Handbook of Food Analysis Instruments

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page i 22.7.2008 6:29am Compositor Name: VBalamugund...

Author: Semih Otles

436 downloads 2268 Views 13MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page i 22.7.2008 6:29am Compositor Name: VBalamugundan

Handbook of

FOOD ANALYSIS INSTRUMENTS

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page ii 22.7.2008 6:29am Compositor Name: VBalamugundan

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page iii 22.7.2008 6:29am Compositor Name: VBalamugundan

Handbook of

FOOD ANALYSIS INSTRUMENTS

Edited by

Semih Ötles¸

Boca Raton London New York

CRC Press is an imprint of the Taylor & Francis Group, an informa business

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page iv

22.7.2008 6:29am Compositor Name: VBalamugundan

CRC Press Taylor & Francis Group 6000 Broken Sound Parkway NW, Suite 300 Boca Raton, FL 33487-2742 © 2009 by Taylor & Francis Group, LLC CRC Press is an imprint of Taylor & Francis Group, an Informa business No claim to original U.S. Government works Printed in the United States of America on acid-free paper 10 9 8 7 6 5 4 3 2 1 International Standard Book Number-13: 978-1-4200-4566-6 (Hardcover) This book contains information obtained from authentic and highly regarded sources. Reasonable efforts have been made to publish reliable data and information, but the author and publisher cannot assume responsibility for the validity of all materials or the consequences of their use. The authors and publishers have attempted to trace the copyright holders of all material reproduced in this publication and apologize to copyright holders if permission to publish in this form has not been obtained. If any copyright material has not been acknowledged please write and let us know so we may rectify in any future reprint. Except as permitted under U.S. Copyright Law, no part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information storage or retrieval system, without written permission from the publishers. For permission to photocopy or use material electronically from this work, please access www.copyright.com (http:// www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe. Library of Congress Cataloging-in-Publication Data Handbook of food analysis instruments / editor, Semih Otles. p. cm. Includes bibliographical references and index. ISBN 978-1-4200-4566-6 (alk. paper) 1. Food--Analysis--Equipment and supplies--Handbooks, manuals, etc. I. Ötles, Semih. II. Title. TX541.H365 2008 664’.07--dc22 Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the CRC Press Web site at http://www.crcpress.com

2008013711

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page v 22.7.2008 6:29am Compositor Name: VBalamugundan

Contents Preface ............................................................................................................................................ vii Acknowledgments ........................................................................................................................... ix Editor ............................................................................................................................................... xi Contributors ................................................................................................................................... xiii

Chapter 1

Data Analysis Techniques ......................................................................................... 1 Michael H. Tunick

Chapter 2

Microextraction Methods in Food Analysis.............................................................. 7 Kathy Ridgway, Sam P.D. Lalljie, and Roger M. Smith

Chapter 3

Supercritical Fluid Extraction in Food Analysis ..................................................... 25 Ruhan Askin, Motonobu Goto, and Mitsuru Sasaki

Chapter 4

Microwave-Assisted Processes in Food Analysis ................................................... 57 Jacqueline M.R. Bélanger and J.R. Jocelyn Paré

Chapter 5

Ultrasound-Assisted Extraction in Food Analysis .................................................. 85 Farid Chemat, Valérie Tomao, and Matthieu Virot

Chapter 6

Advances in High-Performance Liquid Chromatography and Its Application to the Analysis of Foods and Beverages ............................... 105 Peter Varelis

Chapter 7

Gas Chromatography in Food Analysis ................................................................ 119 Jana Hajslova and Tomas Cajka

Chapter 8

Preparative Layer Chromatography in Food Analysis .......................................... 145 Joseph Sherma

Chapter 9

Ion Chromatography in Food Analysis ................................................................. 161 William R. LaCourse

Chapter 10

Mass Spectrometry and Hyphenated Instruments in Food Analysis .................... 197 Tomas Cajka, Jana Hajslova, and Katerina Mastovska

v

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page vi

22.7.2008 6:29am Compositor Name: VBalamugundan

vi

Chapter 11

Instruments to Analyze Food Colors .................................................................... 229 Carmen Socaciu and Horst A. Diehl

Chapter 12

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis of Food and Grain Composition............................................................. 247 Ion C. Baianu and T. You

Chapter 13

Nuclear Magnetic Resonance Spectroscopy in Food Analysis ............................ 281 Francesco Capozzi and Mauro A. Cremonini

Chapter 14

Atomic Absorption, Atomic Emission, and Inductively Coupled Plasma Spectroscopies in Food Analysis .............................................................. 319 John R. Dean and Renli Ma

Chapter 15

Autoﬂuorescence Spectroscopy in Food Analysis ................................................ 347 Charlotte Møller Andersen, Jens Petter Wold, and Søren Balling Engelsen

Chapter 16

Electronic Nose Technology in Food Analysis..................................................... 365 Figen Korel and Murat Ö. Balaban

Chapter 17

Electroanalytical Techniques and Instrumentation in Food Analysis ................... 379 Rubin Gulaboski and Carlos M. Pereira

Chapter 18

Capillary Electrophoresis in Food Analysis .......................................................... 403 Carmen García-Ruiz and Maria Luisa Marina

Chapter 19

Gel Electrophoresis in Food Analysis ................................................................... 423 Reiner Westermeier and Burghardt Scheibe

Chapter 20

Multiplexed Immunoassays in Food Analysis ...................................................... 439 Chien-Sheng Chen, Antje J. Baeumner, and Richard A. Durst

Chapter 21

Rheological Instruments in Food Analysis ........................................................... 461 Nesli Sozer and Jozef L. Kokini

Chapter 22

Scanning Electron and Transmission Electron Microscopies in Food Analysis ................................................................................................... 495 José M. Aguilera and Pedro Bouchon

Index............................................................................................................................................. 513

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page vii 22.7.2008 6:29am Compositor Name: VBalamugundan

Preface The analysis of foods—identiﬁcation, speciation, and determination of components, additives, and contaminants in different raw materials and products—is a critical endeavor in food processing and manufacturing companies since the presence and interactions of various compounds in foods during storage and processing have an impact on all aspects of the quality of food products. The application of proper methods, suitable for analysis of different matrixes with the required method of detection, is crucial for food quality and safety control during production and marketing. In the course of the twenty-ﬁrst century, analytic methods used in food science have evolved considerably. While traditional methods are still used, most analysis now involves the use of increasingly sophisticated instruments. Although there are a number of books that explain the principles of food analysis, describe how to conduct food analysis, and discuss test results, there are few books that focus on understanding the actual instruments used in the analysis. Such instruments are used for a wide variety of tasks, including analyzing the degradation of edible oils or the vitamins in baby food; or quantifying food additives, pesticide residues, or the color in packaging materials; or determining the distinct aroma found in natural products. This handbook has been prepared by a team of food scientists=chemists=biochemists who have extensive personal experience in research of food analysis and practical food control in the industry. This handbook aids the analyst by providing a valuable reference regarding the newly developed instruments and methods of analysis of food components and additives. The handbook, contributed to by 44 leading scientists, many of whom actually developed or reﬁned the techniques and instruments, presents each technique in a uniform format, in a style that can be understood by a reader who is not familiar with the particular technique. Each chapter is structured to provide a description of the information the technique can provide, a simple explanation of how it works, examples of its application, and practical information such as names of instrument vendors, relative costs of instruments and materials, training and education of personnel, and references for more detailed information. This format also facilitates comparison of techniques. The use of different authors to cover a broad spectrum of techniques resulted in some differences of style, but overall the handbook achieved its goal. The handbook comprises a preface, a contributor list, and a subject index and 22 chapters, which take the reader through brief and accessible descriptions of instruments of analysis of food components and additives. Each chapter in the handbook focuses on a speciﬁc type of instrument: capillary electrophoresis, high-performance liquid chromatography (HPLC), nuclear magnetic resonance (NMR), or microwave-assisted process, etc., among many others. Each chapter follows a consistent format, examining the operating principles of a particular technique, its deﬁnitions, theory, and applications to food analysis. Each chapter is introduced by an overview written by the chapter authors. The introductory chapter, ‘‘Data Analysis Techniques,’’ covers topics relevant to all techniques, including calibration, standard addition, internal standards, selectivity, accuracy, precision, detection limit, quantiﬁcation limit, range, robustness, speed, and convenience. The remaining 21 chapters address the major areas of food analysis instruments for sample processing of foods and for food analysis. Chapters 2 and 3 explain sample processing focused on puriﬁcation and enrichment (Chapter 2: microextraction methods in food analysis such as LPME, SPE, SPME, and SBSE) and extraction while Chapters 3 through 5 explain supercritical ﬂuid extraction, microwave-assisted processes, and ultrasound-assisted extraction. The other chapters explain food analysis instruments based on chromatography (Chapters 6 through 9 and Chapter 18: high-pressure liquid, gas,

vii

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page viii 22.7.2008 6:29am Compositor Name: VBalamugundan

viii

preparative layer, ion, and capillary chromatographies); mass spectroscopy and hyphenated techniques (Chapter 10: MS, GC-MS, HPLC-MS, ICP-MS, etc.); physical parameters such as optical (Chapters 11 through 15: color measurements; near infrared; nuclear magnetic resonance; Raman, atomic absorption, emission, and inductively coupled plasma; and autoﬂuorescence spectroscopies); electrical (Chapter 17: electroanalytical techniques and instrumentation); rheological (Chapter 21: rheological instruments); dedicated systems (Chapter 16: electronic nose technology); and based on biological techniques (Chapters 19 through 22: gel electrophoresis, multiplexed immunoassays, and scanning electron and transmission electron microscopies). The handbook addresses primarily food science graduate students, food chemists in industry and food quality control, as well as persons who participate in continuing education systems. Many topics will also be of interest to students of chemistry and biology. Some chapters of the handbook could as well be useful to readers interested in the quality of food. Semih Ötles¸

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page ix

22.7.2008 6:29am Compositor Name: VBalamugundan

Acknowledgments The editor would like to thank all the contributors for the hard work they put into the various chapters of the handbook. The editor also thanks the people at Taylor & Francis Group/CRC Press for their help with the production of this book and also expresses his sincere gratitude to Stephen M. Zollo, Taylor & Francis Group, for his help in preparing the handbook. Finally, special thanks to my wife, Sema Ötles¸, for her patience during the preparation and publication steps of the handbook.

ix

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page x 22.7.2008 6:29am Compositor Name: VBalamugundan

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page xi

22.7.2008 6:29am Compositor Name: VBalamugundan

Editor Semih Ötles¸, a native of Izmir, Turkey, obtained his BSc from the Department of Food Engineering (Ege University, Izmir, Turkey) in 1980. During his assistantship at Ege University in 1985, he received an MS in food chemistry, and in 1989, after completing his thesis research on the instrumental analysis and chemistry of vitamins in foods he received a PhD in food chemistry from Ege University. During 1991–1992, he completed his postdoctoral training on an Organisation for Economic Co-Operation and Development (OECD) postdoctoral fellowship at the Research Center for Meat Technology, Melle, Ghent University, Belgium. Later, he joined the Department of Food Engineering at Ege University as a scientist in food chemistry, and was promoted to associate professor in 1993 and professor in 2000. During the summer of 2005, he was the visiting professor at Kumamoto University, Kumamoto, Japan, as a fellow of the Japan Society of the Promotion of Science (JSPS). During 1996–1998, he was the deputy director at the Ege Vocational School of Higher Studies. Since 2003, he has been the vice dean of the engineering faculty, Ege University. Also, he is a member of the steering committee in the geriatrics department of the medical faculty, Ege University. The research activities of Professor Ötles¸ have focused on the instrumental analysis of food compounds. He began a series of projects on separation and analysis techniques using highperformance liquid chromatography (HPLC), ﬁrst for analysis of vitamins in foods, then proteins, carbohydrates, and most recently carotenoids. Other activities span the ﬁelds of SFE (supercritical ﬂuid extraction), GC (gas chromatography), GC=MS (mass spectrometry) analysis, soy chemistry, aromatics, medical and functional foods, and nutraceutical chemistry, including multiresidue analysis of various foods, and n-3 fatty acids in ﬁsh oils. Professor Ötles¸ has authored or coauthored more than 150 publications (technical papers, book chapters, and books) and has presented seminars in these areas. He is a member of several scientiﬁc societies, associations, and organizations including the Asian Paciﬁc Organization for Cancer Prevention and International Society of Food Physicists. He is a member of the steering committee of the Food Safety Association, Istanbul, Turkey and the APOCP (Asian Paciﬁc Organization for Cancer Prevention) local scientiﬁc bureau, and is a Turkish representative of the International Society of Food Physicists (ISFP), and has organized international congresses on diet=cancer, functional foods, and food physics. Dr. Ötles¸ is a member of the editorial advisory boards of the Asian Paciﬁc Journal of Cancer Prevention (APJCP); FSTA (Food Science and Technology Abstracts) of IFIS (International Food Information Service); Current Topics in Nutraceutical Research; Electronic Journal of Environmental, Agricultural and Food Chemistry (EJEAFChe); Newsline (IUFoST, Corr.); Journal of Oil, Soap, Cosmetics; Turkish World Food; Acta Scientiarum Polonorum; Trends in Food Science and Technology; Pakistan Journal of Nutrition; Journal of Food Technology; Turkish Journal of Toxicology; Electronic Journal of Polish Agricultural Universities; CHI; Plastic and Packaging Technology; Genes & Nutrition; Popular Health Journal; Drink Tech, Advances in Food Sciences; and Keyfood Magazine. He is also a referee=reviewer for The Journal of AOAC International, Journal of Experimental Marine Biology and Ecology, Journal of Medical Foods, die Nahrung, Journal of Alternative and Complementary Medicine, Journal of Harvest, The Analyst, AU Journal of Science and Technology, Turkish Journal of Fisheries and Aquatic Sciences, GOPU Journal of Agriculture Faculty, Journal of Agricultural and Food Chemistry, Electronic Journal of Biotechnology, Industrial & Engineering Chemistry Research, Journal of Food Processing and Preservation, Cancer Causes and Control, and Food Chemistry.

xi

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page xii 22.7.2008 6:29am Compositor Name: VBalamugundan

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page xiii 22.7.2008 6:29am Compositor Name: VBalamugundan

Contributors José M. Aguilera Department of Chemical and Bioprocess Engineering Universidad Católica de Chile Santiago, Chile Charlotte Møller Andersen Department of Food Science University of Copenhagen Frederiksberg, Denmark

Tomas Cajka Department of Food Chemistry and Analysis Institute of Chemical Technology Prague, Czech Republic Francesco Capozzi Department of Food Science University of Bologna Cesena, Italy

Ruhan Askin Department of Applied Chemistry and Biochemistry Kumamoto University Kumamoto, Japan

Farid Chemat Sécurité et Qualité des Produits d’Origine Végétale University of Avignon Avignon, France

Antje J. Baeumner Department of Biological and Environmental Engineering Cornell University Ithaca, New York

Chien-Sheng Chen Department of Food Science National Taiwan Ocean University Keelung, Taiwan

Ion C. Baianu Food Science and Human Nutrition and Nuclear Engineering Departments University of Illinois Urbana, Illinois Murat Ö. Balaban Fishery Industrial Technology Center University of Alaska, Fairbanks Kodiak, Alaska Jacqueline M.R. Bélanger Green Technologies Division Environment Canada Ottawa, Ontario, Canada Pedro Bouchon Department of Chemical and Bioprocess Engineering Universidad Católica de Chile Santiago, Chile

Mauro A. Cremonini Department of Food Science University of Bologna Cesena, Italy John R. Dean School of Applied Sciences Northumbria University Newcastle upon Tyne, United Kingdom Horst A. Diehl Institute of Biophysics University of Bremen Bremen, Germany Richard A. Durst Department of Food Science and Technology Department of Biological and Environmental Engineering Cornell University Ithaca, New York

xiii

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page xiv 22.7.2008 6:29am Compositor Name: VBalamugundan

xiv

Søren Balling Engelsen Department of Food Science The Royal Veterinary and Agricultural University Frederiksberg, Denmark Carmen García-Ruiz Department of Analytical Chemistry University of Alcalá Madrid, Spain Motonobu Goto Department of Applied Chemistry and Biochemistry Kumamoto University Kumamoto, Japan Rubin Gulaboski Faculdade de Ciências Universidade do Porto Porto, Portugal

Renli Ma School of Applied Sciences Northumbria University Newcastle upon Tyne, United Kingdom Maria Luisa Marina Department of Analytical Chemistry University of Alcalá Madrid, Spain Katerina Mastovska Agricultural Research Service Eastern Regional Research Center U.S. Department of Agriculture Wyndmoor, Pennsylvania J.R. Jocelyn Paré Green Technologies Division Environment Canada Ottawa, Ontario, Canada

Jana Hajslova Department of Food Chemistry and Analysis Institute of Chemical Technology Prague, Czech Republic

Carlos M. Pereira Departmento de Química Faculdade de Ciências da Universidade do Porto Porto, Portugal

Jozef L. Kokini Department of Food Science and Center for Advanced Food Technology Rutgers University New Brunswick, New Jersey

Kathy Ridgway Safety and Environmental Assurance Centre Unilever Bedfordshire, United Kingdom

Figen Korel Izmir Institute of Technology Food Engineering Department Urla, Izmir, Turkey

Mitsuru Sasaki Department of Applied Chemistry and Biochemistry Kumamoto University Kumamoto, Japan

William R. LaCourse Department of Chemistry and Biochemistry University of Maryland Baltimore, Maryland

Burghardt Scheibe Protein Sciences GE-Healthcare Europe Munich, Germany

Sam P.D. Lalljie Safety and Environmental Assurance Centre Unilever Bedfordshire, United Kingdom

Joseph Sherma Department of Chemistry Lafayette College Easton, Pennsylvania

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page xv 22.7.2008 6:29am Compositor Name: VBalamugundan

xv

Roger M. Smith Department of Chemistry Loughborough University Leicestershire, United Kingdom

Peter Varelis Department of Biological, Chemical, and Physical Sciences Illinois Institute of Technology Summit Argo, Illinois

Carmen Socaciu Department of Chemistry and Biochemistry University of Agricultural Sciences and Veterinary Medicine Cluj-Napoca, Romania

Matthieu Virot Sécurité et Qualité des Produits d’Origine Végétale University of Avignon Avignon, France

Nesli Sozer Food Engineering Department Gaziantep University Gaziantep, Turkey

Reiner Westermeier Protein Sciences GE-Healthcare Europe Munich, Germany

Valérie Tomao Sécurité et Qualité des Produits d’Origine Végétale University of Avignon Avignon, France

Jens Petter Wold Matforsk Osloveien, Norway

Michael H. Tunick Dairy Processing and Products Research Unit Eastern Regional Research Center U.S. Department of Agriculture Wyndmoor, Pennsylvania

T. You Food Science and Human Nutrition and Nuclear Engineering Departments University of Illinois Urbana, Illinois

Otles/Handbook of Food Analysis Instruments 45660_C000 Final Proof page xvi 22.7.2008 6:29am Compositor Name: VBalamugundan

Otles/Handbook of Food Analysis Instruments 45660_C001 Final Proof page 1 14.7.2008 8:54am Compositor Name: BMani

1 Data Analysis Techniques Michael H. Tunick CONTENTS 1.1 1.2

Introduction .............................................................................................................................. 1 Measurement Techniques ........................................................................................................ 2 1.2.1 Calibration ................................................................................................................... 2 1.2.1.1 Classical Calibration .................................................................................... 2 1.2.1.2 Single-Point Calibration ............................................................................... 2 1.2.1.3 Inverse Calibration ....................................................................................... 2 1.2.2 Standard Addition ....................................................................................................... 3 1.2.3 Internal Standards ........................................................................................................ 3 1.3 Fundamental Criteria ............................................................................................................... 3 1.3.1 Selectivity .................................................................................................................... 3 1.3.2 Accuracy ...................................................................................................................... 3 1.3.3 Precision ...................................................................................................................... 4 1.3.4 Detection Limit ........................................................................................................... 4 1.3.5 Quantiﬁcation Limit .................................................................................................... 4 1.3.6 Range ........................................................................................................................... 4 1.4 Other Considerations ............................................................................................................... 4 1.4.1 Robustness ................................................................................................................... 4 1.4.2 Speed ........................................................................................................................... 5 1.4.3 Convenience ................................................................................................................ 5 1.5 Summary .................................................................................................................................. 5 References ......................................................................................................................................... 5

1.1 INTRODUCTION When a food scientist needs to measure a quantity, such as sample mass or volume, he or she performs a direct measurement, having a good idea of the accuracy and precision involved. But when the concentration of a substance in a sample matrix must be found, the analyst has to make an indirect measurement by calculating the quantity from the measurement of other quantities [1]. Indirect measurements are obtained by correlating a result with sample concentration, which introduces the possibility of decreased accuracy and precision. An equation relating analyte concentration and the instrumental response is formed by using standards and calibrations, and then applied to predict the concentration of the unknown [2]. The procedure must demonstrate traceability, deﬁned as an unbroken chain of comparisons from the measurement to the appropriate national or international standards [3]. This chapter outlines the various techniques available for relating the output of an instrument with the quantity being sought and the necessary criteria involved.

1

Otles/Handbook of Food Analysis Instruments 45660_C001 Final Proof page 2 14.7.2008 8:54am Compositor Name: BMani

2

Handbook of Food Analysis Instruments

1.2 MEASUREMENT TECHNIQUES 1.2.1 CALIBRATION 1.2.1.1

Classical Calibration

The common methods for relating concentration and instrumental response are calibration, standard addition, and the use of internal standards. A classical calibration, also called the standard series method or external standards method, is frequently employed in analytical chemistry. A series of samples containing known concentrations of the substance in question are analyzed, and the resulting responses are plotted against concentration to obtain a calibration curve. The curve is often linear, following the equation y ¼ mc þ b

(1:1)

where y is the instrumental response c is the analyte concentration m is the slope (deﬁned as the sensitivity [3]) b is the y-intercept, which corresponds to the value for the blank When a sample containing an unknown c is analyzed, the response is substituted into the equation to obtain the concentration. Linear regression plots and straight linear plots are most often employed, but quadratic regressions, log plots, etc., are sometimes utilized. For example, the author uses a nitrogen analyzer that is calibrated with an ethylenediaminetetraacetic acid (EDTA) standard that contains 9.56% nitrogen. The EDTA itself was calibrated by the instrument manufacturer against a carbon–hydrogen–nitrogen standard from NIST (National Institute of Standards and Technology), thus establishing traceability. The instrument plots the weights of EDTA against the areas of the response, and calculates ﬁxed and regressed linear, quadratic, and cubic calibration curves. The simplest curve yielding with an acceptable correlation coefﬁcient (0.999 or higher) is selected for the calibration curve. Mitchell et al. [4] detailed a rigorous method for performing this type of calibration. After selecting and analyzing the standards, the regression order is selected, outliers are rejected, and regression equations are obtained with conﬁdence bands. 1.2.1.2

Single-Point Calibration

When a response curve is consistently linear with a zero or analytically insigniﬁcant intercept, a calibration may be obtained using a single reference point standard. Ideally, there is a linear relationship extending from the origin through the calibration point. This single-point calibration offers savings in time and effort. The value for the intercept must be reported [5]. 1.2.1.3

Inverse Calibration

The classical calibration above assumes no errors in c, but sample preparation nowadays may be less accurate than instrumental measurement. An inverse calibration is performed by using c ¼ my þ b

(1:2)

and then comparing with the classical calibration. Errors in sample preparation should be suspected if the results do not match up well [2]. Centner et al. demonstrated that inverse calibration yields more reliable predictions than classical calibration [6], and Grientschnig concluded that this was true regardless of the size of the calibration and test data sets [7].

Otles/Handbook of Food Analysis Instruments 45660_C001 Final Proof page 3 14.7.2008 8:54am Compositor Name: BMani

Data Analysis Techniques

3

1.2.2 STANDARD ADDITION The standard addition method is also known as the additive method, or simply as spiking. A known amount of the constituent being analyzed, the spike, is added to the sample to produce a larger instrumental response. For instance, a sample is analyzed to obtain an estimated result, a small concentrated amount of analyte equal to the amount presumed to be in the sample is added, and the sample is analyzed again to see if the response has doubled. A linear variation between concentration and response is assumed. The method is especially useful if an interfering substance is suspected, since its response will not change when the spike is introduced. The standard addition approach may also be used when a blank sample matrix (without analyte) can be obtained. Saxberg and Kowalski [8] developed and Kalivas [9] extended a generalized standard addition method using multiple linear regressions that allow for simultaneous analysis of different components in a mixture while accounting for interference.

1.2.3 INTERNAL STANDARDS An internal standard is a substance that is similar (but not identical) to the analyte and is added to the sample. The ratio of the responses to the internal standard and the analyte is then compared to a calibration curve. The instrumental responses to the two must be distinguishable. Internal standards are often used when the scientist suspects a loss of analyte when the sample is prepared or when it is introduced into the instrument. A common internal standard in mass spectrometry is the deuterated version of the constituent of interest, since their responses are different, but possible losses before measurement should be identical. Internal standards are also useful when the analyte is not stable enough to be calibrated in other ways, although side reactions or other consequences could occur. Such effects took place when Álvarez del Pino et al. [10] compared internal and external standards for determining tannin in Spanish shrubs. They found that the slopes of the calibration lines were different when puriﬁed tannin was used as an internal or external standard, apparently because the internal standard reacted with other components in the sample.

1.3 FUNDAMENTAL CRITERIA 1.3.1 SELECTIVITY Whichever technique is selected for relating concentration with response, there are several factors that must be considered for the results to be valid. These include selectivity, range, accuracy, precision, detection limit, and quantitation limit. The selectivity of a method is its ability to measure the analyte in the sample matrix in the presence of other sample components. In chromatography, for instance, selectivity refers to the ability of a phase system to retain solutes to signiﬁcantly different extents, resulting in analyte peaks that are completely resolved from other peaks. The term ‘‘speciﬁcity’’ is usually discouraged since it implies that nothing besides the analyte contributes to the result [11].

1.3.2 ACCURACY The reliability of a method is based on its accuracy and precision. Accuracy, the difference between a measured value and the true value, is expressed in terms of error. A consistent error, such as one caused by an improperly prepared reagent, may produce replicate results that are similar but inaccurate by the same amount. This type of error is known as bias [2]. In a narrow sense, the only true values that may be known for certain are obtained in deﬁned quantities and in counting discrete objects. All other measurements are obtained by comparison to a reference standard, such as one provided by NIST, or by comparison to another method known to be reliable.

Otles/Handbook of Food Analysis Instruments 45660_C001 Final Proof page 4 14.7.2008 8:54am Compositor Name: BMani

4

Handbook of Food Analysis Instruments

1.3.3 PRECISION Precision is the amount of scatter in replicate measurements of the same quantity, and is expressed in terms of deviation. Measurement, sampling, and calibration errors all contribute to decreased precision and increased uncertainty [2]. Internal precision is measured by repeatability standard deviation, which reﬂects the results obtained on a test material by the same operator using the same method in the same laboratory with the same equipment within a short period of time. External precision is measured by reproducibility standard deviation and indicates the results obtained on a test material by different operators using the same method in different laboratories with different equipment [12]. In reporting results of a radical scavenging capacity assay for grains and ﬂours, for example, Cheng et al. [13] expressed accuracy as percentage of recovery of the calibration standard and precision as intraday (same operator, same day) and interday (same operator, different days) variabilities.

1.3.4 DETECTION LIMIT Instrumental noise consists of extraneous and unwanted signals which may result from thermal motions of electrons (Johnson–Nyquist noise), random ﬂuctuations of current (shot noise), environmental factors, and other sources [14]. The detection limit, or minimum detectable value, is the lowest concentration of analyte that produces a signal that can be detected above instrumental noise. Usually, a signal-to-noise ratio (S=N) of at least 3 is required for a reportable result.

1.3.5 QUANTIFICATION LIMIT The quantiﬁcation limit, also called the quantitation limit or minimum quantiﬁable value, is the lowest analyte level that can be measured with accuracy and precision. If not determined by experiment, it is often set as the concentration of analyte that leads to S=N ¼ 10 [15].

1.3.6 RANGE The range of a method is the extent of concentrations within which accuracy and precision are retained and the relationship between concentration and response is constant. This frequently means that the calibration curve is linear between the lower and the upper concentration limits. Any results outside of the range would be invalid. The lower end of the calibration range is often the quantiﬁcation limit. A recent example of calibration and use of fundamental criteria is illustrated by a capillary electrophoretic study of olive oil by Carrasco-Pancorbo et al. [16]. Testing seven different analytes, they obtained linear calibration curves with the format of Equation 1.1, and determined recovery of other compounds by standard addition. They also calculated accuracy, internal precision (both intraday and interday), external precision, detection limit, quantiﬁcation limit, and calibration range.

1.4 OTHER CONSIDERATIONS 1.4.1 ROBUSTNESS When choosing a technique to be used for a particular analyte, the scientist should take into account the fundamental characteristics listed above [17]. Three other criteria, robustness, speed, and convenience, are also important. A technique demonstrates robustness or ruggedness if small changes in pH, volume used in the analysis, or other parameters are within a speciﬁed tolerance [3,15]. As an example, Lai et al. [18] included instrumental drift, ambient temperature, and sample aging as robustness factors in their development of Fourier transform infrared spectroscopy for detection of vegetable oil adulteration. Their procedure was robust because minor changes in these variables did not signiﬁcantly affect the accuracy and precision.

Otles/Handbook of Food Analysis Instruments 45660_C001 Final Proof page 5 14.7.2008 8:54am Compositor Name: BMani

Data Analysis Techniques

5

1.4.2 SPEED A necessary characteristic when choosing an analytical procedure is the amount of time required to complete it. Speed of analysis is less critical in a research laboratory, where relatively few samples are to be run, than in a continuous industrial process where timeliness is urgent. However, a time lag may be required between samples or groups of samples so that possible corrective actions may be taken.

1.4.3 CONVENIENCE Aspects of convenience include cost of purchasing and operating instruments, their availability when more than one analyst uses them, the sample size required, reagent stability and preparation time, stafﬁng needs, and ease of performing the analytical method. Automatic sampling, if available, allows for unattended analyses and enables employees to attend to two or more tasks simultaneously.

1.5 SUMMARY The results of an analytical study are only as good as the data used, but the data are only as good as the thoroughness displayed by the analyst in performing the measurements and minimizing errors. An analysis always takes less time to do once properly than to do over again because of carelessness or excessive speed.

REFERENCES 1. Elving, P.J. and Keinitz, H., Methodology of analytical chemistry, in Treatise on Analytical Chemistry, Vol. 1, 2nd ed., Kolthoff, I.M. and Elving, P.J. (Eds.), John Wiley & Sons, New York, 1978, p. 53. 2. Brereton, R.G., Statistical assessment of results of food analysis, in Methods of Analysis of Food Components and Additives, Ötles¸, S. (Ed.), Taylor & Francis, Boca Raton, FL, 2005, Chap. 2. 3. Currie, L.A., Nomenclature in evaluation of analytical methods including detection and quantiﬁcation capabilities, Pure Appl. Chem., 67, 1699, 1995. 4. Mitchell, D.G. et al., Multiple-curve procedure for improving precision with calibration-curve-based analyses, Anal. Chem., 49, 1655, 1977. 5. Cardone, M.J., Palermo, P.J., and Sybrandt, L.B., Potential error in single-point-ratio calibrations based on linear calibration curves with a signiﬁcant intercept, Anal. Chem., 52, 1187, 1980. 6. Centner, V., Massart, D.L., and de Jong, S., Inverse calibration predicts better than classical calibration, Fresenius J. Anal. Chem., 361, 2, 1998. 7. Grientschnig, D., Relation between prediction errors of inverse and classical calibration, Fresenius J. Anal. Chem., 367, 497, 2000. 8. Saxberg, B.E.H. and Kowalski, B.R., Generalized standard addition method, Anal. Chem., 51, 1031, 1979. 9. Kalivas, J.H., Precision and stability for the generalized standard addition method, Anal. Chem., 55, 565, 1983. 10. Álvarez del Pino, M. et al., Comparison of biological and chemical methods, and internal and external standards, for assaying tannins in Spanish shrub species, J. Sci. Food Agric., 85, 583, 2005. 11. Vessman, J. et al., Selectivity in analytical chemistry, Pure Appl. Chem., 73, 1381, 2001. 12. ISO Standard 3534–1993, Statistics—Vocabulary and Symbols, International Organization for Standardization, Geneva, Switzerland, 1993. 13. Cheng, Z., Moore, J., and Yu, L., High-throughput DPPH radical scavenging capacity assay, J. Agric. Food Chem., 54, 7429, 2006. 14. Skoog, D.A. and West, D.M., in Principles of Instrumental Analysis, 2nd ed., Saunders College, Philadelphia, PA, 1980, p. 68. 15. Green, J.M., A practical guide to analytical method validation, Anal. Chem., 68, 305A, 1996.

Otles/Handbook of Food Analysis Instruments 45660_C001 Final Proof page 6 14.7.2008 8:54am Compositor Name: BMani

6

Handbook of Food Analysis Instruments

16. Carrasco-Pancorbo, A. et al., Rapid quantiﬁcation of the phenolic fraction of Spanish virgin olive oils by capillary electrophoresis with UV detection, J. Agric. Food Chem., 54, 7984, 2006. 17. Tunick, M.H., Selection of techniques used in food analysis, in Methods of Analysis of Food Components and Additives, Ötles¸, S. (Ed.), Taylor & Francis, Boca Raton, FL, 2005, Chap. 1. 18. Lai, Y.W., Kemsley, E.K., and Wilson, R.H., Potential of Fourier transform infrared spectroscopy for the authentication of vegetable oils, J. Agric. Food Chem., 42, 1154, 1992.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 7 17.7.2008 6:49pm Compositor Name: VBalamugundan

Methods 2 Microextraction in Food Analysis Kathy Ridgway, Sam P.D. Lalljie, and Roger M. Smith CONTENTS 2.1 2.2 2.3

Introduction .............................................................................................................................. 7 Theory ...................................................................................................................................... 8 Liquid-Phase Microextraction ................................................................................................. 9 2.3.1 Headspace Single-Drop Microextraction .................................................................. 10 2.4 Solid-Phase Extraction ........................................................................................................... 10 2.4.1 Selective Sorbents in SPE ......................................................................................... 12 2.4.1.1 Restricted Access Media ............................................................................ 12 2.4.1.2 Immunosorbents ......................................................................................... 13 2.4.1.3 Molecularly Imprinted Polymers ............................................................... 14 2.5 Solid-Phase Microextraction .................................................................................................. 14 2.5.1 In-Tube SPME .......................................................................................................... 17 2.6 Stir Bar Sorptive Extraction .................................................................................................. 17 2.7 Summary ................................................................................................................................ 18 References ....................................................................................................................................... 20

2.1 INTRODUCTION The accurate determination of food components or residues and contaminants in food is necessary to ensure both the quality and safety of products to consumers. Methods of analysis must always be robust and accurate, but there is also the increasing demand on reducing the time spent on sample preparation and on using more environmentally friendly techniques that use smaller volumes of organic solvents. Minimizing the number of steps in an analytical method results in a reduction not only in time but also in potential sources of error. Ease of automation of techniques is also becoming increasingly important to provide more robust and less labor intensive methods. Food covers a wide range of materials, from solids such as cheese, viscous mixtures such as yogurt, and liquids, including wines and other drinks. The materials may be of natural plant or animal origin and be processed or manufactured. Within this wide range of complex matrices, the analytes may be present at high levels, typical for carbohydrates or fats, or be residues or contaminants at trace levels. As a result, the methods of analysis often need to include extensive sample preparation before instrumental analysis to remove potential interferents, by separating the components of interest from unwanted matrix constituents, or to concentrate the analytes to enable detection at the low levels required. To achieve these goals, a range of extraction and separation techniques have been employed to fractionate the sample, sometimes by a physical separation of vapors or liquids from solid materials but more frequently by employing a comprehensive or selective solvent extraction technique. 7

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 8 17.7.2008 6:49pm Compositor Name: VBalamugundan

8

Handbook of Food Analysis Instruments

Traditionally, concentration of the sample was achieved by evaporation of the extraction solvent until an analyte level suitable for instrumental analysis was obtained. This is wasteful both in the time of the analyst and in solvent usage. This chapter will focus on those sample preparation techniques that provide microextraction=separation and concentration steps resulting in a ﬁnal extract ready for instrumental analysis. In particular, it will examine techniques, such as liquid-phase microextraction (LPME), solid-phase extraction (SPE), and solid-phase microextraction (SPME), which can reduce the time spent on sample preparation and achieve the high concentration factors required for the determination of trace level components, residues, or contaminants in food. Some of the techniques follow an initial extraction stage and can be considered as cleanup=enrichment methods, whereas others offer combined extraction and enrichment in a single step. Another technique that can be utilized to sample volatile analytes in food is headspace analysis. By sampling the headspace above a solid or liquid sample, usually after agitation and heating, a representative proportion of the volatile compounds are separated from the nonvolatile components, which remain in the sample matrix. Although direct static headspace will not be covered in this chapter, some of the techniques can be used to selectively extract analytes from the headspace above food samples, providing not only highly selective extraction, but also in some cases sample enrichment=concentration.

2.2 THEORY In all extraction techniques, the transfer of analytes into the extracting phase (whether a gas, liquid, or solid) is dependent upon the chemical properties of the analytes. Key parameters that must be considered include volatility, solubility (hydrophobicity), molecular weight, ionizability (pKa), and polarity. The analyte is distributed between two immiscible phases and the distribution can be described in terms of equilibrium between these phases. XB

XA

Therefore, using the Nernst distribution law, the distribution=partition coefﬁcient (KD) can be deﬁned as KD ¼

[X]A [X]B

where [X] represents the concentration in each phase at constant temperature (or more accurately, the activity of the analyte in each phase). Usually, the total amount of all forms of the analyte present in each phase at equilibrium is considered. If KD is large, almost all the analyte is transferred into the extracting phase A in each extraction stage and extraction would be considered complete after two or three steps in traditional solvent (liquid–liquid) extraction. A good guide to the distribution coefﬁcient is the n-octanol=water partition coefﬁcient Kow (also referred to as Pow or P) which is a measure of hydrophobicity (the compound’s reluctance to enter a water phase). Kow ¼ KD ¼

[X]o [X]w

A compound with a larger value of Kow is said to be more hydrophobic and will be easier to extract from water using an immiscible, usually organic, phase. Values of Kow are often reported on a

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 9 17.7.2008 6:49pm Compositor Name: VBalamugundan

Microextraction Methods in Food Analysis

9

logarithmic scale (log Kow or log P) and generally a compound with a log P of 3 or above is considered as highly hydrophobic. In contrast, high water solubility is generally characterized by low hydrophobicity. When considering the distribution of an ionizable analyte between phases, the acid dissociation constant (pKa) of a compound must also be considered and the pH at which an extraction is performed can be a key parameter. Thus the efﬁciency of any extraction depends on the distribution ratio of the analyte between the phases and on the volume of each phase. If a large volume of extraction solvent is needed as in conventional liquid–liquid extraction, the extraction solvent may need to be evaporated in an extra step. To achieve a high concentration factor, ideally all the analyte of interest from a large volume of sample should be extracted into a small volume of extracting phase. The extracting phase can be in the form of a free liquid or a solid-supported liquid phase and the same principles of partition between the sample and the liquid-extraction phase apply. In the latter, the liquid-extraction phase can be coated on a ﬁber or a solid surface. Alternatively, the extraction phase can be a solid sorbent where extraction is based on the interaction at the surface only (adsorption). This method is employed in some SPE methods, where the sample or solution is passed through a column and analytes can be exhaustively extracted onto the sorbent. Some methods are not intended to provide exhaustive extraction and are optimized at the equilibrium point of the phase distribution. These include SPME and stir bar sorptive extraction (SBSE) and can involve mechanisms based on both partitioning, where the analytes are partitioned into the matrix and are retained in the bulk phase as in liquid–liquid extraction, and adsorption, where the analyte concentrates onto the surface only. Different mechanisms can be employed depending on the analytes of interest and conditions must be optimized for each application and different food matrix.

2.3 LIQUID-PHASE MICROEXTRACTION Miniaturized versions of liquid–liquid extraction have been devised, including LPME, in which the analyte partitions between the bulk aqueous phase and a very small volume of organic solvent. The extraction can be performed in different modes, including static, dynamic, and headspace LPMEs. This technique was ﬁrst introduced in 1996 [1] and was subsequently reviewed by Wood et al. [2] and by Psillakis and Kalogerakis [3], including a useful comparison with SBSE and SPME (discussed later). Recent developments use only a single droplet of the extraction solvent (single-drop microextraction, SDME), which is suspended at the tip of a needle and exposed to the sample solution. As the extracting phase in this approach is typically only microliters of solvent, large concentration factors are possible even with relatively small sample sizes (a few milliliters). Although the method has potential for liquid samples, reported food applications are limited, but Zhao et al. [4] recently reported an SDME for the analysis of organophosphorus pesticides in orange juice. However, the droplet can only be used with care because it is not rigidly held in position. To overcome this problem, a porous hollow ﬁber membrane can be used to support the organic solvent during the extraction from the aqueous sample. This approach has been reviewed by Rasmussen and Pedersen-Bjergaard [5]. The ﬁber allows the use of vigorous stirring or agitation without loss of the microextract (as can occur in droplet LPME) and as a fresh hollow ﬁber can be used for each extraction, any carryover is avoided. The hollow ﬁber, because of the pores in its walls, also shows some selectivity, preventing the extraction of higher molecular weight materials. This technique has been referred to as hollow ﬁber protected liquid-phase microextraction (HF-LPME). Food applications are limited although LPME using a hollow ﬁber membrane was used for the determination of ochratoxin A in wine [6]. The technique has also been applied to human breast milk [7] and bovine milk [8] (Figure 2.1), but centrifugation of the samples before extraction was necessary to improve analyte extractability. Low recoveries were obtained due to strong analyte interactions with the matrix.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 10

10

17.7.2008 6:49pm Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments 1

3

0

10

4 5

5 mAUFS

2

20 Minutes

FIGURE 2.1 Pesticides of phenoxy acid herbicides extracted from milk sample spiked at 10 ng=mL. Extraction conditions: 1-octanol as the impregnation solvent, 0.5 M HCl in donor phase, 0.1 M NaOH in acceptor phase, extracted for 60 min at 1250 rev=min peak identiﬁcation: (1) 2,4-DCBA (2,4-dichlorobenzoic acid), (2) 2,4-D (2,4-dichlorophenoxyacteic acid), (3) mecoprop (2-(4-chlorophenoxy)-2-methylpropionic acid), (4) 3,5-DCBA (3,5-dichlorobenzoic acid), (5) fenoprop (2-(2,4,5-trichlorophenoxy) propionic acid), mAUFS (milli Absorbance Units full scale). (Reproduced from Zhu, L., Huey Ee, K., Zhao, L., and Lee, H.K., J. Chromatogr. A, 963, 335, 2002. With permission.)

Automation of SDME for liquid samples is difﬁcult, and any agitation of the sample must be carefully controlled to avoid loss of the extracting solvent. However, some manufacturers now provide membrane inserts for in-vial extraction that could make full automation of HF-LPME possible for some samples.

2.3.1 HEADSPACE SINGLE-DROP MICROEXTRACTION In a similar way to the use of droplet liquid–liquid extraction, a single drop of solvent suspended from the tip of a syringe can be used to extract the headspace of a sample [2]. This technique has been used for residual solvent analysis, such as aromatic hydrocarbon and chlorinated solvents in edible oils and pharmaceutical products [9]. The extraction solvent must have a boiling point which is high enough to avoid evaporation during sampling. The use of an internal standard is recommended if the method is performed manually. An automated method has been reported as reasonably robust, although there was some evidence that the extracting drop had fallen off the needle on a few occasions. Practical difﬁculties include a limited choice of solvents because of the viscosity that is required, and further work is needed to prove the reproducibility of this technique.

2.4 SOLID-PHASE EXTRACTION SPE involves the partition of analytes between a solid sorbent (extracting phase) usually held in a short column and the sample matrix (liquid phase). To ensure efﬁcient extraction, the afﬁnity of the analytes for the solid phase must be greater than that for the sample matrix. A comprehensive review covering trends, method development, coupling with liquid chromatography, and all types of SPE sorbent was published by Hennion in 1999 [10] and a number of books have looked at the theory and application of the technique in detail [11,12]. Theoretical aspects of SPE are covered in detail by Poole et al. [13], who discussed the use of computer-aided method development and method optimization. The technique usually involves three or four steps, as illustrated in Figure 2.2.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 11

17.7.2008 6:49pm Compositor Name: VBalamugundan

11

Microextraction Methods in Food Analysis Step 1 Conditioning

Solvation of the sorbent to enable interaction with the sample

Step 2 Retention

Step 3 Rinsing

A C B D

A

B

C

D

Sample is applied and the analyte (A) and some interferences (B, C, and D) adsorb to the solid surface

Selective washing to remove interferences/ unwanted compounds

Step 4 Elution

B

A Selective desorption and collection of analytes for analysis

FIGURE 2.2 Solid-phase extraction.

The sample is loaded onto a pretreated column or cartridge ﬁlled with the required sorbent which traps the analytes and allows most of the matrix, usually an aqueous solution, to pass to waste. After a rinse step, the analyte of interest is eluted with a small volume of a suitable solvent giving a concentrated extract and leaving insoluble interferences on the column. SPE is considered to be an exhaustive technique as the retention of analytes on a sorbent is based on chromatographic retentions where all the analyte is removed from the sample (and subsequently eluted). The sorption process must be reversible. The selectivity of trapping and elution can be obtained by adjusting the pH and solvent polarity. The choice of sorbent in the cartridge is dependent on the food matrix and analyte(s) of interest. Numerous sorbent materials are available using different mechanisms for extraction=retention of analytes, including partitioning, adsorption, and ion exchange interactions based on van der Waals, polar=dipole–dipole, hydrogen-bonding, or electrostatic (ion exchange) interactions. Typical materials include silica bonded with nonpolar alkyl chains, especially C18 (octadecyl) and C8 (octyl) groups or polar chains such as hydroxyl and cyano groups. Other support materials are polymeric resins (polystyrene=divinyl benzene copolymer), Florisil (activated magnesium silicate), and polar sorbents, such as alumina, charcoal, and unbonded silica. Ionic functional groups, such as carboxylic acid or amino groups can also be bonded to silica or polymeric supports to create ion exchange sorbents. Some cartridges use mixed-mode sorbents that use both primary and secondary mechanisms for selective retention of analytes and some very speciﬁc selective sorbents have been designed (Section 2.4.1). The most common SPE system is the syringe barrel cartridge, but thin-porous glass ﬁbers, thin-coated glass ﬁbers, PTFE (polytetraﬂuoroethylene) disks embedded with sorbents, and disposable plastic pipette tips ﬁtted with sorbent beds are all available. One of the drawbacks of SPE is that the packing must be uniform to avoid poor efﬁciency and automated systems can have difﬁculties with reproducibility for some sample types. Analyte sorption is dependent on both the sample volume and sorbent mass, and the theoretical aspects have been reviewed [14]. The presence of particulate matter in the sample can affect the sorption process and, in some cases, ﬁltration of the sample before SPE may be necessary. The sample matrix can also affect the ability of the sorbent to extract the analyte owing to competition for retention. Many traditional sorbents, such as C18 silica, are limited in terms of selectivity and insufﬁcient retention of very

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 12

12

17.7.2008 6:49pm Compositor Name: VBalamugundan

2.922—Acrylamide

Handbook of Food Analysis Instruments

0

2

4

FIGURE 2.3 Liquid chromatography=ultraviolet (LC=UV) chromatogram of acrylamide in a French fries extract using Strata-X-C. LC Conditions: Synergi Polar-R 4 m 150 3.0 mm, mobile phase 94:6 (V:V) water: acetonitrile at 0.4 mL=min, injection volume:10 mL. (Reproduced from Peng, L., Farkas, T., Loo, L., Dixon, A., Teuscher, J., and Kallury, K., Rapid and Reproducible Extraction of Acrylamide in French Fries Using a Single SPE Sorbent—Strata-X-C, Phenomenex, Inc., Torrance, CA, 2007. With permission.)

polar compounds can be a problem. The use of hydrophilic materials for the improved extraction of the most polar compounds by SPE was detailed by Fontanals et al. [15]. More recently, as well as the development of more selective sorbents, the use of monolithic columns and multiwalled carbon nanotubes (MWCNTs) has been investigated [16] for the determination of polybrominated diphenyl ethers in water and milk. Typical examples of the use of SPE in food analysis were given in a review in 2002 [17], and include the determination of folic acid in fruit juices [18], and antioxidants in margarine [19]. More recent examples include the determination of amines in beer [20], veterinary drugs in shrimp [21], acrylamide in French fries [22] (Figure 2.3), and heterocyclic amines in meat [23]. In recent years, automation of SPE has become more widely available and several online systems are now available. Although, in most cases, and particularly for solid or semisolid foods, an initial extraction step is required before cleanup=extraction=concentration by SPE. Carbon-based solid phase extraction tubes have been used for extraction from fruit and vegetables (Figure 2.4).

2.4.1 SELECTIVE SORBENTS

IN

SPE

Most SPE methods are based on the trapping of compounds falling into a broad polarity region, but lack speciﬁcity for selected compounds. More speciﬁc extraction media have been developed to either use two mechanisms in conjunction, as in restricted access media (RAM), or to employ biological speciﬁcity (afﬁnity columns) or their synthetic mimics (molecularly imprinted polymers, MIPs) to trap speciﬁc groups of compounds of interest. 2.4.1.1

Restricted Access Media

The RAM sorbents [24] for SPE were developed particularly for the analysis of biological samples, such as plasma and serum, as they are designed to exclude macromolecules, such as proteins, and

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 13

17.7.2008 6:49pm Compositor Name: VBalamugundan

13

Microextraction Methods in Food Analysis Diphenylamine

Methoxychlor

5

10

15

20

25

30

35

40

45

50

55

60

65

70

Minutes

FIGURE 2.4 Extraction of pesticides from homogenized fruit. Sample (50 g) homogenized with acetonitrile (100 mL) and 10 g for 5 min. Following concentration, extracted with ENVI-Carb, 6 mL, 500 mg SPE tube. Pesticides eluted with acetontirile:toluene (3:1) and extract concentrated with acetone. GC column 14% cyanipropylphenyl=86% dimethylsiloxane, 30 m 0.25 mm ID 0.15 mm ﬁlm. Oven 708C (2 min) to 1308C at 258C=min to 2208C at 28C=min to 2808C at 108C=min, held for 4.6 min. Carrier helium, Mass selective detector (MSD) (2858C), injector in splitless (2 mL). (Reproduced from Supelco Web site—Bulletin 900. With permission.)

allow the trapping of smaller drug molecules. They combine the exclusion of proteins and other high-molecular mass matrix components with the simultaneous enrichment of low-molecular mass analytes at the inner pore surface. Macromolecules are excluded either by a physical barrier (pore diameter) or by a chemical diffusion barrier created by a protein network at the outer surface of the particle. Various RAM sorbents are available with different surface chemistries [25]. Internal surface reversed phase (ISRP) supports are the most popular in which a C4-, C8-, or C18-bonded reversed phase covers the internal pore surface of a glyceryl-modiﬁed silica. The interaction sites within the pores are only accessible to small molecules and the analytes are retained by conventional SPE retention mechanisms, such as hydrophobic or electrostatic interactions. Several food applications are given in a review by Souverain et al. [25], including the direct analysis of pharmaceuticals in milk [26] and tissue [27]. 2.4.1.2

Immunosorbents

Molecular recognition can be used to create highly selective immunosorbents by linking an antibody to a solid support (such as silica), which is then packed into an SPE cartridge or precolumn. This technique uses the very speciﬁc interactions between analytes and a biological system to enable the selective retention of the compounds of interest. The analyte can then be released by elution with solvent or a change in pH. The technique is particularly suited to complex biological and environmental samples. The selectivity is based on the antigen–antibody interaction and immunosorbents can be designed for single analytes. Some antibodies can also bind to other analytes with similar structures to the antigen (known as cross-reactivity), and this can be utilized to develop classselective sorbents. One of the major disadvantages of this technique is the need to initially develop the antibody, which makes it impractical for one-off analyses. The analyte–antibody interaction can also be affected by the sample matrix, leading to low extraction recoveries. A review by Hennion and Pichon [28] describes immuno-based extraction sorbents and also the use of artiﬁcial antibodies. Examples of the use of immunosorbents for food analysis include the determination of pesticides (imazalil and phenylurea herbicides) in fruit juices [29,30]. Methods for the analysis of mycotoxins are now commercially available and methods have been developed and accepted as

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 14

14

17.7.2008 6:49pm Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

valid for several food matrices, including peanut butter, roast coffee, and baby food [31]. Immunosorbents have also been developed for some veterinary drugs, such as ﬂuoroquinolones in chicken liver [32] or corticosteroids in animal feed [33]. 2.4.1.3

Molecularly Imprinted Polymers

Attempts have been made to mimic the speciﬁcity of immunological products with synthetic MIPs. MIPs are created by forming a polymer structure containing selected functional groups, around a template analyte molecule. After removal of the template, the polymer contains highly stable cavities with active sites which are speciﬁc to the shape and functionality of the analyte of interest. Trapping and retention of analytes is due to the shape recognition in the cavities and interactions, such as hydrogen bonding, and hydrophobic interactions. MIP-SPE sorbents allow for larger sample volumes to be used than conventional SPE materials because of their selectivity. They can be heated and are stable in both organic solvents and strong acids and bases, unlike many immunosorbents. However, a separate MIP must be made for each analyte, although they can sometimes perform group trapping if all the analytes contain a common structural feature that has formed the active feature of the template. Because of the nature of their selectivity, once developed, MIPs can often be used for a number of matrices, even though the interaction may be different. MIP-SPE can be used both online and off-line. Coupled with HPLC (high-performance liquid chromatography), the MIP can be packed in a cartridge and used via column switching, before analysis [34,35]. One problem encountered with MIPs is that the selective interactions, which were present in the organic solvent in which the template was prepared, do not always work as well in aqueous solutions. However, it is possible to overcome this problem, by initially retaining the sample by a nonselective interaction, then washing the cartridge with an organic solvent to trap the analyte using selective binding. Alternatively, the analyte can be transferred from an aqueous sample or extract into an organic solvent before SPE. MIPs have been used as selective sorbents for a range of analytes and matrices [36–38]. Food applications to date are limited, but include the determination of triazines in liver [39], nicotine in chewing gum [40], and the detection of Sudan I as a contaminant in food matrices [41]. Currently, the time taken to develop and produce such sorbents is the rate limiting step for new and emerging food contaminants.

2.5 SOLID-PHASE MICROEXTRACTION Although SPE and related methods minimize the use of solvents, they still require an elution stage which effectively dilutes the extract. In contrast, SPME is a solvent-free sample preparation technique. It uses a fused silica ﬁber coated with an appropriate stationary phase as the extraction medium attached to a modiﬁed micro-syringe. The sample is usually released by thermal desorption directly into the injection port of a gas chromatography (GC), but can also be released into an HPLC mobile phase. It was originally developed by Arthur and Pawliszyn [42] in 1990 and a number of books are available on this technique [43–45]. The main advantages of SPME are the combination of sampling and extraction into one step and the ability to examine small sample sizes. It can also have high sensitivity and can be used for polar and nonpolar analytes in a wide range of matrices. SPME is essentially a two-step process. Firstly, the partitioning of analytes between the sample matrix, which can be a liquid sample or headspace, and the ﬁber coating, and then the desorption of the (concentrated) extract from the ﬁber into the analytical instrument. The physical and chemical properties of the extracting phase on the ﬁber and the target analyte molecular weight, volatility, and polarity determine the partition coefﬁcient of the analyte between the ﬁber coating and the sample matrix. For high extraction efﬁciency, the polarity of the phase should match that of the analyte and

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 15

Microextraction Methods in Food Analysis

17.7.2008 6:49pm Compositor Name: VBalamugundan

15

the amount of analyte extracted onto the ﬁber depends on the polarity and thickness of the polymer phase, the extraction time, and the concentration of the analyte in the sample. The yield also depends on the properties of the sample matrix, but generally SPME of the analyte from the matrix is not exhaustive. The maximum sensitivity would be obtained when equilibrium is reached; however, extractions can instead be performed for a deﬁned period of time as long as the yield at that time is reproducible [46]. The speed of extraction can be improved by agitation of the matrix and the equilibrium can be altered by the addition of salt or by changing the pH or temperature. A ﬁber with a thicker coating is best to retain volatile analytes and transfer them to the GC injection port without loss, but a thin coating is used to ensure a rapid release of higher boiling point compounds during thermal desorption. Fibers with different thickness and polarities are available and can generally be classiﬁed into two groups: pure liquid polymer coatings, such as PDMS (polydimethylsiloxane) and PA (polyacrylate), and mixed ﬁlms containing liquid polymers and solid particles, such as Carboxen-PDMS and divinylbenzene (DVB)-PDMS. Extraction can be based on absorption (as with liquid phases, such as PDMS), or adsorption on the surface of the polymer (as with more rigid polymeric structures, such as polystyrene-DVB phases). Mixed ﬁlms combine the absorption properties of the liquid polymer with the adsorption properties of the porous particles, but these phases generally have a more limited lifetime. PDMS is strongly hydrophobic and is particularly suitable for extraction from aqueous matrices. It is a commonly used phase with a generic selectivity for many types of nonpolar analytes. PA and Carbowax (CW)-DVB are better for more polar analytes, such as phenols or alcohols. Carboxen acts as a carbon molecular sieve and is often used in combination with PDMS (Carboxen-PDMS) for lowmolecular weight polar analytes. It is generally better than PDMS, but can give a poorer reproducibility and take longer to equilibrate. DVB is a solid polymer, with slightly larger pores than Carboxen and in combination with PDMS is best suited to semi-polar analytes. More recent papers also detail the production of SPME materials with new solgel coatings. These phases are reported to exhibit high thermal stability and tolerance to organic solvents. Other new coating materials include afﬁnity coatings for target analytes and chiral coatings for optically active analytes [46]. Fibers can be reused and manufacturers claim that under most conditions ﬁbers can provide 50–100 extractions. However, in practice, the ﬁbers can be fragile and can either be broken or the coating can be damaged during injection or agitation. Extractable but nonvolatile compounds in the sample can remain on the ﬁber, which can limit the ﬁber’s lifetime and reproducibility. Proteins can also adsorb irreversibly to the ﬁber, changing the ﬁber properties and making it unusable for more than one sample. Problems with batch to batch variation of ﬁber coatings have also been reported. A recent development is that of superelastic SPME where the ﬁber is a metal alloy with elastic properties and can be coated with PDMS=DVB, Carboxen=PDMS, and DVB=Carboxen-PDMS as well as PDMS [47]. This improves the robustness and overcomes problems caused by the ﬁbers breaking due to misalignment with injection ports or in viscous matrices. To achieve the required throughput with multiple samples automated SPME systems can be used. The extraction temperature, time, and sample agitation rate must be optimized for each application and operating conditions must be consistent. Because of matrix effects, quantitation generally requires matrix matched standards or the method of standard additions can be used. The use of an isotopically labelled internal standard should be considered. The analyte concentration can also inﬂuence the extraction. At low concentrations (<50 ppb), the equilibrium is concentration dependent, so changes in sample volume do not affect response. However, at higher concentrations, volumes become signiﬁcant and must be consistent for samples and standards, especially for compounds with high distribution coefﬁcients. The presence of high concentrations of other matrix components in the sample can result in competitive binding and displacement. In some cases, such as alcoholic beverages, because the ethanol is a competing solvent, the levels can alter the distribution constant.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 16

16

17.7.2008 6:50pm Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

Extraction can be performed by direct immersion of the ﬁber into liquid samples, by extraction of the headspace above the sample, or by using a membrane protected ﬁber. Direct immersion SPME into complex matrices, such as many foods can be difﬁcult, as the ﬁber can be damaged and is therefore more suited to semi- or less-volatile analytes in liquid samples or solutions. For dirty samples, the SPME ﬁber can be rinsed to remove interferences after extraction. Headspace sampling is particularly suitable for many complex food matrices, as the nonvolatile components do not come into contact with the ﬁber and the method can be used with both solid and liquid matrices. This mode of extraction is based on the equilibrium between three phases: the matrix, the headspace, and the ﬁber. Effectively the distribution is between the ﬁber and the matrix. Although raising the temperature increases the volatility of the analyte, it may result in less deposition onto the ﬁber as in the headspace—ﬁber equilibrium, the analyte will again favor the vapor phase. Thus headspace SPME offers a different selectivity as it favors the less volatile compounds in contrast to direct headspace, which favors the more volatile components [48]. Equilibrium is reached faster in headspace analysis than direct ﬁber insertion into the liquid matrix because of better mass transfer. Typical analytes include aromas (Figure 2.5), ﬂavors, and fragrance components [48–53]. An alternative for dirty liquid samples is to protect the ﬁber by placing it inside a hollow cellulose membrane. This can have an added size exclusion effect (e.g., only allowing compounds with molecular weight less than 1000 Da to diffuse through the membrane). However, using this technique requires a much longer extraction time [54], and clogging of the membrane would be an issue for many food matrices.

1. Propane

∗2. Pentane

7

3. Pentanol

∗4. Hexanal

9

15 4

11 1 8

2

0

3

5 6

10

12

10

20

16 13 14

5. 2-Hexenal 6. 2-Heptanone ∗7. 2-Heptenal 8. 1-Octen-3-ol 9. 2-Pentylfuran 10. 3-Octen-2-one 11. 2-Octenal 12. 2-Nonenal 13. 2-Decenal 14. trans, cis-2,4-Decadienal 15. trans, trans-2,4-Decadienal 16. 2-Undecenal 17. BHT ∗Indicators of rancidity

17

30

Minutes

FIGURE 2.5 Flavor compounds extracted from rancid corn oil. SPME ﬁber PDMS, 100 mm ﬁlm, headspace sampling (45 min, 408C), desorption 2508C, 1.5 min. GC column SPB-5, 30 m 0.53 mm ID 5 mm ﬁlm. Oven 408C (5 min) to 2208C at 48C=min. Carrier helium, 5 mL=min Flame Ionisation detector (FID) (3008C), injector in splitless (1 min), 2508C. (Reproduced from Supelco Web site—SPME application note. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 17

Microextraction Methods in Food Analysis

17.7.2008 6:50pm Compositor Name: VBalamugundan

17

In general, SPME provides low recoveries because of the small volume of the stationary phase that can be bound to the ﬁber. It needs to be calibrated carefully to achieve accurate and reproducible quantitative measurements. Derivatization can be used to overcome low extraction efﬁciencies for certain volatile, polar, or thermally unstable analytes, but is mainly used to improve chromatographic behavior. It can be performed before, during, or after the extraction procedure. Typical reactions include the conversion of fatty acids to their methyl esters or addition of functional groups, such as pentaﬂuorobenzene, to enhance detection and the ﬁeld has been reviewed by Stashenko and Martinez [55]. Most SPME methods use GC as the instrument, but it is also possible to interface the method with HPLC using solvent desorption, which can take place in a static or dynamic (ﬂowing eluent) mode. The static mode is preferred for more strongly adsorbed analytes, as the ﬁber is soaked in the mobile phase, or another solvent, for a speciﬁed time, before injection. For automation, in-tube SPME devices (Section 2.5.1) are more suited to HPLC applications. Fiber SPME-HPLC can lead to peak broadening if analytes are slow to desorb, but with in-tube devices the analytes are desorbed before injection. Reviews of the application of SPME in food analysis by Kataoka et al. [46] and by Wardencki et al. [56] give many examples of applications and techniques. These include food components, such as volatile aroma compounds and fatty acids [46], and ﬂavor analysis [51] where commonly headspace-SPME is the preferred method. Recent examples include the determination of furan in baby food [57] and formaldehyde in ﬁsh [58].

2.5.1 IN-TUBE SPME An alternative to the externally coated SPME ﬁber is an internally coated capillary, through which the sample ﬂows, or is drawn repeatedly, and analytes are then eluted or desorbed. This technique was developed due to the difﬁculties of trying to interface SPME with HPLC systems [59]. Several in-tube SPME options can be used with LC, which are suitable for automation and can continuously perform extraction, desorption, and injection. The capillary extraction tube is placed between the injection loop and the injection needle of the HPLC autosampler. A disadvantage of in-tube devices is that particles need to be removed from samples before extraction (by ﬁltration or centrifugation). The amount of analyte extracted by the phase depends on the polarity of the capillary coating, the number and volume of extraction (draw=eject) cycles, and the sample pH. As would be expected, target analytes with lower K values need longer equilibration times. If too many extraction cycles are performed peak broadening can occur. A variant of in-tube SPME termed solid-phase dynamic extraction (SPDE) using wall-coated needles was described by Lipinski [60] for the extraction of liquid samples. Using an adapted syringe needle allows for dynamic extraction, providing high concentration factors and a variety of sorbents are now available. Kataoka [61] reviewed automated in-tube SPME, giving applications for both food contaminants and food component analysis. Examples included heterocyclic amines in beefsteak [62] and endocrine disruptors in fatty foods [63].

2.6 STIR BAR SORPTIVE EXTRACTION SBSE was developed by Baltussen et al. [64] to overcome the small extraction medium volumes used in SPME. Instead of a coated ﬁber, a glass stirrer bar is coated with a bonded adsorbent layer (PDMS) to give a larger volume of the stationary phase, which is generally more robust than SPME ﬁbers. Extraction is achieved by sorption onto the PDMS coating and transfer from the stirrer bar to a GC is then achieved either by thermal desorption or, for HPLC, by elution with a solvent. Because of the larger volumes of PDMS on the stir bar, higher concentration factors with longer extraction times can be achieved. The extraction mechanism from aqueous solutions is based on an absorption

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 18

18

17.7.2008 6:50pm Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments 70,000

3 2

60,000

Abundance

50,000 40,000 30,000 20,000

1 4

10,000 0 10.00

12.00 14.00 16.00

18.00

20.00

22.00

24.00

26.00

28.00

30.00

Time (min)

FIGURE 2.6 Extracted ion chromatogram at m=z 187, 283, and 285 of the (SBSE-Thermal desorption GC-MS) analysis of Italian sparkling wine: (1) vinclozolin, (2) procymidone, (3) (3,5-dichlorophenyl)hydantoin, and (4) iprodione. (Reproduced from Bicchi, C., Cordew, C., Iori, C., Rubiolo, P., and Sandra, P., HRC J. High Resolut. Chromatogr., 23, 539, 2000. With permission.)

process and the octanol=water distribution coefﬁcients can be used to predict recoveries. The technique is generally suited to compounds with a log P > 2 (Kow > 100). It can be used directly in liquid or semisolid complex matrices, such as yogurt, and as with SPME, the stir bar can also be used to sample the volatiles and semi-volatiles in the headspace above the sample. Derivatization can again be used to extend the applicability of the technique, and multiple stir bars can be used to improve sensitivity. The technique is commercialized under the name twister. Currently, only a PDMS coating is commercially available, making the technique most suited to nonpolar analytes from aqueous media. Dual-phase stir bars have been described by Bicchi et al. [65], which consist of a short PDMS tube with an inner cavity that is packed with activated carbon adsorbent. This method combines both sorption and adsorption simultaneously and improved the recovery of volatile or polar compounds when compared to conventional PDMS stir bars for the analysis of coffee and sage (by headspace) and whisky (by immersion). The use of SBSE for food analysis is increasing and the technique has been used for the analysis of coffee brew [50], alcoholic beverages [51], the determination of pesticides in fruits [66] and wine [67] (Figure 2.6), and for examining the headspace of aromatic and medicinal plants [68] and food [69].

2.7 SUMMARY Microextraction methods provide highly selective techniques that are either solvent free or use only small volumes of solvent. They can be used to provide robust, accurate methods that are necessary for food analysis to ensure both the quality and safety of products and ingredients. They require much smaller sample sizes than conventional extraction methods and often enable a selective extraction or extract cleanup technique that can provide enrichment of analytes from complex matrices and thus enable detection down to the levels required for food safety and quality. Many of the approaches can be automated to enable the use of high-throughput methods that are simple, reliable, and more environmentally friendly. A summary of food applications of the techniques described in this chapter is given in Table 2.1.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 19

17.7.2008 6:50pm Compositor Name: VBalamugundan

19

Microextraction Methods in Food Analysis

TABLE 2.1 Food Applications of Microextraction Techniques Liquid-Phase Microextraction Food Type Orange juice Wine Human breast milk Bovine milk Beer

Beer Shrimp Animal tissues Meat Milk Tissue Fruit juice Peanut butter, pistachios, ﬁg paste, and paprika Baby food Milk Roast coffee and baby food Apple juice and puree Chicken liver

Analytes Organophosphorus pesticides Ochratoxin A Basic drugs Phenoxy herbicides Alcohols Solid-Phase Extraction Amines Veterinary drugs (multi-class) Fluoroquinolone residues Heterocyclic amines Pharmaceuticals Pharmaceuticals (including nicardipine, nitrendipine, felodipine, and benzodiazepines) Drugs Aﬂatoxins (B1,B2,G1,G2)

Technique

References

SDME HF-LPME HF-LPME HF-LPME Headspace-LPME

[4] [6] [7] [8] [71]

SPE SPE SPE SPE SPE (RAM) SPE (RAM)

[20] [21] [72] [23] [26] [27]

SPE-Immunosorbents SPE-Immunosorbents

[29] [31]

SPE-Immunosorbents SPE-Immunosorbents SPE-Immunosorbents SPE SPE-Immunosorbents

[31] [31] [31] [31] [32]

SPE-Immunosorbents

[33]

SPE-MIPs SPE-MIPs

[39] [40]

Liver Chewing gum

Aﬂatoxin B1 Aﬂatoxin m1 Ochratoxin A Patulin Fluoroquinolones (ciproﬂoxacin, enroﬂoxacin, saraﬂoxacin, and diﬂoxacin) Corticosteroids (dexamethasone, ﬂumethasone, and triamcinolone) Triazines Nicotine

Beef Fatty foods Baby food Fish Beer Cow’s milk Cheese Food Honey Wines

Solid-Phase Microextraction Heterocyclic amines Endocrine disruptors Furan Formaldehyde Volatiles Phthalate esters Volatiles Volatiles Amitraz Off-ﬂavors

In-tube SPME In-tube SPME Headspace-SPME Headspace-SPME Headspace-SPME Headspace-SPME Headspace-SPME Headspace-SPDE Headspace-SPDE Headspace-SPDE

[63] [64] [58] [59] [50] [73] [74] [75] [76] [77]

Coffee Grape juice Sugarcane juice Honey Alcoholic beverages Oranges Plants Food

Stir Bar Sorptive Extraction Aroma proﬁles Volatiles Pesticides and benzo[a]pyrene Pesticides Flavor proﬁle Pesticides Aroma proﬁle Aroma proﬁle

SBSE, HSSE SBSE, SDE SBSE SBSE SBSE, SPME SBSE HSSE HSSE

[51] [78] [79] [80] [52] [67] [69] [70]

Animal feed (and urine)

Note: SDME, single-drop microextraction; HF-LPME, hollow ﬁber protected liquid-phase microextraction; RAM, restricted access media; MIPs, molecularly imprinted polymers; HSSE, headspace sorptive extraction; SDE, steam distillation extraction.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 20

20

17.7.2008 6:50pm Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

REFERENCES 1. Liu, H.H. and Dasgupta, P.K. Analytical chemistry in a drop. Solvent extraction in a microdrop. Analytical Chemistry, 1996, 68(11), 1817–1821. 2. Wood, D.C., Miller, J.M., and Christ, I. Headspace liquid microextraction. LCGC Europe, 2004, November, 573–579. 3. Psillakis, E. and Kalogerakis, N. Developments in liquid-phase microextraction. TrAC, Trends in Analytical Chemistry, 2003, 22(9), 565–574. 4. Zhao, E., Han, L., Jiang, S., Wang, Q., and Zhou, Z. Application of a single-drop microextraction for the analysis of organophosphorus pesticides in juice. Journal of Chromatography A, 2006, 1114(2), 269–273. 5. Rasmussen, K.E. and Pedersen-Bjergaard, S. Developments in hollow ﬁbre-based, liquid-phase microextraction. TrAC, Trends in Analytical Chemistry, 2004, 23(1), 1–10. 6. Gonzalez-Penas, E., Leache, C., Viscarret, M., Perez de Obanos, A., Araguas, C., and Lopez de Cerain, A. Determination of ochratoxin A in wine using liquid-phase microextraction combined with liquid chromatography with ﬂuorescence detection. Journal of Chromatography A, 2004, 1025(2), 163–168. 7. Bjorhovde, A., Halvorsen, T.G., Rasmussen, K.E., and Pedersen-Bjergaard, S. Liquid-phase microextraction of drugs from human breast milk. Analytica Chimica Acta, 2003, 491(2), 155–161. 8. Zhu, L., Huey Ee, K., Zhao, L., and Lee, H.K. Analysis of phenoxy herbicides in bovine milk by means of liquid–liquid–liquid microextraction with a hollow-ﬁber membrane. Journal of Chromatography A, 2002, 963(1–2), 335–343. 9. Michulec, M. and Wardencki, W. The application of single drop extraction technique for chromatographic determination of solvent residues in edible oils and pharmaceutical products. Chromatographia, 2006, 64(3), 191–197. 10. Hennion, M.C. Solid-phase extraction: Method development, sorbents, and coupling with liquid chromatography. Journal of Chromatography A, 1999, 856(1–2), 3–54. 11. Thurman, E.M. and Mills, M.S. Solid Phase Extraction: Principles and Practice. 1998. Chemical Analysis: A Series of Monographs on Analytical Chemistry and Its Applications. John Wiley and Sons Inc., New York. 12. Fritz, J.S. Analytical Solid-Phase Extraction. 1999. Wiley-VCH, New York. 13. Poole, C.F., Gunatilleka, A.D., and Sethuraman, R. Contributions of theory to method development in solid-phase extraction. Journal of Chromatography A, 2000, 885(1–2), 17–39. 14. Mitra, S. (Ed.). Sample Preparation Techniques in Analytical Chemistry. 2003, Vol. 162. Wiley-Interscience, Hoboken, NJ. 15. Fontanals, N., Marce, R.M., and Borrull, F. New hydrophilic materials for solid-phase extraction. TrAC, Trends in Analytical Chemistry, 2005, 24(5), 394–406. 16. Wang, J.X., Jiang, D.Q., Gu, Z.Y., and Yan, X.P. Multiwalled carbon nanotubes coated ﬁbers for solidphase microextraction of polybrominated diphenyl ethers in water and milk samples before gas chromatography with electron-capture detection. Journal of Chromatography A, 2006, 1137(1), 8–14. 17. Buldini, P.L., Ricci, L., and Sharma, J.L. Recent applications of sample preparation techniques in food analysis. Journal of Chromatography A, 2002, 975(1), 47–70. 18. Breithaupt, D.E. Determination of folic acid by ion-pair RP-HPLC in vitamin-fortiﬁed fruit juices after solid-phase extraction. Food Chemistry, 2001, 74(4), 521–525. 19. Gonzalez, M., Gallego, M., and Valcarcel, M. Gas chromatographic ﬂow method for the preconcentration and simultaneous determination of antioxidant and preservative additives in fatty foods. Journal of Chromatography A, 1999, 848(1–2), 529–536. 20. Molins-Legua, C. and Campins, F. Solid phase extraction of amines. Analytica Chimica Acta, 2005, 546(2), 206–220. 21. Li, H., Kijak, P.J., Turnipseed, S.B., and Cui, W. Analysis of veterinary drug residues in shrimp: A multiclass method by liquid chromatography-quadrupole ion trap mass spectrometry. Journal of Chromatography B, 2006, 836(1–2), 22–38. 22. Peng, L., Farkas, T., Loo, L., Dixon, A., Teuscher, J., and Kallury, K. Rapid and Reproducible Extraction of Acrylamide in French Fries Using a Single SPE Sorbent—Strata-X-C. Phenomenex Applications Note TN-007. 2007. Phenomenex, Inc., Torrance, CA.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 21

Microextraction Methods in Food Analysis

17.7.2008 6:50pm Compositor Name: VBalamugundan

21

23. Toribio, F., Moyano, E., Puignou, L., and Galceran, M.T. Comparison of different commercial solid-phase extraction cartridges used to extract heterocyclic amines from a lyophilised meat extract. Journal of Chromatography A, 2000, 880(1–2), 101–112. 24. Desilets, C.P., Rounds, M.A., and Regnier, F.E. Semipermeable-surface reversed-phase media for highperformance liquid chromatography. Journal of Chromatography A, 1991, 544, 25–39. 25. Souverain, S., Rudaz, S., and Veuthey, J.-L. Restricted access materials and large particle supports for on-line sample preparation: An attractive approach for biological ﬂuids analysis. Journal of Chromatography B, 2004, 801(2), 141–156. 26. Blahova, E., Bovanova, L., and Brandsteterova, E. Direct HPLC analysis of trimethoprim in milk. Journal of Liquid Chromatography and Related Technologies, 2001, 24(19), 3027–3035. 27. Heinig, K. and Bucheli, F. Application of column-switching liquid chromatography-tandem mass spectrometry for the determination of pharmaceutical compounds in tissue samples. Journal of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences, 2002, 769(1), 9–26. 28. Hennion, M.C. and Pichon, V. Immuno-based sample preparation for trace analysis. Journal of Chromatography A, 2003, 1000(1–2), 29–52. 29. Watanabe, E., Yoshimura, Y., Yuasa, Y., and Nakazawa, H. Immunoafﬁnity column clean-up for the determination of imazalil in citrus fruits. Analytica Chimica Acta, 2001, 433(2), 199–206. 30. Pichon, V., Krasnova, A.I., and Hennion, M.C. Development and characterization of an immunoafﬁnity solid-phase-extraction sorbent for trace analysis of propanil and related phenylurea herbicides in environmental waters and in beverages. Chromatographia, 2004, 60(Suppl.), S221–S226. 31. Gilbert, J. and Anklam, E. Validation of analytical methods for determining mycotoxins in foodstuffs. TrAC, Trends in Analytical Chemistry, 2002, 21(6–7), 468–486. 32. Holtzapple, C.K., Buckley, S.A., and Stanker, L.H. Immunosorbents coupled on-line with liquid chromatography for the determination of ﬂuoroquinolones in chicken liver. Journal of Agricultural and Food Chemistry, 1999, 47(7), 2963–2968. 33. Stolker, A.A.M., Schwillens, P.L.W.J., van Ginkel, L.A., and Brinkman, U.A.T. Comparison of different liquid chromatography methods for the determination of corticosteroids in biological matrices. Journal of Chromatography A, 2000, 893(1), 55–67. 34. Lanza, F. and Sellergren, B. The application of molecular imprinting technology to solid phase extraction. Chromatographia, 2001, 53(11–12), 599–611. 35. Caro, E., Marce, R.M., Cormack, P.A.G., Sherrington, D.C., and Borrull, F. On-line solid-phase extraction with molecularly imprinted polymers to selectively extract substituted 4-chlorophenols and 4-nitrophenol from water. Journal of Chromatography A, 2003, 995(1–2), 233–238. 36. Stevenson, D. Immuno-afﬁnity solid-phase extraction. Journal of Chromatography B: Biomedical Sciences and Applications, 2000, 745(1), 39–48. 37. Andersson, L.I. Molecular imprinting for drug bioanalysis: A review on the application of imprinted polymers to solid-phase extraction and binding assay. Journal of Chromatography B: Biomedical Sciences and Applications, 2000, 739(1), 163–173. 38. Andersson, L.I. Molecular imprinting: Developments and applications in the analytical chemistry ﬁeld. Journal of Chromatography B: Biomedical Sciences and Applications, 2000, 745(1), 3–13. 39. Muldoon, M.T. and Stanker, L.H. Molecularly imprinted solid phase extraction of atrazine from beef liver extracts. Analytical Chemistry, 1997, 69(5), 803–808. 40. Zander, A., Findlay, P., Renner, T., Sellergren, B., and Swietlow, A. Analysis of nicotine and its oxidation products in nicotine chewing gum by a molecularly imprinted solid-phase extraction. Analytical Chemistry, 1998, 70(15), 3304–3314. 41. Puoci, F., Garreffa, C., Iemma, F., Muzzalupo, R., Spizzirri, U.G., and Picci, N. Molecularly imprinted solid phase extraction for detection of Sudan I in food matrices. Food Chemistry, 2005, 93(2), 349–353. 42. Arthur, C.L. and Pawliszyn, J. Solid phase microextraction with thermal desorption using fused silica optical ﬁbers. Analytical Chemistry, 1990, 62, 2145–2148. 43. Janusz, P. Solid Phase Microextraction: Theory and Practice. 1997. Wiley, New York. 44. Wercinski, S.A.S. Solid Phase Microextraction: A Practical Guide. 1999. Marcel Dekker, New York. 45. Pawliszyn, J. Applications of Solid Phase Microextraction (RSC Chromatography Monographs). 1999, 1st ed. Royal Society of Chemistry, Cambridge, United Kingdom.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 22

22

17.7.2008 6:50pm Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

46. Kataoka, H., Lord, H.L., and Pawliszyn, J. Applications of solid-phase microextraction in food analysis. Journal of Chromatography A, 2000, 880(1–2), 35–62. 47. Falch, I. (Ed.). SPME-metal ﬁbre assemblies. The Reporter (Europe), 2006, 22 Sigma Aldrich, 12. 48. Miralles-Garcia, J., Ducki, S., and Storey, D.M. 2005. ARF05 presentation. Plymouth, United Kingdom. 49. Liu, M., Zeng, Z., and Xiong, B. Preparation of novel solid-phase microextraction ﬁbers by sol-gel technology for headspace solid-phase microextraction-gas chromatographic analysis of aroma compounds in beer. Journal of Chromatography A, 2005, 1065(2), 287–299. 50. Bicchi, C., Iori, C., Rubiolo, P., and Sandra, P. Headspace sorptive extraction (HSSE), stir bar sorptive extraction (SBSE), and solid phase microextraction (SPME) applied to the analysis of roasted arabica coffee and coffee brew. Journal of Agricultural and Food Chemistry, 2002, 50(3), 449–459. 51. Demyttenaere, J.C.R., Sanchez Martinez, J.I., Verhe, R., Sandra, P., and De Kimpe, N. Analysis of volatiles of malt whisky by solid-phase microextraction and stir bar sorptive extraction. Journal of Chromatography A, 2003, 985(1–2), 221–232. 52. Bicchi, C., Cordero, C., and Rubiolo, P. A survey on high-concentration-capability headspace sampling techniques in the analysis of ﬂavors and fragrances. Journal of Chromatographic Science, 2004, 42(8), 402–409. 53. Wilkes, J.G., Conte, E.D., Kim, Y., Holcomb, M., Sutherland, J.B., and Miller, D.W. Sample preparation for the analysis of ﬂavors and off-ﬂavors in foods. Journal of Chromatography A, 2000, 880(1–2), 3–33. 54. Zhang, Z.Y., Poerschmann, J., and Pawliszyn, J. Direct solid phase microextraction of complex aqueous samples with hollow ﬁbre membrane protection. Analytical Communications, 1996, 33(7), 219–221. 55. Stashenko, E.E. and Martinez, J.R. Derivatization and solid-phase microextraction. TrAC, Trends in Analytical Chemistry, 2004, 23(8), 553–561. 56. Wardencki, W., Michulec, M., and Curylo, J. A review of theoretical and practical aspects of solid-phase microextraction in food analysis. International Journal of Food Science and Technology, 2004, 39(7), 703–717. 57. Bianchi, F., Careri, M., Mangia, A., and Musci, M. Development and validation of a solid phase microextraction-gas chromatography-mass spectrometry method for the determination of furan in baby-food. Journal of Chromatography A, 2006, 1102(1–2), 268–272. 58. Bianchi, F., Careri, M., Musci, M., and Mangia, A. Fish and food safety: Determination of formaldehyde in 12 ﬁsh species by SPME extraction and GC-MS analysis. Food Chemistry, 2007, 100(3), 1049–1053. 59. Lord, H. and Pawliszyn, J. Microextraction of drugs. Journal of Chromatography A, 2000, 902(1), 17–63. 60. Lipinski, J. Automated solid phase dynamic extraction—Extraction of organics using a wall coated syringe needle. Fresenius Journal of Analytical Chemistry, 2001, 369(1), 57–62. 61. Kataoka, H. Automated sample preparation using in-tube solid-phase microextraction and its application—A review. Analytical and Bioanalytical Chemistry, 2002, 373(1–2), 31–45. 62. Kataoka, H. and Pawliszyn, J. Development of in-tube solid-phase microextraction=liquid chromatography=electrospray ionization mass spectrometry for the analysis of mutagenic heterocyclic amines. Chromatographia, 1999, 50(9–10), 532–538. 63. Kataoka, H., Ise, M., and Narimatsu, S. Automated on-line in-tube solid-phase microextraction coupled with high performance liquid chromatography for the analysis of bisphenol A, alkylphenols, and phthalate esters in food contacted with plastics. Journal of Separation Science, 2002, 25(1–2), 77–85. 64. Baltussen, E., Sandra, P., David, F., and Cramers, C. Stir bar sorptive extraction (SBSE), a novel extraction technique for aqueous samples: Theory and principles. Journal of Microcolumn Separations, 1999, 11(10), 737–747. 65. Bicchi, C., Cordero, C., Liberto, E., Rubiolo, P., Sgorbini, B., David, F., and Sandra, P. Dual-phase twisters: A new approach to headspace sorptive extraction and stir bar sorptive extraction. Journal of Chromatography A, 2005, 1094(1–2), 9–16. 66. Blasco, C., Font, G., and Pico, Y. Comparison of microextraction procedures to determine pesticides in oranges by liquid chromatography-mass spectrometry. Journal of Chromatography A, 2002, 970(1–2), 201–212. 67. Sandra, P., Tienpont, B., Vercammen, J., Tredoux, A., Sandra, T., and David, F. Stir bar sorptive extraction applied to the determination of dicarboximide fungicides in wine. Journal of Chromatography A, 2001, 928(1), 117–126.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 23

Microextraction Methods in Food Analysis

17.7.2008 6:50pm Compositor Name: VBalamugundan

23

68. Bicchi, C., Cordero, C., Iori, C., Rubiolo, P., and Sandra, P. Headspace sorptive extraction (HSSE) in the headspace analysis of aromatic and medicinal plants. HRC Journal of High Resolution Chromatography, 2000, 23(9), 539–546. 69. Tienpont, B., Sandra, P., David, F., and Bicchi, C. High capacity headspace sorptive extraction. Journal of Microcolumn Separations, 2000, 12(11), 577–584. 70 Posyniak, A., Zmudzki, J., and Semeniuk, S. Effects of the matrix and sample preparation on the determination of ﬂuoroquinolone residues in animal tissues. Journal of Chromatography A, 2001, 914, 89–94. 71. Tankeviciute, A., Kazlauskas, R., and Vickackaite, V. Headspace extraction of alcohols into a single drop. Analyst, 2001, 126(10), 1674–1677. 72. Feng, Y.L., Zhu, J., and Sensenstein, R. Development of a headspace solid-phase microextraction method combined with gas chromatography mass spectrometry for the determination of phthalate esters in cow milk. Analytica Chimica Acta, 2005, 538(1–2), 41–48. 73. Mallia, S., Fernandez-Garcia, E., and Olivier Bosset, J. Comparison of purge and trap and solid phase microextraction techniques for studying the volatile aroma compounds of three European PDO hard cheeses. International Dairy Journal, 2005, 15(6–9), 741–758. 74. Bicchi, C., Cordero, C., Liberto, E., Rubiolo, P., and Sgorbini, B. Automated headspace solid-phase dynamic extraction to analyse the volatile fraction of food matrices. Journal of Chromatography A, 2004, 1024(1–2), 217–226. 75. Hahn, H., Nothhelfer, A., and Preuss, S. Determination of amitraz in honey by SPDE-GC-MS=MS. 2003. Chromtech Application Note 301. Chromtech, Idstein, Germany. 76. Chokshi, K. and Christ, I. Comparative SPDE and SPME Studies for Analysis of Off-ﬂavors in Wines. 2006. Chromsys LLC, Alexandria, VA. http:==www.chromsys.com=Applications=wine.htm. 77. Caven-Quantrill, D.J. and Buglass, A.J. Comparison of micro-scale simultaneous distillation-extraction and stir bar sorptive extraction for the determination of volatile organic constituents of grape juice. Journal of Chromatography A, 2006, 1117(2), 121–131. 78. Zuin, V.G., Schellin, M., Montero, L., Yariwake, J.H., Augusto, F., and Popp, P. Comparison of stir bar sorptive extraction and membrane-assisted solvent extraction as enrichment techniques for the determination of pesticide and benzo[a]pyrene residues in Brazilian sugarcane juice. Journal of Chromatography A, 2006, 1114(2), 180–187. 79. Blasco, C., Fernandez, M., Pico, Y., and Font, G. Comparison of solid-phase microextraction and stir bar sorptive extraction for determining six organophosphorus insecticides in honey by liquid chromatographymass spectrometry. Journal of Chromatography A, 2004, 1030(1–2), 77–85.

