in oils and fats are perfectly “fit for purpose” and that these methods have been successfully implemented in a large number of laboratories.

Method automation The indirect methods presented in this chapter are perfectly suitable for the monitoring of MCPD and glycidyl esters in a large number of edible oils and fats samples. These methods, however, are based on relatively

Indirect detection techniques

97

complex protocols with a tedious and time-consuming sample preparation demanding precise execution. It is logical that method automation would be welcome by food safety and quality control laboratories specialized in monitoring trace contaminants. Since the adoption of the EU regulatory limits for glycidyl esters [2] and 3-MCPD esters [3] in March 2018 and January 2021, respectively, a monitoring of these process contaminants became routine to ensure regulatory compliance. An increased demand for the analyses of 3-MCPD and glycidyl esters further increased the demand for fully automated applications. Over past few years, several solutions for fully automated determination of 3-MCPD and glycidyl esters in edible oils have been presented [15,31,57]. These are based on the established reliable methods and follow the official protocols with each individual step fully automated. All four ISO methods (see Section Officially adopted methods) have been presented in a fully automated setup. The critical point of the method automation is the ability of the automated system to meet performance criteria of the original (manual) method on a large set of samples. Another point of attention is the sample throughput, which should match, and preferably exceed, the manually performed method. Nowadays, a number of automated, high throughput sample preparation systems based on officially adopted methods are being commercially offered by several vendors.

Analysis of oil-based foodstuffs With a growing interest in monitoring MCPD esters and glycidyl esters in final food products, particularly in those with a higher content of refined vegetable oils and fats (spreads, margarines, mayonnaizes, dressings, etc.), there is an increasing need for developing methods suitable for this purpose. While this section deals with methods for oil-based food products, the methodology for other food matrices, infant formulas in particular, are discussed in detail in Chapter 7 “Methods to detect MCPD and glycidyl esters in complex food matrices”. It should be mentioned here that the indirect methods described earlier in this chapter are applicable solely to oils and fats. Few early attempts were made to quantify 3-MCPD esters or glycidyl esters in complex foodstuffs (see Ref. [18] and references therein). These methods, based on a preliminary step of fat extraction followed by the application of one of

98

Processing Contaminants in Edible Oils

the indirect methods for oils and fats, generally relied on the assumption that native MCPD esters and glycidyl esters were extracted from the food matrix in the same proportion as the internal standard, which is added in the first stages of sample preparation. Unfortunately, this assumption is not always valid, since the complexity of the physicalechemical interactions of native analytes with other components of the food matrix can have a significant impact on the extraction yield and, consequently, accuracy of the results [18].

Challenges in foodstuff analysis There are four fundamental points that must be critically evaluated during development and validation of an analytical procedure for complex food matrices: ensuring quantitative extraction of the analytes from the sample, avoiding degradation or interconversion of the analytes, eliminating interfering compounds co-present in the matrix, and assessing method accuracy [58]. In complex foodstuffs, oils and fats (and lipophilic contaminants present therein) may be present in different forms: as free oil, as a component of an emulsion or as an entrapped fat (e.g., in a coating). The structure of the foodstuffs must be broken down upon homogenization (often facilitated by elevated temperature to melt down high-melting point fats) and the lipophilic substances are then extracted by an organic solvent. The obvious risks are incomplete or selective (when only particular fat fractions are isolated) extraction, or a potential loss of certain analytes in the interface between organic solvent and the sample (e.g., 3-MCPD monoesters possess substantially higher polarity than 3-MCPD diesters or glycidyl esters and tend to concentrate in the interface layer), as depicted in Fig. 4.11.

Figure 4.11 Schematic visualization of oils and fats present in complex foodstuffs and the microstructure break down upon homogenization/heating and solvent extraction.

Indirect detection techniques

99

The difficulties associated with the analyte extraction (one may assume a complete extraction of internal standards which are not hindered by the product microstructure) may lead to underestimation of the results. The reactivity of target analytes, glycidyl esters in particular, has already been mentioned earlier in this chapter. Complex foodstuffs typically contain a variety of ingredients (e.g., water, salt, acidifiers), some of them being nucleophilic reactants or catalysts, and glycidyl esters can be degraded, or even converted to MCPD esters, which would inevitably lead to false results. Thorough optimization of the extraction step and elimination of certain components (e.g., sodium chloride) is critical. The matrix effect, caused by the microstructure of the food product and individual co-present ingredients may have a large impact on method accuracy. Oil-based food products may be generally divided into those based on water-in-oil emulsions (e.g., margarines) and those based on oil-inwater emulsions (e.g., mayonnaizes, salad dressings, low fat spreadable creams, dairy creams, cooking fats). Whereas the stability of water-in-oil emulsions is mainly achieved due to the fat crystal network, oil-in-water emulsions are stabilized solely by emulsifiers (up to 5e10%, w/w) and thickening agents. In addition to this challenge, dairy-based formulations contain high levels of milk proteins (above 5%, w/w) and other surfaceactive compounds such as phospholipids, which are known to negatively affect the fat extraction yield. This illustrates the complexity of the matter and implies that any proposed extraction procedure must be validated for each type of matrix. Verifying parameters of a newly proposed method, namely its accuracy, is another major challenge. A lack of certified reference materials and the absence of officially validated methods suitable for method comparison do not allow standard analytical routines for the determination of accuracy to be applied. Moreover, spiking complex foodstuffs is not an appropriate option since their macrostructure is not homogeneous and, thereby, the spike compounds cannot be delivered into all target destinations within the product structure (e.g., into the micelles). One solution to this problem is to manufacture desired food matrices by formulating the product with oils and fats of known levels of MCPD esters and glycidyl esters. In this fashion, various reference materials with known contamination levels can be manufactured for the purpose of method validation. Development of method for oil-based foodstuffs As of ten years ago, no standard method for the analysis of these contaminants in food products had been established. That, however, did not