Otles/Handbook of Food Analysis Instruments 45660_C002 Final Proof page 24

17.7.2008 6:50pm Compositor Name: VBalamugundan

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 25

17.7.2008 3:20pm Compositor Name: VAmoudavally

Fluid Extraction 3 Supercritical in Food Analysis Ruhan Askin, Motonobu Goto, and Mitsuru Sasaki CONTENTS 3.1 3.2

Introduction .......................................................................................................................... 25 Basic Deﬁnitions .................................................................................................................. 26 3.2.1 Critical Temperature ................................................................................................ 26 3.2.2 Critical Pressure ....................................................................................................... 26 3.2.3 Critical Point ............................................................................................................ 26 3.2.4 Supercritical Fluid ................................................................................................... 26 3.2.5 Reduced Temperature .............................................................................................. 26 3.2.6 Reduced Pressure .................................................................................................... 26 3.2.7 Supercritical Fluid Extraction .................................................................................. 27 3.2.8 Coupled Supercritical Fluid Extraction–Supercritical Fluid Chromatography ....... 27 3.2.9 Cosolvent (Modiﬁer) ............................................................................................... 27 3.3 Supercritical Fluids .............................................................................................................. 27 3.3.1 Background and Historical Perspective .................................................................. 27 3.3.2 Basic Properties and Fundamentals of Supercritical Fluids .................................... 27 3.3.2.1 Phase Transitions...................................................................................... 28 3.3.2.2 Phase Behavior ......................................................................................... 29 3.3.2.3 Solvent Strength ....................................................................................... 31 3.3.2.4 Dispersions in Supercritical Fluids .......................................................... 32 3.3.2.5 Solubility in Supercritical Fluids ............................................................. 32 3.3.2.6 Extraction with Supercritical Fluids ......................................................... 33 3.4 Supercritical Fluid Extraction Mechanism .......................................................................... 36 3.5 Supercritical Fluid Extraction Theory ................................................................................. 37 3.6 Experimental Considerations ............................................................................................... 37 3.7 Applications and Commercial Processes of Supercritical Fluids ........................................ 39 3.7.1 Pharmaceutical Applications ................................................................................... 41 3.7.2 Environmental Applications .................................................................................... 42 3.7.3 Food Applications ................................................................................................... 42 3.7.4 Supercritical Fluid Chromatography ....................................................................... 43 3.8 Instrumentation .................................................................................................................... 45 3.9 Current Trends and Future Expects of Supercritical Fluids ................................................ 51 3.10 Conclusion ........................................................................................................................... 53 References ....................................................................................................................................... 54

3.1 INTRODUCTION This chapter is an overview of the current state of the science and technology of supercritical ﬂuids. The principal objective is to acquaint the reader with the unusual properties of supercritical ﬂuids, and 25

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 26

26

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

with the ways some basic principles are essential in understanding the supercritical ﬂuid extraction (SFE) technique and the independence of relevant process parameters that are exploited for a variety of applications in cases of both SFE and supercritical ﬂuid chromatography (SFC) in the food industry. The unusual solvent properties of supercritical ﬂuids, together with their thermodynamic behavior near a critical point, are explained within the framework of ﬂuid-phase diagrams. Characterizing the behavior of supercritical ﬂuids still offers many challenges to scientists. Engineers have exploited the peculiarities of supercritical ﬂuids to great advantage to design new instruments and processes. These proceedings are an example of the dialogue between scientists and engineers that is needed to deepen the understanding of this interesting medium and to widen the ﬁeld of applications. The basic philosophy of utilization is centered on the fact that the properties of supercritical ﬂuids can be varied from gas-like to liquid-like values by simply adjusting the pressure. These ﬂuids are therefore very attractive as tunable process solvents or reaction media. To summarize, in this chapter the basic knowledge and terminology required for understanding supercritical ﬂuid applications including SFE together with SFC are introduced at an elementary level.

3.2 BASIC DEFINITIONS 3.2.1 CRITICAL TEMPERATURE The critical temperature (Tc) is the maximum temperature at the critical point at which a gas can be converted into a liquid by an increase in pressure.

3.2.2 CRITICAL PRESSURE The critical pressure (Pc) is the minimum pressure that would sufﬁce to liquefy a substance at its critical temperature. Above the critical pressure, increasing the temperature will not cause a ﬂuid to vaporize to give a two-phase system.

3.2.3 CRITICAL POINT The characteristic temperature (Tc) and pressure (Pc) above which a gas cannot be liqueﬁed.

3.2.4 SUPERCRITICAL FLUID The deﬁned state of a compound, mixture, or element above its critical pressure (Pc) and critical temperature (Tc). It is a gas-like, compressible ﬂuid that takes a shape of its container and ﬁlls it. It is not a liquid but has liquid-like densities (0.1–1 g=mL) and solvating power.

3.2.5 REDUCED TEMPERATURE The reduced temperature (Tr) is the ratio of the temperature (T ) in the system to the critical temperature (Tc). Tr ¼ T=Tc

(3:1)

3.2.6 REDUCED PRESSURE The reduced pressure (Pr) is the ratio of the pressure in the system (P) to the critical pressure (Pc). Pr ¼ P=Pc

(3:2)

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 27

17.7.2008 3:20pm Compositor Name: VAmoudavally

27

Supercritical Fluid Extraction in Food Analysis

3.2.7 SUPERCRITICAL FLUID EXTRACTION Extraction of a material using a supercritical ﬂuid. The extracted material is usually recovered by reducing the pressure or increasing the temperature of the extraction ﬂuid and allowing the volatile components of the mobile phase to evaporate. Instrumentally, supercritical ﬂuid extraction can use many of the components of a supercritical ﬂuid chromatographic system. It can be used either as an online sample introduction method for a chromatographic separation or as an ofﬂine sample preparation method.

3.2.8 COUPLED SUPERCRITICAL FLUID EXTRACTION–SUPERCRITICAL FLUID CHROMATOGRAPHY In this system a sample is extracted with a supercritical ﬂuid, which then places the extracted material in the inlet port of a supercritical ﬂuid chromatographic system. The extract is then chromatographed directly using a supercritical ﬂuid.

3.2.9 COSOLVENT (MODIFIER) Organic solvents that are used in small quantities in many SFE procedures have become apparent as the technique has matured. These cosolvents are generally used to increase the solubility of the analyte or possibly to increase the separation of co-extractives. Cosolvents such as ethanol have been used to increase the solubility of phospholipids in supercritical carbon dioxide (SCCO2) [1,2]. Performing SFE with cosolvents usually results in a higher weight percent of fat over that recorded with pure CO2.

3.3 SUPERCRITICAL FLUIDS 3.3.1 BACKGROUND

AND

HISTORICAL PERSPECTIVE

In 1822, Baron Charles Cagniard de la Tour discovered the critical point of a substance in his famous cannon barrel experiments. Listening to discontinuities in the sound of a rolling ﬂint ball in a sealed cannon ﬁlled with ﬂuids at various temperatures, he observed the critical temperature. Above this temperature, the densities of the liquid and gas phases become equal and the distinction between them disappears, resulting in a single supercritical ﬂuid phase. Although their unique solvent properties were ﬁrst reported over 100 years ago, only about 20 years ago did supercritical ﬂuids enter the contemporary technical and industrial scene, with the simultaneous appearance in chemical and engineering journals of reports about applications in decaffeinating coffee and tea, extracting hops ﬂavors used in brewing, and extracting aromas and ﬂavors from spices and herbs; by the early 1980s several huge plants (tens to hundreds of millions of pounds per year) were operating in Europe, United States, and Japan. In the intervening years, supercritical ﬂuids have been applied in the development of new or improved products achieving speciﬁcations that cannot be met by other industrial processing methods. In Table 3.1, the critical properties are shown for some components, which are commonly used as supercritical ﬂuids. Supercritical ﬂuid extraction utilizes the ability of certain chemicals to become excellent solvents for certain solutes under a combination of temperature and pressure [3,4]. The term supercritical ﬂuid describes a gas or liquid at conditions above its critical temperature and pressure, i.e., above the critical point.

3.3.2 BASIC PROPERTIES

AND

FUNDAMENTALS

OF

SUPERCRITICAL FLUIDS

Two researchers, Hannay and Hogarth, at a meeting of the Royal Society (London) in 1879, reported that supercritical ﬂuids have a pressure-dependent dissolving power—the higher the pressure, the higher their dissolving power [5]. They described their work and summarized their ﬁndings as follows: ‘‘We have the phenomenon of a solid dissolving in a gas, and when the solid is precipitated by reducing the pressure, it is brought down as a ‘snow’ in the gas.’’ The researchers

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 28

28

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

TABLE 3.1 Critical Properties of Various Solvents Solvent

Pressure

Molecular Weight (g=mol)

Temperature (K)

(MPa)

(bar)

Density (g=cm3)

44.01 18.02 16.04 30.07 44.09 28.05 42.08 32.04 46.07 58.08

304.1 647.3 190.4 305.3 369.8 282.4 364.9 512.6 513.9 508.1

7.38 22.12 4.60 4.87 4.25 5.04 4.60 8.09 6.14 4.70

73.8 221.2 46.0 48.7 42.5 50.4 46.0 80.9 61.4 47.0

0.469 0.348 0.162 0.203 0.217 0.215 0.232 0.272 0.276 0.278

Carbon dioxide Water Methane Ethane Propane Ethylene Propylene Methanol Ethanol Acetone

referred to supercritical ﬂuids as gases, which, in fact, they are. In the interest of brevity, the term ‘‘gas,’’ or the abbreviation ‘‘SCF’’ for supercritical ﬂuids, will be used liberally throughout this chapter. The solubility behavior was not exploited until many, many years later, but it is of historical interest to relate some of the events surrounding their ﬁndings. There arose serious (but, as were the times, polite) controversy at the October 1879 society meeting. Some of the members who were present said, ‘‘Gases cannot dissolve solid compounds. The researchers must have erred and instead found solubility in superheated liquids.’’ In other carefully planned and executed experiments, the researchers did, however, substantiate their previous ﬁndings. Gases, in other words, SCF, could indeed dissolve many compounds. 3.3.2.1

Phase Transitions

Figure 3.1 shows isotherms and typical behavior of a real gas as it is subjected to different pressures and temperatures. It should be noted that there are no phase transitions above Tc. The isotherms

Pc P

T6 T5 T4 T3 T1 Vc

T2

V

Liquid

Liquid and vapor

Gas

SCF

FIGURE 3.1 (See color insert following page 240.)

Tc

Phase diagram for a typical real gas.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 29

17.7.2008 3:20pm Compositor Name: VAmoudavally

29

Supercritical Fluid Extraction in Food Analysis

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

FIGURE 3.2 Change from two deﬁnite phases to one supercritical phase.

shown in the ﬁgure are smooth; they have no tie lines. Tie lines are the horizontal portions of the isotherms, though these are really not really part of the isotherms. In addition, Figure 3.2 shows three photos of the same system. From left to right, the temperature is increasing. In the upper-left photo, there are two phases present, liquid and gas, and the distinction between them is obvious. The following are near the critical temperature, so the separation of the two phases is becoming obscured. In the photo on the bottom-right, there is no phase distinction, so this is above the critical temperature and is a supercritical ﬂuid as it is also shown in Figure 3.3. 3.3.2.2

Phase Behavior

The observations can be explained by looking at the phase diagram of a pure component, e.g., carbon dioxide. Carbon dioxide was substituted for organic solvents (hexane, benzene, carbon tetrachloride, methylene chloride, methanol, and acetone) used in conventional extraction methods. CO2 is probably the most studied SCF as it is nonﬂammable, harmless, noncorrosive, inexpensive, and nontoxic, and it can be obtained with high purity [6]. In the case of carbon dioxide, the critical

(c) Critical point Fluid (b) Pressure

Gas Liquid (a) Gas Liquid Temperature

FIGURE 3.3

Disappearance of the meniscus at the critical point.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 30

30

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

Pc

Liquid

Solid

Pressure

SCF

Critical point Gas Tc Temperature

FIGURE 3.4

Pressure–temperature phase diagram.

point is at 304.06 K and 7.386 MPa. CO2 is the solvent of choice for use in SFE because it is ‘‘GRAS’’— nonﬂammable, noncorrosive, and inexpensive. In addition, CO2 has a low critical temperature, which can help prevent thermal degradation of food components when they are being extracted. In Figures 3.4 and 3.5, two projections of the phase diagram of carbon dioxide are shown. Drawing from physical chemistry texts, the critical point is located at the end of the vapor pressure curve, and Figure 3.4 shows a generalized vapor pressure curve and its end. The accented region in the ﬁgure denotes the supercritical ﬂuid space where many gases exhibit the propensity to dissolve materials. In the pressure-temperature phase diagram, the boiling line, which separates the vapor and liquid region and ends in the critical point, is observed. At the critical point, the densities of the equilibrium liquid-phase and the saturated vapor-phases become equal, resulting in the formation of a single supercritical phase. This can be observed in the density-pressure phase diagram for carbon dioxide, as shown in Figure 3.5, where the critical point is located at 304.1 K and 7.38 MPa (73.8 bar). With increasing temperatures, the liquid-vapor density gap decreases, up to the critical

1000 280 K 300 K

800

Density (kg/m3)

310 K 600

330 K

400 400 K 200

0 40

60

80

100 Pressure (bar)

FIGURE 3.5

Carbon dioxide density–pressure phase diagram.

120

140

160

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 31

17.7.2008 3:20pm Compositor Name: VAmoudavally

31

Supercritical Fluid Extraction in Food Analysis 9 8 Solubility, wt%

7 6 5 4 3 2 1 0

0

50

100

150

200

250

300

Pressure, atm

FIGURE 3.6 Solubility of naphthalene in supercritical carbon dioxide (458C).

temperature, at which the discontinuity disappears. Thus, above the critical temperature, a gas cannot be liqueﬁed by pressure. By deﬁnition, a supercritical ﬂuid is a substance above both its critical temperature and pressure. In a practical sense, the area of interest in supercritical ﬂuids for processing and separation purposes is limited to temperatures in the vicinity of the critical point, where large gradients in the physical properties are observed. The changes near the critical point are not limited to density. Many other physical properties also show large gradients with pressure near the critical point, e.g., viscosity, the relative permittivity, and the solvent strength, which are all closely related to the density. At higher temperatures, the ﬂuid starts to behave like a gas, as can be seen in Figure 3.5. For carbon dioxide at 400 K, the density increases almost linearly with pressure [3]. In a very brief explanation of the technology, supercritical ﬂuids exhibit a pressure-dependent dissolving power, the higher the pressure, the higher the dissolving power, and this property can be applied to puriﬁcation, extraction, fractionation, and recrystallization of a wide host of materials. Being related to such important properties, pressure-dependent dissolving power is illustrated in Figure 3.6, which shows the solubility of a much-studied model compound, naphthalene, in SCCO2. At pressure levels less than the critical pressure of CO2, the solubility of naphthalene is essentially nil, but as the pressure is raised, the solubility increases to quite high levels. Naphthalene solubility has been studied by at least a dozen groups in a variety of gases, and for an interesting historical aside, Büchner, of Nobel Prize fame, was the ﬁrst person to study the solubility of naphthalene in SCCO2 [7]. The Proceedings of the Royal Society (and other journals) describes much of the work during the early years of supercritical ﬂuids activity, and naphthalene is still studied today for the information its solubility behavior presents to new researchers in the SCF ﬁeld [8]. 3.3.2.3

Solvent Strength

The density of a supercritical ﬂuid is extremely sensitive to minor changes in temperature and pressure near the critical point. The density of ﬂuids is closer to that of organic liquids but the solubility of solids can be 3–10 orders of magnitude higher. The enhancement of solubilities was discovered in 1870s for the potassium iodide-ethanol system. The solvent strength of a ﬂuid can be expressed by the solubility parameter, d, which is the square root of the cohesive energy density and is deﬁned rigorously from ﬁrst principles. A plot of the solubility parameter for carbon dioxide versus pressure would resemble a plot of density versus pressure. This conﬁrms that the solvation strength of a supercritical ﬂuid is directly related to the ﬂuid density. Thus the solubility of a solid can be manipulated by making slight changes in temperatures and pressures.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 32

32

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

Another attractive feature of supercritical ﬂuids is that the properties lie between that of gases and liquids. A supercritical ﬂuid has densities similar to that of liquids, while the viscosities and diffusivities are closer to that of gases. Thus, a supercritical ﬂuid can diffuse faster in a solid matrix than a liquid, yet possess a solvent strength to extract the solute from the solid matrix [9]. 3.3.2.4

Dispersions in Supercritical Fluids

The ability to design surfactants for the interface between water (or organics) and supercritical ﬂuids offers new avenues in protein and polymer chemistry, separation science, reaction engineering, waste minimization, and treatment. Surfactant design, which is reasonably well understood for conventional reverse micelles and water-in-oil microemulsions for alkane solvents, is more difﬁcult for carbon dioxide because the properties of carbon dioxide are much different from those of water or nonpolar organic solvents [10]. Carbon dioxide has no dipole moment and weaker van der Waals forces than hydrocarbon solvents. It is possible, however, to form dispersions of either hydrophilic or lipophilic phases in a carbon dioxide continuous phase. Organic-in-carbon dioxide dispersions may be stabilized using surfactants like ﬂuorinated compounds, which are carbon dioxide-philic. 3.3.2.5

Solubility in Supercritical Fluids

According to the ideal gas law, solubility (g) is the ratio of vapor pressure ( pv) to total pressure ( pt) in an SCF; however, the behavior is nonideal and the solubility raises several orders of magnitude. The reason for this increase in the solubility is due to the increase in the density of the SCF. Increase in solubility is deﬁned by the enhancement factor (E ) that is merely the ratio of actual solubility to the solubility predicted by the ideal gas law. (E) ¼ g

pt pv

(3:3)

Solubility for a given solute also depends on the SCF itself. Different supercritical ﬂuids have different solubilizing efﬁciencies. This difference arises due to various intermolecular interactions occurring between the solvent and the solute, which can be explained by the solvent polarity. Here the ‘‘like dissolves like’’ rule applies. Thus, a polar solvent is expected to dissolve a polar solute more efﬁciently than a nonpolar one. Similarly, the structure similarity of both the solvent and solute plays role in the solubility efﬁciency. As an example from typical basic applications, as expressed in previous parts, the solubility of naphthalene in SCCO2 is shown in Figure 3.6. As one would expect, at low pressure its solubility is essentially nil. As the pressure of the gas is increased to above the critical pressure of carbon dioxide (which is 73 atm), the solubility rises, and for many compounds including naphthalene, the rise is often quite dramatic. For example, at 200 atm and 458C, the solubility is 7%. The solubility behavior shown in Figure 3.6 is the basis of almost all the supercritical ﬂuid extraction=separation processes in operation throughout the world: soluble components are extracted from a substrate by a highpressure gas, and the extracted components that have been dissolved in the gas are precipitated from the gas when the pressure is reduced, for example, across a pressure reduction valve. The solubility of components in SCFs can be further enhanced by the addition of a substance referred to as an entrainer, or cosolvent. As volatility of this additional component is usually intermediate to that of the SCF and the solute, the addition of cosolvent provides a further dimension to the range of solvent properties in a given system by inﬂuencing the chemical nature of the ﬂuid. Cosolvents also provide a mechanism by which the extraction selectivity can be manipulated. The commercial potential of a commercial application of SCF technology can be signiﬁcantly improved through the use of cosolvents. A factor that must be taken into consideration when using cosolvents, however, is that even the presence of small amounts of an additional component to a primary SCF can change the critical properties of the resulting mixture considerably.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 33

17.7.2008 3:20pm Compositor Name: VAmoudavally

33

Supercritical Fluid Extraction in Food Analysis

Starting in the 1960s, many research groups, primarily in Europe, and then later in the United States, examined SCFs for developing advanced extraction processes. European researchers emphasized extraction from botanical substrates, for example, spices, herbs, coffee, tea, and so on, using predominantly SCCO2, and by the 1980s there were several large SCF extraction processes in operation in Germany, the United Kingdom, and the United States, for decaffeinating coffee and tea and extracting ﬂavors and essential oils from hops, spices, and herbs. As an example of size, a coffee decaffeination plant in Bremen processes more than 60,000,000 kg=year. The major motivation for developing these SCF processes was the elimination of residual solvents in the products, especially methylene chloride, which had been previously used to decaffeinate coffee. Solvent residues in pharmaceutical and food products were becoming the focus of regulatory attention in the 1970s, and today increasing regulatory attention is being directed to solvent residues. Besides the elimination of solvent residues, there are also other advantages that accrue from employing supercritical ﬂuids in coffee, spices, and herbs, i.e., enhanced ﬂavor and aroma characteristics that cannot be obtained by the traditional organic solvent extraction processes. Besides the enhanced ﬂavor characteristics and frequently higher yields associated with SFE, some other technical and economic advantages reside in the use of carbon dioxide for the extraction of hop ﬂavors. Organic solvents such as methylene chloride or hexane have previously been the solvents used for the extraction of hops [8]. To obtain the concentrated ﬂavors, it was necessary to distill off the organic solvents, and some of the top note aromas are lost during this step. Carbon dioxide produces a superior product because the top notes are not distilled off, and, as mentioned above, the issue of solvent residues, which is a constant spectre, is eliminated by the use of carbon dioxide. 3.3.2.6

Extraction with Supercritical Fluids

The SFE has been applied only recently to sample preparation on an analytical scale. With advances in process, equipment, and product design, and realization of the potentially proﬁtable opportunities in the production of high value-added products, industries are becoming more and more interested in supercritical ﬂuid technology [11]. The extraction is carried out in high-pressure equipment in a batch or continuous manner as depicted in Figures 3.7 and 3.8, respectively. In both cases, the supercritical solvent is put in contact with the material from which a desirable product is to be separated. Supercritical extraction has been applied to a large number of solid matrices. The desired product can be either the extract or the extracted solid itself. This technique resembles Soxhlet extraction except that the solvent used is a supercritical ﬂuid, a substance above its critical

Pumps

FIGURE 3.7 Schematic diagram of an SCF batch extraction.

Separator

Cosolvent

Cosolvent

SCF

Extractor

Gas

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 34

34

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments CO2 + Solute

Pressure reduction CO2

Extracted material

Pumps CO2

CO2

High pressure

Low pressure Solvent recycling

FIGURE 3.8

Schematic diagram of an SCF continuous extraction.

temperature and pressure. This ﬂuid provides a broad range of useful properties [12]. The advantage of using supercritical ﬂuids in extraction is the ease of separation of the extracted solute from the supercritical ﬂuid solvent by simple expansion. In addition, supercritical ﬂuids have liquid-like densities but superior mass transfer characteristics compared to liquid solvents due to their high diffusion and very low surface tension that enables easy penetration into the porous structure of the solid matrix to release the solute. SFE is a relatively new technique in the ﬁeld of analytical chemistry, having evolved in the last decade as an alternative method of preparing samples before analysis. SFE offers to the analysts many advantages that are not inherent in other sample preparation techniques, such as distillation, extraction with liquid solvents, or low resolution liquid chromatography. The most unique property of supercritical ﬂuids for extraction purposes is the ability to adjust their solubilizing power primarily via mechanical compression (and additionally via temperature), thereby providing the possibility of using one supercritical ﬂuid to extract a host of analytes of varying polarity and molecular size [13]. In addition, solute–ﬂuid binary diffusion coefﬁcients are much greater in supercritical ﬂuid media than in liquid–liquid systems, thereby facilitating fast extraction from a variety of sample matrices. Furthermore, several legislative protocols (such as the EPA Pollution Prevention Act in the USA) have focused on advocating a reduction in the use of organic solvents, which could be harmful to the environment. The proper choice of supercritical ﬂuid can also provide speciﬁc advantages when applied in sample workup before analysis. In addition, the extraction rates are enhanced and less degradation of solutes occurs. Several studies have shown that SFE is a replacement method for traditional gravimetric techniques. In addition, carbon dioxide, which is the most adopted supercritical ﬂuid, has low cost, is a nonﬂammable compound and devoid of oxygen, thus protecting lipid samples against any oxidative degradation. For example, the low critical temperature of supercritical CO2 makes it an excellent candidate for extracting thermally labile compounds under conditions slightly above room temperature. In addition, CO2 provides an extraction environment free from molecular oxygen, thereby limiting potential oxidation of the extracted solutes. Supercritical CO2, unlike many liquid extraction solvents, is a nontoxic extraction medium; hence, its use in a laboratory environment can eliminate the cost and problems associated with solvent disposal as well as longterm exposure of laboratory personnel to potential toxic vapors. In practice, SFE can provide appreciable savings in time and cost associated with sample preparation. In general, large polar compounds exhibit almost no solubility in supercritical CO2,

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 35

17.7.2008 3:20pm Compositor Name: VAmoudavally

35

Supercritical Fluid Extraction in Food Analysis

making it an excellent extraction medium for the separation of nonpolar to moderately polar solutes from such matrices as inorganic solids. However, the solubility of polar analytes can be enhanced in many supercritical ﬂuids by the addition of cosolvents, or modiﬁers, at low levels to the dense gaseous-phase. By far, the most widely used extraction ﬂuid has been supercritical CO2; however, the extractability of polar solutes can be improved by using a more polar supercritical ﬂuid. Taking CO2 into consideration, the problem with most of the ﬂuids besides CO2 is that these are either difﬁcult to handle or obtain in a pure form. The following are the advantages of SFEs: 1. Supercritical ﬂuids have a higher diffusion coefﬁcient and lower viscosity than liquids. 2. Absence of surface tension allows for their rapid penetration into the pores of heterogeneous matrices, which helps enhance extraction efﬁciencies. 3. Selectivity during extraction may be manipulated by varying the conditions of temperature and pressure affecting the solubility of the various components in the supercritical ﬂuid. 4. Supercritical ﬂuid extraction does not leave a chemical residue. 5. Supercritical ﬂuid extractions can use carbon dioxide gas, which can be recycled and used again as part of the unit operation. Supercritical carbon dioxide has been researched for potential applications in many different ﬁelds including food=agriculture, analytical=supercritical ﬂuid chromatography, and the petrochemical= chemical industries. Many of the supercritical ﬂuids would not be suitable for practical extractions due to their unfavorable physical properties, costs, or reactivities. For example, ethylene, which exhibits a subambient critical temperature, has been widely investigated in the laboratory as an extractant. However, its ﬂammability limits its application in many analytical problems. Conversely, most polar ﬂuids have high critical temperatures, which can prove destructive to both the analyte and the extraction system. Other ﬂuids, like ﬂuoroform, are unique in their ability to solubilize basic solutes through intermolecular hydrogen bonding in the supercritical ﬂuid state 4, but the exorbitant cost of the ﬂuid limits its use for SFE. It is useful to compare the physical properties exhibited by CO2, under SFE conditions to those associated with liquid solvents under ambient conditions to gain a better understanding of the advantages, which are attendant to conducting extractions in the supercritical ﬂuid state. Table 3.2 compares the physical properties of CO2 under typical SFE conditions with parameters calculated for three liquid solvents: n-hexane, methylene chloride, and methanol at ambient conditions. The density of CO2 at the above conditions is greater than the corresponding value for n-hexane, but lower than the densities exhibited by methanol or methylene chloride. Although density is only an approximate measure of TABLE 3.2 Comparison of Physical Properties of Supercritical CO2 with Liquid Solvents at 258C COa2 Density (g=mL) Kinematic viscosity (m2=s 107) Diffusivity of benzoic acid (m2=s 109) Pv:sat solvent b Pv:sat solute a b

At 200 atm and 558C. Solute is phenol at 258C.

n-Hexane

Methylene Chloride

Methanol

0.746 1.00 6.0

0.660 4.45 4.0

1.326 3.09 2.9

0.791 6.91 1.8

1.4 105

4.2 102

1.2 103

3.6 102

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 36

36

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

intermolecular attraction, the value for CO2 suggests that near liquid-like densities can be achieved for this gas in its supercritical ﬂuid state. Likewise, kinetic-based properties such as viscosity and solute diffusivity, for CO2, have values that are more typical of gases than those of the liquid state. These gas-like transport parameters contribute to improved rates of mass transfer for solutes in supercritical ﬂuid media, resulting in faster extraction. The ratio of the saturated vapor pressures of the extraction solvents to that exhibited by a typical solute, phenol, at 258C is also tabulated in Table 3.2.

3.4 SUPERCRITICAL FLUID EXTRACTION MECHANISM Liquid-solid extraction techniques are widely used for isolation of analytes from a solid matrix. One such technique, solid extraction, involves repeated solvent distillation through a solid sample to remove the analyte of interest. This technique is often used for extracting additives from polymers and organics from soils. Not only does Soxhlet extraction requires the use of an organic solvent that will eventually require disposal but also the technique is sometimes very slow. A relatively new extraction technique for isolation of analytes from solid samples is SFE. It has been considered in some studies that SFE collects great attention and interest because of providing short sample preparation time and being better than the conventional extraction techniques [14]. Extraction of soluble species (solutes) from solid matrices takes place through four different mechanisms: . .

.

.

If there are no interactions between the solute and the solid phase, the process is simple dissolution of the solute in a suitable solvent that does not dissolve the solid matrix. If there are interactions between the solid and the solute, then the extraction process is termed as desorption and the adsorption isotherm of the solute on the solid in presence of the solvent determines the equilibrium. Most solids extraction processes, such as activated carbon regeneration, fall in this category. Third mechanism is swelling of the solid phase by the solvent accompanied by extraction of the entrapped solute through the ﬁrst two mechanisms, such as extraction of pigments or residual solvents from polymeric matrices. Fourth mechanism is reactive extraction where the insoluble solute reacts with the solvent and the reaction products are soluble hence extractable, such as extraction of lignin from cellulose. Extraction is always followed by another separation process where the extracted solute is separated from the solvent.

Another important aspect in supercritical extraction relates to solvent=solute interactions. Normally, the interactions between the solid and the solute determine the ease of extraction, i.e., the strength of the adsorption isotherm is determined by interactions between the adsorbent and the adsorbate. However, when supercritical ﬂuids are used, interactions between the solvent and the solute affect the adsorption characteristics due to large negative partial molar volumes and partial molar enthalpies in supercritical ﬂuids. The thermodynamic parameters that govern the extraction are found to be temperature, pressure, the adsorption equilibrium constant, and the solubility of the organic in supercritical ﬂuid [9]. Similar to the retrograde behavior of solubility in supercritical ﬂuids, the adsorption equilibrium constants can either decrease or increase for an increase in temperature at isobaric conditions. This is primarily due to the large negative partial molar properties of the supercritical ﬂuids. In addition to the above factors, the rate parameters like the external mass transfer resistances, the axial dispersion in the ﬂuid phase, and the effective diffusion of the organics in the pores also play a crucial role in the desorption process. A thorough understanding of these governing parameters is important in the modeling of SFE process and in the design, development, and future scale-up of the process.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 37

17.7.2008 3:20pm Compositor Name: VAmoudavally

Supercritical Fluid Extraction in Food Analysis

37

3.5 SUPERCRITICAL FLUID EXTRACTION THEORY In an effort to understand the parameters inﬂuencing SFE on the analytical scale, many researchers have studied the thermodynamics of solubility in SCFs and extended this knowledge to supercritical ﬂuid extraction (SCFE) models. There has been a wealth of information relating to SCFE in the chemical engineering and physical chemistry literature. Many simple and relatively fast liquid solvent extraction techniques exist, and when such extractions can be conveniently performed, are quantitative, and do not require concentration for the determination of target analytes, SCFE has few apparent advantages other than the reducing solvent usage. The density of SCF and the Hildebrand solubility parameter (d) increases with increasing pressure. The following semiempirical relationship is deﬁned relating the Hildebrand solubility parameter to the density of an SCF: d ¼ 0:47P1=2 c r

(3:4)

Here, r is the density of the SCF, which is related to pressure and temperature. The equation is used to calculate the Hildebrand solubility parameter for various SCFs. The same procedure is used to calculate the Hildebrand solubility parameter for binary ﬂuids. However, the relationship between the solvent strength of a mixed SCF and its density is no longer valid for binary ﬂuids that contain a polar component (modiﬁer). Hildebrand solubility parameters are fairly good predictors of extraction efﬁciency if the sample matrix has no strong adsorption sites. However, if polar analytes are adsorbed onto a polar sample matrix with relatively strong adsorption sites, small amounts of polar modiﬁer will greatly enhance their desorption. Recovery is vastly improved as compared to the use of pure, unmodiﬁed SCF. The effect of temperature on the solid solubility is different at pressures in the critical range or when the system pressure exceeds the critical value by a factor of two or more. Near the system critical pressure, the ﬂuid density is very sensitive to temperature. Therefore, a moderate increase in temperature leads to a large decrease in ﬂuid density with the consequent reduction in solid solubility. At pressures well above the SCF critical pressure, the solute solubility isotherms exhibit a maximum. It was shown that the maximum is achieved when the partial molar volume of the solute in the ﬂuid phase is equal to the solute solid molar volume [15]. A quantitative correlation and prediction of the solubility of a pure solid in a supercritical gas are possible if the fugacity coefﬁcient of the solid in the gas phase can be obtained from an equation of state.

3.6 EXPERIMENTAL CONSIDERATIONS The extraction concept is not difﬁcult and complex to perform. The process is simple, with the major process parameters being temperature, pressure, and ﬂow rate of the supercritical ﬂuid. Figure 3.9 presents a basic ﬂow diagram for SFE. Mainly, to obtain the desired pressure value, a pump is used and the extraction ﬂuid is supplied to the extraction cell that is placed in an electric oven. The temperature is kept in a value above the critical temperature of SCF. In this case, the supercritical solvent is put into contact with the material from which a desirable product is to be separated. During SFE, the supercritical solvent, saturated with the extracted compound, is expanded to the atmospheric conditions and the solubilized product is recovered in the separation vessel permitting the recycle of the supercritical solvent for further use [16]. A schematic diagram for a typical SFE system is illustrated in Figure 3.10. The system is basically formed of a liquid CO2 cylinder, a pump which is preferred mostly to be of syringe type to keep the pressure at an adjusted value [17]. The pressure is kept above the critical pressure and the temperature of the extraction vessel is controlled in supercritical conditions. The extraction process takes place in extraction vessel. After extraction, SCF is passed at a lower pressure and goes through the receiver. CO2, containing droplets and dissolved substances, leaving the receiver is passed through a demister and a carbon scrubber before being recycled to the

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 38

38

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

CO2 cylinder

Condenser

Pump

Heater

Product

Extraction vessel

Raw material

Product

Separation

FIGURE 3.9

Flow diagram of an SFE system.

liquefaction unit [9]. As a following step, separation of desired component from the stream by SFE can be achieved in different ways. One way is to precipitate the solute from the solvent by reduction of the solvent density that is done by reduction in pressure, increase in temperature, or mixing the extract with atmospheric gases like Ar or N2. Sometimes, the product is recovered from the extract by washing it with a suitable solvent. One should consider the importance of some parameters such as density, diffusivity, critical temperature, critical pressure, etc. so that the SCF can be chosen carefully to carry out the extraction process efﬁciently in which the extraction pressure and temperature are kept constant at desired values for desired extraction time of the materials at prepared sample sizes. Pressure gauge

Pressure gauge Controller Controller Vent

Pressure gauge Flow meter

Six port valve

Dry gas meter

Heater

Heater

Heater

Chiller CO2 cylinder

High-pressure pump

FIGURE 3.10

SFE apparatus.

Extractor

Separator

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 39

17.7.2008 3:20pm Compositor Name: VAmoudavally

39

Supercritical Fluid Extraction in Food Analysis

3.7 APPLICATIONS AND COMMERCIAL PROCESSES OF SUPERCRITICAL FLUIDS During the past 20 years, SCF processing has developed from a laboratory scale to commercial processes. Applications of analytical SFE are numerous and continue to focus on fossil fuels and environmental samples, foods, natural products, and polymers. Many of these applications have adopted the advances in SFE previously discussed. In reviewing the SFE application areas, we choose to classify the work by sample type rather than analyte type. The relatively new processes include coffee decaffeination, hops extraction, catalyst regeneration, extraction of organic wastes from water and soil, and SCF chromatography. These applications complement older technologies such as residuum oil supercritical extraction (ROSE) process, propane deasphalting, and reaction processes for the production of polyethylene and primary alcohols in SCF ethylene. Tables 3.3 and 3.4 show a TABLE 3.3 List of Selected SCCO2 Extraction Plants (Europe) Year

Operator

1978 1982 1984 1984

HAG AG SKW-Trostberg Barth Natal Cane By-Products Ltd. SKW-Trostberg MüllerExtract SKW-Trostberg Barth Messer Griesheim SKW-Trostberg HACO AG FRAVEX Jacobs Suchard HAG AG Raps & Co. SKW-Trostberg Barth SKW-Trostberg Barth SKW-Trostberg Agrisana Barth CEA Calchauvet Barth FLAVEX Essences Arkopharma — — — —

1884 1984 1986 1987 1987 1988 1989 1989 1989 1990 1990 1990 1990 1990 1990 1991 1992 1994 1994 1995 1995 1996 1996 1996 1997 2000 2003 2004 2004

Country

Target Material

Extractor Size

Germany=Bremen Germany=Munchsmuster Germany=Wolnzach South Africa=Merebank

Decaffeine=coffee Hop Hop Hop, red pepper

— 6,500 L 3 500 L 1 1,000 L 2

Germany=Munchsmuster Germany=Coburg Germany=Trostberg Germany=Wolnzach Germany=Krefeld Germany=Munchsmuster Germany Switzerland=Gumlingen Germany=Rehlingen Germany=Bremen Germany=Bremen Germany=Kulmbach Germany=Trostberg Germany=Wolnzach Italy=Venafro Germany=Wolnzach Germany=Trostberg Italy=Roseto di gli Abruzzo Germany=Wolnzach France=Pierrelatte France=Grasse Germany=Wolnzach Germany=Rehlingen Italy=San Marzano France=Carros Switzerland France Spain Great Britain

Decaffeine=coffee Coffee — Hop Hop — Decaffeine=tea — Aroma Coffee Decaffeine=coffee Spices Various products Various products Decaffeine=coffee — — Pharmaceuticals — — Pharmaceuticals — Various products — Pharmaceuticals — Pharmaceuticals — Pharmaceuticals

— 100 L 4 200 L 2 4,000 L 4 200 L 2 — 3,000 L 3 3,000 L 3 — 360 L 14 50,000 ton=year 500 L 3 200 L 2 1,000 L 4=4,000 L 2 20,000 ton=year 4,000 L 2 — — 200 L 2 — — 650 L 1 360 L 3 — — 600 L 2 100 L 1 8,300 L 3 100 L 1

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 40

40

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

TABLE 3.4 List of Selected SCCO2 Extraction Plants (Oceania, Asia, Africa) Year

Operator

Country

1980 1984 1984 1986 1989

Australia=Melbourne South Africa=Merebank Japan=Kawasaki Japan=Kawasaki Japan

Hop Hop, red pepper Colorats, ﬂavor Colorats, ﬂavor Essential oil

India=Chennai China

Spice Aroma

300 L 1 300 L 2

India=Mumbai Malaysia Korea China Korea India=Hyderabad India=Mysore China

Spice Spice Ginseng Nutraceuticals Ginseng Spice Spice Various products

100 L 1 — 170 L 1 500 L 2 100 L 1 500 L 2 200 L 2 500 L 3

China

Various products

500 L 2

1999 1999

CUB Foster Natel Canes Fuji Flavor Fuji Flavor Takasago International — Nan Fang Flour Mill Flavex India — IL HWA — KT & G Novotech Agro South East Agro Guangxia Toothpaste Shaanxi Jia De Agric. Eng. Co Shaanxi —

China China

Hop Chinese medicine

500 L 2 500 L 3

2000 2000 2000

RKS Agro-Tech FiveKingCereals Green Tek21

India=Bangalore Taiwan Korea

Spice Rice cleansing Cosmetics

2001 2001

Mori oil and fat Guangxia Toothpaste Guangxia Toothpaste Guangxia — — — — — UMAX — — GansuYasheng Jiusan Oil & Fat Co. Ottogi

Japan=Matsusaka China

Essential oil Various products

300 L 3 5,800 L 3 100 L1, 50 L2 500 L 1 3,500 L 3

China

Various products

1,500 L 3

China India China New Zealand China China Korea China China China China

Ginseng Spice Health care food Various products Chinese medicine Chinese medicine Sesame oil Petrochemicals Foodstuff Various products Soy lecithin=various

360 L 1 600 L 3 1,000 L 2 1,000 L 3 600 L 2 300 L 2 2,400 L 2 500 L 2 3,000 L 2 200 L 3 1,500 L 2

Korea

Sesame oil

2,300 L 3

1994 1994 1995 1995 1995 1995 1995 1996 1997 1998 1998

2001 2001 2001 2001 2001 2002 2002 2004 2004 2005 2005 2006 2006

Target Material

Extractor Size — 1,000 L 8 300 L 1 300 L 1 420 L 1

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 41

17.7.2008 3:20pm Compositor Name: VAmoudavally

41

Supercritical Fluid Extraction in Food Analysis

TABLE 3.5 Comparison of Some Physical Properties for a Gas, Liquid, and SCF

Gas (1 bar, 208C) Liquid (208C) SCF

Density (kg=m3)

Diffusion Coefﬁcient (m2=s)

Viscosity (Pa s)

0.6–2.0 600–1200 200–900

1–4 105 0.2–2 109 2–7 107

0.01–0.03 0.2–3.0 0.01–0.09

short list of SCF processes that are constructed by several companies and have been taken in operation recently. The very special physical properties of SCF distinguish it from liquid and gases. An SCF has a liquid-like density but its viscosity is more like that of a gas, resulting in diffusion coefﬁcients that are much higher than those in liquids. Table 3.5 shows a comparison of these characteristics for a gas, liquid, and SCF. As mentioned earlier, the ﬁrst paper dealing with the application of SCF dates from 1879. They discussed the ability of an SCF to dissolve low vapor pressure solid materials. Since then, a substantial amount of work has been done by many investigators to understand the basic fundamentals of a ﬂuid in the supercritical region.

3.7.1 PHARMACEUTICAL APPLICATIONS Supercritical ﬂuid technology (SFT) has been used in many ﬁelds for decades, such as the food industry, chemical processing, polymers, textile, forest product industries, and in the cleaning of precision parts other than pharmaceuticals. Pharmaceutical and toxicological applications of SFE are especially challenging because of the following reasons: 1. Standards for recovery and reproducibility are more rigorous than for environmental applications. 2. Analytes are usually at trace levels and are highly polar. 3. Matrices are exceedingly complex and often possess coextractives. Conventional pharmaceutical processing involves extensive use of organic solvents as either antisolvents for recrystallizing drugs from solutions, reaction media in the synthesis of drugs, or extracting agents for selectively isolating drugs from solid matrices. A major research focus in this regard has been the investigation of processes in which the traditional solvents are replaced with SCCO2. Since the residual solvent present in the extracted material is of critical importance in the pharmaceutical industry, supercritical ﬂuid carbon dioxide has found several applications. In the pharmaceutical ﬁeld, it has been widely used for the extraction of natural products like aromatic oils and caffeine, etc. Also, the extraction of vitamin E from soybean oil and a puriﬁcation method for vitamin E have been well studied. Among the reported applications, the formation of drug particles using dense carbon dioxide either as a solvent or nonsolvent and the ‘‘clean’’ synthesis of drug compounds using carbon dioxide as a reaction medium hold immense appeal for large-scale application in the pharmaceutical industry. Newer areas of their application have appeared, such as particle size reduction and designing of novel drug delivery systems [18].

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 42

42

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

3.7.2 ENVIRONMENTAL APPLICATIONS Owing to strict environmental regulations, supercritical ﬂuids are used as replacements for conventional hazardous chemicals such as hexane. At present, an application area of much activity is the environmental remediation and removal of toxic contaminants from soils and industrial waste using supercritical ﬂuids [19]. Also, SFE has been proposed as an alternative technique for activated carbon regeneration. Over 99% of a majority of organics can be removed from contaminated soil. Organics that have been successfully extracted include PAHs, PCBs, DDT, and toxophene. Carbon dioxide has been used with entrainers for the extraction of highly polar compounds. A commercial process to separate oils from reﬁnery sludge and contaminated soil has been developed by CF Systems Corporation, USA. Chelating moieties that dissolve into carbon dioxide have been developed for the extraction of heavy metals from soil.

3.7.3 FOOD APPLICATIONS The food industry is always looking for the best separation technology to obtain natural compounds of high purity, healthy products of excellent quality with several industrial applications. Research into energetically less costly technologies with respect to the environment is required. A summary of commercial applications and examples of recent developments illustrate the different possibilities that SFE has in industrial food processes. One of the ﬁrst commercial applications of SCF technology was the decaffeination of coffee. Initially, green coffee beans are soaked in water to facilitate the extraction process. The wet beans are then contacted with SCCO2, which selectively removes the caffeine. The caffeine-free beans are then roasted causing the release of the aroma components essential to the development of full coffee ﬂavor. These components are unaffected by the extraction process. The caffeine is removed from CO2 by water stripping or adsorption onto activated carbon, after which the solutefree CO2 is recycled. As a result of the increasing therapeutic role of essential fatty acids, there is considerable economic incentive to develop a process for the extraction of these materials from natural sources such as ﬁsh oils. SFE technology has been regarded as an ideal method for this purpose. Supercritical ﬂuid extraction, especially using CO2, is today a popular technology for rapid, contamination-free extraction in the food and pharmaceutical industries. Table 3.6 summarizes some of the known applications and newer applications of the SFE technique.

TABLE 3.6 SFE Applications in Food Products Paprika color (oleoresin) extraction from meats and pickles Decaffeination of coffee and tea Extraction of vegetable oils and fats Extraction of herbal medicines Flavors, fragrances, aromas, and perfumes Food colors from botanicals Antioxidants from plant materials Denicotinization of tobacco Stabilization of fruit juices Hops extraction for bitter De-oiling of fast foods Essential oil extraction Thyme oil extraction from meat and pharmaceutical products

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 43

Supercritical Fluid Extraction in Food Analysis

17.7.2008 3:20pm Compositor Name: VAmoudavally

43

Beyond environmental samples, application of SFE to samples of food interest remains a major emphasis. Several of these applications were comprehensively reviewed. Supercritical ﬂuids have been used to extract a wide range of analytes from botanical samples. These analytes range from essential oils to phytochemicals, and can include lipid extraction. These extracts have been used for analytical, supplementation, and ﬂavor and fragrance purposes. Some companies have even began to market botanical extracts obtained by supercritical CO2 extraction. In case of SFE of lipids, the technique has been utilized to extract lipids from an assortment of matrices. The ﬁrst guiding principle is the optimization of the solubility of lipids in supercritical CO2 and the improvement of the fractionation with respect to a particular lipid species. Some of these extractions have been used to analyze the fat content of different food products. Other extractions have been used to obtain pure lipid extracts or to produce products that contain a reduced amount of certain lipids or other compounds such as cholesterol. The use of SFE for the determination of fats in food products is one of the most prevalent applications in the ﬁeld. For instance, dairy products have been subjected to SFE to fractionate lipids and isolate vitamins for quantiﬁcation. Flavor compounds and other food volatiles are relatively straightforward to extract with SFE. SFE can be favorable due to the temperature and concentration advantages of using CO2, as well as the selectivity advantages and the ability to directly couple to gas chromatography (GC) analysis. Just as SFE found utility in the determination of soil-bound pesticides and herbicides, the technique found application in the determination of these same compounds in food products. For example, the SFE of pesticide residues in fruits and vegetables was applied. Because of the inherent water content in produce samples, where pesticides may be found, drying agents become especially important. As an other example, the use of SCCO2 as a replacement of hexane in soybean oil extraction is being considered recently. Data on the extraction and oil composition of soybean oil have been described [20]. It was shown that the separation of oil from CO2-oil stream at 800 bar can be carried out by dropping the pressure by only 150 or 200 bar at 708C.

3.7.4 SUPERCRITICAL FLUID CHROMATOGRAPHY Supercritical ﬂuid chromatography may be deﬁned as a form of chromatography, i.e., a physical separation method based on the interaction of an analyte in a mobile phase with a stationary phase, in which the mobile phase is subjected to pressures and temperatures near or above the critical point for the purpose of enhancing the mobile-phase solvating power. Typically, one or both parameters (i.e., pressure and temperature) extend into the critical region during a chromatographic run. This deﬁnition encompasses other less-deﬁned forms of chromatography such as dense gas chromatography, hyperpressure gas chromatography, and near (or sub-) critical ﬂuid chromatography. Supercritical ﬂuids can be used as the mobile phase to separate analytes with chromatographic columns. As in SFE, supercritical ﬂuids can have solvating powers similar to organic solvents, but with higher diffusivities, lower viscosity, and lower surface tension. The lower viscosity allows higher ﬂow rates compared to liquid chromatography, and the solvating power can be adjusted by changing the pressure. Gases, supercritical ﬂuids, and liquids have been compared as chromatographic mobile phases, 5–7 and criteria for selecting suitable mobile phases for SFC have been speciﬁed. These considerations include (1) critical pressure; (2) critical temperature; (3) dipole moment; (4) chemical interactions with the stationary phase; (5) chemical interactions with the analyte; (6) compatibility with the detection system; (7) compatibility with seals, tubing, and pumps; (8) environmental and safety considerations; (9) cost; and (10) purity. A major advantage of SFC is that it offers the advantage of liquid-like solubility, with the capability to use a nonselective gas-phase detector such as ﬂame ionization detector. Analytes that cannot be vaporized for analysis by gas chromatography yet have no functional groups for sensitive detection with the usual liquid chromatography detectors can be separated and detected using SFC. Also, compared with high-performance liquid chromatography (HPLC), SFC provides

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 44

44

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

rapid separations without the use of organic solvents. With the desire for environmentally conscious technology, the use of organic chemicals as used in HPLC could be reduced with the use of SFC. Because SFC generally uses carbon dioxide collected as a byproduct of other chemical reactions or is collected directly from the atmosphere, it contributes no new chemicals to the environment. In addition, SFC separations can be done faster than HPLC separations because the diffusion of solutes in supercritical ﬂuids is about 10 times greater than that in liquids (and about 3 times less than in gases). This results in a decrease in resistance to mass transfer in the column and allows for fast high-resolution separations. Compared with GC, capillary SFC can provide high-resolution chromatography at much lower temperatures. This allows fast analysis of thermolabile compounds. The advantages of SFC over gas or high-pressure liquid chromatography have been noted for speciﬁc types of samples, such as oligomeric polymer mixtures or complex mixtures of oleophilic components that can be readily solubilized in SCCO2 [21,22]. In fact, helium is supercritical in gas chromatography and the mobile phase has essentially no solvating power, this cannot be considered SFC. Furthermore, if one investigates the practical operating conditions of SFC, it will be obvious that many of the chromatographic analyses are started at pressures below the critical pressure, and occasionally at subcritical temperatures. In general, three conditions must be met to truly deﬁne SFC: 1. Mobile phase must always be at temperatures and pressures near or above their critical point. 2. Mobile phase must possess solvating power and, thus, be able to contribute to selectivity in the chromatographic process. 3. Mobile phase must be subject to these conditions throughout the full length of the analytical column. Finally, the coupling of extraction methods or multidimensional systems with SFC is possible, as in other forms of chromatography. SFC can be conveniently divided into two categories based on column type: open tubular and packed. The choice of column type is not only due to the obvious chromatographic differences (e.g., sample capacity, resolving power, etc.) but also due to the differences in column pressure drop and volumetric ﬂow, which impose different constraints upon the system. Of course, it is the nature of the mobile phase which is unique to SFC. Preparative SFC is used more and more in research and development laboratories and pilot plants of the pharmaceutical and ﬁne chemical industries. SFC is particularly interesting for the puriﬁcation of . . .

Chiral compounds Actives or intermediates from complex mixtures Lipophilic compounds

Because of the low viscosity and high diffusivity of the SCCO2, preparative SFC ensures faster puriﬁcation than traditional preparative HPLC. The principle is illustrated in Figure 3.11. A supercritical ﬂuid chromatograph consists of a gas supply, usually CO2, a pump, the column in a thermostat-controlled oven, a restrictor to maintain the high pressure in the column, and a detector. The column is usually a capillary GC column, but packed LC columns can also be used. The FID is the most common detector, but other GC or LC detectors can also be used. Overall, the equipment is similar to an HPLC device. Major differences over GC and HPLC are its ability to modify pressure or solvent nature during a run. As an operational principle, cold liquid CO2 is pumped. The pressurized carrier gas enters into a capillary column, which is coated by an adsorbing material, or packed with coated beads, the stationary phase. As the supercritical solution moves through the column, the more strongly adsorbing solutes will be retarded by the stationary phase. On the other hand, larger molecules

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 45

Supercritical Fluid Extraction in Food Analysis

FIGURE 3.11

17.7.2008 3:20pm Compositor Name: VAmoudavally

45

Picture of SCF chromatograph.

cannot sample the lower carrier speeds in boundary regions and thus move faster than small molecules [19]. Before entering the column, it is heated and becomes supercritical. Because of its low viscosity, the pressure at the column outlet is almost identical to the pressure at the column inlet. At the column outlet, the mobile phase is decompressed and heated and becomes gaseous. Products are recovered in cyclones of appropriate design. The gaseous CO2 is then cleaned and cooled down and returned to the tank following the pathway given in Figure 3.12.

3.8 INSTRUMENTATION Supercritical ﬂuid extraction was the ﬁrst application of supercritical ﬂuids. In other words, it can be expressed as the technique of supercritical ﬂuid extraction (SCFE or SFE) that is a new type of green-extracting technique, which is becoming popular in the modern world with wide adaptability and is receiving considerable attention as a method having several kinds of application areas. With this technique, no chemical agent is needed during the process of extraction and the separated substances are also not polluted. Thus, it entirely meets the green requirement of the human race to food, medicine, health protection goods, and cosmetics. The process is the latest technique used to extract oils from natural products, organic pollutants from wastewater, aromatic isomers from

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 46

46

17.7.2008 3:20pm Compositor Name: VAmoudavally

Programmer

Handbook of Food Analysis Instruments

Pressure transducer

Column cross-sections Detector

Oven Column

Pump system

Vent

Solvent preheater

FIGURE 3.12

Injection valve

Restrictor

SCF chromatograph.

mixtures, low molecular weight materials from polymers, and light components from coal. Also, this technique boasts the features of high extraction effect, high quality, low extraction temperature, low energy consumption, and no pollution, and is especially suitable for the extraction of heatsensitive and active substances.

FIGURE 3.13

SCF extractor.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 47

17.7.2008 3:20pm Compositor Name: VAmoudavally

Supercritical Fluid Extraction in Food Analysis

FIGURE 3.14

47

SFE equipment. (From Fluitron, Inc., http:==www.ﬂuitron.com=.)

As an overall technical approach, a supercritical ﬂuid extractor consists of a tank of the mobile phase, usually CO2, a pump to pressurize the gas, an oven containing the extraction vessel, a restrictor to maintain a high pressure in the extraction line, and a trapping vessel as depicted generally in Figures 3.13 and 3.14 [23]. In Figure 3.13, the compressor stroke rate of the system above is 58 rpm. The extraction pressure and ﬂow rate are controlled by a back pressure regulator as is the separation pressure. The extractions and separation temperatures are controlled and indicated independently. A second atmospheric separation may be performed in a glass vessel or cold trap before the gas stream passes through the ﬂow rate indicator and ﬂow totalizer and is vented to the atmosphere [24]. Analytes are trapped by letting the solute-containing supercritical ﬂuid decompress into an empty vial, through a solvent, or onto a solid sorbent material. Extractions are done in dynamic, static, or combination modes. In a dynamic extraction, material is fed in continuously by means of a pump. The discharge of the processed material is also continuous. Shown in Figure 3.15 is a pilot plant for the puriﬁcation=concentration processing of a crude ethanol solution. In a static mode, loading and discharge of the material are carried out by a highpressure batch extractor, with the opening and closing of the lid automated and the feeding of the supercritical ﬂuid continuous. Figure 3.16 shows the supercritical ﬂuid batch extraction device and the vessel with its automatically opening and closing lid. Extraction of odorants or colorants and removal of pesticides from ginseng extract or powder can be given as examples for this system. In the combination mode, a static extraction is performed for some period, followed by a dynamic extraction.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 48

48

FIGURE 3.15

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

Pilot plant with continuous material feeding, for concentration of crude ethanol solutions.

It is being carried out with carbon dioxide on a large scale for the decaffeination of green coffee beans and the extraction of hops for beer production. Some plant extracts are also produced and some of these are a source of pharmaceutical substances. For example, taxicins can be extracted from yew leaves and used as precursors for anticancer drugs. Carbon dioxide is widely used because of its environmental friendly nature and in some cases it is modiﬁed with other solvents to improve its solvating properties. For example, more polar substances are sometimes extracted with carbon dioxide modiﬁed with ethanol. Several ﬁrms are employing the technique in various commercial applications, such as coffee decaffeination or botanical extraction. A sample extraction system is described in Figure 3.17. Components of this pilot plant scale include extraction vessel, separator, high-pressure pump, and complete recycle capability as depicted in Figure 3.18. Among the available standard systems are a 4 L unit and a process development unit (PDU) with maximum allowable working conditions of 5000 psi (333 bar) at 1008C. For decaffeination applications, this system is also used for

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 49

Supercritical Fluid Extraction in Food Analysis

FIGURE 3.16

SFE device (extraction vessel: 390).

FIGURE 3.17

Botanical products SFE system.

17.7.2008 3:20pm Compositor Name: VAmoudavally

49

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 50

50

FIGURE 3.18

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

Multipurpose SFE apparatus.

extraction of spice, hops, and vanilla bean; and defatting of cocoa powder. For chemical processing applications, polymerization, specialty oil fractionation, countercurrent wastewater extraction can be given as examples. Puriﬁcation by removal of lipids, supercritical ﬂuid micronizing, and residual solvent removal are other points of view for various application areas [25]. Another sample for SFE system that is illustrated in Figure 3.19 offers the SFE system as a complete turnkey for extractions up to 9200 psi and separations up to 9200 psi. Gas or ﬂuid from your commercial gas bottle passes through a ﬁlter to the compressor. The advantages of SFE (compared with liquid extraction) are ﬁrstly that it is relatively rapid because of the low viscosities and high diffusivities associated with supercritical ﬂuids. The extraction can be selective to some extent, by controlling the density of the medium. Separation

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 51

17.7.2008 3:20pm Compositor Name: VAmoudavally

Supercritical Fluid Extraction in Food Analysis

FIGURE 3.19

51

SFE equipment.

of the ﬂuid substance from the product is relatively easy and the solvent residues in the product are small and of a being nature. Trial experiments are carried out on pilot scale using a system which is installed to be used for both pure and modiﬁed ﬂuids and can recycle the ﬂuid. Figures 3.20 and 3.21 show the various additional SFE commercial scale extraction mechanism equipments [26,27].

3.9 CURRENT TRENDS AND FUTURE EXPECTS OF SUPERCRITICAL FLUIDS If one have a conclusive look through supercritical ﬂuids from historical point of view, as a starting point for scientiﬁc discoveries, critical point was discovered by Baron Charles Cagniard de la Tour in 1822. However, supercritical state was properly described only in the year 1870 by Thomas Andrews, who named the ‘‘critical point’’ for the ﬁrst time. At the origin of new processes, during 1920s, application studies were done in petrochemistry ﬁelds. During 1960s, natural product extraction with SCCO2 was developed. For innovative industrial applications, SCCO2 found its ﬁrst industrial application with coffee decaffeination, at

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 52

52

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

FIGURE 3.20 Tea decaffeination plant (Switzerland). (From Uhde High Pressure Technolgies, http:==www. uhde.hpt.com=.)

FIGURE 3.21

Sesame seed-oil SFE system (2400 L 2). (From U-MAX, Inc., http:==www.iumax.co.kr=.)

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 53

Supercritical Fluid Extraction in Food Analysis

17.7.2008 3:20pm Compositor Name: VAmoudavally

53

the end of 1970s. Hop extraction and tea decaffeination processes were developed during the same time. Supercritical ﬂuids have also been used as unique solvents in a number of analytical techniques, such as nuclear magnetic resonance spectroscopy and thin-layer chromatography (TLC); chemists, however, normally associate the preﬁx ‘‘supercritical ﬂuid’’ with chromatographic or extraction methodologies that have been extensively developed during the past 15 years. During the 1980s, a ﬁrst strong development step occurred for supercritical technology with the building of huge industrial units dedicated to solid extraction, in Europe, in the United States, and in Australia, and also with the building of ﬁrst development unit for liquid fractionation. First spice extracts started to be delivered onto the market. During the 1990s, building of industrial units dedicated to toll extraction and world extension of the technology revealed development of new applications and new markets. Also the ‘‘renaissance’’ of SFC occurred in the mid-1970s, largely as a result of improvements in injection and pumping devices, enhanced column efﬁciencies, and the reﬁnement of transport mechanisms to deliver the separated solutes to modiﬁed gas (CC) and liquid chromatographic (LC) detectors. By contrast, SFE, despite a long history as a physicochemical phenomenon and a recent plethora of applications in chemical engineering, has developed as an analytical technique only since the mid-1980s and is presently in an evolutionary state. The current practice of analytical SFE is divided between ofﬂine and online methods, despite their common physicochemical basis. Such deﬁnitions refer to the mechanism of conducting the extraction. Current trends in analytical SFE are diverse and worthy of comment. The recent introduction of instrumentation capable of performing extractions on larger and more representative samples is one current trend. As a result, instrumentation manufacturers have had to consider the design of supercritical ﬂuid delivery systems with respect to higher ﬂuid ﬂow rates and extraction pressures. Likewise, the development of multisample extractors for the simultaneous processing of large numbers of samples has further catalyzed the creation of new instrumentation. In case of future developments, optimal SFE system has yet to be created. Extraction systems need to be developed that offer the ﬂexibility of operating at both higher and lower pressure ranges. SFE is an excellent technique for examining volatile components because the extractions can be conducted at relatively low temperatures and in a non-oxidative environment. Emergence of novel applications for supercritical ﬂuids occurred from the late 1990s: precision cleaning, aerogels, impregnation, particle generation, microencapsulation, etc. Industrial developments in these ﬁelds are presently going on. Reaction processes such as oxidation with supercritical water (SCOW process) or chemical= biochemical synthesis promise industrial successes. Also to be noticed are the innovative processes using subcritical water for natural products extraction together with SFC. Certainly, SFE is a viable alternative to headspace techniques, which depend on thermal energy to volatilize analytes; hence, the authors can envisage a bright future for SFE in sensory analysis problems.

3.10 CONCLUSION In the last decade, new trends have emerged in the food industry. These trends include an enhanced concern for the quality and safety of food products, increased preference for natural products over synthetic ones, and broadened regulations related to nutritional and toxicity levels of active ingredients. These trends have driven supercritical ﬂuid technology to become the primary alternative to traditional solvent extraction for the extraction and fractionation of active compounds. Supercritical ﬂuid extraction is an extraction process using a supercritical ﬂuid as a solvent. SFE utilizes the ability of certain chemicals to become excellent solvents for certain solutes under a combination of temperature and pressure. The physiochemical properties of a ﬂuid in the supercritical state are in between those of a typical gas and liquid. For example, the density of a supercritical ﬂuid can be changed by varying the pressure on the ﬂuid. Carbon dioxide is certainly the most

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 54

54

17.7.2008 3:20pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

popular ﬂuid because of its physiological compatibility, nontoxicity, inﬂammability, easy availability, convenient critical parameters (Tc ¼ 318C, Pc ¼ 7.38 MPa), inexpensiveness, and environmental friendliness. Supercritical ﬂuid extraction has proved effective in the separation of essential oils and its derivatives for use in the food, cosmetics, pharmaceutical, and other related industries, producing high-quality essential oils with commercially more satisfactory compositions (lower monoterpenes) than obtained with conventional hydro distillation. Whether it is used as a solvent for extraction in analytical methods such as gravimetric determination of fat content or in large-scale extractions such as decaffeinating coffee, SCCO2 has proven its usefulness as a replacement for organic solvents. Supercritical carbon dioxide proved to be highly selective for caffeine, prompting its use as the selected solvent in the commercial decaffeination of coffee and black tea. Recent investigations have demonstrated the potential exploration of solvent and antisolvent properties of carbon dioxide in the recovery of alkaloids such as theophylline, theobromine, and pilocarpine, among others. Supercritical ﬂuid extraction with CO2 delivers the most natural-smelling and -tasting extracts because there are no volatiles removed in a residual solvent-removal post-processing step. This has beneﬁts for an array of products. For instance, many spices are known for their therapeutic value. Supercritical ﬂuid technology offers tremendous advantages, such as the absence of any organic solvent residues and selective extraction and fractionation of different compounds. All of these advantages are almost impossible to obtain easily from conventional processes at low operating costs. Therefore, supercritical ﬂuid technology is an ideal tool for the processing of active compounds for use in food products and dietary supplements. To conclude, with increasing concern about the use of organic solvents and their disposal, SFE is gaining popularity faster than ever before. The future looks promising for the use of supercritical ﬂuids, with new methods of extraction constantly being developed, as with other novel uses for the food processing industry.

REFERENCES 1. Montanari, L., King, J.W., List, G.R., and Rennick, K.A., Selective extraction of phospholipid mixtures by supercritical CO2 and cosolvents, J. Food Sci., 61, 1230, 1996. 2. Temelli, F.J., Extraction of triglycerides and phospholipids from canola with supercritical carbon dioxide and ethanol, J. Food Sci., 57, 440, 1992. 3. Reid, R.C., Prausnitz, J.M., and Poling, B.E., The Properties of Gases and Liquids, 4th ed., McGraw-Hill, New York, 1987. 4. Rozzi, N.L. and Singh, R.K., Supercritical ﬂuids and the food industry, Compr. Rev. Food Sci. Food Safety, 1, 33, 2002. 5. Hannay, J.B. and Hogarth, J., Solubility of solids in gases, Proc. Roy. Soc., London, 29, 324, 1879. 6. Brogle, H., Carbon dioxide as a solvent: Its properties and applications, Chem. Ind., 19, 385, 1982. 7. Buchner, E.G., Die beschrankte Mischbarkeit von Flussigkeiten das System Diphenyamin und Kohlensaure, Z. Phys. Chem., 56, 257, 1906. 8. Krukonis, V., Brunner, G., and Perrut, M., Industrial operations with supercritical ﬂuids: Current processes and perspectives on the future, Proceedings of 3rd International Symposium on Supercritical Fluids, Strasbourg, 1994, p. 1. 9. Dixon, D.J. and Johnston, K.P., Supercritical ﬂuids, in Encyclopedia of Separation Technology, Ruthven, D.M., Ed., John Wiley, 1997, p. 1544. 10. Akgerman, A. and Giridhar, M., Fundamentals of solids extraction by supercritical ﬂuids, in Supercritical Fluids—Fundamentals for Applications, Sengers, J.M.H. and Kiran, E., Eds., Klüwer Academic Publishers, 1994, p. 669. 11. Sihvonen, M., Jarvenpaa, E., Hietaniemi, V., and Huopalahti, R., Advances in supercritical carbon dioxide technologies, Trends Food Sci. Technol., 10, 217, 1999. 12. Tanaka, Y. et al., Extraction of lipids from salmon toe with supercritical carbon dioxide, J. Oleo Sci., 52, 295, 2003.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 55

Supercritical Fluid Extraction in Food Analysis

17.7.2008 3:20pm Compositor Name: VAmoudavally

55

13. Gopalan, B., Motonobu, G., Akio, K., and Tsutomu, H., Supercritical carbon dioxide extraction of turmeric (Curcuma longa), J. Agric. Food Chem., 48(6), 2189, 2000. 14. Gomez, A.M., Lopez, C.P., and Ossa, E.M., Recovery of grape seed oil by liquids and supercritical carbon dioxide extraction: A comparison with conventional solvent extraction, Chem. Eng. J., 61, 227, 1996. 15. Kurnik, R.T. and Reid, R.C., Solubility extreme in solid–ﬂuid equilibria, AIChE J., 27, 861, 1981. 16. Hawthorne, S.B., Analytical-scale supercritical ﬂuid extraction, Anal. Chem., 62, 633, 1990. 17. Snow, N.H., Dunn, M., and Patel, S., Determination of crude fat in food products by supercritical ﬂuid extraction and gravimetric analysis, J. Chem. Educ., 74, 1108, 1997. 18. Subramanyam, B., Rajeswski, R., and Snavely, K., Pharmaceutical processing with supercritical carbon dioxide, J. Pharm. Sci., 86, 885, 1997. 19. Kiran, E. and Brennecke, J.F., Supercritical Fluid Engineering Science: Fundamentals and Applications, American Chemical Society, Washington, DC, 1993, p. 7. 20. Friedrich, J.P. and Pryde, E.H., Supercritical CO2 extraction of lipid-bearing materials and characterization of the products, J. Am. Oil Chem. Soc., 61, 223, 1984. 21. Andersen, M.R. et al., Supercritical ﬂuid extraction as a sample introduction technique, J. Chromafogr. Sci., 27, 371, 1989. 22. Fukuzato, R., Proceedings of the 6th International Symposium on Supercritical Fluids, Versailles (France), April 28–30, 2003. 23. Fluitron, Inc., http:==www.ﬂuitron.com=. 24. Newport Scientiﬁc, Inc., http:==www.newport-scientiﬁc.com=. 25. Pressure Products Industries, Inc., http:==www.pressureproductsindustries.com=. 26. U-MAX, Inc., http:==www.iumax.co.kr=. 27. Uhde High Pressure Technologies, http:==www.uhde-hpt.com=.

Otles/Handbook of Food Analysis Instruments 45660_C003 Final Proof page 56

17.7.2008 3:20pm Compositor Name: VAmoudavally

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 57

17.7.2008 4:02pm Compositor Name: VAmoudavally

Processes 4 Microwave-Assisted in Food Analysis Jacqueline M.R. Bélanger and J.R. Jocelyn Paré CONTENTS 4.1 4.2

Introduction ............................................................................................................................ 57 Theoretical Considerations .................................................................................................... 58 4.2.1 Liquid-Phase Extraction ............................................................................................ 58 4.2.2 Gas-Phase Extraction ................................................................................................. 60 4.3 Instrumentation ...................................................................................................................... 61 4.4 Applications to Food Analysis .............................................................................................. 64 4.4.1 Meat, Poultry, and Fish ............................................................................................. 65 4.4.2 Dairy and Eggs .......................................................................................................... 66 4.4.3 Fruits and Vegetables ................................................................................................ 66 4.4.4 Cereals and Oilseeds ................................................................................................. 67 4.4.5 Herbs and Spices ....................................................................................................... 67 4.4.6 Food Ingredients ........................................................................................................ 68 4.4.7 Characterization, Reaction, and Bakery Products ..................................................... 69 4.4.8 Food Safety ............................................................................................................... 70 References ....................................................................................................................................... 78

4.1 INTRODUCTION Microwave-assisted extraction (MAE) is a relatively new tool. Dating back to the mid-1980s, it was the ﬁrst application within the microwave-assisted processes (MAP*), a family of technologies pioneered and patented by Canada’s Department of Environment [1–11]. They were developed in response to growing demand for accrued environmental sustainability through enhanced cost and performance efﬁciency from a chemical and energy standpoints. Only the extraction component of MAP is the subject matter of this handbook. It is appropriate to do so especially that the technology had originally been conceptualized to serve the food industry, especially targeting the ﬂavor and fragrance sectors. MAP extraction can be divided into two main streams of activities, namely liquid-phase extraction and gas-phase extraction. While the former has seen a tremendous level of development, the latter is still at the stage of equipment design and commercialization, thus limiting its access on a day-to-day basis. This chapter will address both technologies from a theoretical standpoint and an instrumentation standpoint in addition to provide a thorough review of application types in the food analysis sector. While the listing of applications will not be exhaustive from a number of publications standpoint, it is believed that it will present a fair and representative snapshot in time of the state of the art of the technology in terms of range of applications as well as range of foodstuff types where the technology has found practical applications. * MAP is a trademark of Her Majesty the Queen in Right of Canada as represented by the Minister of the Environment.

57

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 58

58

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

The goal is to provide the readers with a single-document handbook that will inform them rapidly on the applicability of the technology for their intended applications. To achieve this, we have organized the section dealing with applications by foodstuff. We added two subsections dealing with purpose as opposed to foodstuff, namely on characterization and reaction and on food safety. The chapter comprises a relatively large table that summarizes all the applications section. The latter was included to provide for a rapid check as to whether the intended application has been reported by itself or for other foodstuff. Conversely, readers looking for a speciﬁc analyte will ﬁnd the table well adapted for that purpose, despite the fact that it is not presented in alphabetical order of analytes.

4.2 THEORETICAL CONSIDERATIONS Microwaves are a source of electromagnetic energy that interacts to various degrees with materials according to some fundamental physical properties characteristic of these materials themselves. Hence, microwaves are inherently unable to heat any mixture in a uniform fashion. To the noninitiated, this statement may appear opposite to ﬁrst impressions, ﬁrst impressions often resulting from the even more widely spread use of the domestic microwave ovens at home. While these home appliances make use of similar frequency and devices to generate and apply the microwaves, they are also equipped with a number of devices that allow for signiﬁcant time periods where the actual emissions of microwave are turned off—even though the turntable and the fan are still operational—thus allowing for thermal transfer to occur between the various chemical constituents. This deceptively simple approach makes for the popularity of these appliances as they provide unusually fast means to heat, or reheat, ﬁnished or raw foodstuff that at the time of being consumed offer relatively uniform temperature proﬁles. But the reality is otherwise and MAE is based upon more fundamental principles, namely that the basic physical properties of materials known as dielectric properties lead to different heating characteristics for each and every chemical substance. The most important dielectric properties are the static dielectric constant, represented as «0 , and the loss factor, represented as «00 . The dielectric constant can be described as the degree of opposition to the passage of the electrical ﬁeld component of the microwaves exerted by that material. On the other hand, the loss factor can be described as the ability the material has to convert this electrical energy into heat. Hence, if one accepts that there must ﬁrst be an interaction of the microwaves with the material before the latter can transform the energy contained in the microwaves into heat, then one could say that the dielectric constant is a prime factor to consider in the determination of microwave applicability. MAE, as is the case for all MAP technologies, is based upon the natural ability microwaves have to create nonequilibrium, thermal gradients within complex materials such as foodstuff. This statement holds true for both liquid-phase extraction as well as gas-phase extraction. In fact, all MAP technologies share a common development logic, namely to see whether microwaves create unique operational conditions, whether these conditions provide for novel of additional and enhanced beneﬁts, and whether that can be reproduced in a controllable fashion.

4.2.1 LIQUID-PHASE EXTRACTION By liquid-phase extraction, we are referring to the steps necessary to dissolve target analytes contained within a matrix of interest into a liquid material to make the resulting solution amenable to subsequent separation and chemical identiﬁcation and determination of each substances making up that solution. Conventional and traditional liquid-phase extraction techniques include hot solvent extraction, liquid-liquid extraction, and Soxhlet extraction to name a few. All of them are characterized by relatively tedious procedures, long extraction times, or a signiﬁcant demand on solvents used to perform the procedures. More recent techniques such as supercritical ﬂuid extraction and pressurized

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 59

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

59

ﬂuid extraction have come and gone but, unfortunately from an environmental standpoint, failed to become signiﬁcant replacement to the long-established techniques. Everyone of these techniques, apart form microwaves, is based on common basic principles from an energy standpoint. It can be summarized as follows: a source of energy is provided to the extractant whether liquid as for Soxhlet or gas as for carbon dioxide, the energized extractant is put into contact with the matrix that contains the target analyte, the extractant diffuses into the matrix, the extractant dissolves the soluble components including the target analytes, the extractant diffuses out of the matrix, and ﬁnally, a separation step is applied to recover the target-containing extract from the matrix. One can see that basic requirements and drawbacks include the following: the energy is applied randomly to the matrix, the container, and the extractant, hence large amounts of energy are used as usually as the ratio of extractant-to-matrix is high; the selected extractant must be able to diffuse into the matrix, irrespective of its ability to dissolve the material, hence the most adequate solvent may not be suitable for extraction; and a relatively large amount of time is required since the process is based upon diffusion in and out of the matrix in addition to the associated partition equilibration times required to reach maximal extraction efﬁciency (or repeated extraction steps that lead to even greater quantities of waste extractants). In contrast, when used under its originally intended form [1], the MAE process follows very distinct steps. They can be summarized as follows: the matrix of interest is put into contact with the extractant into a microwave-transparent container; the extractant is selected for its ability to dissolve the target analyte and its transparency to microwaves relative to the matrix, irrespective of its diffusivity into the matrix; and ﬁnally, a separation step is applied to recover the target-containing extract from the matrix. One can see that the basic requirements are fundamentally different from those associated with the conventional technologies. In fact, this technology is the only extraction technology that makes use of selective heating and where the energy is deposited directly into the matrix as opposed to heating randomly the container, the solvent, and the matrix. This approach brings economic advantages from an energy-efﬁciency standpoint in addition to offer in many cases signiﬁcant reductions in the quantities of solvent used that in turn lead to reduced operational costs associated with solvent recovery and disposal. Early Refs. [1–16] from our laboratories highlighted the versatility and the high degree of suitability of the technology for the ﬁeld of food analysis. This technology is deceptively simple, and as such it has been confronted with non-negligible levels of inertia in its acceptance. Times of extraction were now in seconds as opposed to hours or even days for Soxhlet. The issue of relative transparency to microwaves is not intuitive and consequently, many applications were derived from this core approach whereby one simply uses microwaves as a source of heat. This alternative approach still falls within the realm of the basic technology, but it offers reduced advantages or simply advantages of lesser impact. In fact, a deceptively large number of applications were developed making use of a solvent that absorbs relatively well microwaves while being relatively efﬁcient at converting microwaves into heat, say for example, acetone. The fact that the selected solvent may be less transparent than need be brings about increased energy consumption and longer extraction times. These two parameters are still much better than conventional technologies if we agree that better is characterized by lower energy consumption and faster extraction, but they are nowhere nearly as efﬁcient as when one uses transparent solvents. The approach of simply heating the mixture as opposed to apply energy selectively is by far more intuitive as one can extrapolate almost without modiﬁcations the parameters used in conventional technologies. If one extends this approach one step further, then one gets into very similar conditions as those found historically in the evolution of inorganic digestion over a hot plate versus the now widely accepted and highly more energy-efﬁcient microwave-assisted digestion. Still further along the lines of the similarity of evolution, it is not surprising then to see that a majority of references deal with the so-called closed-vessel applications where the matrix is put into contact

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 60

60

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

with the solvent and both are contained into sealable microwave-transparent container. The extraction work is then carried out under high operating temperature and pressure conditions thus reducing somewhat the extraction times, albeit they are still at least two orders of magnitudes larger than open-vessel applications. Energy efﬁciency does not even come close either. Another parameter within the approach of selective heating associated with open-vessel heating is the possibility of doing in situ chemistry when required whether to enhance the extraction efﬁciency from a yield standpoint or simply to provide an improvement for the subsequent separation and analysis steps. Again, despite an early application to food safety by our collaborators and ourselves [17] not much has been done since in food analysis. This chapter will cover both approaches irrespective as to whether one heats selectively or not. Similarly, we will cover both closed- and open-vessel applications as a whole as we opted to survey applications from a foodstuff standpoint as opposed to a method standpoint.

4.2.2 GAS-PHASE EXTRACTION By gas-phase extraction, we are referring to the steps necessary to vaporize target analytes, whether in part or completely, contained within a matrix of interest into a gaseous environment to make the resulting gaseous mixture amenable to subsequent separation and chemical identiﬁcation and determination of each substances making up that gaseous mixture. Conventional and traditional gas-phase extraction techniques include static or dynamic headspace (HS) sampling, purge-and-trap, and other adsorption=desorption techniques to name a few. They are characterized by the need to vaporize the target analytes and recover it, that process being in constant competition with the vaporization of other substances, not the least of which is the main body of the matrix itself which can be signiﬁcant if it is relatively volatile such as water for example. More recent technologies include solid-phase microextraction (SPME) which has not gained the anticipated broad acceptance, mostly as the result of its serious limitations in terms of target analyte quantiﬁcation. Static techniques such as HS sampling are based upon the partition that occurs between the various gaseous components when maintained at constant temperature and pressure for a time long enough that one can assume to be near the equilibrium. That partition is then dependent upon the ability the product has to vaporize, or more simply, on its partial vapor pressure. This nonlinear behavior between two distinct substances being present in equal amounts in the matrix makes quantitative application of the technology a real challenge. Consequently, there are basically no ofﬁcial methods based on this technology. It is mostly used for qualitative proﬁling, an activity of special importance to food analysis. Dynamic techniques such as purge-and-trap are not only more amenable to quantitative determinations, but are also plagued by relatively tedious procedures where errors are a major concern. Operational costs can also be a concern according to the sorbent used or to the coolant used. Consequently, again, limited use is made of the technology. MAP has been applied to gas-phase extraction mostly as a result of its speed, ease of use, and broad range of applicability [3,5–11,18]. The technology is based again on selective heating. In fact, microwaves can heat a liquid or solid matrix selectively over the surrounding gaseous environment. Hence, if one draws a parallel with say static HS, then the fundamental principles are as follows: the matrix, say wine, is enclosed into a sealable container, the container is sealed, the container and its contents are subjected to microwave energy, and the microwaves heat selectively the liquid phase while the gaseous phase remains relatively cold as it is only heated as a result of thermal diffusion from the liquid. The liquid matrix along with its volatile components partition with the gaseous phase at an accelerated rate because of the energy nonequilibrium of the system; the gaseous sample can be recovered and injected directly onto an analytical separation and characterization device such as a gas chromatography-mass spectrometry (GC-MS) for example.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 61

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

61

The analogy to conventional HS ends there. In fact, the key word in the preceding paragraph was ‘‘nonequilibrium.’’ MAP-HS as we termed it is a kinetics-driven technology as opposed to a purely thermodynamics, equilibrium-based technology. This approach opens a whole new area of possibilities. For example, in the case above, it is possible to heat the wine matrix at a much more rapid rate than the surrounding gas. Let us assume that we apply microwaves until the liquid is at 958C while the gas is still at 408C. Then, at the moment where we stop the application of the microwaves, the system intends to revert back to equilibrium through a number of relatively slow thermal diffusion, chemical, and physical processes. The actual phenomena occurring are numerous and complex. Under these conditions (liquid at 958C and gas at 408C), the rate of vaporization is greatly enhanced compared to a conventional 958C overall temperature. The vapor pressure of each component still plays a role, but their relative behavior may be affected if these substances do not afford similar vapor pressure dependent upon temperature. Other factors that have an impact on the separation of liquid and gas are the fact that water, in our wine example, has a heat capacity of about twice the value of that of other organic substances in addition to often afford a lower vapor pressure. In other words, based on this heat capacity factor alone, for every one mass unit of water that evaporates and recondenses it provides enough energy to vaporize two mass units of the other components. The net result is an enhancement of the organic components over that of water over what would be observed for a conventional HS sampling carried out under thermal equilibrium at 958C. This behavior, which is synonymous with saying that Raoult’s law does not apply under these kinetics driven conditions, is creating lots of discussions around its ﬁnal potential as a food analysis tool. This is only one fundamental phenomenon that is novel in this technology. There are several others, a few of which are still upon investigation to determine their relative importance and potential use in further reﬁning the technology. This approach clearly presents some serious challenge from an end-user acceptance standpoint. It may well be the ﬁrst kinetics-driven analytical technology available to the food analyst and as such it will require exceptionally reliable instrumentation before it becomes a standard analytical technology in every food laboratory. Despite that, we opted to include it in this chapter as it is the subject of intense product development at this time and that there is no doubt that it will become a major tool for foods, foodstuff, and food ingredients proﬁling in the very near future.

4.3 INSTRUMENTATION The two previously described approaches of applying microwave energy, namely the mere bulk heating of a mixture through the use of absorbing containers and solvents, and the more reﬁned, albeit demanding, selective heating of the target materials led to the development of two main types of laboratory extraction instruments. The most widely used is still the oven-type apparatus that makes use of long-proven bulk heating technology. These instruments present little novel technology for the end user as they simply provide an efﬁcient heating of the mixture along with its container. This approach makes use of the so-called multimode cavities and for the purpose of this handbook it is sufﬁcient to simply compare it to a conventional oven in which the source of energy has been changed from thermal to microwave. Generally, these apparatus operate under closed-vessel conditions, namely that the extraction vessel is closed and subjected to microwaves. The heating of the mixture gives rise to an elevation in the temperature, which is now compounded by the fact that there is an accompanying rise in pressure. It is intuitive to see the limitations inherent to this approach, namely the relatively long extraction times because of the need to wait for the pressure to decrease signiﬁcantly before being able to operate the vessel and proceed with further sample processing and analysis. This brings about risks associated with the degradation of thermally labile components, with the potential for the

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 62

62

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

analytes to readsorb onto the matrix during the cool down=depressurization period, and with the occurrence of potentially explosive conditions. This approach is far from being a technological advancement in food analysis. Despite this, it does bring about signiﬁcant reductions in terms of processing times (usually minutes instead of hours), consumption of solvents, and solvent exposure to laboratory personnel. One can assume that this limited commercial development lies on two main reasons, namely the possibility of directly using existing core laboratory technology and the lack of competition. The former is the result of relatively inexpensive and readily available multimode ovens dedicated to laboratory work in the area of digestion (more recently, the same trend was applied to chemical synthesis). The latter is the result that the basic technology was patented by Canada’s Department of the Environment who licensed the Hewlett-Packard Co., CEM Corp., and Société Prolabo to commercially exploit the technology. Hewlett-Packard to date did not produce any commercial apparatus and CEM effectively acquired Société Prolabo. This led to the monopoly by CEM. Until today CEM offers only closed-vessel, multimode technology to its users. The second approach is based on selective heating and makes use of containers and solvents that are transparent to microwaves relative to the matrix to be subjected to extraction. In this case, one can operate safely and much more efﬁciently under open-vessel conditions (i.e., atmospheric pressure). Operating temperatures remain low, in fact, in its most basic form such instruments do not need temperature probes as the judicious application of microwaves will only give rise to temperature increases up to the boiling point of the solvent selected. Once ﬁtted with an appropriate reﬂux column, the system is auto-controlled in terms of temperature. The advantages and superiority of this approach are obvious. To name a few: reduced extraction time (in seconds) leading to reduced exposure to high temperature, no pressure hence allowing for immediate downstream processing capacity, lower instrument costs (no need for pressurized containers and temperature sensors), enhanced selectivity, wider range of suitable solvents, and ease to modify the matrix (e.g., by adding water) and enhance further the selectivity. Unfortunately, with the closure of the Société Prolabo, this type of extraction apparatus has ceased to be commercially available, thus limiting greatly the potential of the technology in the hands of the end user. For the purpose of comparison, it may be worth noting that since this approach made use of focused microwave applicators, the power requirements were reduced signiﬁcantly. An apparatus would operate in the 200–300 W only. Extraction times for a 5 g sample would be of the order of 15–30 s. Even when in situ chemical derivatization is necessary in which case the power has to be further reduced [19], the overall extraction and derivatization procedures would require 2–3 min. It is plausible to assume that this lack of open-vessel technology combined to the fact that no major analytical instrument manufacturer makes microwave apparatus has contributed to the slow acceptance and use of the technology. In fact, to date, no manufacturer provides a commercial version of an automated apparatus whereby the samples would be extracted, ﬁltered, and subsequently transferred onto an analytical device, say a GC for example. This again is deplorable and hinders dramatically the potential for method development. There is a third line of instrumentation, namely that dealing with gas-phase extraction such as MAP-HS. In this latter case, the technology has been licensed by Environment Canada for commercial exploitation to Hewlett-Packard Co. and Shimadzu Corporation. Instrumentation is in development and it is expected that commercially available, fully interfaced apparatus will reach market in the near future. To the end user these instruments, which will make use of novel and stateof-the-art technology such as focused mono-mode cavities and even solid-state generators, will appear as simple enhancements on their current technologies, but offering vastly superior performance in addition to expand signiﬁcantly on the types of samples amenable to this type of analysis. One cannot end this section without commenting brieﬂy on the advent of large-scale extraction systems making use of liquid solvents and liqueﬁed gases as solvents. These are in operation now on a proprietary basis, but the technology that was used in their design and construction will be made

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 63

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

63

available soon. Similarly, apparatus used to monitor the progression of an extraction process for example will also be introduced shortly. It is hoped that, albeit it is unusual for a large-scale apparatus to come to be before an equivalent laboratory system was proven and validated, their release will stimulate the development of a new generation of MAP apparatus for food analysis. The following are photographs of typical MAP extraction devices currently in use, whether for laboratory or for industrial purposes (Figures 4.1 and 4.2).

(a)

(b)

FIGURE 4.1 (See color insert following page 240.) Typical commercial MAP extraction apparatus suitable for food analysis laboratories. (a) shows a focused mono-mode open-vessel apparatus for multi-step fast extraction and (b) shows a multi-mode closed-vessel multi-sample apparatus.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 64

64

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

(b)

(a)

(c)

FIGURE 4.2 Typical industrial-scale MAP extraction apparatus for food processing. Photograph (a) shows a batch-type apparatus, photograph (b) shows a continuous mode extractor, and photograph (c) shows a liqueﬁed gas pressurized system.

4.4 APPLICATIONS TO FOOD ANALYSIS This section will cover applications published from circa 1995 to date. That coverage was selected to follow the publication of the ﬁrst full chapter on this technology [11] applied to food analysis and as such it represents a seminal date in the expansion of the use of the technology in this area. As mentioned previously, this section does not claim to be an exhaustive list of publications, rather it is an attempt at providing a list of application types with the goal to be a good representation of the state of the art as of today. In other words, when reference is made to extraction of say paprika, we do not infer that the cited work constitutes all the work published on using MAP to extract paprika. Rather, it only presents one such example, albeit efforts were made to cite papers that used different approaches while using the technology so as to provide broad range of use and modes of use.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 65

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

65

The applications are grouped primarily by foodstuff, although some applications that were more process in nature such as food safety, food characterization, and chemical reaction performed to enhance the extraction to the subsequent characterization and determination are presented under a process HS.

4.4.1 MEAT, POULTRY, AND FISH In their continued applications of the development of the MAP family of technologies, Paré et al. [19] reported the ﬁrst full method making use of the most basic parameters available when using MAP. They reported on an elegant open-vessel atmospheric pressure method for the extraction of fat from a variety of foodstuff including meat and meat products. The fat content of samples such as Kam lunch meat, picnic ham, salami, chicken wieners, sausages, and bacon was extracted with recoveries similar to or better than ofﬁcial methods. The most important beneﬁts of the technology are the ability to perform the extraction without any predrying or preprocessing of the sample, its ease of use, low solvent consumption, low energy consumption, and good reproducibility and recoveries. Piñeiro-Avila et al. [20] reported using the microwave-assisted technique in meat samples such as pork fat and codﬁsh liver oil for the saponiﬁcation of animal greases for cholesterol determination. Compared to the conventional reﬂux method that takes 1 h at 908C, this technique offered a fast and safe alternative. The MAP technology was labeled as focused microwave-assisted Soxhlet extraction (FMASE) by Luque-García et al. [21] and was used as a fast technique to monitor oil quality in two types of food samples, namely chicken nuggets and hake ﬁngers. The same group [22] also used this technique for the extraction of lipids in three types of sausage products, the common material in these products being meat and fat from pork. Both papers report advantages over the conventional Soxhlet including time savings (less than 1 h instead of 8 h), minimum amount of solvent used (and associated recycling of solvents), no need to adjust the moisture content of the sample for analysis, and efﬁciency and reproducibility better or comparable to conventional Soxhlet extraction, while offering values in fat composition similar to those obtained by conventional Soxhlet. Martìn-Calero et al. [23] used the MAE for the extraction of 11 different heterocyclic amines in ﬁve different kinds of meat extracts including granulated meat extract, nongranulated meat extract, chicken extract, and two different brands of meat soup cubes. The optimum microwave parameters used were selected using a factorial design model. These authors report achieving the extraction in 5 min and being able to analyze both polar and less polar amines in a single HPLC (highperformance liquid chromatography) run, again adding to the time saving of the method. In aquaculture, there was a reported need for a technique that is rapid and simple for sensory analysis of off-ﬂavors. In this context, Conte et al. [24,25] have developed a microwave distillation-solid phase adsorbent trapping device and reported on the determination of the off-ﬂavors geosmin and methylisoborneol in catﬁsh tissues. In a subsequent paper, this group [26] replaced the solid phase extraction (SPE) step with SPME. Rapid analysis and low detection limits are reported as being the major advantages in these techniques that could replace methods for sample preparation where steam distillation was traditionally used. Lloyd and Grimm [27] also reported on another version of the use of microwave distillation-SPME for the same off-ﬂavor compounds in catﬁsh. Grim et al. [28] have also reported on the use of this technique for the analysis of volatile compounds from ﬁsh tissues and present a qualitative listing of 174 compounds observed in the HS. Batista et al. [29] developed an alternative to the Bligh and Dyer method for the extraction of lipids from ﬁsh and the subsequent determination of their fatty acid composition. The matrices they chose were ﬁllets of mackerel with low lipid and high water contents and livers of cod which represent ﬁsh tissue with very high lipid content, and extracted the lipids from these matrices using

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 66

66

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

MAE. As they reported, both methods gave good reproducibility, but MAP requires less material and solvent as well as less toxic solvent, therefore creating less pollution. Also, the analysis is more rapid and easier to perform. As well, in the Bligh and Dyer method, the amount of water in the sample has a large inﬂuence on the accuracy of the results and special care needs to be placed on the optimum conditions, where this is not the case with MAP.

4.4.2 DAIRY AND EGGS Dairy and egg products have been extracted using the MAP by Paré et al. [19]. The fat content was determined gravimetrically after the extraction in organic solvents that are transparent to microwaves relative to the sample. This reference is especially noteworthy because the authors showed that, for these types of matrices, hydrolysis can be performed in situ thus simplifying the overall procedure while enhancing widely the range of matrices that can be treated by the general procedure. Total hydrolysis and extraction times were 1 min at 60 W for dairy products and 4 min at 30 W for egg powder products. It is important to note that in these extractions, the authors applied the same chemistry principles as in conventional techniques, but replaced heating and extraction steps with microwave treatment of materials immersed in solvents that are relatively transparent to microwave so that most, if not all, of the microwave energy is imparted to the sample. The net result was enhanced efﬁciency. Recoveries were similar or better than in ofﬁcial methods, and advantages were short extraction times, small volumes of solvent used, reduced energy consumption, and temperature control. Luque de Castro et al. [30–32] have also used their MAP-derived extraction procedures (FMASE) for cheese and milk products and showed that there were great advantages. For example, for fat in cheese [30], the hydrolysis time was decreased from 1 h to 10 min, with no neutralization step required and the extraction time was decreased from 6 h to 40 min. For the extraction of lipids from milk samples [31], the results were quantitatively similar to the ones using the conventional Weibull-Berntrop extraction method, but the milk fat extracted using the microwave technique underwent lesser chemical transformation of the triglycerides during the process. Also the microwave extraction time was 50 min as compared to 10 h with the conventional method. Papadakis and Polychroniadou [33] reported using the MAE for 13 different organic acids in Greek cheeses and in sheep milk yoghurt. They found that this method was superior to conventional ones in terms of accuracy and repeatability.

4.4.3 FRUITS

AND

VEGETABLES

Bureau et al. [34] have reported on the use of microwave for the extraction of glycosides from grape juice and grapes. Besides the rapidity and ease of use of the method, another advantage is the extraction of the berries without the need to take out the seeds. Kratchanova et al. [35] exposed orange peels to microwave energy as a pretreatment step. This led to changes in the plant tissue, as shown with scanning electron micrographs, which facilitated the extraction of pectin by conventional extraction method and provided higher extraction yields and improvement in the quality of pectin. These authors [35] and others [1,10,11,36] also reported on the structural changes occurring in the plant cells when submitted to microwave irradiation. Pectin extraction from lime using the MAP was also reported by Fishman et al. [37]. Essential oil extraction from orange peels was reported by Ferhat et al. [38], where they used what they term a microwave Clavenger or microwave-accelerated distillation (MAD) technique and compared it to the conventional hydrodistillation. Using MAD, the extraction time was reduced to 30 min instead of 3 h, yield and product quality were also better. Wang et al. [39] developed another version of solvent-free microwave extraction (SFME) for essential oils, where they used a carbonyl iron powder (CIP) as the absorption medium, therefore making the extraction of essential oil from dried plant material possible without any pretreatment.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 67

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

67

Sun et al. [40] used an MAE for anthocyanins in red raspberries and developed the optimal conditions for their experimental protocol using a central composite design and presented scanning electron micrographs of the disruption in the fruit tissue. Ai et al. [41] used a similar type of approach, i.e., the quadratic general spinning design to develop their microwave extraction parameters to obtain polyphenols from apple pomace. Wang et al. [42] also used apple pomace but for the extraction of pectin. They also used a response surface methodology to set their optimum parameters and concluded that the experimental and predicted yields were in close agreement. A recent critical review by Wang and Weller [43] reported on various techniques, including MAE, for the extraction of nutraceuticals from plants.

4.4.4 CEREALS AND OILSEEDS A microwave hydrolysis procedure that is simple and rapid for furosine determination in cereal and dairy foods was developed by Acquistucci et al. [44]. The short hydrolysis time makes the method appropriate for routine analysis. Furosine found in acid hydrolysates of cereal and dairy products was suggested to be a suitable marker of process and food quality. The MAP-derived FMASE reported by Garcia-Ayuso et al. [32,45,46] has also been used to extract the oil content of different seeds, and the qualitative and quantitative results were in agreement with ofﬁcial methods. Li et al. [47] also used MAP extraction to enhance the oil yields from soybeans. Duvernay et al. [48] also used the technology for antioxidant components from rice bran. Matthäus and Brühl [49] also used the technique in a comparative study of extraction methodologies that involved the determination of the oil content of rapeseed, sunﬂower seed, and soybean. Choi et al. [50] compared the extraction of soluble proteins from various cultivars of soybeans using a microwave-assisted procedure, which proved to be better than the conventional extraction with a shaking water bath. In the microwave-assisted system, the dynamic process was monitored by a response surface methodology. Scanning electron micrographs also show the disruption of the microstructure of the soybean cells. HS analysis is the usual method for the analysis of dimethyl sulﬁde from cereals and canola. Ren [51] used microwave energy to release this volatile compound from the matrix into the HS and obtained results comparable with conventional analytical procedures.

4.4.5 HERBS

AND

SPICES

The use of MAE has been used quite extensively for the extraction of various constituents of plant material, such as pigments, ﬂavors, essential oils, and others, as indicated by the numerous references that can be found on the subject. In fact, these constituted the original applications and appear to still be the most widely used to date in food analysis. Kiss et al. [52] have reported using microwaves for the extraction of pigments from paprika powders. They used a spectral mapping technique to establish the relationship between the efﬁcacy and the selectivity of the extraction and also evaluated 30 extracting solvent mixtures, and concluded that the efﬁcacy and selectivity of MAE depend a lot on the dielectric constant of the solvent. These results were in agreement with the fundamental work carried out in that area [11,13]. Pan et al. [53] studied various parameters such as extraction time, ethanol and ammonia concentrations, liquid=solid ratios, and preleaching time before applying microwaves for the extraction of glycyrrhizic acid from licorice root. The technique proved equivalent to conventional methods, but is preferred since it saves time and solvent and is less labor intensive. Kwon et al. [54,55] reported on using MAP for the fast extraction of ginseng saponins, taking into account parameters such as time, yield, quality, and nature of the ginsenosides extracted. The method proved to be very rapid compared to the 12 h required for the conventional reﬂux methods. The groups of Shu et al. [56] and Yang et al. [57] also reported on the extraction of ginsenosides

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 68

68

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

from ginseng root and made similar observations. Zhao et al. [58] selectively separated two types of ginsenosides by combining macroporous resin adsorption with MAD, using different microwave conditions to achieve the respective desorptions of the ginsenosides. Liu et al. [59] reported on the analysis of saponins as well as ﬂavonoids from Acanthopanax senticosus leaves. Kerem et al. [60] used microwaves for the extraction of saponins from chick peas (Cicer arietinum L). The fast extraction of capsaicinoids from capsicum fruits was extracted with microwave irradiation by Williams et al. [61], who also reported on the main advantages of this technique as being savings in time and energy, and its reliability. Various authors have reported on the extraction of essential oils using MAP. Lucchesi et al. [62–65] used one MAP-HS-derived method [3,5–9], which they labeled SFME, in combination with dry distillation using microwave heating to extract essential oil from spices such as ajowan, cumin, and star anise [62]. They have also extracted essential oils from aromatic herbs such as basil, garden mint, and thyme [63,64], and cardamone essential oil, using a central composite design to evaluate parameters such as extraction time, irradiation power, and moisture content of the seeds [65]. Chemat et al. [66] and Iriti et al. [67], who also reported on the plant structural changes with scanning electron microscopy (SEM), have used what they call microwave-accelerated steam distillation (MASD) for the extraction of essential oil from lavender ﬂowers. Tigrine-Kordjani et al. [68] and Lo Presti et al. [69] used MASD to extract essential oils from rosemary. All of these versions of using microwave energy to perform extractions of essential oils conﬁrmed earlier reports that MAP [1–11] offers the advantages of being rapid, giving a product of similar quality to conventional techniques without the use of solvent, and being environmentally friendly approaches. Chemat et al. [70] studied the MAP kinetics of carvone and limonene from caraway seeds and presented some SEM to show the structural changes in the plant material. Wang et al. [71] reported using SFME [39] combined with CIP for the extraction of essential oils from dried Cuminum cyminum L. and Zanthoxylum bungeanum Maxim. They also used the technique for the extraction of essential oils from dried Illicium verum Hook. f. and Zingiber ofﬁcinale Rosc. [72]. Following our early work [12], piperine from black pepper has been extracted rapidly and selectively using MAP by Raman and Gaikar [73]. They also indicated that the method can be used as a quality control tool for rapid screening of raw pepper. The ﬂavoring agent, vanillin [74] has been extracted using MAP and comparison was made with ultrasound-assisted extraction and conventional extraction. Among advantages of MAP over other techniques are the rapidity of the method as well as giving a better yield for vanillin. Longares-Patrón and Cañizares-Macías [75] also reported on the same observations for the extraction of vanillin and p-hydroxybenzaldehyde from vanilla fragrances. Dandekar and Gaikar [76] used microwaves to extract rapidly and selectively curcuminoids from Curcuma longa (turmeric). In another type of experiment, Chyau and Mau [77] used microwave heating to release and isolate volatile compounds from garlic juice with 2,4-decadienals. Stashenko et al. [78] used microwave hydrodistillation to extract the essential oils from Xylopia aromatica (Lamarck). For the same quantity of oil extracted, the microwave technique took only one quarter of the time used for hydrodistillation. Dai et al. [79,80] reported on the extraction of azadirachtin-related limonoids in neem seed kernel.

4.4.6 FOOD INGREDIENTS Two types of approaches were used to extract fat from chocolate. Simoneau et al. [81] used closedvessel microwave for the extraction of fat from a variety of black chocolate, while ElKhori et al. [82] extracted fat from cocoa power and cocoa nibs using an open-vessel system. Both papers conclude that the MAE for fat content gave similar results as the conventional AOAC (Association of Ofﬁcial Analytical Chemists) method without affecting the triacylglycerol and fatty acid proﬁle and that the extraction could be performed in much shorter time and with less solvent and energy consumption.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 69

17.7.2008 4:02pm Compositor Name: VAmoudavally

Microwave-Assisted Processes in Food Analysis

69

Wang et al. [83] investigated the use of SPME coupled to MAE for the analysis of Veltol (2-methyl 3-hydroxy 4-pyrone) and Veltol-Plus (2-ethyl 3-hydroxy 4-pyrone) (two patented ﬂavor ingredient) in food products. The technique allowed the detection of trace amounts of these ﬂavor ingredients in a variety of food products. Japón-Luján et al. [84] used the microwave extraction technique for biophenols from olive leaves. They report that the extraction time took 8 min compared to the 24 h that takes the conventional method and that the extracts from MAP were clean enough to be injected directly on the liquid chromatograph. Pan et al. [85] used microwaves for the extraction of tea polyphenols and tea caffeine from green tea leaves. They reported better yield than any other conventional method, in only 4 min. Various food constituents were extracted using microwaves: Hong et al. [86] reported on the extraction of phenolic compounds from grape seeds, Liang et al. [87] used the technology to extract polysaccharide from Opuntia milpa Aha, and Liu et al. [88] extracted polysaccharides from Porphyra yezoensis. Shorter extraction times, reduced solvent usage, and better yields are among some of the advantages reported in these papers.

4.4.7 CHARACTERIZATION, REACTION,

AND

BAKERY PRODUCTS

This section will cover some other applications where microwave energy was used for the characterization of food products, including bakery products. Joergensen and Thestrup [89] used microwave heating to hydrolyze proteins in pure and real protein samples with carbohydrates, fats, nucleic acids, and minerals. Their technique reduced the hydrolysis time from 24 h to 10–30 min and the microwave technique gave similar or better results than conventional method. They also reported on the effect of degassing and stabilizing agents. Kovács et al. [90] also used microwaves for the extraction of free amino acids from food, from animal and plant origins and reported on obtaining 10% better yield and reducing the sample preparation time by 66% when using MAE instead of the conventional methods. Yaylayan et al. [91,92] reported on a two-stage MAP procedure where they used focused microwave irradiation to selectively synthesize (Microwave-Assisted Synthesis (MAS)) and quantitatively separate (MAP) Maillard reaction products using L-phenylalanine and glycine=D-glucose were used as model systems in these synthesis=extraction. Stenberg et al. [93] examined the hydrolysis-induced racemization comparing conventional hydrolysis method and microwave techniques. They used the dipeptide aspartame as their model, aspartame at different pH, a soft drink containing aspartame, and a Maillard reaction containing lysine and glucose. Jun and Chun [94] reported on designing a U-column MAP system and verifying its performance on Cape jasmine for the extraction of edible yellow pigments used as a natural coloring agent in Korean food. Their systems gave extraction yields 50% higher than conventional method when thermal energy input and ﬂow rate were identical. Caballo-López and Luque de Castro [95] used a focused microwave digestor to accelerate the removal of free sugars from ﬂour and bread before the analysis of starch, also assisted using microwave energy. The advantages of this technique over the conventional method are as follows: faster (15 min compared to 67 min), shortening of the hydrolysis time (15 min compared to 4 h), reduction in time for the overall method, and the hydrolysis efﬁciencies and precision are similar to the conventional method. Alfaro et al. [96] used the MAP for the extraction of ginger. This study aimed at demonstrating the inﬂuence of the dielectric properties of the matrix vis-à-vis the solvent used to perform the extraction. The authors demonstrated that it is possible to produce extracts in greater quantity and of similar quality to those obtained by Soxhlet and in shorter times by varying the dielectric nature of the extracting medium. This manuscript is especially noteworthy as it provided some further fundamental understanding of the processes involved in MAP. The report demonstrated

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 70

70

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

unequivocally for the ﬁrst time the effect and potential value of controlling the energy density of the system and of using chemicals as a lens to enhance MAP in extraction (as well as in other applications such as chemical synthesis). The primary importance of the dielectric constant over the loss factor was also evidenced. Zhao et al. [97] studied the effect of microwaves on the stability of (all-E)-astaxanthin used as a model compound and concluded that microwaves induced the isomerization of (all-E)-astaxanthin to its Z analogues (preferentially to (13Z)-astaxanthin). This percentage of isomerization increased with increased microwave power and treatment time. This was also observed in the comparison study using ultrasound, but it is reported that ultrasound probably degrades this pigment into colorless compounds. The group of Luque de Castro has published various papers on the extraction of fat from bakery products. Their procedures using the MAP-derived FMASE technique and the types of samples analyzed can be found in Table 4.1 and in the following Refs. [98–101].

4.4.8 FOOD SAFETY A wide range of pesticides within diverse food matrices have been extracted using MAE. Among these, Akhtar et al. reported on the use of MAP for roxarsone in pig tissues [17], for chloramphenicol residues found in egg albumen and yolk [102], and for salinomycin residues found in chicken tissues and eggs [103]. The methods developed were efﬁcient, rapid, economical, and environmentally friendly. A multi-residue screening technique using closed-vessel MAE was developed by Pylypiw et al. [104] for several crop matrices and concluded that the MAP data compared well with conventional data. Prados-Rosales et al. [105] used FMASE for the determination of organochlorine pesticide (OCP) residues in sunﬂower seeds. The extraction parameters were optimized using a factorial design and obtained similar or even better efﬁciencies by comparison to the reference extraction method. The method is reported to be fully automated and has the advantage of being shorter (45 min) than the reference conventional technique that takes 7 h, no sample manipulation before or during extraction procedure is required, higher recoveries, and use of samples as received, and no moisture adjustment are necessary. MAE has also been used for the determination of organophosphorus pesticides in oranges by Bouaid et al. [106], and optimizing the extraction parameters with a factorial experimental design. Falqui-Cao et al. [107] used the technique for the determination of pesticide residues in strawberries and Sanusi et al. [108] used the technique for the extraction of pyrethroid residues in strawberries. Chen et al. [109] used microwaves for the extraction of residual dichlorvos in vegetable and fruit samples, whereas Padrón-Sanz et al. [110] reported on applying microwaveassisted micellar extraction methodology for a mixture of eight organophosphorus pesticides from tomato, lettuce, and pepper samples. El-Saeid et al. [111] did a comparative study of various extraction methods for the determination of atrazine in frozen vegetables, fruit juice, and jam. Among the advantages reported in these studies using MAP are the rapidity of the method, its simplicity of usage, and similar or better results than the conventional technologies. Other food matrices that were extracted using microwave energy are corn [112] and wheat [113] samples for the removal of zearalenone, beans for the extraction of fenitrothion residues [114], and sesame seeds [115] for the determination of 16 organochlorine insecticides. Fish tissues have also been subjected to MAP. Weichbrodt et al. [116] reported using focused open- and closed-vessel MAEs for organochlorine compounds in cod liver and ﬁsh ﬁllets. For the open-vessel protocol, the solvent mixture removed the water from the sample matrix via an azeotropic distillation. As for closed-vessel applications, the extraction had to be performed in two stages: the ﬁrst one to remove the co-extracted water content manually and the second step for the quantitative extraction of the organochlorine compounds with the pure solvent. Wittmann et al. [117] developed a procedure for the determination of trichlorobenzenes in ﬁsh samples and reported

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 71

17.7.2008 4:02pm Compositor Name: VAmoudavally

71

Microwave-Assisted Processes in Food Analysis

TABLE 4.1 Survey of MAE Applications in Food Analysis (1995–2006) Foodstuff

Analysis

References

Meat, Poultry, and Fish Kam lunch meat Picnic ham Salami Chicken wieners Sausages Bacon Pork fat Codﬁsh liver oil Chicken nuggets Hake ﬁngers Sausage products (containing meat and fat from pork) Granulated meat extract Nongranulated meat extract Chicken extract Meat soup cubes

Catﬁsh tissues Fish tissue

Fillets of mackerel Livers of cod

Fat

[11,19]

Saponiﬁcation of animal greases

[20]

Oil quality

[21]

Lipids

[22]

11 heterocyclic amines 2-Aminodipirydo[1,2-a:30 ,20 -d]imidazole (Glu-P2) 2-Amino-3-methylimidazo[4,5-f]quinoline (IQ) 2-Amino-3,4-dimethylimidazo[4,5-f]quinoline (MeIQ) 2-Amino-3,8-dimethylimidazo[4,5-f]quinoxaline (MeIQx) 3-Amino-1-methyl-5H-pyrido[4,3-b]indole (Trp-P-2) 3-Amino-1,4-dimethyl-5H-pyrido[4,3-b]indole (Trp-P-1) 2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine (PhIP) 2-Amino-9H-pyrido[2,3-b]indole (AaC) 2-Amino-3-methyl-9H-pyrido[2,3-b]indole (MeAaC) 9H-Pyrido[4,3-b]indole (NH) 1-Methyl-9H-pyrido[4,3-b]indole (H) Geosmin Methylisoborneol This reference contains a qualitative listing of 174 compounds Geosmin Methylisoborneol Lipids

[23]

Fat

[11,19]

Fat

[30,32]

Lipids Organic acids Citric acid Orotic acid Tartaric acid Pyruvic acid Succinic acid Lactic acid

[31,32] [33]

[24–27] [28]

[29]

Dairy and Eggs Milk powder Egg powder Cheddar cheese Powder cheese (semicured) Cream cheese (fresh cheese) Cured cheese Milk (cow, goat, sheep) Greek cheeses Sheep milk yoghurt

(continued )

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 72

72

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

TABLE 4.1 (continued) Survey of MAE Applications in Food Analysis (1995–2006) Foodstuff

Dairy Products Pasteurized milk UHT milk UHT skimmed milk Milk powder Concentrate milk Mozzarella cheese Imitation mozzarella cheese Processed cheese

Analysis

References

Acetic acid Fumaric acid Propionic acid Butyric acid Isovaleric acid Valeric acid Hippuric acid Furosine

[44]

Glycosides

[34]

Pectin Pectin Essential oils Essential oils

[35,36] [37] [38] [39]

Anthocyanins Polyphenols Pectin Nutraceuticals

[40] [41] [42] [43]

Furosine

[44]

Oil Oil

[45] [46]

Oil Antioxidant components Oil content

[47] [48] [49]

Soluble proteins Dimethyl sulﬁde

[50] [51]

Fruits and Vegetables Grapes Grape juice Orange peels Lime Orange peels Dried menthol mint Orange peels Red raspberries Apple pomace Plants Cereals and Oilseeds Cereals Products Semolina Spaghetti Short pasta Pasta with eggs Rusks Biscuits Olives Soybean Sunﬂower Rape seeds Soybean Rice bran Rapeseed Sunﬂower Soybean Soybean Wheat Paddy

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 73

17.7.2008 4:02pm Compositor Name: VAmoudavally

73

Microwave-Assisted Processes in Food Analysis

TABLE 4.1 (continued) Survey of MAE Applications in Food Analysis (1995–2006) Foodstuff

Analysis

References

Barley Canola Herbs and Spices Paprika (Capsicum annuum L.) Licorice root Ginseng root

Acanthopanax senticosus leaves Chickpea (Cicer arietinum L.) Capsicum fruit Ajowan (Carum ajowan, Apiaceae) Cumin (Cuminum cyminum, Umbelliferae) Star anise (Illicium anisatum, Illiciaceae) Basil (Ocimum basilicum L.) Garden mint (Mentha crispa L.) Thyme (Thymus vulgaris L.) Cardamone (Elletaria cardamomum L.) Lavender ﬂowers (Lavandula angustifolia Mill, Lamiaceae) Rosemary (Rosmarinus ofﬁcinalis L.) Caraway seeds (Carum carvi L.) Cuminum cyminum L. Zanthoxylum bungeanum Maxim Illicium verum Hook. f.

Zinziber ofﬁcinale Rosc.

Black pepper (Piper nigrum) Vanilla planifolia Curcuma longa (turmeric) Garlic juice with 2,4-decadienals

Pigments Glycyrrhizic acid Ginsenosides 20(S)-Protopanaxdiol and 20(S)-protopanaxtriol type Ginsenosides Flavonoids Saponins Capsaicinoids Essential oil

[11,13,52] [53] [54–57]

Essential oil

[1–9,11,63,64]

Essential oil Essential oil

[65] [66,67]

Essential oil Carvone Limonene Essential oil

[68,69]

Essential oils (E)-anethole Limonene Linalool a-Pinene Essential oils Zingiberene ar-Curcumene b-Bisabolene b-Sesquiphellandrene, camphene b-Phellandrene Piperine Vanillin p-Hydroxybenzaldehyde Curcuminoids (curcumin, demethoxy curcumin, bis(demethoxy) curcumin) Volatile compounds Sulfur dioxide Allyl mercaptan Isopropyl alcohol

[58] [59] [60] [61] [62]

[70] [71] [72]

[72,96]

[12,73] [74,75] [75] [76] [77]

(continued )

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 74

74

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

TABLE 4.1 (continued) Survey of MAE Applications in Food Analysis (1995–2006) Foodstuff

Xylopia aromatica (Lamarck) Neem seed kernel

Analysis Hexanal Allyl alcohol Allyl sulﬁde 2-Pentylfuran Methyl allyl disulﬁde Dimethyl trisulﬁde Nonanal (E)-2-Octenal Dithio(1-propenyl)propionate Diallyl disulﬁde 1,2-Dithiacyclopent-3-ene n-Hexanethiol Methyl benzyl sulﬁde Dihydro-2(3H)-thiophenthione 3-Vinyl-4H-1,2-dithiin (E,Z)-2,4-Decadienal 2-Vinyl-1,3-dithiane (E,E)-2,4-Decadienal Hexanoic acid 2-Vinyl-4H-1,3-dithiin Essential oil Azadirachtin-related limonoids

References

[78] [79,80]

Food Ingredients Chocolate and chocolate formulations Cocoa powder Theobroma cocoa nibs Coffee Beverages (coca cola) Chewing gums Potato chips Olive leaves

Green tea leaves Grape seeds Opuntia milpa Aha Porphyra yezoensis

Fat (dark chocolate, cocoa products, cocoa liquor, deodorized cocoa butter) Fat

[81]

Veltol and Veltol-Plus

[83]

Oleuropein Verbacoside Apigenin-7-glucoside Luteolin-7-glucoside Tea polyphenols Tea caffeine Phenolic compounds Polysaccharides Polysaccharides

[84]

[82]

[85] [86] [87] [88]

Characterization, Reaction, and Bakery Products Casein Gelatine BioProtein Salami (Kaisers salami) Cheese (Trappista and Szekszard) Broccoli

Proteins

[89]

Free amino acids Asp

[90]

Thr

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 75

17.7.2008 4:02pm Compositor Name: VAmoudavally

75

Microwave-Assisted Processes in Food Analysis

TABLE 4.1 (continued) Survey of MAE Applications in Food Analysis (1995–2006) Foodstuff Cauliﬂower

Maillard reactions L-Phenylalanine model

Glycine model

Dipeptide aspartame Diet coke Maillard reaction Cape jasmine (Gardenia jasminoides Ellis) Flour Bread Ginger (Zingiber ofﬁcinale) (all-E)-Astaxanthin Bakery products (cookies and snacks)

Analysis

References

Ser Glu Gly Ala Val Met Ile Leu Tyr Phe Gaba His Lys Arg Pro Benzeneacetaldehyde 2-(50 -(Hydroxymethyl)-20 -formylpyrrol-10 -yl)-3phenylpropionic acid lactone 3,5-diphenylpyridine 1,6-Dimethyl-2(1H)-pyrazinone 1,5,6-Trimethyl-2(1H)-pyrazinone 5-Hydroxy-1,3-dimethyl-2(1H)-qunoxalinone Racemization of amino acids

[91,92]

Pigments Free sugars

[94] [95]

Total ginger extracts Stability study Total fat and trans fatty acids

[96] [97] [98–101]

Chloramphenicol Salinomycin

[102] [103]

Dacthal Chlorpyrifos Chlorpyrifos Chlorothalonil Diazinon Permethrin Methoxychlor Azinphos-methyl a-HCH b-HCH g-HCH (lindane) d-HCH Endosulfan I

[104]

[93]

Food Safety Egg albumen and yolk Chicken Egg Crops of Beets Cucumber Lettuce Peppers Tomatoes

Sunﬂower seeds

[105]

(continued )

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 76

76

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

TABLE 4.1 (continued) Survey of MAE Applications in Food Analysis (1995–2006) Foodstuff

Orange peel

Strawberries

Strawberries

Tomatoes Strawberries Pakchoi Tomatoes Lettuce Pepper Frozen vegetables Fruit juice Jam Corn Wheat Beans, white, and black (Vigna unguiculata L.) Sesame seeds

Cod liver Fish ﬁllets

Analysis Endosulfan II Endosulfan sulfate Atrazine Parathion-methyl Chlorpyriphos Fenamiphos Methidathion Carbendazim Diethofencarb Azoxystrobine Napropamide Bupirimate Acrinathrin Bifenthrin L-cyhalothrin deltamethrin Dichlorvos

References

[106]

[107]

[108]

[109]

Organophosphorus pesticides

[110]

Atrazine

[111]

Zearalenone

[112,113]

Fenitrothion (O,O-dimethyl-O-4-nitrom-tolyl phosphorothionate) a-HCH b-HCH g-HCH d-HCH Heptachlor Aldrin Heptachlor epoxide Endosulfan I Dieldrin p,p0 -DDE Endrin Endosulfan II p,p0 -DDD Endosulfan sulfate p,p0 -DDT Methoxychlor Polychlorinated biphenyls DDT Toxaphene Chlordane Hexachlorobenzene Hexachlorocyclohexanes Dieldrin

[114] [115]

[116]

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 77

17.7.2008 4:02pm Compositor Name: VAmoudavally

77

Microwave-Assisted Processes in Food Analysis

TABLE 4.1 (continued) Survey of MAE Applications in Food Analysis (1995–2006) Foodstuff Cod

Salmon Eel Sea bass Green mussels (Perna viridis) Turbot Salmon

Sausage Lamb liver Pumpkin seed oils

Analysis

References

1-Bromo-4-chlorobenzene (4-BCB) 1,4-Dibromobenzene (4-DBB) 1,3,5-Trichlorobenzene 1,2,4-Trichlorobenzene 1,2,3-Trichlorobenzene Polybrominated diphenyl ethers (PBDE) congeners (47, 99, and 100)

[117]

Benz[a]anthracene Benzo[b]ﬂuoranthene Benzo[k]ﬂuoranthene Benzo[a]pyrene Dibenz[a,h]anthracene Indene[1,2,3-cd]pyrene Benzo[a]pyrene

[119]

16 PAHs Naphthalene Acenaphthylene Acenaphthene 9H-Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benzo[a]anthracene Chrysene Benzo[b]ﬂuoranthene Benzo[k]ﬂuoranthene Benzo[a]pyrene Indeno[1,2,3-cd]pyrene Dibenzo[ah]anthracene Benzo[ghi]perylene

[121]

[118]

[120]

that the results were signiﬁcantly dependent on the extraction parameters, i.e., they obtained better results at 5 bar than at 1 bar, but in the former case sulfur formation was observed. Bayen et al. [118] developed a method that is rapid, sensitive, and gives good quantitative results for the analysis of polybrominated diphenyl ethers (PBDE) congeners in marine biological tissues. Pena et al. [119] developed an optimized MAP method that is quick and efﬁcient for all regulated polycyclic aromatic hydrocarbons (PAHs) in ﬁsh samples. They simultaneously hydrolyzed the fats in the samples and extracted the hydrocarbons of interest in hexane. García-Falcón et al. [120] focused on a speciﬁc PAH and developed a simple, fast, and inexpensive technique for benzo[a]pyrene in foods with a high fat content. In the same context of human exposure to PAHs, Gfrerer and Lankmayr [121] developed a microwave-assisted saponiﬁcation method for pumpkin seed oil pretreatment, for the determination of 16 priority PAHs followed by a liquid-liquid extraction step. The method offers a fast and safe alternative to traditional saponiﬁcation processes.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 78

78

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

Other papers dealing with the comparison of extraction technologies that include MAE in their studies for the extraction of various contaminants in food systems include the works of Ahmed [122], Björklund et al. [123], and Rosenblum et al. [124].

REFERENCES 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.

12. 13. 14. 15. 16.

17.

18.

19.

20. 21.

22.

23.

24.

Paré, J.R.J., Sigouin, M., and Lapointe, J., U.S. Patent 5,002,784, 1991. Paré, J.R.J., U.S. Patent 5,338,557, 1994. Paré, J.R.J., U.S. Patent 5,377,426, 1995. Paré, J.R.J., U.S. Patent 5,458,897, 1995. Paré, J.R.J., U.S. Patent 5,519,947, 1996. Paré, J.R.J., U.S. Patent 5,675,909, 1997. Paré, J.R.J., U.S. Patent 5,732,476, 1998. Paré, J.R.J., U.S. Patent 5,884,417, 1999. Paré, J.R.J., Bélanger, J.M.R., and Punt, M.M., U.S. Patent 6,061,926, 2000. Paré, J.R.J., Bélanger, J.M.R., and Stafford, S.S., Microwave-assisted process (MAPe): A new tool for the analytical laboratory, Trends Anal. Chem., 13(4), 176, 1994. Paré, J.R.J. and Bélanger, J.M.R., Microwave-assisted process (MAPe): Principles and applications, in Instrumental Methods in Food Analysis, J.R.J. Paré and J.M.R. Bélanger (Eds.), Elsevier Science, Amsterdam, the Netherlands, Chapter 10, pp. 395–420, 1997. Bélanger, J.M.R., Bélanger, A., and Paré, J.R.J., Procédé assisté par micro-ondes (MAPe): Oléorésines de graines de céleri et de poivre noir, Riv. Ital. EPPOS, 5, 114, 1994. Bélanger, J.M.R. and Paré, J.R.J., Procédé assisté par micro-ondes (MAPe): Application au paprika, Riv. Ital. EPPOS, 5, 126, 1994. Bélanger, J.M.R., Aubry, C., Chagonda, L., and Paré, J.R.J., Procédé assisté par micro-ondes (MAPe): Extraction d’huiles essentielles au Zimbabwe, Riv. Ital. EPPOS, 5, 136, 1994. Bélanger, A., Landry, B., Dextraze, L., Bélanger, J.M.R., and Paré, J.R.J., Extraction et détermination de composés volatils de l’ail (Allium sativum), Riv. Ital. EPPOS, 2, 455, 1991. Paré, J.R.J. and Bélanger, J.M.R., Microwave-assisted process (MAPe): Applications to the extraction of natural products, Proc. 28th Microwave Power Symp., International Microwave Power Institute, Manassas, VA, 62, 1993. Akhtar, M.H., Croteau, L., Bélanger, J.M.R., and Paré, J.R.J., High performance liquid chromatography determination following microwave-assisted extraction of 3-nitro-4-hydroxyphenylarsonic acid from swine liver, kidney and muscle, J. Liq. Chromatogr., 17, 2971, 1994. Paré, J.R.J., Bélanger, J.M.R., Bélanger, A., and Ramarathnam, N., A rapid method for monitoring food volatiles, in Food Science and Nutrition, G. Charalambous (Ed.), Elsevier, Amsterdam, the Netherlands, 141, 1992. Paré, J.R.J., Matni, G., Bélanger, J.M.R., Li, K., Rule, C., Thibert, B., Yaylayan, V., Liu, Z., Mathé, D., and Jacquault, P., Use of the microwave-assisted process in extraction of fat from meat, dairy, and egg products under atmospheric pressure conditions, J. AOAC Int., 80(4), 928, 1997. Piñeiro-Avila, G., Salvador, A., and de la Guardia, M., Microwave-assisted saponiﬁcation of animal greases for cholesterol determination, Anal. Chim. Acta, 371, 297, 1998. Luque-García, J.L., Velasco, J., Dobarganes, M.C., and Luque de Castro, M.D., Fast quality monitoring of oil from prefried and fried foods by focused microwave-assisted Soxhlet extraction, Food Chem., 76, 241, 2002. Priego-Lòpez, E., Velasco, J., Dobarganes, M.C., Ramis-Ramos, G., and Luque de Castro, M.D., Focused microwave-assisted Soxhlet extraction: An expeditive approach for the isolation of lipids from sausage products, Food Chem., 83, 143, 2003. Martìn-Calero, A., Pino, V., Ayala, J.H., Gonzalez, V., and Afonso, A.M., Focused microwave-assisted extraction and HPLC with electrochemical detection to determine heterocyclic amines in meat extracts, J. Liq. Chromatogr. Relat. Technol., 30, 27, 2007. Conte, E.D., Shen, C.Y., and Miller, D.W., Microwave distillation—solid phase adsorbent trapping device for the determination of off-ﬂavors, geosmin and methylisoborneol, in catﬁsh tissue below their rejection levels, Anal. Chem., 68, 2713, 1996.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 79

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

79

25. Conte, E.D., Shen, C.Y., Perschbacher, P.W., and Miller, D.W., Determination of geosmin and methylisoborneol in catﬁsh tissue (Ictalurus punctatus) by microwave-assisted distillation-solid phase adsorbent trapping, J. Agric. Food Chem., 44, 829, 1996. 26. Zhu, M., Aviles, F.J., Conte, E.D., Miller, D.W., and Perschbacher, P.W., Microwave mediated distillation with solid-phase micro extraction: determination of off-ﬂavors, geomin and methylisoborneol, in catﬁsh tissue, J. Chromatogr. A, 833, 223, 1999. 27. Lloyd, S.W. and Grimm, C.C., Analysis of 2-methylisoborneol and geosmin in catﬁsh by microwave distillation-solid-phase micro extraction, J. Agric. Food Chem., 47, 164, 1999. 28. Grimm, C.C., Lloyd, S.W., Batista, R., and Zimba, P.V., Using microwave distillation-solid-phase micro extraction-gas chromatography-mass spectrometry for analyzing ﬁsh tissue, J. Chromatogr. Sci., 38(7), 289, 2000. 29. Batista, A., Vetter, W., and Luckas, B., Use of focused open vessel microwave-assisted extraction as prelude for the determination of the fatty acid proﬁle of ﬁsh—A comparison with results obtained after liquid–liquid extraction according to Bligh and Dyer, Eur. Food Res. Technol., 212, 377, 2001. 30. Garcia-Ayuso, L.E., Velasco, J., Dobarganes, M.C., and Luque de Castro, M.D., Accelerated extraction of the fat content in cheese using a focused microwave-assisted Soxhlet device, J. Agric. Food Chem., 47, 2308, 1999. 31. Garcia-Ayuso, L.E., Velasco, J., Dobarganes, M.C., and Luque de Castro, M.D., Double use of focused microwave irradiation for accelerated matrix hydrolysis and lipid extraction in milk samples, Int. Dairy J., 9, 667, 1999. 32. Garcia-Ayuso, L.E. and Luque de Castro, M.D., Employing focused microwaves to counteract conventional Soxhlet extraction drawbacks, Trends Anal. Chem., 20(1), 28, 2001. 33. Papadakis, E.N. and Polychroniadou, A., Application of a microwave-assisted extraction method for the extraction of organic acids from Greek cheeses and sheep milk yoghurt and subsequent analysis by ionexclusion liquid chromatography, Int. Dairy J., 15, 165, 2005. 34. Bureau, S., Razungles, A., Baumes, R., and Boyonove, C., Glycosylated ﬂavor precursor extraction by microwaves from grape juice and grapes, J. Food Sci., 61(3), 557, 1996. 35. Kratchanova, M., Pavlova, E., and Panchev, I., The effect of microwave heating of fresh orange peels on the fruit tissue and quality of extracted pectin, Carbohydr. Polym., 56, 181, 2004. 36. Zhongdong, L., Guohua, W., Yunchang, G., and Kennedy, J.F., Image study of pectin extraction from orange skin assisted by microwave, Carbohydr. Polym., 64, 548, 2006. 37. Fishman, M.L., Chau, H.K., Hoagland, P.D., and Hotchkiss, A.T., Microwave-assisted extraction of lime pectin, Food Hydrocolloids, 20, 1170, 2006. 38. Ferhat, M.A., Meklati, B.Y., Smadja, J., and Chemat, F., An improved microwave Clevenger apparatus for distillation of essential oils from orange peel, J. Chromatogr. A, 1112, 121, 2006. 39. Wang, Z.M., Ding, L., Wang, L., Feng, J., Li, T.C., Zhou, X., and Zhang, H.Q., Fast determination of essential oil from dried menthol mint and orange peel by solvent free microwave extraction using carbonyl iron powder as the microwave absorption medium, Chin. J. Chem., 24, 649, 2006. 40. Sun, Y., Liao, X., Wang, Z., Hu, X., and Chen, F., Optimization of microwave-assisted extraction of anthocyanins in red raspberries and identiﬁcation of anthocyanin of extracts using high-performance liquid chromatography-mass spectrometry, Eur. Food Res. Technol., DOI 10.1007=s00217-006-0447-1, 2006. 41. Ai, Z., Quo, J., Wang, Y., Liu, Y., and Zhao, Q., Microwave-assisted extraction technique of apple polyphenols in apple pomace, Nongye Gongcheng Xuebao=Trans. Chin. Soc. Agric. Eng., 22(6), 188, 2006. 42. Wang, S., Chen, F., Wu, J., Wang, Z., Liao, X., and Hu, X., Optimization of pectin extraction assisted by microwave from apple pomace using response surface methodology, J. Food Eng., 78, 693, 2007. 43. Wang, L. and Weller C.L., Recent advances in extraction of nutraceuticals from plants, Trends Food Sci. Technol., 17, 300, 2006. 44. Acquistucci, R., Panﬁli, G., and Marconi, E., Application of the microwave hydrolysis to furosine determination in cereal and dairy foods, J. Agric. Food Chem., 44, 3855, 1996. 45. Garcia-Ayuso, L.E. and Luque de Castro, M.D., A multivariate study of the performance of a microwaveassisted Soxhlet extractor for olive seeds, Anal. Chim. Acta, 382, 309, 1999. 46. Garcia-Ayuso, L.E., Velasco, J., Dobarganes, M.C., and Luque de Castro, M.D., Determination of the oil content of seeds by focused microwave-assisted Soxhlet extraction, Chromatographia, 52(1=2), 103, 2000.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 80

80

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

47. Li, H., Pordesimo, L.O., Weiss, J., and Wilhelm, L.R., Microwave and ultrasound assisted extraction of soybean oil, Trans. ASAE, 47(4), 1187, 2004. 48. Duvernay, W.H., Assad, J.M., Sabliov, C.M., Lima, M., and Xu, Z., Microwave extraction of antioxidant components from rice bran, Pharm. Eng., 25(4), 126, 2005. 49. Matthäus, B. and Brühl, L., Comparison of different methods for the determination of the oil content of oilseeds, J. Am. Oil Chem. Soc., 78(1), 95, 2001. 50. Choi, I., Choi, S.J., Chun, J.K., and Moon, T.W., Extraction yield of soluble protein and microstructure of soybean affected by microwave heating, J. Food Process. Preservation, 30, 407, 2006. 51. Ren, Y.L., Comparison of solvent extraction and microwave extraction for release of dimethyl sulﬁde from cereals and canola, J. Agric. Food Chem., 49, 1737, 2001. 52. Kiss, G.A.C., Forgács, E., Cserháti, T., Mota, T., Morais, H., and Ramos, A., Optimisation of the microwave-assisted extraction of pigments from paprika (Capsicum annuum L.) powders, J. Chromatogr. A, 889, 41, 2000. 53. Pan, X., Liu, H., Jia, G., and Shu, Y.Y., Microwave-assisted extraction of glycyrrhizic acid from licorice root, Biochem. Eng. J., 5, 173, 2000. 54. Kwon, J.-H., Bélanger, J.M.R., Paré, J.R.J., and Yaylayan, V.A., Application of the microwave-assisted process (MAPe) to the fast extraction of ginseng saponins, Food Res. Int., 36, 491, 2003. 55. Kwon, J.-H., Lee, G.-D., Bélanger, J.M.R., and Paré, J.R.J., Int. J. Food Sci. Technol., 38, 615, 2003. 56. Shu, Y.Y., Ko, M.Y., and Chang, Y.S., Microwave-assisted extraction of ginsenosides from ginseng root, Microchem. J., 74, 131, 2003. 57. Yang, Y., Chen, L., Zhang, X.-X., and Guo, Z., Microwave assisted extraction of major active ingredients in Panax quinquefolium L., J. Liq. Chromatogr. Relat. Technol., 27(20), 3203, 2004. 58. Zhao, Y., Chen, B., and Yao, S., Separation of 20(S)-protopanaxdiol type ginsenosides and 20(S)protopanaxtriol type ginsenosides with the help of macroporous resin adsorption and microwave assisted desorption, Sep. Purif. Technol., 52, 533, 2007. 59. Liu, Z., Yan, G., Bu, F., Sun, J., Hu, X., Zhang, H., and Liu, Z., Analysis of chemical composition of Acanthopanax senticosus leaves applying high-pressure microwave-assisted extraction, Chem. Analityczna, 50(5), 851, 2005. 60. Kerem, Z., German-Shashoua, H., and Yarden, O., Microwave-assisted extraction of bioactive saponins from chickpea (Cicer arietinum L.), J. Sci. Food Agric., 85, 406, 2005. 61. Williams, O.J., Raghavan, G.S.V., Orsat, V., and Dai, J., Microwave-assisted extraction of capsaicinoids from capsicum fruit, J. Food Biochem., 28(2), 113, 2004. 62. Lucchesi, M.E., Chemat, F., and Smadja, J., An original solvent free microwave extraction of essential oils from spices, Flavour Fragrance J., 19, 134, 2004. 63. Lucchesi, M.E., Chemat, F., and Smadja, J., Solvent-free microwave extraction of essential oil from aromatic herbs: comparison with conventional hydro-distillation, J. Chromatogr. A, 1043, 323, 2004. 64. Lucchesi, M.E., Chemat, F., and Smadja, J., Solvent-free microwave extraction: An innovative tool for rapid extraction of essential oil from aromatic herbs and spices, J. Microwave Power Electromagnetic Energy, 39(3–4), 135, 2004. 65. Lucchesi, M.E., Smadja, J., Bradshaw, S., Louw, W., and Chemat, F., Solvent free microwave extraction of Elletaria cardamomum L.: A multivariate study of a new technique for the extraction of essential oil, J. Food Process Eng., 79, 1079, 2007. 66. Chemat, F., Lucchesi, M.E., Smadja, J., Favretto, L., Colnaghi, G., and Visinoni, F., Microwave accelerated steam distillation of essential oil from lavender: A rapid, clean and environmentally friendly approach, Anal. Chim. Acta, 555, 157, 2006. 67. Iriti, M., Colnaghi, G., Chemat, F., Smadja, J., Faoro, F., and Visinoni, F., Histo-cytochemistry and scanning electron microscopy of lavender glandular trichomes following conventional and microwaveassisted hydrodistillation of essential oils: A comparative study, Flavour Fragrance J., 21, 704, 2006. 68. Tigrine-Kordjani, N., Meklati, B.Y., and Chemat, F., Microwave ‘‘dry’’ distillation as a useful tool for extraction of edible essential oils, Int. J. Aromatherapy, 16, 141, 2006. 69. Lo Presti, M., Ragusa, S., Trozzi, A., Dugo, P., Visinoni, F., Fazio, A., Dugo, G., and Mondello, L., A comparison between different techniques for the isolation of rosemary essential oil, J. Sep. Sci., 28, 273, 2005. 70. Chemat, F., Aït-Amar, H., Lagha, A., and Esveld, D.C., Microwave-assisted extraction kinetics of terpenes from caraway seeds, Chem. Eng. Process, 44, 1320, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 81

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

81

71. Wang, Z., Ding, L., Li, T., Zhou, X., Wang, L., Zhang, H., Liu, L., Li, Y., Liu, Z., Wang, H., Zeng, H., and He, H., Improved solvent-free microwave extraction of essential oil from dried Cuminum cyminum L. and Zanghoxylum bungeanum Maxim, J. Chromatogr. A, 1102, 11, 2006. 72. Wang, Z., Wang, L., Li, T., Zhou, X., Ding, L., Yu, Y., Yu, A., and Zhang, H., Rapid analysis of the essential oils from dried Illicium verum Hook. f. and Zingiber ofﬁcinale Rosc. by improved solvent-free microwave extraction with three types of microwave-absorption medium, Anal. Bioanal. Chem., 386, 1863, 2006. 73. Raman, G. and Gaikar V.G., Microwave-assisted extraction of piperine from Piper nigrum, Ind. Eng. Chem. Res., 41, 2521, 2002. 74. Sharma, A., Verma, S.C., Saxena, N., Chadda, N., Singh, N.P., and Sinha, A.K., Microwave- and ultrasound-assisted extraction of vanillin and its quantiﬁcation by high-performance liquid chromatography in Vanilla planifolia, J. Sep. Sci., 29, 613, 2006. 75. Longares-Patrón, A. and Cañizares-Macías, M.-P., Focused microwaves-assisted extraction and simultaneous spectrophotometric determination of vanillin and p-hydroxybenzaldehyde form vanilla fragans, Talanta, 69, 882, 2006. 76. Dandekar, D.V. and Gaikar, V.G., Microwave assisted extraction of curcuminoids from Curcuma longa, Sep. Sci. Technol., 37(11), 2669, 2002. 77. Chyau, C.-C. and Mau, J.-L., Release of volatile compounds from microwave heating of garlic juice with 2,4-decadienals, Food Chem., 64, 531, 1999. 78. Stashenko, E.E., Jaramillo, B.E., and Martínez, J.R., Analysis of volatile secondary metabolites from Colombian Xylopia aromatica (Lamarck) by different extraction and headspace methods and gas chromatography, J. Chromatogr. A, 1025, 105, 2004. 79. Dai, J., Yaylayan, V.A., Raghavan, G.S.V., and Paré, J.R.J., Extraction and colorimetric determination of azadirachtin related limonoids in neem seed kernel, J. Agric. Food Chem., 47, 3738, 1999. 80. Dai, J., Yaylayan, V.A., Raghavan, G.S.V., Paré, J.R.J., Liu, Z., and Bélanger, J.M.R., Inﬂuence of operating parameters on the use of the microwave-assisted process (MAP) for the extraction of azadirachtin related limonoids from neem under atmospheric pressure conditions, J. Agric. Food Chem., 49, 4584, 2001. 81. Simoneau, C., Naudin, C., Hannaert, P., and Anklan, E., Comparison of classical and alternative extraction methods for the quantitative extraction of fat from plain chocolate and the subsequent application to the detection of added foreign fats to plain chocolate formulations, Food Res. Int., 33, 733, 2000. 82. ElKhori, S., Paré, J.R.J., Bélanger, J.M.R., and Pérez, E., The microwave-assisted process (MAPe) : Extraction and determination of fat from cocoa powder and cocoa nibs, J. Food Eng., 79, 1110, 2007. 83. Wang, Y., Bonilla, M., McNair, H.M., and Khaled, M., Solid phase microextraction associated with microwave assisted extraction of food products, J. High Resolution Chromatogr., 20, 213, 1997. 84. Japón-Luján, R., Luque-Rodríguez, J.M., and Luque de Castro, M.D., Multivariate optimisation of the microwave-assisted extraction of oleuropein and related biophenols from olive leaves, Anal. Bioanal. Chem., 385, 753, 2006. 85. Pan, X., Niu, G., and Liu, H., Microwave-assisted extraction of tea polyphenols and tea caffeine from green tea leaves, Chem. Eng. Process, 42, 129, 2003. 86. Hong, N., Yaylayan, V.A., Raghavan, V.G.S., Paré, J.R.J., and Bélanger, J.M.R., Microwave-assisted extraction of phenolic compounds from grape seed, Nat. Prod. Lett., 15(3), 197, 2001. 87. Liang, Y., Ying, M., Lu, Y., Lu, X., and Zhang, Y., Microwave-assisted technology for extracting Opuntia polysaccharide, Nongye Gongcheng Xuebao=Trans. Chin. Soc. Agric. Eng., 22(7), 159, 2006. 88. Liu, Q., Yang, X., Deng, H., Yang, L., and Wu, Y., Technology for extraction of polysaccharide by using microwave technology from Porphyra yezoensis, Nongye Gongcheng Xueboa=Trans. Chin. Soc. Agric. Eng., 21(2), 153, 2005. 89. Joergensen, L. and Thestrup, H.N., Determination of amino acids in biomass and protein samples by microwave hydrolysis and ion-exchange chromatography, J. Chromatogr. A, 706, 421, 1995. 90. Kovács, Á., Ganzler, K., and Simon-Sarkadi, L., Microwave-assisted extraction of free amino acids from foods, Z. Lebensmittel Untersuchung Forsch. A, 207, 26, 1998. 91. Yaylayan, V.A., Matni, G., Paré, J.R.J., and Bélanger, J.M.R., Microwave-assisted synthesis and extraction of selected Maillard reaction products, J. Agric. Food Chem., 45, 149, 1997. 92. Yaylayan, V.A. and Keyhani, A., Application of microwave-assisted process and Py-GC-MS to the analysis of Maillard reaction products, ACS Symp. Ser., 705, 223, 1998.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 82

82

17.7.2008 4:02pm Compositor Name: VAmoudavally

Handbook of Food Analysis Instruments

93. Stenberg, M., Marko-Varga, G., and Öste, R., Racemization of amino acids during classical and microwave oven hydrolysis—Application to aspartame and a Maillard reaction system, Food Chem., 74, 217, 2001. 94. Jun, S.J. and Chun, J.K., Design of U-columns microwave-assisted extraction system and its application to pigment extraction from food, Trans. IChemE. 76, Part C, 231, 1998. 95. Caballo-López, A. and Luque de Castro, M.D., Fast microwave-assisted free sugars washing and hydrolysis pre-treatment for the ﬂow injection determination of starch in food, Talanta, 59, 837, 2003. 96. Alfaro, M.J., Bélanger, J.M.R., Padilla, F.C., and Paré, J.R.J., Inﬂuence of solvent, matrix dielectric properties, and applied power on the liquid-phase microwave-assisted processes (MAPe) extraction of ginger (Zingiber ofﬁcinale), Food Res. Int., 36, 499, 2003. 97. Zhao, L., Zhao, G., Chen, F., Wang, Z., Wu, J., and Hu, X., Different effects of microwave and ultrasound on the stability of (all-E)-astaxanthin, J. Agric. Food Chem., 54, 8346, 2006. 98. Priego-Capote, F., Ruiz-Jiménez, J., García-Olmo, J., and Luque de Castro, M.D., Fast method for the determination of total fat and trans fatty-acids content in bakery products based on microwave-assisted Soxhlet extraction and medium infrared spectroscopy detection, Anal. Chim. Acta, 517, 13, 2004. 99. Priego-Capote, F. and Luque de Castro, M.D., Focused microwave-assisted Soxhlet extraction: a convincing alternative for total fat isolation from bakery products, Talanta, 65, 81, 2005. 100. Ruiz-Jiménez, J., Priego-Capote, F., and Luque de Castro, M.D., FT-midIR determination of fatty acid proﬁles, including trans fatty acids, in bakery products after focused microwave-assisted Soxhlet extraction, Anal. Bioanal. Chem., 385, 1532, 2006. 101. Priego-Capote, F., Ruiz-Jiménez, J., and Luque de Castro, M.D., Identiﬁcation and quantiﬁcation of trans fatty acids in bakery products by gas chromatography-mass spectrometry after focused microwave Soxhlet extraction, Food Chem., 100, 859, 2007. 102. Akhtar, M.H., Croteau, L.G., Dani, C., and Elsooud, K.A., Development and validation of microwaveassisted extraction of fortiﬁed and incurred chloramphenicol residues in egg albumen and yolk, Spectroscopy, 13(1), 33, 1997. 103. Akhtar, M.H., Comparison of microwave assisted extraction with conventional (homogenization, vortexing) for the determination of incurred salinomycin in chicken eggs and tissues, J. Environ. Sci. Health B, B39(5–6), 835, 2004. 104. Pylypiw, H.M., Jr., Arsenault, T.L., Thetford, C.M., and Incorvia Mattina, M.J., Suitability of microwave-assisted extraction for multiresidue pesticide analysis of produce, J. Agric. Food Chem., 45, 3522, 1997. 105. Prados-Rosales, R.C., Luque Garcia, J.L., and Luque de Castro, M.D., Rapid analytical method for the determination of pesticide residues in sunﬂower seeds based on focused microwave-assisted Soxhlet extraction prior to gas chromatography-tandem mass spectrometry, J. Chromatogr. A, 993, 121, 2003. 106. Bouaid, A., Martín-Esteban, A., Fernández, P., and Cámara, C., Microwave-assisted extraction method for the determination of atrazine and other organophosphorus pesticides in oranges by gas chromatography (GC), Fresenius J. Anal. Chem., 367, 291, 2000. 107. Falqui-Cao, C., Wang, Z., Urruty, L., Pommier, J.-J., and Montury, M., Focused microwave assistance for extracting some pesticide residues from strawberries into water before their determination by SPME=HPLC=DAD, J. Agric. Food Chem., 49, 5092, 2001. 108. Sanusi, A., Guillet, V., and Montury, M., Advanced method using microwaves and solid-phase microextraction coupled with gas chromatography-mass spectrometry for the determination of pyrethroid residues in strawberries, J. Chromatogr. A, 1046, 35, 2004. 109. Chen, Y.-I., Su, Y.-S., and Jen, J.-F., Determination of dichlorvos by on-line microwave-assisted extraction coupled to headspace solid-phase microextraction and gas chromatography-electron-capture detection, J. Chromatogr. A, 976, 349, 2002. 110. Padrón-Sanz, C., Sosa-Ferrera, Z., and Santana-Rodríguez, J.J., An approach to the application of microwave-assisted micellar extraction and liquid chromatography with ultraviolet detection to the extraction and determination of organophosphorus pesticides in tomato samples, J. AOAC Int., 88(5), 1485, 2005. 111. El-Saeid, M.H., Kanu, I., Anyanwu, E.C., and Saleh, M.A., Impacts of extraction methods in the rapid determination of atrazine residues in foods using supercritical ﬂuid chromatography and enzyme-linked immunosorbent assay: microwave solvent vs. supercritical ﬂuid extractions, Sci. World J. (Electronic Resource), 5(14), 11, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 83

Microwave-Assisted Processes in Food Analysis

17.7.2008 4:02pm Compositor Name: VAmoudavally

83

112. Pallaroni, L. and von Holst, C., Comparison of alternative and conventional extraction techniques for the determination of zearalenone in corn, Anal. Bioanal. Chem., 376, 908, 2003. 113. Pallaroni, L., von Holst, C., Eskilsson, C.S., and Björklund, E., Microwave-assisted extraction of zearalenone from wheat and corn, Anal. Bioanal. Chem., 374, 161, 2002. 114. Diagne, R.G., Foster, G.D., and Khan, S.U., Comparison of Soxhlet and microwave-assisted extractions for the determination of fenitrothion residues in beans, J. Agric. Food Chem., 50, 3204, 2002. 115. Papadakis, E.N., Vryzas, Z., and Papadopoulou-Mourkidou, E., Rapid method for the determination of 16 organochlorine pesticides in sesame seeds by microwave-assisted extraction and analysis of extracts by gas chromatography-mass spectrometry, J. Chromatogr. A, 1127, 6, 2006. 116. Weichbrodt, M., Vetter, W., and Luckas, B., Microwave-assisted extraction and accelerated solvent extraction with ethyl acetate-cyclohexane before determination of organochlorines in ﬁsh tissue by gas chromatography with electron-capture detection, J. AOAC Int., 83(6), 1334, 2000. 117. Wittmann, G., Huybrechts, T., Van Langenhove, H., Dewulf, J., and Nollet, H., Trace analysis of trichlorobenzenes in ﬁsh by microwave-assisted extraction and gas chromatography-electron-capture detection, J. Chromatogr. A, 993, 71, 2003. 118. Bayen, S., Lee, H.K., and Obbard, J.P., Determination of polybrominated diphenyl ethers in marine biological tissues using microwave-assisted extraction, J. Chromatogr. A, 1035, 291, 2004. 119. Pena, T., Pensado, L., Casais, C., Mejuto, C., Phan-Tan-Luu, R., and Cela, R., Optimization of a microwave-assisted extraction method for the analysis of polycyclic aromatic hydrocarbons from ﬁsh samples, J. Chromatogr. A, 1121, 163, 2006. 120. García-Falcón, M.S., Simal-Gándara, J., and Carril-González-Barros, S.T., Analysis of benzo[a]pyrene in spiked fatty foods by second derivative synchronous spectroﬂuorimetry alter microwave-assisted treatment of samples, Food Addit. Contam., 17(12), 957, 2000. 121. Gfrerer, M. and Lankmayr, E., Microwave-assisted saponiﬁcation for the determination of 16 polycyclic aromatic hydrocarbons from pumpkin seed oils, J. Sep. Sci., 26, 1230, 2003. 122. Ahmed, F.E., Analyses of pesticides and their metabolites in foods and drinks, Trends Anal. Chem., 20(11), 2001. 123. Björklund, E., Holst, C., and Anklam, E., Fast extraction, clean-up and detection methods for the rapid analysis and screening of seven indicator PCDBs in food matrices, Trends Anal. Chem., 21(1), 39, 2002. 124. Rosenblum, L., Garris, S.T., and Morgan, J.N., Comparison of ﬁve extraction methods for determination of incurred and added pesticides in dietary composites, J. AOAC Int., 85(5), 1167, 2002.

Otles/Handbook of Food Analysis Instruments 45660_C004 Final Proof page 84

17.7.2008 4:02pm Compositor Name: VAmoudavally

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 85 14.7.2008 9:45am Compositor Name: BMani

Extraction 5 Ultrasound-Assisted in Food Analysis Farid Chemat, Valérie Tomao, and Matthieu Virot CONTENTS 5.1 5.2

Introduction ............................................................................................................................ 85 Basic Principles ..................................................................................................................... 86 5.2.1 Importance of the Extraction Step ............................................................................ 86 5.2.2 Ultrasound Cavitation ................................................................................................ 86 5.2.3 Instrumentation .......................................................................................................... 88 5.3 Ultrasound-Assisted Extraction: Important Parameters and Mechanism .............................. 89 5.3.1 Inﬂuence of Operating Conditions ............................................................................ 89 5.3.2 Inﬂuence of the Food Matrix .................................................................................... 89 5.4 Ultrasound-Assisted Extraction: Main Applications in Food Analysis ................................ 91 5.4.1 Flavors and Fragrances .............................................................................................. 91 5.4.2 Metals ........................................................................................................................ 91 5.4.3 Antioxidants .............................................................................................................. 93 5.4.4 Oil and Fat ................................................................................................................. 96 5.5 Comparison with Traditional and Recent Extraction Techniques ........................................ 96 5.5.1 Soxhlet ....................................................................................................................... 97 5.5.2 Supercritical Fluid Extraction .................................................................................... 98 5.5.3 Accelerated Solvent Extraction ................................................................................. 98 5.5.4 Microwave-Assisted Extraction ................................................................................. 98 5.6 Ultrasound-Assisted Extraction: Environmental Impact ....................................................... 99 5.7 Future Trends ......................................................................................................................... 99 References ....................................................................................................................................... 99

5.1 INTRODUCTION Food products are complex mixtures of vitamins, sugars, proteins and lipids, ﬁbers, aromas, pigments, antioxidants, and other organic and mineral compounds. Before such substances can be analyzed, they have to be extracted from the food matrix. Direct analyses are generally not possible to achieve due to the complexity of food samples and necessitate the introduction of samples under a liquid form to the analysis detector. Different methods can be used for this purpose, e.g., Soxhlet extraction, maceration, elution, steam distillation, cold pressing, and simultaneous distillation-extraction. Nevertheless, many food ingredients are well known to be thermally sensitive and vulnerable to chemical changes. Losses of some compounds, low extraction efﬁciency, time- and energy-consuming procedures (prolonged heating and stirring in boiling solvent, use of large volumes of solvents, etc.) may be encountered using these extraction methods. These shortcomings have led to the use of new sustainable ‘‘green’’ techniques in extraction, which typically involve less solvent and energy, such as ultrasound-assisted extraction (UAE) [1], supercritical ﬂuid extraction [2], headspace method [3], microwave extraction [4], controlled pressure drop process [5], accelerated solvent extraction [6], 85

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 86 14.7.2008 9:45am Compositor Name: BMani

86

Handbook of Food Analysis Instruments

and subcritical water extraction [7]. Extraction under extreme or nonclassical conditions is currently a dynamically developing area in applied research and industry. Alternatives to conventional extraction procedures may increase production efﬁciency and contribute to environmental preservation by reducing the use of solvents, fossil energy, and generation of hazardous substances. Ultrasound is a key technology in achieving the objective of sustainable green chemistry. Ultrasound is well known to have a signiﬁcant effect on the rate of various processes in the chemical and food industry. Much attention has been given to the application of ultrasound for the extraction of natural products that typically needed hours or days to reach completion with conventional methods. Using ultrasound, full extractions can now be completed in minutes with high reproducibility, reducing the consumption of solvent, simplifying manipulation and workup, giving higher purity of the ﬁnal product, eliminating post-treatment of wastewater, and consuming only a fraction of the fossil energy normally needed for a conventional extraction method such as Soxhlet extraction, maceration, or steam distillation. Several classes of food components such as aromas, pigments, antioxidants, and other organic and mineral compounds have been extracted and analyzed efﬁciently from a variety of matrices (mainly animal tissues, food, and plant materials). UAE is a research area that has an impact in several ﬁelds of modern chemistry. The main beneﬁts are decrease of extraction time, energy, and solvent used. The advantages of using ultrasound energy for extraction also include more effective mixing and micromixing, faster energy and mass transfer, reduced thermal and concentration gradients and extraction temperature, selective extraction, reduced equipment size, faster response to process extraction control, faster start-up, increased production, and elimination of process steps. Extraction processes performed under the action of ultrasound are believed to be affected in part by cavitation phenomena and mass transfer enhancement. This chapter presents a complete picture of current knowledge on UAE in food analysis. It provides the necessary theoretical background and some details about extraction by ultrasound, the technique, the mechanism, some applications, and environmental impacts.

5.2 BASIC PRINCIPLES 5.2.1 IMPORTANCE OF

THE

EXTRACTION STEP

In general, any analytical procedure for food components from vegetables, fruits, spices, or other complex food matrices comprises two steps: extraction (e.g., single-step solvent extraction, Soxhlet extraction, steam distillation, and simultaneous distillation–extraction) and analysis (e.g., gas chromatography, gravimetry, etc.). While the analysis step is complete after only 15 to 30 min, extraction takes at least several hours. It is frequently carried out by prolonged heating and stirring in boiling solvent. Thus, the principal limiting step of a food analysis operation is the extraction of the analyte from the matrix, which consists of transferring the desired compounds into solvent. The conventional solvent extraction procedure represents 70% of the total processing time (Figure 5.1). It is thus important to shorten this limiting step. The choice of the technique is the result of a compromise between efﬁciency and reproducibility of extraction, ease of procedure, together with considerations of cost, time, degree of automation, and safety.

5.2.2 ULTRASOUND CAVITATION The ultrasound frequency range can be divided basically into diagnostic and power ultrasound. Diagnostic ultrasound plays a very important role in modern measuring techniques. It involves highfrequency ultrasound in the range 2–10 MHz. A typical application is to measure the velocity and absorption coefﬁcient of the acoustic wave in a medium. It is an easy, fast, noninvasive, and nondestructive way of gaining structural and chemical information. Low-power ultrasound can be used to characterize acoustic properties of foodstuffs like, for instance, velocity of sound, attenuation, reﬂection, and scattering. In pure food compounds (oil, water, sugar, etc.), the attenuation and

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 87 14.7.2008 9:45am Compositor Name: BMani

87

Ultrasound-Assisted Extraction in Food Analysis Analysis 5

5 Data management 25

25

5

5

Time saved

70 Sample collection

50 15

Extraction Conventional extraction

Ultrasound extraction

FIGURE 5.1 Relative consuming time of different steps for a food analytical procedure.

velocity of sound can be measured relatively easily and the adiabatic compressibility can be calculated. With multiphase products (most food products) it is not as straightforward and a lot of computational data treatment is required for useful results. Ultrasonic spectroscopy is a technique where a very short ultrasonic pulse (broadband) is transmitted into a product. The ultrasonic spectra of the original pulse and its echoes are recorded and the change in frequency is a result of various physical properties like particle size, concentration, temperature, etc. Ultrasonic imaging is being used to scan fruits for bruises and diseases. For instance, certain diseases cause the inside of pears to turn brown and hollow. A laboratory scale apparatus has been developed with which fruits are knocked on with small hammer-like devices. The reverberation is recorded with polymeric transducers and analyzed. The system was capable of measuring the ripeness of mangos in a nondestructive way [8–11]. Power ultrasounds, having frequencies between 20 kHz and 100 MHz, are now well-known to have signiﬁcant effects on the rate of various physical and chemical processes (Figure 5.2). Cleaning and solubilization are the more developed applications and a large variety of ultrasound baths exist for chemical laboratory use. The effect of ultrasonic waves on solid samples is widely used for the extraction of aromas from plant materials or metal impurities from soils. Degassing and stripping are widely used for ﬂavor analysis and in environmental and polymer research. Other interesting ultrasound applications involve homogenization, emulsiﬁcation, sieving, ﬁltration, and crystallization. The most interesting effect of ultrasound-based operational units is the reduction of processing time and increase of product quality. All these effects are attributed to acoustic cavitation: When a liquid is irradiated by ultrasound, micro-bubbles form, grow, and oscillate extremely fast, and eventually collapse powerfully if the acoustic pressure is high enough. These collapses, occurring near a solid surface, generate micro-jets and shock waves that result in cleaning, erosion, and fragmentation of the surface. Power ultrasound involves the mechanical and chemical effects of cavitation. The mechanism can be explained by two competing theories. The hot spot theory

0

10 102 103 104 105 106 107 108

Human hearing

16 Hz–18 kHz

Power ultrasound

20 –100 kHz

Extented power range

100 kHz–2 MHz

Diagnostic ultrasound

5 –10 MHz

FIGURE 5.2 Frequency ranges of sound.

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 88 14.7.2008 9:45am Compositor Name: BMani

88

Handbook of Food Analysis Instruments

assumes that high pressures and temperatures generated in the bubbles during the last nearly adiabatic compression, just before collapse, are responsible for the breakage of molecular bonds and formation of radicals. On the other hand, the electrical theory involves micro-discharges due to high electrical ﬁelds generated by deformation and fragmentation of the bubbles [12–14].

5.2.3 INSTRUMENTATION The two most common ultrasound equipments that are used for extraction are the ultrasonic cleaning bath and the more powerful probe system. For small extraction volumes, an ultrasound horn with the tip submerged in the ﬂuid can be sufﬁcient. Large volumes of ﬂuids have to be sonicated in an ultrasound bath or in continuous or recycled-ﬂow sonoreactors (Figure 5.3). Recently, the new methodology of continuous-ﬂow systems has been used in analytical chemistry. Most UAE applications have been developed in discrete systems using a bath or an ultrasonic probe, particularly in extraction of food samples. Less frequent has been the design of online UAE systems in the same ﬁeld [15]. However, it is noted that the last approach is considerably faster. It consists in an open system, in which fresh solvent ﬂows continuously through the sample. This induces the displacement of mass transfer equilibria toward the solubilization of analytes into the liquid phase. The coupling of the extraction step to the analytical steps, which would overcome the dilution effect, has not been performed yet despite its ease of implementation. The extract is then driven to the continuous manifold for online achievement of the analytical process, which involves preconcentration, derivatization, ﬁltration, and detection (using ﬂame atomic-absorption spectrometer [FAAS], gas chromatography-mass spectrometry (GC-MS), or other techniques). The main advantages of using online UAE are the reduction of sample contamination as well as analyte losses (because less sample manipulation is needed), a reduction of reagent consumption and concentration, compared with off-line (batch or discontinuous reactor) UAE. In addition, the online method removes the centrifugation or ﬁltration step required to separate the liquid phase from the sample particles and thus signiﬁcantly reduces the duration of sample preparation.

US bath

US probe Sampling Cooler H2O

Tank Seeds + Solvent

Transducers

Transducers

FIGURE 5.3

Some current concepts of UAE.

H2O

US continuous

Pump Tank

Ultrasound transducer

US probe

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 89 14.7.2008 9:45am Compositor Name: BMani

Ultrasound-Assisted Extraction in Food Analysis

89

While most of the research effort in UAE has concentrated on ultrasound itself, some studies have also examined the coupling between ultrasound and other techniques. For instance, UAE is being employed in combination with microwave energy [16], supercritical ﬂuid extraction [17], or simply with conventional methods such as Soxhlet extraction [18]. When combined with supercritical ﬂuid extraction, UAE enhances the mass transfer of the species of interest from the solid phase to the solvent used for extraction. Soxhlet extraction can also be improved by ultrasound when applied at the cartridge zone before siphoning, thus permitting the removal of lipid fractions from very compact matrices. The efﬁciency of combining microwave and ultrasound has been clearly shown in applications such as extraction of copper and the Kjeldahl method for determination of total nitrogen in food [19].

5.3 ULTRASOUND-ASSISTED EXTRACTION: IMPORTANT PARAMETERS AND MECHANISM 5.3.1 INFLUENCE

OF

OPERATING CONDITIONS

Proper selection of the solvent is the key to successful UAE. Solvent choice is dictated by the solubility of the analytes of interest, the interactions between the solvent and matrix, and the intensity of ultrasound cavitation phenomena in the solvent. Important physical parameters related to UAE are presented in this section. Ultrasound power, temperature, and extraction time affect not only the extraction yield but also the composition of the extract. According to Palma and Barroso [20], a higher temperature for UAE means a higher efﬁciency in the extraction process due to the increase of the number of cavitation bubbles and a larger solid-solvent contact area. However, this effect is decreased when the temperature is near the solvent’s boiling point. It is also important to prevent the degradation of thermolabile compounds. For instance, polyphenols and isocyanates are conventionally extracted at 48C and increasing the extraction temperature will automatically decrease the quantity and quality of the extract. Wu et al. [21] suggest that the optimal duration for the UAE of ginseng saponins from ginseng root is about 2 h. Short ultrasound treatment (less then 30 min) was found to improve the extraction process [22]. The ultrasound power is one of the parameters to optimize to reach a compromise between extraction time and solvent volume. Li et al. [23] pointed that the relative yield of soybean oil at a power density of 47 W=cm2 was approximately ﬁve times higher than at 16 W=cm2. Generally, the highest efﬁciency of UAE, in terms of yield and composition of the extracts, can be achieved by increasing the ultrasound power, reducing the moisture of food matrices to enhance solvent-solid contact, and optimizing the temperature to allow a shorter extraction time.

5.3.2 INFLUENCE

OF THE

FOOD MATRIX

Food tissues consist of cells surrounded by walls. Some cells exist in the form of glands (external or internal) that are ﬁlled with the target products (generally secondary metabolites). A characteristic of such glands (when external) is that their skin is very thin and can be very easily destroyed. For internal glands, it is the degree of milling of the plant material that plays an important role. Conventional solvent extraction may be thought of as a transfer of solutes from one phase (e.g., a solid phase) into another (the solvent). The food matrix can be compared to a grain constituted of an impermeable core covered by a solvent boundary layer (Figure 5.4). Secondary metabolites are extracted in three steps: desorption from the matrix surface or release from internal glands, diffusion through the boundary layer to the boiling solvent, and solubilization in the solvent. The extraction recovery can be limited by one or several steps. The phenomena at play in UAE could be visualized by referring to our original investigations where we designed a series of solid–liquid extraction steps using various extraction procedures on a single source carvi, a food spice material [24]. The effects of such extraction processes on the

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 90 14.7.2008 9:45am Compositor Name: BMani

90

Handbook of Food Analysis Instruments 1: Desorption 2: Diffusion 3: Solubilization

1 2

FIGURE 5.4

3

Schematic representation of the individual steps in the extraction process.

physical microstructure of the material being extracted were closely monitored using scanning electron microscopy. The various extraction methods (Soxhlet, maceration, and ultrasound) produced distinguishable physical changes on the extracted matrix (caraway seeds). Figure 5.5a is a micrograph of the untreated seeds, broken cryogenically, which can be compared with structures of the treated seeds in Figure 5.5b and c. After a few hours of conventional extraction or maceration in hexane (698C), the cell walls seemed thicker but intact and most of the cells were totally free from any component released out of the cell. After 30 min of ultrasound extraction (208C), cells and cell walls were affected to different degrees. We observed a huge perforation of the particles’ external surface and some waste material is dispersed, showing that all the cell walls were ﬁnally broken and converted into undeﬁned cell shapes. There was clear evidence of explosions occurring at the cell level as a consequence of the sudden enhancement in micromixing, generated in that case by localized mass transfer caused by ultrasound power. Ultrasound has focused its power, at the beginning of extraction, on cuticular layer destruction and oil exudation. Then, it deﬂected this power against cell walls perforation mainly due to the high resistance of the particles in the medium toward ultrasound energy. When the glands were subjected to more severe thermal stresses and localized high pressures induced by cavitation, as in the case of UAE, the pressure buildup within the glands could have exceeded their capacity for expansion, and caused their rupture more rapidly than in control experiment. In general, the SEM observations pointed two distinct extraction mechanisms for conventional and ultrasound procedures, respectively. The ﬁrst involves diffusion of the plant extract components across the unbroken gland wall due to the temperature increase in the medium, and the other one, exudation of oil from damaged cell walls and even cells, due to a strong ultrasonic mechanical effect, which generally triggers an instantaneous release of the plant extract components into the surrounding solvent.

Without treatment

FIGURE 5.5

Conventional extraction

Ultrasound extraction

Electron micrograph of carvi seeds (untreated, conventional extraction, and UAE).

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 91 14.7.2008 9:45am Compositor Name: BMani

Ultrasound-Assisted Extraction in Food Analysis

91

5.4 ULTRASOUND-ASSISTED EXTRACTION: MAIN APPLICATIONS IN FOOD ANALYSIS Among newer techniques used in extraction technology, UAE of food components has been employed as a new tool to improve the yield and quality of extraction products and to reduce the duration of analytical procedures. The ﬁrst applications were related to the determination of metals in foods. Since then, numerous other compounds have been efﬁciently extracted such as aromas, antioxidants, oils, pigments, etc.

5.4.1 FLAVORS AND FRAGRANCES Natural ﬂavors and fragrances have been used probably since the discovery of ﬁre. Egyptians, Phoenicians, Jews, Arabs, Indians, Chinese, Greeks, Romans, and even Mayas and Aztecs all possessed a fragrance culture of great reﬁnement. Fragrances are complex mixtures of volatile substances generally present in low concentrations. Flavors or aromas are obtained from a variety of aromatic plant materials including ﬂowers, buds, seeds, leaves, twigs, bark, herbs, wood, fruits, and roots. Their yield and quality depend mainly on the cultivar (chemotype and genetic variability), environment (fertilization, climatic conditions, and crop protection), and physiological stage (plant development stage). These aromatic compounds are produced by plants as by-products or indeed as ﬁnal metabolites and stored in certain organs of the plant: . . . .

Thyme, sage, and rosemary (Lamiaceae family): in glandular cells, hairs, and scales. Cinnamon, laurel, and cassia (Lauraceae family): in essential oil and resin cells. Caraway, anis, and coriander (Umbellifers family): in essential oil channels occurring in the intercellular space of plant tissue. Lemon, orange, and bergamot (Rutaceae family): in lysigenous secretory reservoirs formed inside the plant.

Conventional ﬂavor and fragrance extraction techniques have important drawbacks, such as low yields and formation of by-products because of the low stability of the target compounds. For steam distillation and hydrodistillation methods, the steam is percolated through the ﬂask containing the aromatic plants and the aromas evaporate. The elevated temperatures and prolonged extraction time can cause chemical modiﬁcations of the aromatic components and often a loss of the most volatile molecules. To obtain high-quality extracts from aromatic plant materials, several innovative methods are available and a large number of studies have dealt with the beneﬁts of ultrasonic power. Typical analytes, extraction conditions, and detection devices are summarized in Table 5.1. Prior to extraction, there are many parameters to optimize: plant moisture, particle size, ultrasonic design, temperature, and solvent [25]. Vanillin has been extracted and quantiﬁed by sonoelectroanalysis with ethylacetoacetate in two different samples [26]. Vanillin concentrations were in close agreement with HPLC-UV quantiﬁcation. Vanillin was also extracted from Vanilla planifolia in dry ethanol in 0.99% yield in comparison with a maximal yield of vanilla ‘‘crude’’ extract (14.3%) obtained in EtOH=H2O (40=60). UAE is a useful tool for the rapid quantiﬁcation of several aromatic compounds in wine [25,27,28], honey, citrus ﬂowers [29,30], and carvone [24]. In addition, US assistance enabled the precise quantiﬁcation of 37 volatile compounds in brandies and 16 in alcoholic oak extracts [31]. These good results can be explained by the mild conditions and speciﬁc mechanical effects of sound waves, according to Chemat et al. [24] and Shotipruk et al. [32].

5.4.2 METALS Metals occurring in food are important in the ﬁelds of nutrition, toxicity, and control of the manufacturing process. Several metal ions have a nutritional value since they are involved in

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 92 14.7.2008 9:45am Compositor Name: BMani

92

Handbook of Food Analysis Instruments

TABLE 5.1 Flavor and Fragrance Extracts Matrix

Analyte

Aged brandies and red wine

Aroma compounds

Caraway seeds

Carvone and limonene

Citrus ﬂowers

Volatile compounds

Vanilla

Vanillin

Greek saffron

Safranal

Honey

Aroma compounds

Must and wine

Aroma compounds

Peppermint leaves

Menthol

White wine

Aroma compounds

Wine

Volatile compounds

Extraction Conditions and Remarks USB, 208C, 3-step extraction in CH2Cl2. ST: 310 min. Mean values (mg=L) for monoterpenoids: 291 (linalool), 248 (a-terpineol), 397 (citronellol) USH, 20 kHz, 150 W, 208C, n-hexane. ST: 60 min. Yd (mg=g): 17 (carvone) 16 (limonene). USAE gave a better quality of extracts with an increased yield for carvone USB, 258C, n-pentane: Et2O. ST: 10 min. Among extracted compounds, linalool was the major (% of the total peak area): 51.6 (orange), 11.3 (lemon), 75.2 (tangerine), 80.6 (sour orange) USH, 20 kHz, 750 W, 258C, EtOH or EtOH=H2O. ST: 1–2 min USB, 35 kHz, 258C, H2O: Et2O. ST: 510 min. Safranal ranged between 40.7 and 647.7 mg=100 g saffron USB, 258C, H2O, n-pentane: Et2O, ST: 210 min. UAE allows isolation of several compounds and is the best extraction technique USB, 48 kHz, 208C, 3-step extraction with CH2Cl2. ST: 310 min. Results were higher than those obtained by traditional method and several compounds were extracted USB, 40 kHz, 228C, H2O, ST: 60 min, Yd: 17.8 mg=g (2% of total product). The amount of menthol released can be enhanced (12%) with temperature increasement (398C) USB, n-pentane: Et2O, MgSO4, ST: 30 min. The method described enables the rapid quantiﬁcation of 24 wine compounds USB, 40 kHz, 258C, CH2Cl2, ST: 15 min. 12 compound concentrations are done and extend from 0.422 to 168 mg=L. Linalool and a-terpineol were not detected

Detection (Connection)

References

GC-FID (off-line) GC-MS (off-line)

[31,33]

GC-FID (off-line) GC-MS (off-line) SEM (off-line)

[24]

GC-MS (off-line)

[30]

HPLC-DAD (off-line) Sonoelectro-analysis (online) GC-FID (off-line) GC-MS (off-line)

[26,34]

GC-MS (off-line)

[29,30]

GC-FID (off-line)

[27]

GC-FID (off-line) SEM (off-line)

[32]

GC-FID (off-line)

[28]

GC-FID (off-line) GC-MS (off-line)

[25]

[35]

USB, ultrasonic bath; USH, ultrasonic horn; ST, sonication time; Yd, yield.

various biological mechanisms such as enzyme functioning. However, in elevated concentrations, metal ions may have adverse and toxic effects. The qualitative and quantitative knowledge of metals such as Sn or Cu is of great importance because they are possibly involved in cancer and cardiovascular diseases. An actual trend of great health interest is Al speciation because this element

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 93 14.7.2008 9:45am Compositor Name: BMani

Ultrasound-Assisted Extraction in Food Analysis

93

might favor the development of Alzheimer’s disease. Another important point is the determination of metal traces in seafood samples which can be used as biomarkers to monitor the environmental pollution. Extracting metals from food samples is complicated by the strong interactions typically occurring between the food matrix and analytes. Several methods are available to transfer the desired analytes into the liquid phase, the most common being leaching. Other methods such as digestion, calcination, ashing, etc. can also be used when leaching is not sufﬁcient. These methods infer some drawbacks and drastic conditions like high temperatures or pressures, use of concentrated acids, sample losses by manipulation, or volatilization. In addition, these methods are generally hazardous and time consuming. Table 5.2 shows a sample of US-assisted metal extractions carried out on several matrices. The acid-leaching procedure assisted by ultrasound appears rapid, accurate, and effective [36–38]. It offers rapid sample preparation and mild conditions compared with the tedious and time-consuming acid digestion. Filgueiras et al. [39] described a fast UAE (only 7 min) for Mg, Mn, and Zn. According to Krishna and Arunachalam [40], the procedure required only 15 min to estimate trace elements in mussel and lichen and allows the preparation of about 35 samples by working day. Another alternative is to combine the beneﬁts of UAE with a ﬂow injection (FI) manifold coupled with an FAAS. FI permits a preconcentration step for a higher sensitivity. Thus, the handling and analytical steps are shortened. The risk of contamination is reduced and the centrifugation step required in the off-line technique is totally removed. For faster procedures, the sampling frequency is another key parameter to optimize. According to Del Carmen Yebra et al. [41], continuous UAE coupled with an FI-FAAS allows a total sample frequency of 46 and 18 samples per hour for copper and iron determination in seafood samples, respectively. Calcium determination in seafood samples can also be carried out at a frequency of 40 samples per hour with the same technique [42]. Another focus in the current works is the method described by Cava-Montesinos et al. [43], where an online FI method coupled with Cold vapour atomic absorption spectrometry (CV-AFS) was used in order to determine mercury contents in ﬁsh samples. The method presented by Šuchman and Bednár [44] allows direct chloride analysis in meat products on the average extraction time of 7 min. The US leaching technique avoids the use of strong acids, which is a great advantage for US probes and analytical instruments. It also presents the advantage of requiring only little solvent and sample (as low as a few micrograms) while offering a rapid, precise, and reproducible metal extraction procedure in different food samples.