100

Processing Contaminants in Edible Oils

stop laboratories from generating data using various in-house methods. These data were even collected for the purpose of an exposure assessment [59]. As expected, the quality of these data aroused serious concerns [4]. Amid the uncertainty regarding the quality of occurrence data in oil-based products Unilever commenced a study, which aimed to develop the first method for the simultaneous quantification of MCPD esters and glycidyl esters in oil-based food products. The novel method developed by Ermacora and Hrncirík [18] consisted of a series of extraction steps by organic solvent (heptane/methyl tert-butyl ether, 1:2, v/v) and water, assisted by homogenization, heating (60 C) and sonication, followed by dissolving the isolated oil residue in anhydrous tetrahydrofuran prior to application of the Unilever method (see Section Officially adopted methods) for the quantification of MCPD esters and glycidyl esters. Since accuracy is a fundamental requisite of a method, it was thoroughly tested on products manufactured in in-house test production facilities. Such an approach allowed for the control of manufacturing parameters and accounted for any food matrixeanalyte interactions. In addition, sampling could be performed at various stages of the production process. For each individual formulation tested, the analyses of the initial mixture of fatsoluble ingredients (so called “fat-phase”), the mixture of water and fatphase before processing (called “a premix”) and the final product, were carried out simultaneously. The application window of the method was tested on food products based on both water-in-oil and oil-in-water emulsions, with a wide range of fat levels (28e98%, w/w) and other ingredients present in the formulations. The levels of sodium chloride and surface-active compounds were found to have no impact on method performance. The method showed good accuracy (with recoveries of 97e106% for bound 3-MCPD and 2-MCPD and 88%e115% for bound glycidol) and sensitivity (LOD below 0.05 mg/kg), and the repeatability and reproducibility were satisfactory (RSD below 2% and 5%, respectively) for all analytes [18]. AOCS collaborative study Soon after the method was published, AOCS together with Unilever launched a collaborative study with the intention to validate the method by evaluating its performance in an interlaboratory comparison [60].

Indirect detection techniques

101

In total, 21 laboratories from nine countries participated in the study. The participants were obliged to follow the novel extraction procedure [18] followed by application of the Unilever method and/or other official methods for the determination of MCPD esters and glycidyl esters in oils and fats (see Section Officially adopted methods). Only the results obtained by the Unilever method (17 participants) were used for evaluation of the overall method performance, while data obtained by the SGS method (3 participants), the DGF method (2 participants) or other methods (1 participant) were used only for comparative purposes. Five samples, greatly varying in composition and a microstructure, were manufactured: a brick margarine (75% fat), two soft margarines (40% and 75% fat), a mayonnaize (75% fat) and a dressing (30% fat). The desired level of contamination was achieved by blending oils and fats with known levels of MCPD esters and glycidyl esters. The assigned values (Table 4.2) were determined according to established protocols. The contamination ranges varied from 0.1 to 1.1 mg/kg for bound 3-MCPD, from below the LOQ to 0.5 mg/kg for 2-MCPD and from below the LOD to 3.0 mg/kg for bound glycidol. Four out of five samples contained less than 0.5 mg/kg of each analyte, which reflected realistic levels at the time the study was performed. Prior to distribution of the test material to the participants, the samples were subjected to a homogeneity check according to ISO/IEC 17043:2010 Table 4.2 Overview of test materials used for the collaborative study leading to the adoption of the Official AOCS Method Cd 30e15 [60].

Sample description/ID

Brick margarine (75% fat), AOCS 2015e07/10 Soft margarine (75% fat), AOCS 2015 - 06 Soft margarine (40% fat), AOCS 2015 - 11 Mayonnaize (75% fat), AOCS 2015 - 09 Mayonnaize (30% fat), AOCS 2015e05/08 a b

Bound 3-MCPD (mg/kg)

Bound 2-MCPD (mg/kg)

Bound glycidol (mg/kg)

1.06  0.06

0.53  0.02

3.03  0.20

0.45  0.05

0.23  0.03

0.37  0.04

0.23  0.04

0.11  0.02

0.13  0.02

0.43  0.05

0.21  0.03

0.19  0.03

0.11  0.02

0.06  0.01a