5.4.3 ANTIOXIDANTS Plant and food antioxidants are able to rapidly scavenge free radicals, thereby inhibiting deleterious oxidative processes such as lipid peroxidation that are responsible for food deterioration, accumulation of toxic products, and off-ﬂavor compounds. Thus, the knowledge of their properties and concentrations in food is most desirable. Traditional extraction methods such as maceration, mix-stirring, or reﬂuxing require large volumes of solvent and are often time consuming. In addition, they often require drastic conditions (high temperature or pressure) that are not fully compatible with the general chemical instability of potent antioxidants. Hence, special care (in terms of light exposure, temperature, pH, etc.) is needed during handling in order to prevent antioxidant-rich extracts from oxidation. For instance, some authors have performed direct HPLC analysis of wine or cider for a simple and rapid determination of polyphenols [57,58]. However, direct analysis is not always possible and a preliminary preparation step is often necessary. The feasibility of UAE in the analysis of polyphenols and other antioxidants has been investigated on many matrices (Table 5.3). In most cases, the yields are increased with US assistance. According to Tsanova-Savova et al. [59], 5 min of sonication is equivalent to 1 h of mechanical stirring for

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 94 14.7.2008 9:45am Compositor Name: BMani

94

Handbook of Food Analysis Instruments

TABLE 5.2 Metal Extraction Matrix

Analyte

Fish and mussel

Cd, Cu, Zn

Fish samples

Hg

Fruits and vegetables

Cd

Juices and soft drinks

Al

Lettuce and cabbage

Ca, Mg, Mn, Zn

Meat

Fe, Zn

Meat products

Chlorides

Mussel

Cd, Pb

Mussel

Cd, Pb, Cu, Fe

Mussel tissue samples Plant samples

Hg

Raw pork meat

Rice

Pb, Ca, Cu, Cd, Cr, Fe, Zn, Mg As

Seafood

Ca, Cu, Fe

Seafood

Se

Mg, Mn, Zn

Extraction Conditions and Remarks USB, 568C, HNO3: HCl: H2O2. ST: 30 min. Yd (mg=kg), for ﬁsh and mussel, respectively: 0.06, 0.55 (Cd), 1.16, 4.24 (Cu) and 15.45, 52.30 (Zn) USB, 50 Hz, 50 W, 508C, HCl: H2SO4: HNO3: H2O2. Total Hg concentrations ranged between 0.74 (anchovy) and 6.1 mg=kg (mussel) USB, 208C, HNO3, ST: 1–2.5 min, ﬂow rate: 3.5 mL=min. Concentrations of Cd found in fruits and vegetables samples ranged between 0.118 (banana) and 0.640 mg=g (lettuce) USB, 35 kHz, 808C, HNO3: H2SO4: H2O2, ST: 20 min, Yd (mg=mL) extended from 2.15 to 12.0 for 18 different juice and soft drink samples USB, 47 kHz, 258C, H2O, HNO3, detergent, ST: 10 min, Yd for lettuce and cabbage, respectively: 1.68, 0.89% (Ca), 0.280, 0.180% (Mg), 165.69, 32.35 mg=g (Mn), 112.26, 26.36 mg=g (Zn) USB, 40 kHz, 208C, HNO3 and=or HCl, ST: 0.5–5 min, ﬂow rate: 3.5 mL=min. Contents (mg=g) ranged from 58.8 (pig muscle) to 277.8 (rabbit liver) for Fe, and from 58.8 (chicken muscle) to 195.7 (lamb muscle) for Zn USB, 500 W, 608C, H2O, ST: 5 min. Yd (g NaCl=kg) extend from 16.5 (salami) to 41.2 (herkules salami) for different thermal meat products USH, HNO3. ST: 15 s at 10% amplitude for Cd and 180 s at 60% for Pb. Contents ranged from 0.60 and 0.79 mg=g for Cd, and from 2.03 to 2.81 for Pb for ﬁve samples USB, 40 kHz, 208C, HNO3 for Cd and Pb, or HNO3=HCl for Cu and Fe. ST: 2–5 min, ﬂow rate: 3.5 mL=min. Yd (mg=g): 0.383–0.559 (Cd), 0.49–1.0 (Pb), 1.2–3.6 (Cu), 212.5–257.1 (Fe) USH, HCl. ST: 3–5 min, amplitude: 20%–70%. Content of methylmercury and inorganic mercury ranged from 0.053 to 0.243 mg=g for four samples USH, 20 kHz, 100 W, <508C–608C, HCl, 30% amplitude, ST: 3 min. Concentration (mg=g) ranges found in various samples: 20.1–46.5 for Zn, 28.4–731 for Mn and 1439–2989 for Mg USH, 20 kHz, 400 W, ambient temperature, HNO3, duty cycle 0.1 s, amplitude 70%, ST: 10 min, Yd (mg=kg) for 1 sample: 0.17 (Pb), 0.022 (Cd), 1.8 (Cu), 9.1 (Cr), 166 (Ca), 219 (Mg), 16.6 (Fe), 19.4 (Zn) USH, 258C, H2O þ enzymes (10 mg of a-amylase than 30 mg of protease). ST: 1–2 min. Yield of total as extracted: 149 ng=g for Spanish white rice and 56 ng=g for Indian basmati rice USB, 40 kHz, 208C, HNO3, ST: 0.5–3 min, ﬂow rate: 6 mL=min. Values (mg=g) ranged from 5.9 (hake) to 18.8 (mussel) for Cu, 37.2 (crab) to 256.2 (mussel) for Fe, 1576.73 (tuna) to 6849.35 mg=kg (prawn) for Ca USH, HNO3, ST: 3 min at 50% amplitude. Contents ranged from 0.95 (meagrin) to 2.61 mg=g (edible crab)

Detection (Connection) GFAAS (offline) FAAS (off-line) CV-AFS (online)

References [36]

[43]

FI-FAAS (online)

[45,46]

ETAAS (off-line)

[47]

FAAS (off-line)

[48]

FI-FAAS (online)

[49,50]

Potentiometry (online)

[44]

ETAAS (off-line)

[38]

FI-FAAS (online)

[51–53]

FI-CV-AAS (off-line)

[37]

FAAS (off-line)

[39]

GFAAS (offline) FAAS (off-line)

[54]

ICP-MS (offline) HPLCICP-MS (off-line) FI-FAAS (online)

[55]

ETAAS (off-line)

[56]

[41,42]

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 95 14.7.2008 9:45am Compositor Name: BMani

95

Ultrasound-Assisted Extraction in Food Analysis

TABLE 5.3 Antioxidant Extracts Matrix

Analyte

Amaranthus caudatus seed

Tocopherols, vitamin E isomers

Bulgarian fruits

(þ)-Catechin, ()-epicatechin

Eucommia ulmodies Oliv. (E. ulmodies)

Chlorogenic acid

Olive leaves

Oleuropein

Extra virgin olive oil

Polyphenols

Fruit juices

Ascorbic acid (vitamin C)

Tea leaves, grape seeds, and soybeans

Catechins, epicatechins, and isoﬂavones

Pistachio hulls, coconut shells

Phenolic compounds

Rosemary leaves

Carnosic acid

Strawberries

Phenolic compounds

Yellow sweet clover

Coumarins

Pastinaca sativa fruits

Furanocoumarins

Extraction Conditions and Remarks USB, 258C, 2-step extraction: ST: 30 or 60 min in MeOH than 30 min in hexane. Total tocopherols content varies from 51.81 to 63.7 mg=kg USB, MeOH=H2O, ST: 5 min, 15 fruits were studied and total catechins ranged between 4.3 and 195.3 mg=kg. No catechins were detected in melon USB, 50 kHz, 160 W, MeOH. ST: 330 min. Concentration ranged from 0.07% (fresh bark) to 0.71% (fresh leaves) with a good recovery USH, 20 kHz, 450 W, 408C, EtOH=H2O, duty cycle: 70%, amplitude: 30%, ﬂow rate: 5 mL=min, ST: 25 min. The main found compounds ranged from 488 (verbacoside) to 22610 mg=kg (oleuropein) USH, 20 kHz, 400 W, n-hexane, amplitude: 10%, duty cycle: 30%, ﬂow rate: 2.4 mL=min. Total Yd of polyphenols in various oil samples ranged between 124 and 1267 mg CAE=mL USH, 20 kHz, 25 W=cm, 258C, Phosphate buffer solution, ST: <2 min, Yield: 31 mg=100 mL USH, 200 W, 24 kHz, 608C, MeOH or EtOH, ST: 10–20 min. Content ranges: 0.22–0.46 mg=g (catechin), 0.07–3.36 mg=g (epicatechin) (Ref. [15]) and 72.58–550.45 mg=g (isoﬂavones) (Ref. [13]) USB, H2O, ST: 45 min, Yd: 34.2 mg TAE=gdw for pistachio hulls USB, 150 W, 25 kHz, 308C, EtOH=H2O. ST: 60 min, Yd: 406.93 mg TAE=L for coconut shells USB, Acetone. ST: 35 min, Yd: 26.2 mg=g USB, 40 kHz, 478C–538C, EtOH. ST: 45 min, Yd is about 18 mg=g USH, 20 kHz, 100 W, HCl, ST: 2 min, duty cycle and radiation amplitude: 20%. Contents ranged from 0.12 (syringic acid) to 566 mg=kg (gallic acid) for strawberries USB, EtOH. ST: 60 min, Yd (mg=g): 3.569 (coumarin), 1.269 (o-coumaric acid), 8.092 (melilotic acid) USB, 608C, ST: 330 min in petroleum ether than 330 min in MeOH. Total content ranged from 0.264 (phellopterin) to 14.444 mg=g (imperatorin)

Detection (Connection)

References

HPLC-UV (off-line)

[61,62]

HPLC-FLD (off-line)

[59]

HPLC-UV (off-line)

[63]

GC-MS (offline) HPLCDAD (online)

[64]

DAD-UV-vis spectrometer (online)

[65]

UV-vis spectrometer (off-line) HPLC-DADFLD, HPLCDAD or -MS (off-line)

[66]

UV-vis spectrometer (off-line)

[69,70]

MS (off-line) HPLC-UV or DAD (off-line) HPLC-DAD (off-line)

[60,71]

[67,68]

[72, 73]

HPLC-DAD (off-line)

[74]

HPLC-UV (off-line)

[75]

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 96 14.7.2008 9:45am Compositor Name: BMani

96

Handbook of Food Analysis Instruments

TABLE 5.4 Oil Extraction Matrix

Analyte

Adlay seed

Oil

Almond

Oil

Almond, apricot, and rice bran Bakery products

Oil

Oleaginous seeds

Fat

Soybean

Oil

Fat

Extraction Conditions and Remarks USH, 20 kHz, 110 W, 408C, sCO2, P: 20 MPa, ﬂow rate: 3.0 L=h. ST: 210 min. Extraction yields were 96.36% for oil USH, 20 kHz, 50 W, 558C, sCO2, P: 280 Bar, ﬂow rate: 20 kg=h. ST: 8.5 h. Yield is enhanced by about 20% and kinetics by about 30% USB, 42 kHz, 80 W, H2O, ammonium sulphate, t-butanol. ST: 6–8 min. Recovery (%): 87–89 (almond), 72–80 (apricot), 76–88 (rice bran) USH, 20 kHz, 400 W, n-hexane. ST: 6 min, duty cycle: 0.8 s, amplitude: 100%, ﬂow rate: 2 mL=min. Found values ranged from 4.81 (Egg cakes) to 35.22% (Snack corn) USH, 20 kHz, 100 W, 758C, n-hexane, duty cycle: 0.5 s, amplitude: 40%. Yd: 99% of fat recoveries in 90 min on sunﬂower seeds USH, 20 kHz, 20–50 W=cm, 258C, hexane: isopropanol. ST: 3 h. Oil yield was 12.21%

Detection (Connection)

References

GC-MS (off-line)

[78,79]

Gravimetric determination (off-line) Gravimetric determination (off-line) FTIR, gravimetric determination (off-line)

[80]

GC-FID (off-line)

[76]

GC-FID, SEM, gravimetric (off-line)

[77,84]

[81]

[82,83]

the extraction of catechins (ﬂavonols) from Bulgarian fruits. By using an ultrasonic horn, 15 min are also sufﬁcient to extract carnosic acid from rosemary leaves compared with 3 h of shaking in a water bath [60].

5.4.4 OIL AND FAT Many papers have reported on the UAE of oil and fat from various food samples (Table 5.4). According to Luque-García and Luque de Castro [76], extraction from oleaginous seeds is difﬁcult. Indeed, only 75%–85% of the oil is solubilized in the solvent. The rest of the oil content is strongly bound to the matrix and cannot be extracted without additional treatments. For instance, Li et al. [77] described the UAE of soybean oil in optimized yield and reduced operating time in comparison with conventional maceration. Another growing trend in the development of a new extraction process is to combine traditional and novel techniques. Luo et al. and Hu et al. [78,79] performed a US-assisted supercritical ﬂuid extraction (SFE) device for adlay seed extraction. The procedure resulted in decreased temperature and operating time and a higher extraction yield. For instance, US-assisted SFE required a temperature of 408C, a pressure of 20 MPa, and a ﬂow rate of 3 L=h during 3.5 h instead of 458C, 25 MPa, 3.5 L=h during 4 h for conventional SFE. Finally, Luque-García et al. [76] have reported a device consisting of a U.S.-assisted Soxhlet apparatus that is an attractive alternative to traditional Soxhlet extraction. Using this device, only 90 min are needed to obtain the same yield (99% of fat recovery) as in 12 h of traditional Soxhlet extraction.

5.5 COMPARISON WITH TRADITIONAL AND RECENT EXTRACTION TECHNIQUES Advantages and drawbacks of UAE have been compared to traditional and recent extraction techniques as conventional Soxhlet extraction, SFE, accelerated solvent extraction (ASE), and microwave-assisted extraction (MAE) (Table 5.5).

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 97 14.7.2008 9:45am Compositor Name: BMani

97

Ultrasound-Assisted Extraction in Food Analysis

TABLE 5.5 Advantages and Drawbacks of Traditional and Recent Extraction Techniques Supercritical Fluid Extraction

Accelerated Solvent Extraction

Sample is contained in an extraction cartridge and percolated with recondensed vapors of the solvent 3–48 h

Sample is placed in a high pressure vessel and crossed continuously by the supercritical ﬂuid

Sample is heated by a conventional oven and crossed by the extraction solvent under pressure

Sample is immersed in solvent and submitted to microwave energy

Sample is immersed in solvent and submitted to ultrasound using a US probe or US bath

10–60 min

10–20 min

3–30 min

10–60 min

1–30 g 150–500 mL

1–5 g 2–5 mL (solid trap) 30–60 mL (liquid trap) High Fast extraction, low solvent consumption, concentration of the extract, no ﬁltration necessary, possible high selectivity Many parameters to optimize

1–30 g 15–60 mL

1–10 g 10–40 mL

1–30 g 50–200 mL

High Fast extraction, no ﬁltration necessary, low solvent consumption

Moderate Fast extraction, easy to handle, moderate solvent consumption

Low Easy to use

Possible degradation of thermolabile analytes

Extraction solvent must absorb microwave energy, ﬁltration step required

Large solvent volume, ﬁltration step required

Name Brief description

Extraction time Sample size Solvent use

Soxhlet

Investment Advantages

Low Easy to handle, no ﬁltration necessary, high matrix capacity

Drawbacks

Long extraction time, large solvent volume

MicrowaveAssisted Extraction

UltrasoundAssisted Extraction

5.5.1 SOXHLET Extraction of solid material is traditionally performed by Soxhlet extraction. This method proceeds by iterative percolation of the sample with recondensed vapors of solvent. It has been one of the most used solid-liquid extraction techniques for a long time and is currently the principal reference method. Soxhlet extraction of solid materials has undeniable advantages such as uninterrupted extraction with repeated percolation with fresh solvent, non-necessary ﬁltration step, and possible recycling of solvent. This technique nevertheless displays some disadvantages: poor extraction of polar lipids, long operating time, large solvent volumes, operation at the solvent’s boiling point, and inadequacy for thermolabile analytes [85]. Using a UAE versus conventional Soxhlet extraction presents the advantage that the whole extraction process is accelerated. In fact, the extraction efﬁciency can be equivalent [86] or higher [87] than that obtained with conventional Soxhlet extraction and, in some cases, allows the extraction of thermolabile compounds that are typically degraded during the conventional procedure [85]. However, some drawbacks must be pointed out: the renewal of the solvent is not possible during the process and a ﬁltering step is necessary.

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 98 14.7.2008 9:45am Compositor Name: BMani

98

Handbook of Food Analysis Instruments

5.5.2 SUPERCRITICAL FLUID EXTRACTION Supercritical ﬂuid extraction (SFE) is an important method for the extraction of solid materials using a supercritical ﬂuid [88,89]. Carbon dioxide is the most widely used solvent in SFE because it is nontoxic, nonﬂammable, cheap, easily eliminated after extraction, and endowed with a high solvating capacity for nonpolar molecules. Other possible solvents are Freon, ammonia, and some organic solvents [90]. In a typical SFE procedure, the supercritical ﬂuid continuously enters the solid matrix where it dissolves the material of interest. The extraction can be achieved with a remarkably high selectivity by adjusting the solvating capacity of the supercritical ﬂuid by changing the pressure and temperature. Major advantages of SFE include preconcentration effects, cleanness and safety, quantitativeness, expeditiousness, and simplicity [91]. The application of SFE from plant, animal, and oil has been reviewed [92]. Compared to UAE, SFE processes are more precise in many cases [71]. The use of CO2 as supercritical ﬂuid extractor limits the polluting hazards. The drawback of SFE versus UAE is the need of more expensive equipment with the difﬁculty of extracting polar molecules without adding modiﬁers to CO2. Indeed, UAE permits the extraction of a wide variety of compounds using polar or nonpolar solvents and much simpler equipment.

5.5.3 ACCELERATED SOLVENT EXTRACTION Accelerated solvent extraction (ASE) makes use of the same solvents as traditional extraction methods while operating at elevated temperatures and pressures. ASE, generalized to pressurized ﬂuid extraction, is now well accepted as an alternative to Soxhlet extraction [93]. The sample is contained in an extraction cartridge heated in a conventional oven and crossed by the extraction solvent. The extraction is performed statically for a short period. When the extraction is complete, compressed gas shifts the solvent from the cartridge to the collecting vessel. ASE allows rapid extraction with small solvent volumes by using high temperatures (up to 2008C) for increased solvent diffusivity [94,95] and high pressure (up to 20 MPa) to keep the solvent in its liquid state [96]. ASE provides a wide range of applications but the high extraction temperature may lead to degradation of thermolabile compounds [94]. Fisher et al. [97] pointed out that the main drawbacks of ASE are a strong background interference and high detection limit. Moreover, the equipment is expensive. Compared to ASE, UAE allows an important reduction of extraction time, solvent volume, and sample manipulations [98].

5.5.4 MICROWAVE-ASSISTED EXTRACTION Microwave-assisted extraction (MAE) has attracted growing interest as it allows the efﬁcient use of microwave energy to extract valuable compounds from solid samples. There are two types of laboratory MAE systems: closed microwave extraction vessels under controlled pressure and temperature and focused microwave ovens at atmospheric pressure [95]. The fundamentals of microwave-enhanced chemistry, including the theory behind sample preparation and new instruments, has been described by Kingston and Haswell [99]. MAE consists in subjecting a solid sample to microwave irradiation in a solvent. The process is based upon the fast localized heating of the solid sample without heating the vessels. The boiling point of the solvent is then rapidly reached, thus resulting in a short extraction time. The main beneﬁts of MAE are short extraction time, low solvent consumption, high extraction yield, and a simple process that ensures reproducibility at low cost. Pan et al. [100] have found that MAE extraction improves the efﬁciency of the extraction of polyphenols and caffeine from green tea leaves. However, the moisture content of samples is a deﬁning parameter for the recovery yield. When using dried samples, the recovery yield drops dramatically. Furthermore, Molins et al. [101] have reported that the use of hexane as the sole solvent in presence of a completely dry sample was not satisfactory. Indeed, the efﬁciency of MAE is typically low when the solvent lacks a signiﬁcant dipole moment for microwave energy

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 99 14.7.2008 9:45am Compositor Name: BMani

Ultrasound-Assisted Extraction in Food Analysis

99

absorption. The methods are therefore limited in terms of solvents and nature of the solid material. Compared to MAE, UAE may eventually be simpler [102] and faster [103]. In addition, UAE is not restricted by the solvent and type of matrix used, or by the moisture content.

5.6 ULTRASOUND-ASSISTED EXTRACTION: ENVIRONMENTAL IMPACT UAE is a clean method that avoids the use of large quantity of solvent and voluminous extraction vessels like Soxhlet and maceration. The reduced environmental impact of UAE is clearly advantageous in terms of energy and time. The energy required to perform the three extraction methods are, respectively, 6 kW h for maceration at the solvent’s boiling point (electrical energy for mechanical mixing and for heating), 8 kW h for Soxhlet (electrical energy for heating), and 0.25 kW h for UAE (electrical energy for ultrasound supply). The power consumption was determined with a Wattmeter at the ultrasound generator supply and the electrical heater power supply. Regarding environmental impact, the calculated quantity of carbon dioxide rejected in the atmosphere is higher in the case of Soxhlet (6400 g CO2=100 g of extracted solid material) and maceration (3600 g CO2=100 g of extracted solid material) than for UAE (200 g CO2=100 g of extracted solid material). These calculations have been carried out based on the following assumptions: To obtain 1 kW h from coal or fuel, 800 g of CO2 will be rejected in the atmosphere during combustion of fossil fuel. UAE is thus proposed as an ‘‘environmentally friendly’’ extraction method suitable for extraction prior to food analysis.

5.7 FUTURE TRENDS UAE of food components is increasingly efﬁcient at directly transferring knowledge into technology for commercial development. UAE makes use of physical and chemical phenomena that are fundamentally different from those applied in conventional extraction techniques. This novel process can extract analytes under a concentrated form (low volumes of solvent) and free from any contaminants or artefacts. The new systems developed to date clearly indicate that UAE offers net advantages in terms of yield, selectivity, operating time, energy input, and preservation of thermolabile compounds.

REFERENCES 1. Vinatoru, M., Toma, M., and Mason, T.J., Ultrasonically assisted extraction of bioactive principles from plant and their constituents, in Advances in Sonochemistry, Mason, Ed., 5, 209, 1999. 2. Reverchon, E., Supercritical ﬂuid extraction and fractionation of essential oils and related products, Journal of Supercritical Fluids, 10, 1, 1997. 3. Shu, Y.Y. et al., Analysis of polychlorinated biphenyls in aqueous samples by microwave-assisted headspace solid-phase microextraction, Journal of Chromatography A, 1008, 1, 2003. 4. Pare, J.R.J. and Belanger, J.M.R., Microwaves-assisted process (MAPTM): Principles and applications, in Pare, J.R.J. and Belanger, J.M.R. (Eds.), Instrumental Methods in Food Analysis, Elsevier Science: Amsterdam, The Netherlands, Chap. 10, pp. 395–420, 1997. 5. Rezzoug, S.A., Boutekedjiret, C., and Allaf, K., Optimization of operating conditions of rosemary essential oil extraction by a fast controlled pressure drop process using response surface methodology, Journal of Food Engineering, 71, 9, 2005. 6. Brachet, A. et al., Optimisation of accelerated solvent extraction of cocaine and benzoylecgonine from coca leaves, Journal of Separation Science, 24, 865, 2001. 7. Ozel, M.Z., Gogus, F., and Lewis, A.C., Subcritical water extraction of essential oils from Thymbra spicata, Food Chemistry, 82, 381, 2003. 8. McClements, D.J., Advances in the application of ultrasound in food analysis and processing, Trends in Food Science and Technology, 6, 293, 1995. 9. McClements, D.J. and Povey, M.J.W., Ultrasonic analysis of edible fats and oils, Ultrasonics, 30, 383, 1992.

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 100 14.7.2008 9:45am Compositor Name: BMani

100

Handbook of Food Analysis Instruments

10. McClements, D.J., Povey, M.J.W., and Dickinson, E., Absorption and velocity dispersion due to crystallization and melting of emulsion droplets, Ultrasonics, 31, 433, 1993. 11. Bjørnø, L., Ultrasonics International, Conference Proceedings, 1991. 12. Mason, T.J., Chemistry with Ultrasound, Elsevier Applied Science, New York, 1990. 13. Povey, M. and Mason, T.J., Ultrasound in Food Processing, Povey, M. and Mason, T.J. (Eds.), Blackie Academic & Professional, London, 1998. 14. Suslick, K.S., Ultrasound, Its Chemical, Physical and Biological Effects, VCH Publishers, New York, 1988. 15. Priego-Capote, F. and de Castro, L., Ultrasound-assisted digestion: A useful alternative in sample preparation, Journal of Biochemical and Biophysical Methods, 70, 299, 2007. 16. Lagha, A. et al., Microwave-ultrasound combined reactor suitable for atmospheric sample preparation procedure of biological and chemical products, Analusis, 27, 452, 1999. 17. Hu, A. et al., Ultrasound assisted supercritical ﬂuid extraction of oil and coixenolide from adlay seed, Ultrasonics Sonochemistry, 14, 219, 2007. 18. Luque-García, J.L. and Luque de Castro, M.D., Ultrasound-assisted Soxhlet extraction: An expeditive approach for solid sample treatment: Application to the extraction of total fat from oleaginous seeds, Journal of Chromatography A, 1034, 237, 2004. 19. Chemat, S. et al., Ultrasound assisted microwave digestion, Ultrasonics Sonochemistry, 11, 5, 2004. 20. Palma, M. and Barroso, C.G., Ultrasound-assisted extraction and determination of tartaric and malic acids from grapes and winemaking by-products, Analytica Chimica Acta, 458, 119, 2002. 21. Wu, J., Lin, L., and Chau, F., Ultrasound-assisted extraction of ginseng saponins from ginseng roots and cultured ginseng cells, Ultrasonics Sonochemistry, 8, 347, 2001. 22. Ebringerová, A. and Hromádková, Z., Effect of ultrasound on the extractability of corn bran hemicelluloses, Ultrasonics Sonochemistry, 9, 225, 2002. 23. Li, H., Chen, B., and Yao, S., Application of ultrasonic technique for extracting chlorogenic acid from Eucommia ulmodies Oliv. (E. ulmodies), Ultrasonics Sonochemistry, 12, 295, 2005. 24. Chemat, S. et al., Comparison of conventional and ultrasound-assisted extraction of carvone and limonene from caraway seeds, Flavour and Fragrance Journal, 19, 188, 2004. 25. Cabredo-Pinillos, S. et al., Ultrasound-assisted extraction of volatile compounds from wine samples: Optimisation of the method, Talanta, 69, 1123, 2006. 26. Hardcastle, J.L., Paterson, C.J., and Compton, R.G., Biphasic sonoelectroanalysis: Simultaneous extraction from, and determination of vanillin in food ﬂavoring, Electroanalysis, 13, 899, 2001. 27. Cocito, C., Gaetano, G., and Delﬁni, C., Rapid extraction of aroma compounds in must and wine by means of ultrasound, Food Chemistry, 52, 311, 1995. 28. Hernanz Vila, D. et al., Optimization of an extraction method of aroma compounds in white wine using ultrasound, Talanta, 50, 413, 1999. 29. Alissandrakis, E. et al., Evaluation of four isolation techniques for honey aroma compounds, Journal of the Science of Food and Agriculture, 85, 91, 2005. 30. Alissandrakis, E. et al., Ultrasound-assisted extraction of volatile compounds from citrus ﬂowers and citrus honey, Food Chemistry, 82, 575, 2003. 31. Caldeira, I. et al., Improved method for extraction of aroma compounds in aged brandies and aqueous alcoholic wood extracts using ultrasound, Analytica Chimica Acta, 513, 125, 2004. 32. Shotipruk, A., Kaufman, P.B., and Wang, H.Y., Feasibility study of repeated harvesting of menthol from biologically viable Mentha x piperata using ultrasonic extraction, Biotechnology Progress, 17, 924, 2001. 33. Peña, R.M. et al., Comparison of ultrasound-assisted extraction and direct immersion solid-phase microextraction methods for the analysis of monoterpenoids in wine, Talanta, 67, 129, 2005. 34. Sharma, A. et al., Microwave- and ultrasound-assisted extraction of vanillin and its quantiﬁcation by high-performance liquid chromatography in Vanilla planifolia, Journal of Separation Science, 29, 613, 2006. 35. Kanakis, C.D. et al., Qualitative determination of volatile compounds and quantitative evaluation of safranal and 4-hydroxy-2,6,6-trimethyl-cyclohexene-1-carboxaldehyde (HTCC) in Greek saffron, Journal of Agricultural and Food Chemistry, 52, 4515, 2004. 36. Manutsewee, N. et al., Determination of Cd, Cu, and Zn in ﬁsh and mussel by AAS after ultrasoundassisted acid leaching extraction, Food Chemistry, 101, 817, 2007.

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 101 14.7.2008 9:45am Compositor Name: BMani

Ultrasound-Assisted Extraction in Food Analysis

101

37. Río-Segade, S. and Bendicho, C., Ultrasound-assisted extraction for mercury speciation by the ﬂow injection-cold vapor technique, Journal of Analytical Atomic Spectrometry, 14, 263, 1999. 38. Lavilla, I., Capelo, J.L., and Bendicho, C., Determination of cadmium and lead in mussels by electrothermal atomic absorption spectrometry using an ultrasound-assisted extraction method optimized by factorial design, Fresenius’ Journal of Analytical Chemistry, 363, 283, 1999. 39. Filgueiras, A.V. et al., Comparison of ultrasound-assisted extraction and microwave-assisted digestion for determination of magnesium, manganese and zinc in plant samples by ﬂame atomic absorption spectrometry, Talanta, 53, 433, 2000. 40. Krishna, M.V.B. and Arunachalam, J., Ultrasound-assisted extraction procedure for the fast estimation of major, minor and trace elements in lichen and mussel samples by ICP-MS and ICP-AES, Analytica Chimica Acta, 522, 179, 2004. 41. Del Carmen Yebra, M., Moreno-Cid, A., and Cancela, S., Flow injection determination of copper and iron in seafoods by a continuous ultrasound-assisted extraction system coupled to FAAS, International Journal of Environmental Analytical Chemistry, 85, 315, 2005. 42. Moreno-Cid, A. and Yebra, M.C., Continuous ultrasound-assisted extraction coupled to a ﬂow injectionﬂame atomic absorption spectrometric system for calcium determination in seafood samples, Analytical and Bioanalytical Chemistry, 379, 77, 2004. 43. Cava-Montesinos, P. et al., On-line speciation of mercury in ﬁsh by cold vapour atomic ﬂuorescence through ultrasound-assisted extraction, Journal of Analytical Atomic Spectrometry, 19, 1386, 2004. 44. Šucman, E. and Bednár, J., Determination of chlorides in meat products with ion-selective electrode using the batch injection technique, Electroanalysis, 15, 866, 2003. 45. Yebra, M.C., Cancela, S., and Moreno-Cid, A., Continuous ultrasound-assisted extraction of cadmium from vegetable samples with on-line preconcentration coupled to a ﬂow injection-ﬂame atomic spectrometric system, International Journal of Environmental Analytical Chemistry, 85, 305, 2005. 46. Yebra, M.C. and Cancela, S., Continuous ultrasound-assisted extraction of cadmium from legumes and dried fruit samples coupled with on-line preconcentration-ﬂame atomic absorption spectrometry, Analytical and Bioanalytical Chemistry, 382, 1093, 2005. 47. Jalbani, N. et al., Application of factorial design in optimization of ultrasonic-assisted extraction of aluminum in juices and soft drinks, Talanta, 70, 307, 2006. 48. Nascentes, C.C., Korn, M., and Arruda, M.A.Z., A fast ultrasound-assisted extraction of Ca, Mg, Mn and Zn from vegetables, Microchemical Journal, 69, 37, 2001. 49. Moreno-Cid, A. et al., Flow injection on-line ultrasound-assisted extraction of iron in meat samples coupled to a ﬂame atomic absorption spectrometric system, Analytical and Bioanalytical Chemistry, 377, 730, 2003. 50. Yebra-Biurrun, M.C., Moreno-Cid, A., and Cancela-Pérez, S., Fast on-line ultrasound-assisted extraction coupled to a ﬂow injection-atomic absorption spectrometric system for zinc determination in meat samples, Talanta, 66, 691, 2005. 51. Yebra, M.C. and Moreno-Cid, A., Continuous ultrasound-assisted extraction of iron from mussel samples coupled to a ﬂow injection-atomic spectrometric system, Journal of Analytical Atomic Spectrometry, 17, 1425, 2002. 52. Yebra-Biurrun, M.C., Cancela-Pérez, S., and Moreno-Cid-Barinaga, A., Coupling continuous ultrasound-assisted extraction, preconcentration and ﬂame atomic absorption spectrometric detection for the determination of cadmium and lead in mussel samples, Analytica Chimica Acta, 533, 51, 2005. 53. Moreno-Cid, A. and Yebra, M.C., Flow injection determination of copper in mussels by ﬂame atomic absorption spectrometry after on-line continuous ultrasound-assisted extraction, Spectrochimica Acta— Part B Atomic Spectroscopy, 57, 967, 2002. 54. García-Rey, R.M., Quiles-Zafra, R., and De Castro, M.D.L., New methods for acceleration of meat sample preparation prior to determination of the metal content by atomic absorption spectrometry, Analytical and Bioanalytical Chemistry, 377, 316, 2003. 55. Sanz, E., Muñoz-Olivas, R., and Cámara, C., A rapid and novel alternative to conventional sample treatment for arsenic speciation in rice using enzymatic ultrasonic probe, Analytica Chimica Acta, 535, 227, 2005. 56. Méndez, H. et al., Ultrasonic extraction combined with fast furnace analysis as an improved methodology for total selenium determination in seafood by electrothermal-atomic absorption spectrometry, Analytica Chimica Acta, 452, 217, 2002.

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 102 14.7.2008 9:45am Compositor Name: BMani

102

Handbook of Food Analysis Instruments

57. Vian, M.A. et al., Simple and rapid method for cis- and trans-resveratrol and piceid isomers determination in wine by high-performance liquid chromatography using Chromolith columns, Journal of Chromatography A, 1085, 224, 2005. 58. Suárez, B. et al., Liquid chromatographic method for quantifying polyphenols in ciders by direct injection, Journal of Chromatography A, 1066, 105, 2005. 59. Tsanova-Savova, S., Ribarova, F., and Gerova, M., (þ)-Catechin and ()-epicatechin in Bulgarian fruits, Journal of Food Composition and Analysis, 18, 691, 2005. 60. Albu, S. et al., Potential for the use of ultrasound in the extraction of antioxidants from Rosmarinus ofﬁcinalis for the food and pharmaceutical industry, Ultrasonics Sonochemistry, 11, 261, 2004. 61. Bruni, R. et al., Wild Amaranthus caudatus seed oil, a nutraceutical resource from ecuadorian ﬂora, Journal of Agricultural and Food Chemistry, 49, 5455, 2001. 62. Bruni, R. et al., Rapid techniques for the extraction of vitamin E isomers from Amaranthus caudatus seeds: Ultrasonic and supercritical ﬂuid extraction, Phytochemical Analysis, 13, 257, 2002. 63. Li, H., Chen, B., and Yao, S., Application of ultrasonic technique for extracting chlorogenic acid from Eucommia ulmodies Oliv. (E. ulmodies), Ultrasonics Sonochemistry, 12, 295, 2005. 64. Japón-Luján, R., Luque-Rodríguez, J.M., and Luque De Castro, M.D., Dynamic ultrasound-assisted extraction of oleuropein and related biophenols from olive leaves, Journal of Chromatography A, 1108, 76, 2006. 65. Ruiz-Jiménez, J. and Luque De Castro, M.D., Flow injection manifolds for liquid–liquid extraction without phase separation assisted by ultrasound, Analytica Chimica Acta, 489, 1, 2003. 66. Akkermans, R.P., Wu, M., and Compton, R.G., A Comparison between pulsed sonovoltammetry and low power laser activated voltammetry for the electroanalysis of ascorbic acid in a commercial fruit drink, Electroanalysis, 10, 814, 1998. 67. Rostagno, M.A., Palma, M., and Barroso, C.G., Ultrasound-assisted extraction of soy isoﬂavones, Journal of Chromatography A, 1012, 119, 2003. 68. Piñeiro, Z., Palma, M., and Barroso, C.G., Determination of catechins by means of extraction with pressurized liquids, Journal of Chromatography A, 1026, 19, 2004. 69. Goli, A.H., Barzegar, M., and Sahari, M.A., Antioxidant activity and total phenolic compounds of pistachio (Pistachia vera) hull extracts, Food Chemistry, 92, 521, 2005. 70. Rodrigues, S. and Pinto, G.A.S., Ultrasound extraction of phenolic compounds from coconut (Cocos nucifera) shell powder, Journal of Food Engineering, 80, 869, 2007. 71. Tena, M.T. et al., Supercritical ﬂuid extraction off natural antioxidants from Rosemary: Comparison with liquid solvent sonication, Analytical Chemistry, 69, 521, 1997. 72. Herrera, M.C. and Luque De Castro, M.D., Ultrasound-assisted extraction for the analysis of phenolic compounds in strawberries, Analytical and Bioanalytical Chemistry, 379, 1106, 2004. 73. Herrera, M.C. and Luque De Castro, M.D., Ultrasound-assisted extraction of phenolic compounds from strawberries prior to liquid chromatographic separation and photodiode array ultraviolet detection, Journal of Chromatography A, 1100, 1, 2005. 74. Martino, E. et al., Microwave-assisted extraction of coumarin and related compounds from Melilotus ofﬁcinalis (L.) Pallas as an alternative to Soxhlet and ultrasound-assisted extraction, Journal of Chromatography A, 1125, 147, 2006. 75. Waksmundzka-Hajnos, M. et al., Inﬂuence of the extraction mode on the yield of some furanocoumarins from Pastinaca sativa fruits, Journal of Chromatography B: Analytical Technologies in the Biomedical and Life Sciences, 800, 181, 2004. 76. Luque-García, J.L. and Luque de Castro, M.D., Ultrasound-assisted Soxhlet extraction: An expeditive approach for solid sample treatment: Application to the extraction of total fat from oleaginous seeds, Journal of Chromatography A, 1034, 237, 2004. 77. Li, H., Pordesimo, L., and Weiss, J., High intensity ultrasound-assisted extraction of oil from soybeans, Food Research International, 37, 731, 2004. 78. Luo, D. et al., Ultrasound assisted extraction of coix lacryma-jobi seed oil in supercritical CO2, Journal of Shaanxi University of Science and Technology, 23, 24, Dec. 2005. 79. Hu, A. et al., Ultrasound assisted supercritical ﬂuid extraction of oil and coixenolide from adlay seed, Ultrasonics Sonochemistry, 14, 219, 2007. 80. Riera, E. et al., Mass transfer enhancement in supercritical ﬂuids extraction by means of power ultrasound, Ultrasonics Sonochemistry, 11, 241, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 103 14.7.2008 9:45am Compositor Name: BMani

Ultrasound-Assisted Extraction in Food Analysis

103

81. Sharma, A. and Gupta, M.N., Oil extraction from almond, apricot and rice bran by three-phase partitioning after ultrasonication, European Journal of Lipid Science and Technology, 106, 183, 2004. 82. Ruiz-Jiménez, J. et al., Use of chemometrics and mid infrared spectroscopy for the selection of extraction alternatives to reference analytical methods for total fat isolation, Analytica Chimica Acta, 525, 159, 2004. 83. Ruiz-Jiménez, J. and Luque De Castro, M.D., Forward-and-back dynamic ultrasound-assisted extraction of fat from bakery products, Analytica Chimica Acta, 502, 75, 2004. 84. Li, H. et al., Microwave and ultrasound assisted extraction of soybean oil, Transactions of the American Society of Agricultural Engineers, 47, 1187, 2004. 85. Luque de Castro, M.D. and García-Ayuso, L.E., Soxhlet extraction of solid materials: An outdated technique with a promising innovative future, Analytica Chimica Acta, 369, 1, 1998. 86. Hechler, U., Fischer, J., and Plagemann, S., Comparison of different extraction methods for the determination of polycyclic aromatic hydrocarbons in soil, Fresenius’ Journal of Analytical Chemistry, 351, 591, 1995. 87. Beard, A., Naikwadi, K., and Karasek, F.W., Comparison of extraction methods for polychlorinated dibenzo-p-dioxins and dibenzofurans in ﬂy ash using gas chromatography–mass spectrometry, Journal of Chromatography, 589, 265, 1992. 88. Luque de Castro, M.D., Valcárcel, M., and Tena, M.T., Analytical Supercritical Fluid Extraction, Springer-Verlag, Heidelberg, 1994. 89. Chester, T.L., Pinkston, J.D., and Raynie, D.E., Supercritical ﬂuid chromatography and extraction, Analytical Chemistry, 70, 301R, 1998. 90. Luque de Castro, M.D., Valcárcel, M., and Tena, M.T., Analytical Supercritical Fluid Extraction, Springer-Verlag, New York, 1994. 91. Luque de Castro, M.D. and Jiménez-Carmona, M.M., Where is supercritical ﬂuid extraction going? Trends in Analytical Chemistry, 19, 223, 2000. 92. Lang, Q. and Wai, C.M., Supercritical ﬂuid extraction in herbal and natural product studies—A practical review, Talanta, 53, 771, 2001. 93. Majors, R.E., Modern techniques for the extraction of solid materials—An update, LC-GC North America, 24, 73, 2006. 94. Brachet, A. et al., Optimisation of accelerated solvent extraction of cocaine and benzoylecgonine from coca leaves, Journal of Separation Science, 24, 865, 2001. 95. Kaufmann, B. and Christen, P., Recent extraction techniques for natural products: Microwave-assisted extraction and pressurised solvent extraction, Phytochemical Analysis, 13, 105, 2002. 96. Richter, B.E. et al., Accelerated solvent extraction: A technique for sample preparation, Analytical Chemistry, 68, 1033, 1996. 97. Fisher, J.A., Scarlett, M.J., and Stott, A.D., Accelerated solvent extraction: An evaluation for screening of soils for selected U.S. EPA semivolatile organic priority pollutants, Environmental Science and Technology, 31, 1120, 1997. 98. Assis Jacques, R. et al., Chemical composition of mate tea leaves (llex paraguariensis): A study of extraction methods, Journal of Separation Science, 29, 2780, 2006. 99. Kingston, S. and Haswell, S.J., Microwave-Enhanced Chemistry Fundamentals, Sample Preparation, and Applications, Edited by H.M., American Chemical Society, 1997. 100. Pan, X., Niu, G., and Liu, H., Microwave-assisted extraction of tea polyphenols and tea caffeine from green tea leaves, Chemical Engineering and Processing, 42, 129, 2003. 101. Molins, C. et al., Microwave assisted solvent extraction (MASE) of organochlorine pesticides from soil samples, International Journal of Environmental Analytical Chemistry, 68, 155, 1997. 102. Luque-García, J.L. and Luque De Castro, M.D., Ultrasound: A powerful tool for leaching, Trends in Analytical Chemistry, 22, 41, 2003. 103. Lopez-Avila, V., Young, R., and Teplitsky, N., Microwave-assisted extraction as an alternative to Soxhlet, Sonication, and Supercritical Fluid Extraction, Journal of AOAC International, 79, 142, 1996.

Otles/Handbook of Food Analysis Instruments 45660_C005 Final Proof page 104 14.7.2008 9:45am Compositor Name: BMani

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 105 15.7.2008 8:19am Compositor Name: BMani

in High6 Advances Performance Liquid Chromatography and Its Application to the Analysis of Foods and Beverages Peter Varelis CONTENTS 6.1 6.2 6.3

Solvent Delivery Systems .................................................................................................... 105 Mass Spectrometric Detection ............................................................................................. 106 Sample Preparation Strategies ............................................................................................. 110 6.3.1 Immunoafﬁnity Columns ........................................................................................ 110 6.3.2 Molecularly Imprinted Polymers ............................................................................. 110 6.3.3 Online Sample Preparation ...................................................................................... 111 6.3.4 Solid-Phase Extraction ............................................................................................ 112 References ..................................................................................................................................... 115

6.1 SOLVENT DELIVERY SYSTEMS Although there have been considerable improvements in the design of autosamplers and solvent delivery systems, arguably, the most signiﬁcant advancement in high-performance liquid chromatography (HPLC) in more than a decade has been the introduction of so-called ultra-performance liquid chromatographs (UPLC). In contrast to conventional HPLC systems that have a pressure limitation of around 5500 psi, UPLC systems are capable of operating at more than twice this pressure. The distinct advantage that UPLC systems have over conventional systems is that sub-twomicron particle size columns can be used on these systems. The main beneﬁt of such columns is that, unlike larger particle size columns, these can be operated at comparatively higher ﬂow rates without signiﬁcantly compromising column efﬁciency (Figure 6.1). In practical terms, this means that analytical run times can be signiﬁcantly shortened—in some case by as much as an order of magnitude—while maintaining good resolution. Furthermore, in addition to the considerable economic advantages, short analytical runs signiﬁcantly improve the signal-to-noise ratio by reducing chromatographic dispersion effects such as peak broadening. For instance, the UPLCmass spectrometric response of 16 priority pesticides was such that their signal-to-noise ratios were signiﬁcantly better than the corresponding ratios obtained using conventional HPLC. A review of the literature shows that UPLC has been applied to the analysis of pesticides [1–3], veterinary drugs [4,5], mycotoxins [6–8], and heat-derived toxins [9,10] in foods.

105

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 106 15.7.2008 8:19am Compositor Name: BMani

106

Handbook of Food Analysis Instruments

10 µm 5 µm HETP

3.5 µm

Sub 2 µm

Linear velocity

FIGURE 6.1 Qualitative relationship between column efﬁciency (height equivalent theoretical plates, HETP) and linear velocity for columns with various particle sizes.

6.2 MASS SPECTROMETRIC DETECTION The coupling of mass spectrometers to liquid chromatographic systems is, arguably, the most important recent development in the detection of solutes in liquid solutions. In particular, the introduction of tandem mass spectrometric and ion trap detectors with atmospheric pressure ionization sources for HPLC has signiﬁcantly simpliﬁed and improved the selectivity, and in general the sensitivity, of qualitative and quantitative methods for the analysis of foods. For instance, in our own work with the mycotoxin ochratoxin A in wine, the analytical run time was reduced by a factor of 9 when the method was changed from ﬂuorescence detection [11] to tandem mass spectrometric [12] detection. Multiresidue analysis of foods and drinking water by HPLC-tandem mass spectrometry is, perhaps, a more illustrative example of the impact that this hyphenated technique has had on food safety [13,14]. Indeed, a recent GC–LC mass spectrometric comparative study of 500 high priority pesticides found that, with the exception of the organochlorine pesticides, LC–MS shows wider scope and better sensitivity [14]. A further advantage that HPLC-mass spectrometric detectors have over more conventional and less-selective detectors such as ultraviolet–visible and ﬂuorescence detectors is that they are well suited to automated online extraction and concentration of analytes such as pesticide residues in drinking water and biological ﬂuids [15–17]. This is particularly the case for tandem mass spectrometers. These systems can be operated in several highly selective modes, of which the so-called selected reaction monitoring is the more common and, generally, the most selective. In this mode, which is also referred to as multiple reaction monitoring, there are two stages of mass ﬁltering. The ﬁrst occurs after ionization and the second occurs after the initially ﬁltered ions are fragmented and immediately before detection. In the case of triple-stage quadrupole instruments, fragmentation is achieved by accelerating the ions and causing them to collide with an inert gas such as argon in the second quadrupole, which only has an ion transmission capability. Consequently, it is only the ﬁrst and the third quadrupoles of triple-stage quadrupoles instruments that have a mass resolution capability. In the case of ion trap mass spectrometers, which are capable of multiple fragmentation experiments or so-called msn, ions are trapped, fragmented, and ejected using a combination of voltages and radiofrequencies. It is through this consecutive mass resolution process that tandem mass spectrometers achieve their unsurpassed selectivity in that the chance of isobaric masses having the same mass-to-charge ratio after fragmentation is unlikely. This, however, may not necessarily be the case for structurally isomeric compounds such as diastereoisomers, geometric or positional isomers, and it is certainly not the case for enantiomers. For example, the structurally isomeric lycopenes all give the same collision-induced dissociation spectra. In other words, the

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 107 15.7.2008 8:19am Compositor Name: BMani

107

Advances in High-Performance Liquid Chromatography

parent–daughter ion transitions are the same for all isomers. Consequently, chromatographic separation is necessary to distinguish between the lycopene isomers [18]. While mass spectrometric detection offers distinct advantages over conventional HPLC detectors, these instruments have their unique problems that can severely limit their application. Two particularly notable problems with mass spectrometric detection are matrix effects and interchannel cross-talk. In the case of matrix effects, the mass spectrometric response is either enhanced or, more usually, suppressed by matrix components in the sample extract. As the total ion chromatograms in Figure 6.2 demonstrate, matrix effects can be severe. The top chromatogram was obtained from the analysis of a coffee extract for the probable human carcinogen acrylamide and the bottom is that of a standard in which the concentration of acrylamide is similar to that in the coffee extract. The phenomenon is thought [19] to be caused by co-eluting nonvolatile solutes such as proteins changing the colligative properties of the mobile phase, and thereby, preventing the transfer of ions from solution into the gas phase. Of the two commonly used ion sources, electrospray ionization is more prone to this vexing problem than is atmospheric pressure chemical ionization [19,20]. In the absence of internal standards, matrix effects may go unnoticed in that chromatograms may otherwise look unremarkable. In this respect, it should be emphasized that co-eluting internal standards such as those used in isotope dilution techniques may not necessarily or adequately compensate for matrix effects [21]. Indeed, we found this to be the case in the analysis of coffee and cocoa for acrylamide using an HPLC-MS-MS stable isotope method [22]. After changing from a

RT: 0.00–3.60 SM: 7G 100 90

Acrylamide

80

NL: 1.68 E5 TIC MS ICIS 22B

AA: 522823

70 60

Relative abundance

50 40 30 20 10 0 100

AA: 8418054

90 80 70

NL: 1.36 E6 TIC MS ICIS cal03

60 50 40 30 20 10 0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Time (min)

FIGURE 6.2 Ion suppression caused by matrix effects. The top chromatogram was obtained by spiking a coffee extract with acrylamide so that its concentration was equivalent to the standard used to obtain the bottom chromatogram.

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 108 15.7.2008 8:19am Compositor Name: BMani

108

Handbook of Food Analysis Instruments Acrylamide

Acrylamide

Y = –0.0112001 + 0.000922844∗X R2 = 0.9760 W: Equal

Y = 0.0001889 + 0.000798272∗X R2 = 1.0000 W: Equal

0.14 0.12 0.10

0.10 Area ratio

Area ratio

0.12 0.08 0.06 0.04

0.08 0.06 0.04 0.02

0.02 0.00

0.00 0

50

(A)

100

150

0

50

100

150

ng

(B)

ng

FIGURE 6.3 Calibration curves obtained by interspersion of the standards among the sample extracts before (A) and after (B) puriﬁcation of the extracts using two alternative sample preparation method.

nonpolar C18 column to a polar graphite analytical column, we observed a dramatic improvement in both the abundance and peak shape of the analyte and its carbon-13 labeled internal standard that suggested we had overcome the matrix effects we encountered in earlier analyses. However, the calibration curve (Figure 6.3A) suggested otherwise in that the lower calibration standards were considerably more dispersed than usual. That matrix effects were the cause of the problem was conﬁrmed by the excellent linearity (Figure 6.3B) of the lower calibration standards after the coffee and cocoa extracts were puriﬁed using an alternative sample preparation technique that allowed for the selective extraction of acrylamide form coffee and cocoa [22]. Similarly, Zollner et al. [23] demonstrated that there were less matrix effects in the analysis of ochratoxin A in wines after they were processed using a highly selective sample preparation technique that was based on immunoafﬁnity chromatography. The need to assess for matrix effects in qualitative and quantitative HPLC-MS methods cannot be overemphasized—particularly when nonselective extraction techniques such as protein precipitation and solid-phase extraction are used to prepare samples for the analysis of trace components in foods and beverages. Interestingly, of the three commonly used sample preparation techniques, liquid–liquid extraction, solid-phase extraction, and protein precipitation, liquid–liquid extraction was reported [24] to cause the least ion suppression in the electrospray ionization mode. An elegant and convenient method for assessing matrix effects has been described by Bonﬁglio and his colleagues [24]. The method involves postcolumn infusion of the analyte into the mass spectrometer to establish a baseline above the chemical noise (Figure 6.4). An aliquot of the sample extract is then injected on to the column and a region where the baseline drops below the arbitrary zero mark is an Mass spectrometer Autosampler Pump

T-piece

Ion source

Column

Syringe-infusion pump ensemble

FIGURE 6.4

Schematic for assessing matrix effect using the method described by Bonﬁglio et al. [24].

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 109 15.7.2008 8:19am Compositor Name: BMani

109

Advances in High-Performance Liquid Chromatography

Intensity

a 28,00,000 27,00,000 26,00,000 25,00,000 24,00,000 23,00,000 22,00,000 21,00,000 20,00,000 19,00,000 18,00,000 17,00,000 16,00,000 15,00,000 14,00,000 13,00,000 12,00,000 11,00,000 10,00,000 9,00,000 8,00,000 7,00,000 6,00,000 5,00,000 4,00,000 3,00,000 2,00,000 1,00,000 0 0.0

c

b

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

2.6

2.8

3.0

Time (min)

FIGURE 6.5 Typical ion suppression curves obtained by injecting extracts that were obtained using two alternative sample preparation techniques (a and b). The third (c) was obtained by injecting the solvent used to reconstitute the residues that were obtained from the two sample preparations.

indication of ion suppression. On the other hand, a region where the baseline increases above the arbitrary zero mark is an indication of signal enhancement. The results of typical ion suppression experiments using this technique are shown in Figure 6.5. The traces were obtained by injecting wine extracts obtained using two alternative solid-phase extraction techniques [11]. The third trace is the ion suppression caused by injecting the solvent used to reconstitute the residue obtained after evaporation of the wine extracts. These experiments demonstrate that the ion suppression caused by one of the sample preparation techniques is no worse than that caused by the solvent, whereas the other sample preparation technique caused signiﬁcant ion suppression. It is also worth noting that these experiments demonstrate that there is considerable ion suppression at the solvent front and so elution in this region of the chromatogram should, where possible, be avoided. A practical limitation of assessing matrix effects is that the analyte may be present in the extract. Consequently, using such extracts to determine matrix effects will either underestimate ion suppression or produce the artifact of ion enhancement or overestimate this phenomenon. In such situations, analogues such as homologues and stable isotopes can be used to assess matrix effects. For example, in our laboratory we used [13C3]acrylamide to assess matrix effects in coffee and cocoa extracts. Interchannel cross-talk is an issue that is exclusively restricted to triple-stage quadrupole mass spectrometers. This undesirable phenomenon, which produces a false-positive signal, arises when two consecutive selected reaction monitoring (SRM) channels have a common product ion and the ion transit time through the collision cell is longer than the scan time. In other words, ions from the previous SRM experiment are detected in the subsequent experiment because insufﬁcient time has been allocated for the ions to exit the collision cell. Interchannel cross-talk is a potential problem when short chromatographic run times are used for multicomponent analyses in which some of the analytes belong to the same chemical class and have common features in their MS-MS spectra. This is particularly the case for multiresidue analyses such as pesticides. In such high throughput analyses where chromatographic resolution is sacriﬁced to decrease analytical run times, the potential for cross-talk must be carefully assessed to avoid false positives.

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 110 15.7.2008 8:19am Compositor Name: BMani

110

Handbook of Food Analysis Instruments

6.3 SAMPLE PREPARATION STRATEGIES 6.3.1 IMMUNOAFFINITY COLUMNS Since the ﬁrst reported [25] application of immobilized antibodies to isolate aﬂatoxin B1 from foods and feeds in 1990, there has been a consistent trend towards using this highly selective solid-phase extraction technique for the analysis of food contaminants. In particular, the use of so-called immunoafﬁnity columns, i.e., antibodies that have been immobilized onto beads, has been applied to the extraction of mycotoxins [26,27] from foods, feeds, and beverages. Although, their use to extract other trace residues such as pesticides [28,29] and antibiotics [30,31] has also been reported, these will, undoubtedly, become an important application of this selective solid-phase extraction technique. The major advantage of immunoafﬁnity solid-phase extraction is its ability to very selectively and expediently extract trace amounts of analytes from within complex matrices such as wine, eggs, milk and, to a lesser extent, unlike conventional solid-phase extraction, immunoafﬁnity chromatography does not involve the use of large volumes of toxic organic solvents. The selectivity of this solid-phase extraction technique is evident from the impressive chromatograms that are obtained from the analysis of complex samples in that there is often no other compounds in the region where the analyte elutes. Indeed, the chromatograms are not unlike those obtained using mass spectrometric detectors. However, this technique is not without its problems. For instance, immunoafﬁnity columns have, when compared with conventional columns, very limited capacities, are generally expensive, and have a limited shelf-live. Furthermore, their performance is dependent upon the chemistry used to immobilize the protein antibodies to the support matrix, the nature of this matrix, and the method used to generate the antibodies [32]. Consequently, the performance of immunoafﬁnity columns can vary considerably between suppliers. Additionally, the subtle mechanism by which the protein antibodies recognize the antigen can be affected under conditions that either denature proteins such as heat, ionic strength, or by matrix components that interfere with binding. With respect to the last point, the most well-recognized phenomenon is, arguably, the so-called cross-reactivity in which compounds that are structurally related to the antigen such as metabolites compete with the analyte for binding to the antibodies. For instance, the ethyl ester of ochratoxin A is known to strongly cross-react with the antibodies to ochratoxin A and to a lesser extent with ochratoxin B. Consequently, because of this phenomenon, there is a potential to underestimate the amount of analyte when selective chromatographic methods such as HPLC are used to analyze the extract obtained using immunoafﬁnity chromatography. Furthermore, because of their capacity limitation, the possibility that a negative result is a consequence of cross-reactivity must be considered. The recent [33] commercial availability of stable isotope-containing mycotoxins that can be used as internal standards addresses some of these issues and, in some cases, can improve the versatility of a method by adequately compensating for preparative losses associated with the matrix. However, such standards are generally not readily available and, where they are used, mass spectrometric detection is required. In some respects, it can be argued that the combined use of immunoafﬁnity chromatography and HPLC-mass spectrometric detection is unnecessary for routine analyses in so far as these are two highly selective techniques, which in the case of immunoafﬁnity chromatography work well with less-selective and expensive detectors and in the case of tandem mass spectrometric detection less-selective sample preparation methods can be used to extract the analyte. However, it must be emphasized that the combination of these two highly selective techniques, particularly in the case where internal standards are used, improves the validity of an assay. It is worth noting that extracts obtained using immunoafﬁnity chromatography are unlikely to cause the vexing problem of ion suppression in LC–MS assays.

6.3.2 MOLECULARLY IMPRINTED POLYMERS Although this novel and exciting sample preparation technique has not yet come into the main stream of analytical chemistry, it is worth brieﬂy discussing because of its emerging importance as a selective solid-phase extraction technique.

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 111 15.7.2008 8:19am Compositor Name: BMani

111

Advances in High-Performance Liquid Chromatography

Molecularly imprinted polymers (MIPs) are, as the name suggests, polymeric materials in which a three-dimensional shape that corresponds to the molecular topography of the analyte has been imprinted into a polymer using a suitable template [34]. The most commonly used method of producing MIPs is to use functional monomers that form weak covalent bonds with the template, which may be the analyte or a compound that has a molecular analogy with the analyte of interest. The monomers are then polymerized in the presence of a cross-linking agent to form a crosslinked polymer. The template is then extracted using an organic solvent to create a deﬁned threedimensional cavity, which is capable of selectively binding the analyte through weak covalent or ionic bonds with the functional components that were designed into the polymer [35]. In this manner, an MIP can be designed to recognize a single compound or a group of compounds such as pesticides. It is worth noting that, in principle, the mechanism by which MIPs extract analytes from solution mimics that of immunoafﬁnity chromatography. Consequently, it should not be surprising that the extracts obtained using MIPs are more reﬁned than the extracts obtained using conventional sorbents. Some important advantage that MIPs have over their immunoafﬁnity counterparts is that they are more robust and have higher loading capacities. A major limitation of MIPs is bleeding of residual template from the polymer. To some extent, this can be overcome by using structural analogues of the analytes of interest as templates. However, the selectivity of the sorbent may be compromised and analogues of complex natural products such as mycotoxins and antibiotics may not be readily available. Despite these limitations, several MIP sorbents are commercially available [36] for extracting some important food chemical contaminants.

6.3.3 ONLINE SAMPLE PREPARATION Online sample preparation is a process that integrates the extraction and concentration of analytes into the analytical method. In its simplest form, this may involve direct injection of the sample onto an analytical column. For example, in the case of the analysis of ochratoxin A in red wines, the samples were directly injected onto an analytical column using an initial mobile-phase composition that allowed for the extraction and concentration of the analyte. After this initial period, the analyte was eluted from the column into a tandem mass spectrometric detector by changing the mobilephase composition [37]. However, it is more usual in online methods to use a separate column or cartridge to perform the extraction and concentration of the analyte. In this situation, a six port valve and an additional solvent delivery system are required to pump mobile phase through the secondary column (Figure 6.6). The sample is loaded onto the column using a mobile-phase composition that allows for the separation of the analyte from the bulk matrix components while retaining the analyte as close to the head of the column as possible. The valve is then switched, and the ﬂow from the second

B A

Waste

Analytical column Detector

A

Detector

Waste

Online spe column (A)

Load-extraction position

(B)

Analysis position

FIGURE 6.6 Typical conﬁguration for online sample extraction using a six port valve. A, solvent delivery system; B, autosampler.

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 112 15.7.2008 8:19am Compositor Name: BMani

112

Handbook of Food Analysis Instruments

pump is used to back ﬂush the analyte onto the analytical column. The mobile-phase composition used at this stage should be such that the analyte elutes onto and is, preferably, initially retained by, the analytical column as a narrow band so as to prevent peak broadening. Although conventional analytical columns can and have been used to perform online sample processing, they are not particularly suited to this task because of their low porosity and the high back pressures that can be generated as the columns begin to foul with bulk matrix components such as proteins and oligosaccharides. Several columns are commercially available that overcome these issues. A common feature of these columns is that they have irregular macroporous particles that allow for high ﬂow rates, low back pressures, and good mass transfer. Furthermore, the pores of these particles are such that small molecules diffuse into the so-called mesopores faster than macromolecules, which are ﬂushed out of the column. In other words, these columns have size exclusion-like properties that allow for the separation of small molecules from macromolecules. A further feature of these particles, which improves their selectivity, is that the chemistry within the mesopores can be selected so that the analytes interact more strongly with the phase than with other small molecules. Consequently, small molecules that enter the mesopores and have a low afﬁnity for the phase will diffuse out of pores faster than those compounds that have a higher afﬁnity for the phase. An alternative to using a secondary column to perform online sample preparation is in-tube solid-phase microextraction [38]. In this particular technique, a short length of coated capillary column of the type used in gas chromatography is attached to the autoinjector unit of an HPLC. Extraction of the analyte is then effected by repeatedly aspirating the sample onto the column, ideally, until equilibrium is reached. The analyte is subsequently eluted onto the column by either directly passing mobile phase through the capillary column or by aspirating a solvent of choice from a second vial. With modern autoinjectors that have pre- and post-treatment capabilities, there is an opportunity to reﬁne and expedite the extraction process by using multiple solvents. One notable beneﬁt of this technique when the mobile phase is used to desorb the analyte is that the solvent peak is signiﬁcantly reduced or eliminated. A review of the literature reveals that in-tube solid-phase microextraction has been used for the analysis of pesticides in water [39], antibiotics in ﬁsh and eggs [40,41], phytoestrogens from soybean foods [42], endocrine disruptors in foods [43], and catechin and caffeine in tea [44]. With the exception of highly selective extraction columns such as molecularly imprinted polymers and immobilized antibodies or comparatively clean matrices such as drinking water, online sample processing is best suited to analyses that use tandem mass spectrometric detection. Indeed, online sample processing can, in general, be considered as an alternative to protein precipitation and liquid–liquid extraction. Consequently, the extracts are crude. However, the beneﬁts of online sample processing are signiﬁcant and numerous and include reduction of matrix effects, an increase in sample throughput, improvement in reproducibility, and consistency of method performance in different matrices [45].

6.3.4 SOLID-PHASE EXTRACTION With few exceptions, the chemical complexity of foods is such that sample preparation is required to separate the analytes from matrix components that may either be a source of interference or compromise the performance of the analytical column. This is particularly the case for trace analyses where large amounts of sample may be required to achieve a low limit of detection or the sample matrix is not compatible with the analytical technique, i.e., it is not in the liquid form in the case of HPLC. Indeed, even in the case of the highly selective and sensitive tandem mass spectrometric detector, sample preparation is highly desirable not only to avoid matrix effects but also to avoid frequent maintenance of the instrument. Traditionally, the extraction and concentration of organic compounds from food were performed using an organic solvent. The procedure, in general, involves partitioning between an organic and aqueous phase. Although the analytes can be selectively extracted through careful choice of solvent,

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 113 15.7.2008 8:19am Compositor Name: BMani

113

Advances in High-Performance Liquid Chromatography

the technique is not without its problems. For instance, solvent extraction may involve the use of (relatively) large volumes of toxic and environmentally unfriendly solvents, emulsions can form, and the technique is less amenable to automation. An alternative to liquid–liquid extraction is solid-phase extraction. In this technique, a solution containing the analytes is passed through a solid sorbent that, through various mechanisms, retains the compounds, which are subsequently eluted from the sorbent in a suitable solvent. The advantages of solid-phase extraction over liquid–liquid extraction are numerous and include convenience, amenable to both sample throughput and automation, and the use of smaller volumes of organic solvents. However, the most important advantage is the opportunity that solid-phase extraction presents to selectively extract analytes from solution. For instance, in the case of nonpolar sorbents, the polarity of the solution containing the analytes can be adjusted so that matrix components more polar than the analytes are not retained by the sorbent. The analytes can then be selectively eluted from more lipophilic extraneous components such as triglycerides and fats using a solvent with an appropriate polarity. This strategy has been used to isolate ochratoxin A from wines. In this example, the polarity of the wine was adjusted with methanol before the application of the sample on a nonpolar solid-phase sorbent [11]. In addition to selectively extracting the analyte, the other beneﬁt of adding methanol to the sample was it dramatically improved the recovery of ochratoxin A from wine. This was attributed to a mass-action effect in that the analyte did not have to compete with more abundant polar matrix components for the limited capacity of the sorbent. A similar strategy has also been used to reﬁne cocoa and coffee extracts for the analysis of acrylamide in these beverages [22]. The extraction of ionizable compounds such as organic acids and amines using ion exchange sorbents and vicinal and 1,3-diols using phenylboronic acid sorbents are further examples where solid-phase sorbents have been used to selectively extract organic compounds from solution. A list of some of the common sorbents and their applications are shown in Table 6.1. TABLE 6.1 Some Common Sorbents Used for Solid-Phase Extraction Sorbent

Retention Mode

C18 and C8

Hydrophobic

Phenyl and divinylbenzene (DVB)

Hydrophobic and p–p interaction

Silica gel

Adsorption

Aminopropyl

Adsorption and weak ionic

SCX

Ionic

SAX

Ionic

Comments Hydrophobic to moderately polar compounds such as pesticides, antibiotics, and mycotoxins. Hydrophobic to polar compounds, particularly those that contain an aromatic ring. The phenyl phase is more polar than both C18 and DVB and enhances the retention of basic compounds. DVB is more hydrophobic than C18. Polar to moderately hydrophobic compounds such as mycotoxins, alcohols, ketones, and pesticides. This sorbent is incompatible with aqueous solutions. Polar normal-phase sorbent that has weak anion exchange properties. This sorbent is compatible with aqueous solutions and can be used to extract organic acids and glycosides. Strong cation exchange resin for the extraction of cations. This is a stronger and more selective mode than both adsorption and hydrophobic. Strong anion exchange sorbent for the extraction of anions. This is a stronger and more selective mode than both adsorption and hydrophobic. (continued )

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 114 15.7.2008 8:19am Compositor Name: BMani

114

Handbook of Food Analysis Instruments

TABLE 6.1 (continued) Some Common Sorbents Used for Solid-Phase Extraction Sorbent

Retention Mode

Florisil

Adsorption

Graphite

Adsorption

Alumina

Adsorption and ionic

Phenylboronic acid

Covalent bonding

Comments A more polar sorbent than silica. Generally used to remove polar compounds from nonpolar organic solvent extracts of environmental samples. Polar sorbent with a high surface area for the extraction of both polar and nonpolar compounds. This sorbent is compatible with aqueous solutions. A polar normal-phase sorbent, which depending on how it has been treated can have ionic properties. Selectively forms a cyclic boronate with vicinal and 1,3-diols, and catechols. The boronates are hydrolyzed under acid conditions.

1.0

0.8

0.8

0.6

0.6

0.4

0.4

Volts

0.2

0.2

OTA

0.0

0.0

1.0 Name

1.0

0.8

0.8

0.6

0.6

0.4 OTA 0.28 µg/L

0.2

Volts

1.0 Name

0.4

Volts

Volts

The use of solid-phase extraction methods that are so-called orthogonal to the analytical separation mode, in general, provides the best results. In other words, the retention mode used to extract the analytes from solution should be as different as possible to the analytical separation mode. Intuitively, this can be rationalized in the following way: the retention mode of the solidphase extraction sorbent exploits a property of the analytes that is very different to the analytical separation mode. Consequently, it is less likely that the chromatographic behavior of the analytes and any extracted extraneous matrix components will be similar. The chromatograms shown in Figure 6.7 demonstrate the effectiveness of this sample preparation strategy. The chromatograms

0.2

0.0

0.0 0

1

2

3

4

5

6

7

8

9

10

Minutes

FIGURE 6.7 Overlay of two chromatograms (top) that were obtained by processing wine using nonpolar solid-phase extraction sorbent. The other (bottom) was obtained by reﬁning one of the extracts using a normalphase sorbent.

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 115 15.7.2008 8:19am Compositor Name: BMani

Advances in High-Performance Liquid Chromatography

115

shown in the top panel were obtained from the analysis of two red wine samples that were extracted using two alternative methods, both of which utilized a hydrophobic solid-phase extraction sorbent to isolate the mycotoxin ochratoxin A (OTA) from the samples. The chromatogram in the bottom panel was obtained by analysis of one of the above extracts after it had been further processed using a normal-phase sorbent. The analytical column for this method had the same retention mode (reverse phase) as the hydrophobic solid-phase extraction sorbent that was initially used to process the wines.

REFERENCES 1. Leandro, C.C., Hancock, P., Fussell, R.J., and Keely, B.J., Comparison of ultra-performance liquid chromatography and high-performance liquid chromatography for the determination of priority pesticides in baby foods by tandem quadrupole mass spectrometry. Journal of Chromatography, A 2006, 1103(1), 94–101. 2 Kovalczuk, T., Jech, M., Poustka, J., and Hajslova, J., Ultra-performance liquid chromatography–tandem mass spectrometry: A novel challenge in multiresidue pesticide analysis in food. Analytica Chimica Acta 2006, 577(1), 8–17. 3. Rontree, S., Ryan, C., Kearney, G., and Winkler, M., Fast UPLC=MS=MS for simultaneous analysis of multiple pesticide residues in agricultural products. Lebensmittel- & Biotechnologie, 2005, 22(3), 93–96. 4. Kaufmann, A., Butcher, P., Maden, K., and Widmer, M., Ultra-performance liquid chromatography coupled to time of ﬂight mass spectrometry (UPLC-TOF): A novel tool for multiresidue screening of veterinary drugs in urine. Analytica Chimica Acta 2007, 586(1–2), 13–21. 5. Cui, X., Shao, B., Zhao, R., Yang, Y., Hu, J., and Tu, X., Simultaneous determination of seventeen glucocorticoids residues in milk and eggs by ultra-performance liquid chromatography=electrospray tandem mass spectrometry. Rapid Communications in Mass Spectrometry 2006, 20(15), 2355–2364. 6. Ren, Y., Zhang, Y., Shao, S., Cai, Z., Feng, L., Pan, H., and Wang, Z., Simultaneous determination of multi-component mycotoxin contaminants in foods and feeds by ultra-performance liquid chromatography tandem mass spectrometry. Journal of Chromatography, A 2007, 1143(1–2), 48–64. 7. Ventura, M., Guillen, D., Anaya, I., Broto-Puig, F., Lliberia, J.L., Agut, M., and Comellas, L., Ultraperformance liquid chromatography=tandem mass spectrometry for the simultaneous analysis of aﬂatoxins B1, G1, B2, G2 and ochratoxin A in beer. Rapid Communications in Mass Spectrometry 2006, 20(21), 3199–3204. 8. Kearney, G.C., Spanjer, M.C., and Romano, J., Multi mycotoxins analysis by UPLC MS MS. Abstracts of Papers, 232nd ACS National Meeting, San Francisco, CA, United States, Sept. 10–14, 2006. 9. Barcelo-Barrachina, E., Moyano, E., Galceran, M.T., Lliberia, J.L., Bago, B., and Cortes, M.A., Ultraperformance liquid chromatography–tandem mass spectrometry for the analysis of heterocyclic amines in food. Journal of Chromatography, A 2006, 1125(2), 195–203. 10. Zhang, Y., Jiao, J., Cai, Z., Zhang, Y., and Ren, Y., An improved method validation for rapid determination of acrylamide in foods by ultra-performance liquid chromatography combined with tandem mass spectrometry. Journal of Chromatography, A 2007, 1142(2), 194–198. 11. Varelis, P., Leong, S.-L.L., Hocking, A., and Giannikopoulos, G., Quantitative analysis of ochratoxin A in wine and beer using solid phase extraction and high performance liquid chromatography–ﬂuorescence detection. Food Additives and Contaminants 2006, 23(12), 1308–1315. 12. Leong, Su-lin L., Hocking, A.D., Varelis, P., Giannikopoulos, G., and Scott, E.S., Fate of ochratoxin A during viniﬁcation of semillon and shiraz Grapes. Journal of Agricultural and Food Chemistry 2006, 54(17), 6460–6464. 13. Soler, C. and Pico, Y., Recent trends in liquid chromatography–tandem mass spectrometry to determine pesticides and their metabolites in food. TrAC, Trends in Analytical Chemistry 2007, 26(2), 103–115. 14. Alder, L., Greulich, K., Kempe, G., and Vieth, B.., Residue analysis of 500 high priority pesticides: Better by GC–MS or LC–MS=MS? Mass Spectrometry Reviews 2006, 25(6), 838–865. 15. Hernandez, F., Sancho, J.V., Pozo, O., Lara, A., and Pitarch, E., Rapid direct determination of pesticides and metabolites in environmental water samples at sub-mg=l level by on-line solid-phase

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 116 15.7.2008 8:19am Compositor Name: BMani

116

16.

17. 18.

19.

20.

21.

22.

23.

24.

25. 26. 27. 28.

29. 30.

31. 32. 33. 34. 35. 36. 37. 38.

Handbook of Food Analysis Instruments extraction–liquid chromatography–electrospray tandem mass spectrometry. Journal of Chromatography, A 2001, 939(1–2), 1–11. Hernandez, F., Sancho, J.V., and Pozo, O.J., Critical review of the application of liquid chromatography= mass spectrometry to the determination of pesticide residues in biological samples. Analytical and Bioanalytical Chemistry 2005, 382(4), 934–946. Henion, J., Brewer, E., and Rule, G., Sample preparation for LC=MS=MS: Analyzing biological and environmental samples. Analytical Chemistry 1998, 70(19), 650A–656A. Fang, L., Pajkovic, N., Wang, Y., Gu, C., and van Breemen Richard, B., Quantitative analysis of lycopene isomers in human plasma using high-performance liquid chromatography–tandem mass spectrometry. Analytical chemistry 2003, 75(4), 812–817. King, R., Bonﬁglio, R., Fernandez-Metzler, C., Miller-Stein, C., and Olah, T., Mechanistic investigation of ionization suppression in electrospray ionization. Journal of the American Society for Mass Spectrometry 2000, 11(11), 942–950. Matuszewski, B.K., Constanzer, M.L., and Chavez-Eng, C.M., Strategies for the assessment of matrix effect in quantitative bioanalytical methods based on HPLC-MS=MS. Analytical Chemistry 2003, 75(13), 3019–3030. Jemal, M., Schuster, A., and Whigan, D.B., Liquid chromatography=tandem mass spectrometry methods for quantitation of mevalonic acid in human plasma and urine: Method validation, demonstration of using a surrogate analyte, and demonstration of unacceptable matrix effect in spite of use of a stable isotope analog internal standard. Rapid Communications in Mass Spectrometry 2003, 17(15), 1723–1734. Aguas, P.C., Fitzhenry, M.J., Giannikopoulos, G., and Varelis, P., Analysis of acrylamide in coffee and cocoa by isotope dilution liquid chromatography–tandem mass spectrometry. Analytical and Bioanalytical Chemistry 2006, 385(8), 1526–1531. Zollner, P., Leitner, A., Berner, D., Kleinova, M., Jodlbauer, J., Mayer, B.X., and Lindner, W., Improving LC-MS=MS analyses in complex food matrices, Part I—sample preparation and chromatography. LC-GC Europe 2003, 16(3), 163–171. Bonﬁglio, R., King, R.C., Olah, T.V., and Merkle, K., The effects of sample preparation methods on the variability of the electrospray ionization response for model drug compounds. Rapid Communications in Mass Spectrometry 1999, 13(12), 1175–1185. Candlish, A.A., Smith, J.E., and Stimson, W.H., Aﬂatoxin monoclonals: Academic development to commercial production. Letters in Applied Microbiology 1990, 10(4), 167–169. Scott, P.M. and Trucksess, M.W., Application of immunoafﬁnity columns to mycotoxin analysis. Journal of AOAC International 1997, 80(5), 941–949. Gilbert, J., Recent advances in analytical methods for mycotoxins. Food Additives and Contaminants 1993, 10(1), 37–48. Sanchez, F.G., Diaz, A.N., Herrera, R.G., and San Jose, L.P., Development and characterization of an immunoafﬁnity chromatographic column for the on-line determination of the pesticide triclopyr. Talanta 2007, 71(3), 1411–1416. Nunes, G.S. and Barcelo, D., Analysis of carbamate insecticides in foodstuffs using chromatography and immunoassay techniques. TrAC, Trends in Analytical Chemistry 1999, 18(2), 99–107. Zhang, S., Zhou, J., Shen, J., Ding, S., and Li, J., Determination of chloramphenicol residue in chicken tissues by immunoafﬁnity chromatography cleanup and gas chromatography with a microcell electron capture detector. Journal of AOAC International 2006, 89(2), 369–373. Stidl, R. and Cichna-Markl, M., Sample clean-up by sol-gel immunoafﬁnity chromatography for determination of chloramphenicol in shrimp. Journal of Sol-Gel Science and Technology 2007, 41(2), 175–183. Subramannian, A., Immunoafﬁnity chromatography. Molecular Biotechnology 2002, 20, 41–47. Analytix, 2007, issue 1 (available at: Analytix). Ensing, K. and De Boer, T., Trends in Analytical Chemistry 1999, 18, 138–145. Ensing, K., Berggren, C., and Majors, R.E., Selective sorbents for solid-phase extraction based on molecularly imprinted polymers. LC-GC Europe 2002, January, 2. http:==www.miptechnologies.se=. Kearney, G. and Bernsmann, T., A method for the rapid and sensitive detection of ochratoxin A in red wines. Application Note, Waters Corporation, Library Number: 720000792EN. Kataoka, H., Automated sample preparation using in-tube solid-phase microextraction and its application— a review. Analytical and Bioanalytical Chemistry 2002, 373(1–2), 31–45.

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 117 15.7.2008 8:19am Compositor Name: BMani

Advances in High-Performance Liquid Chromatography

117

39. Chafer-Pericas, C., Herraez-Hernandez, R., and Campins-Falco, P., In-tube solid-phase microextraction– capillary liquid chromatography as a solution for the screening analysis of organophosphorus pesticides in untreated environmental water samples. Journal of Chromatography, A 2007, 1141(1), 10–21. 40. Wen, Y., Wang, Y., and Feng, Y.-Q., Simultaneous residue monitoring of four tetracycline antibiotics in ﬁsh muscle by in-tube solid-phase microextraction coupled with high-performance liquid chromatography. Talanta 2006, 70(1), 153–159. 41. Huang, J.-F., Lin, B., Yu, Q.-W., and Feng, Yu-Qi., Determination of ﬂuoroquinolones in eggs using intube solid-phase microextraction coupled to high-performance liquid chromatography. Analytical and Bioanalytical Chemistry 2006, 384(5), 1228–1235. 42. Mitani, K., Narimatsu, S., and Kataoka, H., Determination of daidzein and genistein in soybean foods by automated on-line in-tube solid-phase microextraction coupled to high-performance liquid chromatography. Journal of Chromatography, A 2003, 986(2), 169–177. 43. Kataoka, H., Ise, M., and Narimatsu, S., Automated on-line in-tube solid-phase microextraction coupled with high performance liquid chromatography for the analysis of bisphenol A, alkylphenols, and phthalate esters in foods contacted with plastics. Journal of Separation Science 2002, 25(1=2), 77–85. 44. Wu, J., Xie, W., and Pawliszyn, J., Automated in-tube solid-phase microextraction coupled with HPLCES-MS for the determination of catechins and caffeine in tea. Analyst (Cambridge, United Kingdom) 2000, 125(12), 2216–2222. 45. Guegel, A. and Herman, J.L., Turbo ﬂow chromatography (TFC): Online sample preparation for LC-MS=MS. GIT Labor-Fachzeitschrift 2005, 49(10), 862–864.

Otles/Handbook of Food Analysis Instruments 45660_C006 Final Proof page 118 15.7.2008 8:19am Compositor Name: BMani

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 119 17.7.2008 4:04pm Compositor Name: VAmoudavally

Chromatography 7 Gas in Food Analysis Jana Hajslova and Tomas Cajka CONTENTS 7.1 7.2

Introduction .......................................................................................................................... 119 Sample Introduction ............................................................................................................. 120 7.2.1 Split=Splitless Injection............................................................................................ 120 7.2.2 Cold On-Column Injection ...................................................................................... 122 7.2.3 Programmable Temperature Vaporization Injection................................................ 123 7.2.4 Direct Sample Introduction=Difﬁcult Matrix Introduction ...................................... 125 7.2.5 Solid-Phase Microextraction.................................................................................... 125 7.3 Sample Separation................................................................................................................ 126 7.3.1 Capillary Columns for GC....................................................................................... 126 7.3.2 Fast Gas Chromatography ....................................................................................... 126 7.3.3 Comprehensive Two-Dimensional Gas Chromatography....................................... 130 7.3.3.1 GC GC Setup ........................................................................................ 131 7.3.3.2 Optimization of Operation Conditions and Instrumental Requirements in GC GC ....................................................................... 131 7.3.3.3 Advantages of GC GC .......................................................................... 133 7.4 Sample Detection ................................................................................................................. 134 7.4.1 Flame Ionization Detector ....................................................................................... 135 7.4.2 Thermal Conductivity Detector ............................................................................... 136 7.4.3 Electron Capture Detector ....................................................................................... 136 7.4.4 Nitrogen–Phosphorus Detector ................................................................................ 136 7.4.5 Flame Photometric Detector and Pulsed Flame Photometric Detector ................... 136 7.4.6 Photo-Ionization Detector ........................................................................................ 136 7.4.7 Electrolytic Conductivity Detector .......................................................................... 136 7.4.8 Atomic-Emission Detector....................................................................................... 136 7.4.9 Mass Spectrometric Detector................................................................................... 137 7.5 Matrix Effects ...................................................................................................................... 137 7.6 Food Analysis Applications................................................................................................. 140 7.7 Conclusion and Future Trends............................................................................................. 142 Acknowledgments......................................................................................................................... 142 References ..................................................................................................................................... 142

7.1 INTRODUCTION In food analysis, gas chromatography (GC) represents one of the key separation techniques for many groups of (semi)volatile compounds. The high separation power of GC in a combination with a wide range of the detectors makes GC an important tool in the determination of various components that may occur in such complex matrices as food crops and products. 119

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 120 17.7.2008 4:04pm Compositor Name: VAmoudavally

120

Handbook of Food Analysis Instruments

Sample preparation

Data analysis

Very fast GC

Cold on-column

Heart-cut GC Comprehensive twodimensional GC

2D-GC

Pulsed splitless

Ultra-fast GC

Programmable temperature vaporiser Direct sample introduction/ Difficult matrix introduction Solid-phase microextraction

Flame ionisation Thermal conductivity

Fast GC Conventional

Classical splitless

Detection

Conventional GC 1D-GC

Split

Separation

Electron capture Nitrogen–phosphorus (Pulsed) flame photometric Photo-ionisation Electrolytic conductivity Atomic-emission

Mass spec.

Advanced

Conventional

Sample introduction

Quadrupole Quadrupole ion trap Magnetic sector Time-of-flight Hybrid instruments

FIGURE 7.1 Basic steps typically involved in the determinative step of gas chromatographic analysis of organic food compounds.

In practice, a GC-based method consists typically of the following steps: (1) isolation of analytes from a representative sample (extraction); (2) separation of co-extracted matrix components (cleanup); (3) identiﬁcation and quantiﬁcation of target analytes (determinative step), and if the need is important enough, this is followed by (4) conﬁrmation of results by an additional analysis (Figure 7.1). In any case, the sample preparation practice plays a crucial role for obtaining required parameters of a particular analytical method. Under some conditions, especially when polar analytes are to be analyzed, derivatization is carried out prior to the GC step to avoid hydrogen bonding, hence increasing the analyte volatility and reducing interaction with active sites in the system. In Figure 7.2, the interrelationship between solute amounts and instrumental options (inlet, column, and detector) is illustrated. GC users should examine the relationship of analyzed samples to the operating range of the instrument system. If the analyte concentration lies outside this range, a different injection technique, column dimension, or detector may be appropriate.

7.2 SAMPLE INTRODUCTION There are a number of options available for GC inlet systems; the most common (characterized below) being split=splitless, programmed temperature vaporizer, and cold on-column (COC) injector. The choice of an optimum sample introduction strategy depends mainly on the concentration range of target analytes, their physico-chemical properties, and the amount and nature of matrix co-extracts present in the sample.

7.2.1 SPLIT=SPLITLESS INJECTION Split=splitless injection remains the main sample introduction technique in the analysis of GC-amenable food components mainly due to its easy operations.

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 121 17.7.2008 4:04pm Compositor Name: VAmoudavally

121

Column diameter and film thickness

Splitless, direct, on-column Split

Mass and concentration

Gas Chromatography in Food Analysis Femtograms parts per trillion 1

10

100

Picograms parts per billion 1

10

100

1

10

100

Percentage:

1

10

100

1000

0.1%

1%

10%

100%

10–6

10–5

10–4

10–3

10–6

10–5

10–4

10–3

10–5

10–4

10–3

Solute mass (g) 10–15 10–14

10–13

10–12

10–11

10–10

10–9

10–8

10–7

Split 100:1 10–15 10–14

10–13

10–12

10–11

10–10

10–9

10–8

10–7

Splitless dc

df

530 µm

0.1 µm 0.1 µm

320 µm 250 µm

0.1 µm 0.1 µm

180 µm

0.1 µm

100 µm Detector minimum detection limit and dynamic range

Micrograms parts per thousand

Nanograms parts per million

5.0 µm 5.0 µm

1.0 µm 0.5 µm

0.5 µm Thermal conductivity

Flame ionization Nitrogen–phosphorus Electron–capture MS (scan) MS (single-ion monitoring) 10–15 10–14

10–13

10–12

10–11

10–10 10–9 10–8 Solute mass (g)

10–7

10–6

FIGURE 7.2 GC dynamic range nomogram. Concentrations expressed in grams per microliter (g=mL). (Reproduced from Hinshaw, J.V., LC GC Eur., 20, 138, 2007. With permission.)

In a split injection mode, typically small volume of sample extract (0.1–2 mL) is rapidly delivered into a heated glass liner followed by its splitting into two streams: the larger part is vented, while the smaller part is transferred onto the column. Considering that the most of injected sample is lost, this technique is obviously not suitable for trace analysis, where very low detection limits are required. Another problem associated with split injection is a potential discrimination due to the heating of the syringe during its introduction into a hot injector resulting in a change of relative abundances of sample components when a mixture of analytes largely differing in boiling points is analyzed. Another adverse phenomenon related to this technique is nonlinear splitting due to adsorption of sample components on liner surfaces or deposited matrix ‘‘dirt.’’ Nowadays, hot splitless injection represents the most commonly used injection technique in trace quantitative analysis since entire injected sample is introduced onto the GC capillary. The major limitation of this inlet is that it suffers from the potential thermal degradation and adsorption of susceptible analytes that may result either in matrix-induced response enhancement or its diminishment. In addition, the volume of injected sample=solvent is typically limited to 1 mL (for some solvents even less) due to the expansion volume of solvent used (Table 7.1), since the total liner volume is in a range of 150–1000 mL and the safety limit is typically 75% in maximum of the total liner volume.

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 122 17.7.2008 4:04pm Compositor Name: VAmoudavally

122

Handbook of Food Analysis Instruments

TABLE 7.1 Expansion Volume of Solvents Used in GC Expansion Volume (mL) Solvent Water Methanol Acetonitrile Acetone Ethyl acetate Toluene Hexane Isooctane

1 mL at 2508C and 69 kPa (10 psig)

1 mL at 2508C and 345 kPa (50 psig)

5 mL at 2508C and 345 kPa (50 psig)

1414 631 487 347 261 241 195 155

540 241 186 133 100 92 75 59

2700 1205 929 663 498 460 373 295

Source: From Hewlett-Packard FlowCalc 2.0 software. Available at http:==www.chem.agilent. com=cag=servsup=usersoft=ﬁles=GCFC.htm via the Internet. Accessed July 1, 2007. Note: Calculated using Hewlett-Packard FlowCalc 2.0 software.

To overcome, or at least partly compensate for these problems, pulsed splitless injection can be applied. Increased column head pressure for a short period during the sample injection (usually 1–2 min) leads to a higher carrier gas ﬂow rate through the injector (8–9 versus 0.5–1 mL=min during classical splitless injection), thus faster transport of sample vapors onto the GC column. In this way, the residence time of analytes and, consequently, their interaction with active sites in the GC inlet is fairly reduced [3]. The detection limits of troublesome compounds obtained with pulsed splitless injection are thus lower and their further improvement can be obtained by injection of higher sample volumes (for most liners up to 5 mL) without the risk of backﬂash (Table 7.1) [4]. It should be noted that for injections >1–2 mL, a retention gap prior to the analytical column is generally required to avoid excessive contamination of separation column and peak distortion (Figure 7.3).

7.2.2 COLD ON-COLUMN INJECTION In COC injection, a sample aliquot is directly introduced by a special syringe onto the analytical column or a retention gap at temperatures lower (608C–808C) than those typically used in hot

Pulsed splitless

Pulsed splitless

3 µL

4 µL

2 µL 5 µL

1 µL

3 µL 2 µL

Splitless 1 µL

(A)

1 µL (B)

FIGURE 7.3 Peak shapes obtained by pulsed splitless injections of different volumes of standard solution onto the GC column (A) without a retention gap; (B) with an installed retention gap. (Reproduced from Godula, M., Hajslova, J., and Alterova, K., J. High Resolut. Chromatogr., 22, 395, 1999. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 123 17.7.2008 4:04pm Compositor Name: VAmoudavally

Gas Chromatography in Food Analysis

123

split=splitless mode (2008C–3008C). COC is therefore expected to cause less thermal stress on analytes during the injection process. This low-temperature injection eliminates both syringe needle and inlet discrimination and is suitable namely for high-boiling analytes. On the other hand, the introduction of the entire sample (both analytes and matrix components co-isolated from food matrix) into the GC system is associated with increased demands for cleaning and maintenance when such complex samples as food is analyzed [5]. There are two alternative approaches available to perform on-column injection. 1. Small volume on-column injection: In this approach, a small volume of the sample (up to 1–2 mL) is injected onto the separation column, or preferably, in the case of dirty samples, onto a retention gap [6]. 2. Large volume on-column injection: In this mode, a large volume of the sample (up to 1000 mL) can be introduced into the GC system [7]. The bulk of solvent is usually eliminated via a special solvent vapor exit. Once the venting is ﬁnished, the solvent vapor exit is closed and analytes, together with remaining traces of solvent, are transferred onto the analytical column. However, a modiﬁcation of the GC system is required in this case to include a large diameter retention gap (10–15 m 0.53 mm), connected to a retaining precolumn (3–5 m 0.32 mm) assisting in retention of volatile analytes. In addition, injection speed has to be slowed down to prevent ﬂooding of the system during large volume injections (LVI) with injection speeds in LVI-COC of 20–300 mL=min as compared to LVI-PTV at 50–1500 mL=min [8].

7.2.3 PROGRAMMABLE TEMPERATURE VAPORIZATION INJECTION A programmable temperature vaporization (PTV) injector represents the most versatile GC inlet offering signiﬁcant reduction of most problems typically present when using hot vaporizing devices (splitless and=or cool on-column inlets) in trace analysis. The most important fact is that a PTV injector chamber is cool at the moment of injection. A rapid temperature increase, following withdrawal of the syringe from the inlet, allows efﬁcient transfer of the volatile analytes onto the GC column while leaving behind nonvolatiles in the injection liner. With regard to these operational features, PTV is ideally suited for thermally labile analytes and samples with a wide boiling range (when needed, PTV operating temperature can be programmed even higher than the usual column temperature allowing injection of analytes that would not be vaporized through a classic split=splitless inlet). In addition eliminating a discrimination phenomenon and diminishing adverse affects of nonvolatile matrix deposits on the recovery of injected analytes, PTV enables to introduce large sample volumes (up to hundreds of microliters) into the GC system. No retention gaps or precolumns are needed for this purpose; instead of that, the liner size is increased. This feature makes PTV particularly suitable for trace analysis and also enables its online coupling with various enrichment and cleanup techniques such as automated solid-phase extraction (SPE) approaches. From practical point of view, PTV is compatible with any capillary GC column diameter including microbore columns. However, to attain optimal PTV performance in particular application, many parameters have to be optimized (e.g., initial and ﬁnal injector temperature, inlet heating rate, venting time, ﬂow and pressure, transfer time, injection volume, type of liner). Due to the inherent complexity of this inlet, method development might become on some occasions a rather demanding task. Despite that, the use of PTV in food analysis is rapidly growing. The paragraphs below describe two most commonly used PTV operation modes. 1. PTV splitless injection. The sample is introduced at a temperature below or close to the boiling point of the solvent. A split exit is closed during the sample evaporation and solvent vapors are vented via a GC column. PTV splitless injection can be employed for both LVI of up to 20 mL of sample and for conventional small volume injections [9]. The

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 124 17.7.2008 4:04pm Compositor Name: VAmoudavally

124

Handbook of Food Analysis Instruments

(A) 10

12.5

15

17.5

20

22.5

Phosalone 31.943

25.803 24.749

20.906 Carbaryl 21.156 21.656 22.185 22.592

19.182

Dimethoate 17.742

Omethoate 15.750

Acephate 13.975 14.194

12.255

pA 1400 1200 1000 800 600 400 200 0

Methamidophos 12.328

advantage of this technique is that no losses of volatile analytes occur. Operating parameters have to be carefully optimized to avoid inlet overﬂow by sample vapors (losses of volatile compounds) as well as column ﬂooding by excessive solvent (poor peak shapes of more volatile analytes). It has been reported that for some analytes the PTV splitless injection may produce better stability of responses and less matrix inﬂuence [10]. 2. PTV solvent vent injection. When employing this technique, a sample is injected at temperatures well below the boiling point of the solvent, holding the temperature of the injection port at low a value, thus enabling elimination of solvent vapors via a split exit. After the venting step, the inlet is rapidly heated and analytes are transferred onto the front part of a GC column. In this way, sample volumes of up to hundreds of microliters can be injected [11,12]. For injection of large volumes, injector liner is often packed with various sorbents (e.g., Tennax, polyimide, Chromosorb, glass wool, glass beads, PTFE, Dexsil) in order to protect solvent from reaching bottom of injector what may lead to column ﬂooding with liquid sample [13]. However, some labile compounds can be prone to degradation= rearrangement due to the catalytic effects of the sorbent; alternatively, strong binding onto the packing material causes poor desorption (Figure 7.4). If selection of a suitable inactive sorbent fails to prevent these adverse effects, then the only viable solution is the use of an empty or open liner. Under these conditions, rather smaller volumes (in maximum about 50 mL) of sample can be injected, typically employing a concept of multiple injections to get a larger injection volume. Obtaining good performance of the PTV injector in solvent vent mode requires thorough experimental optimization of all relevant parameters as described above.

25

27.5

30

27.5

30

32.5

min

15

17.5

20

22.5

25

31.950

25.801

22.600

21.664

20.915

19.188

12.5

17.739

(B) 10

14.205

1400 1200 1000 800 600 400 200 0

12.041

pA

32.5

min

FIGURE 7.4 Chromatograms obtained by programmable temperature vaporization (PTV) injection into (A) empty liner and (B) liner packed with glass wool plug. The differences in responses of sensitive analytes when injections were carried out into empty multibafﬂe liner and into single-bafﬂe liner packed with glass wool. (Reproduced from Godula, M. et al., J. Sep. Sci., 24, 355, 2001. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 125 17.7.2008 4:04pm Compositor Name: VAmoudavally

125

Gas Chromatography in Food Analysis

7.2.4 DIRECT SAMPLE INTRODUCTION=DIFFICULT MATRIX INTRODUCTION Direct sample introduction (DSI) or its fully automated version, difﬁcult matrix introduction (DMI), represents a novel LVI technique. The DSI approach involves adding up to 30 mL of the extract to a microvial that is placed in the adapted GC liner. The solvent is evaporated and vented at a relatively low temperature. After that, the injector is ballistically heated to volatilize the GC-amenable compounds, which are then focused at the front of a relatively cold GC column. The column then undergoes normal temperature programming to separate the analytes and cool to initial conditions, at which time the microvial is removed and discarded along with the nonvolatile matrix components that it contains. Only those compounds with the volatility range of the analytes enter the column [14]. In the commercial DMI approach, the entire liner along with the microvial is replaced after each injection [15]. In this way, time-consuming and expensive puriﬁcation steps can be omitted=signiﬁcantly reduced for some matrices [15,16]. Since the bulk (semi)volatile matrix components introduced from the sample into the injector may inﬂuence the quantitative aspects of the injection process and interfere in analytes detection, instruments with MS analyzers (single or tandem) providing more accurate results should be preferably used [17]. In Figure 7.5 the distinct improvement obtained by sample cleanup is illustrated. Regardless sample preparation strategy, reduced demands for the GC system maintenance represents a positive feature of this technique.

7.2.5 SOLID-PHASE MICROEXTRACTION Solid-phase microextraction (SPME) represents a solvent-free sampling technique employing a fused-silica ﬁbre that is coated on the outside with an appropriate stationary phase. Volatile analytes emitted from the analyzed sample are isolated from the headspace or by direct immersion into the liquid sample and concentrated in ﬁbre coating. After the extraction, thermal desorption in the hot GC injection port follows [18]. The main features of SPME include unattended operation via robotics (if a fully automated option is available) and the elimination of maintenance of the liner and column. However, this sample introduction technique is associated with strong matrix effects,

Metalcap (septum inside) O-rings 1

2 Needle guide

Microvial

Liner 3

FIGURE 7.5 Difﬁcult matrix introduction (DMI) liner after injection of (1) puriﬁed baby-food extract, (2) crude extract, and (3) detail of microvials used for introduction of crude extracts. (Reproduced from Cajka, T. et al., J. Sep. Sci., 28, 1048, 2005. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 126 17.7.2008 4:04pm Compositor Name: VAmoudavally

126

Handbook of Food Analysis Instruments

thus complications in quantiﬁcation. In addition, variability of limit of detections for different analytes depends on the equilibrium between the coating material and the matrix.

7.3 SAMPLE SEPARATION To be amenable for the GC analysis, an analyte should possess not only appreciable volatility at temperatures below 3508C–4008C, but also must be able to withstand relative high temperatures without degradation and reaction with other compounds present in the GC system. With regard to a typically complex mixture of matrix components occurring in food extracts (often even after its puriﬁcation), the optimization of GC separation requires careful attention to a number of important variables and their interaction. Both physical (column length, internal diameter, and stationary phase including its ﬁlm thickness), and parametric (temperature and ﬂow velocity) column variables affect the separation process.

7.3.1 CAPILLARY COLUMNS

FOR

GC

To illustrate a wide range of combinations to be considered when selecting capillary GC column, the overview of commonly available internal diameters is shown in Table 7.2. Table 7.3 shows relative polarities of commercially available stationary phases. The range of stationary phases including also those dedicated for speciﬁc applications (e.g., volatiles, fatty acids, dioxins) is growing and also their quality characterized by reduced bleed and increased upper temperature limit is improving. In addition, the limits of inlet pressure, sampling system, and mass spectrometric detector (MSD) parameters have to be involved into consideration.

7.3.2 FAST GAS CHROMATOGRAPHY A higher sample throughput together with the need to reduce laboratory operating costs has brought attention of many laboratories to the implementation of high-speed GC (HSGC) systems. Although the basic principles and theory of HSGC were formulated as early as in the 1960s, its practical development occurred at the end of last century from introduction of novel technologies such as new methods of fast and reproducible column heating, inlet devices allowing large sample volume introduction, and MS detectors with fast acquisition rates. It should be noted that full exploitation of the potential of this technique in routine practice is conditioned by reduction of sample TABLE 7.2 Classiﬁcation of Capillary Column Category Megabore Wide bore Narrow bore Microbore Sub-microbore

Column Diameter Range (mm)

Standard Commercial Column Diameters (mm)

Max Flow Rate (mL=min)a

0.5 0.3 to <0.5 0.2 to <0.3 0.1 to <0.2 <0.1

0.53 0.32, 0.45 0.20, 0.25, 0.28 0.10, 0.15, 0.18 Various

660 85 to <660 17 to <86 1 to <17 <1

Source: Reproduced from Mastovska, K. and Lehotay, S.J., J. Chromatogr. A, 1000, 153, 2003. With permission. a

Flow rate calculated using helium carrier gas at 690 kPa, 2008C oven, vacuum outlet conditions, and 10 m column length.

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 127 17.7.2008 4:04pm Compositor Name: VAmoudavally

127

Gas Chromatography in Food Analysis

TABLE 7.3 Characterization of Stationary Phases Used in GC Analysis Polarity Nonpolar

Phase Composition

Commercial Description

100% Dimethylpolysiloxane

5% Diphenyl–95% dimethylpolysiloxane

20% Diphenyl–80% dimethylpolysiloxane 6% Cyanopropyl-phenyl–94% dimethylpolysiloxane 35% Diphenyl–65% dimethylpolysiloxane

Moderately polar

50% Diphenyl–50% dimethylpolysiloxane

Polar

14% Cyanopropyl-phenyl–86% dimethylpolysiloxane 50% Cyanopropyl-phenyl–50% dimethylpolysiloxane Polyethylene glycol

Highly polar

70% Cyanopropyl-phenyl–30% dimethylpolysiloxane 100% Cyanopropylsiloxane

DB-1, DB-1 ms, HP-1, HP-1 ms, Ultra 1, DB-1ht, Equity-1, SPB-1, AT-1, AT-1MS, Optima 1, Optima1 ms, BP-1, VF-1MS, CP Sil 5 CB, CP Sil 5 CB MS, ZB-1, 007-1, Elite-1, Rxi-1 ms, Rtx-1, Rtx-1MS DB-5, HP-5, DB-5 ms, HP-5 ms, Ultra 2, DB-5ht, Equity-5, SPB-5, AT-5, AT-5MS, Optima 5, Optima-5 ms, BP-5, BPX-5, VF-5MS, CP Sil 8 CB, CP Sil 8 CB MS, ZB-5, 007-2, PE-2, Rxi-5 ms, Rtx-5, Rtx-5 ms, Rtx-5Sil MS Rtx-20, SPB-20, At-20, 007-7 DB-1301, HP-1301, Rtx-1301, SPB-1301, AT-624, Optima 1301, 007-1301 DB-35, HP-35, DB-35 ms, Rtx-35, Rtx-35MS, SPB-35, AT-35, AT-35MS, BPX-35, VF-35MS, ZB-35, 007-11, PE-11 DB-17, DB-17 ms, HP-50 þ , DB-17ht, Rtx-17, VF-17MS, SPB-50, AT-50, AT-50MS, Optima 17, BPX-50, CP Sil 24 CB, ZB-50, 007-17, PE-17 DB-1701, HP-1701, SPB-1701, AT-1701, Optima 1701, BP-10, CP Sil 19 CB, ZB-1701, 007-1701, PE-1701 DB-23, DB-225, DB-225 ms, Rtx-225, AT-225, Optima 225, BP-225, CP Sil 43 CB, 007-225, PE-225 DB-WAX, HP-INNOWax, Rtx-WAX, Stabilwax, Supelcowax-10, AT-Wax, Optima WAX, BP-20, CP Wax 52 CB, ZB-WAX, 007-CW, PE-CW BPX-70 SP-2340

preparation time and other operations limiting laboratory throughput. Using approximate terms, the classiﬁcation of GC analyses on the basis of their speed is summarized in Table 7.4. Alike in conventional GC, the separation time is deﬁned as the retention time (tR) for the last target component peak eluting from the column:

TABLE 7.4 Classiﬁcation of GC Analyses Based on Speed of Sample Separation Type of GC Analysis

Typical Separation Time (min)

Full Width at Half-Maximum

Conventional Fast Very fast Ultra-fast

>10 1–10 0.1–1 <0.1

>1 s 200–1000 ms 30–200 ms 5–30 ms

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 128 17.7.2008 4:04pm Compositor Name: VAmoudavally

128

Handbook of Food Analysis Instruments

L tR ¼ (k þ 1) u

(7:1)

where k is the solute capacity ratio (capacity factor, retention factor) for the last compound L is the column length u is the average linear carrier gas velocity On the basis of this equation, the faster GC separation can be achieved by following ways [19,20]: . .

.

Reduced column length (# L) Decreased retention factor (# k): (1) increased isothermal temperature, (2) faster temperature programming, (3) altered stationary phase to improve selectivity, (4) thinner ﬁlm of the stationary phase, (5) larger diameter capillary column (for ﬁxed length) Increased carrier gas velocity (" u): (1) higher than optimum carrier gas velocity, (2) increased optimum carrier gas velocity (hydrogen as a carrier gas and vacuum outlet operation)

The increase in separation speed generally requires a compromise in terms of reduced resolution (R) and=or sample capacity (Qc). The acceptability of these losses has to be considered for each particular case separately. Availability of compatible sample introduction technique and detection parameters play an important role in selection of the analytical strategy. Reduction of column length represents a very simple approach to decrease time of GC analysis. In practice, almost all fast GC analyses are performed with short columns (usually 10 m) in a combination with other approaches (Figure 7.6). p On this account, reduction of the length of a given column results in reduced resolution (R ~ L), which can be compensated to some extent by suitable MS detector (spectral resolution). Use of a column with a small internal diameter is another attractive way towards faster GC analysis. However, the instrumental requirements especially the difﬁculties with the sample introduction of larger sample volumes and also the lower sample capacity limit their application in many real-world analyses. Use of a column with a thin ﬁlm of stationary phase results in the decrease of the capacity (retention) factor and thus in the faster GC analysis. In addition, due to the decreased contribution of mass transfer in the stationary phase, separation efﬁciency is increased. On the other hand, reduced ruggedness and sample capacity are the fees for analysis speed. Fast temperature programming is the most popular approach in application of fast GC in food analysis. Either convection heating facilitated by a conventional GC oven or resistive heating can be employed. If ‘‘fast’’ separation in terms of classiﬁcation shown in Table 7.4 is required, a conventional GC oven can be used. At faster programming rates, heat losses from the oven to the surrounding may cause poor oven temperature proﬁle, hence lower reproducibility of analyte elution. Operation of column outlet at low pressure (low-pressure GC) is another fast GC alternative that may ﬁnd a wide use in routine laboratories concerned with food analysis. Because of operating a megabore separation column (typically 10 m length 0.53 mm internal diameter 0.25–1 mm phase) at low pressure, optimum carrier gas linear velocity is attained at higher value because of increased diffusivity of the solute in the gas phase. Consequently, faster GC separations can be achieved with a disproportionately smaller loss of separation power [22]. The main attractive features of LP-GC–MS involve (1) reduced peak tailing and width (Figure 7.7) thus their improved detection limits, (2) increased sample capacity of megabore column allowing injection of higher sample volume resulting in lower detection limits for compounds not limited by matrix interferences, and (3) reduced thermal degradation of thermally labile analytes [23,24].

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 129 17.7.2008 4:04pm Compositor Name: VAmoudavally

129

Gas Chromatography in Food Analysis

50000

40000

Intensity

20 28

30000

20000

14

27

10000 1

(A)

6

2

0

3

8

4

5

10

9 11 13 10 12

30 16 19 22 25 26 29 23 31 18 21 15

7

20

30

40

50

min

80000 20 28

Intensity

60000

40000 14

27

20000 1

(B)

0.0

2 0.5

6 4

8

9 11 10 12 13

7 1.0

1.5

24 22 16 25 23 21 15 18 19 2.0

30 29 31 min

FIGURE 7.6 GC–FID chromatograms of fatty acid methyl esters obtained under conditions of (A) conventional (column: Rtx-WAX, 30 m 0.25 mm 0.25 mm; injection: split 1:100; oven temperature program: 508C, 38C=min to 2808C; acquisition rate: 12.5 Hz) and (B) fast GC (column: Supelcowax, 10 m 0.10 mm 0.10 mm; injection: split 1:200; oven temperature program: 508C, 808C=min to 1508C, 708C=min to 2508C, 508C=min to 2808C (1 min); acquisition rate: 50 Hz). (Reproduced from Mondello, L. et al., J. Chromatogr A., 1035, 237, 2004. With permission.)

Hydrogen can be used as a carrier gas because with the highest diffusion coefﬁcient it is obviously the best carrier gas for fast GC. Its low viscosity also results in lower inlet pressure requirements. In practice, however, helium is usually preferred as a carrier gas ﬂow for safety and inertness reasons.

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 130 17.7.2008 4:04pm Compositor Name: VAmoudavally

130

Handbook of Food Analysis Instruments Abundance

Abundance

4000 3000

m /z 201 3000

m /z 201

2000 2000 1000

1000

0 Time−> 9.50 Abundance

0 10.50 min Time−> 3.00 Abundance

4.00 min

4000 3000

m /z 283

m /z 283 3000

2000

2000

1000 0 Time−> 9.50 (A) Conventional GC–MS

1000

10.50 min

0 Time−> 3.00

4.00 min

(B) LP-GC–MS

FIGURE 7.7 Comparison of peak shapes of thiabendazole (m=z 201) and procymidone (m=z 283) obtained by (A) conventional GC–MS (column: Rtx-5MS, 30 m 0.25 mm 0.25 mm; injection: splitless, 1 mL; oven temperature program: 908C (0.5 min), 208C=min to 2208C, 58C=min to 2408C, 208C=min to 2908C (6.5 min)) and (B) LP-GC–MS (column: Rtx-5Sil MS, 10 m 0.53 mm 1.0 mm coupled to 3 m 0.15 mm restriction column; injection: splitless, 1 mL; oven temperature program: 908C (0.5 min), 608C=min to 2908C (3.0 min)). (Reproduced from Mastovska, K., Lehotay, S.J., Hajslova, J., J. Chromatogr. A, 926, 291, 2001. With permission.)

7.3.3 COMPREHENSIVE TWO-DIMENSIONAL GAS CHROMATOGRAPHY In the analysis of complex mixtures, such as food extracts, by one-dimensional chromatography (1D-GC), overlap of some sample components unavoidably occurs. To achieve a considerable increase in peak capacity, two independent separation processes with peak capacities n1 and n2 can be employed in the sample analysis. Supposing that separations are based on two different mechanisms (orthogonality criterion), the maximum peak capacity calculated as n1 n2 is typically enhanced by at least one order of magnitude. Most of the successful applications reported in food analysis since 1960 up to 1990 employed so-called heart-cut mode, in which only narrow fraction(s) containing analytes of interest is (are) transported for further separation onto the second column. However, this approach has limitations. Increasing the width of the ﬁrst column fraction or isolating too many parts of 1D-GC analysis to subject them to two-dimensional gas chromatography (2D-GC) separation becomes troublesome. Also, time-demanding reconstruction of generated chromatograms may become a serious problem. The introduction of systems that allow the entire sample from the ﬁrst column to be analyzed on the second column has enabled and improved both target and nontarget screening of food components in a wide range of matrices. This approach, called comprehensive two-dimensional (GCGC), is introduced in the following sections in a greater detail.

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 131 17.7.2008 4:04pm Compositor Name: VAmoudavally

131

Gas Chromatography in Food Analysis

7.3.3.1

GCGC Setup

The heart of the GCGC system is a modulator that connects the ﬁrst-dimension conventional-size column with a short microbore column in the second dimension (Figure 7.8). There are three fundamental functions of this interface: (1) trapping of small adjacent fractions (typically 2–10 s) of the efﬂuent from the ﬁrst separation column, (2) refocusing these fractions (either in time or in space), and (3) injection of the refocused fractions as narrow pulses into the second-dimension column. The separation on the latter column is extremely fast and takes only 2–10 s versus 20–120 min for the ﬁrst dimension, and is therefore performed under essentially isothermal conditions. A large series of high-speed chromatograms as the outcome of the transfer of chromatographic band from the ﬁrst to the second dimension are generated during the GCGC run. As shown in Figure 7.9, these adjacent pulses are usually stacked side-by-side by a special software to form a 2D chromatogram with one dimension representing the retention time on the ﬁrst column (tR1) and the other, the retention time on the second column (tR2). The most convenient way to visualize GCGC data is as contour plots representing the bird’s eye view, where peaks are displayed as spots on a plane using colors and shading to indicate the signal intensity (Figure 7.9). 7.3.3.2

Optimization of Operation Conditions and Instrumental Requirements in GCGC

Compared to conventional 1D-GC, the optimization of GCGC analysis requires a more complex approach. The changes in operational parameters such as oven temperature or carrier gas ﬂow rate have different impacts on the performance of separation columns since these differ both in their geometry and separation mechanism. Furthermore, new parameters such as modulation frequency and modulator temperature have to be optimized [25]. Conventional columns, typically 15–30 m length 0.25–0.32 mm internal diameter 0.1–1 mm ﬁlm thickness, are used in the ﬁrst dimension. This allows application of virtually all sample introduction techniques (split=splitless, on column, LVI-PTV, DMI=DSI, and=or SPME). Stationary phases commonly used in ﬁrst-dimension columns are typically 100% dimethylpolysiloxane or (5% phenylene)-dimethylpolysiloxane. The separation on these nonpolar columns is governed mainly by analyte volatility. The size of columns for second dimension is commonly in a range of 0.5–2 m length 0.1 mm internal diameter 0.1 mm ﬁlm thickness. More polar stationary phases such as 35%–50% phenylene–65%–50% dimethylpolysiloxane, polyethylene glycol, carborane, and=or cyanopropyl–phenyl–dimethylpolysiloxane are employed. Analytes interact with these mediumpolar=polar phases via various mechanisms such as p–p interactions, hydrogen bonding, etc., hence the requirement for different, independent separation principle is met. In most applications, orthogonality is achieved using nonpolar polar separation mechanisms. To obtain acceptable separation in both dimensions, a compromise has to be made with regard to both columns. The linear velocity of the carrier gas in the (narrow bore) ﬁrst-dimension column is usually rather lower than optimal (about 30 cm=s) while, at the same time, the linear velocity in the (microbore) second-dimension capillary is relatively high, typically exceeding 100 cm=s. Also when setting the temperature programming rate, the requirement for obtaining at least four modulations Injector

Modulator

Detector First column

FIGURE 7.8 GCGC instrument conﬁguration.

Second column

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 132 17.7.2008 4:05pm Compositor Name: VAmoudavally

132

Handbook of Food Analysis Instruments 1D chromatogram (first column outlet)

Modulation Raw 2D chromatogram (second column outlet)

Transformation

Second dimensional chromatograms

n

io

s en

m

d

Firs

t dim

ens

n co

ion

di

Se

Visualization

3D plot

First dimension

First dimension

Second dimension

Second dimension

2D plot

FIGURE 7.9 Generation and visualization of a GCGC chromatogram. (Reproduced from Zrostlikova, J., Hajslova, J., and Cajka, T., J. Chromatogr. A, 1019, 173, 2003. With permission.)

over each ﬁrst-dimension peak (so-called modulation criterion) has to be taken into account. In most analyses, this is achieved by using programming rates as low as 0.58C–58C=min, which is less than in conventional 1D-GC [26]. It should be noted, however, that even steeper programming rates (thus faster GC separation) can be employed (108C–208C=min) in GCGC, which typically results in two modulations over each ﬁrst-dimension peak. Under these conditions, the separation accomplished in the ﬁrst column might be lost. However, because of different activity coefﬁcients on the second column, the analytes can be completely separated (with higher chromatographic resolution than in the case of 1D-GC) in the second dimension [27,28]. For better tuning of the GCGC setup, systems with a programmable second oven are preferred. Effective and robust modulation is a key process in the GCGC analysis. Thermal modulation in a capillary GC can be performed by both heating and cooling. While heated modulators use a thickﬁlm modulation capillary to trap subsequent sample fractions eluting from the ﬁrst column by means

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 133 17.7.2008 4:05pm Compositor Name: VAmoudavally

Gas Chromatography in Food Analysis

133

of stationary phase focusing (compounds are released by the temperature increase), the cryogenically cooled modulators do not use a modulation capillary. Instead, they trap and focus the sample fractions eluting from the ﬁrst column at the front part of the secondary column itself. Initially, moving heated=cooled modulators were used but they exhibited relatively low robustness (fragile capillary can be easily broken). These shortcomings of moving modulators have been overcome by two-stage jet modulators that use a stream of nitrogen or carbon dioxide for cooling a short section of the second column for trapping=focusing of the analytes eluting from the ﬁrst column [29,30]. In practice, ﬁxed modulation frequency, typically in a range of 0.1–10 Hz is employed during the analysis. Under ideal experimental conditions, the retention time of the most retained compound in the second dimension is shorter than a modulation time. If this is not the case, i.e., analytes do not elute in their modulation cycle, so-called wrap-around, which may cause co-elutions, occurs. Avoiding this phenomenon can be achieved, e.g., by an increase of the second-dimension column temperature (if an independent oven is available). In any case, optimal function of modulator is essential for the quality of separation and detection process. The fast separation on a short and microbore second-dimension column results in very narrow peaks with widths of 50–1000 ms at the baseline. Although fast analogue detectors such as a ﬂame ionization detector or electron capture detector (ECD) are fully compatible with fast chromatography and provide reliable peak recognition, they do not provide structural information. Coupling GCGC separation with MS detector results into the three-dimensional system that may contribute to the identiﬁcation of 2D separated peaks and brings a deeper understanding of structured chromatograms [26]. However, conventional scanning MS detectors are typically too slow and do not provide reliable spectra and peak reproduction. At present, only time-of-ﬂight mass spectrometers (see Chapter 10) can acquire the 50 or more mass spectra per second, which are required for the proper reconstruction of the chromatogram and for quantiﬁcation in GCGC [35]. 7.3.3.3

Advantages of GCGC

A number of characteristics of GCGC have been reported that documents superiority of this technique over conventional 1D-GC [26]. High peak capacity. The peak capacity, characterized as a maximal number of chromatographic peaks that can be placed side by side into the available separation space (chromatogram), is signiﬁcantly enhanced. Under the real-world conditions, the total peak capacity in GCGC is rather lower than the calculated value due to the imperfections in the sample transfer between the two columns; however, it still greatly exceeds the limits of conventional GC. As an example, the merit in pesticide residue analysis resulting from the separation power is shown in Figure 7.10. Enhanced sensitivity. Compared to 1D-GC separation, pronounced improvement of detection limits in GCGC system is obtained; thanks to compressing the peak in the modulation capillary and front part of the second column (following fast chromatography avoids band broadening of focused peaks). Furthermore, thanks to improved separation of analytes and matrix interferences (chemical noise) in the GCGC system, the signal to noise ratio is also improved. An example is given in Figure 7.11 that illustrates differences in 1D-GC versus GCGC analysis of limonene. Structured chromatograms. Thanks to complementary separation mechanisms occurring in both columns, the chromatograms resulting from particular GCGC setup are ordered, i.e., molecules have their deﬁnite locations in the retention space based on their structure. In the reconstructed 2D contour plots, characteristic patterns are obtained, in which the members of homological series differing in their volatility are ordered along the ﬁrst-dimension axis (nonpolar capillary is typically employed in ﬁrst dimension), whereas the compounds differing by polarity are spread along the second-dimension axis. The formation of clusters of the various subgroups of compounds in a GCGC contour plot may be useful for the group type analysis. Improved identiﬁcation of unknowns. Nontarget screening allows obtaining of overview of the sample constituents. This approach consists from: (1) peak ﬁnding and deconvolution (algorithm for

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 134 17.7.2008 4:05pm Compositor Name: VAmoudavally

134

Handbook of Food Analysis Instruments

2

20000 18000

250000

16000 14000

200000

12000 10000 8000

150000

1

6000 4000

100000

2000 Time (seconds) spectrum #

50000

Time (seconds) 640 spectrum # 697 (A)

650 747

660 797

79

650 747

660 797

670 847 109

79

680 897

690 947

670 847 109

680 897

690 947

186

700 997

185 Peak true–sample

2 30000

1000

25000

500

109

79 71 87

20000

1000

1

500 47

5000

79

662 0.6 32520

662 0.8 32560

(B)

662 1 32600

662 1.2 32640

662 1.4 32680

79

109

185

93

60 60

1st Time (seconds) 2nd Time (seconds) spectrum #

220

60 80 100 120 140 160 180 200 220 240 Library Hit - similarity 727, “Phosphoric acid, 2,2-dichlorovinyldimethylester” 109

15000 10000

185 128 145

662 1.6 32720

80

145

185

220

100 120 140 160 180 200 220 240

662 1.8 32760

662 2 32800

FIGURE 7.10 Separation of dichlorvos (1) in apple extract at 0.01 mg=kg from matrix co-extract 5-(hydroxymethyl)-5-furancarboxaldehyde (2). Plotted are three most abundant ions in the mass spectrum of dichlorvos (79, 109, and 185). Chromatogram of (A) 1D-GC analysis of zoomed section shows the peak of dichlorvos (m=z 185) and matrix interference (m=z 79 and 109); and (B) GCGC analysis (DB-XLBDB-17 columns); matrix interference resolved on medium polar DB-17 column. Data acquired by TOFMS at acquisition rates 5 spectra=s and 250 spectra=s, respectively.

recognizing of partly co-eluting peaks in the GC–MS chromatogram and obtaining their ‘‘pure’’ mass spectra), (2) library searching, and (3) further post-processing. Since a large amount of data have to be processed, automated data processing is employed.

7.4 SAMPLE DETECTION Depending upon the type of food compounds being measured several different detectors are available for this purpose (Table 7.5), each with its own advantages and drawbacks. The following sections brieﬂy introduce various GC detectors most commonly in use today [32,33].

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 135 17.7.2008 4:05pm Compositor Name: VAmoudavally

135

Gas Chromatography in Food Analysis

S/N = 19570

S/N = 639

3 ⫻104 3⫻104

Abundance

Abundance

4⫻104

2⫻104

1⫻104

2 ⫻104

1⫻104 570

(A)

590 610 Retention time (s)

(B)

590.8 0.22

593.8 0.22

596.8 1t (s) R 0.22 2 tR (s)

FIGURE 7.11 Improvement of detectability of limonene (m=z 93) isolated from honey headspace by SPME under the conditions of (A) 1D-GC and (B) GCGC (DB-5 msSupelcowax-10 columns). Data acquired by TOFMS at acquisition rates of 10 spectra=s and 300 spectra=s, respectively. (Cajka, T. et al., J. Sep. Sci., 30, 534, 2007.)

7.4.1 FLAME IONIZATION DETECTOR Flame ionization detector (FID) represents one of the most widely used detectors. The efﬂuent from an analytical column is mixed with hydrogen and air, and is directed into a ﬂame, which breaks down organic molecules and produces ions. A voltage potential is applied across the gap between the burner tip and an electrode located just above the ﬂame. The resulting current is then measured and is proportional to the concentration of the components present. TABLE 7.5 Overview of GC Detectors Applicable for the Determination of Food Components Detector

Selectivity

Detectability

Linearity 107 104–6 104 104–7 104 N: 105, P: 104

104–7 102

Flame ionization detector (FID) Thermal conductivity detector (TCD) Electron capture detector (ECD) Nitrogen–phosphorus detector (NPD) Halogen-speciﬁc detector (XSD) Thermionic ionization detector (TID)

No No Halogens N, P Halogens N, P

Photoionization detector (PID) Flame photometric detector (FPD) Pulsed ﬂame photometric detector (PFPD)

Aromatics S, P Tuneable for 28 elements

Atomic-emission detector (AED) Electrolytic conductivity detector (ELCD) or Hall electrolytic conductivity detector Mass spectrometric detector (MSD) Fourier transform infrared (FTIR)

Tuneable for any element S, N, halogens

2 pg C=s 300 pg=mL fg=s fg–pg N, P=s pg Cl=s 100 fg N=s, 100 fg P=s pg pga pg S=s, 100 pg P=sa pg=sa pg

Yes Yes

fg–pg pg

a

The detectability considerably varies among particular elements.

pg, picogram; fg, femtogram.

106 S: 103, P: 105 S, P: 103 103–4 106

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 136 17.7.2008 4:05pm Compositor Name: VAmoudavally

136

Handbook of Food Analysis Instruments

7.4.2 THERMAL CONDUCTIVITY DETECTOR Thermal conductivity detector (TCD) consists of an electrically heated wire or thermistor. The temperature of the sensing element depends on the thermal conductivity of the gas ﬂowing around it. Changes in thermal conductivity cause a temperature rise in the element, which is sensed as a change in resistance.

7.4.3 ELECTRON CAPTURE DETECTOR In ECD, the sample is introduced into the detector through an analytical column and passes over a 63 Ni radioactive source emitting b particles, which causes ionization of the carrier gas and the subsequent release of electrons. When organic molecules containing electronegative functional atoms or groups pass by the detector, they capture some of the electrons and reduce the current measured between the electrodes.

7.4.4 NITROGEN–PHOSPHORUS DETECTOR In nitrogen–phosphorus detector (NPD), a glass bead containing an alkali metal is electrically heated until electrons are emitted. These electrons are then captured by stable intermediates to form a hydrogen plasma, which ionizes compounds from the column efﬂuent. A polarizing ﬁeld directs these ions to a collector anode creating a current.

7.4.5 FLAME PHOTOMETRIC DETECTOR

AND

PULSED FLAME PHOTOMETRIC DETECTOR

In ﬂame photometric detector (FPD), a sample is burned in a hydrogen=air ﬂame to produce molecular products that emit light by means of chemiluminescent chemical reactions. The emitted light is then isolated from background emissions by narrow bandpass wavelength-selective ﬁlters and is detected by a photomultiplier and then ampliﬁed. Unfortunately, the detectability of the FPD is limited by light emissions of the continuous ﬂame burning products. This disadvantage is eliminated by pulsed ﬂame photometric detector (PFPD), where a hydrogen=air mixture ﬂows into the FPD so low that a continuous ﬂame could not be sustained. By inserting a constant ignition source into the gas ﬂow, the hydrogen=air mixture would ignite, propagate back through a quartz combustor tube to a constriction in the ﬂow path, extinguish, then reﬁll the detector, ignite, and repeat the cycle.

7.4.6 PHOTO-IONIZATION DETECTOR In photo-ionization detector (PID), the column efﬂuent is ionized by ultraviolet light and the current (proportional to the concentrations of the ionized material) produced by the ion ﬂow is measured.

7.4.7 ELECTROLYTIC CONDUCTIVITY DETECTOR In electrolytic conductivity detector (ELCD), compounds eluting from an analytical column are swept into a nickel reaction tube at the temperature up to 9008C. The components are stripped off their halogenated atoms and these atoms are carried into a conductivity cell. As the concentrations of the halogens change in this cell, the measured conductivity of a solution in the cell changes proportionally.

7.4.8 ATOMIC-EMISSION DETECTOR In atomic-emission detector (AED), eluted compounds from an analytical column are transported into a microwave powered plasma (or discharge) cavity where those compounds are destroyed and their atoms are excited by the energy of the plasma. The emitted light by the excited particles is separated into individual lines via a photodiode array. The individual emission lines are then sorted

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 137 17.7.2008 4:05pm Compositor Name: VAmoudavally

Gas Chromatography in Food Analysis

137

and produce chromatograms consisting of peaks from eluants that contain only a speciﬁc element. In this way, elemental composition can be estimated. It should be noted that the intensity of signal largely varies among the elements and it is relatively low for oxygen, nitrogen, chlorine, bromine while higher sensitivity is obtained for carbon, phosphorus, and sulphur.

7.4.9 MASS SPECTROMETRIC DETECTOR The mass spectrometer (MS) is by far the most powerful and ﬂexible of the detectors used in the analysis of GC-amenable food components today. The advantage over all GC detectors described above is a possibility to obtain, in addition to selective detection of analyte eluted at certain retention time, also structural information, enabling either conﬁrmation of target compound or identiﬁcation of unknown species. The character of data obtained largely depends on the type of mass analyzer employed. The principles of this type of detection are thoroughly discussed in Chapter 10.

7.5 MATRIX EFFECTS Under the real-world conditions, some residues of matrix co-extractives unavoidably remain in the puriﬁed sample prepared for examination by GC analysis. Inaccurate quantiﬁcation, decreased method ruggedness, low analyte detectability, and even reporting of false positive or negative results are the most serious matrix-associated problems, which can be encountered [3]. The extent of these phenomena depend on a wide range factors including sample composition and injection technique employed. Matrix-induced chromatographic response enhancement, ﬁrst described by Erney et al., is presumably the most discussed matrix effect adversely impacting quantiﬁcation accuracy of certain, particularly more polar analytes [34]. Its principle is as follows: During injection of particular compounds in neat solvent, adsorption and=or thermo-degradation of susceptible analytes on the active sites (mainly free silanol groups) present in the injection port and in chromatographic column may occur. On this account, the number of analyte molecules reaching GC detector is reduced. This is, however, not the case when a real-world sample is analyzed. Co-injected matrix components tend to block the active sites in GC system thus reducing the analyte losses and, consequently, enhancing their signals as compared to the injection in neat solvent (Figures 7.12 and 7.13). If these facts are ignored and calibration standards in solvent only are used for calculation of target analytes concentration, recoveries as high as even several hundred percent might be obtained [3]. It is worth noticing that hydrophobic, nonpolar substances, such as persistent organochlorine contaminants (with some exceptions such as DDT that may thermally degrade in a dirty hot injector), are not prone to these hot injection-related problems. Repeated injections of nonvolatile matrix components, which are gradually deposited in the GC inlet and=or front part of the GC column, can give rise to successive formation of new active sites, which might be responsible for the effect, sometimes called matrix-induced diminishment [36]. Gradual decrease in analyte responses associated with this phenomenon together with distorted peak shapes (broadening, tailing) and shifting the retention times towards higher values negatively impact ruggedness, i.e., long-term repeatability of analyte peak intensities, shapes, and retention times, performance characteristic of high importance in routine trace analysis [24]. Three basic approaches and their combination should be considered as way to improved quality assurance [3]: (1) elimination of primary causes, (2) optimization of calibration strategy enabling compensation, and (3) optimization of injection and separation parameters. Unfortunately, the ﬁrst concept of the GC system free of active sites is in principle hardly viable—not only because of commercial unavailability of virtually inert materials stable even under long-term exposure to high temperatures that typically occur in a GC inlet port, but also due to impossibility to control formation of new active sites from deposited nonvolatile matrix. In this content, a more conceivable alternative might be based on avoiding sample matrix to be introduced

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 138 17.7.2008 4:05pm Compositor Name: VAmoudavally

138

Handbook of Food Analysis Instruments

Standard

Sample

C

C Injection

X

Y

Liner

Transfer onto the GC column

C–X

C–Y

FIGURE 7.12 Illustration of the cause of matrix-induced chromatographic enhancement effect; (C) number of injected analyte molecules; (X, Y) number of free active sites for their adsorption in injector; (*) molecules of analyte in injected sample; (.) portion of analyte molecules adsorbed in GC injector; (~) molecules of matrix components in injected sample; (~) portion of matrix compounds adsorbed in GC liner; (C X) < (C Y). (Reproduced from Hajslova, J. and Zrostlikova, J., J. Chromatogr. A, 1000, 181, 2003. With permission.)

260 240 220 200 180 160 140 120 100

(P)

(S)

80 60 Time−> 12.20

12.30

12.40

12.50

FIGURE 7.13 Matrix-induced enhancement response effect: 1 pg of 2-nitronaphthalene (m=z 173) injected in pure solvent (S) and in puriﬁed sample of pumpkin seed oil (P). (Reproduced from Dusek, B., Hajslova, J., and Kocourek, V., J. Chromatogr. A, 982, 127, 2002. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 139 17.7.2008 4:05pm Compositor Name: VAmoudavally

139

Gas Chromatography in Food Analysis

into the GC system. Unfortunately, again, none of the common isolation and=or cleanup techniques are selective enough (mainly in the case of broad scope methods) to avoid the presence of residual sample components in the analytical sample. Since an effective elimination of the sources of the matrix effects is not likely to occur in practice, their compensation by using alternative calibration methods is obviously the most feasible option. Several strategies are conceivable for this purpose: (1) addition of isotopically labeled internal standards, (2) the use of standards addition method, (3) the use of matrix-matched standards, and (4) the use of analyte protectants (introduced only recently). The main disadvantages=requirements of these methods are summarized in Table 7.6 [37]. As regards analyte protectants, these compounds are capable to strongly interact with active sites in the GC system, thus decreasing degradation and adsorption of target analytes [37]. The same amount of the analyte protectants is added to both sample extracts and matrix-free standards, which results in maximization and equalization of the matrix-induced response enhancement effect and avoids overestimation of results, which can occur with standards in neat solvent [38]. A wide range of compounds containing multiple polar=ionizable groups such as various polyols and their derivatives, carboxylic acids, amino acids, and derivatives of basic nitrogen containing heterocycles have been experimentally evaluated as analyte protectants. In a study concerned with the analysis of multiple pesticide residues using hot splitless injection, a mixture of 3-ethoxypropane-1,2-diol, L-gulonic acid g-lactone, and D-glucitol (in acetonitrile extracts) was found to most effectively cover a wide volatility range of GC-amenable analytes [38]. This analyte protectant mixture worked also very well in the multi residue GC analysis of pesticides using DMI [15], which has more active glass surfaces that need effective deactivation during each injection. Figure 7.14 shows chromatograms TABLE 7.6 Quantiﬁcation Strategies and Their Critical Assessment Method Standard additions Isotopically labeled standards

Matrix-matched standards

Analyte protectants

Comments Extra labor effort required for preparation Inaccuracies may occur because the matrix effect is concentration dependent Only a limited number of certiﬁed isotopically labeled standards is currently commercially available; not available in wide scope methods Restriction in the use of detection techniques other than mass spectrometry Additional labor=time burden of developing analytical conditions for so many more compounds Need for enough blank matrix (ideally identical as the samples) and its longterm storage Extra time, labor, and expense for preparing the blank extracts for calibration standards needed Greater amount of matrix material injected onto the column in a sequence, which leads to greater requirements for GC maintenance Greater potential for analyte degradation in the matrix solution Following criteria have to be met for analyte protectants: Unreactiveness with analytes in solution and the GC system Sufﬁcient stability under the GC conditions (no thermodegradation, re-arrangement, etc.) No deterioration of the GC column or detector performance (e.g., due to accumulation) No interference with the detection process of analytes (i.e., low intensity, low mass ions in its MS spectra) Good availability, low cost, no toxicity Good solubility in the solvent of interest

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 140 17.7.2008 4:05pm Compositor Name: VAmoudavally

140

Handbook of Food Analysis Instruments

A) Without analyte protectants

B) With analyte protectants

1.12x CI

Injection in: Matrix Solvent

CI

CI

CI

CI CI

(A) Lindane

S

CI O

3.98x

S P O O

N O

(B) Phosalone

2.16x OH

10.6x (C) o-Phenylphenol

FIGURE 7.14 Comparison of peak shapes and intensities of 100 ng=mL lindane (m=z 219), phosalone (m=z 182), and o-phenylphenol (m=z 170) obtained by injection in matrix (mixed fruit extract) and solvent (MeCN) solutions (A) without and (B) with the addition of analyte protectants (3-ethoxypropane-1,2-diol, L-gulonic acid g-lactone, and D-glucitol at 10, 1, and 1 mg=mL in the injected sample, respectively). (Reproduced from Mastovska, K., Lehotay, S.J., and Anastassiades M., Anal. Chem., 77, 8129, 2005. With permission.)

for three pesticides lindane, phosalone, and o-phenylphenol obtained by hot splitless injection in solvent and matrix-matched standards without and with the above mixture of analyte protectants, demonstrating dramatic improvement in peak shapes and intensities with the use of analyte protectants [38]. Besides the above compensation approaches, also careful optimization of injection and separation parameters (including the choice of suitable injection technique, temperature, and volume; liner size and its design; solvent expansion volume; column ﬂow rate; column dimensions) can reduce to some extent the number of active sites available for interaction (lower surface area) and its duration [37].

7.6 FOOD ANALYSIS APPLICATIONS Since a large range of food compounds are (semi)volatile compounds, the GC is widely used for their determination. The choice of an optimal GC setup depends on the requirements for the performance characteristics of methods used, cost, speed, and several other factors. In Table 7.7, the current GC methods for several groups of food constituents are summarized with special attention paid to applicability of recent advances in the ﬁeld of this technique for their analysis [39–43].

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 141 17.7.2008 4:05pm Compositor Name: VAmoudavally

141

Gas Chromatography in Food Analysis

TABLE 7.7 Overview of Typical Conditions of Most Common Applications Employing GC for Separation Injection Natural substances

Food contaminants

Lipids (fatty acids, mostly derivatized)

Split

Aroma and ﬂavor compounds

Split, splitless, SPME

Modern pesticides

Splitless, PTV, DSI=DMI, SPME

Polychlorinated biphenyls

Splitless, PTV

Polychlorinated dibenzo-p-dioxins and dibenzofurans

Splitless, PTV

Brominated ﬂame retardants

Splitless, PTV

Polycyclic aromatic hydrocarbons

Splitless, PTV

Veterinary drugs (derivatized)

Splitless

Mycotoxins (derivatized) Acrylamide

Splitless

Chloropropanols (derivatized) Heterocyclic amines (derivatized)

Splitless, PTV

Splitless, PTV, DSI

Splitless

Typical GC Column Phase Polyethylene glycol 70% Cyanopropyl-phenyl–30% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane Polyethylene glycol 5% Diphenyl–95% dimethylpolysiloxane 50% Diphenyl–50% dimethylpolysiloxane 6% Cyanopropyl-phenyl–94% dimethylpolysiloxane 35% Diphenyl–65% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane 50% Diphenyl–50% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane 50% Cyanopropyl-phenyl–50% dimethylpolysiloxane other special phases 100% Dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane 14% Cyanopropyl-phenyl–86% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane 50% Diphenyl–50% dimethylpolysiloxane 100% Dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane (derivatized form) Polyethylene glycol (nonderivatized form) 5% Diphenyl–95% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane 50% diphenyl–50% dimethylpolysiloxane

Detection FID, MS

MS, FID

MS, ECD, NPD, FPD, PFPD, AED, PID, ELCD

ECD, MS

MS

MS, ECD

MS, PID

MS

MS, ECD MS (both forms), ECD (derivatized form) MS MS

(continued )

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 142 17.7.2008 4:05pm Compositor Name: VAmoudavally

142

Handbook of Food Analysis Instruments

TABLE 7.7 (continued) Overview of Typical Conditions of Most Common Applications Employing GC for Separation Injection Food contaminants

Phthalate and adipate esters Epoxy-compounds (derivatized)

Splitless, SPME Splitless

Typical GC Column Phase 5% Diphenyl–95% dimethylpolysiloxane 5% Diphenyl–95% dimethylpolysiloxane

Detection MS, ECD MS

Note: AED, atomic-emission detector; DMI, difﬁcult matrix introduction; DSI, Direct sample introduction; ECD, electron capture detector; ELCD, electrolytic conductivity detector; FID, ﬂame ionization detector; NPD, nitrogen– phosphorus detector; PFPD, pulsed ﬂame photometric detector; PID, photo-ionization detector; PTV, programmable temperature vaporization; MS, mass spectrometry; SPME, solid-phase microextraction.

7.7 CONCLUSION AND FUTURE TRENDS After several decades of GC on the market, the technology and its applications have improved signiﬁcantly. Despite that they have not reached an end to the possibilities, which are conceivable. There are always new challenges for further improvements of performance and extending the scope of applications. Considering future uses of GC in food analysis, the main trend foreseen is successive replacement of conventional detection approaches by MSDs employing various types of mass analyzers. Fast GC–MS can be introduced in many applications; thanks to the spectral resolution of co-eluting compounds that can compensate for lower GC resolution obtained in highspeed separations.

ACKNOWLEDGMENTS This chapter was ﬁnancially supported by the Ministry of Education, Youth and Sports of the Czech Republic (project MSM 6046137305).

REFERENCES 1. Hinshaw, J.V., Setting realistic expectations for GC optimization, LC GC Eur., 20, 138, 2007. 2. Hewlett-Packard FlowCalc 2.0 software. Available at http:==www.chem.agilent.com=cag=servsup=usersoft= ﬁles=GCFC.htm via the Internet. Accessed July 1, 2007. 3. Hajslova, J. and Zrostlikova, J., Matrix effects in (ultra)trace analysis of pesticide residues in food and biotic matrices, J. Chromatogr. A, 1000, 181, 2003. 4. Godula, M., Hajslova, J., and Alterova, K., Pulsed splitless injection and the extent of matrix effects in the analysis of pesticides, J. High Resolut. Chromatogr., 22, 395, 1999. 5. Zrostlikova, J. et al., Performance of programmed temperature vaporizer, pulsed splitless and on-column injection techniques in analysis of pesticide residues in plant matrices, J. Chromatogr. A, 937, 73, 2001. 6. Stan, H.J. and Müller, H.M., Evaluation of automated and manual hot-splitless, cold-splitless (PTV), and on-column injection technique using capillary gas chromatography for the analysis of organophosphorus pesticides, J. High Res. Chromatogr., 11, 140, 1988. 7. Grolimund, B. et al., Solvent trapping during large volume injection with an early vapor exit, Part 2: Chromatographic results and conclusions, J. High Res. Chromatogr., 21, 378, 1998. 8. Bailey, R., Injectors for capillary gas chromatography and their application to environmental analysis, J. Environ. Monit., 7, 1054, 2005. 9. Grob, K. and Li, Z., PTV splitless injection of sample volumes up to 20 mL, J. High Res. Chromatogr., 11, 626, 1988.

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 143 17.7.2008 4:05pm Compositor Name: VAmoudavally

Gas Chromatography in Food Analysis

143

10. Grob, K. and Läubli, Th., Splitless injection-development and state of the art. Including a comparison of matrix (‘‘dirt’’) effects in conventional and PTV splitless injection, J. High Res. Chromatogr., 11, 462, 1988. 11. Termonia, M., Lacomblez, B., and Munari, F., Optimization of the cold split-splitless injector in the solvent-split mode, J. High Res. Chromatogr., 11, 890, 1988. 12. Staniewski, J. and Rijks, J.A., Potential and limitations of differently designed programmed-temperature injector liners for large volume sample introduction in capillary GC, J. High Res. Chromatogr., 16, 182, 1993. 13. Godula, M. et al., Optimization and application of the PTV injector for the analysis of pesticide residues, J. Sep. Sci., 24, 355, 2001. 14. Jing, H. and Amirav, A., Pesticide analysis with the pulsed-ﬂame photometer detector and a direct sample introduction device, Anal. Chem., 69, 1426, 1997. 15. Cajka, T. et al., Use of automated direct sample introduction with analyte protectants in the GC–MS analysis of pesticide residues, J. Sep. Sci., 28, 1048, 2005. 16. Patel, K. et al., Evaluation of large volume-difﬁcult matrix introduction–gas chromatography–time of ﬂight–mass spectrometry (LV-DMI-GC-TOF-MS) for the determination of pesticides in fruit-based baby foods, Food Addit. Contam., 21, 658, 2004. 17. Lehotay, S.J., Analysis of pesticide residues in mixed fruit and vegetable extracts by direct sample introduction=gas chromatography=tandem mass spectrometry, J. AOAC Int., 83, 680, 2000. 18. Kataoka, H., Lord, H.L., and Pawliszyn, J., Applications of solid-phase microextraction in food analysis, J. Chromatogr. A, 880, 35, 2000. 19. Mastovska, K. and Lehotay, S.J., Practical approaches to fast gas chromatography–mass spectrometry, J. Chromatogr. A, 1000, 153, 2003. 20. Mastovska, K., Instrumental aspects and application of (ultra)fast gas chromatography–mass spectrometry, in Niessen, W. (Ed.), Encyclopedia of Mass Spectrometry, Elsevier, Oxford, 2006, p. 73. 21. Mondello, L. et al., Evaluation of fast gas chromatography and gas chromatography–mass spectrometry in the analysis of lipids, J. Chromatogr A., 1035, 237, 2004. 22. de Zeeuw, J. et al., A simple way to speed up separations by GC–MS using short 0.53 mm columns and vacuum outlet conditions, J. High Resol. Chromatogr., 23, 677, 2000. 23. Mastovska, K., Lehotay, S.J., and Hajslova, J., Optimization and evaluation of low-pressure gas chromatography–mass spectrometry for the fast analysis of multiple pesticide residues in a food commodity, J. Chromatogr. A, 926, 291, 2001. 24. Mastovska, K., Lehotay, S.J., and Hajslova, J., Ruggedness and other performance characteristics of lowpressure gas chromatography–mass spectrometry for the fast analysis of multiple pesticide residues in food crops, J. Chromatogr. A, 1054, 335, 2004. 25. Dallüge, J. et al., Optimization and characterization of comprehensive two-dimensional gas chromatography with time-of-ﬂight mass spectrometric detection (GCGC-TOF MS), J. Sep. Sci., 25, 201, 2002. 26. Dallüge, J., Beens, J., and Brinkman, U.A.Th., Comprehensive two-dimensional gas chromatography: A powerful and versatile analytical tool, J. Chromatogr. A, 1000, 69, 2003. 27. Zrostlikova, J., Hajslova, J., and Cajka, T., Evaluation of two-dimensional gas chromatography–time-ofﬂight mass spectrometry for the determination of multiple pesticide residues in fruit, J. Chromatogr. A, 1019, 173, 2003. 28. Cajka, T. et al., Solid phase microextraction–comprehensive two-dimensional gas chromatography–timeof-ﬂight mass spectrometry for the analysis of honey volatiles, J. Sep. Sci., 30, 534, 2007. 29. Adahchour, M. et al., Recent developments in comprehensive two-dimensional gas chromatography (GCGC): I. Introduction and instrumental set-up, TrAC-Trend Anal. Chem., 25, 438, 2006. 30. Adahchour, M. et al., Recent developments in comprehensive two-dimensional gas chromatography (GCGC): II. Modulation and detection, TrAC-Trend Anal. Chem., 25, 540, 2006. 31. Cajka, T. and Hajslova, J., Gas chromatography–time-of-ﬂight mass spectrometry in food analysis, LC GC Eur., 20, 25, 2007. 32. Larson, P., Detectors for quantitative gas chromatography, in Handley, A.J. and Adlard, E.R. (Eds.), Gas Chromatographic Techniques and Applications, Shefﬁeld Academic Press, Shefﬁeld, 2001, p. 122. 33. Powell, M., Detectors for compound identiﬁcation, in Handley, A.J. and Adlard, E.R. (Eds.), Gas Chromatographic Techniques and Applications, Shefﬁeld Academic Press, Shefﬁeld, 2001, p. 140. 34. Erney, D.R. et al., Explanation of the matrix-induced chromatographic response enhancement of organophosphorus pesticides during open tubular column gas chromatography with splitless or hot on-column injection and ﬂame photometric detection, J. Chromatogr. A, 638, 57, 1993.

Otles/Handbook of Food Analysis Instruments 45660_C007 Final Proof page 144 17.7.2008 4:05pm Compositor Name: VAmoudavally

144

Handbook of Food Analysis Instruments

35. Dusek, B., Hajslova, J., and Kocourek, V., Determination of nitrated polycyclic aromatic hydrocarbons and their precursors in biotic matrices, J. Chromatogr. A, 982, 127, 2002. 36. Fajgelj, A. and Ambrus A. (Eds.), Principles and Practices of Method Validation, Royal Society of Chemistry, Cambridge, United Kingdom, 2000. 37. Anastassiades, M., Mastovska, K., and Lehotay, S.J., Evaluation of analyte protectants to improve gas chromatographic analysis of pesticides, J. Chromatogr. A, 1015, 163, 2003. 38. Mastovska, K., Lehotay, S.J., and Anastassiades M., Combination of analyte protectants to overcome matrix effects in routine GC analysis of pesticide residues in food matrixes, Anal. Chem., 77, 8129, 2005. 39. Hajslova, J. and Cajka, T., Gas chromatography–mass spectrometry (GC–MS), in Pico, Y. (Ed.), Food Toxicants Analysis, Elsevier, Oxford, 2006, p. 419. 40. Careri, M., Bianchi, F., and Corradini, C., Recent advances in the application of mass spectrometry in food-related analysis, J. Chromatogr. A, 970, 3, 2002. 41. Santos, F.J. and Galceran, M.T., Modern developments in gas chromatography–mass spectrometry-based environmental analysis, J. Chromatogr. A, 1000, 125, 2003. 42. Mastovska, K., Food & nutritional analysis: (q) Pesticide residues, In Worsfold, P., Townshend, A., and Poole, C. (Eds.), Encyclopedia of Analytical Science, Elsevier, Oxford, 2005, p. 251. 43. Covaci, A., Voorspoels, S., and de Boer, J., Determination of brominated ﬂame retardants, with emphasis on polybrominated diphenyl ethers (PBDEs) in environmental and human samples—a review, Environ. Int., 29, 735, 2003.

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 145 16.7.2008 11:24pm Compositor Name: MSubramanian

Layer 8 Preparative Chromatography in Food Analysis Joseph Sherma CONTENTS 8.1 8.2 8.3 8.4

Introduction .......................................................................................................................... 145 Layers for PLC .................................................................................................................... 146 Sample Application ............................................................................................................. 146 Development of the Layer ................................................................................................... 148 8.4.1 Capillary Flow Development in CPLC ................................................................... 148 8.4.2 FFPLC ..................................................................................................................... 150 8.4.2.1 OPLC ........................................................................................................ 150 8.4.2.2 RPC .......................................................................................................... 152 8.5 Detection of Zones .............................................................................................................. 152 8.5.1 Viewing in White and UV Light ............................................................................ 152 8.5.2 Post-Chromatographic Derivatization ..................................................................... 153 8.5.3 Instrumental Detection and Documentation ............................................................ 153 8.5.4 Detection of Radioactive Zones .............................................................................. 155 8.6 Recovery of Separated Zones from the Layer .................................................................... 155 8.7 Applications of PLC in Food Analysis ............................................................................... 155 8.8 Conclusions and Future Prospects ...................................................................................... 156 References .................................................................................................................................... 157

8.1 INTRODUCTION Preparative layer chromatography (PLC) is a method used to separate and isolate larger amounts of material (e.g., 1–1000 mg) than are applied in analytical thin layer chromatography (TLC) or high performance TLC (HPTLC). Micropreparative TLC is the term used when lower amounts in the preparative range are chromatographed on thinner preparative layers, most often 0.5 [1], or 0.25 mm analytical layers. The purpose of PLC is to obtain pure compounds for such tasks as further chromatographic (e.g., gas chromatography [2] and column high performance liquid chromatography [HPLC] [3]) or spectrometric (e.g., infrared, ultraviolet [UV], nuclear magnetic resonance [NMR], and mass spectrometry [MS] [4,5]) analysis; investigations of physical, chemical, pharmacological, or biological properties; obtaining analytical standards; or additional reactions in a synthetic sequence. Simple and inexpensive ascending capillary ﬂow development in a closed chamber is most widely used (classical PLC, CPLC), but ﬂow achieved by application of pressure or centrifugal force can also be used (forced ﬂow PLC, FFPLC). This chapter brieﬂy

145

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 146 16.7.2008 11:24pm Compositor Name: MSubramanian

146

Handbook of Food Analysis Instruments

reviews the procedural steps and instrumentation of modern PLC and gives selected applications to food analysis. For more details on techniques and applications of CPLC, readers should consult a recent comprehensive book [6].

8.2 LAYERS FOR PLC Commercial precoated plates with 0.5–2 mm layer thickness are normally used for increased loading capacity compared to 0.1–0.25 mm layers for TLC and HPTLC. Precoated preparative plates are available with silica gel, aluminum oxide (alumina) [7], kieselguhr, cellulose, octadecyl (C-18) and ethyl (C-2) [8] reversed-phase bonded silica gel, and other sorbents. Normal phase silica gel is most widely used and provides group selectivity and separation of positional isomers; reversed phase sorbents give separations of homologous compounds and slower mobile phase velocity. Preparative layers are often formulated with gypsum (CaSO4 hydrate) binder (termed as a G-plate), which leads to a soft layer that can be removed by scraping more easily than if an organic polymer binder of the type that is typical for analytical layers was used. Bulk sorbents are also available for self-making preparative layers in the laboratory, but they are less reproducible than precoated layers; these bulk sorbents may be designated P-type and often contain gypsum binder. Silica gel is by far the most widely used layer for PLC. Correct prewashing and storage of plates are critical for recovery of compounds without interferences from the layer. Precoated plates with an inert pre-adsorbent or concentrating zone composed of diatomaceous earth or large pore silica gel 50,000 facilitate the application of samples and formation of narrow developed bands with higher resolution. Sample compounds spotted onto the pre-adsorbent zone migrate with the mobile phase front and are focused into a tight, concentrated initial band at the preadsorbent=sorbent interface. Analtech (Newark, DE) offers a PLC plate with a 0.7 mm thick pre-adsorbent and tapered silica gel layer (0.3–1.7 mm, bottom to top) that gives an improved mobile phase ﬂow pattern, less vertical band spreading, and much increased resolution in PLC. Layers can be impregnated to improve separations in PLC. For example, fatty acids were isolated from soybean oil using silver nitrate impregnated silica gel plates [9] and fatty acids from albacore on silica gel layers impregnated with boric acid [10]. The layers and bulk sorbents discussed above are used in CPLC; they are available from companies such as Merck (Darmstadt, Germany), Whatman (Florham Park, NJ), Analtech, and Macherey-Nagel (Dueren, Germany) and are described more fully by Hauck and Schulz [11]. Layers for FFPLC have the same sorbents but are of special design.

8.3 SAMPLE APPLICATION The sample dissolved (typically at a concentration of 5–10%, w=v) in a weak (nonpolar for silica gel), volatile solvent is applied as a uniform, straight, narrow band across the plate. Sample solubility in the chosen solvent is very important. Ideally, the sample volume and concentration should allow the sample to be absorbed within the entire layer thickness, not just the surface. Manual application can be achieved by hand with a pipet or syringe guided by a ruler. Manual application is aided by the use of layers with an inert pre-adsorbent zone. Modern sample application instruments are available commercially for applying narrow bands of sample. The Analtech semiautomatic TLC sample streaker with a 100 or 500 mL syringe (Figure 8.1) applies a 1 mm wide band across a PLC plate up to 40 cm wide, without scratching the layer, in about one-tenth the time needed for hand spotting. The syringe plunger is continually pressed downward manually to force out the liquid as it travels laterally across the sloping stainless steel bar. The Camag (Muttenz, Switzerland) Linomat 5 (Figure 8.2) applies a sample in a nonpolar or polar solvent for PLC by spraying from the tip of a 100 or 500 mL syringe by means of compressed

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 147 16.7.2008 11:24pm Compositor Name: MSubramanian

Preparative Layer Chromatography in Food Analysis

147

FIGURE 8.1 Analtech TLC sample streaker.

nitrogen ﬂow; operation is automated except for manual ﬁlling of the syringe. The tip of the syringe is adjusted to be about 1 mm above the layer, and the stage movement and application rate are controlled under winCATS software to apply a compact, uniform band of selectable length across the layer. Operation is compliant with the requirements of good manufacturing practice

FIGURE 8.2

Camag Linomat 5 sample applicator.

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 148 16.7.2008 11:24pm Compositor Name: MSubramanian

148

Handbook of Food Analysis Instruments

(GMP)=good laboratory practice (GLP). The Linomat can also be programmed to apply shorter, low volume bands for analytical TLC. The Desaga (Wiesloch, Germany) AS-30 is a similar semiautomated gas-stream spray-on applicator in which a stepping motor depresses the plunger of a 100 mL syringe downward and a second motor moves the syringe tower sideways across the plate. The Camag ATS4 is a fully automatic sample applicator that can apply bands with similar spray-on operation as the Linomat 5. Added features include unattended syringe reﬁlling for the same sample, syringe rinsing between different samples, faster application rate of larger sample volumes by a heated spray nozzle (308C–608C), over-spotting onto the same zone, and selectable sample volume up to 1 mL. Initial zones in the form of round spots can also be applied automatically for analytical TLC.

8.4 DEVELOPMENT OF THE LAYER 8.4.1 CAPILLARY FLOW DEVELOPMENT

IN

CPLC

The mobile phase is usually selected by trial and error guided by prior experience or by performing preliminary analytical separations of the sample in a saturated chamber. PLC separations will be inferior to analytical TLC separations using the same mobile phase because of the thicker layer, larger particle size (e.g., 5–40 mm for Merck plates), and overloaded sample conditions. A good general rule is that analytical TLC should achieve separations with at least 0.1 Rf value difference if the PLC separations are to be adequate with the transferred mobile phase. Isocratic development is usually used, but gradient development has been applied in certain situations for increased resolution. Development times are slower in PLC than in analytical TLC, but sample loading capacity is higher. Rectangular glass tanks (N-chambers) with inner dimensions of 21 21 9 cm are used most frequently for the ascending, capillary ﬂow development of PLC plates, which usually measure 20 20 cm. The tank is lined with thick chromatography paper (e.g., Whatman 3 MM) soaked in the mobile phase and allowed to equilibrate with the mobile phase vapor for up to 2 h before development for a maximum distance of 18 cm. A saturated chamber provides faster capillary ﬂow of the mobile phase, more uniform bulk and alpha solvent fronts, and higher separation efﬁciency. A plate angle of 758 from horizontal is recommended for the fastest development with minimum zone distortion. A single ascending development is most often used, but multiple and two-dimensional development can be used for PLC in special situations. The Camag twin-trough chamber (Figure 8.3) is an N-chamber for 20 20 cm plates with a glass ridge along the base separating it into two compartments. The trough opposite to the one used for mobile phase development can be left empty or ﬁlled with the mobile phase or a different solvent to provide various modes of plate conditioning and chamber saturation before or during development. The MAT-DC is an instrument for automated rather than manual ascending linear development that is available from Baron (Baron Laborgeraete, Insel Reichenau, Germany; Figure 8.4). It is supplied with a large polypropylene solvent trough that is suitable for PLC plates. The instrument has a sensor for mobile phase front recognition and exactly indicates the development time. Automated development is carried out in a light- and air-protected chamber and stopped after a deﬁned time by removal of solvent from the trough. After development, the mobile phase vapor is exhausted from the chamber by an integrated blower [12]. Horizontal capillary action linear development of 20 20 cm PLC plates has been carried out mostly using the Dzido-Soczewinski (DS)-II chamber (Chromdes, Lublin, Poland) [8]. The mobile phase in a shallow horizontal container positioned at about the same level as the PLC plate is supplied to the layer by shifting a small glass cover plate to the edge of the PLC plate. A vertical meniscus of mobile phase forms in the container and moves in the direction of the PLC plate during development. The latest model (Figure 8.5) is equipped with a mobile phase distributor that

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 149 16.7.2008 11:24pm Compositor Name: MSubramanian

Preparative Layer Chromatography in Food Analysis

FIGURE 8.3 Camag twin-trough chamber.

FIGURE 8.4 Baron MAT-DC automatic development chamber adjustable for preparative purposes.

149

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 150 16.7.2008 11:24pm Compositor Name: MSubramanian

150

Handbook of Food Analysis Instruments

FIGURE 8.5

Chromdes DS-II horizontal development chamber.

allows the chamber to be covered with a glass plate during introduction of the mobile phase into the container and at the start of chromatogram development. Chambers with glass distributors also allow online application of large volumes of samples [1]. Development can be carried out in normal and sandwich (absence of vapor phase) modes, saturated or unsaturated condition. Development in the DS-chamber can be superior to ascending development in an N-chamber because of very low solvent consumption, the mobile phase is not ﬂowing against gravity, and the free volume around the plate is minimal, leading to rapid and uniform equilibration. Various methods can be performed, including gradient elution, multiple development, two-dimensional (2-D) development, sample preconcentration, and band application. A higher separation efﬁciency can be achieved by using higher temperature in some applications [13]. The DS chamber is described in detail by Dzido [14]. Instead of the common single one-dimensional (1-D) development, double [15] or triple [16] 1-D development with the same mobile phase or different mobile phases can be used to increase resolution of the mixture components. As a speciﬁc example, PLC on C-18 bonded silica gel was carried out using development with water-acetonitrile-tetrahydrofuran-acetic acid (85:12:12:1) and then (81:14:2:3) for blood orange juice authentication based on cinnamic acid derivatives [17]. The 2-D PLC is another development method for increasing resolution. A single sample is applied in the corner of a plate, which is then developed at right angles to each other with two different mobile phases having unique selectivities for the mixture components. An example is the 2-D semipreparative TLC of phenolic extracts of brandies on cellulose [18].

8.4.2 FFPLC FFPLC, in which ﬂow of mobile phase is achieved by external pressure (overpressured layer chromatography: OPLC) or centrifugal force (rotation planar chromatography: RPC), might lead to better resolution than CPLC because an optimum mobile phase velocity can be used and compounds migrate over the entire separation distance. Both types of FFPLC can be used as an online technique, with connection to a ﬂow detector, recording of chromatograms, elution from the stationary phase, and recovery in a fraction collector, but complex, relatively expensive instrumentation is needed. 8.4.2.1

OPLC

The major instrument for preparative OPLC (POPLC) is the Personal OPLC 50 system (OPLCNIT Ltd., Budapest, Hungary; Figure 8.6). The system components are the liquid delivery unit

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 151 16.7.2008 11:24pm Compositor Name: MSubramanian

Preparative Layer Chromatography in Food Analysis

151

FIGURE 8.6 OPLC-NIT personal OPLC 50 instrument.

with microprocessor controlled programming and the OPLC separation chamber. The separation chamber is pressurized to 5 MPa or 50 bars (725 psi) via an on-board hydraulic system using a water–glycerol mixture driven by a single-piston pump. Layers used in POPLC are 20 20 cm, up to 0.5 mm thick, on glass backing; the layer is sealed with an elastic polymer membrane under the pressure of the hydraulic system. Mobile phase ﬂow is controlled using a typical column HPLC pump in the range of 10 mL=min to 10 mL=min. Unidirectional, bidirectional, bidimensional linear, or circular separations can be performed with single or multiple runs; linear development has been used in the few preparative applications reported so far. Samples can be streaked on the layer in an ofﬂine mode manually or automatically, or a sample can be injected online into the stream of mobile phase ﬂowing through the layer closed by external pressure into the chamber using manual injection or an autosampler. For ofﬂine detection, the cassette including the layer can be removed from the chamber and the separation examined as in CPLC; for online detection, the chamber can be interfaced with all HPLC detectors, e.g., UV, ﬂuorescence, radiometric, evaporative light scattering, MS, or NMR. A paper on preparative OPLC [19] reported use of a 20 20 cm silica gel 60 layer (0.2 mm thickness) sealed on four sides in conjunction with a loading of 25–150 mg, which is not as high as could be obtained if a thicker layer is used. A later paper [20] reported the combination of OPLC with the ﬂowing eluent wall (FEW) technique and the BioArena bioautography system for the detection of antimicrobial activity and compound isolation based on peak collection; online FEW–OPLC isolation of compounds and further ofﬂine OPLC investigation of fractions were illustrated. Details of POPLC are available in book chapters by Nyiredy [21] and Tyihak and Mincsovics [22].

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 152 16.7.2008 11:24pm Compositor Name: MSubramanian

152

Handbook of Food Analysis Instruments

FIGURE 8.7

8.4.2.2

Analtech Cyclograph 1 RPC instrument.

RPC

All preparative RPC separations have been performed in the circular mode up to this time. The CycloGraph with RAVE (radially accelerated variable engine) technology is a centrifugal PLC instrument from Analtech for performing PLC (Figure 8.7) that has had much more use than POPLC. Separations occur within 20 min, there is no need to scrape off and elute separated bands as in CPLC, and repeated use of layers is possible. The sample solution is applied inside the adsorbent ring of a precast rotor (1–8 mm silica gel G and 2 or 4 mm alumina rotors with UV indicator are available) using a solvent pump or hand-held syringe. The mobile phase is pumped at an appropriate speed (100–1400 rpm) through the adsorbent layer to separate mixture components. As the individual rings reach the outer rim of the rotor, they are spun off the edge of the glass and collected in a circular trough; the angle of the trough allows the eluent to collect at the bottom of the trough and drip out of the collection port. An integrated 4 W UV lamp allows viewing of separations. A set of rotor scrapers with scraper device are available for use in place of online collection. RPC was applied to the preparative puriﬁcation of ergosterol peroxide with hexane-ethyl acetate mobile phase at a ﬂow rate of 4 mL=min [23].

8.5 DETECTION OF ZONES After plate development and drying of the mobile phase (usually in a vacuum desiccator rather than by heating, which could change labile compounds), detection of separated zones is made by nondestructive methods because separated compounds must be recovered in an unchanged, pure state. The methods described in this section meet this goal.

8.5.1 VIEWING

IN

WHITE AND UV LIGHT

Zones containing separated compounds can be detected by viewing their natural color in daylight (white light). Naturally, ﬂuorescent compounds can be detected by inspection under 254 or 366 nm light in a viewing box containing these UV source lamps. Compounds that absorb UV light at or

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 153 16.7.2008 11:24pm Compositor Name: MSubramanian

Preparative Layer Chromatography in Food Analysis

153

near 254 or 366 nm can be detected on preparative layers containing a ﬂuorescent indicator (phosphor), which are marked ‘‘F’’ or ‘‘UV’’ on commercial plates. This method, termed ﬂuorescence quenching, gives dark zones on a ﬂuorescent background.

8.5.2 POST-CHROMATOGRAPHIC DERIVATIZATION Post-chromatographic chromogenic or ﬂuorogenic reagents can be used to detect compounds that are not naturally visible or ﬂuorescent and do not quench ﬂuorescence. One of the most widely used reagents is iodine vapor, which reversibly detects many classes of chemical compounds as brown zones on a pale yellow-brown background [24]. Destructive universal or selective chromogenic or ﬂuorogenic reagents must be applied only to the side edges of the layer (the rest of the layer is covered with a glass plate) to locate the areas from which to recover the separated, pure compounds. Analtech sells Prep-scored Uniplates that scored 2.5 cm from each vertical edge for use with destructive detection reagents; after developing, the strips are snapped away from the center portion and the reagent is applied, followed by realignment with the center portion to locate the zones to be recovered by scraping off. Among the most common detection reagents are the ﬂuorescent dyes 20 ,70 -dichloroﬂuorescein [25] and rhodamine B [26], followed by zone visualization under UV light. Other detection reagents that have been used in PLC are listed in Section 8.7. The plate is exposed to iodine vapor by placing it inside a closed, equilibrated chamber that contains some iodine crystals. Reagent solutions are almost always applied to the layer by manual spraying or dipping. Commercial instruments are available for automated spraying (e.g., the Desaga ChromaJet DS 20) or dipping (the Camag Chromatogram Immersion Device), but these are designed mostly for use with TLC and HPTLC plates in quantitative applications where uniform, consistent reagent application is critical, rather than in PLC.

8.5.3 INSTRUMENTAL DETECTION

AND

DOCUMENTATION

Various instruments are available for documentation of PLC separations, detection of zones, and measurement of spectra for determination of band purity and compound identiﬁcation. An image of an entire layer with colored, ﬂuorescent, or ﬂuorescence-quenched zones that were detected by the methods described in Sections 8.5.1 and 8.5.2 on different tracks can be photographed for documentation by use of an instrument such as the Camag DigiStore 2 system with high resolution 12-bit charged couple device (CCD) digital camera and winCATS software (Figure 8.8). The system consists of the Reprostar 3 for illumination of the plate in white light or using 254 or 366 nm UV lamps and a 12-bit camera with a highly linear CCD chip for accurate color reproduction of preparative chromatograms for archival purposes. The captured image can be annotated. Substances that quench ﬂuorescence of an F-plate are photographed under shortwave 254 nm UV light as dark zones, longwave 366 nm UV light is used to excite substances that ﬂuoresce (a cutoff ﬁlter must be used to prevent the illuminating UV light from reaching the camera while the emitted UV light of different wavelengths passes through), and white light is used to visualize colored substances. Three illumination modes are used: reﬂectance, transmission, and reﬂectance þ transmission. Reﬂectance gives an image that is very close to that seen by eye in white light, while transmission adds information on weak sample zones with molecules located deeper inside the layer. A slit-scanning densitometer such as the Camag TLC Scanner 3 (Figure 8.9) can be used to record chromatograms with colored, UV absorbing, and ﬂuorescent zones at an optimum visible or UV wavelength (190–800 nm spectral range), chosen by setting the monochromator, to produce a densitogram [8]. UV-absorbing bands that are not visible by eye on plates without a ﬂuorescent indicator can be detected in this way. The ability of a slit scanner to measure in situ absorption or ﬂuorescence spectra by scanning through all wavelengths from a visible or UV source gives information for zone identiﬁcation; spectra of unknowns are compared automatically with spectra

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 154 16.7.2008 11:24pm Compositor Name: MSubramanian

154

Handbook of Food Analysis Instruments

FIGURE 8.8

Camag DigiStore 2 documentation system.

FIGURE 8.9

Camag TLC Scanner 3.

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 155 16.7.2008 11:24pm Compositor Name: MSubramanian

Preparative Layer Chromatography in Food Analysis

155

of known standards chromatographed on the same plate or a library of computer-stored spectra. Densitometry can better deﬁne the exact location of zone boundaries that may not be clearly ascertained visually, thereby allowing more accurate zone removal (see below). Comparing scanned spectra from various locations within a band can prove its homogeneity and purity.

8.5.4 DETECTION

OF

RADIOACTIVE ZONES

Radioactive PLC zones can be detected by use of various instruments, such as spark chambers, radioscanners, linear analyzers, multiwire proportional counters, and bioimaging=phosphor-imaging analyzers [27], as well as radioautography. Speciﬁc examples are the detection of radioactive zones of alfaprostol and metabolites on silica gel with a spark chamber [28] and 14C-labeled deltamethrin by radioautography [29], but little use of this technology has been reported so far for PLC in food analysis.

8.6 RECOVERY OF SEPARATED ZONES FROM THE LAYER The zones containing the desired compounds are scraped from the plate backing into small tubes or funnels plugged with glass wool, the compounds are eluted with a strong solvent, any remaining sorbent particles are separated, and the solution is concentrated as required. Scraping off is done manually with a spatula, scalpel, or razor blade; there is no commercial scraping instrument suitable for PLC. The details of compound recovery for CPLC by scraping of sorbent zones and elution with solvent are given in Ref. [30]. As stated above, elution and collection of separated compounds in FFPLC can be carried out online, or separated compounds can be recovered by traditional scraping off and elution methods. The ChromXtractor (ChromAn, Holzhausen, Germany) is a new instrument for quantitative elution of substances on TLC plates. A specially designed stamp isolates the area of the spot and seals with a cut against the plate. The substances in the spot will be eluted with a small amount of a suitable liquid, which can be delivered at a ﬂow rate of about 0.1 mL=min from a linked HPLC pump. The extract is then available for further analysis in an LC=MS system as well as by diode array detection, or fraction collection. There is only a small dead volume (ca. 100 mL is enough for elution), and the detector or collector is directly coupled. The instrument is not suitable for elution of long bands in PLC, but parts of bands can be analyzed online for identiﬁcation and conﬁrmation of identity.

8.7 APPLICATIONS OF PLC IN FOOD ANALYSIS The following are selected applications of PLC in food analysis (determination of natural food component or ingredient, degradation product, intentional additive, or contaminant in food crop or ﬁnished food). Each was carried out on a silica gel F-layer with single, capillary ﬂow ascending mobile phase development, and detection by UV absorbance at 254 nm (ﬂuorescence quenching) unless otherwise noted: . . . . . . . . .

Alkylresorcinol in crude latex extract [31] Antibacterial constituents in honey (agar plate assay detection) [32] Butylated hydroxyanisole (BHA) and butylated hydroxytoluene (BHT) in polyethylene food packaging [33] Buntan extract in fruit juices and essential oil food products [34] Carboxylic acids in rye [35] Characteristic minor components in vanilla [36] Cholesterol in soybean oil [37] Chlorogenic acid in almond (natural product detection reagent: 1% methanolic diphenylboric acid ethylamino ester followed by 5% ethanolic polyethylene glycol 4000) [38] Chlorogenic acid derivatives in bamboo (silica gel and C-18 bonded silica gel layers) [39]

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 156 16.7.2008 11:24pm Compositor Name: MSubramanian

156 . . . . . . . . . . . . . . . . . . . . . . . . .

Handbook of Food Analysis Instruments

Colorants in paprika [40] Curcumin, demethoxy curcumin, and bis-demethoxy curcumin in curcuma (detection under 366 nm UV light) [41] Deoxynivalenol puriﬁcation (sulfuric acid and aluminum chloride detection reagents) [42] Deoxynivalenol in grains and snack foods (aluminum chloride detection reagent) [4] Fatty acid diesters in goat milk [43] Fatty acids in seed oil (20 ,70 -dichloroﬂuorescein detection reagent) [44] Gingerols and slogaols in ginger [45] Glycosides in edible daylily ﬂowers [46] Glycosides in squill (sulfuric acid, antimony chloride, and anisaldehyde-phthalic acid detection reagents) [47] Hydrocarbons in honey (20 ,70 -dichloroﬂuorescein detection reagent) [48] 2-Hydroxycinnamaldehyde in cinnamons (detection by ﬂuorescence under 254 nm UV light) [49] Hydroxycinnamic acid esters in rye leaves (cellulose layer) [50] Lipids in dairy foods [51] Morantel related residues in milk [52] Myosmine in nuts [53] Phenols in olive oil [54] Phenols in tea (double development with two different mobile phases and sulfuric acidformaldehyde detection reagent) [55] Phenols in basil (ferric chloride detection reagent) [56] Phytosterol food additives [57] Polyphenols in tea (formaldehyde-sulfuric acid and anisaldehyde-sulfuric acid detection reagents) [58] Quercetin and kaempferol in various foods [59] Saffron secondary metabolites (aluminum oxide layer) [60] Sildenaﬁl in health foods (MS detection) [61] Triacylglycerols in sesame seeds (20 ,70 -dichloroﬂuorescein detection reagent) [62] Xanthophylls, chlorophylls, and carotenes in vegetables (C-18 bonded silica gel layer and detection by natural color) [63]

8.8 CONCLUSIONS AND FUTURE PROSPECTS Most preparative planar chromatography continues to be done today using CPLC, and most applications have been for pharmaceutical research and natural mixtures in plants (phytochemistry) [64]. A new book on CPLC [6] has recently been published containing detailed, comprehensive coverage of theory and fundamentals, techniques, and applications. This book should lead to much greater use of CPLC in the future relative to other preparative chromatography methods, such as HPLC, ﬂash chromatography, and gas chromatography. PLC will be applied to a wider assortment of analytes and food matrices with increased success because of the practical information available in this ﬁrst book devoted to PLC. More use of preparative silica gel taper plates with pre-adsorbent zone (Analtech), instrumental application of initial bands, automated ascending development, horizontal development, and detection and documentation of separated zones by densitometry is inevitable as their advantages are better appreciated. The availability of precoated plates with smaller particle size average and range would improve separations. A silica gel 1 mm preparative layer with 5 mm average particle size and 4.5–5.5 mm range (Mallinckrodt Baker, Phillipsburg, New Jersey) was shown to give better performance compared to other commercially available 1 mm PLC plates on the basis of theoretical plate number and resolution [65]. This layer [No. Si1000HPF-PA (19C)] was listed in the Mallinckrodt Baker 2004=2005 catalog but not in the online catalog of TLC products in

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 157 16.7.2008 11:24pm Compositor Name: MSubramanian

Preparative Layer Chromatography in Food Analysis

157

August 2006, so it must have been discontinued. A wider variety of stationary phases on commercial precoated PLC plates, such as other normal phase and reversed phase bonded phases, would also lead to wider use. The theoretical advantages of POPLC for resolution would suggest its wider use in the future relative to capillary ﬂow PLC. However, for this to happen, present-day OPLC instruments and procedures will have to become simpliﬁed and improved (e.g., high-quality commercial precoated plates for OPLC). Publication of additional practical POPLC techniques by the manufacturer and other current users to serve as examples would be an important catalyst leading to more general application for preparative purposes. The combination of parallel and serially connected layers for online POPLC separations, as envisioned by Nyiredy [21], is an interesting prospect for multidimensional FFPLC. Preparative RPC offers high separating power in terms of the amount of samples and number of compounds to be resolved, and good stationary phases are available. Again, additional published applications to serve as models for other users would be a great advantage in expanding applications in food analysis. Nyiredy [21] predicted the combination of PLC and preparative column LC, and such coupled or hyphenated techniques with their higher resolution capabilities would have signiﬁcant value in food analysis. Overall, because of better understanding of the classical method and the new developments in it and the forced ﬂow modes, PLC should be used in the future to an increasing degree for isolation and puriﬁcation of compounds from many classes in a wide assortment of food samples.

REFERENCES 1. Tuzimski, T., Two-stage fractionation of a mixture of pesticides by micropreparative TLC and HPLC, J. Planar Chromatogr.—Mod. TLC, 18, 39, 2005. 2. Molkentin, J. and Precht, D., Optimized analysis of trans-octadecanoic acids in edible fats, Chromatographia, 41, 267, 1995. 3. Arnold, A., Corain, E.A., Scaglioni, L., and Ames, J., New colored compounds from the Maillard reaction between xylose and lysine, J. Agric. Food Chem., 45, 650, 1997. 4. Brumley, W.C., Trucksess, M.W., Adler, S.N., Cohen, C.K., White, U.D., and Spohn, J.A., Negative ion chemical ionization mass spectrometry of deoxynivalenol (DON): Application to identiﬁcation of DON in grains and snack foods after quantitation=isolation by TLC, J. Agric. Food Chem., 33, 326, 1985. 5. Kozyra, M., Glowniak, K., Zabza, A., Zgorka, G., Mroczek, T., Cierpicki, T., Kulesza, J., and Mudio, I., Column chromatography and preparative TLC for isolation and puriﬁcation of coumarins from Peucadanum verticillare L. Koch ex DC, J. Planar Chromatogr.—Mod. TLC, 18, 221, 2005. 6. Kowalska, T. and Sherma, J. (Eds.), Preparative Layer Chromatography, CRC=Taylor & Francis, Boca Raton, FL, 2006. 7. Caballero-George, C. and Vlietinck, A.J., Inhibitory activity on binding of speciﬁc ligands to the human angiotensin II AT1 and endothelin 1 ETA receptors: Bioactive benzophenanthridine alkaloids from the root of Bocconia frutescens, Planta Med., 68, 770, 2002. 8. Bartnik, M., Glowniak, K., Maciag, A., and Hajnos, M.L., Use of reversed phase and normal phase preparative thin layer chromatography for isolation and puriﬁcation of coumarins from Peucadanum tauricum Bieb. leaves, J. Planar Chromatogr.—Mod. TLC, 18, 244, 2005. 9. Mossoba, M.M., McDonald, R.E., Armstrong, D.J., and Page, S.W., Hydrogenation of soybean oil: A thin layer chromatography and gas chromatography=matrix isolation=Fourier transform infrared study, J. Agric. Food Chem., 39, 695, 1991. 10. Aubourg, S.P., Sotelo, C.G., and Gallordo, J.M., Zonal distribution of fatty acids in albacore (Thunnus alalunga), J. Agric. Food Chem., 38, 255, 1990. 11. Hauck, H.E. and Schulz, M., Sorbents and precoated layers, in Preparative Layer Chromatography, Kowalska, T. and Sherma, J. (Eds.), CRC=Taylor & Francis, Boca Raton, FL, Chap. 3, 2006. 12. Morlock, G.E., Sample application and chromatogram development, in Preparative Layer Chromatography, Kowalska, T. and Sherma, J. (Eds.), CRC=Taylor & Francis, Boca Raton, FL, Chap. 5, 2006.

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 158 16.7.2008 11:24pm Compositor Name: MSubramanian

158

Handbook of Food Analysis Instruments

13. Dzido, T.H., Golkiewicz, W., and Pilat, J.K., The effect of temperature on the separation of some test solutes in preparative layer chromatography, J. Planar Chromatogr.—Mod. TLC, 15, 258, 2002. 14. Dzido, T.H., Modern TLC chambers, in Planar Chromatography, Nyiredy, Sz. (Ed.), Springer Scientiﬁc Publisher, Budapest, Hungary, Chap. 4, 2001. 15. Kong, L.-Y., Qin, M.-J., and Niwa, M., New cytotoxic bis-labdanic diterpenoids from Alpina calcarata, Planta Med., 68, 813, 2002. 16. Guedes, D.N., Silva, D.F., Barbosa-Filho, J.M., and Medeiros, I.A., Muscarinic agonist properties involved in the hypotensive and vasorelaxant response of rotundifolone in rats, Planta Med., 68, 700, 2002. 17. Mouly, P.P., Gaydou, E.M., Faure, R., and Estienne, J.M., Blood orange juice authentiﬁcation using cinnamic acid derivatives. Variety differentiations associated with ﬂavanone glycoside content, J. Agric. Food Chem., 45, 373, 1997. 18. Gomez-Cordoves, C., Bartolome, B., and Jimeno, M.L., Identiﬁcation of 2,3-dihydroxy-1-guaiacylpropan-1-one in brandies, J. Agric. Food Chem., 45, 873, 1997. 19. Mincsovics, E., Flowing wall processes in OPLC: Using segmentation of non-segmented adsorbent layer for single and parallel separations, J. Planar Chromatogr.—Mod. TLC, 17, 411, 2004. 20. Mincsovics, E., Katay, Gy., Ott, P.G., Kiraly-Veghely, Zs., Moricz, A.M., and Tyihak, E., Isolation of some antimicrobial compounds of red wine by OPLC ﬂowing eluent wall technique, Chromatographia, 62, S51, 2005. 21. Nyiredy, Sz., Preparative layer chromatography, in Handbook of Thin Layer Chromatography, 3rd ed., Sherma, J. and Fried, B. (Eds.), Marcel Dekker, Inc., New York, Chap. 11, 2003. 22. Tyihak, E. and Mincsovics, E., Overpressured layer chromatography (optimum performance laminar chromatography), in Planar Chromatography, Nyiredy, Sz. (Ed.), Springer Scientiﬁc Publisher, Budapest, Hungary, Chap. 8. 23. Davis, N.D., Cole, R.J., Dorner, J.W., Weete, J.D., Backman, P.A., Clark, E.M., King, C.C., Schmidt, S.P., and Diener, U.L., Steroid metabolites of Acremonium coenophialum, an endophyte of tall fescue, J. Agric. Food Chem., 34, 105, 1986. 24. Fatope, M.O., Adoum, A., and Takeda, Y., C18 acetylenic fatty acids of Ximerica americana with potential pesticide activity, J. Agric. Food Chem., 48, 1872, 2000. 25. Utzmann, C.M. and Lederer, M.O., Identiﬁcation and qualiﬁcation of aminophospholipid linked Maillard compounds in model systems and egg yolk products, J. Agric. Food Chem., 48, 1000, 2000. 26. Ramadan, M.F. and Morsel, J.T., Oil goldenberry (Physalis peruviana L.), J. Agric. Food Chem., 51, 4646, 2003. 27. Hazai, I. and Klebovich, I., Thin layer radiochromatography, in Handbook of Thin Layer Chromatography, 3rd ed., Sherma, J. and Fried, B. (Eds.), Marcel Dekker, Inc., New York, Chap. 12. 28. Kaykaty, M., Weiss, G., and Barbalas, M., Metabolism of the synthetic prostaglandin alfaprostol in the cow, J. Agric. Food Chem., 34, 688, 1986. 29. Zhang, L.Z., Khan, S.U., Akhtar, M.H., and Ivarson, K.C., Persistence, degradation, and distribution of deltramethrin in an organic soil under laboratory conditions, J. Agric. Food Chem., 32, 1207, 1984. 30. Sherma, J., Additional detection methods and removal of zones from the layer, in Preparative Layer Chromatography, Kowalska, T. and Sherma, J. (Eds.), CRC=Taylor & Francis, Boca Raton, FL, Chap. 8, 2006. 31. Bandyopadhyay, C., Gholap, A.S., and Mamdapur, V.R., Characterization of alkylrescorcinol in mango (Mangifera indica L.) latex, J. Agric. Food Chem., 33, 377, 1985. 32. Gopalakrishnan, M., Narayanan, C.S., and Granz, M., Identiﬁcation of some antibacterial constituents of New Zealand Manuka honey, J. Agric. Food Chem., 38, 10, 1990. 33. Steiner, I., Changes of a polyethylene foil for food packaging after sterilization with ozone, Dtsch. Lebensmittel-Rundsch., 87, 107, 1991. 34. Mokbel, M.S. and Hashinaga, F., Evaluation of the antioxidant activity of extracts from buntan (Citrus grandis Osbeck) fruit tissues, Food Chem., 94, 529, 2006. 35. Shillinc, D.G., Jones, L.A., Worsham, A.D., Parker, C.E., and Wilson, R.F., Isolation and identiﬁcation of some phytotoxic compounds from aqueous extracts of rye (Secale cereale L.), J. Agric. Food Chem., 34, 633, 1986. 36. Kaunzinger, A., Juchelka, D., and Mosandl, A., Progress in the authenticity assessment of vanilla. 1. Initiation of authenticity proﬁles, J. Agric. Food Chem., 45, 1752, 1997.

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 159 16.7.2008 11:24pm Compositor Name: MSubramanian

Preparative Layer Chromatography in Food Analysis

159

37. Kajimoto, G., Kanomi, Y., Tsushima, Y., Kamata, Y., Yoshida, H., and Shibahara, A., Inﬂuence of cholesterol on the deterioration of oil and decomposition of cholesterol migrating in frying oil, Nippon Eiyo Shokuryo Gakkaishi, 46, 167, 1993. 38. Takeoka, G.R. and Dao, L.T., Antioxidant constituents of almond [Prunus dulcis (Mill.) D.A. Webb], J. Agric. Food Chem., 50, 496, 2002. 39. Kweon, M.-H., Hwang, H.-J., and Sung, H.-C., Identiﬁcation and antioxidant activity of novel chlorogenic acid derivatives from bamboo (Phyllostachus edulis), J. Agric. Food Chem., 49, 4646, 2001. 40. Goda, Y., Nakanishi, T., Sakamoto, S., Sato, K., Maitani, T., and Yamada, T., Analyses of coloring constituents in commercial paprika by HPLC, J. Food Hyg. Soc. Jpn., 37, 20, 1996. 41. Gupta, A.P., Gupta, M.M., and Kumar, S., Simultaneous determination of curcuminoids in curcuma samples using high performance thin layer chromatography, J. Liq. Chromatogr. Relat. Technol., 22, 1561, 1999. 42. Witt, M.F., Hart, L.P, and Pestka, J.J., Puriﬁcation of deoxynivalenol (vomitoxin) by water-saturated silica gel chromatography, J. Agric. Food Chem., 33, 754, 1985. 43. Cerbulis, J., Parks, O., Liu, R., Protrowsky, E., and Farrell, H., Occurrence of diesters of 3-chloro1,2-propanediol in the neutral lipid fraction of goats’ milk, J. Agric. Food Chem., 32, 474, 1984. 44. Daultatabad, C.D., Mulla, G.M., and Mirajkar, A.M., Venolic and cyclopropenoic fatty acids in Piper nigrum seed oil, Fat Sci. Technol., 97, 453, 1995. 45. Chen, C.C., Kuo, M.C., Wu, C.M., and Ho, C.T., Pungent compounds of ginger (Zingiber ofﬁcinale Roscae) extracted by carbon dioxide, J. Agric. Food Chem., 34, 477, 1986. 46. Cichewicz, R.H. and Nair, M.G., Isolation and characterization of stelladerol, a new antioxidant naphthalene glycoside and other antioxidant glycosides from edible daylily (Hemerocallis) ﬂowers, J. Agric. Food Chem., 50, 87, 2002. 47. Verbiscar, A.J., Patel, J., Banigan, T.F., and Schatz, R.A., Scilloriside and other scilla compounds in red squill, J. Agric. Food Chem., 34, 973, 1986. 48. Bonaga, G., Grumanini, A.G., and Gliozzi, G., Chemical composition of chestnut honey: Analysis of the hydrocarbon fraction, J. Agric. Food Chem., 34, 319, 1986. 49. Kiridena, W., Miller, K.G., and Poole, C.F., Identiﬁcation of 2-hydroxycinnamaldehyde in the cinnamons of commerce, J. Planar Chromatogr.—Mod. TLC, 8, 177, 1995. 50. Strack, D., Engel, U., Weissenbock, G., Grotjahn, L., and Wray, V., Ferulic acid esters of sugar carboxylic acids from primary leaves of rye (Secale cereale), Phytochemistry, 25, 2605, 1986. 51. Olsson, N.U., Advances in planar chromatography for the separation of food lipids, J. Chromatogr., 624, 11, 1992. 52. Lynch, M.J., Burnett, D.M., Mosher, F.R., Dimmock, M.E., and Bartolucci, S.R., Determination of morantel related residues in bovine milk by electron capture gas chromatography, J. Assoc. Off. Anal. Chem., 69, 646, 1986. 53. Zwickenpﬂug, W., Meger, M., and Richter, E., Occurrence of the tobacco alkaloid myosmine in nuts and nut products of Arachus hypogea and Corylus avellana, J. Agric. Food Chem., 46, 2703, 1998. 54. Capasso, R., Evidente, A., and Scognamiglio, F., A simple thin layer chromatographic method to detect the main polyphenols occurring in olive oil vegetation, Phytochem. Anal., 3, 70, 1992. 55. Ferreira, D., Kamara, B.I., Brandt, E.V., and Joubert, E., Phenolic compounds from Cyclopia intermedia (honeybush tea), J. Agric. Food Chem., 46, 3406, 1998. 56. Jayasinghe, C., Botoh, N., Aoki, T., and Wada, S., Phenolics, composition and antioxidant activity of sweet basil (Ocimum basilicum L.), J. Agric. Food Chem., 51, 4442, 2003. 57. Johnsson, L. and Dutta, P.C., Characterization of side chain oxidation products of sitosterol and campesterol by chromatographic and spectroscopic methods, J. Am. Oil Chem. Soc., 80, 767, 2003. 58. Kamara, B.I., Brandt, E.V., Ferreira, D., and Joubert, E., Polyphenols from honeybush tea (Cyclopia intermedia), J. Agric. Food Chem., 51, 3874, 2003. 59. Bilyk, A. and Sapers, G.M., Distribution of quercetin and kaempferol in lettuce, kall, chive, garlic chive, leek, horseradish, red radish, and red cabbage tissues, J. Agric. Food Chem., 33, 226, 1985. 60. Iborra, J.L., Castellar, M.R., Canovas, M., and Manjon, A., TLC preparative puriﬁcation of picrocrocin, HTCC, and crocin from saffron, J. Food Sci., 57, 714, 1992. 61. Moriyasu, T., Shigeoka, S., Kishimoto, K., Ishikawa, F., Nakajima, J., Kamimura, H., and Yasuda, I., Identiﬁcation system for sildenaﬁl in health foods, J. Pharm. Soc. Jpn., 121, 765, 2001. 62. Nikolova-Damyanova, B., Velikova, R., and Kuleva, L., Quantitative TLC for determination of triacylglycerol composition of sesame seeds, J. Liq. Chromatogr. Relat. Technol., 25, 1623, 2002.

Otles/Handbook of Food Analysis Instruments 45660_C008 Final Proof page 160 16.7.2008 11:24pm Compositor Name: MSubramanian

160

Handbook of Food Analysis Instruments

63. Khachik, F., Beecher, G.R., and Whittaker, N.F., Separation, identiﬁcation, and quantiﬁcation of the major carotenoid and chlorophyll constituent in extracts of several green vegetables by liquid chromatography, J. Agric. Food Chem., 34, 603, 1986. 64. Waksmundzka-Hajnos, M. and Wawryzynowicz, T., Strategy of preparative separations by thin layer chromatographic methods, J. Liq. Chromatogr. Relat. Technol., 25, 2351, 2002. 65. Campbell, A.N. and Sherma, J., Comparative evaluation of precoated silica gel plates for preparative layer chromatography, Acta Chromatogr., 13, 102, 2003.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 161 16.7.2008 10:47pm Compositor Name: MSubramanian

Chromatography 9 Ion in Food Analysis William R. LaCourse CONTENTS 9.1

Introduction .......................................................................................................................... 161 9.1.1 Ions and Terminology.............................................................................................. 162 9.1.2 Theory of Ion Chromatography............................................................................... 163 9.1.2.1 Principles of Separation ........................................................................... 163 9.1.2.2 Properties of Ion-Exchange Phase ........................................................... 163 9.1.2.3 Inﬂuence of the Mobile Phase ................................................................. 165 9.2 Ion Chromatography Instrumentation ................................................................................. 166 9.2.1 Mobile Phase Reservoir and Solvent Delivery System .......................................... 167 9.2.1.1 Online Reagent Generation ...................................................................... 168 9.2.2 Sample Introduction Device .................................................................................... 168 9.2.3 Column .................................................................................................................... 169 9.2.4 Post-Column Apparatus .......................................................................................... 169 9.2.5 Detectors .................................................................................................................. 170 9.2.5.1 Refractive Index Detectors ...................................................................... 170 9.2.5.2 Absorbance-Based Detectors ................................................................... 170 9.2.5.3 Mass Spectrometric Detection ................................................................. 172 9.2.5.4 Electrochemical Detectors ........................................................................ 172 9.2.6 Data Collection and Output System ........................................................................ 182 9.2.7 Post-Detection Eluent Processing ............................................................................ 183 9.2.8 Connective Tubing and Fittings .............................................................................. 183 9.2.9 Related Separation Techniques................................................................................ 183 9.3 Ion Chromatography Applications to Food Analysis ......................................................... 184 9.3.1 Inorganic Anions and Cations ................................................................................. 185 9.3.2 Organic Acids .......................................................................................................... 188 9.3.3 Amines and Other Organic Bases ........................................................................... 189 9.3.4 Carbohydrates and Other Oligosaccharides ............................................................ 189 9.4 Future Directions ................................................................................................................. 194 Acknowledgment .......................................................................................................................... 194 References .................................................................................................................................... 194

9.1 INTRODUCTION The development of assays for foods is often hindered by the complicated nature of the samples. Following a plethora of sample preparation treatments to extract and partially fractionate the compounds of interest, chromatographic techniques are often used to further treat the sample in an attempt to isolate the analyte or analytes for eventual detection. Ions are separated using ion chromatography (IC), which is a type of liquid chromatography (LC), vide infra. To make matters 161

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 162 16.7.2008 10:47pm Compositor Name: MSubramanian

162

Handbook of Food Analysis Instruments

worse, many of the major components (e.g., inorganic ions, organic acids, organic bases, carbohydrates, oligosaccharides, and polysaccharides) of food samples have poor optical detection properties. Although signiﬁcant advances have occurred in the use of pre-injection and post-column chemical derivatization to produce photometrically or electrochemically active adducts, the simplicity of sensitive direct detection in IC will always be preferred whenever available. As such, conductivity is typically the detection mode of choice when either inorganic or organic ions enter the detector cell. Pulsed electrochemical detection (PED) is a revolutionary approach to the direct detection of numerous polar aliphatic compounds. This technique exploits the electrocatalytic activity of noble metal electrode surfaces to oxidize various polar functional groups. Electrode activity is maintained by the application of potential-time waveforms, which combine amperometric detection with online cleaning and reactivation. The full potential of PED is often best realized when combined with IC. In this chapter, we review IC in relation to food analysis, especially in regard to inorganic ions, organic acids and bases, and carbohydrates and related compounds. The fundamental aspects of IC and its instrumentation are reviewed, and an abridged summary of food applications in regard to ions are presented to highlight the analytical utility IC.

9.1.1 IONS AND TERMINOLOGY An ion can be deﬁned as an atom (monatomic), a group of atoms (polyatomic), or a molecule that has acquired a net electric charge by gaining or losing electrons from an initially neutral electronic conﬁguration. Ions with negative (gain of electrons relative to the neutral species) and positive (loss of electrons relative to the neutral species) charges are known as anions and cations, respectively. Anions are attracted to charges or electrodes (anodes) of positive charge due to coulombic attraction. Cations behave similarly, but they are attracted to charges or electrodes (cathodes) of negative charge. Cations are typically formed from metals and anions from nonmetals. Sodium chloride is an example of an ionic compound—a compound that contains both positively and negatively charged ions held together by coulombic (electrical) attraction to give an overall neutral charge. Ionic compounds are arranged in three-dimensional structures as solids, and these solids are said to dissociate into their component ions as they dissolve. Although water (H2O) is an electrically neutral molecule, water is a very effective solvent due to one end (the O atom) of the molecule being rich in electrons possessing a partial negative charge and the other end (the H atoms) having a partial positive charge. As an ionic compound dissolves, the ions are surrounded by H2O molecules that help to stabilize the ions in solution and prevent the cations and anions from recombining. In addition, because the ions and their shells of surrounding water are free to move about, the ions become dispersed uniformly throughout the solution. It is important to note that the overall solution must maintain electrical neutrality, or, in other words, the total charge of all the cations must be equal to the total charge of all the anions in solution. Anion and cations of strong acids, bases, or salts tend to dissociate (separation of preexisting ions) to nearly 100% in water, and they are known as strong electrolytes (e.g., chloride, sulfate, triﬂuoroacetate, sodium, and potassium ions). Weak acids and bases tend to ionize (the process of forming ions) in water. The extent to which these compounds ionize is reﬂected in their acid or base ionization constants, which are denoted as Ka and Kb, respectively. Water self-ionizes to form the hydronium (H3Oþ) and hydroxide ions (OH) with an ion product constant known as Kw. Ka, Kb, and Kw are all equilibrium constants, and by taking the –log of these values one obtains the pKa, pKb, and pKw, respectively. It is important to note that the relative fraction of ionized to respectively unionized species can be controlled through pH, which will play an important role both in the separation and detection of ions. Ions can be singly charged (monovalent) or multiply charged (polyvalent). Mixtures of analytes with different valence states can complicate the separation process due to the broad range of

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 163 16.7.2008 10:47pm Compositor Name: MSubramanian

Ion Chromatography in Food Analysis

163

coulombic attraction between the different analytes for the stationary phase. Zwitterionic molecules contain both cationic and anionic functional groups. Amino acids are good examples. They contain both ammonium=cationic and carboxylic acid=anionic functional groups. Since each functional group of a zwitterionic compound can exist in a different state of percent ionization depending on the pH, the overall charge of the zwitterions represents a compilation of these groups. The pI of a zwitterionic compound represents the pH at which the molecule has no net charge. Since the –log [I] is not real, the pI is a contrived term. If pH > pI, the overall charge is negative and vice versa.

9.1.2 THEORY OF ION CHROMATOGRAPHY Liquid chromatography, vide infra, is a physical separation technique in which substances in a mobile phase are separated because of their different afﬁnities for a stationary phase. The mobile phase is a solution that is driven by some type of pumping device through a column that is packed with a material known as the stationary phase. An injection valve is used to introduce a dissolved sample into the mobile phase before the analytical column. As the sample enters the column, the constituent molecules and ions in the mobile phase interact to varying degrees with the stationary phase. The greater the afﬁnity of the sample constituents for the stationary phase, the longer a substance is retained and vice versa. Separation in liquid chromatography is based on physical interactions, which means that the original species are unaltered in the separation process. Finally the eluted species are passed through a detector, which converts the amount (e.g., concentration) of the sample constituents to an electrical signal that can be output as data. The separation of ions in a dissolved sample is accomplished using ion chromatography or ion-exchange chromatography. Ion-exchange chromatography has been used to separate amino acids in their ionic forms since 1956 [1]. 9.1.2.1

Principles of Separation

Ion chromatography uses ion-exchange resins to separate charged species (i.e., atomic or molecular ions) based on their coulombic interaction with the resin. Coulombic interaction is the physical process of the attraction of opposite charges and repulsion of like charges over distance. To effect a separation, a stationary phase capable of ion exchange must have electric charges on its surface. þ þ Hence, ionic groups (e.g., SO2 3 , COO , NH3 , or NR3 ) are incorporated into the resin or gel. Figure 9.1a shows a cartoon representation of a stationary phase capable of separating anions (e.g., Cl). Electroneutrality is maintained by counterions in the mobile phase, which is continually ﬂowing at a particular rate over and through the stationary phase. Upon injection, ionic sample molecules compete with ions in the mobile phase for sites on the surface of the stationary phase (see Figure 9.1b). This interaction for the most part is a physical, charge–charge interaction. A sample ion that competes ineffectively, or has a weak interaction, is only slightly retained if at all, and those with a strong interaction are retained longer. In fact, sample ions can be retained so strongly that an increase in concentration (gradient) of, or the addition of, a ‘‘pusher’’ ion to the mobile phase may be required to dislodge the sample ion from the column sites and push them off the column. Figure 9.1c shows the analyte leaving the column and passing on to the detector. 9.1.2.2

Properties of Ion-Exchange Phase

Table 9.1 lists the most common ion-exchange resins. All the listed functional groups can be bonded to either silica or an organic resin (e.g., styrene divinyl benzene or S-DVB). Silica-based phases suffer from limited chemical stability, in that they can only be used over a deﬁned pH range of 2–8, which can be a drawback in IC. The capacity of a resin is typically given in milliequivalents per gram (meq=g) of resin. Although many commercial resins have ion-exchange capacities greater than ~3 meq=g, the bulk resins employed in ion chromatography usually use low-capacity ion exchangers (0.01–0.2 meq=g) [2]. Resin-based phases typically have higher capacities per gram than silica-based phases, but silica has a higher density, which results in similar capacities per unit volume.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 164 16.7.2008 10:47pm Compositor Name: MSubramanian

164

Handbook of Food Analysis Instruments Sample ion in the mobile phase

OH –

Counterion in the mobile phase OH–

CI– OH– +

NR3

OH –

OH –

+

NR3

+

NR3

(a)

Ion-exchange resin OH –

OH– OH– OH –

CI–

OH – +

+

NR3

+

NR3

NR3

(b)

Ion-exchange resin OH–

CI– OH– OH –

OH– +

+

NR3

NR3

(c)

OH – +

NR3

Ion-exchange resin

FIGURE 9.1 Diagram of the ion-exchange process on an anion-exchange resin. (a) The sample ion (i.e., Cl) is injected onto the resin, which is in equilibrium with the counterions (e.g., OH) in the mobile phase. (b) The negatively charged Cl interacts with the stationary phase in competition with the counterions in the mobile phase. (c) The sample ion elutes from the column at a rate reﬂective of its interaction with the resin, which manifests itself as retention time.

TABLE 9.1 Summary of Ion-Exchange Resins Exchanger

Type

Resin

Description

Functional Groups

Cation

Strong Weak Strong

SA CM QA QAE AE DEA DEAE DMAE PEI

Sulfonic acid Carboxymethyl Quaternary amine Quaternary aminoethyl Aminoethyl Diethylamine Diethylaminoethyl Dimethylaminoethanol Polyethyleneimene

–SO 3 –CH2–COO þ –NCHþ 3 or –NR3 –CH2–CH2–N(CH2 CH3 )þ 3 –CH2–CH2–NHþ 3 –NH(CH2 CH3 )þ 2 –CH2–CH2–NH(CH2 –CH3 )þ 2 –O–CH2–CH2–NH(CH3 )þ 2 –(NH–CH2–CH2–)n–NHþ 3

Anion

Weak

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 165 16.7.2008 10:47pm Compositor Name: MSubramanian

165

Ion Chromatography in Food Analysis

–NR2H+ Exchange capacity

–COO⫺

–NR2

–COOH

1

2

3

4

5

6

7

8

9

10

11

12

pH

FIGURE 9.2 Ion-exchange capacities of (——) WCX, pH 4.5; (– – –) WAX, 9.0; (– – –) SCX; and ( SAX resin types versus pH of the mobile phase. .

.

)

. . . .

Strong cation and anion exchangers, or SCX and SAX, respectively, maintain their charges and ion-exchange capacity over the entire pH range that is typical in IC (i.e., 2–10). These phases are used for the separation of weak acids and bases, where retention and selectivity are established by the extent of ionization (or overall charge) of the analytes of interest, which is pH dependent. In contrast to the strong ion exchangers, the capacity of weak ion exchangers (cation, WCX and anion, WAX) is dependent on the pH of the mobile phase due to ionization of the stationary phase functional groups. For instance, at a pH signiﬁcantly lower than the pKa of a WCX resin, the resin is fully protonated, and its capacity to retain sample ions is near zero. On the other hand, accurate control of the mobile phase pH near to the pKa of the resin will allow both retention of the analyte and a high degree of selectivity versus other analyte species. Alternatively, weak anion exchangers (WAX) retain at pH’s lower than the pKa of the functional group of the resin. Figure 9.2 shows the normalized ion-exchange capacities of each of the four resin types versus pH of the mobile phase. Most applications in IC use strong ion exchangers. Weak ion exchangers are primarily used for biological samples where a high degree of selectivity or reduced retention of polyvalent species is required. 9.1.2.3

Inﬂuence of the Mobile Phase

The retention of an ion can be expressed by the following equations: R Naþ þ Xþ Ð R Xþ þ Naþ

(cation-exchange process)

Rþ Cl þ X Ð Rþ X þ Cl

(anion-exchange process)

where R (cation exchanger)=Rþ (anion exchanger) is the stationary phase Xþ (cation)=X (anion) is the sample ion Naþ=Cl are the counterions in the mobile phase Numerous equilibria between the mobile phase constituents and the stationary phase are occurring simultaneously, and changes in pH or ionic strength will affect the retention of ions, especially for analytes with different charge states.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 166 16.7.2008 10:47pm Compositor Name: MSubramanian

166

Handbook of Food Analysis Instruments

The factors that control the retention of an ion and subsequent separation of a mixture of different ions are as follows: (1) type of ion exchanger of the stationary phase; (2) composition of the mobile phase including its pH, ionic strength (concentration of ions), type of counterions, and organic modiﬁer content; and (3) temperature. These parameters must be carefully selected and optimized to effect a separation. For a particular column type (i.e., SCX, SAX, WCX, and WAX), the following generalities about the mobile phase can be made to assist in the development of an IC method. .

.

. .

. .

pH (strong ion-exchange systems)—An increase in pH will lead to greater sample ionization of acids and their subsequent retention in SAX chromatography and, conversely, a decrease in pH favors the retention of bases in SCX systems. pH (weak ion-exchange systems)—An increase in pH will lead to less stationary phase capacity in WAX chromatography and elution will be faster. Similarly, a decrease in pH leads to increased protonation of the WCX stationary phase, which results in reduced analyte interaction and shorter retention times. Counterion concentration—An increase in counter concentration in the mobile phase tends to reduce retention in IC. Counterion type—Counterions vary in charge, hydrodynamic radius, and polarizability (i.e., ease of dipole induction). Hence, ions can be designated as either strong or weak displacers or ‘‘pushers’’ of analyte ions. In anion-exchange chromatography, the strength of counterions is F (weak) < OH < acetate < Cl < SCN < Br < CrO 4 < NO3 < I 3 < oxalate < SO4 citrate (strong). In cation-exchange chromatography, the strength of þ þ þ þ 2þ < Zn2þ counterions is Liþ (weak) < Hþ < Naþ < NHþ 4 < K < Rb < Cs < Ag < Mg 2þ 2þ 2þ 2þ 2þ 2þ 2þ < Co < Cu < Cd < Ni < Ca < Pb < Ba (strong). In general, a strong displacer ion reduces analyte retention more than an isomolar concentration of a weaker displacer ion. It is important to note that the type of counter can also affect selectivity. Organic solvents—Organic solvents in the mobile phase tend to decrease retention in IC. Common solvents are methanol and acetonitrile. Temperature—Coulombic interactions tend to decrease as temperature is increased, as such, retention time of analytes decreases with increases in column temperature. Simultaneously, higher column temperatures can be used to improve mass transfer, which may lead to improve chromatographic efﬁciencies.

Although not as prominent, other modes of separation (e.g., adsorption processes or steric=size effects) may play a role in the separation of the sample ions. With so many factors, the elution order of ions can sometimes be difﬁcult to predict. Commercially available software is available that can assist method developers predict the retention of ions under various conditions. Dionex (Sunnyvale, California) offers its users Virtual Column Separation Simulator, which models all of their most popular anion-exchange and cation-exchange columns, using retention data for hundreds of analytes. Other chromatography companies offer similar products.

9.2 ION CHROMATOGRAPHY INSTRUMENTATION Modern high-performance ion chromatography uses high pressure to force the mobile phase and an analyte through a closed column packed with micron-size particles, which constitute the stationary phase. LC, or IC, instrumentation is made up typically of nine basic components: mobile phase= solvent reservoir, solvent delivery system, sample introduction device, column, post-column apparatus, detector, data collection and output system, post-detector eluent processing, and connective tubing and ﬁttings. Although all components except for the post-column apparatus are essential to

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 167 16.7.2008 10:47pm Compositor Name: MSubramanian

167

Ion Chromatography in Food Analysis

#1

#2

Active temperature control

#3

Guard column 1.2 mL/min Analytical column Solvent delivery system

Injector or autosampler Post-column suppressor unit

Output device Conductivity detector

Control Data

To waste or fraction collector

FIGURE 9.3 Schematic of a typical LC system outﬁtted to perform ion chromatography.

performing IC, the use of post-column suppression often offers signiﬁcantly improved performance. Figure 9.3 shows a schematic diagram of a generic ion chromatographic system.

9.2.1 MOBILE PHASE RESERVOIR

AND

SOLVENT DELIVERY SYSTEM

The mobile phase reservoir can be any clean, inert container. It usually contains from 0.5 to 2 L of solvent, and it should have a cap that allows for a tubing inlet line, which feeds mobile phase to the solvent delivery system. The cap also serves to keep out dust, reduce solvent evaporation, allow for pressurization of the bottle, offer ports for additional inlet lines, and sparging (i.e., dispersing He or Ar into the mobile phase to reduce dissolved air). All mobile phases=solvents should be freshly ﬁltered and preferably degassed. Online degassers, which are primarily used to remove small gas bubbles and reduce dissolved air, are now popular additions to many IC systems, and they eliminate the need to degas the mobile phase off line. An additional ﬁlter is often placed at the end of the mobile phase inlet line to remove any precipitants that may form in the mobile phase during its use. Sparging control and the ability to blanket the solvents with inert gases is highly recommended to eliminate carbonate formation in alkaline solvents and to maintain extremely low levels of dissolved O2, when performing electrochemical or ﬂuorescence detection. The high-pressure pump can operate at pressures from 500 to 5,000 psi. The purpose of the pump is to deliver a precise, accurate, reproducible, constant, and pulse-free ﬂow of mobile phase to the column. Three major classes of IC pumps are currently in use: constant pressure pumps, syringebased or displacement pumps, and constant ﬂow pumps. Both constant pressure and syringe-based pumps are not easily adapted to gradient solvent delivery; hence, their use has generally been restricted to post-column reagent addition. The majority of commercial high-pressure pumps available today are designed around a simple reciprocating piston pump. The rotational energy of a motor is transferred into the reciprocal movement of the piston by an eccentric cam or gear. The piston is driven in and out of a solvent chamber in the pump head, which typically has a volume of 10–100 mL. A pair of check valves control the direction of ﬂow through the pump head. A piston seal keeps the mobile phase from leaking out of the pump head. With a twin-head reciprocating pump, two pump heads operate simultaneously but at 1808 out-of-phase with each other. As a result, mobile phase ﬂows to the column 100% of the time. The twin-head design gives

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 168 16.7.2008 10:47pm Compositor Name: MSubramanian

168

Handbook of Food Analysis Instruments

essentially pulseless ﬂow as compared to the single head design. To fully accommodate the pH ranges and high salt concentration utilized in IC, ﬂow paths including pump heads require the use of polymeric materials (e.g., PEEK). Pistons, piston seals, and check values also need to be pH- and corrosion-resistant. Many separations can be done isocratically, which means that eluent being delivered to the column is not changing in composition over the course of the separation. For more complex separations, gradient elution is required. Most commonly, gradient elution is performed by altering the proportion of the eluents over the course of the separation. In doing so, the early eluting compounds remain well resolved, while the more highly retained compounds elute quicker. Gradient elution is simply the programming, or changing of the solvent strength over the course of a separation. A gradient can be linear, convex, concave, stepped, or a complex sequence of each to achieve the desired separation. Computer-controlled pumping is required to generate a gradient ﬂow. The following are the predominating approaches used to produce a gradient ﬂow: (1) proportioning valve, which is controlled by a microprocessor, regulates the amount of up to four eluents and the eluent mixture is sent to the high-pressure pump and (2) delivery of eluents from multiple high-pressure pumps, which are controlled individually by a programming device, are mixed together in a high-pressure mixing chamber after the pump. The former approach is known as low-pressure mixing and the latter as high-pressure mixing. Low-pressure mixing is less expensive than high-pressure mixing since only one high-pressure pump is used versus the two or more required for high-pressure mixing, and the maintenance of one pump is much easier than the maintenance of two or more pumps. The main problem with low-pressure mixing is that it is more susceptible to bubble formation because the solvents are being mixed at atmospheric pressure. Hence, the common use of online degassers. 9.2.1.1

Online Reagent Generation

More recently, manufacturers are offering automated eluent generation. Counterions of either potassium or methanesulfonate diffuse across a semipermeable membrane into a high-pressure chamber in which water is being electrolyzed into OH and Hþ to form KOH (for anion separations) or methanesulfonic acid (for cation separations). By-products of electrolysis include hydrogen or oxygen gas, which are removed by a built-in degasser. Eluent concentrations of KOH and methanesulfonic acid are a function of the applied voltage, eluent ﬂow rate, concentration of the species in the counterion reservoir, and other factors. These factors are controlled by the associated software, which allows the user to simply ‘‘dial in’’ a desired concentration. The online eluent generators are placed after the pump, and the instantaneous process of electrolysis allows for changes of eluent concentration (i.e., gradients) during the chromatographic run. The automatic generation of high-quality IC eluents on demand offers the following beneﬁts: (1) longer pump life, (2) no eluent preparation, (3) improved reproducibility, and (4) better gradient performance. The use of eluent generation has been determined to be compliant with EPA requirements.

9.2.2 SAMPLE INTRODUCTION DEVICE In many instances, the limiting factor in the precision of an IC system lies in the reproducibility of the sample introduction system. The sample introduction device, also known as a sample injector, is used to introduce the sample into the IC system without depressurizing it. The most widely used method of sample injection is based upon a sample loop that can be placed in and out of the mobile phase ﬂow path by merely switching a valve. When the valve is in the load position, the sample loop is ﬁlled at atmospheric pressure. Sample sizes often range from 5 to 500 mL. For best results, an excess of sample (i.e., two to ﬁve times the injection volume) is ﬂushed through the loop to ensure that no trace of the previous sample remains. By turning the valve from the load to the inject position, the sample loop is connected to the high-pressure mobile phase stream and the sample is

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 169 16.7.2008 10:47pm Compositor Name: MSubramanian

Ion Chromatography in Food Analysis

169

then carried to the column. It is imperative that the rotor seals in the injection valve be pH tolerant. Often different seal materials are required for highly basic versus highly acidic mobile phases. The valve-based sample introduction system is easily automated using simple robotic technology. The use of autoinjectors not only improves injection reproducibility, but they allow for the continuous processing of numerous (i.e., tens to thousands) samples at a time. Autoinjectors have also been used for the implementation of pre-column derivatization protocols, especially for amino acid analyses. Pre-column derivatization is used to improve the chromatographic or detection properties of analytes. Autoinjectors may include temperature control of the sample chamber.

9.2.3 COLUMN The column is the part of the IC in which the separation occurs. IC columns are mainly constructed from plastic, especially PEEK. If pH compatibility is not an issue, columns are sometimes constructed from titanium and smooth-bore stainless steel. Common dimensions for analytical scale columns are in the range of 10–30 cm long and 2–10 mm i.d. The common particle sizes of packings are 3, 5, and 10 mm. Columns of the above dimensions often have efﬁciencies of 40,000–60,000 plates per meter. The current trend has been the use of higher performance, highspeed columns, which have smaller dimensions then those described above. Such columns have efﬁciencies of 1,00,000 plates per meter and have the advantage of speed and minimal solvent consumption. Hundreds of packed columns in differing size and packing material are available from numerous manufacturers. It is important to read the manufacturers’ literature relating to the maintenance, handling, and limitations of the column (e.g., silica-based columns are only compatible with pH values from 2 to 7). In addition to chemical limitations of the packing material, columns are easily degraded by the irreversible adsorption of impurities from samples and solvents. Hence, a guard column is often used to protect the integrity of the analytical column, which is much more expensive. Also, for analytes, which may contain particulates, an in-line ﬁlter can be placed between the injector and guard column. It should be noted that with the addition of each component after the injector, the efﬁciency of the separation is degraded. Hence, judicious use of in-line devices is necessary. For many applications, close control of column temperature is not necessary and IC separations are performed under ambient conditions. However, temperature control can enhance chromatographic reproducibility and afford opportunities to improve separation efﬁciency. In addition, conductivity detection is signiﬁcantly improved using active temperature control. Modern instruments can be equipped with column heaters=ovens that control column temperature to a few tenths of a degree from ambient to 1508C.

9.2.4 POST-COLUMN APPARATUS If post-column modiﬁcation of the mobile phase or analyte is required, then the system will have a post-column apparatus in the ﬂow path. Modiﬁcation of the mobile phase (e.g., addition of buffers, changing the pH, and solvent strength) may be needed to enhance the compatibility of the mobile phase with the detector, while post-column derivatization (commonly used for amino acid and transition metal analyses) of the analyte may be needed to improve the detection properties of the analytes after their separation. In either case, a post-column addition system consists simply of a reagent delivery pump, a mixing tee, and a mixing coil. Typically, a pressurized reservoir is used to deliver a pulseless ﬂow of the reagent. The vessel is usually ﬁtted with a check value to prevent reagent backup. The major drawback of using a pressure-based delivery system is that it cannot handle a great deal of system back-pressure. Hence, close attention must be paid to the minimization of post-column back-pressure sources. Any single piston pump, even with extensive pulse dampening, is usually inadequate for high sensitivity work. Delivery of the post-column reagent to the chromatographic eluent ﬂow is accomplished via a mixing tee. The mixing tee should be a low-dead-volume ﬁtting. Probably the most crucial

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 170 16.7.2008 10:47pm Compositor Name: MSubramanian

170

Handbook of Food Analysis Instruments

component in the post-column system is the mixing coil, which connects the mixing tee to the detector. It is essential that the mixing coil produces a homogeneous solution in the most efﬁcient manner, in other words, with minimal band-broadening. The best mixing is obtained with a woven=knitted reaction coil. The three-dimensional weave achieves efﬁcient mixing and effectively reduces band-broadening effects by preventing laminar ﬂow patterns. In addition, their open-tubular nature produces less back-pressure than packed-bed reactors, and woven reactors are easy to make using commercially available Teﬂon tubing.

9.2.5 DETECTORS By passing the column efﬂuent through the detector, some chemical or physical property of the analyte is converted into an electrical signal, and the solutes are monitored as they are eluted from the column. The electrical signal, which can be ampliﬁed and manipulated by suitable electronics, is proportional to the level of some property of the mobile phase or solutes. Chromatographic detectors are classiﬁed as either bulk property detectors, which respond to a bulk property of the eluent such as refractive index or conductivity, or solute property detectors, which respond to some property of the analyte such as UV absorbance. In either case, the response of the detector is modulated by the presence and amount of the analyte. Solute property detectors tend to be more sensitive than bulk property detectors, on the order of 1000 times or more. Ideal characteristics of any detector are high sensitivity, good stability, linearity, short response time, reliability, nondestructiveness, ease of use, and low dead volume. Many types of analytical techniques have been applied to IC with varying degrees of success. In ion chromatography, electrochemical detection approaches dominate with conductivity being the most popular due to the inherent conductivity of ions. Many metal ions are electroactive and lend themselves to amperometric detection, and polar aliphatic compounds (e.g., carbohydrates) are amenable to PED. Optical detection methods are used for ions with natural chromophores and for systems that use post-column derivation to enhance the detection properties of transition metals. Only a brief review of the most common detectors is presented here. 9.2.5.1

Refractive Index Detectors

Refractive index (RI) detectors monitor the difference in refractive index between the column eluent containing analyte and a reference stream containing mobile phase only. These detectors are the closest to universal detectors in HPLC and IC because any solute can be detected as long as its refractive index is different from that of the mobile phase. Numerous work has been performed using RI detection in IC [3,4]. RI offers a wide latitude in the selection of mobile phase, eluent pH, and ionic strength. Unfortunately, the sensitivity of RI detection is poor due to the small differences in the absolute refractive indices of many substances commonly analyzed by IC. RI detection is also sensitive to both temperature and pressure changes, and, as a consequence, strict temperature control of the detector and pulseless ﬂow are mandatory. Since RI is a bulk property detector, it is sensitive to changes in the mobile phase composition, as well as analyte concentration, and it is not amenable to gradient elution without signiﬁcant additional effort and complexity. 9.2.5.2

Absorbance-Based Detectors

In UV=vis absorbance detectors, the mobile phase is passed through a small ﬂow cell, where the radiation beam of a UV=vis photometer or spectrophotometer is located. As a UV-absorbing solute passes through the ﬂow cell, a signal is generated that is proportional to the solute concentration. Absorption of radiation is a function of concentration, c, as described by the Beer–Lambert law: A ¼ « bc

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 171 16.7.2008 10:47pm Compositor Name: MSubramanian

171

Ion Chromatography in Food Analysis

where A is the absorbance « is the molar extinction coefﬁcient b is the ﬂow cell path length UV-absorbing compounds (e.g., alkenes and aromatics) are typically compounds that have multiple bonds between C and O, N, or S. Compounds that absorb light at visible wavelengths require analytes to possess extended conjugation systems or color complexes. The mobile phase components should be selected carefully so that they absorb little or no radiation. Although alkali metals are not detected by UV, numerous anions (see Table 9.2) absorb at lower wavelengths [5]. Bromate, iodide, and nitrate absorb at longer wavelengths, and several methods have been described with detection limits at the nanogram level. Also, nitrite and nitrate in drinking water and cured meat have been reported using direct UV absorption [6]. Aromatic acids absorb well, and methods for detecting these ions are quite common. The utility of direct UV absorbance detection is limited by the lack of speciﬁcity at lower wavelengths, which leads to complex chromatograms and extensive sample cleanup. More importantly, several important anions (e.g., sulfate, chloride, and phosphate) are not detected. Another approach is to add a UV=vis-absorbing counterion to the mobile phase. By monitoring a wavelength where the eluent absorbs, a negative peak is observed in the chromatogram where a sample ion elutes. Unfortunately, these additional chromatographic equilibria further complicate the method to a point where optimizing the detection process reproducibly is often difﬁcult. UV absorbance methodology can be expanded by the exploitation of the absorption of metal chloride complexes (lmax 335 nm) formed before injection. This approach has been used extensively to detect metal ions in solution [7]. Metal EDTA complexes can also be used [8]. A classic approach for the determination of many metal ions is with the use of post-column color-forming reagents, especially 4-(2-pyridylazo)-resorcinol (PAR). PAR works with 32 metals (e.g., Pb(II), Hg(II), Cd(II), Cu(II), Fe(II), Fe(III), Bi(III), Zr(IV), Th(IV)) [9]. In addition to PAR, Arsenazo I and Arsenazo III have been used to visualize metals. A spectrophotometric method of

TABLE 9.2 Ions for Direct Absorbance Detection and by Complex Formation Ion or Ion Complex Wavelength (nm) 190–199 200–209 210–219

220–249 250–299

Direct Absorbance Br, BrO 3 , Cl , ClO2 , ClO3 , HCOO , CH3COO , 3 2 2 2 IO3 , N3 , PO4 , SCN , SeO3 , SO3 , SO4 3 2 AsO3 4 , AsO3 , Br , BrO3 , C2 O4 , citrate, CN 2 2 CNO, IO , MoO , NO , NO , S O 2 3 4 2 3 3 Br, Au(CN) 2 , I , metal–Cl complexes, metal– CN complexes, metal–EDTA complexes, organoarsenic acids, S 2 Cr, Pt, Au–Cl complexes, I

300–349 350–400

Fe(ClO4)3 Color-Forming Complex

CrO2 4 , Cr(III)

3 2 S 2 , PO3 , H2 PO2 , IO3 , CO3 , Br , CN , 2 2 B4 O7 , BrO3 , SiO3 3 2 CrO24 , SCN, Fe(CN)4 6 , Fe(CN)6 , SO4 , Cl , 4 2 3 2 P2 O7 , S2 , S2 O8 , PO4 , SO3 NO 2

Source: Adapted from tables in Fritz, J.S. and Gjerde, D.T., Ion Chromatography, 3rd ed., Wiley-VCH, New York, 2000.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 172 16.7.2008 10:47pm Compositor Name: MSubramanian

172

Handbook of Food Analysis Instruments

detection of many inorganic anions was developed by Imanari et al. [10] using a post-column reactor that mixed a stream of ferric perchlorate under acidic conditions. Ferric perchlorate is essentially colorless, but most anions will complex the iron and form colored species that can be detected at 330–340 nm, see Table 9.2. Unfortunately, nitrate, ﬂuoride, and chlorate ions are not detected by this method. In addition to the highly popular detectors mentioned above, numerous other spectrometric detectors have been used in IC. Flame photometric has been used for alkali and alkaline earth metals, and atomic absorption (AA) for a variety of arsenic species. Both of these approaches are very sensitive (~10 ppb limits of detection). 9.2.5.3

Mass Spectrometric Detection

Owing to innovations in LC-mass spectrometry (MS) interfacing, MS has become a very important HPLC detector because of its ability to generate structural and molecular weight information about the eluted solutes. The combination of IC and mass spectrometry allows for both separation and identiﬁcation in the same step, an advantage few of the other detectors provide. Three major problems must be addressed when interfacing MS with IC. .

. .

Flow rate in IC with conventional 4.6 mm i.d. columns is often >1 mL=min, which is much larger than the ﬂow that can be taken by the conventional mass spectrometer vacuum systems. MS has difﬁculty vaporizing nonvolatile and thermally labile molecules without degrading them. Mobile phase constituents found in IC eluents are often not compatible with MS.

In recent years, the technical problems have been addressed by using small bore columns and postcolumn desalting systems. IC-MS allows for the determination of ionic and polar compounds with greater conﬁdence, in that MS may provide oxidation state speciation, molecular mass identiﬁcation, isotope ratios, and fragmentation patterns (for ﬁngerprinting) while avoiding coelution and background interferences. 9.2.5.4

Electrochemical Detectors

Electrochemical detectors can be classiﬁed according to the three fundamental parameters of voltage or potential (E), resistance (R), or current (i). These terms are related via Ohm’s law, which is E ¼ iR. The conductance of a solution (G) is the inverse of resistance (G ¼ 1=R). Potentiometric detectors measure potential in volts (V) under conditions where i essentially equals zero; conductimetric detectors measure solution conductance in siemens (S); and amperometric detectors measure current in amperes (A) as a function of applied potential. Conductivity detectors have the potential (pun intended) to measure all ions, while potentiometric and amperometric detectors are more selective. Electrochemical detectors offer many advantages such as high sensitivity, high selectivity, and wide linear range. They are easily adapted to microchromatographic and electrophoric separation systems due to the response being dependent on electrode area and not pathlength as in optical absorbance methods. The detectors are often simple, rugged, and relatively inexpensive. Unfortunately, electrochemical detectors are also sensitive to ﬂow rate, mobile phase, or eluent constituents including dissolved oxygen, which is difﬁcult to eliminate or control. Conductivity detection is least affected, especially using eluent suppression technology, which has facilitated its popularity. Potentiometric techniques often suffer from slow response times, and they are sometimes difﬁcult to use due to their heterogeneous detection process. In other words, analytes must diffuse to an electrode surface in potentiometric and amperometric techniques, which can lead to ‘‘poisoning’’ of the electrode surface. Many amperometric techniques require daily polishing of the electrode surface. PED is designed to mitigate these types of problems.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 173 16.7.2008 10:47pm Compositor Name: MSubramanian

173

Ion Chromatography in Food Analysis

9.2.5.4.1 Potentiometric Detection Potentiometric detectors typically measure the potential difference (DE) across a membrane that originates from the difference in analyte concentration in the eluent versus an internal reference solution. The most common potentiometric measuring device is a pH electrode, in which a glass membrane responds to hydronium ion concentration in the test solution. Other ion-selective, or indicator, electrodes are also available commercially. The attribute of an indicator electrode to be highly selective for a particular species is also its drawback, in that a different electrode is needed for each type of ion. Halides and sulfates can be monitored using silver=silver salt and lead=lead salt electrodes, respectively [11]. 9.2.5.4.2 Conductimetric Detectors Conductivity detectors are used primarily to detect ions in conjunction with IC. This type of detector monitors the ability of the eluent to conduct electricity. To understand conductivity detection (CD), it is important to understand how individual ions contribute to the total conductance of the solution. The conductivity of a dilute solution is the sum of the conductivity of all the ions in solution multiplied by their concentrations. In other words, conductivity is proportional to concentration, which is summarized in Kohlrausch’s law of independent migration; ki ¼

Si loi ci 1000

where ki (S=cm) is the measured conductivity ci (eq=L) is the concentration loi (Scm2=eq) is the ionic limiting equivalent conductivity of each speciﬁc ion Table 9.3 lists the limiting equivalent conductivities for several organic and inorganic ions in aqueous solution at 258C [12]. As the concentration increases beyond 1 mM, the direct proportionality between conductivity and concentration is generally lost. For example, the equivalent conductivity at 258C of KCl at inﬁnite dilution versus 1 mM is 149.9 and 146, respectively, which is a decrease of ~2% [13]. If the electrolyte is a weak acid or base with only partial dissociation, then ci is replaced by the concentration of only the dissociated ions. TABLE 9.3 Partial List of Limiting Equivalent Ionic Conductances at 258C Cations Hþ Liþ Naþ Kþ NHþ 4 Mg2þ Ca2þ Ba2þ Cu2þ Fe3þ Ce3þ CH3 NHþ 3 N(CH3 CH2 )þ 4

l8i (Scm2=eq)

Anions

l8i (Scm2=eq)

350 39 50 74 73 53 60 60 55 68 70 58 33

OH F Cl Br I NO 3 HCO 3 CO2 3 SO2 4 PO3 4 Formate Acetate Benzoate

198 54 76 78 77 71 45 72 80 69 55 41 32

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 174 16.7.2008 10:47pm Compositor Name: MSubramanian

174

Handbook of Food Analysis Instruments

The limited equivalent conductivity of an ion, l8i, reﬂects the mobility of an ion, which is affected by the properties of the solvent (e.g., viscosity). Ions with large hydration spheres (e.g., l8chloride ¼ 76) are less mobile, and therefore less conductive than ions with small hydration spheres (e.g., l8ﬂuoride ¼ 54). Fluoride is known to be less hydrated. One of the most important factors is temperature. Conductivity increases ~2% per 8C. Hence, it is necessary to measure the conductivity independent of temperature, which is typically accomplished by normalizing the measured conductivity to that which would be measured at 258C. This correction is accomplished by measuring the conductivity cell temperature with a thermistor and multiplying the conductivity by a value known as the temperature compensation factor, and it is expressed in units of % per 8C [13]. The design of the mobile phase also plays a signiﬁcant role in the detection process, which is usually performed by a conductivity detector. A conductivity detector applies an electrical signal to the efﬂuent (i.e., the mobile phase exiting the chromatography column) as it passes through a specially designed cell, and measures its ability to carry a current. The greater the concentrations of ions in the efﬂuent, the higher the conductivity of the solution, and the larger the analytical signal reported in units of siemens. In non-suppressed-ion chromatography, the conductivity of the eluent is minimized via the careful selection of reagents and control of their concentration. Under these conditions, charged analyte ions are more conductive than the eluent and a signal is generated as they are pumped through the detector. In suppressed-ion chromatography, the response of the background is neutralized using a post-column suppressor (e.g., eluent hydroxide ions in anion-exchange chromatography and eluent hydrogen ions in cation-exchange chromatography are neutralized to water). Under these conditions, only the charged analyte ions are detected. Suppressed IC inherently and in practice has superior detection limits as compared to non-suppressed IC, despite its having a larger dead volume. Conductivity is a bulk property detector in that it cannot distinguish the origins or source of the ions. Hence, both mobile phase ions and sample ions are detected as a total. One of the consequences of how an ion separation occurs is that the conductivity of the mobile phase itself is quite high. This results in a large background (i.e., not the target analyte) signal that must be dealt with to make sensitive measurements. One approach, known as unsuppressed-ion chromatography, is to use mobile phase constituents that exhibit low ionic strengths and a specially designed conductivity detector that can electronically suppress the large background signals postdetection. The other approach is to use a suppressor unit after the column to neutralize the mobile phase ions thereby lowering the background signal relative to the analyte signal. Suppressed-ion chromatography has proven to be more sensitive, reliable, and less limiting than the unsuppressed approach. Figure 9.4 shows a simple schematic of suppressed-ion anion chromatography. For the sake of illustration, a sample containing potassium ﬂuoride (KF) and lithium chloride (LiCl) dissolved in water is injected into the separator column, which is an anion-exchange column in the hydroxide (OH) form. The ﬂuoride (F) and Cl anions equilibrate with the ion-exchange resin and are slowly displaced from the column by the hydroxide ion (OH) in the mobile phase. The potassium (Kþ) and lithium (Liþ) cations are not retained on the anion-exchange column and simply wash through. After a period of time, F and Cl elute from the column dependent on their individual afﬁnities for the ion-exchange resin as shown in Figure 9.4d. These ions cannot be easily detected due to the high background concentration of Naþ and OH ions, which is very deleterious to trace analysis. To remedy this problem, the efﬂuent is passed through an electronic suppressor, which generates electrolytically hydronium (Hþ) ions, see Figure 9.5. The membranes in the suppressor are speciﬁcally designed to allow only cations to pass freely. Hence, Hþ ions can pass through the membrane from the suppressor into the efﬂuent to replace any Naþ counterions, which ﬂow out of the efﬂuent stream. The anions of interest cannot pass through the membrane, and thus are restricted to the efﬂuent stream. The overall effect is the removal of the Naþ from the efﬂuent and the neutralization of OH to form water—a nonelectrolyte. When the analyte is present, HF and HCl with higher conductivity are produced and detected. Figure 9.6 illustrates the resolving power of

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 175 16.7.2008 10:47pm Compositor Name: MSubramanian

175

Ion Chromatography in Food Analysis

Mobile phase

(a)

Stationary phase

Mobile phase

(b)

Stationary phase

Mobile phase

(c)

Stationary phase

Mobile phase

(d)

Stationary phase

FIGURE 9.4 Representation of an IC separation. (a) The sample is injected as a mixture onto the head of the column. (b) The substances in the sample are distributed between the mobile phase and the stationary phase according to each substance’s afﬁnity for the stationary phase. (c) The compound, which resides in the mobile phase to the greater degree, is pushed down the column faster than the compound that is retained by the stationary phase. (d) The compounds are separated, and each elutes from the column at its own rate, which is reﬂected as the compounds retention time. Stationary phase: anion-exchange resin. Mobile phase: hydroxide, OH based eluent. (4) F from injected KF (.) Cl from injected LiCl.

suppressed-ion anion chromatography. Detection limits using suppressed-ion chromatography are routinely in the low ppb range. For cation chromatography, the suppressor membranes are anion-exchange polymers that allow anions to pass freely but exclude cations. Dilute acids (e.g., methanesulfonic acid) are used in the mobile phase. In the suppressor, the methanesulfonate counterions are replaced by hydroxide, neutralizing the acidic mobile phase and providing a highly conductive hydroxide counterion to the analytes cation. Post-column suppression offers the following beneﬁts: .

. .

Increased signal-to-noise. Increased analyte conductivity increases the signal, and decreasing background eluent conductivity lowers noise. As a result, suppressed conductivity often provides lower detection limits, a wider dynamic range, and dirty samples can be diluted more thereby extending column life. Improved chromatograms. The elimination of interference from eluent and sample counterions leads to fewer system peaks and baseline artifacts. Improved system performance. In addition to improved gradient compatibility, suppressed conductivity allows for the use of more concentrated eluents, which provides a greater range of elution control and the ability to use larger sample volumes.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 176 16.7.2008 10:47pm Compositor Name: MSubramanian

176

Handbook of Food Analysis Instruments Analyte and Na+ OH⫺ eluent Anode

Cathode Waste/vent

Waste/vent

+

– Na+ OH⫺ and H2

H2O and O2

OH⫺

Na+

H+

H+ + OH⫺ Cation-exchange membrane

H2O

H2O

Cation-exchange membrane

H+ + O2

H2 + OH⫺

H2O

H2O

H2O Analyte and H2O to detector

FIGURE 9.5

Autosuppression with the anion self-regenerating suppressor (ASRS).

10

10

7

1 3

4

5

1 µS

µS

2

6 0

5

3

6

0 0

(a)

7 4

2

2

4

6 8 Minutes

10

12

0

14 (b)

4

8

12 16 Minutes

20

24

FIGURE 9.6 Common anions as determined by EPA method 300.0. (a) Column: IonPac AG14, AS14. Eluent: 3.5 mM Na2CO3=1.0 mM NaHCO3. Flow rate: 1.2 mL=min. Injection volume: 50 mL. Detection: suppressed conductivity. (b) Column: IonPac AG9HC, AS9HC. Eluent: 9.0 mM Na2CO3. Flow rate: 1.0 mL=min. Injection volume: 50 mL. Detection: suppressed conductivity. Peaks (ppm): (1) ﬂuoride, 2; (2) chloride, 3; (3) nitrite, 5; (4) bromide, 10; (5) nitrate, 10; (6) phosphate, 15; (7) sulfate, 15. (Reprinted from Dionex Corporation. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 177 16.7.2008 10:47pm Compositor Name: MSubramanian

177

Ion Chromatography in Food Analysis

Suppressor-ion chromatography is widely accepted across the scientiﬁc community. It has been used routinely in thousands of laboratories for the past twenty years, and methods using suppressedion anion chromatography have been accepted and approved by governmental regulatory agencies. In fact, the US EPA has incorporated this technique in many of its own methods of analysis including EPA 300.1 for anions in drinking water. 9.2.5.4.3 Amperometric=Coulometric Detection Electrochemical detection (ED) can exploit the property of a compound to undergo either oxidation or reduction at an electrode to which a potential has been applied. The rate of the electrochemical reaction is observed as current, and, hence, these techniques fall under the heading of amperometry, or amperometric detection. The output from an electrochemical detector may be measured in either amperes or coulombs, if the signal is integrated over time. The conversion efﬁciency, or the percent of analyte converted to product, is typically less than 5% [14]. If all the analyte is oxidized or reduced to product, then the technique is referred to as coulometry. The quantity of analyte can be determined via Faraday’s law; Q ¼ nFN where Q is the number of coulombs N is the number of moles converted to product n is the number of electrons per reaction F is the Faraday constant Since 100% of the analyte is consumed, the need for a standard can theoretically be avoided. In ED following LC or IC, the eluent ﬂows into the electrochemical cell. As the eluent band of an analyte passes over the working electrode, it acts as an electron sink to either accept or donate electrons to the analyte should it be oxidized or reduced, respectively. The electroactivity of a compound is dependent on several factors, including molecular structure, accessibility of ﬁlled and unﬁlled orbitals, and functional groups present. Since the eluent is in motion over the electrode surface, diffusional mass transfer is assisted by forced convection to bring the analyte to the electrode surface. As a hydrodynamically controlled system, the limiting current (ilim) of the analyte can be described by the following equation: ilim ¼ nFAD

Cb d

where A is the area of the electrode D is the diffusion coefﬁcient of the analyte Cb is the concentration of the analyte d is the thickness of the diffusion layer n and F are the same as for Faraday’s law described above In addition to sample concentration, the limiting current or reaction rate is dependent on applied potential and the physical and chemical properties of both the eluent and the electrode material. Table 9.4 shows the useful potential limits for some common electrodes under various pH conditions. The simplest potential-time waveform that is applied to an electrode is that of a constant potential, which is known as DC amperometry. The high sensitivity and selectivity of ED is ideally suited for complex samples, as evinced by its application to the determination of neurotransmitters in complex biological samples (e.g., brain extracts). Neurotransmitters are typically aromatic

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 178 16.7.2008 10:47pm Compositor Name: MSubramanian

178

Handbook of Food Analysis Instruments

TABLE 9.4 Useful Potential Limits for Common Electrodes under Various pH Conditions Potential Limit (V)

Electrode Material

Supporting Electrolyte

Negative

Positive

DV

Pt

1 M H2SO4 pH 7 buffer 1 M NaOH 1 M NaOH 1 M H2SO4 1 M KCl 1 M NaOH 1 M HClO4 0.1 M KCl

0.3 0.7 1.0 0.9 1.1 1.9 2.0 0.2 1.3

þ1.2 þ1.0 þ0.6 þ0.8 þ0.3 þ0.1 þ0.1 þ1.5 þ1.0

1.5 1.7 1.6 1.7 1.4 2.0 2.1 1.7 2.3

Au Hg

C

Source: Adapted from a table within LaCourse, W.R., Pulsed Electrochemical Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997.

compounds (e.g., phenols, aminophenols, catecholamines, and other metabolic amines), which are detected easily by anodic reactions at a constant (DC) applied potential at inert electrodes [15,16]. Hypochlorite, ascorbate, hydrazine, arsenite, thiosulfate, nitrite, nitrate, cobalt, and iron are a partial list of the ions that have been detected using electrochemical detection [17]. The most common ions determined by amperometric detection in IC include inorganic anions forming complexes with cyanide, sulﬁde, and iodide at a silver electrode [18]. The relevant chemical reaction is shown below: Ag0 (the electrode) þ X ! AgX(s) þ e where Ag is the electrode X is the halide ion Silver is used as a sacriﬁcial electrode. A sacriﬁcial electrode is one where the electrode undergoes the redox reaction, and the electrode is consumed in the process. Other ions detected by this electrode include ﬂuoride, chloride, and bromide. 9.2.5.4.4 Pulsed Electrochemical Detection The most common electrode materials in ED are Au, Pt, and C. Electronic resonance in aromatic molecules stabilizes free-radical intermediate products of anodic oxidations, and, as a consequence, the activation barrier for the electrochemical reaction is lowered signiﬁcantly. Even for reversible redox couples that are considered to be well behaved, DC amperometry is often accompanied by the practice of disassembling the electrochemical cell and mechanically polishing the working electrode—daily. In this manner, fouling from nonspeciﬁc adsorption processes or mechanistic consequences is physically removed from the electrode surface. In contrast to aromatic moieties, absence of p-resonance for aliphatic compounds results in very low oxidation rates even though the reactions may be favored thermodynamically. Stabilization of free-radical products from aliphatic compounds can be achieved alternatively via their adsorption to the surface of noble metal electrodes. Unfortunately, adsorption of organic molecules and free radicals also has the consequence of fouling of the electrode and loss of its activity [19]. The historical perspective of non-reactivity for aliphatic compounds at noble metal electrodes can be attributed to surface fouling as a result of high, but transient, catalytic activity. An alternate

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 179 16.7.2008 10:47pm Compositor Name: MSubramanian

179

Ion Chromatography in Food Analysis ia F 50 µA

A

B

E +0.4

–0.8 V

+0.8 V

–0.4

D C

ic

FIGURE 9.7 Cyclic voltammogram of glucose at gold electrode in 0.1 M NaOH. Conditions: 900 rpm rotation speed, 200 mV=s scan rate. Solutions (.....) aerated 0.1 M NaOH, (——) deaerated 0.1 M NaOH, and (-----) 0.4 mM glucose. A, oxide formation; B, O2 evolution; C, oxide dissolution; D, dissolved O2 reduction; E, aldehyde oxidation; F, hydroxyl oxidation.

approach is to combine electrochemical detection with online cleaning. Hence, to maintain uniform and reproducible electrode activity at noble metal electrodes for polar aliphatic compounds, PED was developed [20]. A detailed review of all aspects of PED has been published [21]. Although increased sensitivity and reproducibility have also been reported for pulsed potential cleaning at carbon electrodes by several researchers [22,23], these electrodes have not generally been successful for the detection of polar aliphatic compounds. This effect is attributable to the absence of appropriate electrocatalytic properties of carbon surfaces to support the anodic oxygentransfer reaction mechanisms of polar aliphatic compounds. Most frequently, PED is applied at Au electrodes under alkaline conditions (pH >12). Figure 9.7 shows a cyclic voltammogram of (——) glucose at an Au electrode in (——) deaerated 0.1 M NaOH using an Ag=AgCl reference electrode. All aldehydes, including reducing sugars like glucose, are anodically detected during the positive potential excursion at the oxide-free surface in the region of approximately 0.6 to þ0.2 V. Large anodic signals are obtained for alcohols, polyalcohols, and nonreducing sugars in the region of approximately 0.3 to þ0.2 V (Mode I: oxide-free detections) with an attenuated signal for most compounds from approximately þ0.2 to þ0.6 V. Nitrogen and sulfur-containing compounds, for which a nonbonded electron pair is present, are adsorbed at oxide-free Au surfaces for E less than approximately þ0.1 V and can be anodically detected by oxidecatalyzed (Mode II) reactions during the positive scan or E greater than approximately þ0.1 V. Detections at E greater than approximately þ0.8 V are not recommended because of the deleterious effects of evolution of O2. Although there is little or no individual compound selectivity, functional group selectivity is clearly evident. Electroinactive surface-adsorbing species can be detected by suppression of the oxide formation process (Mode III) at potentials greater than approximately þ0.2 V. The presence of dissolved O2 can be a problem for detection in PED. Figure 9.7 shows also the signal for () dissolved O2 in the supporting electrolyte. Gold electrodes are commonly used in PED due to the presence of a background-free region from 0.1 V to þ0.2 V in 0.1 M NaOH. Voltammetric resolution of complex mixtures is futile since electrocatalytic-based detection of various members within a class of compounds is controlled primarily by the dependence of

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 180 16.7.2008 10:47pm Compositor Name: MSubramanian

180

Handbook of Food Analysis Instruments

the catalytic surface state on the electrode potential rather than by the redox potentials (E o ) of the reactants. Therefore, general selectivity is achieved via chromatographic separation before the electrocatalytic detection. This conclusion does not preclude limited selectivity from control of detection parameters. 9.2.5.4.4.1

Pulsed Amperometric Detection

Oxide-free detections are often implemented with a quadruple-pulse potential-time waveform (Table 9.5) at a frequency of ~2–0.5 Hz, which is appropriate for most IC applications to maintain chromatographic peak integrity. The detection potential (Edet) is chosen to be appropriate for the desired functional group, and the faradaic signal can be sampled during a short time (e.g., 16.7 ms) after a delay of tdel. Typical values of tdet are in the range of 100–600 ms. Following the detection step, the electrode surface is subjected to three additional potential excursions: 1. Reductive cleaning by a large negative potential pulse, or Ered (1.6 to 2.0 V) for a period of tred. In addition, any Au ions in the diffusion layer are reduced to elemental gold, which greatly extends the life of the electrode. 2. Oxidative cleaning by a positive potential pulse, or Eoxd (þ0.6 to þ0.8 V) for a period of toxd. 3. Reductive activation and adsorption by a large negative step to Eads (0.8 to þ0.1 V) for tads before the next detection cycle. This last step can also be used to increase adsorption, or preconcentrate, of the analyte to the electrode surface.

Potential

TABLE 9.5 Typical Waveforms for (a) PAD and (b) IPAD

Time (minutes) (b)

(a)

Parameter Edet

Ered Eoxd Eads

PAD

IPAD

Time (min)

Potential (V ) versus Ag=AgCl Reference

Time (min)

Potential (V ) versus pH Reference

0.00 0.20

þ0.10 þ0.10

0.40 0.41 0.42 0.43 0.44 0.50

þ0.10 2.0 2.0 þ0.6 0.1 0.1

0.00 0.16 0.17 0.41 0.42 0.51 0.52 0.53 0.54 0.55 0.60

þ0.33 þ0.33 þ0.55 þ0.55 þ0.33 þ0.33 1.67 1.67 þ0.93 þ0.13 þ0.13

Integration

Begin

End

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 181 16.7.2008 10:47pm Compositor Name: MSubramanian

Ion Chromatography in Food Analysis

181

Amperometric detection under the control of a simple potential-time waveform is known as pulsed amperometric detection (PAD), which is a subset of PED. Table 9.5 lists the typical quadruple-pulse potential-time waveform used in PAD for carbohydrates in 0.1. M NaOH. In addition to carbohydrates and alcohols, compounds which also have amine or sulfur moieties (e.g., aminoalcohols, aminosugars, thiosugars) rely on the detection of the –OH groups to take advantage of the simplicities of PAD waveforms for Mode I detections. The amine or sulfur moieties are exploited to increase the adsorption of the analyte to the electrode surface. Strong adsorption of reacting molecules is considered very beneﬁcial because the residence time of the molecule on the electrode surface is increased substantially, thereby increasing the probability for a successful detection reaction. In the case of compounds containing only amine or sulfur groups, only Mode II detections are often available. Numerous organic and inorganic sulfur compounds are adsorbed at the oxide-free surfaces of Au and Pt electrodes and can be detected by Mode II [24]. These compounds include thioalcohols, thioethers, thiophenes, thiocarbamates, organic thiophosphates, and numerous inorganic compounds. Adsorption is a prerequisite to detection and therefore at least one non-bonded electron pair must reside on the S-atom. The kinetics for detection of adsorbed S-compounds are quite favorable at pHs from 0 to 14. Since alcohol and amine groups are detected only under highly acidic or alkaline conditions, the detection of sulfur compounds under mildly acidic conditions is highly selective. 9.2.5.4.4.2

Integrated Pulsed Amperometric Detection

Anodic detection of amine- and sulfur-containing compounds occurs in a potential region where there is a signiﬁcant signal for the concurrent formation of surface oxide. As a result, a large baseline signal is often encountered for amino acids and sulfur compounds (Mode II). Furthermore, the large baseline current is frequently observed to drift to large anodic values, especially for new or freshly polished electrodes. This drift is the consequence of a slow growth in the true electrode surface area as a result of surface reconstruction caused by the oxide on–off cycles in the applied multistep waveforms. Mode II detections performed with PAD are subject to a number of disadvantages due to the formation of surface oxide, which is required and concomitant with the detection of amineand sulfur-based compounds. These disadvantages are as follows: .

.

.

Baseline sensitivities—Any changes or gradients in pH, organic modiﬁers, ionic strength, and temperature may lead to baseline drifting. The baseline drift is attributable to variations in the extent of surface oxide formation. Poor signal-to-noise ratio—The sample current is only a fraction of the total signal. The majority of signal for Mode II detections is derived from surface oxide formation. Oxideformation signal tends to be noisy. Post-peak ‘‘dips’’—The presence of the analyte at the electrode surface interferes with surface oxide formation. Hence, the background is different in the presence and absence of the analyte, which often results in a dip after the chromatographic peak.

The disadvantages listed above are either alleviated or greatly diminished by the use of integrated pulsed electrochemical detection (IPAD), which is also a form of PED. Table 9.5 shows the IPAD waveform that is generally used for amino acid detection. Here, the electrode current is integrated electronically throughout a rapid cyclic or square wave scan of the detection potential (Edet) within the pulsed waveform. The potential excursion proceeds into oxide formation (positive scan or step) and back out of the oxide formation (negative scan or step) region of the oxide-catalyzed reaction for detection by Mode II. The anodic charge for oxide formed on the positive sweep tends to be compensated by the corresponding cathodic charge (opposite polarity) for dissolution of the oxide on the negative sweep. Hence, the background signal on the electronic integration at the end of the detection period can be virtually zero and is relatively unaffected by the gradual change of

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 182 16.7.2008 10:47pm Compositor Name: MSubramanian

182

Handbook of Food Analysis Instruments

electrode area. IPAD combines cyclic voltammetry with potential pulse cleaning to maintain uniform electrode activity. 9.2.5.4.4.3

IC-PED

Ion chromatography separations are controllable via changes in pH, ionic strength, and organic modiﬁers. For example, carbohydrates are readily separated based on anion formation in alkaline media on anion-exchange columns [25] using pH and salt gradients. The separation of amino acids using AAA-Direct technology from Dionex (Sunnyvale, California) uses a complex quaternary gradient system (i.e., pH, salt, and organic modiﬁer) on an anion-exchange column to achieve baseline resolution of all the natural amino acids, vide supra. For a particular column type as described above, the following generalities about the mobile phase can be made to assist in the development of IC methods when using PED. .

.

.

pH—The effect of pH on the oxide formation process is attributable to the pH-dependent nature of Au oxide formation, i.e., Au(H2O)ads ! Au–OH þ Hþ þ e, and the potential for onset of oxide formation at Au electrodes shifts to more negative values with increases in pH at a rate of approximately 60 mV pH1. The negative shift in oxide formation with increasing pH can be reﬂected by a large baseline change in IC-PED under pH gradient elution when Edet remains constant throughout the gradient [26]. In isocratic separations, the effect of pH can be mitigated by changing the waveform potentials proper for a particular pH. This effect can also be alleviated to a great extent by substitution of a pHsensitive glass-membrane electrode for the Ag=AgCl reference electrode in the PED cell. Because the response of the glass-membrane electrode is approximately 60 mV pH1, the value of Edet is automatically adjusted during the execution of pH gradients. Ionic strength—Under ionic strength conditions suitable for electrochemical detection (i.e., m > 50 mM), the effect of changing ionic strength is reﬂected as minor perturbations in the background signal from oxide formation. This effect is not noticeable under isocratic IC conditions. Under gradient conditions (e.g., increasing acetate concentration), both positive and negative baseline drifts have been observed. Organic modiﬁer concentration—In comparison to ionic strength effects, changes in the concentration of organic modiﬁers can have a much greater effect on the baseline signal in IC-PED. This can occur, even for electroinactive organic additives, because the modiﬁers are frequently adsorbed at the electrode surface with a resulting suppression of the oxide formation process (Mode III). In addition to alteration of the IC-PED baseline, adsorbed organic modiﬁers can severely attenuate the analytical signal for carbohydrates by interfering with access to speciﬁc adsorption sites on the electrode needed for the reaction to occur.

The current produced by analyte detection based on an oxide-catalyzed detection is accompanied by current from surface oxide formation. Consequently, variables (e.g., pH, organic modiﬁer, and ionic strength) that affect the rates of oxide formation and dissolution will be reﬂected as drift in the baseline in IC-PED. For these reasons, IPAD was designed to apply a waveform that coulometrically rejects the oxide background by summing the charges due to oxide formation and oxide dissolution, which are expected to be of equivalent magnitude but opposite polarity. It is important to note that IPAD can virtually eliminate drift and changes associated with variations in composition (e.g., ionic strength and organic modiﬁer) of the mobile phase, as well as changes in the total surface area of the noble metal electrode surface. Changes in pH may require the simultaneous use of a pH reference electrode.

9.2.6 DATA COLLECTION

AND

OUTPUT SYSTEM

A data collection and output device (e.g., computer, integrator, or recorder) is connected to the electronic output of the detector or detectors. The data collection device takes the electronic signal

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 183 16.7.2008 10:47pm Compositor Name: MSubramanian

Ion Chromatography in Food Analysis

183

produced by the detector and outputs a plot of response versus time. This resulting chromatogram can then be evaluated for both qualitative and quantitative information. Recorders are rarely used today on their own. Both integrators and computers can integrate the peaks of a chromatogram and have the added advantage of being able to store chromatograms for post-collection processing. The data from computer-based collection systems can also be exported to other software programs and around the world by e-mail or over the Internet. In addition, the computer is typically able to communicate and control the entire HPLC system. Hence, the majority of modern IC systems are outﬁtted with computer control, data collection, and output systems.

9.2.7 POST-DETECTION ELUENT PROCESSING After the eluent passes through the detector, it is often directed to a waste container. The waste container should be properly labeled for eventual disposable by environmentally acceptable means. If isocratic chromatography is being performed, a mobile phase recycling device can be added after the detector. When the instrument is running and no injections are being made or during periods of the separation when no peaks are being eluted, the solvent recycling device will direct clean mobile phase back to the mobile phase reservoir. If collection of the eluted peaks is needed, a fraction collection device is added to the system after the detector. The collection of the peaks by the fraction collector can be done on the basis of a ﬁxed interval of time or mobile phase volume or via the output of the signal from the detector. Fraction collectors can also be under the control of an IC’s computer for ease of use.

9.2.8 CONNECTIVE TUBING

AND

FITTINGS

The connective tubing in the IC system should be made of material inert to the solvents and constituents of the mobile phase. Often, tubing is made of stainless steel or polymer-based materials (e.g., PEEK, Teﬂon, etc.). The connections between tubings and IC components are made with ﬁttings and unions designed to minimize excess volume, or dead volume. Zero-dead-volume (ZDV) and low-dead-volume (LDV) ﬁttings are necessary to reduce band-broadening effects. Also, great care should be used when assembling columns and ﬁttings so that they match properly. Tubing should be cut ﬂat and ﬁt ﬂush, and all ﬁttings should be zero-dead-volume. Any extra-column volume compromises separation efﬁciency. Polymer-based, or PEEK, tubing and ﬁttings throughout the chromatographic ﬂow path are becoming very popular. PEEK tubing is available in a wide range of sizes, is color-coded for ease of size selection, is easy to work with, is tolerant to a wide range of buffer and solvent conditions, and is impermeable to oxygen. The effects of band-broadening are not important before the injection; therefore, the i.d. of the tubing from the pump to the injector=autosampler should be as large (i.e., typically 0.030 in. internal diameter) as possible (i.e., to reduce system back-pressure) while maintaining high pressure strength for isocratic systems. In gradient systems, the dead volume should be minimized by using narrow bore tubing to avoid delays in the onset of gradient proﬁles. From the injector to the column to the detector, all tubing should be as short as possible and of the narrowest diameter available to minimize extra-column effects. Post-injector tubing diameters are often 0.01–0.005 i.d.

9.2.9 RELATED SEPARATION TECHNIQUES The IC system described above was based on the dimensions of a typical analytical system, otherwise known as normal bore chromatography. The same basic components and designs are used in both micro- (i.e., column dimensions of 2 mm i.d.) and preparative systems, which typically use larger column dimensions (>10 mm i.d.). Microchromatographic techniques include microbore, nanobore, and capillary chromatography, which have column i.d. of 1–2, 0.3–1 mm, and <300 mm, respectively. Most importantly, microchromatographic systems require smaller extra-column volumes than normal bore systems. Hence, great pains should be taken to use the smallest i.d. tubing available,

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 184 16.7.2008 10:47pm Compositor Name: MSubramanian

184

Handbook of Food Analysis Instruments

zero-dead-volume ﬁttings, smaller injection volumes, and smaller detection cell volumes. Microchromatographic systems are employed when sample volume is limited. MicroIC systems are of limited availability from IC manufacturers. If compound puriﬁcation or isolation is intended, preparative chromatography is used. Since a large quantity of material is to be injected in order to isolate a signiﬁcant quantity of analyte, column capacity and, hence, columns diameters are larger. Column diameters in preparative IC can be on the order of meters. The high ﬂow rates needed for these large-scale systems require the use of special pumps and larger tubing throughout.

9.3 ION CHROMATOGRAPHY APPLICATIONS TO FOOD ANALYSIS Food and beverage matrices are complex mixtures of compounds ranging in concentration from trace levels to percentages for target analytes. Methods often require extensive cleanup procedures to remove interferences. Since the inception of IC, the food industry has been both a pioneer and benefactor of this technology. The selectivity derived from highly efﬁcient chromatographic separations combined with high sensitivity detection leads to improved chromatograms derived from methods of minimal sample cleanup. Table 9.6 shows a partial listing of ofﬁcially accepted methods that take advantage of either CD or PED [27]. The following brief overview of several important food and beverage applications has been designed to help the reader understand and explore the breadth and analytical utility of IC. TABLE 9.6 Partial List of Ofﬁcially Accepted IC Methods Agency AOAC International Methods

Method AOAC Method 993.30 AOAC Method 996.04 AOAC Method 995.13

Analytica-EBC International Methoda

AOAC Method 997.08 —

International Standards Organization

ISO 11292

Int. Comm. for Uniform Methods of Sugar Analysis American Society for Testing Materials US National Institute for Occupational Safety and Health U.S. Environmental Protection Agency

ISO 10304-1 — ASTM D4327-91 NIOSH 4110 Method 300.0 Method 218.6

a

Title of Method Determination of Inorganic Anions in Water by Ion Chromatography Determination of Sugar in Molasses Determination of Carbohydrates in Soluble (Instant) Coffee: Anion Exchange Chromatographic Method with Pulsed Amperometric Detection Determination of Fructans in Food Products Determination of Anions in Beer by Ion Chromatography Instant Coffee: Determination of Free and Total Carbohydrates—Methods by High Performance Anion Exchange Chromatography Anions in Natural and Contaminated Waters Determination of Sugar in Molasses Anions in Water by Chemically Suppressed Ion Chromatography Determination of Anions by Ion Chromatography Determination of Inorganic Anions in Water by Ion Chromatography Determination of Dissolved Hexavalent Chromium in Drinking Water, Ground Water, and Industrial Efﬂuents by Ion Chromatography

This method was collaboratively tested and approved by the American Society for Brewing Chemists, the European Brewing Convention, and the Brewery Convention of Japan.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 185 16.7.2008 10:47pm Compositor Name: MSubramanian

185

Ion Chromatography in Food Analysis

9.3.1 INORGANIC ANIONS

AND

CATIONS

Ion chromatography offers the food chemist an important avenue to address the needs of food labeling requirements, food processing, and quality control. The concentration of ions may inﬂuence the ﬂavor and quality of food and beverages, including the health concerns associated with ions such as nitrite, bromide, bromate, iodide, cyanide, and chromium(IV). Many of these ions require their determination at trace levels in the presence of high concentrations of other similar ions. At the heart of many food and beverage products is the incoming water used as a raw material. The analysis of the water for many of the standard cations (group I and group II) and anions (e.g., 3 2 F, Cl, Br, NO 2 , NO3 , PO4 , and SO4 ) commonly found in municipal drinking water is of critical importance. Ion levels of the incoming water can act as monitors for quality control, ﬂavor, and nutritional considerations. Alkali and alkaline earth metal cations are easily and rapidly determined on an IonPac CS16 column using a 26 mM methanesulfonic acid eluent and conductivity detection, see Figure 9.8a. Figure 9.8b shows analysis of tap water with IC-CD [28]. Column technology allows for the determination of most cations even in the presence of signiﬁcant amounts of other ions, such as sodium ion. In addition to water, the cation proﬁles of many beverages, such as diet cola and wine, have been produced [27]. Owing to high column speciﬁcity and detection sensitivity, sample preparation is reduced to degas, dilute, and shoot. Figure 9.9 shows the separation of the common anions that are easily separated with baseline resolution using anion separation technology [29]. Since nitrate can be reduced to nitrite, which can react in the human body to form carcinogenic nitrosamines, their determination is an important health concern. Nitrate and nitrite can be determined at the low ppb level in drinking water by IC with either UV detection or suppressed conductivity detection using EPA Method 300.1. This same approach can be used for the determination of nitrates and nitrites in ham, see Figure 9.10, using IC with UV absorbance detection [29]. In this method, homogenized meat samples are extracted with 30 mM Methanesulfonic acid 1.00

2

µS

1.50

µS

23 6 1 5 7 8 4

6

3

1

0.10 5

0 11

12

10 15 20 48 mM Methanesulfonic acid

1.00

9

2

µS

10

0.80

5

4

6

8

10

25

3

1

0.10 0

(a)

5

4

10

20

Minutes

0

30 (b)

2

4

6 Minutes

12

FIGURE 9.8 (a) Amines and common cations separated on an IonPac CS16 using suppressed conductivity in the recycle mode. (b) Tap water sample at various eluent strengths on the same column. (left) Eluent: 26 mM methanesulfonic acid. Flow rate: 1.0 mL=min. Injection volume: 25 mL. Temperature: 658C. Detection: suppressed conductivity. Peaks (ppm): (1) lithium, 0.05; (2) sodium, 0.20; (3) ammonium, 0.25; (4) ethanolamine, 0.50; (5) methylamine, 0.50; (6) potassium, 0.50; (7) dimethylamine, 1.00; (8) 5-amino-1-pentanol, 2.00; (9) morpholine, 2.00; (10) trimethylamine, 1.50; (11) magnesium, 0.25; (12) calcium, 0.50. (right) Flow rate: 1.0 mL=min. Injection volume: 25 mL. Temperature: 408C. Detection: suppressed conductivity. Peaks (ppm): (1) lithium, 0.002; (2) sodium, 19.730; (3) ammonium, 0.065; (4) potassium, 0.987; (5) magnesium, 7.210; (6) calcium, 18.544. (Reprinted from Dionex Corporation. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 186 16.7.2008 10:47pm Compositor Name: MSubramanian

186

Handbook of Food Analysis Instruments 100

µS

2

7 3 4

1

5 6

0 0

5

10

2

7

2 µS

4

1

3

56

0 0

5 Minutes

10

FIGURE 9.9 Anion proﬁle of a municipal drinking water sample. Column: IonPac AG14A, AS14A, 5-mm (3 150 mm). Eluent: 8.0 mM Na2CO3=1.0 mM NaHCO3. Flow rate: 0.5 mL=min. Injection volume: 12.5 mL. Temperature: 308C. Detection: suppressed conductivity. Peaks (ppm): (1) ﬂuoride, 0.06; (2) chloride, 42.0; (3) bromide, 0.03; (4) nitrate, 3.2; (5) chlorate, 0.14; (6) phosphate, 1.05; (7) sulfate, 61.0. (Reprinted from Dionex Corporation. With permission.)

water at 708C to 808C for 15 min followed by centrifugation and ﬁltration. The sample is directly injected without further cleanup. Similarly, nitrate=nitrite and iodide in dairy products are particularly important because of potential health implications. Figure 9.11 shows the determination of I in (1) 2% milk and (2) infant formula by IC followed by DC amperometry at an Ag electrode with pulsed potential cleaning [29]. Only a simple protein precipitation was required. Another health concern is bromate, which is a by-product of the ozonation disinfection process for drinking water. Bromate has been cited as a potential carcinogen even at the low ppb level by both the US EPA and the World Health Organization. Bromate is also used as a dough stabilizer for bread and other baked goods, with trace amounts remaining in the ﬁnal products. Bromate can be determined at target levels of low ppb in the presence of much higher concentration of chloride and other common inorganic ions [27].

0.02

AU

1

2

⫺0.01

0

5 Minutes

10

FIGURE 9.10 Determination of nitrate and nitrite in ham extract. Column: IonPac AG11, AS11. Eluent: 5 mM NaOH. Flow rate: 1.0 mL=min. Injection volume: 25 mL. Detection: UV, 225 nm. Peaks (ppm): (1) nitrite, 1.16 and (2) nitrate, 0.54. (Reprinted from Dionex Corporation. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 187 16.7.2008 10:47pm Compositor Name: MSubramanian

187

Ion Chromatography in Food Analysis 30

nC

nC

30

1

1

–2

–2 0

(a)

2

4 6 Minutes

8

0

2

(b)

4 6 Minutes

8

FIGURE 9.11 Analysis of iodide in (a) 2% milk and (b) infant formula. Column: IonPac AG7, AS7. Eluent: 50 mM HNO3. Flow rate: 1.5 mL=min. Injection volume: 50 mL. Detection: amperometry, þ0.8 V. Pt electrode. Peaks (ppb): (1) iodide, (a) 81, (b) 33. (Reprinted from Dionex Corporation. With permission.)

Sulﬁte is a food additive that is used as a preservative. Unfortunately, many in the general population exhibit allergic reactions. Hence, the US FDA regulates the concentration to be <10 mg/kg otherwise it requires it to be noted on the label. On the basis of a method by Kim and Kim [30], IC was used to determine sulﬁte in dried apricots. It was later shown that PAD at a Pt electrode improved reproducibility as compared to DC amperometry as used previously by Kim and Kim. Wagner and McGarrity [31] used ion chromatography followed by PAD to determine sulﬁte in beer. Correlation coefﬁcients of 0.997 or better were obtained with excellent spike recoveries. HPAECPAD results were in good agreement with the standard para-rosaniline method. In fact, this method was subsequently automated resulting in excellent precision and accuracy [32]. Traditionally, transition metals in food products have been determined by atomic absorption or inductively coupled plasma spectroscopy. Chelation IC using PDCA as a mobile phase additive followed by post-column derivatization with PAR allows for the separation and detection of seven different metals, see Figure 9.12 [33]. In contrast to AA and ICP, interference from calcium and magnesium is not a problem even at transition metal ion concentrations in the low ppb level. The power of IC is clearly demonstrated by the determination of commercial polyphosphates. Polyphosphates are widely used additives in products such as fruit juices and canned goods to prevent discoloration and off-ﬂavors. They are also used in curing ham, tenderizing vegetables, as emulsion stabilizers for cheese, and to retain moisture in frozen entrees. Commercial polyphosphates are

2

5

2

4

AU

1

7 3 6

0 0

2

4

6

8 10 Minutes

12

14

FIGURE 9.12 Separation of a mixture of seven transition metals followed by post-column complexation with PAR and UV absorbance detection. Column: IonPac AG5A, AS5A (2 mm). Eluent: PDCA. Eluent ﬂow rate: 0.3 mL=min. Injection volume: 30 mL with TCC-2 concentrator column. Total run time: 30 min. Detection: 530 nm. Peaks (1 ppb each): (1) iron; (2) copper; (3) nickel; (4) zinc; (5) cobalt; (6) cadmium; and (7) manganese. (Reprinted from Dionex Corporation. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 188 16.7.2008 10:47pm Compositor Name: MSubramanian

188

Handbook of Food Analysis Instruments Bad batch 10

2

1

8 6 µS

4

5 6

4 2

3

0 0

5

10

15

(a)

20

25

30

35

40

30

35

40

Minutes Good batch

10

1 3

8

µS

6

6

4 5

2

4 2 0 0

(b)

5

10

15

20

25

Minutes

FIGURE 9.13 Phosphonates in (a) bad and (b) good batches of cheese products. Column: IonPac AG11, AS11 (2 mm). Eluent: NaOH gradient. Flow rate: 0.3 mL=min. Injection volume: 10 mL. Detection: suppressed conductivity. Peaks: (1) PO4; (2) P2O7; (3) P3O9; (4) P3O10; (5) P4O12; and (6) P4O13. (Reprinted from Dionex Corporation. With permission.)

mixtures of polyphosphates of various chain lengths. Baluyot and Hartford [34] used IC-CD to troubleshoot a critical problem in the processing of cheese products. Figure 9.13 shows the chromatographic proﬁle of two 50% sodium metaphosphate solutions that were prepared from the same lot of dry powder. The signiﬁcant difference in the chromatograms indicated that there was a critical problem in the processing. Note the high degree of resolution provided by the chromatographic separation.

9.3.2 ORGANIC ACIDS Organic acids in fruit juices are important for establishing freshness and for detecting adulteration. Ratios of organic acid often can be used to ﬁngerprint a particular type of juice. For instance, quinic acid is a speciﬁc marker for cranberry juice and it is used as a measure of its purity and authenticity [27]. Other fruit juices show characteristic organic acid proﬁles. Organic acids are important ﬂavor constituents in brewing liquors; whereas, the inorganic ions affect physical appearance. Both organic acids and inorganic ions can be determined simultaneously (Figure 9.14) [29]. Food dyes fall into one of four classes: azo(mono-, di-, and tri-), indol, triphenylmethane, and methin dyes. For the most part, these dyes are acidic or anionic due to the presence of sulfonates, carboxyl, or phenolic groups. By using a mixed-mode column technology that combines reversed phase and ion-exchange mechanisms of interaction, complex mixture of many common food dyes can be separated and detected using UV absorbance spectrophotometry [27].

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 189 16.7.2008 10:47pm Compositor Name: MSubramanian

189

Ion Chromatography in Food Analysis

10

20

µS

9 12

15

8 2 1 3

4

7 5

14

13 11

16

6

0 0

5

10

15

20

25

30

35

40

Minutes

FIGURE 9.14 Determination of anions and organic acids in grape juice. Column: IonPac AG11, AS11. Eluent: KOH gradient=10% MeOH. Flow rate: 1.5 mL=min. Injection volume: 10 mL. Temperature: 308C. Sample preparation: 1–10 dilution with deionized water. Detection: suppressed conductivity. Peaks (ppm): (1) quinate, 55; (2) ﬂuoride, 0.8; (3) lactate, 46.2; (4) galacturonate, 60.0; (5) chloride, 1.0; (6) nitrate, 0.6; (7) glutarate, 0.7; (8) succinate, 1.0; (9) malate, 116; (10) malonate=tartrate, 190; (11) maleate, 2.0; (12) sulfate, 15.0; (13) oxalate, 12.9; (14) phosphate, 27,0; (15) citrate, 80.0; and (16) iso-citrate, 1.0. (Reprinted from Dionex Corporation. With permission.)

9.3.3 AMINES

AND

OTHER ORGANIC BASES

Low molecular weight amines such as methyl-, dimethyl-, and trimethylamine are indicators of the quality in ﬁsh and other food products. Using IC with CD, these low molecular weight amines can be determined simultaneously with numerous other common cations. Other studies have focused on the detection of aliphatic amines by IC-PAD [35], but amine detections are oxide-catalyzed and are best performed using IPAD. A ﬁne example involves the use of HPLC-IPAD to directly determine biogenic amines as indicators of seafood spoilage [36]. They separated putrescine, histidine, cadaverine, and histamine using a cation-exchange column with an ACN (acetonitrile) gradient. Figure 9.15 shows the chromatograms of amines extracted from spoiled canned herrings [28]. Detection was performed by IPAD following post-column addition of NaOH. Simple mixtures of amino acids can be separated isocratically using anion-exchange chromatography [37]. For complex mixtures, it is essential to perform separations with gradient-elution chromatography. Figure 9.16 shows the chromatogram of 22 amino acid residues using IC with IPAD with a pH reference electrode [29]. The amino acid mixture was separated using an anion-exchange column with a quaternary gradient, which incorporated both a pH and an organic modiﬁer gradient. Presently, detection limits (i.e., S=N ¼ 3) for IC-IPAD for amino acids are typically 1–50 pmole injected with comparable sensitivities for primary and secondary amino acids. Naturally occurring alkaloids (e.g., theobromine, theophylline, and caffeine) are bitter ﬂavor constituents in coffee, tea, cola, and similar beverages. A mixture of 10 alkaloids have been separated and detected using a mixed-mode cationexchange column followed by UV absorbance detection, respectively [27].

9.3.4 CARBOHYDRATES

AND

OTHER OLIGOSACCHARIDES

High-performance anion-exchange chromatography (HPAEC)-PAD is now well established for the direct detection of monosaccharides, disaccharides, oligosaccharides, polysaccharides, and sugar alcohols. The technique of HPAEC-PAD has been reviewed extensively by LaCourse [18]. The analysis of the components derived from the hydrolysis of complex carbohydrates and dietary ﬁber has also been reviewed [38]. The following summary of food applications is only a glimpse of the applications covered by this technique for food and beverage analysis.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 190 16.7.2008 10:47pm Compositor Name: MSubramanian

190

Handbook of Food Analysis Instruments

µC

3

1 2

3 4

0 (a) 3

1 2

µC

3

4 5

0 0

5

10

(b)

15

20

Minutes

FIGURE 9.15 Determination of biogenic amines in (a) canned herring and (b) same sample spiked with 300 mg=g of each amine. Column: IonPac CG11, CS11. Eluent: gradient of (1) acetonitrile–water 90:10 (v=v); (2) HClO4, 0.5 M; (3) NaClO4, 1.0 M; and (4) water. Flow rate: 1.0 mL=min. Detection: IPAD. Peaks (mg=g): (1) putrescine, 16; (2) histidine, 103; (3) cadaverine, 187; (4) histamine, 172; and (5) spermidine. (Reprinted from Dionex Corporation. With permission.)

Carbohydrates are basic constituents in food and beverage products. In addition to their analysis as sweeteners, carbohydrates act as bulking agents and fat substitutes. Their determination is important for quality control monitoring (e.g., fermentation of alcoholic beverages), authenticity, and food labeling claims. HPAEC offers the food chemists a powerful method to separate sugar alcohols, monosaccharides, oligosaccharides, and polysaccharides. The direct (without derivatization) and sensitive detection is made possible with PAD. Figure 9.17 shows the separation of 18 carbohydrates, alditols, alcohols, and glycols found in a common fermentation broth [29]. As discussed above, carbohydrate analysis of foods is an effective means of identifying cases of adulteration. Low [39] describes several projects in which adulteration was easily detected using the carbohydrate proﬁle from both the pure food and the sweeteners used in the adulteration process. 275 17

nC

1

21

2 3 18

5

10

7 6 8

4

9

22 20 19

14

11 12 13

15 16

50 0

5

10

15 Minutes

20

25

30

FIGURE 9.16 Separation and detection of 22 amino acids (100 pmole each) by IC-IPAD with a pH reference electrode. (Reprinted from Dionex Corporation. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 191 16.7.2008 10:47pm Compositor Name: MSubramanian

191

Ion Chromatography in Food Analysis 140 5 4 nC

23 7 1

6

8 9 10 12 11 13 15 14

16

17

18

0 0

10

20

30 40 Minutes

50

60

70

FIGURE 9.17 Assay of a common fermentation broth for carbohydrates, alditols, alcohols, and glycols. Column: CarboPac MA1, MA1 guard. Eluent: 480 mM NaOH. Flow rate: 0.4 mL=min. Injection volume: 10 mL. Detection: PAD, Au electrode. Peaks: (1) 2,3-butanediol; (2) ethanol; (3) methanol; (4) glycerol; (5) erythritol; (6) rhamnose; (7) arabitol; (8) sorbitol; (9) galactitol; (10) mannitol; (11) arabinose; (12) glucose; (13) galactose; (14) lactose; (15) ribose; (16) sucrose; (17) rafﬁnose; and (18) maltose. (Reprinted from Dionex Corporation. With permission.)

Grapefruit juice and honey adulteration can be determined by comparing the ﬁngerprint obtained for these products using HPAEC-PAD with the ﬁngerprint obtained for sweeteners such as high fructose corn syrup. Using these proﬁles, adulteration can be easily identiﬁed. Analysis of honey for carbohydrate content is a challenging problem. Glucose and fructose are the major components, but in addition, honey contains at least 11 disaccharides, 11 trisaccharides, and other higher oligosaccharides. Swallow and Low [40] analyzed honey from different sources to evaluate the applicability of HPAEC-PAD for this analysis. Glucose and fructose were determined by simply diluting the honey with water and analyzing by HPAEC-PAD. Analysis of the puriﬁed oligosaccharide samples revealed a unique ﬁngerprint for honeys from different natural sources. Figure 9.18 shows a chromatogram of the carbohydrate proﬁle obtained for a pure honey. 2

17

3

nC

5 6 13 7 14 8

18 19

9 10 11

1

0

12

30

60 Minutes

FIGURE 9.18 HPAEC-PAD of carbohydrates in alsike honey. Peaks: (1) neotrehalose; (2) glucose; (3) fructose; (5) isomaltose, matulose; (6) sucrose; (7) kojibiose; (8) turanose=genitobiose; (9) palatinose; (10) melezitose; (11) isomaltotriose; (12) nigerose; (13) maltose, 1-kestose; (14) theanderose; (17) erlose; (18) panose; and (19) maltotriose. (Reprinted from Dionex Corporation. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 192 16.7.2008 10:47pm Compositor Name: MSubramanian

192

Handbook of Food Analysis Instruments

Fingerprints of this type can be used to detect adulteration of honey with inexpensive sweeteners. In a later paper, Swallow and Low [41] demonstrated the use of HPAEC-PAD for the determination of honey adulteration. In this work, samples from 44 different honey sources were analyzed as above and although the oligosaccharide proﬁles were similar, a unique ﬁngerprint was discerned for each one. The sources of the variation in the honey are numerous, ranging from the botanical source to the storage temperature. Honey samples were then purposely adulterated with high fructose corn syrup and invert syrup to determine changes in the proﬁle. In both cases, the presence of the adultering substance was easily detected. Determination of adulteration of maple syrup by HPAECPAD was also shown by Stuckel and Low [42]. Monosaccharides and disaccharides in molasses were determined by HPAEC-PAD, which is the International Commission for Uniform Methods of Sugar Analysis ofﬁcial method, see Table 9.6. This was an interlaboratory study that involved 11 laboratories. The reproducibility of this method was excellent for the different laboratories and the results agreed with a GC method. Bernarl et al. [43] also determined monosaccharides present in instant coffee. A C-18 SPE step was again used before the analysis to protect the chromatography system. Two of the coffees used in this study were adulterated. This was easily detectable because the levels of glucose and fructose were elevated compared with the unadulterated coffees. Free carbohydrates and total carbohydrate content of soluble coffee were determined in a multiple laboratory study [44]. Eleven different laboratories analyzed six different coffee samples and the results were compared to determine the precision of the HPAEC-PAD method in the different laboratories. When the carbohydrate content in the coffee was greater than 0.3%, the precision of the HPAEC-PAD method was excellent. Lactose and sucrose were determined in a chocolate milk sample using HPAEC-PAD [45]. The sample was prepared by simply diluting 100 mg chocolate milk in 1 L water. The resulting chromatogram was very clean with the analyte peaks clearly detectable. No background peaks were observed. These papers demonstrate the ease of application of PAD to a very complex milk matrix. One of the earliest papers using HPAEC-PAD showed the analysis of high fructose corn syrup and potato chips [46]. The corn syrup assay was of immediate interest to manufacturers, because the presence of higher saccharides affects the taste of the syrup by imparting a bitter taste. The use of HPAEC-PAD for quality control of product was demonstrated with the analysis of glucose, sucrose, and lactose to determine the level of seasoning present in the ﬁnished product. Rocklin and Pohl published a more detailed study of determining glucose, fructose, and lactose from extracts of ﬂavored potato chips in the presence of high concentrations of salt [47]. Artiﬁcial sweeteners include sucrose derivatives as well as sugar alcohols. Sucralose is a sweetener which is 400–800 times as sweet as sucrose. It is a selectively chlorinated sucrose. Ichiki and coworkers determined sucralose in food [48], and trace impurities 4-Cl-galactose and 1,6-Cl-fructose were detected in a sucralose sample, see Figure 9.19. Other sugar substitutes such as kestose, maltrin, and inulin are also separated and detected using HPAEC-PAD. Sugar alcohols including glycerol, sorbitol, and mannitol have been determined in hard candies and chewing gum. Sample preparation involves either dissolution or sonication followed by dilution. The determination of lactose, galactose, and dextrose in grated cheeses was improved by using anion-exchange chromatography with PAD [49]. Previous LC methods suffered from a lack of sensitivity, failure to resolve galactose and dextrose, and interference from the presence of salt in the cheese in the determination of dextrose. Koizumi and coworkers used HPAEC-PAD to determine D-gluco-oligosaccharides and D-glucopolysaccharides with degree of polymerization (DP) 50 [50]. They also determined that sensitivity is not independent of molecular weight; rather, it increases for each HCOH group present, which makes PAD a good detection choice for oligosaccharides. In addition, Koizumi and coworkers demonstrated the separation of depolymerization products of the amylopectin portion of starch

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 193 16.7.2008 10:47pm Compositor Name: MSubramanian

193

Ion Chromatography in Food Analysis 780

nA

3

1 2

0 0

10

20 30 Minutes

40

50

FIGURE 9.19 Determination of impurities in sucralose. Column: CarboPac MA1. Eluent: 150 mM sodium acetate=0.2% (v=v) acetic acid, pH 5.5. Flow rate: 0.4 mL=min. Injection volume: 25 mL. Detection: PAD, Au electrode. Peaks (ppm): (1) 4-Cl-galactose, 7.8; (2) 1,6-di-Cl-fructose, 2.4. and (3) sucralose. (Reprinted from Dionex Corporation. With permission.)

from different crops including rice, corn, sweet potato, and edible canna. Baseline separation of amylopectins from DP ¼ 6 to DP ¼ 60 is accomplished in 40 min [51]. Under gradient conditions, acetate ion is typically the preferred ‘‘pusher’’ ion, in that it offers rapid equilibration, is PAD-inactive, and allows for near maximal resolution of carbohydrate moieties. Wong and Jane compared the effects of acetate and nitrate as pushing agents in HPAEC-PAD for the separation and detection of debranched amylopectin [52]. They found that in comparison with the commonly used pushing agent, nitrate offered greater reproducibility, accuracy, and lower limits of detection. Figure 9.20 shows the HPAEC-PAD chromatographic proﬁles of the enzymatically debranched tapioca and wheat amylopectins using a nitrate gradient. Impressively, baseline resolution of peaks up to a DP of 66 was achieved. The separations were designed to be completed within 100 min.

6000 5500 PAD signal (nA)

10 5000 4500

20

4000 30

3500

40

50

3000 (a)

FIGURE 9.20

0

10

20

30

40

50

60

70

80

90

100

Chromatographic proﬁles of the enzymatically debranched (a) tapioca amylopectin and (continued )

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 194 16.7.2008 10:47pm Compositor Name: MSubramanian

194

Handbook of Food Analysis Instruments 6000

PAD signal (nA)

10

5000

20 4000 30

3000 0

10

(b)

20

30

40

50

50

40

60

70

80

90

100

Retention time (min)

FIGURE 9.20 (continued) (b) wheat amylopectin using HPAEC-PAD. Peak numbers indicate the degree of polymerization. (Reprinted from Wong, K.S. and Jane, J., J. Liq. Chromatogr., 18, 63, 1995. With permission.)

9.4 FUTURE DIRECTIONS Ion chromatography has matured rapidly over the last 50 years. Three major advancements have allowed this technology to reach its full potential. First, the development of polymeric-based phases allowed the full range of pH to be exploited for highly efﬁcient separations. In addition, the deep understanding of this technology allowed for the tailoring of separations to handle uniquely difﬁcult problems of critical signiﬁcance. Second, the invention of suppressed conductivity detection afforded the analyst both high sensitivity and ease of use, especially with electrochemically generated suppression. Third, the revolutionary development of PED that allowed for the sensitive and direct determination of carbohydrates, amines, and sulfur-containing compounds without derivatization. The future continues to be bright for IC. Evolutionary maturity is obvious in the latest commercial systems. Eluent generators allow one to dial in eluent concentration, unique column chemistries (e.g., immobilized metal afﬁnity, chelation, and cryptand based) and designs (e.g., monolithic), improve separations, and detections are performed by MS, which allows for both compound characterization and quantitation. In addition, software is becoming available to simplify the task of selecting phase and predicting separations. Reﬁnements continue in all aspects of IC, microchromatographic systems are on the rise, and the last frontier continues to be sample preparation technology.

ACKNOWLEDGMENT This author would like to acknowledge the support and assistance by Dionex in the preparation of this chapter.

REFERENCES 1. Campbell, P.N., Separation of amino acids on columns of anion-exchange resins, Biochim. Biophys. Acta, 21, 167, 1956. 2. Fritz, J.S. and Gjerde, D.T., Ion Chromatography, 3rd ed., Wiley-VCH, New York, 2000, p. 38. 3. Buytenhuys, F.A., Ion chromatography of inorganic and organic species using refractive index detection, J. Chromatogr., 218, 57, 1981.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 195 16.7.2008 10:47pm Compositor Name: MSubramanian

Ion Chromatography in Food Analysis

195

4. Haddad, P.R. and Heckenberg, A.L., High-performance liquid chromatography of inorganic and organic ions using low-capacity ion-exchange columns with indirect refractive index detection, J. Chromatogr., 252, 177, 1982. 5. Fritz, J.S. and Gjerde, D.T., Ion Chromatography, 3rd ed., Wiley-VCH, New York, 2000, p. 67. 6. Jackson, P.E., Haddad, P.R., and Dilli, S., Determination of nitrate and nitrite in cured meats using highperformance liquid chromatography. J. Chromatogr., 295, 471, 1984. 7. Goodkin, L., Seymour, M.D., and Fritz, J.S., Ultraviolet spectra of metal ions in 6M hydrochloric acid, Talanta, 22, 245, 1975. 8. Matsushita, S., Simultaneous determination of anions and metal cations by single-column ion chromatography with ethylenediaminetetraacetate as eluent and conductivity and ultraviolet detection, J. Chromatogr., 312, 327, 1984. 9. Fritz, J.S. and Gjerde, D.T., Ion Chromatography, 3rd ed., Wiley-VCH, New York, 2000, p. 69. 10. Imanari, T., Tanabe, S., Toida, T., and Kawanishi, T., High performance liquid chromatography of inorganic anions using iron(3þ) as a detection reagent, J. Chromatogr., 250, 55, 1982. 11. Hershcovitz, H., Yarnitsky, Ch., and Schmuckler, G., Ion chromatography with potentiometric detection, J. Chromatogr., 252, 113, 1982. 12. Reeve, R.N., Determination of inorganic main group anions by high-performance liquid chromatography, J. Chromatogr., 177, 393, 1979. 13. Rocklin, R.D., Conductivity and Amperometry, Dionex, Sunnyvale, CA, 1989, pp. 9–10. 14. LaCourse, W.R., Pulsed Electrochemical Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997, p. 61. 15. Adams, R.N., Electrochemistry at Solid Electrodes, Marcel Dekker, New York, 1969. 16. Kissinger, P.T., Laboratory Techniques in Electroanalytical Chemistry, Kissinger, P.T and Heineman, W.R. (Eds.), Marcel Dekker, New York, 1984. 17. Fritz, J.S. and Gjerde, D.T., Ion Chromatography, 3rd ed., Wiley-VCH, New York, 2000, p. 72. 18. LaCourse, W.R., Pulsed Electrochemical Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997, p. 70. 19. Gilman, S., Electroanalytical Chemistry, vol. 2, Bard, A.J. (Ed.), Marcel Dekker, New York, 1967. 20. Hughes, S., Meschi, P.L., and Johnson, D.C., Anal. Chim. Acta 132, 11, 1981. 21. LaCourse, W.R., Pulsed Electrochemical Detection in High Performance Liquid Chromatography, John Wiley & Sons, New York, 1997. 22. Ewing, A.G., Dayton, M.A., and Wightman, R.M., Anal. Chem., 53, 1842, 1981. 23. Tengyl, J., Electrochemical Detectors, Ryan, T.H. (Ed.), Plenum Press, New York and London, 1984. 24. Vandeberg, P.G., Kowagoe, J.L., and Johnson, D.C., Anal. Chim. Acta, 260(1), 1, 1992. 25. Paskach, T.J., Lieker, P.J., and Thielecke, K., Carbohydr. Res., 215, 1, 1991. 26. LaCourse, W.R., Jackson, W.A., and Johnson, D.C., Anal. Chem., 61(22), 2466, 1989. 27. Better Solutions for Food and Beverage Analysis, 2nd ed., Dionex Corporation, Sunnyvale, CA, 1997. 28. Dionex Technical Manual, A Tradition in Innovation, Dionex Corporation, Sunnyvale, CA, 2001, p. 33. 29. Dionex Technical Manual, Ion Chromatography into the 21st Century, Dionex Corporation, Sunnyvale, CA, 1999, p. 4. 30. Kim, H.J. and Kim, Y.K., J. Food Sci., 51, 1380, 1986. 31. Wagner, H.P. and McGarrity, M.J., J. Am. Soc. Brew. Chem., 50, 1, 1992. 32. Wagner, H.P., J. Am. Soc. Brew. Chem., 53, 82, 1995. 33. IICS Workshop Manual, Ion Chromatography, Dionex Corporation, Sunnyvale, CA, 2005, p. 18. 34. Baluyot, E. and Hartford, C.G., J. Chromatogr. A, 739, 217, 1996. 35. Dobberpuhl, D.A., Hoekstra, J.C., and Johnson, D.C., Anal. Chim. Acta, 322, 55, 1996. 36. Draisci, R., Cavalli, S., Lucentini, L., and Stacchini, A., Chromatographia, 35, 9, 1993. 37. Welch, L.E., LaCourse, W.R., Mead Jr, D.A., and Johnson, D.C., Talanta, 37, 377, 1990. 38. Henshall, A., High-performance anion exchange chromatography with pulsed amperometric detection (HPAE-PAD): A powerful tool for the analysis of dietary ﬁber and complex carbohydrates, Complex Carbohydrates in Foods, Cho, S.S., Prosky, L., and Dreher, M. (Eds.), Marcel Dekker, Inc., New York, 1999, p. 267. 39. Low, N.H., Am. Lab., 28(6), 35M, 1996. 40. Swallow, K.W. and Low, N.H., J. Agric. Food Chem., 38(9), 1828, 1990. 41. Swallow, K.W. and Low, N.H., J. AOAC Int., 77(3), 695, 1994.

Otles/Handbook of Food Analysis Instruments 45660_C009 Final Proof page 196 16.7.2008 10:47pm Compositor Name: MSubramanian

196 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52.

Handbook of Food Analysis Instruments Stuckel, J.G. and Low, N.H., J. Agric. Food Chem., 43(2), 3046, 1995. Bernarl, J.L., Del Nozal, M.J., Toribio, L., and Del Alamo, M., J. Agric. Food Chem., 44(2), 507, 1996. Prodolliet, J., Bugner, E., and Feinberg, M., JAOAC, 78(3), 768, 1995. Peschet, J.L. and Giacalone, A., Ind. Aliment Agric., 108(7–8), 583, 1991. Edwards, P. and Haak, K.K., Am. Lab., April, 78, 1993. Rocklin, R.D. and Pohl, C.A., J. Liq. Chromatogr., 6(9), 1577, 1983. Ichiki, H., Semma, M., Sekiguchi, Y., Nakamura, M., and Ito, Y., Nippon Shokuhin Kagaku Gakkaishi, 2(2), 119, 1996. Pollman, R.M., JAOAC, 72(3), 425, 1989. Koizumi, K., Kubota, Y., Tanimoto, T., and Okada, Y., J. Chromatogr., 464(2), 365, 1989. Koizumi, K., Fukuda, M., and Hizukuri, S., J. Chromatogr., 585(2), 233, 1991. Wong, K.S. and Jane, J., J. Liq. Chromatogr., 18, 63, 1995.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 197 16.7.2008 10:49pm Compositor Name: MSubramanian

Spectrometry and 10 Mass Hyphenated Instruments in Food Analysis Tomas Cajka, Jana Hajslova, and Katerina Mastovska CONTENTS 10.1

Introduction ........................................................................................................................ 198 10.1.1 Mass Spectrum ..................................................................................................... 198 10.2 Instrumentation .................................................................................................................. 199 10.2.1 Ion Sources for Gas Chromatography.................................................................. 201 10.2.1.1 Electron Ionization............................................................................... 201 10.2.1.2 Chemical Ionization ............................................................................. 202 10.2.2 Ion Sources for Liquid Chromatography ............................................................. 203 10.2.2.1 Electrospray Ionization ........................................................................ 204 10.2.2.2 Atmospheric Pressure Chemical Ionization......................................... 204 10.2.2.3 Atmospheric Pressure Photoionization ................................................ 205 10.2.3 Ion Source for ICP-MS ........................................................................................ 205 10.2.4 Ion Sources for Direct MS Analysis .................................................................... 205 10.2.4.1 Matrix-Assisted Laser Desorption Ionization ...................................... 205 10.2.4.2 Desorption Electrospray Ionization and Desorption Atmospheric Pressure Chemical Ionization .............................................................. 206 10.2.4.3 Direct Analysis in Real Time .............................................................. 206 10.2.5 Mass Analyzers .................................................................................................... 207 10.2.5.1 Quadrupole........................................................................................... 211 10.2.5.2 Quadrupole Ion Trap (3D Trap) .......................................................... 213 10.2.5.3 Linear Quadrupole Ion Trap (2D Trap)............................................... 213 10.2.5.4 Time-of-Flight...................................................................................... 213 10.2.5.5 Magnetic Sector ................................................................................... 215 10.2.5.6 Fourier Transform Ion Cyclotron Resonance ...................................... 217 10.2.5.7 Orbitrap ................................................................................................ 218 10.2.5.8 Hybrid Instruments .............................................................................. 218 10.2.6 Detectors............................................................................................................... 219 10.2.7 Miscellaneous ....................................................................................................... 220 10.2.7.1 High-Field Asymmetric Waveform Ion Mobility Spectrometry ......... 220 10.2.7.2 Supersonic Molecular Beam MS Interface.......................................... 220 10.2.7.3 Imaging Mass Spectrometry ................................................................ 220 10.3 Food Analysis Applications............................................................................................... 221 10.3.1 Natural Substances in Food.................................................................................. 221 10.3.2 Food Contaminants .............................................................................................. 222

197

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 198 16.7.2008 10:49pm Compositor Name: MSubramanian

198

Handbook of Food Analysis Instruments

10.3.3 Processing Contaminants...................................................................................... 224 10.3.4 Migrants from Packaging ..................................................................................... 225 10.4 Conclusion and Future Trends........................................................................................... 225 Acknowledgments......................................................................................................................... 226 References ..................................................................................................................................... 226

10.1 INTRODUCTION Mass spectrometry (MS) has come a long way since the record of the ﬁrst mass spectra of a simple low-molecular weight substance by Thomson in 1912 [1]. Especially over the past decades, MS has been the subject of many developments. Particularly, the hyphenation of MS to capillary gas chromatography (GC) and liquid chromatography (LC) and also the development of novel ionization techniques caused extensive spreading of MS to food analysis. Also, the introduction of relatively inexpensive quadrupole and ion trap mass analyzers and, at the end of the last century, the rediscovery of time-of-ﬂight mass analyzers allowed the use of this sophisticated instrumental technique in both research as well as routine applications [2–4]. In this chapter, we would like to give the readers basic information of crucial components of various MS instrumentation, introducing also the up-to-date trends and advancements in this rapidly developing area.

10.1.1 MASS SPECTRUM MS is an instrumental technique based on the separation of ions in vacuum, in the gas phase, according to their mass-to-charge ratios (m=z). The analyzed molecules have to be ﬁrst ionized, forming a molecular or pseudomolecular ions, depending on the ionization technique (see below for details). If the (pseudo)molecular ion contains sufﬁcient internal energy, it undergoes further fragmentation, which leads to formation of fragment ion(s). Generally, each fragment ion derived from the molecular ion can again undergo fragmentation. All these formed ions are separated in the mass spectrometer according to their m=z and are detected. A mass spectrum of the molecule is then produced, which represents a distribution chart of the abundances (y-axis) of ions versus their m=z values (x-axis). The peak with the highest intensity in the spectrum is called the base peak and the spectrum is generally normalized to the abundance (intensity) of this peak (i.e., its intensity is assigned 100%). A mass spectrum obtained under the conditions of electron ionization (EI) can be interpreted to provide structural information as demonstrated in Figure 10.1 on an example of a pesticide dichlorvos. Molecular ion is undoubtedly the single most important piece of information because it determines the molecular weight (MW). Mass spectrometrists use isotopic MW, which is based on the isotopic weights of the most common (abundant) isotope of each element in the molecule. Whereas, average MW, based on the elements’ average atomic weights in nature, is typically used in other ﬁelds. For instance, the isotopic MW of dichlorvos is 220 Da, whereas the average MW is 221 Da (when expressed in integer numbers). It should be noted that, for high-resolution mass spectrometric instruments, the isotopic MW is typically expressed in a decimal form (e.g., 219.9459 Da for the example of dichlorvos given previously). The molecular ion is usually the most abundant peak that appears in the heaviest mass isotope pattern in the spectrum (ion m=z 220 in Figure 10.1). In some cases, however, fragmentation can be too extensive, leaving little or no trace of a molecular ion, which makes the determination of MW difﬁcult or even impossible. In such cases, ‘‘soft’’ ionization techniques are typically applied leading to enhanced formation of the (pseudo)molecular ion (Section 10.2.1.2). Isotopic peaks, resulting from the natural isotope abundances of the individual elements, also provide very useful information. As an example can be used natural chlorine that consists of 75% 35 Cl-isotopes and 25% 37Cl-isotopes and, consequently, each fragment ions containing chlorine can

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 199 16.7.2008 10:49pm Compositor Name: MSubramanian

199

Mass Spectrometry and Hyphenated Instruments in Food Analysis

100 Normalized abundance

Molecular ion 220

Base peak 109

222

220 185

224 50

Fragment ions

O

185

79

47 83 93

0 40

60

128

145

80 100 120 140

174

CI O

Fragment ions

79 60

109

P

O

CI

O

220

160 180 200 220 m/z

FIGURE 10.1 Mass spectrum of dichlorvos (MW ¼ 220), obtained under the conditions of EI, and its main characteristics.

be identiﬁed by its typical chlorine isotopic pattern. In Figure 10.1, these ion patterns are represented by the ions m=z 185 and 187 of the fragment containing one chlorine atom, i.e., [C4 H7 35 ClO4 P]þ and [C4 H7 37 ClO4 P]þ , and the ions m=z 220, 222, and 224 in the molecular ion pattern containing two chlorine atoms, i.e., [C4 H7 35 Cl2 O4 P]þ , [C4 H7 35 Cl37 ClO4 P]þ , and [C4 H7 37 Cl2 O4 P]þ . On the other hand, the base peak m=z 109, i.e., [C2 H6 O3 P]þ , does not contain any chlorine atom and consists only of elements that do not offer characteristic ion patterns, thus only one dominating peak is present (the other isotopic peaks have intensity <2%). The isotopic peaks can be obtained also for other elements such as bromine, sulphur, carbon, etc., therefore introduce very characteristic patterns, as shown in Figure 10.2. On the contrary, mono-isotopic elements such as ﬂuorine, iodine, phosphorus, and also hydrogen provide only a single peak in the mass spectrum.

10.2 INSTRUMENTATION A mass spectrometer consists of three fundamental parts: an ion source, a mass analyzer, and a detector (Figure 10.3). Initially, the analyzed sample has to be introduced into the ion source of the instrument, where the sample molecules undergo the ionization process. These ions are then extracted into the analyzer where they are separated according to their m=z. The separated ions

100 80 60 40 20 0 CI

CI2

CI3

CI4

Br

Br2

Br3

Br4

100 80 60 40 20 0

FIGURE 10.2

Isotopic patterns for 1–4 chlorine and bromine atoms in an ionized molecule=fragment.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 200 16.7.2008 10:49pm Compositor Name: MSubramanian

200

Handbook of Food Analysis Instruments

Sample

Data analysis

Chemical ionization

LC

Electrospray ionzation Atmospheric pressure chemical ionization Atmospheric pressure photo ionization

Direct

Matrix-assisted laser desorption ionization

Scanning

Electron ionization

Mass analyzer

Non-scanning

GC

Ion source

Detector

Quadrupole

Electron multiplier

Quadrupole ion trap

Photon multiplier

Linear quadrupole ion trap

Microchannel plate

Magnetic sector Time-of-flight Fourier transform ion cyclotron resonance Orbitrap

Desorption electrospray ionization Desorption atmospheric pressure chemical ionization Direct analysis in real time

FIGURE 10.3

Overview of mass spectrometric technique setups.

are detected and the signal containing information about the m=z values is stored together with their relative abundance (intensity) as a mass spectrum. To allow the ions traveling from one end of the instrument to the other part without any hindrance originated from air molecules, both the analyzer and detector (and often the ion source as well) are kept under high vacuum. The sample can be introduced to the ion source directly, or after its previous separation involving either GC or high-performance LC (HPLC); thus, the sample is separated into a series of components, which then enter the mass spectrometer for their subsequent analysis. There are several ionization methods available, each having its own advantages and limitations. The ﬁtness of the ion source depends on factors such as the polarity, molecular weight, thermal lability of the analytes, and complexity of the examined sample (Figure 10.4). In GC–MS, electron ionization and chemical ionization (CI) represent the fundamental ionization techniques. On the basis of the scientiﬁc literature abstracted in SciFinder Scholar the EI was used in approx. 95% of all food GC–MS applications, while the rest of applications (5%) employed CI. In LC–MS, electrospray ionization (ESI), atmospheric pressure chemical ionization (APCI), and atmospheric pressure photoionization (APPI) represent ionization techniques widely employed. On the basis of SciFinder Scholar, ESI was the most often reported ionization technique in food LC–MS applications (approx. 80%), while APCI and APPI were more rarely employed (18% and 2%, respectively). In addition, the direct ionization techniques such as matrix-assisted laser desorption ionization (MALDI), desorption electrospray ionization (DESI), desorption atmospheric pressure chemical ionization (DAPCI), and direct analysis in real time (DART) can be employed for sample characterization. A wide range of mass analyzers are nowadays available. While quadrupole, quadrupole ion trap, linear quadrupole ion trap, and magnetic sector represent scanning instruments, in which the mass spectrum is recorded by gradual changing of electrical and magnetic ﬁeld; the other group of mass

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 201 16.7.2008 10:49pm Compositor Name: MSubramanian

201

Mass Spectrometry and Hyphenated Instruments in Food Analysis 100000 LC–MS ESI

Molecular weight

10000

1000

LC–MS APPI LC–MS APCI

GC–MS EI, CI

100

10 Nonpolar

Very polar Polarity of analyte

FIGURE 10.4

Application scope of various ionization techniques coupled to GC and LC.

analyzers such as time-of-ﬂight, Fourier transform ion cyclotron resonance (FT-ICR) analyzer, and orbitrap are nonscanning instruments, in which entire mass spectra are obtained simultaneously. The detector allows to monitor the ion current and, after its ampliﬁcation, to record the data in the form of mass spectra. The most common types of MS detectors include the electron multiplier, the photon multiplier, and the microchannel plate (MCP) detector.

10.2.1 ION SOURCES 10.2.1.1

FOR

GAS CHROMATOGRAPHY

Electron Ionization

The EI (EI, formerly called electron impact) process starts by the acceleration of electrons through an electric ﬁeld. These energetic electrons interact with neutral molecules (M). Upon collision, the molecule loses an electron and becomes a particle with an odd number of electrons and a positive charge. This radical cation (M .þ ) is typically called molecular ion. M þ e ! M .þ þ 2 e If the molecular ion contains sufﬁcient internal energy, it undergoes fragmentation, which leads to the formation of fragment ion(s); either a radical (R ) and an ion with an even number of electrons (EE þ), or a neutral molecule (N) and a new (odd ion) radical cation (OE þ) can be considered. Each fragment ion derived from the molecular ion can undergo further fragmentation. .

.

M .þ

! !

EE þ þ R. OE .þ þ N

The fragmentation reactions are initiated at the site of the unpaired electron and the positive charge in the precursor ion. The most preferred radical and charge site in the molecular ion is assumed to be a loss of the molecule’s electron of lowest energy in the order of s < p < n electrons originating from sigma bonds, double bonds, or nonbonding electron pairs. The energy required for the ionization process is called the ionization energy and can vary between 0 and over 100 eV. A positively charged species (i.e., molecular ions) begin to appear at

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 202 16.7.2008 10:49pm Compositor Name: MSubramanian

202

Handbook of Food Analysis Instruments

low intensities at around 10 eV. Increasing of this ionization potential leads to the formation of fragment ions. Standard mass spectra are obtained typically at 70 eV because maximum ion intensity is observed at this value, and mass spectra are reproducible and characteristic independently of the type of instrument. Under standardized conditions, the electrons are produced in the EI source from a heated ﬁlament and are energized by accelerating them through a potential of 70 eV. The moving electrons directed across the source are forced into a beam by a magnetic ﬁeld. This electron beam then interacts with sample molecules that have been vaporized. The ions are extracted from the source by an electric ﬁeld and passed into the analyzer as an ion beam or as an ion packet. Under these circumstances, virtually identical spectra can be obtained regardless of mass spectrometers equipped with the EI source as long as the electron energy is the same. This fact has led to the compilation of extensive mass spectra libraries (e.g., the NIST—National Institute of Standard and Technology library and the Wiley library). EI is therefore preferred for the identiﬁcation of unknowns, determination of molecular structure, and conﬁrmation of target analyte identity through consistent ion abundance ratios and library spectra matching. 10.2.1.2

Chemical Ionization

CI represents a low energy or ‘‘soft’’ ionization technique and is therefore very suitable for those less volatile or thermally labile molecules that do not yield molecular ions by EI. For CI, a suitable reagent gas is introduced into the ion source at a concentration that largely excesses the amount of the analytes (e.g., 104:1). The reagent gas is usually ionized as in EI. The formation of primary ions is followed by reactions between those primary ions and the noncharged molecules of gas producing the CI reagent ions as well as the thermal electrons. In positive chemical ionization (PCI), the ion source is ﬁlled with a reagent gas (e.g., methane), at a relatively high pressure (0.1–100 Pa), which undergoes EI, producing an excess of reagent ions. CH4 þ e ! CH4.þ þ 2e Sample molecules are subsequently ionized by the reagent gas ions via proton transfer, producing pseudomolecular ions [M þ H]þ and, depending on the choice of a reagent gas, adduct ions may be formed. . CH4.þ þ CH4 ! CHþ 5 þ CH3

þ M þ CHþ 5 ! [M þ H] þ CH4

The pseudomolecular ions generally have low internal energy and, consequently, they are less prone to the fragmentation than the molecular ions generated under EI conditions; in this way unambiguous molecular weight information can be obtained. Due to little or no fragmentation, PCI is less suitable for conﬁrmation. This can be, however, highly appreciated in some analyses, since the pseudomolecular ion is more intensive and speciﬁc than any lower-mass fragment ions. In other words, PCI can offer both increased sensitivity and improved detectability due to reduced chemical noise from background or co-eluting analytes, resulting in increased signal to noise ratio (S=N). The production of a large population of low-energy electrons during the CI operation utilizes another ionization technique: negative chemical ionization (NCI), alternatively called electron capture negative ionization (ECNI) or negative ion chemical ionization (NICI). The basic mechanism of this technique is similar to that of an electron capture detector: a low-energy electron is captured by an electronegative sample molecule (ABC), forming the molecular anion (by the resonance capture, dissociative capture, or ion-pair formation mechanisms), which may undergo fragmentation, depending on its structure.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 203 16.7.2008 10:49pm Compositor Name: MSubramanian

203

Mass Spectrometry and Hyphenated Instruments in Food Analysis 79

152

79

150

32,000

20,000

400,000

12,000

24,000 115 16,000

107 149 4,000 119 264 51 182 236 0 100 200 300

(A)

Mass range

Abundance

Abundance

Abundance

264

45

35 200,000

8,000 238

70

300 0

(B)

100

200

0

300

Mass range

(C)

182

200 100 Mass range

300

FIGURE 10.5 The (A) EI, (B) PCI, and (C) NCI mass spectra of a pesticide captan (MW ¼ 299). The ionization techniques inﬂuence the extent of fragmentation and ion intensities. Methane was used as a CI reagent gas.

ABC þ e ! ABC ABC þ e ! AC þ B ABC þ e ! ABþ þ C þ e The main advantages of NCI compared to EI and PCI include a possibility of up to 100-fold improvement in sensitivity, and higher degree of selectivity, since only a limited number of analytes, such as those containing a halogen atom, a nitro group, or an extended aromatic ring system, are prone to efﬁcient electron capture. Figure 10.5 compares the effect of EI, PCI, and NCI ionization techniques on the fragmentation and ion intensities in the resulting GC–MS spectra of a selected analyte.

10.2.2 ION SOURCES

FOR

LIQUID CHROMATOGRAPHY

Compared to GC–MS, the LC–MS coupling was slower in its development mainly due to the technical problems that derived from the introduction of liquids into a mass spectrometer with a high vacuum required for separation of ions. The online LC–MS coupling requires the availability of an interface or restrictor that allow to keep the total mass ﬂow entering the mass analyzer at values compatible with the pumping capacity of the vacuum system. The main interfaces adopted for this purpose include the use of the moving belt interface, particle beam interface, direct liquid introduction, continuous ﬂow fast atom bombardment, thermospray, and atmospheric pressure ionization (API). In the following paragraphs, the attention will be paid to the most frequently used LC–MS ion sources coupled in current practice: ESI, APCI, and APPI. It should be noted that LC–API-MS interface is considerably inﬂuenced by the composition of liquid entering the detector, i.e., the type and amount of organic mobile phase modiﬁers and volatile buffers, and also the type and amount of sample matrix components. These substances present in the injected sample can cause serious quantiﬁcation problems when co-eluted with the analyte of interest; either by suppression or enhancement of the analyte signal. It is assumed that matrix components inﬂuence the efﬁciency of the ionization processes in API interface (causing a mutual positive or negative effect in the amount of ions formed from the target analyte). Those components may also inﬂuence the ion formation in the ionization process by altering the surface tension of electrospray droplets and by building adduct ions or ion pairs with the analytes. As a result of matrix suppression=enhancement phenomena, the response of an analyte in pure solvent standard may differ signiﬁcantly from that in matrix sample. Therefore for quantiﬁcation purposes, calibration using solvent-based external standards can provide biased results, especially in the analysis of complex samples, such as food [5].

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 204 16.7.2008 10:49pm Compositor Name: MSubramanian

204

Handbook of Food Analysis Instruments

10.2.2.1

Electrospray Ionization

ESI represents a very soft ionization technique, resulting typically in a single peak of (de) protonated molecular ion ([M þ H]þ in positive-ion mode and [M H] in negative-ion mode) or molecular ion adducts with, e.g., sodium, potassium, or ammonia. ESI is suitable for ionic analytes and other relatively polar compounds with MW ranging from less than 100 Da to more than 1,000,000 Da. In the analysis of large molecules, such as proteins, ESI generates multiply charged ions, which reduces the m=z ratio, effectively extending the mass range of mass spectrometers. During ESI, the sample is dissolved in a polar solvent and introduced though a narrow stainless steel capillary (nebulizing needle, ESI needle) either from a syringe pump or as the efﬂuent ﬂow from a LC. Flow rates are typically in a range of 1 mL=min to 1 mL=min. Using a high voltage (typically 2.5–4 kV), applied to the tip of the capillary, it results in the formation of a strong electric ﬁeld and the sample emerging from the tip is dispersed into an aerosol, which consists of highly charged droplets. A nebulizing gas (usually nitrogen) ﬂowing around the outside of the capillary helps to direct the spray towards the mass analyzer. By solvent evaporation, the charged droplets diminish their size by the assistance of a drying gas (warm ﬂow of nitrogen), which passes across the front part of the ion source (Figure 10.6). After the droplets reach the point that the surface tension cannot sustain the charge (so-called the Rayleigh limit), ‘‘Coulombic explosion’’ occurs and the droplets are disintegrated. This process repeats until analyte ions evaporate from the droplet. The (multiply) charged analyte ions then pass through a sampling cone or oriﬁce into a vacuum region, and from there through a small aperture into the mass analyzer. 10.2.2.2

Atmospheric Pressure Chemical Ionization

APCI is a somewhat similar ionization technique to CI in GC–MS. However, while the latter ionization requires a vacuum, the APCI occurs at atmospheric pressure. In APCI, the analyte solution is introduced from a direct inlet probe or an LC eluate at a ﬂow rate between 2 mL=min and 2 mL=min into a pneumatic nebulizer, where it is converted into a vapor by a nitrogen beam. Created droplets are then displaced by the gas ﬂow through a heated quartz tube (desolvation=vaporization chamber). After desolvation, the gas-phase solvent molecules are ionized by primary ions (such as N2.þ ) produced by the corona discharge, which has the same function as the electron ﬁlament in CI. The primary ions collide with the vaporized solvent molecules to form secondary reactant gas ions, which then undergo repeated collisions with the analyte resulting in the proton transfer and formation of analyte ions. Similarly to ESI, the APCI spectra are characterized by predominant molecular species and adduct ions with very little fragmentation. The ions then enter the mass spectrometer through a tiny inlet and are focused towards the analyzer.

Sample introduction

FIGURE 10.6

Multiply charged droplet

Multiply charged droplets Analyte + ++ + + + + + ++ + + ions + + + + + + + + + ++ Coulombic Solvent To mass + + + + + analyzer + explosion + + + + evaporation ++ + + + ++ + + + + ++ + + + ESI needle Drying gas Analyte molecule

Nebulizing gas

Illustration of droplet and ion formation in electrospray ionization (ESI).

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 205 16.7.2008 10:49pm Compositor Name: MSubramanian

205

Mass Spectrometry and Hyphenated Instruments in Food Analysis

10.2.2.3

Atmospheric Pressure Photoionization

APPI represents a novel, alternative ionization method for LC–MS. The ionization process in APPI is initiated by 10 eV photons, which are emitted by a krypton discharge lamp. The photons can ionize compounds that possess ionization energies below their energy (10 eV), which includes most of the larger molecules (analytes), but leaves out most of the typically used gases and solvents. Under these conditions, the analytes can be ionized selectively, with minimum background interference. In addition, as the ionization of the analytes is dependent on the ionization energy of the analyte rather than its proton afﬁnity like in ESI and APCI, the ionization of molecules of relatively low polarity is also possible.

10.2.3 ION SOURCE

FOR

ICP-MS

The sample for inductively coupled plasma (ICP)-MS has to be introduced into the plasma as an aerosol, usually as a liquid sprayed through a nebulizer. Under the conditions of high temperature (6,000–10,000 K) in the plasma, the sample is atomized and ionized, creating positively charged atomic ions. While the larger aerosol droplets are removed from the gas stream by a spray chamber, the remaining smaller droplets are swept into the central channel of an argon plasma followed by their drying, decomposition, and dissociation into individual atoms in the plasma. These atoms are converted to positively charged ions before their extraction into the vacuum system for the detection creating the ICP-MS spectrum, which represents the elemental composition of the sample. The most important features of ICP-MS, compared to other elemental analysis techniques, are extreme sensitivity, selectivity, and simultaneous multi-element capability. The only elements that cannot be directly measured by this technique are hydrogen, helium, neon, argon, and ﬂuorine [6].

10.2.4 ION SOURCES 10.2.4.1

FOR

DIRECT MS ANALYSIS

Matrix-Assisted Laser Desorption Ionization

MALDI is mainly suitable for thermolabile and=or nonvolatile organic compounds with high molecular mass (e.g., peptides, proteins, glycoproteins, oligoproteins, and oligonucleotides). Before the analysis, the sample is pre-mixed with a solvent containing highly UV-absorbing matrix compound. This mixture is dried before the analysis removing the liquid solvent used in preparation. The obtained deposit of matrix–analyte solid solution then undergoes the ablation of bulk portions of this solid solution by intense pulses of laser for a short duration (Figure 10.7). The irradiation by the laser induces rapid heating of the crystals as a result of the accumulation of a large amount of energy in the condensed phase through excitation of the matrix molecules. The rapid heating leads

Analyte/matrix solid solution

Laser beam Matrix ion Proton transfer

⫺

+ Desorption

Desolvation Analyte ion

Sample plate

FIGURE 10.7

Principle of matrix-assisted laser desorption ionization (MALDI).

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 206 16.7.2008 10:49pm Compositor Name: MSubramanian

206

Handbook of Food Analysis Instruments

to sublimation of the deposits and their expansion into the gas phase followed by evaporation of the matrix molecules away from the clusters to leave the free analyte ions in the gas phase. The analyte ions, created by proton transfer from ionized matrix molecules, are then extracted into the mass spectrometer for the analysis [7]. In most MALDI instruments, the desorption=ionization process takes place in vacuum, thus the sample is not accessible during the analysis (similarly to EI or CI in GC–MS). Although atmospheric pressure MALDI (APMALDI) has been recently introduced, it still requires sample dilution=coating with a UV-absorbing matrix compound and operation under enclosed conditions to protect the operator from potential exposure to laser radiation [8]. Also, APMALDI and atmospheric pressure ionization techniques, such as ESI, APCI, or APPI, require that samples have to be exposed to elevated temperatures and electrical potentials, ultraviolet irradiation, laser radiation, or a high-velocity gas stream. Recently introduced ionization methods described below in Sections 10.2.4.2 and 10.2.4.3 do not pose these limitations, requiring essentially no sample preparation and allowing full access to the sample while mass spectra are being recorded. 10.2.4.2

Desorption Electrospray Ionization and Desorption Atmospheric Pressure Chemical Ionization

DESI is carried out by directing electrosprayed charged droplets and ions of solvent onto the analyzed sample surface (Figure 10.8). The spray is directed at an insulating sample or an analyte deposited on an insulating surface such as polytetraﬂuoroethylene. The desorbed ions are sampled with a commercial mass spectrometer equipped with an atmospheric interface connected via a transfer line made either of metal or an insulator. Similarly to ESI, the resulting mass spectra show mainly singly or multiply charged molecular ion of the analyte [9]. DAPCI offers an alternative option for those compounds that do not provide sufﬁcient ion intensity by DESI. DAPCI has been shown to provide increased sensitivity for compounds of moderate polarity. The DAPCI technique uses nitrogen sheath gas and a solvent from which ions are produced by a corona discharge. Reagent ions formed in the corona discharge region react with desorbed analyte molecules, creating protonated or deprotonated molecular ions depending on the polarity of the ionization mode [10]. 10.2.4.3

Direct Analysis in Real Time

DART is based on the atmospheric pressure interactions of long-lived electronic excited-state atoms or vibrionic excited-state molecules with the sample and atmospheric gases. A gas (typically helium or nitrogen) ﬂows through a chamber where an electrical discharge produces ions, electrons, and excited-state (metastable) atoms and molecules. Most of the charged particles are removed as

HV power supply Solvent N2

Solvent capillary

Ion transfer line

V

Desorbed ions Sample

Gas capillary Spray droplets Surface

FIGURE 10.8

Desorption electrospray ionization (DESI) schematic.

Inlet of MS

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 207 16.7.2008 10:49pm Compositor Name: MSubramanian

207

Mass Spectrometry and Hyphenated Instruments in Food Analysis OOO OOL

Linoleic acid

LLL LLO Linolenic acid 296

Oleic acid

298

300

200

302

m /z

400

840

600

860

880

900

800

920

m /z

940

m /z

1000

FIGURE 10.9 DART mass spectrum of olive oil. (Reproduced with permission from JEOL(Europe), Zaventem, Belgium.)

the gas passes through perforated lenses or grids and only the neutral gas molecules (including metastable species) remain. A perforated lens or grid at the exit of DART acts as an electrode to promote ion drift toward the oriﬁce of the atmospheric pressure interface of the mass spectrometer. DART produces relatively simple mass spectra characterized by [M]þ or [M þ H]þ in positive-ion mode, and [M] or [MH] in negative-ion mode. As opposed to DESI, DART does not use any solvent, but its applicability is probably limited to smaller organic molecules [11]. Figure 10.9 shows the DART mass spectrum of olive oil characterizing its lipid composition.

10.2.5 MASS ANALYZERS Once the molecules are converted into charged particles, they undergo mass analysis in a mass analyzer. The choice of an optimal mass analyzer is determined by several key operating parameters that are brieﬂy summarized below [12–14]. The general speciﬁcations and features of selected mass analyzers hyphenated to both GC and LC are shown in Tables 10.1 and 10.2. .

.

Mass range represents the range of m=z over which a mass analyzer can separate and then detect=record ions. In GC, analyte volatility=thermolability effectively dictates the upper mass limit, thus the majority of mass spectrometers combined with GC operate typically up to m=z 1000. In LC, however, the mass range is often increased up to several orders higher m=z. Mass resolution=mass resolving power is the ability of a mass analyzer to separate two ions of similar mass. On the basis of the recent IUPAC (the International Union of Pure and Applied Chemistry) provisional recommendation on standard deﬁnitions of MS terms (drafted in 2006 [15]), mass resolution is deﬁned as the smallest mass difference Dm between two equal magnitude peaks, so that the valley between them is a speciﬁed fraction of the peak height. Mass resolving power is then deﬁned as the observed mass divided by the difference between two masses that can be separated, m=Dm. The procedure by which Dm was obtained and the mass at which the measurement was

þ MSn, n ¼ 2–10

EI, PCI, NCI Unit mass

4–5

Up to 1,000 Da 0.1–0.2 Da 5 scans=s (m=z 50–550 Da) Full scan, SRM, MRM >pg in full scan

Ion Trap

EI, PCI, NCI >10,000 (10% valley deﬁnition) þþþþ Only with special conﬁguration

>5

fg in SIM

Up to 4,000 Da <5 ppm 7 scans=s (m=z 50–550 Da) Full scan, SIM

DF Magnetic Sector

þþþ None

EI Unit mass

4

pg

500 spectra=s (m=z 50–550 Da) Full spectra

Up to 1,000 Da

High-Speed TOF

þþþ None

EI, PCI, NCI >7,000 (FWHM)

4

fg–pg

Up to 1,500 Da <5 ppm 20 spectra=s (m=z 50–550 Da) Full spectra

High-Resolution TOF

Note: EI, electron ionization; MS, mass spectrometry; NCI, negative chemical ionization; PCI, positive chemical ionization; SIM, selected ion monitoring; SRM, selected reaction monitoring; MRM, multiple reaction monitoring; pg, picogram; fg, femtogram.

þþþ MS2

þ None

>5

Linear dynamic range (orders of magnitude) Versatility Mass resolution=mass resolving power Cost MS=MS EI, PCI, NCI Unit mass

>pg in full scan fg in SIM

Sensitivity

Full scan, SIM, SRM, MRM >pg in full scan 5

Up to 1,500 Da 0.1–0.2 Da

Triple Quadrupole

208

EI, PCI, NCI Unit mass

Up to 1,200 Da 0.1–0.2 Da 12,500 amu=s (i.e., theoretically 25 Hz for m=z 50–550 Da) Full scan, SIM

Quadrupole

Mass range Mass accuracy Maximal spectral acquisition speed Acquisition mode

Criteria

TABLE 10.1 General Speciﬁcations and Features of Selected Mass Analyzers Coupled to GC

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 208 16.7.2008 10:49pm Compositor Name: MSubramanian

Handbook of Food Analysis Instruments

þþþ MS2

þ Nonea

þ MSn, n ¼ 2–10

4–5 ESI, APCI Unit mass

2 scans=s (m=z 50–1,000 Da) >pg in full scan

Up to 2,000–6,000 Da 0.1–0.2 Da Full scan, SRM, MRM

Ion Trap

þþþ Nonea

4 ESI, APCI, APPI >12,000 FWHM

>3 ESI >2,000 FWHM þþþ Nonea

20 spectra=s (m=z 50–1,000 Da) pg

Up to 30,000 Da <3 ppm Full spectra

High-Resolution TOF

100 spectra=s (m=z 50–1,000 Da) >pg

Up to 6,000 Da <15 ppm Full spectra

High-Speed TOF

Note: APCI, atmospheric pressure chemical ionization; APPI, atmospheric pressure photoionization; ESI, electrospray ionization; FWHM, full width at half maximum; MS, mass spectrometry; SIM, selected ion monitoring; TOF, time-of-ﬂight; SRM, selected reaction monitoring; MRM, multiple reaction monitoring; pg, picogram; fg, femtogram. a MS=MS option only when hyphenated, such as Q-TOF.

Linear dynamic range Versatility Mass resolution=mass resolving power Cost MS=MS

>pg in full scan 5 ESI, APCI, APPI Unit mass

6 scans=s (m=z 50–1,000 Da) >pg in full scan 5 ESI, APCI Unit mass

Maximal spectral acquisition speed Sensitivity

Up to 1,000–3,000 Da 0.1–0.2 Da Full scan, SIM, SRM, MRM

Triple Quadrupole

Up to 1,000–3,000 Da 0.1–0.2 Da Full scan, SIM

Quadrupole

Mass range Mass accuracy Acquisition mode

Criteria

TABLE 10.2 General Speciﬁcations and Features of Selected Mass Analyzers Coupled to LC

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 209 16.7.2008 10:49pm Compositor Name: MSubramanian

Mass Spectrometry and Hyphenated Instruments in Food Analysis 209

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 210 16.7.2008 10:49pm Compositor Name: MSubramanian

210

Handbook of Food Analysis Instruments ∆m

Normalized abundance

1.0

Maximum ∆m

0.5

50% of Maximum FWHM 10% Valley 5% of Maximum ∆m 0.0 99

FIGURE 10.10

.

.

.

.

100

101

102

m /z

Relationship between the two deﬁnitions of mass resolution.

made should be reported. It should be noted that mass resolution is also often expressed as m=Dm ratio in current MS practice, i.e., not as Dm as recommended by the IUPAC. Figure 10.10 demonstrates two different ways that are used in practice to obtain Dm and, consequently, to evaluate mass resolution and mass resolving power. The 10% valley deﬁnition is based on the mass difference between two mass peaks separated by a 10% valley, while full width at half maximum (FWHM) deﬁnition expresses Dm as the peak width of a given mass peak measured (in mass units) at 50% of its height. The FWHM deﬁnition usually gives a number twice the magnitude of the 10% valley deﬁnition. The most mass analyzers operate at so-called unit mass resolution (or actually at mass resolving power <1000). High-resolution MS instruments provide improved selectivity, thus reduced chemical noise and, consequently, increased S=N in complex samples. Mass accuracy is the deviation between measured mass (accurate mass) and calculated mass (exact mass) of an ion expressed as an error value (mDa, ppm). This parameter is important for structural interpretation allowing conﬁrmation of the target analyte identity and the calculation of elemental composition of ‘‘unknowns.’’ The unit-resolution mass spectrometers provide mass accuracy of approx. 0.1–0.2 Da (e.g., 200.0 and 200.1 Da), while the high-resolution instruments operate at mass accuracy <5 ppm (e.g., 200.000 and 200.001 Da). Spectral acquisition speed represents time required for recording of a mass spectrum or selected ion(s). For scanning instruments terms scans=s, s=scan, or s=decade are commonly used; for nonscanning instruments a term spectra=s is preferred. Maximal spectral acquisition speed is a critical parameter in detection of very narrow peaks generated during fast chromatographic separation. Selection of acquisition mode allows the use of different modes of mass spectra collection in scanning instruments. Both magnetic sector and quadrupole instruments can be operated in full scan and selected ion monitoring (SIM); in the latter case structural information is sacriﬁced in favor of detection sensitivity. Detectability (or often expressed as sensitivity) of the instrument, which is characterized by the instrument manufacturers as minimal S=N ratio at a given concentration of a reference

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 211 16.7.2008 10:49pm Compositor Name: MSubramanian

Mass Spectrometry and Hyphenated Instruments in Food Analysis

.

.

. .

211

compound. For this purpose, compounds such as octaﬂuoronaphthalene or hexachlorobenzene are typically used in GC–MS, and reserpine in LC–MS. Linear dynamic range is the range of analyte concentrations, over which the detector proportionally responds to concentration changes. For common scanning instruments, this range varies between 5–6 orders of magnitude; the nonscanning instruments, such as TOFMS, offer linear dynamic range of typically 4 orders of magnitude in maximum. Narrow dynamic range makes the quantiﬁcation complicated, especially if samples with largely varying concentrations of target analytes have to be analyzed. Availability of tandem MS function, which is a method involving at least two stages of mass analysis, either in conjunction with a dissociation process or a chemical reaction that causes a change in the mass or charge of particular ion. MS=MS methods involve activation of selected ions (called precursor or parent ions), typically by collision with an inert gas, sufﬁcient to induce fragmentation (resulting in ions called product or daughter ions). Basically, two different approaches in MS=MS exist: in space by coupling of two physically distinct parts of instrument (e.g., in triple quadrupole, see Section 10.2.5.8); or in time by performing a sequence of events in an ion storage device (e.g., in ion trap, see Section 10.2.5.2). The main tandem MS=MS scan modes are (Figure 10.11) (1) product ion scan, which involves selection of an ion of interest, its activation, and mass analysis of the product ions in full scan mode, (2) precursor ion scan represents opposite process compared to the product ion scan, providing information about all precursor ions that react to produce a selected product ion, (3) neutral loss scan is a scan that determines all precursor ions that react to the loss of a selected neutral mass, (4) selected reaction monitoring (SRM) monitors a single transition from a precursor to a product ion, (5) multiple reaction monitoring (MRM) is used if several different reactions are monitored in one time window, and (6) MSn scans, which is commonly applied on ion trap analyzers (Section 10.2.5.2). Versatility, in terms of the use of alternative ionization techniques, extends the scope of applications. Cost is undoubtedly an important factor when considering an MS system purchase. In routine practice, less expensive scanning instruments (quadrupole and ion trap analyzers) are widely used. Although there has been a decrease in cost of other mass analyzers, such as TOF or sector, during the recent years, unfortunately, their cost is still substantially higher (approx. 2–3 times for TOF and approx. 6 times for sector) compared to lowresolution quadrupole or ion trap instruments.

10.2.5.1

Quadrupole

The quadrupole represents the most popular mass analyzer mainly due to its relatively low cost, ruggedness, reliability, and the simplicity of operation. The quadrupole mass analyzer (often considered as a ‘‘mass ﬁlter’’) consists of four (hyperbolic or cylindrical) rods placed in a radial array (Figure 10.12). Opposite rods are charged by positive direct-current (DC) voltage, while adjacent rods have the opposite (negative) charge applied. Ions are introduced into the quadrupole ﬁeld by means of a low accelerating potential. An appropriate combination of DC and radiofrequency (RF) ﬁelds on the quadrupole rods allows to pass only the ions of one particular m=z at a time. Ions with a nonstable trajectory through the quadrupole collide with the quadrupole rods, thus are not detected using the given DC and RF potential settings. The quadrupole can be operated in two modes: (1) full scan of a selected mass range (e.g., m=z 50–550) and (2) SIM. In the latter mode, sensitivity of an analyte is enhanced by monitoring only a few selected m=z ions (typically 1–20), but the spectral information is sacriﬁced, which must be taken into account.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 212 16.7.2008 10:49pm Compositor Name: MSubramanian

212

Handbook of Food Analysis Instruments

Product ion scan

Selected m /z

CID

Scanned

CID

Selected m /z

CID

Scanned m /z = x–a

CID

Selected fragment

Precursor ion scan

Scanned Neutral Ioss scan

Scanned m /z = x Selected reaction monitoring

Selected precursor m /z = a MSn

m /z = b

(MS3)

Selected m /z (MS1)

FIGURE 10.11

CID

Selected m /z (MS2)

CID

Scanned (MS3)

Various setups in tandem MS (CID stands for collision-induced dissociation).

Detector

Quadrupole rods

Exit slit

Ion source

Stable trajectory (detected ions) Source slit

FIGURE 10.12

Nonstable trajectory (not detected ions)

Quadrupole mass analyzer.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 213 16.7.2008 10:49pm Compositor Name: MSubramanian

213

Mass Spectrometry and Hyphenated Instruments in Food Analysis Inlet end-cap electrode

Ring electrode

Exit end-cap electrode

Detector

Ion source

~ V

FIGURE 10.13

10.2.5.2

Quadrupole ion trap mass analyzer.

Quadrupole Ion Trap (3D Trap)

The quadrupole ion trap consists of three cylindrically symmetric electrodes (two end-caps and a ring), see Figure 10.13. The ions of selected m=z range are trapped by an RF potential, which is formed between end-cap electrodes and a ring electrode. Increasing of the RF potential leads to the formation of unstable trajectories of ions that rapidly deﬂect in the direction of the exit end-cap electrode followed by their detection in increasing m=z order. Although the quadrupole ion traps allow both full scan and SIM acquisition, there is no increase in detection sensitivity in SIM mode that is typically observed with quadrupole or sector instruments. Enhanced selectivity (increased of S=N) can be obtained in MSn mode (mostly MS=MS because orders n > 3 are not practical for relatively small molecules). The MS=MS is often performed by means of collision-induced dissociation (CID), where an RF voltage is applied to the end-cap electrodes to isolate precursor ions of a selected m=z value. After ionization, the application of an excitation voltage together with collision with helium buffer gas lead to the formation of monitored product ions. The main advantage of using MS=MS is the signiﬁcant reduction of the chemical noise, which can originate from different sources (e.g., matrix compounds, co-eluting analytes, column bleed, contamination from an ion source). The effect of multiple MS=MS steps on S=N ratio illustrates Figure 10.14. Although each MS step leads to a loss in the analyte signal intensity, the noise decreases even more rapidly, resulting in signiﬁcant improvements in S=N ratio as the number of MS steps increases in MSn [16]. 10.2.5.3

Linear Quadrupole Ion Trap (2D Trap)

The linear quadrupole ion trap differs from the 3D ion trap as it traps ions in the radial dimension by an RF quadrupole ﬁeld and by static DC potentials applied to the ends of a four-rod structure. Upon ejection, ions emerge radially over the length of the quadrupole-rod structure and can be detected using conventional detection system. The 2D trap has several advantages over the 3D trap. Since there is no quadrupole electric ﬁeld in the z direction the ion injection and extraction efﬁciencies can be nearly 100%. In addition, as compared to the 3D trap, which has a limited capacity for ion storage due to its small strapping volume, the 2D trap has almost one order of magnitude increase in ion capacity, thus also in the linear dynamic range [17]. 10.2.5.4

Time-of-Flight

The (orthogonal acceleration) time-of-ﬂight (TOF) mass analyzer consists of a pulsing electrode, a ﬂight tube, and a reﬂectron. The generated ions are accelerated to get constant kinetic energy

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 214 16.7.2008 10:49pm Compositor Name: MSubramanian

214

Magnitude

Handbook of Food Analysis Instruments

Signal

S /N

Noise

1

2 3 MS step

4

FIGURE 10.14 Effect of the number of MS steps in MSn on signal (S), noise (N), and S=N. (Reproduced from Cooks, R.G. and Busch, K.L., J. Chem. Edu., 59, 926, 1982. With permission.)

followed by their ejection into a mass analyzer’s ﬂight (drift) tube using pulsed electric-ﬁeld gradient, which is oriented orthogonally to the ion beam. This orthogonal acceleration has a positive inﬂuence on mass resolving power, which is further enhanced by using a reﬂectron (ion mirror). This device consists of a series of ring electrodes with increasing voltage creating retarding ﬁelds. The ions with higher energy penetrate more deeply inside it, extending the time of their reﬂection. Consequently, the ions of the same m=z value with different initial energies hit the detector at almost the same time (Figure 10.15). In addition, the mass resolving power is substantially improved by making the ions pass twice along the TOF ﬂight tube before reaching the detector. The times of arrival at the detector are proportional to the square root of respective m=z values [14,18]. It should be noted that the mass analyzer efﬁciency of TOFMS instruments is as high as 25% in full spectra acquisition. This value is signiﬁcantly higher as compared to that obtained by scanning instruments such as quadrupole (~0.1%) [19]. This fact implies availability of full spectra information even at ultra-trace levels of a particular compound and, consequently, the possibility to carry out its identiﬁcation on the basis of library search. Furthermore, since the TOF is nonscanning mass analyzer, all ions are recorded simultaneously, thus there are no changes in the ratios of analyte ions across the peak during the acquisition of the mass spectrum and, consequently, no spectral skewing

(1)

(2) Flight tube (3)

Ion source Reflectron

Detector (4)

FIGURE 10.15 Time-of-ﬂight mass analyzer. (1) Pulse of ions from the orthogonal accelerator (spatial focusing); (2) separation of ions according to their ﬂight times; (3) focusing of kinetic energy of ions; and (4) separation of focused ions according to their ﬂight times, which correspond to their weight.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 215 16.7.2008 10:49pm Compositor Name: MSubramanian

215

Mass Spectrometry and Hyphenated Instruments in Food Analysis

Peak true-sample “KM21 - STD 1–1.5 min splitless 50 psig: 1”, peak 106 at 635.975 s (Spec # 1978)

Peak true - sample “KM21 - STD 1 - 1.5 min splitless 50 psig: 1”, peak 104 at 632.775 s (Spec # 1962)

Peak 1

1000

246

500 60

176

100

246

500 50

75 105

50

300

200

250

300

174

125

Peak 2

85

1000 500

58

302 TIC

50

100

1000

NIST

500

79

1000

0 Time (s) TIC⫻0.4 246

150

50

250

300

350

350

104

634

632 201

636 145

638

640 104

50 100 150 200 Library Hit-similarity 769, “Folpet”

260 250

104

1000

100

350

260

130 147

50

300

NIST

76 50 63

200

300

Peak 4

20000

36 111 146 100

250

Peak true-sample “KM21 - STD 1–1.5 min splitless 50 psig: 1”, peak 107 at 636.975 s (Spec # 1983)

40000

500

50

200

147 178

201

45

150

129 350

350

125

28

63

50 100 150 200 250 300 Library Hit - similarity 878, “Thiabendazole”

300

NIST

145

80000

500 90

63

50 100 150 200 250 Library Hit-similarity 910, “Methidathion”

100000

350

201 500

Peak 3

145

500

4

60000

281

Peak true - sample “KM21 - STD 1 - 1.5 min splitless 50 psig: 1”, peak 105 at 634.975 s (Spec # 1973) 1000

3

318 210

150

2

302 350

NIST

176 140

100

1 283 318

210

50 100 150 200 250 Library Hit - similarity 801, “o,p’.DDE” 1000

85

1000

150

178 200 232 200

250

296 300

350

FIGURE 10.16 Demonstration of spectral deconvolution of four closely co-eluted compounds eluting within 8 s. Peak 1 ¼ o,p0 -DDT (m=z 246), peak 2 ¼ thiabendazole (m=z 201), peak 3 ¼ methidathion (m=z 145), and peak 4 ¼ folpet (m=z 104).

(observed commonly by scanning instruments) is encountered. This allows TOF feature-automated deconvolution of partially overlapped peaks on the basis of increasing=decreasing ion intensities in collected spectra and background subtraction followed by identiﬁcation using library search (Figure 10.16). Although the deconvolution function (employing software correction for spectral skewing) is currently available also for scanning instruments, e.g., in the AMDIS software (Automated Mass Spectral Deconvolution and Identiﬁcation System) provided by the NIST free of charge [20]), the low signal intensity during full spectra acquisition as well as relatively low acquisition rate of common scanning instruments can be a drawback that limits potential of this function when coupled to fast GC separations. Very narrow chromatographic peaks (<2 s at the baseline) generated under these conditions require a detector with high acquisition rates and, at the same time, with high sensitivity during the acquisition of full mass spectra, such as TOF, to fully utilize the advantage of spectral deconvolution [12,13]. Nowadays, two types of TOFMS instruments differing in their basic characteristics are available. One type uses high-resolution analyzers (7000 FWHM) providing only moderate acquisition speed (up to 20 spectra=s), and the second type are unit-resolution instruments that feature high acquisition speeds (up to 500 spectra=s) [14]. Figure 10.17 illustrates the main advantage of using a high-resolution TOFMS instrument. As mentioned earlier, the use of high mass resolving power signiﬁcantly reduces chemical noise that can originate from various sources (mostly from matrix co-extractives), resulting in improved limit of detection (LOD). It should be also noted that the mass measurement accuracy allows determination of the mass of molecular and fragment ions. Thus, using this approach, the identiﬁcation of the analyte can be based not only on retention time and mass spectrum, but also on elemental composition determination, which brings a new dimension to the identiﬁcation=conﬁrmation process. 10.2.5.5

Magnetic Sector

The single focusing magnetic sector mass analyzer consists of a magnetic sector, in which the magnetic ﬁeld is used for separation of the ions with different m=z ratios. The ions entering the mass

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 216 16.7.2008 10:49pm Compositor Name: MSubramanian

216

Handbook of Food Analysis Instruments

4.12

Normalized abundance

100

4.12 100

4.00 4.05 3.85 3.83 3.92 3.71

4.18

4.33

4.44 4.54

0

0

3.80

4.00

(A)

4.20

3.80

4.40

Time (min)

4.00

(B)

4.20

4.40

Time (min)

FIGURE 10.17 GC–HRTOFMS chromatograms of phosalone (tR ¼ 4.12 min) at concentration 0.01 mg=kg in baby food. Target ion m=z 182.0009 extracted using different mass windows. While by using (A) 1 Da mass window (setting corresponds to a unit mass resolution instrument) the peak-to-peak (PtP) S=N ratio was only 7:1, setting the mass window as narrow as (B) 0.02 Da led to a distinctly improved PtP S=N value of 63:1. Mass measurement accuracy allowed determination of the mass of phosalone’s quantiﬁcation ion [C8 H5 35 ClNO2 ]þ with the error as low as 0.9 mDa.

analyzer are initially accelerated to a high velocity using an electric ﬁeld. The ions then pass through a magnetic sector where the magnetic ﬁeld is applied in a direction perpendicular to the direction of ion motion. Although a magnetic sector separates the ions according to their m=z, the mass resolution is affected by the fact that ions leaving the ion source do not have exactly the same energy thus do not have exactly the same velocity. Coupling a magnetic sector with an electrostatic sector (double focussing magnetic sector mass analyzer) (Figure 10.18) allows to focus the kinetic energy of the ions. As a result, higher mass resolving power is achieved. The simplest way to operate a magnetic sector mass analyzer is to keep the accelerating potential and the electric sector at a constant potential and vary the magnetic ﬁeld. Ions with a constant kinetic energy but differing in their m=z are focused at the detector slit at different magnetic ﬁeld strengths. Usually, the electric

Lens

M

or ct

se

ag n

et

ic

ic at

st

se

ro

ct

or

t ec El

Mass filtering

Energy focussing Detector slit

Lens Detector

Ion source

FIGURE 10.18

Double-focusing magnetic sector mass analyzer.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 217 16.7.2008 10:49pm Compositor Name: MSubramanian

217

Mass Spectrometry and Hyphenated Instruments in Food Analysis

sector is held constant at a value, which passes only ions with the speciﬁc kinetic energy. The mass resolving power of a magnetic sector is inﬂuenced by the setting of slit widths. Higher mass resolving power is obtained by decreasing the slit widths, thus the number of ions that reach the detector is reduced. Double focussing magnetic sector mass analyzers provide very high reproducibility, high mass resolving power, high sensitivity, and a high dynamic range. However, their use is limited due to their size and higher cost compared to other mass analyzers, such as quadrupole and ion trap. A double focussing magnetic sector mass analyzer has been employed in a number of applications for many years and is still being used to solve speciﬁc problems in different application areas. A very high capacity to remove the signals originating from matrix interfering compounds in the determination of the analytes typically present at ultra-trace concentrations (e.g., in the analysis of polychlorinated dibenzo-p-dioxins and -furans) is achieved when employing SIM mode at a mass resolving power of 10,000 (10% valley deﬁnition). Under these conditions, the presence of abundant matrix components in the analyzed extracts does not interfere with the detection process and high level of mass accuracy can still be obtained. Another use represents the determination of complex mixtures of contaminants, such as polychlorinated terphenyls. In this particular case, not only the matrix interferences, but also problems due to interferences between fragment ions of congeners with different degree of chlorination may occur. In order to eliminate these speciﬁc interferences, a mass resolving power higher than 10,000 (10% valley deﬁnition) is required [3]. 10.2.5.6

Fourier Transform Ion Cyclotron Resonance

The FT-ICR mass analyzer performs the mass analysis in a cubic cell that consists of pair of trapping plates, excitation plates, and detector plates (Figure 10.19). After trapping the ions in the cell, they are excited to a larger cyclotron radius by an oscillating electric ﬁeld perpendicular to the ﬁxed magnetic ﬁeld. A packet of the ions is also formed during the excitation. The signal is then detected as an image current on a pair of receivers. The resulting signal consists of a superposition of sine waves, from which a regular mass spectrum is obtained by applying Fourier transformation. The most important features of this type of mass analyzer are its extremely high mass resolving power (even 107 is possible) and sensitivity.

Induced alternating current

Detector plates Excitation plates Trapping plates

Magnetic field

FIGURE 10.19

FT-ICR mass analyzer.

RF

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 218 16.7.2008 10:49pm Compositor Name: MSubramanian

218

Handbook of Food Analysis Instruments r

z

FIGURE 10.20

10.2.5.7

Orbitrap mass analyzer.

Orbitrap

The orbitrap represents a special case of an ion trap. Contrary to a conventional ion trap, which uses RF to hold the ions inside the trap, the moving ions in orbitrap are trapped in an orbit around a spindle-shaped electrode by an electrostatic ﬁeld (Figure 10.20). The electrode conﬁnes the ions, which orbit around the central electrode and oscillate back and forth along long axis of the central electrode. This oscillation can be detected as an image current on the two halves of an electrode encapsulating the orbitrap. Similarly to the FT-ICR MS, Fourier transformation is employed to obtain oscillation frequencies for ions differing in their m=z. The features of the orbitrap include high mass resolving power (up to 150,000), large space charge capacity, high mass accuracy (2–5 ppm), and high mass range (m=z 6000) [21,22]. 10.2.5.8

Hybrid Instruments

Normalized abundance

Different types of mass analyzers can be coupled together forming hybrid instruments. Triple quadrupole (QqQ) represents the most common known example. Other commercially available hybrid instruments include quadrupole=linear ion trap (Q-LIT), quadrupole=time-of-ﬂight (Q-TOF), magnetic sector=time-of-ﬂight, and time-of-ﬂight=time-of-ﬂight (TOF-TOF). The widely used triple quadrupole consists of two single quadrupoles with a collision cell in between. The ions are directed from the ion source into the ﬁrst quadrupole, where the precursor ion is selected for MS=MS reaction in the collision cell (usually a hexapole). The product ions are then separated in the second quadrupole and recorded by the detector. Figure 10.21 illustrates the

(A)

100

8.61

9.30

100

8.59

9.29

9.72

100

9.41

8.37

9.15 8.70 8.85

9.69

9.72

9.80 10.21

10.36

0

0 9.00

10.00

Time (min)

0 9.00

(B)

10.00

Time (min)

9.00 (C)

10.00

Time (min)

FIGURE 10.21 Liquid chromatography (LC)–mass spectrometry (MS[=MS]) analysis of thiodicarb (tR ¼ 9.72 min) in baby food at a concentration of 0.1 mg=kg. (A) MS1, data acquired in full scan of m=z 50–1000 (m=z 355 displayed); (B) MS1, data acquired in SIM mode (m=z 355 displayed); and (C) MS=MS, data acquired in MRM (transition m=z 355 > 87.9 displayed). The PtP S=N values 6.4:1, 9.4:1, and 125:1, respectively.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 219 16.7.2008 10:49pm Compositor Name: MSubramanian

219

Mass Spectrometry and Hyphenated Instruments in Food Analysis

TABLE 10.3 Comparison of Selected Single and Tandem Mass Spectrometry Instruments Mass Analyser Q IT (MS1) Sector HRTOF HSTOF QqQ IT (MS2) QTOF

Sensitivity in Full Spectra Acquisition

Selectivity

Mass Accuracy

Linear Dynamic Range

þ þþ þ þþþ þþþ þ þ þþ

þ þ þþþ þþ þ þþþ þþþ þþþ

þ þ þþþ þþþ þ þ þ þþþ

þþþ þþ þþþ þþ þþ þþþ þþ þþ

Note: HRTOF, high-resolution time-of-ﬂight; HSTOF, high-speed time-of-ﬂight; MS, mass spectrometry; Q, quadrupole; QqQ, triple quadrupole; Q-TOF, quadrupole=time-of-ﬂight; IT, ion trap.

improvement in selectivity, when employing tandem MS versus single MS (both in full scan and SIM) in the LC–MS (=MS) analysis. If the last quadrupole in the series is replaced by TOF mass analyzer, the quadrupole=time-ofﬂight mass spectrometer is obtained, which combines the simplicity of quadrupole with high efﬁciency of a TOF mass analyzer. As in QqQ, the sample is introduced to the ion source, followed by focusing of the ions using the hexapole ion bridge into the ﬁrst quadrupole, where the precursor ion is selected. The ions are ejected into the collision cell (a hexapole), where argon is typically used for their fragmentation. The product ions are collected into the TOF region, where separation of ions similar to that of a single TOF analyzer occurs. Due to the high mass analyzer efﬁciency, the detection of ions across the full mass range of the Q-TOF instrument is signiﬁcantly higher (approx. 10–100 times) compared to that provided by QqQ under the same conditions. In addition, the Q-TOF systems offer high-resolution capability of up to 10,000 FWHM mass resolving power, allowing high-resolution and accurate mass measurement of the product ions generated in the collision cell [23]. Table 10.3 summarizes the advantages and limitations of single and tandem MS instruments in terms of sensitivity, selectivity, mass accuracy, and linear dynamic range.

10.2.6 DETECTORS All mass analyzers separate the ions for their individual m=z values, what is followed by the recording of the number (abundance) of ions at each m=z value to give a mass spectrum. The ion detectors can be divided into three classes: (1) point ion collectors that detect the arrival of all ions sequentially at one point (e.g., electron multiplier, photon multiplier), (2) array collectors that detect the arrival of all ions simultaneously along a plane (e.g., MCP detector), and (3) detectors that consists of a pair of metal surfaces within the mass analyzer region, which the ions only pass near as they oscillate and only a weak image current is produced in a circuit between the electrodes (this principle is employed only for FT-ICR MS and orbitrap). In an electron multiplier, the ions reach the ﬁrst plate (dynode) of an electron multiplier and then the ejected electrons are accelerated through an electric potential to a second dynode. This process is typically repeated 10–12 times (according the number of used dynodes), which gives the ampliﬁcations of 106. The ﬁnal ﬂow of electrons provides an electric current that can be further increased by electronic ampliﬁcation.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 220 16.7.2008 10:49pm Compositor Name: MSubramanian

220

Handbook of Food Analysis Instruments

The photon multiplier is made up of two conversion dynodes, a phosphorescent screen, and a photomultiplier. Considering the positive mode, secondary ions are accelerated towards the dynode that holds the negative potential. Secondary electrons that are generated are accelerated towards the phosphorescent screen, where conversion into photons occurs, followed by their detection by the photomultiplier. The ampliﬁcation ranges between 104–105. Both electron multiplier and photon multiplier are detectors typically used for quadrupole, ion trap, and sector instruments. The MCP detector consists of a large number of miniature electron multiplier elements placed side by side over a plane. The surface of the MCP is metal coated, serving as electrodes. A voltage applied between the electrodes produces an electric-ﬁeld gradient. When an ion hits the front part of the inner walls of the channel at this point, multiple secondary electrons are emitted. These electrons are then accelerated by the electric-ﬁeld gradient inside the channel, and repeatedly hit the walls on the opposite side, emitting secondary electrons. At the end of this process, the electrons are captured by the anode, which produces an electrical signal. The ampliﬁcation factor of the MCP is a few thousand at maximum. To achieve an ampliﬁcation factor of 106, two MCPs are usually used (Dual MCP). The MCP is commonly used for detection of ions in TOFMS.

10.2.7 MISCELLANEOUS 10.2.7.1

High-Field Asymmetric Waveform Ion Mobility Spectrometry

High-ﬁeld asymmetric waveform ion mobility spectrometry (FAIMS) represents a new technique that separates gas-phase ions at atmospheric pressure based on the difference between ion mobility at high and low electric ﬁelds. In FAIMS, a mixture of ions is introduced between two electrodes, to which a high-voltage asymmetric waveform is applied to oscillate the ions in the alternating strong and weak electric ﬁelds. The ions drift toward an electrode depending on their differences in mobility in those electric ﬁelds. A compensation voltage is then applied to the electrodes to stop the drift of selected ions, which are transmitted to mass analyzer, while the others collide with the electrodes. FAIMS coupled to MS provides a rapid, sensitive, and selective detection of ions. When coupled with MS, the FAIMS is located between an atmospheric pressure ion source (e.g., ESI) and the inlet of the mass spectrometer. This approach allows to separate isobaric ions and isotopes, reduce the chemical background, and simplify spectra of complex mixtures [24,25]. 10.2.7.2

Supersonic Molecular Beam MS Interface

Supersonic molecular beam (SMB) interface is a unique instrumentation that is currently in the process of commercialization for the GC–EI-MS. In GC–SMB-MS, a nozzle of 0.1 mm is placed between the GC outlet and the MS. As eluted compounds from the separation column pass through the small opening, they form the SMBs. SMBs are characterized by intramolecular vibration supercooling, unidirectional molecular motion with controlled hyperthermal kinetic energy (1–20 eV), mass focusing, and capability to handle very broad range of column ﬂow rates (up to 240 mL=min is possible with a prototype instrument) compared to a ‘‘conventional’’ ﬂow of 1 mL=min typical for GC. The low thermal energy creates unique mass spectral properties that have several advantages compared to conventional GC–EI-MS: (1) increased selectivity as a result of enhanced intensity of molecular ion that occurs for most molecules at the low temperatures of SMB, (2) increased speed of analysis due to the use of very high gas ﬂow rates, (3) extended scope of analyzed compounds including the analysis of both thermally labile and low-volatility compounds, (4) versatility in selection of injection techniques and column dimensions for fast GC separation, and (5) better peak shapes since the tailing effects that typically occur in the MS ion source are suppressed [26,27]. 10.2.7.3

Imaging Mass Spectrometry

Imaging mass spectrometry (IMS) is a procedure used to form chemically selective images of objects based on the MS detection of ions desorbed from its surface. This is achieved through ionization from a clearly identiﬁed point on a ﬂat sample and performing a raster of the analyzed

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 221 16.7.2008 10:49pm Compositor Name: MSubramanian

Mass Spectrometry and Hyphenated Instruments in Food Analysis

221

sample by moving the point of ionization over the sample surface. The collected data (positional data and m=z intensities) are converted into images that map distribution of target compounds in tissues or in various other materials [28,29]. For MS imaging, either secondary ion mass spectrometry (SIMS) or MALDI (Section 10.2.4.1) can be used as desorption=ionization methods. In SIMS, the surface of the sample is bombarded by high energy ions (typically Csþ, Arþ, Oþ 2,O , þ and Ga at 1–30 keV), which leads to emission of both neutral and charged particles from the test piece. The emitted secondary ions are then extracted by an electrical potential and analyzed using a mass spectrometer. Basically, there are three different variants of this technique: (1) static SIMS used for submonolayer elemental analysis, (2) dynamic SIMS used for obtaining compositional information as a function of depth below the surface, and (3) imaging SIMS used for obtaining compositional images of the surface. While imaging SIMS provides information on the spatial distribution of the elements as well as molecular structures of low-molecular mass compounds (<1000 Da), MALDI allows to obtain spatial distribution and molecular structure information about higher molecular mass compounds (<100,000 Da) [29].

10.3 FOOD ANALYSIS APPLICATIONS A food matrix is very complex; in addition to major components such as lipids, proteins, and saccharides, a wide range of other natural minor compounds are contained (e.g., vitamins, aroma and ﬂavor compounds, pigments). Under certain conditions, contaminants and other hazardous compounds may be present in food matrices as a consequence of human activities or due processing practices. During the recent years, MS techniques have proved to be an excellent tool for qualitative characterization and quantitative determination of various food components because of their high sensitivity and speciﬁcity. In this part, we brieﬂy discuss the application of MS, mainly coupled to chromatographic techniques (GC–MS, LC–MS), in the analysis of the most important natural and contaminant substances in food. It should be noted, however, that the sample preparation practice plays a crucial role in obtaining required parameters of particular analytical method. Not only the method performance characteristics such as detection limits, accuracy and ruggedness, but also its speed, labor demands, and (consequently) the cost of an analytical procedure depend on selected extraction and cleanup strategy and their efﬁciency.

10.3.1 NATURAL SUBSTANCES

IN

FOOD

Lipids. The applicability of both LC–APCI-MS and LC–APPI-MS has been successfully evaluated in the analysis of mixtures of neutral lipids, such as triacylglycerols and sterols. Those large neutral molecules are typically troublesome in obtaining the structural characterization by using other methods, such as GC–MS and LC–ESI-MS [2,30]. For the analysis of polar lipids (e.g., phospholipids), both LC–ESI-MS and LC–APCI-MS are the methods of choice. Applications of GC–EI-MS to lipid analysis include mainly the characterization of acylglycerol and sterol fractions [2,30,31]. Peptides and proteins. The predominant analytical approaches applicable to protein food research (proteomics) involve MALDI-TOFMS and LC–ESI-MS. The MALDI-MS is useful for obtaining ﬁngerprint of the protein composition of various food samples, thus, rapid and accurate evaluation of the authenticity is possible. The LC–ESI-MS employing TOF, represents a tool for accurate mass determination of proteins. In addition, LC–ESI-MS=MS (or MALDI-TOF=TOF) allows analysis of proteins for elucidation purposes as well as the characterization of low-molecular mass peptides [32]. Saccharides. MS techniques employing different ionization modes and the hyphenated techniques play an important role in the structural elucidation of sugars. MALDI-TOFMS represents a powerful tool for the characterization of polysaccharides in various food samples. LC–MS with fast atom bombardment (FAB) interface, although not so frequently used anymore, has been proved valuable for precise isotopic measurements for saccharides and for their structural characterization. For the

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 222 16.7.2008 10:49pm Compositor Name: MSubramanian

222

1

2

3 4 56

Second-dimension retention time (s)

Handbook of Food Analysis Instruments

?

Abundance

5 ⫻ 106 4 ⫻ 106 3 ⫻ 106 2 ⫻ 106 1 ⫻ 106 640 (A)

645

650

Retention time (s)

655 (B)

8 1.04 5+6 2 3

0.84

1

9 7 4

0.64 635

645

655

665

First-dimension retention time (s)

FIGURE 10.22 Comparison of separation of selected honey volatiles in two GC systems: (A) 1D-GC– TOFMS, and (B) GCGC–TOFMS (TIC records shown). Marked compounds: (1) nonan-2-one, (2) linalool oxide, (3) dehydro-p-cymene, (4) undecane, (5) nonan-2-ol, (6) linalool, (7) terpinolen, (8) hotrienol, (9) nonanal; (?) complete co-elution of hotrienol and nonanal in 1D-GC system. Compound (7) not identiﬁed in 1DGC. Compounds (5) and (6) partially co-eluted in GCGC. (Reproduced from Cajka, T. et al., J. Sep. Sci., 30, 534, 2007. With permission.)

characterization of mono- and oligosaccharides by GC–MS, a derivatization procedure is required to remove hydrogen bonding [2]. Vitamins. Many HPLC-based methods that employ conventional detectors for detection of different groups of vitamins lack sensitivity and=or selectivity. The LC–MS technique with ESI and APCI (depending of the examined compounds) followed by their MS detection enables reliable identiﬁcation=quantiﬁcation of these nonvolatile, thermally labile biologically active compounds [2]. Aroma and ﬂavor compounds. GC–MS has a considerable potential in the separation and characterization of food aroma and ﬂavor compounds. The potential of this technique can be further increased by coupling with solid phase microextraction (SPME). This solvent-free, inexpensive sampling technique enables isolation of a wide range of analytes present in food crops and products by their extraction from its headspace and concentration in the ﬁbre coating. During the last few years, GC–MS equipped with various analyzers, such as quadrupole and ion trap, allowed obtaining information on aroma and ﬂavor compounds [33]. Recently, the application of high-speed TOF mass analyzer combined with comprehensive two-dimensional (GC GC) has been used as a tool for rapid and comprehensive analysis of these compounds (Figure 10.22) [34,35]. Typically, EI is preferred for the characterization of the volatile fraction, since it allows library searching based on the EI mass spectra. CI is used mostly to conﬁrm MW of the compounds of interest. It should be noted that under certain conditions, the use of retention indexes is important for identiﬁcation=conﬁrmation of aroma and ﬂavor compounds.

10.3.2 FOOD CONTAMINANTS Modern pesticides represent a group of compounds possessing a wide range of physico-chemical properties. With regards to this fact and considering a number of registered active ingredients (almost 1000), the analysis of multiple residues in various food matrices is obviously a difﬁcult task [36]. Nowadays, both GC–MS and LC–MS=MS are of a growing popularity in this ﬁeld. The types of MS instruments used for pesticide analysis include single mass analyzers (quadrupole, ion trap, and TOF) as well as hybrid instruments (triple quadrupole, quadrupole=time-of-ﬂight). When employing SIM and MS=MS modes, settings of time segments are typically needed, which limits

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 223 16.7.2008 10:49pm Compositor Name: MSubramanian

Mass Spectrometry and Hyphenated Instruments in Food Analysis

223

the number of targeted analytes that can be detected in a particular time. The trade-off with SIM relates to the difﬁculty of identifying analytes due to fewer ions monitored and higher chance of matrix interferences as compared to MS=MS [37]. A recent progress in instrumentation design as well as the use of fast recording electronics together with improvements of signal processing techniques has led to rediscovery of TOF mass analyzers (both types, high-resolution and highspeed) for the determination of a wide range of pesticide residues [14,18,38]. Various ionization techniques are also possible in GC–MS; nevertheless, in the case of multiresidue analysis, EI is commonly preferred. CI (both PCI and NCI) as a softer ionization technique tends to give lower LODs depending on the pesticide, but it is not as widely applicable in multiclass pesticide methods and does not provide as much structural information about the analyte as EI. Regarding the LC–MS=MS, among the various interfacing systems developed during the past years, ESI and APCI have improved the feasibility of identiﬁcation of pesticides of different chemical structures in foods at concentrations lower or comparable to those obtainable by GC–MS [39,40]. LC–MS=MS provides a complimentary tool to GC–MS for the analysis of more polar or thermolabile pesticides, which were previously analyzed by GC–MS with a various degree of difﬁculty or only after derivatization. Polychlorinated biphenyls (PCBs). GC–MS represents a reliable technique for PCBs quantiﬁcation, particularly given by the availability of 13C-labeled PCB standards. Although the EI mode is often used, the NCI allows obtaining lower LODs of these analytes. Both single quadrupole (SIM) and ion trap (MS=MS) are frequently employed in the routine analysis of PCBs [41]. The use of highresolution MS permits quantiﬁcation of lower PCBs differing by two chlorines because of the high mass resolving power that allows unbiased measurement of ions. In addition, the application of both high-resolution and high-speed TOFMS (the latter in combination with GCGC) represents a tool successfully applied in the analysis of PCBs [42,43]. Polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDDs=PCDFs). Quantitative determination of PCDDs=PCDFs occurring in biotical matrices at ultra-trace levels is typically performed using GC coupled to high-resolution MS (sector analyzer) [43]. Although PCI=NCI techniques can be used for their determination, a majority of laboratories employ EI. High-resolution systems (mass resolving power of 10,000) provide higher selectivity compared to unit mass resolution instruments especially when the levels of potentially interfering compounds are too great. However, this instrumentation is very expensive, bulky, and requires operation by a highly trained specialist. Therefore, alternate analytical instruments (less expensive) have been investigated for dioxin analysis in several laboratories. GC–MS=MS employing ion trap analyser [44,45] and GCGC– TOFMS [43,46] have been reported as a valuable technique for improved selectivity in dioxin analysis. In the case of GC–MS=MS, the high selectivity is obtained due to formation of characteristic dioxin fragment ions produced by the secondary ionization, while in GCGC– TOFMS the improvement of selectivity is achieved employing the secondary column with different polarity that can better separate the target compounds from co-eluting matrix components. Brominated ﬂame retardants (BFRs) are represented by various groups of compounds, the most known being polybrominated diphenyl ethers (PBDEs), hexabromocyclododecane (HBCD), and tetrabromobisphenol A (TBBPA). The determination of PBDEs is performed by GC–MS operated either in EI or NCI mode. The low-resolution MS is more routinely applied compared to the highresolution MS that requires more experienced users and is much more costly and labor intensive [47]. The high-resolution MS (sector) has several advantages over low-resolution MS (e.g., increased sensitivity and selectivity), but is almost exclusively operated in EI mode. For lowresolution MS, NCI, in addition of EI, can be applied to obtain an increased sensitivity for higher-brominated BDE congeners. Recently, application potential of high-resolution TOFMS under NCI conditions in the analysis of PBDEs has been demonstrated [48]. The EI is preferred in the analysis of PBDEs, whenever the identiﬁcation of mixed organohalogenated compounds

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 224 16.7.2008 10:49pm Compositor Name: MSubramanian

224

Handbook of Food Analysis Instruments

occurring in sample extracts has to be carried out. Another advantage of EI mode is the possibility to use 13C-labelled internal standards. This is not applicable in NCI, since generally only the [Br] ions (m=z 79 and 81) are monitored. The main beneﬁts of NCI include efﬁcient ionization, lower LODs, and less fragmentation compared with EI. Special attention has to be paid to analysis of BDE-209 (decabromodiphenyl ether). This high-molecular weight congener is sensitive for higher temperatures and the higher susceptibility for degradation in the GC system. For that reason, analysis of BDE-209 is typically carried out on a relatively short GC column (10–15 m) with a thin ﬁlm of stationary phase (0.1–0.2 mm), which reduces the residence time of this congener [49]. Traditionally, analysis of HBCD has been carried out using GC–MS, typically in NCI mode by monitoring the [Br] ions for a higher sensitivity. However, this technique does not allow quantiﬁcation of individual diastereomers of HBCD (a-, b-, and g-HBCD, each having two enantiomers) since they are not separated using common GC stationary phases; moreover, they undergo the interconversion above 1608C. Contrary to GC, reversed-phase LC–ESI-MS=MS or LC–APCIMS=MS employing nonpolar (C18, C30) or chiral HPLC columns for their separation represent a versatile tool for the isomer-speciﬁc determination of HBCD isomers [49]. As regards tetrabromobisphenol A (TBBPA), acidiﬁcation and derivatization are required before the GC–MS analysis, while LC–ESI-MS allows its direct determination [49]. Polycyclic aromatic hydrocarbons (PAHs). GC–EI-MS operated in SIM mode represents probably the most common GC technique for determination of PAHs in food matrices [50], although HPLC with a ﬂuorescence detector (FLD) is also often routinely used. The problem encountered in the analysis of PAHs is separation of isomers and limited EI fragmentation, which does not allow reliable conﬁrmation at ultra-trace levels. Alternatively, the LC–APCI-MS allows determination of PAHs without derivatization (post column), which is typically required in LC–ESI-MS [51]. The recently developed APPI enhances the ionization of the PAH analytes, thus, decreasing LODs [52]. Veterinary drug residues. In recent years, LC–MS-based strategies have been widely applied for determination and conﬁrmation of veterinary drug residues in food samples. LC–MS revolutionizes this ﬁeld, enabling quantitative and conﬁrmatory multiresidue, multiclass analysis of veterinary drug residues, independent of their chemical structure or biological activity, thus replacing traditional single residue=class immunochemical or microbial methods [53]. The use of GC–MS-based methods is also possible, although their main limitation is the need of extensive cleanup and time-consuming derivatization procedures prior to their GC analysis since most of these compounds are relatively polar [2]. Mycotoxins. The GC–MS determination of mycotoxins is performed either by direct analysis or (mostly) employing various chemical derivatizations. The derivates can be detected selectively to very low levels using NCI, but also EI or PCI. Although quadrupole MS (SIM) is mostly used, the high-molecular weight of derivatized analytes predetermines their molecules for MSn-measurements, especially in complex matrices like foodstuff and beverages. Recently, LC–ESI-MS=MS and LC–APCI-MS=MS methods, allowing the screening of mycotoxins of various classes within a single run, have been developed. These LC–MS-based methods represent a popular tool in mycotoxin analysis due to the improvement of sensitivity and avoiding extensive cleanup procedures and derivatization steps compared to GC–MS methods [54–56].

10.3.3 PROCESSING CONTAMINANTS Acrylamide. GC–MS methods employing either quadrupole or ion trap=triple quadrupole (tandem MS) can be used in EI or CI mode for the analysis of acrylamide in foods. Acrylamide is often derivatized (brominated) prior to the GC–MS analysis. A direct GC–MS determination without a derivatization step is also possible when a suitable sample preparation method is employed removing acrylamide precursors (asparagine and reducing sugars) from the analyzed sample, thus

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 225 16.7.2008 10:49pm Compositor Name: MSubramanian

Mass Spectrometry and Hyphenated Instruments in Food Analysis

225

preventing potential acrylamide formation from these precursors during the GC determinative step [57,58]. Alternatively, HRMS (sector or TOF) instruments are applied in the acrylamide analysis, since due to the high mass resolution power interfering compounds (chemical noise) with ions close to those of acrylamide can be signiﬁcantly reduced. In this way improvement of S=N ratio can be effectively achieved. Besides GC–MS, recent studies pay more attention to methods employing LC–ESI-MS=MS techniques with triple quadrupole for the routine analysis because this instrumental technique applied for the quantitative analysis of acrylamide has high sensitivity and avoids potential problems with acrylamide formation that may occur in the direct GC–MS analysis [59]. Chloropropanols. The analysis of 1,3-dichloropropan-2-ol (1,3-DCP), 3-monochloropropane-1,2diol (3-MCPD), and 2-monochloropropane-1,2-diol (2-MCPD) is based on GC–MS after their previous derivatization. Quadrupole mass analyser operated in SIM mode or ion trap analyser in MS=MS are typically used for their detection [60,61]. Heterocyclic amines. Determination of these mutagenic and carcinogenic substances classiﬁed as the IQ type (aminoimidazoazaarenes) and non-IQ type (pyrolytic heterocyclic amines) is based on the GC–MS methods. However, since most of these compounds are polar and nonvolatile, a derivatization step is usually performed before their analysis. This step can be effectively avoided using LC–MS with a single quadrupole or HRTOFMS, which allows selective and sensitive detection. The selectivity can be further enhanced by using LC–MS=MS with either ion trap or triple quadrupole MS instruments. Both ESI and APCI can be used for ionization [62,63].

10.3.4 MIGRANTS FROM PACKAGING Phthalate and adipate esters. GC–MS represents the key method for the determination of these packaging contaminants [64]. Epoxy compounds. The most widely used technique for the analysis of bisphenol A diglycidyl ether (BADGE), bisphenol F diglycidyl ether (BFDGE), and their hydrolyzed and chlorinated derivates is reversed-phase HPLC coupled to FLD. Since this is not a conﬁrmatory technique, GC–MS methods were developed to conﬁrm the presence of these compounds. However, this technique requires a derivation step. The use of HPLC methods with either APCI or ESI for the identiﬁcation of these substances has been recently reported [65].

10.4 CONCLUSION AND FUTURE TRENDS Over the past few years, there has been a substantial progress in technologies employing the MS technique in food analysis. In this context, a wide range of analytical methods involving both GC–MS and LC–MS(=MS) have been reported to detect, identify, quantify, and conﬁrm various naturally occurring as well as xenobiotic substances in food chain. The most important developments in MS include the implementation of new ion sources and also mass analyzers. The multimode (online) LC-sources with simultaneous ESI and APCI ion generation provide new possibilities in terms of obtaining more structured information about the analyzed samples within a single run. The multimode GC sources applicable for EI and CI (not operated simultaneously) without a need for instrument’s hardware modiﬁcation may allow conﬁrmation of analytes that can be troublesome under EI conditions. A distinctive trend in GC–MS analysis involves the rapidly growing interest in the applications that employ GC–TOFMS for both target and nontarget analysis of many classes of (semi)volatile organic compounds occurring in food. The high-speed TOFMS instruments, that are ideal for combining with fast GC, GCGC, and ultra-performance LC (UPLC) provide a powerful technique for the identiﬁcation and quantiﬁcation of a wide range of compounds present in complex food matrices. On the other hand, the high-resolution TOFMS instruments allow also accurate mass measurements, which can be appreciated for calculation of analyte elemental

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 226 16.7.2008 10:49pm Compositor Name: MSubramanian

226

Handbook of Food Analysis Instruments

composition for the identiﬁcation of ‘‘unknown’’ compounds and=or conﬁrmation target analyte identity. It can be expected that in the near future some GC–TOFMS instruments will replace mass analyzers such as sector since the former ones are easier to operate and less costly. In LC–MS, both single (high-resolution TOFMS) and hybrid instruments (QqQ and Q-TOF) are becoming increasingly popular, mainly due to their enhanced selectivity and improved quantiﬁcation and conﬁrmation capabilities as compared to single quadrupole analyzers and ion traps, respectively. Finally, the availability of sophisticated data systems and data-processing algorithms enabling automated and faster data handling represents another challenge, important for implementation of MS technique into the routine use.

ACKNOWLEDGMENTS This chapter was ﬁnancially supported by the Ministry of Education, Youth and Sports of the Czech Republic (project MSM 6046137305).

REFERENCES 1. Thomson, J.J., Further experiments on positive rays, Phil. Mag., 24, 209, 1912. 2. Careri, M., Bianchi, F., and Corradini, C., Recent advances in the application of mass spectrometry in food-related analysis, J. Chromatogr. A, 970, 3, 2002. 3. Santos, F.J. and Galceran, M.T., Modern developments in gas chromatography–mass spectrometry-based environmental analysis, J. Chromatogr. A, 1000, 125, 2003. 4. Nunez, O., Moyano, E., and Galceran, M.T., LC–MS=MS analysis of organic toxics in food, TrAC-Trend Anal. Chem., 24, 683, 2005. 5. Hajslova, J. and Zrostlikova J., Matrix effects in (ultra)trace analysis of pesticide residues in food and biotic matrices, J. Chromatogr. A, 1000, 181, 2003. 6. Todoli, J.L. and Mermet, J.M., Sample introduction systems for the analysis of liquid microsamples by ICP-AES and ICP-MS, Spectrochim. Acta B, 61, 239, 2006. 7. Karas, M. and Hillenkamp, F., Laser desorption ionization of proteins with molecular masses exceeding 10,000 daltons, Anal. Chem., 60, 2299, 1988. 8. Laiko, V.V., Baldwin, M.A., and Burlingame, A.L., Atmospheric pressure matrix-assisted laser desorption=ionization mass spectrometry, Anal. Chem., 72, 652, 2000. 9. Takats, Z. et al., Mass spectrometry sampling under ambient conditions with desorption electrospray ionisation, Science, 306, 471, 2004. 10. Cotte-Rodriguez, I. et al., Desorption electrospray ionization of explosives on surfaces: Sensitivity and selectivity enhancement by reactive desorption electrospray ionization, Anal. Chem., 77, 6755, 2005. 11. Cody, R.B., Laramee, J.A., and Durst, H.D., Versatile new ion source for the analysis of materials in open air under ambient conditions, Anal. Chem., 77, 2297, 2005. 12. Mastovska, K. and Lehotay, S.J., Practical approaches to fast gas chromatography–mass spectrometry, J. Chromatogr. A, 1000, 153, 2003. 13. Mastovska, K., Instrumental aspects and application of (ultra)fast gas chromatography–mass spectrometry, in Niessen, W. (Ed.), Encyclopedia of Mass Spectrometry, Elsevier, Oxford, 2006, p. 73. 14. Cajka, T. and Hajslova, J., Gas chromatography–time-of-ﬂight mass spectrometry in food analysis, LC GC Eur., 20, 25, 2007. 15. Murray, K.K. et al., Standard Deﬁnition of Terms relating to Mass Spectrometry (UIPAC Recommendations 2006). Available: http:==www.iupac.org=reports=provisional=abstract06=murray_prs.pdf via the Internet. Accessed on Feb 1 2007. 16. Cooks, R.G. and Busch, K.L., Counting molecules by desorption ionization and mass spectrometry=mass spectrometry, J. Chem. Edu., 59, 926, 1982. 17. Hopfgartner, G. et al., Triple quadrupole linear ion trap mass spectrometer for the analysis of small molecules and macromolecules, J. Mass Spec., 39, 845, 2004. 18. Cajka, T. and Hajslova, J., Gas chromatography–high-resolution time-of-ﬂight mass spectrometry in pesticide residue analysis: Advantages and limitations, J. Chromatogr. A, 1058, 251, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 227 16.7.2008 10:49pm Compositor Name: MSubramanian

Mass Spectrometry and Hyphenated Instruments in Food Analysis

227

19. Guilhaus, M., Selby, D., and Mlynski, V., Orthogonal acceleration time-of-ﬂight mass spectrometry, Mass Spec. Rev., 19, 65, 2000. 20. AMDIS Download Page. Available: http:==chemdata.nist.gov=mass-spc=amdis= via the Internet. Accessed on Feb 1 2007. 21. Scigelova, M. and Makarov, A., Orbitrap mass analyzer—overview and applications in proteomics, Proteomics, 6, 16, 2006. 22. Hu, Q. et al., The Orbitrap: A new mass spectrometer, J. Mass Spec., 40, 430, 2005. 23. Ferrer, I. and Thurman, E.M., Liquid chromatography=time-of-ﬂight=mass spectrometry (LC=TOF=MS) for the analysis of emerging contaminants, TrAC-Trend Anal. Chem., 22, 750, 2003. 24. Guevremont, R., High-ﬁeld asymmetric waveform ion mobility spectrometry: A new tool for mass spectrometry, J. Chromatogr. A, 1058, 3, 2004. 25. Guevremont, R. et al., Ion trapping at atmospheric pressure (760 Torr) and room temperature with a highﬁeld asymmetric waveform ion mobility spectrometer, Int. J. Mass Spectrom., 193, 45, 1999. 26. Amirav, A., Gordin, A., and Tzanani, N., Supersonic gas chromatography=mass spectrometry, Rapid Comm. Mass Spec., 15, 811, 2001. 27. Fialkov, A.B. et al., Sensitivity and noise in GC–MS: Achieving low limits of detection for difﬁcult analytes, Int. J. Mass Spec., 260, 31, 2007. 28. Stoeckli, M. et al., Imaging mass spectrometry: A new technology for the analysis of protein expression in mammalian tissues, Nat. Med., 7, 493, 2001. 29. Todd, P.J. et al., Organic ion imaging of biological tissue with secondary ion mass spectrometry and matrix-assisted laser desorption=ionisation, J. Mass. Spec., 36, 355, 2001. 30. Cai, S. and Syage, J.A., Atmospheric pressure photoionization mass spectrometry for analysis of fatty acid and acylglycerol lipids, J. Chromatogr. A, 1110, 15, 2006. 31. Canabate-Diaz, B. et al., Separation and determination of sterols in olive oil by HPLC-MS, Food Chem., 102, 593, 2007. 32. Careri, M. and Mangia, A., Analysis of food proteins and peptides by chromatography and mass spectrometry, J. Chromatogr. A, 1000, 609, 2003. 33. Kataoka, H., Lord, H.L., and Pawliszyn, J., Applications of solid-phase microextraction in food analysis, J. Chromatogr. A, 880, 35, 2000. 34. Adahchour, M. et al., Comprehensive two-dimensional gas chromatography with time-of-ﬂight mass spectrometric detection for the trace analysis of ﬂavour compounds in food, J. Chromatogr. A, 1019, 157, 2003. 35. Cajka, T. et al., Solid phase microextraction–comprehensive two-dimensional gas chromatography–timeof-ﬂight mass spectrometry for the analysis of honey volatiles, J. Sep. Sci., 30, 534, 2007. 36. Mastovska, K., Food & Nutritional Analysis: (q) Pesticide residues, in Worsfold, P., Townshend, A., and Poole, C., (Eds.), Encyclopedia of Analytical Science, Elsevier, Oxford, 2005, 251. 37. Lehotay, S.J. and Mastovska, K., Determination of pesticide residues, in Otles, S. (Ed.), Methods of Analysis of Food Components and Additives, Taylor & Francis, Boca Raton, 2005, p. 329. 38. Zrostlikova, J., Hajslova, J., and Cajka, T., Evaluation of two-dimensional gas chromatography–time-ofﬂight mass spectrometry for the determination of multiple pesticide residues in fruit, J. Chromatogr. A, 1019, 173, 2003. 39. Pico, Y. et al., Control of pesticide residues by liquid chromatography-mass spectrometry to ensure food safety, Mass Spec. Rev., 25, 917, 2006. 40. Alder, L. et al., Residue analysis of 500 high priority pesticides: Better by GC–MS or LC–MS=MS? Mass Spec. Rev., 25, 838, 2006. 41. Ahmed, F.E., Analysis of polychlorinated biphenyls in food products, TrAC-Trend Anal. Chem., 22, 170, 2003. 42. Dallüge, J., Roose, P., and Brinkman, U.A.Th., Evaluation of a high-resolution time-of-ﬂight mass spectrometer for the gas chromatographic determination of selected environmental contaminants, J. Chromatogr. A, 970, 213, 2002. 43. Focant, J. et al., Comprehensive two-dimensional gas chromatography with isotope dilution time-of-ﬂight mass spectrometry for the measurement of dioxins and polychlorinated biphenyls in foodstuffs: Comparison with other methods, J. Chromatogr. A, 1086, 45, 2005. 44. Eppe, G. et al., PTV-LV-GC=MS=MS as screening and complementary method to HRMS for the monitoring of dioxin levels in food and feed, Talanta, 63, 1135, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C010 Final Proof page 228 16.7.2008 10:49pm Compositor Name: MSubramanian

228

Handbook of Food Analysis Instruments

45. Loran, S. et al., Evaluation of GC-ion trap-MS=MS methodology for monitoring PCDD=Fs in infant formulas, Chemosphere, 67, 513, 2007. 46. Hoh, E., Mastovska, K., and Lehotay, S.J., Optimization of separation and detection conditions for comprehensive two-dimensional gas chromatography–time-of-ﬂight mass spectrometry analysis of polychlorinated dibenzo-p-dioxins and dibenzofurans, J. Chromatogr. A, 1145, 210, 2007. 47. Covaci, A., Voorspoels, S., and de Boer, J., Determination of brominated ﬂame retardants, with emphasis on polybrominated diphenyl ethers (PBDEs) in environmental and human samples—a review, Environ. Int., 29, 735, 2003. 48. Cajka, T. et al., Challenges of gas chromatography–high-resolution time-of-ﬂight mass spectrometry for simultaneous analysis of polybrominated diphenyl ethers and other halogenated persistent organic pollutants in environmental samples, J. Sep. Sci., 28, 601, 2005. 49. Covaci, A. et al., Recent developments in the analysis of brominated ﬂame retardants and brominated natural compounds, J. Chromatogr. A, 1153, 145, 2007. 50. Stolyhwo, A. and Sikorski, Z.E., Polycyclic aromatic hydrocarbons in smoked ﬁsh—a critical review, Food Chem., 91, 303, 2005. 51. Takino, M. et al., Determination of polycyclic aromatic hydrocarbons by liquid chromatography– electrospray ionization mass spectrometry using silver nitrate as a post-column reagent, J. Chromatogr. A, 928, 53, 2001. 52. Itoh, N., Aoyagi, Y., and Yarita, T., Optimization of the dopant for the trace determination of polycyclic aromatic hydrocarbons by liquid chromatography=dopant-assisted atmospheric-pressure photoionization=mass spectrometry, J. Chromatogr. A, 1131, 285, 2006. 53. Gentili, A., Perret, D., and Marchese, S., Liquid chromatography-tandem mass spectrometry for performing conﬁrmatory analysis of veterinary drugs in animal-food products, TrAC-Trend Anal. Chem., 24, 705, 2005. 54. Nielsen, K.F. and Thrane, U., Fast methods for screening of trichothecenes in fungal cultures using gas chromatography–tandem mass spectrometry, J. Chromatogr. A, 929, 75, 2001. 55. Sforza, S., Dall’Asta, C., and Marchelli, R., Recent advances in mycotoxin determination in food and feed by hyphenated chromatographic techniques=mass spectrometry, Mass Spec. Rev., 25, 54, 2006. 56. Zöllner, P. and Mayer-Helm, B., Trace mycotoxin analysis in complex biological and food matrices by liquid chromatography–atmospheric pressure ionisation mass spectrometry, J. Chromatogr. A, 1136, 123, 2006. 57. Mastovska, K. and Lehotay, S.J., Rapid sample preparation method for liquid and=or gas chromatographicmass spectrometric analysis of acrylamide in various food matrices, J. Agric. Food Chem., 54, 7001, 2006. 58. Dunovska, L. et al., Direct determination of acrylamide in food by gas chromatography–high-resolution time-of-ﬂight mass spectrometry, Anal. Chim. Acta, 578, 234, 2006. 59. Zhang, Y., Zhang, G., and Zhang, Y., Occurrence and analytical methods of acrylamide in heat-treated foods: Review and recent developments, J. Chromatogr. A, 1075, 1, 2005. 60. Chung, W., Hui, K., and Cheng, S., Sensitive method for the determination of 1,3-dichloropropan-2-ol and 3-chloropropane-1,2-diol in soy sauce by capillary gas chromatography with mass spectrometric detection, J. Chromatogr. A, 952 185, 2002. 61. Hamlet, C.G. and Sutton, P.G., Determination of the chloropropanols, 3-chloro-1,2-propandiol and 2-chloro-1,3-propandiol, in hydrolysed vegetable proteins and seasonings by gas chromatography=ion trap tandem mass spectrometry, Rapid Com. Mass Spec., 11, 1417, 1997. 62. Barcelo-Barrachina, E. et al., Evaluation of reversed-phase columns for the analysis of heterocyclic aromatic amines by liquid chromatography—electrospray mass spectrometry, J. Chromatogr. B, 802, 45, 2004. 63. Toribio, F. et al., Ion-trap tandem mass spectrometry for the determination of heterocyclic amines in food, J. Chromatogr. A, 948, 267, 2002. 64. Balafas, D., Shaw, K.J., and Whitﬁeld, F.B., Phthalate and adipate esters in Australian packaging materials, Food Chem., 65, 279, 1999. 65. Pardo, O. et al., Determination of bisphenol diglycidyl ether residues in canned foods by pressurized liquid extraction and liquid chromatography–tandem mass spectrometry, J. Chromatogr. A, 1107, 70, 2006.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 229

22.7.2008 6:30am Compositor Name: VBalamugundan

to Analyze 11 Instruments Food Colors Carmen Socaciu and Horst A. Diehl CONTENTS 11.1 11.2

Introduction ........................................................................................................................ 229 Light and Color ................................................................................................................. 230 11.2.1 Physics of Light and Color .................................................................................. 230 11.2.2 Chemistry of Color .............................................................................................. 232 11.2.3 Biology of Color .................................................................................................. 233 11.3 Color Measurement ........................................................................................................... 233 11.3.1 Principles of Color Measurement ........................................................................ 233 11.3.2 Optical Spectrometry ........................................................................................... 235 11.3.2.1 Absorption=Transmission Spectrometry ............................................. 236 11.3.2.2 Reﬂection Spectrometry ...................................................................... 236 11.3.3 Colorimetry .......................................................................................................... 236 11.3.4 Color Imaging Analysis ....................................................................................... 238 11.3.5 Reference Methods .............................................................................................. 238 11.4 Color Measurements as Food Quality Indicators .............................................................. 239 11.4.1 Complexity of Food Color Analysis ................................................................... 239 11.4.2 Applications of Food Color Measurements ......................................................... 240 11.5 Conclusion ......................................................................................................................... 242 References ..................................................................................................................................... 243

11.1 INTRODUCTION Color is an intrinsic property of food, characterizing its identity and quality. Consumers are attracted or disgusted by the color of a food product, drawing a conclusion to the food freshness and quality [1–7]. Considering that the evaluation of food color is an essential topic in food technology and food quality management, this chapter focuses on fundamental and practical aspects of color perception and instrumental measurement. Food producers are interested in keeping the food color under strict control by optical inspection as a quality indicator of the raw material, as a control for a correct food processing, and as a marker of appropriate addition of colorant additives [8,9]. Section 11.2 deals with basic scientiﬁc issues of light and color. Color is a perceptive sensation by humans, although rather different between various individuals. To analyze food colors, ﬁrst of all, is a physical issue. The physical deﬁnition of color provides the theoretical base for standardized measurements, and physical rules govern the technology of the measuring instruments. To assess food quality by the appearance of food colors and its components, as well changes, is a chemical issue. Food decay processes are disintegration processes on a molecular level. Any instrumental food color measurement is to be interpreted in terms of quality standards. 229

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 230

230

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

To mediate attraction to food is a biological issue. Eye receptors linked to the brain transmit the color perception. Men appreciate food quality by food color and thus give reasons to food producers to add colors to food, either natural or artiﬁcial ones. Many animals ﬁnd their food resources or sexual partners by color identiﬁcation. Section 11.3 deals with instrumental details of color measurement devices. The instruments to analyze food colors are engineered on physical principles and their ability to simulate human color perception is limited. This limitation is based mainly on two reasons: (1) process of human color perception is so complicated that it can only partially be reﬂected by a physical setup and (2) individual differences in color perception abilities are considerably high, and only on a statistical base a standard human color perceptor is deﬁned. Two basic types of color-measuring instruments are generally used: spectrophotometers and colorimeters. Both are based on two modes of operation: reﬂectance and absorption=transmittance. The measurement principles and applications will be presented. Section 11.4 includes data and interpretations of objective physical=instrumental food color measurements and gives examples for raw, processed, and stored food. Rational automation of quality control is in progress and examples of food quality sorting according to color are given as case studies.

11.2 LIGHT AND COLOR The term color is used to describe at least three aspects of reality: an object attribute (blue sky), a characteristic of a light (red light), or a sensation given by an eye perception [9–11]. Color results as a consequence of physical interaction of light with matter. Before Newton’s discovery of the light spectrum no distinctions were made between color and light, the color being considered as visible light but after this era, the science of colors began. Spectral colors were represented by seven perceived hues in a color wheel (violet-indigo-blue-green-yellow-orange-red) having white in the center of the circle. Color is a consequence of eye perception followed by a brain interpretation, while light is electromagnetic energy (being considered as an electromagnetic wave or as a bunch of photons) characterized by frequency (n) and wavelength (l). Optical instruments (spectral light meters) identify colored matter depending on the physical structure of the matter either by light absorption or transmittance or reﬂection or scattering. Food materials usually are characterized by a combination of them and the spectral intensity distribution will differ between them [12,13].

11.2.1 PHYSICS

OF

LIGHT

AND

COLOR

The dual nature of light, to be a wave and to be a stream of particles (photons), has to be considered when interactions with different matters are investigated. Light does transmit energy to food. This can cause a warming up effect (wave nature of light) or it can destruct punctually molecular complexes or even molecules themselves (quantum nature of light). When light interacts with matter, . . . .

It can be reﬂected at the surface. This may be a specular reﬂection or a diffuse reﬂection. It can be absorbed or transmitted by the matter. It can be scattered from the bulk matter (not only from the surface) if the matter is inhomogeneous. It may excite ﬂuorescence which can be detected at longer wavelengths (Stokes shift). Natural food rarely contains ﬂuorescent matter, but food additives may contain it.

Detecting and analyzing methods work along these principles [14,15]. If the surface of the matter is smooth, a mirror-like (specular) reﬂection occurs, while a rough matter (e.g., a colloid solution) causes diffusion (scattering) of the light at the surface as well as in the bulk matter. This may result

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 231

22.7.2008 6:30am Compositor Name: VBalamugundan

231

Instruments to Analyze Food Colors

in different perceptions of color in hues and values. When the matter surface is extremely rough (e.g., powdered matter as such or suspended in liquids), light becomes completely scattered and no color is seen. To give some examples, a clear apple juice has a light and bright yellow-brownish color, while a turbid raw juice appears glossy white-brown and with less brightness because of scattering of light on colloid particles. The same phenomenon we observe on the beer foam which is perceived as white because of multiple scattering effects of light on the air bubbles. It is generally assumed that the energy of the incident light (I) is conserved and converted to reﬂected (specular or diffuse), scattered, absorbed, or transmitted light. For clear solutions, absorption or transmission is dominant while for highly colloidal systems, scattering becomes dominant. To speak about color of an object is to speak simultaneously about the illuminating light source, the light transmitting medium, object properties, eye sensitivity, and conventions about color scales [12,13,16,17]. Matter can be isotropic (gases, clear liquids, crystals) or anisotropic (heterogenic macrocrystals, polymers). The appearance of food and its color depends in detail on its transparency and homogeneity. Food is a high anisotropic system which interacts with light and its color is determined by reﬂected or scattered, absorbed or transmitted, remanent light. Optical phenomena modulate the color impression of food and can be measured either by objective optical evaluation (qualitative and quantitative spectrometry) or by eye perception. The sensorial perception of color by individuals is not easy to describe in physical terms because the light receptors differ considerably between species and individuals and cannot be standardized. Consequently, an automation of sensorial visual evaluation is limited to cases of exactly reproducible frame conditions. Objects which are identical substantially, but show a varying surface ripple structure or contain a varying degree of moisture, will yield different results. Light shining on the object may be diffused or more focussed, or the angle between the illuminating light beam to the object and the light beam from the object to the viewer’s position may change; all this will impact error on measurements. Colorimetry is the method to mimic and to measure the perception of color, considering the factors mentioned [9,10,18]. Principles of these two ways of color measurement are shown in Figure 11.1. Natural food shows no or weak ﬂuorescence. But food may contain added ﬂuorescent pigments which impart bright colors to the matter when irradiated by blue or ultraviolet (UV) light, where usually ﬂuorophores are excited to emit light in the visible region. To compare food colors under daylight and under UV light helps to identify artiﬁcial color additions. Here we do not consider ﬂuorescence as a natural coloring property.

Absorption Specular reflection Regular transmission Scattered reflection

Scattered transmission

FIGURE 11.1 Principles of light interaction with matter. Left: At the surface of the matter, part of the impinging light is reﬂected, either specularly or by scattering. Center: Part of the incident light is absorbed. Right: Transmitted light remains in the geometrical line with the incident beam (regular transmission) or is scattered. Scattering can happen at the matter surface as well as from the bulk matter.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 232

232

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

11.2.2 CHEMISTRY

OF

COLOR

Food can be considered a supramolecular matrix, mainly formed of organic molecules, water, or their complexes with inorganic groups (metal cations, anions). While the colors of transition metal complexes are explained by the ligand ﬁeld theory (donor–acceptor electronic transitions), the colored organic molecules (e.g., pigments and dyes) contain conjugated p–p or n–p electronic bonds related to various chromophores (polyene chains –C¼C–(C¼C)n–C¼C–, or azo –N¼N–, thio –C¼S, nitroso –N¼O) or color enhancers (auxochromes) which consist of electron-donor groups (–NH2, –NHR) coupled with electron acceptors –NO2, –Br, –OH. Each of these conjugated systems, depending on the extension of electronic delocalization in resonance, absorbs at speciﬁc wavelengths (labs), which undergoes a hypsochromic shift toward longer wavelengths when the number of conjugated double bonds increases [5]. The intensity of absorption is determined by auxochromes and the number of double bonds. The color is determined by the electronic transitions n–p* and p–p*. For example, increasing the number of double bonds from 6 or 8 to 10 or 12 up to 16 shifts the absorption gradually from the UV to the Vis to the red spectral region. Retinol (vitamin A) containing ﬁve double bonds absorbs in the UV region (325 nm), while the pigment lutein containing 10 double bonds absorbs at 445 nm, b-carotene and zeaxanthin (11 double bonds) at 450 nm, and lycopene (12 double bonds) at 470 nm. As these pigments absorb in the blue spectral region, their color appears yellow-orange-red (Y-O-R) to the observer (Figure 11.2). Other polyenes, involving cyclic, non-benzoic conjugated systems include porphyrins like chlorophyll and heme. They both contain nine double bonds, but different metal ions complexed with the porphirinic cycle, Mg2þ and Fe2þ, respectively. Chlorophylls absorb in the violet (420 nm) and red (650 nm) regions, transmitting the green light, while heme absorbs in the green region, transmitting the red light. In algae, the Cu-phthalocyanine complexes (26 double bonds) absorbing in the red region (>550 nm) transmit the blue-green light. Many synthetic colorants (malachite green, crystal violet, phenolphthalein) containing triphenylmethane conjugated chromophores linked to auxochromes (donor–acceptor groups) have different colors dependent of different factors (pH and ionic environment) and therefore they can be used as analytical indicators.

Chlorophyll a Chlorophyll b Carotenoids

400 Violet

500 Indigo

Blue

600 Green

Yellow Orange

700 Red

A: Absorbed colors

Yellow Orange Red Violet Indigo Blue Green T: Colors complementory to a, seen after transmission or reflection of light

FIGURE 11.2 (See color insert following page 240.) Chemical identiﬁcations by complementary display of color absorption and transmission. The absorption spectra of chlorophyll a, chlorophyll b, and carotenoids are plotted over the wavelength scale. Below the wavelength scale: A—the absorbed colors (not perceivable by the eye) are named and T—the complemetary colors are named which are transmitted or reﬂected and are perceived by the eye.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 233

Instruments to Analyze Food Colors

22.7.2008 6:30am Compositor Name: VBalamugundan

233

In general, the Y-O-R colorants (b-carotene, lycopene) have single absorption bands in the B-V-G region, the V-B colorants (indigo) have single absorption bands in the R-O-Y region, the purple colorants absorb in the central B-G-Y region. Yet, the green colorants have two absorption bands, one in the blue region and one in the red region. Therefore, it is difﬁcult to prepare homogeneous natural green colorants. They are obtained from mixtures of yellow and blue colors but with different fading susceptibilities. Many natural pigments found in food are chromophores, e.g., carotenoids, chlorophylls, ﬂavonoids (quercetin—yellow), anthocyanins (cyanidin—red, purple, or blue in acidic, neutral, or alkali environments, respectively), and betalains. Flavonoids and especially anthocyanins show different colors dependent on pH, which inﬂuence their stability. Most biological colors of raw food products (fruits, vegetables, green leaves, ﬂowers) as well some animal food (meat, ﬁsh, seafood) or ﬁnal products or added food dyes can fade, can be degraded, or can be converted to less colored derivatives depending on environmental and technological factors (temperature, pH, light, humidity, chemicals, irradiation, etc.). To measure the intact color and such modiﬁcations, chemists use spectrometry as the main method of choice, being quantitative, accurate, and objective. Various separation methods are coupled with spectrometry to achieve an accurate evaluation of colored molecules. These methods are used as reference methods for colorimetric measurements. Colorimetry mimics the eye perception of color, using primary colors or mixed colors, but is not an accurate, quantitative method. A set of deﬁnitions and statistical evaluations have been developed to standardize human perceptions, as will be discussed in detail later.

11.2.3 BIOLOGY

OF

COLOR

The human eye cannot sharply differentiate the following three characteristics of color: its (1) hue— red, yellow, green, blue, etc., (2) chroma or saturation, and (3) lightness=darkness. On average, the eye perceives hue differences ﬁrst, chroma or saturation differences second, and lightness=darkness last. Hue is the quality that normally identiﬁes a color such as red, green, and blue. Saturation is the clarity of a color, expressed as the intensity of hue compared with its own brightness. A saturated color looks clear and bright, but an unsaturated color appears pale, muddy, or dull. The light value is represented by its lightness or brightness. Lightness considers color as a source of reﬂected light by an illuminated surface and is related to the connotations light and dark. Brightness represents the total light from an illuminated object or reﬂected from a surface. Colorimetry seeks to simulate the human color perception. The evaluation of colorimetric measurements has to consider the physiological color perception mechanisms of the human eye. Rod-shaped receptors in the eye are responsible for night vision. Cone-shaped receptors are responsible for daylight and color vision. There are three types of cone-shaped receptors which are sensitive to red, green, and blue colors. The so-called tristimulus color values are coordinated to these spectral eye sensitivities. Neurologically, the red, green, and blue cone responses are mixed into opponent coders as they move up the optic nerve to the brain. The so-called opponent colors theory is based on that objection.

11.3 COLOR MEASUREMENT 11.3.1 PRINCIPLES

OF

COLOR MEASUREMENT

Appearance is the manifestation of the nature of objects and materials through visual attributes such as size, shape, chroma, color, texture, glossiness, haze, transparency, opacity, hue, luster, orange peel, translucency, etc. The detecting instruments generally fall into one of the four categories: colorimeters, densitometers, spectral cameras, and spectrophotometers [9,11,12,17,18]. Colorimeters measure color using three or four ﬁlters that match human color receptors. Colorimeters can show L, a, b

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 234

234

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

or L*, a*, b* numbers, but can only measure in one light source. Densitometers measure the density of layers using one or more ﬁlters. Densitometers do not give color information, but are useful for the control of juice concentrations. Spectral cameras provide measurements with full spectral and spatial information. Spectrophotometers operate on the principle of absorption=transmittance or reﬂected light. They give reproducible complete spectral and intensity informations under standardized conditions. Human perceptions are not considered. Food color is analyzed either by direct inspection (sensorial analyses) or by instrumental methods. The direct inspection determines the sensorial attribute of the color, usually combined with food odor=smell, taste=ﬂavors. Visual color assessment is subjective and may be done by experts if reliable visual evaluations with multiple variables can be controlled. The instrumental color measurements eliminate interindividual errors and are better reproducible. Two basic types of color-measuring instruments (spectrophotometers and colorimeters) are used based either on absorption=transmittance or on reﬂectance [16,19,20]. Absorption spectrophotometers measure the speciﬁc spectral absorption or transmittance of the sample from the entire spectrum of light (UV, visible, and infrared), the residual light being quantiﬁed. Rather transparent solutions are best measured by spectrophotometers in the absorption mode at a spectral bandwidth of a few nanometers. Often the full visible wavelength range, approximately 300 nm, is examined. Absorption intensity versus wavelength is plotted (Figure 11.2). The reﬂectance principles for spectrometric and colorimetric measurement of color are presented in Figure 11.3a and b. The principles of reﬂection spectrometry are presented in Figure 11.3a. Reﬂection measurements are more complex than transmission measurements. Materials possessing a high degree of scatter (turbid material) show light scattering as well as absorption. It determines the collection geometry for detecting the reﬂected light. Normally, integrating spheres are used where the specimen is part of the reﬂecting sphere wall. Mostly the sample is positioned in a way that the ﬁrst reﬂection of the incident light from the sample is excluded, or other geometrical arrangements are used, e.g., 458=08 means the light shines to the specimen surface under an angle of 458 to the normal and is observed under 08 or vice versa (08=458) [12]. These instruments work well when the surface of the sample being tested has uniform texture and gloss, e.g., paints and some types of foods. If signiﬁcant texture or granulation is present on the surface of the sample, however, some reﬂected light may be scattered at different angles and escape detection.

Diode array detector Specimen

Data display

Specimen I

Light source

Data processor

Light source

Light dispersion element Wavelength Data display

(a) Reflection spectrometry

Spectral filters

X Y Z

X Y Z

Photodetectors (b) Tristimulus colorimetry

FIGURE 11.3 Color measurement by optical reﬂection techniques. (a) Reﬂection spectrometry—the light reﬂected from the specimen is dispersed into its spectral composition by a dispersing element (grating or prism), spectrally registered by a diode array (or any other light sensor), processed and displayed. (b) Tristimulus colorimetry—the light reﬂected from the specimen is directed to the three ﬁlters (red, green, and blue), and as such detected and processed.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 235

Instruments to Analyze Food Colors

22.7.2008 6:30am Compositor Name: VBalamugundan

235

Also colorimetry is based on reﬂection measurements (Figure 11.3b). The monochromatic colorimeter measures the amount of light (in arbitrary units) which is reﬂected in a narrow spectral area of the visible light. It is basically color-blind and sees only one color, e.g., red, yellow, or green. The tristimulus colorimeter measures the true colors and correlates them to what the eye sees, using specialized glass color ﬁlters and light detectors (up to 10 million different shades of color can be quantiﬁed) [21,22]. A 16- or 31-data point (wavelengths) spectrophotometer or a tristimulus colorimeter is an appropriate tool for accurately measuring the color of foods. A tristimulus colorimeter concentrates on three major colors: red, green, and blue. When connected to a personal computer equipped with the appropriate software, the instrument can provide three-dimensional color plots and other information that is more relevant to the way humans perceive the reﬂected color of a food product.

11.3.2 OPTICAL SPECTROMETRY Optical spectrometry is based on the irradiation of the probe with UV, Vis, or infrared light, which is absorbed and induces molecular interactions (electronic excitations, molecular vibrations, and rotations). The absorption informs about the identity (absorption wavelengths) and the number of such molecules (absorption intensity). The complementary relation between the absorbed light and the transmitted color is presented in Section 11.2.3. (Figure 11.2) With respect to food colorants, the Vis region is the most relevant one, useful to identify the colored molecules absorbing in the 360–800 nm region, to quantify their concentration and to qualify the food. For practical purposes one has to discriminate between measurements in the absorbance mode and those in the reﬂection mode. To characterize pure or extracted food color, the absorption mode is usually used [13,23]. The most important are spectroscopic methods in the visible spectral region, where food colorants are perceived by men. The ideal light source is a lamp emitting white light, with uniformly distributed intensity over the whole visible spectral region. Such light sources are not available, but incandescent tungsten lamps approach this request partially, less well in the blue and UV regions, where high pressure mercury lamps are recommended. High pressure xenon arc lamps emit well over the whole spectral region, simulating, better than any other source, the sunlight spectral emission. Special light sources, applied particularly in colorimetry, use additional ﬁlter systems. To identify a certain colored molecule, monochromatic light is required. For this either monochromatic light emitting lasers are needed or a dispersion of light into monochromatic regions (red, green, and blue) has to be achieved. To obtain monochromatic light, one of the three different monochromating devices is used: 1. Filters of colored glass or special interference ﬁlters that allow very narrow spectral cuts. 2. Glass or quartz prisms (which are as well transmissible by UV light) that disperse the spectrum colors and separate the color wanted. 3. Scattering diffraction gratings. By means of optical lenses or mirrors, the scattered light is brought to interference that determines the spectral decomposition into monochromatic light. To detect the light transmitted after the absorption or reﬂection by the sample, different optoelectric devices (detectors) that transform light signals into electric signals are used: photocells and photomultipliers, based on the photo-electric effect (1), mainly used in laboratory spectrophotometers, or small-sized solid state photo-detectors (semiconductors) working at low voltages, suited for portable instruments (2). Details about their construction and use are presented elsewhere [14,16,23].

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 236

236

11.3.2.1

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

Absorption=Transmission Spectrometry

The most suitable spectroscopic method for transparent colored food is UV-Vis absorption= transmission spectrometry. To quantify the concentration of a food colorant, the light transmission (T ) is expressed as T ¼ I=I0, where I0 is the incident light intensity and I is the transmitted light intensity after absorption. No linear correlation between T and the concentration c of the colorant is observed, but the logarithm of the reciprocal transmission (log1=T ), called extinction or absorption (A), is directly correlated to the colorant concentration (c), the optical distance of the light crossing the sample (d) (cm), and the coefﬁcient of molecular absorption («), expressed as mol1 L cm1: log 1=T ¼ log I0 =I ¼ A ¼ « c d Details about the use of spectrometry to measure color are presented extensively for carotenoidcontaining plants and food [24,25]. In relation to the UV-Vis spectrometry, the ﬂuorescence spectrometry is useful to identify ﬂuorescent (synthetic) food colorants that have been incorporated into food. These colorants can be traced by its spectral pattern (ﬂuorescence excitation and emission spectra), which in special cases may be modiﬁed in the food matrix according to the molecular interactions with the environment. Molecular ﬂuorescence is observed in molecules that contain aromatic, heterocyclic, and condensed ring systems. For example, betaxanthins have been identiﬁed as pigments responsible for visible ﬂuorescence in ﬂowers [26]. 11.3.2.2

Reﬂection Spectrometry

Three visible=near-infrared reﬂection (VNIR) spectrophotometers (same manufacturer and model) were used to measure L*a*b* color values for 50 food and agricultural products including spices and seasonings (e.g., turmeric, mustard, chili powder, salt), dessert powders, seeds (e.g., canola, poppy, lentil), cornmeal, wheat ﬂour, powdered cheese, drink crystals, soup mix, and ground breakfast cereal. The L*a*b* values were calculated from the visible spectral data as deﬁned by the Commission Internationale de I’Eclairage (CIE). The color values from the VNIR instruments were in very good agreement with reference values collected from three Minolta instruments designed speciﬁcally to measure color [27]. For practical measurements various reﬂection spectrometers are used like ColorFlex, Miniscan, and especially LabScan XE, a xenon-ﬂash portable color spectrometer which delivers advanced features and optimum accuracy, available in d=88 or 458=08 optical geometry, with 25 to 7.6 mm measurement areas, one-handed operation with customizable LCD (liquid crystal display), and storing up to 999 measurements. It contains interfaces for quality control and color formulation software.

11.3.3 COLORIMETRY Color is basically speciﬁed by the geometry and spectral distributions of three elements: the light source, the reﬂectivity of the sample, and the visual sensitivity of observer. Each of these was deﬁned since 1931 by the CIE and improved in time. The deﬁnition was aimed at simulating the human color perception based on a 28 ﬁeld of view, a set of primaries (red, green, and blue), and color-matching functions [22]. Under the guidance of the CIE, several color description systems and notions have been developed which determine the presently used procedures: 1. CIE system (1931) introduced the standard observer looking to the object under the angle of 28 and later under the angle of 108. The later CIE systems proceed from there. The 108

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 237

Instruments to Analyze Food Colors

22.7.2008 6:30am Compositor Name: VBalamugundan

237

standard observer is recommended for its better correlation with average visual assessments, made with large ﬁelds of view that are typical for most commercial applications. Any color can be described by a combination of X, Y, and Z primaries. By deﬁnition, X, Y, and Z speciﬁcations of objects are always made relative to the luminosity of a perfect white object (reﬂectance ¼ 100 at each wavelength). The Y for the perfect white is always 100. The magnitudes of X and Z for the perfect white change with the color of the illuminant being used [17]. 2. The Hunter Lab system (1958) was the ﬁrst system to use the opponent-color theory which states that the red, green, and blue cone responses are remixed into opponent coders as they move up the optic nerve to the brain [17]. On the basis of the opponent-color theory, the Hunter L,a,b color space is three-dimensional rectangular, where L (lightness) varies from 0 (black) to100 (white), a (red–green axis) with positive (red) and negative (green) values, and b (yellow–blue axis) with positive (yellow) and negative (blue) values. L*a*b* values used in the system are calculated from the tristimulus values (X, Y, and Z), which are the backbone of all mathematical color models. The location of a color in the CIE color space is deﬁned by a three-dimensional Cartesian (rectangular) coordinate system. The lightness value (L*) indicates how light or dark the color is. The a* value indicates the position on the red–green axis, and b* is the position on the yellow–blue axis. Once the L*a*b* position of a standard color is determined, a rectangular tolerance box can be drawn around the standard [18]. 3. The CIELAB system (1976) standardizes strictly the light source and the observer, and it recommends three standard sources and observers with normal, standard color vision. L*C*H* color-difference calculations are derived from the L*a*b* values. L*a*b* values are converted from the rectangular coordinate system to a cylindrical coordinate system. The L* value is the same as the L* value in the L*a*b* color space and represents the lightness plane on which the color resides. The C* value is the calculated vector distance from the center of the color space to the measured color. Larger C* values indicate higher chroma. H* measures hue. Using the L*C*H* polar coordinate system allows instrumental readings to match more closely the color perception of human observers. Further system improvements have been made with regard to color difference formula [21] and with regard to chromatic adaptation formula to describe color appearance under different viewing conditions [28,29]. It is based on the consideration that color is changing according to direction and is adopted as British Standard BS6923: 1988. The scale system currently in vogue is the CMC (Colour Measurement Committee) colortolerancing scale, developed by the Colour Measurement Committee of the Society of Dyes & Colourists in Great Britain. CMC is not a new color space, but a modiﬁcation of CIE which provides better agreement between visual assessment and instrumentally measured color differences. The CMC mathematical calculation deﬁnes an ellipsoid around the standard color with the same semi-axis corresponding to hue, chroma, and lightness. The ellipsoid represents the volume of acceptance and automatically varies in size depending on the position of the color in the color space. CMC is considered the ‘‘best system to evaluate human color perception’’ (according to Teunis, X-Rite Inc., Grandville, Michigan), because human color perception is more tolerant in the green region, but less tolerant in the dark blue region. Food shapes and consistency vary, the heterogeneity of composition and color distribution cause difﬁculties for colorimetric evaluations. To meet these problems, appropriate colorimeters were produced using spherical geometry that applies diffuse sample illuminations, eliminating the directionality of the light. In these instruments, detectors mounted on top of a sphere at 88 from center are used. One application for spherical geometry-based instruments is measuring the color of granulated powders. Colorimeters that use spherical geometries tend to be more expensive. Deciding which type of instrument to use requires experience.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 238

238

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

The principles of these systems have been put into mathematical formulae that can be transferred to each other, but the different measuring conditions have to be taken into account. A system of providing uniformity of color measurement in accordance with the state veriﬁcation scheme (GOST 8.205-90) was provided for studying the metrological characteristics of a contemporary class of spectrocolorimeters from the ﬁrms GretagMacbeth Co. LLC, Windham, NH acquired recently by X-Rite Grand Rapids, M1 (USA) [20]. The British Standards Institute and the International Standards Organization have edited general guidances and test methods for the assessment of the color of foods.

11.3.4 COLOR IMAGING ANALYSIS The conventional optical spectrophotometry is going to be replaced by new optical analytical methods, mainly digital color analysis (DCA), or computer vision systems (CVS) using digital and video cameras. The digital color analyzer is a hand-held size instrument for measuring colors, it transforms the color information into analytical information (numerical values, color library data), further than spectrophotometry can do. By handling colors as digital information, the semiquantitative analysis made by colorimetry can be upgraded and serve as an accurate quantitative method [30,31]. Such analyzers (e.g., Colortron, Light Source Computer Images, Inc., Larkspur, California, 1994) do determine colors, calculate the tristimulus values (X, Y, and Z) with the obtained spectrum data, and the XYZ values can be converted into a variety of numerical color data and color library data. The color digital cameras are mostly based on light-sensitive single array elements placed on a CCD (Charge-Coupled Device) chip, with a ﬁlter for red (R), green (G), and blue (B), and the relative intensities are adjustable by a white balance. Thus, a digital color image is represented in RGB form with three components per pixel and conventionally stored using eight bits per color. These three intensity images are electronically combined to produce a digital color picture [32]. These systems do not only offer a methodology for color measurement, but they can also be applied to measure other attributes of food appearance [2,33]. CVS was recently implemented to quantify standard colors of fruit and vegetables in RGB, HSV (Hue-Saturation-Value), and L*a*b* color spaces, and image capture conditions affecting the results were evaluated [33]. The three color spaces were compared in terms of their suitability for color quantiﬁcation in curved surfaces. The CVS showed to be robust to changes in sample orientation, resolution, and zoom. However, the measured average color was shown to be signiﬁcantly affected by the background, by the surface curvature, and by gloss. The L*a*b* system is suggested as the best color space for quantiﬁcation in foods with curved surfaces. The new CVS systems are used more and more to investigate the food surface color modiﬁcations [32,34–37].

11.3.5 REFERENCE METHODS To evaluate accurately the color of food or the colorants added to food, a detailed physical–chemical analysis is done in specialized laboratories. Such detailed analysis includes at least four successive steps: sample preparation by extraction and puriﬁcation, followed by separation of target molecules and their identiﬁcation or quantitative evaluation. The sample preparation for color analysis is a critical step of analysis since it is needed to minimize the color modiﬁcations by heat, light, enzymatic, or oxidative degradations. The actual protocols for food color analysis avoid organic solvents, using better physical methods like ultrasound, microwaves, supercritical ﬂuids (CO2), high pressure liquid extraction, and solid-phase extractions [5]. The identiﬁcation of the target colorant is done either directly by Vis spectrometry [38] or after a previous separation from a mixture, using techniques like planar chromatography (paper chromatography and thin layer chromatography) [39], counter-current chromatography [40], highperformance liquid chromatography (HPLC) [41–44], ion pair chromatography [45], and capillary

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 239

22.7.2008 6:30am Compositor Name: VBalamugundan

Instruments to Analyze Food Colors

239

electrophoresis techniques like capillary zone electrophoresis (CZE) or micellar electrokinetic chromatography [46,47]. Nondestructive spectroscopic methods, like FT-IR (Fourier transform-infrared) and NIR (nearinfrared) spectrometries, and Raman spectroscopy are yet widely applied to identify directly in food matrix the speciﬁc color ﬁngerprint, these techniques being increasingly applied in food analysis, as recently reviewed [48–51]. Mass spectrometry is used as well, either coupled with HPLC for the detection of separated molecules, or for the identiﬁcation of a ﬁngerprint based on fragmentation patterns [52]. Details about the general use of these techniques in food colorant analysis are given elsewhere [5]. All techniques mentioned aim for the absolute evaluation of color and are actually used as reference methods to compare the performance and accuracy of color measurement methods (based on reﬂection spectrometry or colorimetry). Most of the published data about food color evaluation compare the reference values obtained by reference methods with the performance given by colorimetry or spectrometry.

11.4 COLOR MEASUREMENTS AS FOOD QUALITY INDICATORS In food industries and food engineering, research color is considered a fundamental physical property and a quality component of raw or ﬁnal food products, since it correlates with other physical, chemical, and sensorial indicators of food quality [4,7,30,35]. Color measurement techniques evaluate quality factors of raw food, such as degree of ripeness and spoilage during shipping, storage, shelf-life, up to the consumer. Color measurements serve as quality index of raw and processed food for use in quality control, determination of conformity of food quality into a claimed speciﬁcation, and analysis of quality change as a result of food processing and storage. Quality control involving color evaluation is carried out in numerous ways. Consumers rely primarily upon their vision to evaluate product color. Because color perception differs from person to person, and depends upon lighting and numerous other factors, many industries rely on human vision coupled with an instrumental system of color measurement.

11.4.1 COMPLEXITY

OF

FOOD COLOR ANALYSIS

Generally, spectrophotometers and colorimeters work better for paints and dyes, for homogeneous beverages or food extracts than they do for food products. The key problem that prevents accurate, reproducible color measurements is that most foods have nonuniform surfaces and heterogenic composition, with profound effects on light and color reﬂection and perception. For example, the color measurement of a fruit depends on where on the curvature of its surface the measurement is made. If the angle of measurement is different from reading to reading, the quantitative color reading will be different. Measuring the color of porous material, like a cake, is challenging because any sample compression during the color-measurement process turns it darker. Meat products that color fade during storage are also difﬁcult to investigate accurately. Both developing industry and potential buyers need analytical methods that are simple, quick, and readily available, enabling a rapid evaluation of product quality. The most commonly used technique for evaluating the color of food is colorimetry by the measurement of CIE tristimulus XYZ values, which are transformed mathematically to a coordinate system that describes color, such as Hunter Lab and CIE L*a*b* (described earlier in Section 11.3.3). For both of these coordinate systems, lightness is measured by the ﬁrst value (L, L*), redness is measured using the value (a, a*), and yellowness is measured using the value (b, b*). Lightness and redness are the more important values for color fading evaluation. However, instrumental metamerism is a common and serious defect of colorimeters. In addition, as mentioned before, CIE L*a*b* measurements are inﬂuenced by geometrical conditions of

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 240

240

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

illumination and observation. In production environments, it can be difﬁcult to ensure that the geometrical conditions are accurately reproduced, so an alternative approach, e.g., the spectral reﬂectance from the surface of food was elaborated. Optical ﬁber based spectrometer methods, previously been used for the prediction of texture, can be used for color determination using the reﬂection in the visible and NIR region (400–2200 nm) [53,54]. These instruments attempt to simulate the manner in which the average human eye sees the color of an object, under speciﬁed illumination conditions, and provide a quantitative measurement. The reﬂected spectral data are transformed or ﬁltered to provide reproducible color values in accordance with standards developed by the CIE. Frequently, color and ﬂavor are directly related. However, food processors are often limited in their ability to adjust color in the ﬁnal product. Therefore, they pay strict attention to the color of ingredients and to color changes that appear during each step of production.

11.4.2 APPLICATIONS

OF

FOOD COLOR MEASUREMENTS

Color measurement systems are used to evaluate a broad range of food products. These include fresh and processed fruits and vegetables, formulated or compounded foods, dairy products, meats and meat products (including ﬁsh and poultry), spices and ﬂavors, cereals and grains, oils, syrups, sugar, and beverages. In the agricultural and food industries, the most popular numerical color-space system is the L*a*b*, which is also referred to as the CIELAB system [21,22], as described earlier. Such colormeasurement instruments are widely used throughout the food and agricultural industries to monitor color of products such as meat [55], ﬁsh [56], tea [57], honey [58], tomato [59], vegetables [60], beverages [61], and ﬂour [62]. Spectrocolorimetry in the CIE L*a*b* color space was a useful tool to monitor the ripening process and the quality of French red-smear soft cheeses [63]. The estimation of wheat color by parallel measurements was recently reported [64] using an automatic reﬂection CIE (color-space L*, a*, b*) system, absorption measurement of a water-saturated butanol extract, and HPLC determinations. Carotenoids are responsible for the color and were best determined by HPLC, whereas the colorimetric reﬂection method is fast and safe but provides only relative values. Measuring the color of processed foods can be a critical quality control tool. Also, color measurements can be a useful indicator of the quality of incoming raw materials and ingredients used to prepare processed foods [65]. Fruit and vegetable processors often base their decision to harvest or not, on color measurements. Furthermore, measuring the color of various parts of the plant may provide information that allows farmers to optimize applications of fertilizers and herbicides. Meat processors also use color measurements to assess the quality of their products. Measuring the extent of red and brown is a good indicator of meat freshness, as well as consumer appeal and acceptance. Changes in color are usually analyzed by means of the three numerical data corresponding to the chromatic color indices of CIELAB color measurement system [21]. But these three values taken independently do not provide information about tone and intensity of the measured color. Consequently, some new color indices have been developed. Most important methods are those of the American Spice Trade Association, of the ASTA-20.1 [66] for extractable color, and of the chromatic attributes proposed by CIE [22]. Such methods support color measurements directly or after a previous chromatographic separation. To give some examples, a rapid spectrophotometric method for the determination of total carotenoid content of paprika oleoresins has been developed to estimate the concentration of the red and yellow isochromic fractions and their ratios and the total carotenoid content as a quality index [67,68]. Another example, in the olive fruit the ratio of total chlorophylls to total carotenoid content (Chl=Car) is used to characterize cultivars and to establish the ripening stage of the fruit. Normal values for Chl=Car in olive fruits are 3–3.5, while in the virgin olive oil extracted from these fruits the ratio takes a value of 1 because of the preferential extraction of carotenoids versus chlorophylls

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 241

Instruments to Analyze Food Colors

22.7.2008 6:30am Compositor Name: VBalamugundan

241

[65]. The relative concentrations of some carotenoids such as violaxanthin, lutein, and b-carotene have also been used to characterize and authenticate virgin olive oils [69]. The carotenoid HPLC ﬁngerprint coupled with their spectrometric proﬁles have been applied for the authentication and determination of geographical origin of ‘‘Valencia’’ orange juice and for the adulteration of orange juice with paprika extract and tangerine juice [70]. Other color indices are used to measure the maturity stage of fruits. Apart from the pigment ratios (such as R=Y), the esteriﬁcation of xanthophylls has been measured during the ripening of several fruits (orange, red pepper, apple, peach, etc.). Commonly, a gradual accumulation of total xanthophylls is accompanied by a concomitant increase in xanthophyll esters, in some cases up to 60%–75% of the total xanthophylls. Producers aim for a uniform fruit maturity stage, to ensure good product quality, e.g., crucial when red pepper fruits are used for paprika production [71,72]. The use of such an index would ensure uniformity at the ripening stage of fruit and take maximum advantage of their carotenogenic capacity, with a direct impact on the quality of the processed products, paprika and oleoresins. The anthocyanin color proﬁle (red to violet and blue) usually changes during fruit ripening and maturation. It has been extensively studied in pomegranate fruit, where the diglucoside derivatives (cyanidin 3,5-diglucoside) are the predominant pigments during the early ripening stages, while the monoglucoside derivatives (cyanidin 3-glucoside) prevail in the later stages. The red color of pomegranate juice intensiﬁes throughout storage at 28C–58C [73]. The use of the a=b Hunter ratio was highly functional in measuring red to yellow changes, during degreening of lemons, and during postharvest storage of brocolli [74] or ripening of tomato [75] including the saturation index chroma (the distance from the coordinate’s origin to the determined color point). The hue angle index representing the tone of color that commonly decreases during degreening (908 represents a yellow color, higher values indicate green and lower orange) was also determined in green vegetables [76]. The citrus color index, as a speciﬁc color index for citrus fruit, shows a high correlation to visual appreciation of the ﬂavedo color [77]. The total increment of color or total color difference, formerly proposed, is still used to determine the color changes in vegetables [76]. These indices have been applied to represent changes in color throughout the continuous degreening system of lemons, grapefruit, and oranges with 5 ppm of ethylene, to simulate a maximum marketing period, while other quality attributes such as titratable acidity, pH, total soluble solids content, and maturity index did not show signiﬁcant changes. A combination of digital camera, computer and graphics software provided a versatile and inexpensive technique for measuring nonhomogeneous food surface, such as pizza color [78]. Near-infrared reﬂectance spectrophotometers (NIRS) are also widely used throughout the food and agricultural industries to measure chemical constituents such as protein, oil, starch, ﬁber, and moisture, as well as color of meat [79], tea [51], cheese [63,80], and tomatoes [81]. Approximately 15 years ago, NIR instruments became available with an extended spectral range which includes the visible region (VNIR instruments). The visible region allowed for the measurement of pigments such as chlorophyll in immature canola seed [82] and carotenoids in wheat [83] by measuring absorbances or reﬂectances at speciﬁc wavelengths speciﬁc to these pigments. VNIR instruments have been used to estimate colorspace values of cotton [84]. However, these color measurements are based on standard NIR predictive equations of L*, a*, and b*, rather than spectral transformations in accordance with the CIE deﬁnitions. The only use of a VNIR instrument to directly measure color of an agricultural product, in accordance with the principles of the CIE, was recently reported, using a ﬁber-optic equipped VNIR instrument to monitor in-line color and chemical composition of an extruded corn-feed mix [85]. However, no attempt was made to determine the accuracy of the VNIR-determined color values through comparison with instruments designed speciﬁcally to measure color. Recent researches studied the feasibility of using a VNIR instrument to measure color of food and agricultural products, in accordance with the principles of the CIE and to compare the color values with a colorimeter speciﬁcally designed to provide colorspace values. The color measurements were done using three different VNIR instruments (NIRSystem 6500, Foss North America, Eden Prairie, MN) and three different colorimeters=spectrophotometers designed to measure color (Minolta CR310 models equipped with granular material attachment) [27].

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 242

242

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

Tristimulus colorimetry is used often for juice color evaluation [86–88]. A recently developed food, the ultrafrozen orange juice (UFOJ), has been characterized in terms of carotenoid pigments, ascorbic acid, and color [89,90], using reﬂection measurements with a CAS 140B spectroradiometer, using the illuminant D65 and a 108 observer and in parallel a reference method (HPLC). Sliced ham products undergo discoloration from their original pink color to a pale grey color when exposed to a combination of oxygen and light, unappealing to consumers who expect a pink color for sliced ham. To monitor the initial color and fading status of cured ham before packaging, an optical ﬁber-based sensor was applied. Two methods to evaluate the fading were used by CIE L*a*b* measurements and analysis of the spectral reﬂectance of the ham color [91]. The reproducibility of the CIE L*a*b* values proved to be quite difﬁcult and signiﬁcant overlapping of the L* (lightness) and a* (redness) values measured for pink- and grey-colored hams was observed. The variations of these values can be attributed to intensity differences of the reﬂected light from different products. L*a*b* measurements are sensitive to light intensity and pigment concentration. Analysis of the spectral reﬂectance readings did not encounter these problems as the spectral response was normalized (to reduce intensity errors) before data analysis was carried out on the spectral shape or pattern using principal component analysis (PCA) and artiﬁcial neural networks (ANN). A classiﬁer based on PCA and ANN was successfully implemented that can discriminate various fading stages of the ham slices. This case study showed that the sensor system can better discriminate between a light initial color and a dark faded color than the CIE L*a*b* color measurement system. A potentially useful application of color manipulation of taste was initiated for elderly population [6] whose smell and taste tend to deteriorate. Food processors’ interest in this area of color application is likely to escalate since the average age of consumers is increasing.

11.5 CONCLUSION In spite of the signiﬁcant changes in the conventional laboratory spectrometry and colorimetry instrumentation, correlated with high-performance reference methods applied in high-standard laboratories, for measuring color in food-related areas of application, new developments have been implemented in recent years using handheld instruments and especially those based on CVS and DCA. To summarize the most important points of development in this area, we point out the following: .

.

.

.

Creation of new, user-friendly, simpliﬁed software or upgrades by color instrument manufacturers. The recorded color-measurement data can easily be exported to quality control programs that include other types of quality control data such as pH, fat, moisture, etc. Portability of powerful handheld colorimeters, convenient for color measurements online (orchard or processing lines). The collected data can be extensively analyzed in the lab with a computer. New-generation instruments are improved with regard to user-friendliness and accuracy, providing easily understandable outputs (pass=fail, good=bad, overripe=underripe), avoiding number values. Measurements can be made at various angles simultaneously and give averaging results. For example, ColorTec handheld colorimeters use an unusual light source, a series of 12 light emitting, long-life diodes (light emitting diodes, LEDs) that cover the entire visible wavelength range. Increasing demand by food companies in measuring color because of accuracy improvements, and availability for user-friendly handheld instruments at decreasing prices. More and more, food companies realize how useful and practical color-measuring instruments can be as a reliable food quality control tool.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 243

Instruments to Analyze Food Colors

22.7.2008 6:30am Compositor Name: VBalamugundan

243

REFERENCES 1. Francis, F.J. and Clydesdale, F.M. (Eds.), Food Colorimetry: Theory and Applications, AVI Publishing Co., Westport, CT, p. 477, 1975. 2. Hutchings, J.B. (Ed.), Food Color and Appearance, Aspen Publishers, Inc., Gaithersburg, MD, p. 593, 1999. 3. Hutchings, J.B., Luo, R., and Ji, W., Calibrated colour imaging analysis of food, In: MacDougall, D. (Ed.), Colour in Food, Woodhead Publishing, and CRC Press LLC, Boca Raton, Boston, Chapter 14, pp. 352–366, 2002. 4. McDougall, D.B. (Ed.), Color in Food, Improving Quality, Woodhead Publishing Ltd. and CRC Press LLC, Boca Raton, FL and Boston, MA, 2002. 5. Socaciu, C. (Ed.), Food Colorants: Chemical and Functional Properties, CRC Press-Taylor & Francis group, Boca Raton, FL, 2008. 6. Clydesdale, F.M., Color as a factor in food choice, Crit. Rev. Food Sci. Nutr., 33, 83, 1993. 7. Mackinney, G. and Little, A.C., Color of Foods, The Avi Publishing Company, Inc., Westport, CT, 1962. 8. Kress-Rogers, E. and Brimelow, C.J.B. (Eds.), Instrumentation and Sensors for the Food Industry, 2nd ed., Woodhead Publishing Ltd., 2001. 9. Hunt, R.W.G. (Ed.), Measuring of Colour, Ellis Horwood Ltd., New York, 1991. 10. Billmeyer, F.W. and Saltzman, M., Principles of Color Technology, John Wiley & Sons, New York, 1981. 11. Nasau, K., The Physics and Chemistry of Color, the Fifteen Causes of Color, 2nd ed., JohnWiley & Sons, NewYork, 2001. 12. Grum, F. and Bartleson, C.J., Optical radiation measurements, Vol. 2, Color Measurement, Academic Press, NewYork, 1980. 13. Hammes, G.C., Spectroscopy for the Biological Sciences, 1st ed., Wiley-Vch, Weinheim, Germany, 2005. 14. Diehl, H.A., Physics of color, in Socaciu, C. (Ed.), Food Colorants: Chemical and Functional Properties, CRC Press-Taylor & Francis group, Boca Raton, FL, 2008. 15. Mirabella, F.M., Modern Techniques in Applied Molecular Spectroscopy, John Wiley & Sons, New York, 1998. 16. Banwell, C.N. and Mccash, E.M., Molekülspektroskopie, R. Oldenbourg Verlag, München, Germany, 1999. 17. Hunter, R.S. and Harold, R.W. (Eds.), The Measurement of Appearance, John Wiley & Sons, New York, p. 411, 1987. 18. HunterLab, Insight on Colour, Vol. 12, Technical Services Department, Number 6, 2000. 19. Gremlich, H.-U. and Yan, B., Infrared and Raman Spectroscopy of Biological Materials, Marcel Dekker, Inc., Basel, Switzerland, 2000. 20. Gorshkova, T.B., Provision of uniformity for measurement of color characteristics in the paint, food, textile, and other branches of industry, Measurement Techniques, 48, 1096, 2005. 21. CIE, Colorimetry, Ofﬁcial Reccomandations of the Commission Internationale De L’éclairage, Central Bureau Vienna, 2nd ed., 27 A-1030 Publication 15.2, 1986. 22. CIE Publication No. E308–99, Standard practice for computing the colors of objects by using the CIE system, in ASTM Standards on Color and Appearance Measurement, 6th ed., CIE Central Bureau Kegelgasse 27 A-1030, Wien, Austria, 2000. 23. Schmidt, W., Optical Spectroscopy in Chemistry and Life Sciences, 1st ed., Wiley-Vch, Weinheim, Germany, p. 370, 2005. 24. Delgado-Vargas, F. and Paredes-Lopez, O., Natural Colorants for Food and Nutraceutical Uses, CRC Press LLC, Boca Raton, FL, Chapter 2, 2003. 25. Britton, G., Liaaen-Jensen, S., and Pfander, H. (Eds.), Carotenoids Handbook, Birkhäuser Verlag, Basel= Boston=Berlin, 2004. 26. Gandia-Herrero, F., Escribano, E., and Garcia-Carmona, F., Betaxanthins as pigments responsible for visibile ﬂuorescence in ﬂowers, Planta, 222, 586, 2005. 27. McCaig, T.N., Extending the use of visible=near-infrared reﬂectance spectrophotometers to measure colour of food and agricultural products, Food Res. Int., 35, 731, 2002. 28. Li, C., Luo, M.R., Rigg, B., and Hunt, R.W.G., CMC 2000 chromatic adaptation transform: CMCCAT 2000, Color Res. Appl., 27, 49, 2002. 29. Luo, M.R., Cui, G., and Rigg, B., The development of the CIE 2000 colour-difference formulae: CIEDE 2000, Color Res. Appl., 26, 340, 1986.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 244

244

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

30. Abdullah, M.Z., Guan, L.C., Lim, K.C., and Karim, A.A., The applications of computer vision system and tomographic radar imaging for assessing physical properties of food, J. Food Eng., 61, 125, 2001. 31. Hirayama, E., Sugiyama, T., Hisamoto, H., and Suzuki, K., Visual and colorimetric lithium ion sensing based on digital color analysis, Anal. Chem., 72, 465–474, 2000. 32. Russ, J.C. (Ed.), Image Analysis of Food Microstructure, CRC Press LLC, New York, p. 369, 2005. 33. Mendoza, F., Dejmek, P., and Aguilera, J.M., Calibrated color measurements of agricultural foods using image analysis, Postharvest Biol. Technol., 41, 285, 2006. 34. Du, C.-J. and Sun, D.-W., Recent developments in the applications of image processing techniques for food quality evaluation, Trends Food Sci. Technol., 15, 230, 2004. 35. Segnini, S., Dejmek, P., and Öste, R., A low cost video technique for color measurement of potato chips, Lebensmittel-Wiss. Technol., 32, 216, 1999. 36. Yam, K.L. and Papadakis, S.E., A simple digital imaging method for measuring and analyzing color of food surfaces, J. Food Eng., 61, 137, 2004. 37. O’Sullivan, M.G., Byrne, D.V., Martens, H., Gidskehaug, L.H., Andersen, H.J., and Martens, M., Evaluation of pork color: Prediction of visual sensory quality of meat from instrumental and computer vision methods of color analysis, Meat Sci., 65, 909, 2003. 38. Wilson, R. (Ed.), Spectroscopic Methods for Food Analysis, VCH Publishers, New York, 1994. 39. Sherma, J., Planar chromatography, Anal. Chem., 78, 3841, 2006. 40. Degenhardt, A., Knapp, H., and Winterhalter, P., Separation of natural food colorants by high-speed countercurrent chromatography, in Ames, J.M. and Hofmann, T. (Eds.), Chemistry and Physiology of Selected Food Colorants, ACS Symposium Series 775, American Chemical Society, Washington DC, p. 22, 2001. 41. Nollet, L.M.L., Food Analysis by HPLC, 2nd ed., Marcel Dekker Inc., New York, 2000. 42. Gennaro, M.C., Abrigo, C., and Cipolla, G., HPLC analysis of food colors and its relevance in forensic chemistry, J. Chromatogr. A, 674, 281, 1994. 43. Gratzfeld-Huesgen, A. and Schuster, R., HPLC for Food Analysis: A Primer, Hewlett-Packard Company, 1996. 44. Oliver, J. and Palou, A., Chromatographic determination of carotenoids in foods, J. Chromatogr. A, 881, 543, 2000. 45. Henshall, A., Use of ion chromatography in food and beverage analysis, Cereal Foods World, 42, 414, 1997. 46. Sadecka, J. and Polonsky, J., Electrophoretic methods in the analysis of beverages, J. Chromatogr. A, 880, 243, 2000. 47. Watanabe, T. and Terabe, S., Analysis of natural food pigments by capillary electrophoresis, J. Chromatogr. A, 880, 311, 2000. 48. Wetzel, D.L.B., Analytical near infrared spectroscopy, in Wetzel, D.L.B. and Charalambous, G. (Eds.), Instrumental Methods in Food and Beverage Analysis, Elsevier, Amsterdam, the Netherlands, 1998. 49. Osborne, B.G., Near-infrared spectroscopy in food analysis, in Meyers, R.A. (Ed.), Encyclopedia of Analytical Chemistry, John Wiley & Sons Ltd., Chichester, United Kingdom, 2000. 50. Schulz, H., Baranska, M., and Baranski, R., Potential of NIR-FT-Raman spectroscopy in natural carotenoid analysis, Biopolymers, 77, 212, 2005. 51. Schulz, H., Engelhardt, U.H., Wegent, A., Drews, H.-H., and Lapczynski, S., Application of near-infrared reﬂectance spectroscopy to the simultaneous prediction of alkaloids and phenolic substances in green tea leaves, J. Agric. Food Chem., 47, 5064, 1999. 52. Smedsgaard, J. and Frisvad, J.C., Using direct electrospray mass spectrometry proﬁling of crude extracts, J. Microbiol. Methods, 25, 5, 1996. 53. Garcia-Rey, R.M., García-Olmo, J., De Pedro, E., Quiles-Zafra, R., and Luque de Castro, M.D., Prediction of texture and colour of dry-cured ham by visible and near infrared spectroscopy using a ﬁber optic probe, Meat Sci., 70, 357, 2005. 54. Ortiz, M.C., Sanabia, L., Garcia-Ray, R., and Luque De Castro, M.D., Sensitivity and speciﬁcity of PLSclass modelling for ﬁve sensory characteristics of dry-cured ham using visible and near infrared spectroscopy, Anal. Chim. Acta, 558, 125, 2006. 55. Wulf, D.M. and Wise, J.W., Measuring muscle color on beef carcasses using the L*a*b* color space, J. Anim. Sci., 77, 2418, 1999. 56. Skrede, G. and Storebakken, T., Characteristics of color in raw, baked and smoked wild and pen-reared Atlantic salmon, J. Food Sci., 51, 3, 804, 1986. 57. Joubert, E., Tristimulus colour measurement of rooibos tea extracts as an objective quality parameter, Int. J. Food Sci. Technol., 30, 783, 1995.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 245

Instruments to Analyze Food Colors

22.7.2008 6:30am Compositor Name: VBalamugundan

245

58. Terrab, A., Díez, M.J., and Heredia, F.J., Chromatic characterisation of Moroccan honeys by diffuse reﬂectance and tristimulus colorimetry—Non-uniform and uniform colour spaces, Food Sci. Technol. Int., 8, 189, 2003. 59. Arias, R., Lee, T.C., Logendra, L., and Janes, H., Correlation of lycopene measured by HPLC with the L*, a*, b* color readings of a hydroponic tomato and the relationship of maturity with color and lycopene content, J. Agric. Food Chem., 48, 1697–1702, 2000. 60. Ihl, M., Shene, C., Scheuermann, E., and Bifani, V., Correlation for pigment content through colour determination using tristimulus values in a green leafy vegetable, swiss chard, J. Sci. Food Agric., 66, 527, 1994. 61. Eagerman, B.A., Clydesdale, F.M., and Francis, F.J., Comparison of color scales for dark colored beverages, J. Food Sci., 38, 1051, 1973. 62. Oliver, J.R., Blakeney, A.B., and Allen, H.M., Measurement of ﬂour color in color space parameters, Cereal Chem., 69, 546, 1992. 63. Dufossé, L., Galaup, P., Carlet, E., Flamin, C., and Valla, A., Spectrocolorimetry in the CIE L*A*B* color space as useful tool for monitoring the ripening process and the quality of PDO red-smear soft cheeses, Food Res. Int., 38, 919, 2005. 64. Fratianni, A., Irano, M., Panﬁli, G., and Acquistucci, R., Estimation of color of durum wheat. Comparison of WSB, HPLC, and reﬂectance colorimeter measurements, J. Agric. Food Chem., 53, 2373, 2005. 65. Roca, L.M. and Mínguez-Mosquera, M.I., Change in the natural ratio between chlorophylls and carotenoids in olive fruit during processing for virgin olive oil, J. Am. Oil Chem. Soc., 78, 133, 2001. 66. ASTA, Ofﬁcial Analytical Methods of the American Spice Trade Association, 2nd ed., Englewood Chiffs, New York, 1986. 67. Hornero-Méndez, D. and Míguez-Mosquera, M.I., Rapid spectrophotometric determination of red and yellow isochromic carotenoid fractions in paprika and red pepper oleoresins, J. Agric. Food Chem., 49, 3584–3588, 2001. 68. Míguez-Mosquera, M.I. and Pérez-Gálvez, A., Color quality in paprika oleoresins, J. Agric. Food Chem., 46, 5124, 1998. 69. Gandul-Rojas, B., López-Cepero, M.G.R., and Mínguez-Mosquera, M.I., Use of chlorophyll and carotenoids pigment composition to determine authenticity of virgin olive oil, J. Am. Oil Chem. Soc., 77, 853–858. 70. Mouly, P.P., Gaydou, E.M., Lapierre, L., and Corsetti, J., Differentiation of several geographical origins in single-strength Valencia orange juices using quantitative comparison of carotenoid proﬁles, J. Agric. Food Chem., 47, 4038, 1999. 71. Mínguez-Mosquera, M.I., Gandul-Rojas, B., and Gallardo-Guerrero, L., Rapid method of quantiﬁcation of chlorophylls and carotenoids in virgin olive oil by HPLC, J. Agric. Food Chem., 40, 60, 1992. 72. Mínguez-Mosquera, M.I., Pérez-Gálvez, A., and Garrido-Fernández, J., Carotenoid content of the varieties jaranda and jariza (Capsicum annuum L.) and response during the industrial slow drying and grinding steps in paprika processing, J. Agric. Food Chem., 48, 2972, 2000. 73. Gil, M.I., García-Viguera, C., Soler, C., Arté, S.F., and Tomá, S.-B.F.A., Evolution of pomegranate juice anthocyanins during fruit ripening and postharvest storage, Bull. Liaison Groupe Polyphenols (Inra) Colloq., 69, 213, 1995. 74. Shewfelt, R.L., Heaton, E.K., and Batal, K.M., Nondestructive color measurement of fresh broccoli, J. Food Sci., 49, 1612, 1987. 75. Shewfelt, R.L., Thai, C.N.L., and Davis, J.W., Prediction of changes in colour of tomatoes during ripening at different constant temperatures, J. Food Sci., 53, 1433, 1988. 76. Gnanasekharan, V., Shewfelt, R.L., and Chinnan, M.S., Detection of color changes in green vegetables, J. Food Sci., 57, 149, 1992. 77. Jiménez, M., Cuquerella, J., and Martínez, J.M., Determination of a color index for citrus fruit degreening, Proc. Int. Soc. Citriculture, 2, 750, 1981. 78. Papadakis, S., Abdul-Malek, S., Kamdem, R.E., and Jam, K.L., A versatile and inexpensive technique for measuring color of foods, Food Technol., 54, 48, 2000. 79. Rodbotten, R., Nilsen, B.N., and Hildrum, K.I., Prediction of beef quality attributes from early post mortem near infrared reﬂectance spectra, Food Chem., 69, 427, 2000. 80. Sorensen, L.K. and Jepsen, R., Assessment of sensory properties of cheese by near-infrared spectroscopy, Int. Dairy J., 8, 863, 1998.

Otles/Handbook of Food Analysis Instruments 45660_C011 Final Proof page 246

246

22.7.2008 6:30am Compositor Name: VBalamugundan

Handbook of Food Analysis Instruments

81. Peiris, K.H.S., Dull, G.G., Lefﬂer, R.G., and Kays, S.J., Near-infrared (NIR) spectrometric technique for nondestructive determination of soluble solids content in processing tomatoes, J. Am. Soc. Hortic. Sci., 123, 1089, 1998. 82. Williams, P.C. and Sobering, D.C., Comparison of commercial near-infrared transmittance and reﬂectance instruments for analysis of whole grains and seeds, J. Near Infrared Spectrosc., 1, 25, 1993. 83. McCaig, T.N., McLeod, J.G., Clarke, J.M., and Depauw, R.M., Measurement of durum pigment with a near-infrared instrument operating in the visible range, Cereal Chem., 69, 671, 1992. 84. Thomasson, J.A. and Shearer, S.A., Correlation of NIR data with cotton quality characteristics, Trans. Am. Soc. Agric. Eng., 38, 1005, 1995. 85. Apruzzese, F., Balke, S.T., and Diosady, L.L., In-line colour and composition monitoring in the extrusion cooking process, Food Res. Int., 33, 621, 2000. 86. Meléndez-Martínez, A.J., Vicario, I.M., and Heredia, F.J., Application of tristimulus colorimetry to estimate the carotenoids content in ultrafrozen orange juices, J. Agric. Food Chem., 51, 7266, 2003. 87. Meléndez-Martínez, A.J., Vicario, I.M., and Heredia, F.J., Correlation between visual and instrumental colour measurements of orange juice dilutions. Effect of the background, Food. Qual. Prefer., 16, 471, 2004. 88. Meléndez-Martínez, A.J., Vicario, I.M., and Heredia, F.J., Instrumental measurement of orange juice colour: A review, J. Sci. Food Agric., 85, 894, 2005. 89. Meléndez-Martínez, A.J., Vicario, I.M., and Heredia, F.J., Carotenoids, color, and ascorbic acid content of a novel frozen-marketed orange juice, J. Agric. Food Chem., 55, 1347, 2007. 90. Meléndez-Martínez, A.J., Vicario, I.M., and Heredia, F.J., Inﬂuence of white reference measurement and background on the colour speciﬁcation of orange juices by means of diffuse reﬂectance spectrophotometry, J. AOAC Int., 89, 452, 2006. 91. Sheridan, C., O’Farrell, M., Lewis, E., and Flanaga, C., Comparison of CIE L*A*B* and spectral methods for the analysis of fading in sliced cured ham, J. Opt. A: Pure Appl. Opt., 9, S32, 2007.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 247 14.7.2008 10:33am Compositor Name: BMani

Near-Infrared 12 High-Resolution and Nuclear Magnetic Resonance Analysis of Food and Grain Composition Ion C. Baianu and T. You CONTENTS 12.1 12.2

Introduction and Literature Review .................................................................................. 248 NIR and NMR Principles and Techniques ........................................................................ 248 12.2.1 NIR Principles ...................................................................................................... 248 12.2.2 NIR Techniques and Instruments ........................................................................ 249 12.2.3 NIR Calibrations .................................................................................................. 250 12.2.3.1 Regression Methods ............................................................................ 252 12.3 Introducing NMR Techniques for Developing Novel Reference Methods of Food and Grain Composition Analysis ........................................................................ 252 12.3.1 Protein Analysis ................................................................................................... 252 12.3.1.1 Nuclear Magnetic Resonance Spectroscopic Analysis Methods for Protein Determination .................................................... 253 12.3.2 Oil Determination ................................................................................................ 254 12.3.3 Moisture Determinations ..................................................................................... 254 12.3.4 Nuclear Magnetic Resonance Spectroscopic Analysis of Food Carbohydrates: Total Sugars and Fibers .............................................................. 255 12.3.4.1 NMR Techniques for Sugar Determination ........................................ 255 12.3.4.2 Fiber Determination by HR-NMR ...................................................... 255 12.4 Development of Diode-Array Calibrations or Determination of Protein, Oil, Sugars, Fiber, and Moisture Contents in Soybean Seeds ................................................. 255 12.5 Single-Seed Soybean Composition Determination by NIRS ............................................ 256 12.5.1 Calibration Procedures ......................................................................................... 257 12.5.1.1 Spectral Pre-Processing ....................................................................... 257 12.5.2 Single Soybean Seed Calibration Results ............................................................ 257 12.6 DA-NIR Analysis of the USDA-UIUC Germplasm Soybean Collection: A Validation Example of DA-NIR Calibrations ............................................................... 260 12.7 Conclusions and Discussion .............................................................................................. 261 Acknowledgments ........................................................................................................................ 266 References ..................................................................................................................................... 266 Appendix A: Standard Reference Methods for Component Analysis ......................................... 270 Appendix B: A Literature Review of Soybean Chemistry .......................................................... 278

247

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 248 14.7.2008 10:33am Compositor Name: BMani

248

Handbook of Food Analysis Instruments

12.1 INTRODUCTION AND LITERATURE REVIEW The food industry and nutritional sciences have a great need for rapid analytical techniques that are economical, accurate, reproducible, and preferably nondestructive. Is FT-NIR spectroscopy capable of rapid, reproducible analyses of food grains and foods? Can this result in signiﬁcant savings of time and money in research and food industry laboratories? The standard chemical methods for protein, oil, and sugar determination are time consuming and expensive, inappropriate for sample sizes as large as 10,000–1,00,000 or more samples. Therefore, a rapid, accurate, and inexpensive measurement method is highly desirable. Nuclear magnetic resonance (NMR) is a nondestructive method and can provide rapid, accurate, and inexpensive measurements. It has been applied to protein and oil determination in intact wheat, [1,4,54,55], soybean seeds [11,12,53–55], and sugar determination in fruit [17], but so far it has not been routinely applied to soy foods [43] although it was, and still is, employed industrially inline for certain other foods such as meat, butter, soft cheese, Texturized Vegetable Products (TVP), and chocolate products. Near-infrared spectroscopy (NIRS) is another nondestructive method, which is, however, faster and also much less expensive than NMR in most small-grain related applications. NIRS has been widely tested for chemical analysis of seed composition, but it does require calibration by more expensive, time-consuming, and usually destructive chemical analysis; it has already been in use for a long time in the area of new grain development, genetic selection, and cross-breeding. However, NIRS has not been extensively tested for total sugar determinations in either grains or foods. Developmental laboratories and grain laboratories in industry have been reluctant to use NIR because of the low quality of instruments available (until ~6 years ago), and the lack of proper calibrations. Because of its high sensitivity, NIR is useful as a rapid and inexpensive screening tool, despite not having usually very high resolution. Recent NIR spectrometers that utilize Fourier transform (FT) fulﬁll all these conditions, and are also capable of very high resolution, limited only by sample component spectral overlaps, but require precalibration by AOCS-approved wet chemistry techniques, using well-deﬁned and stable sample standard sets of 50–100 different samples. The development of NIR calibrations for determination of protein, oil, moisture, total sugars, and ﬁbers in foods and grains allows the subsequent use of such NIR calibrations to measure rapidly, economically, and accurately, for example, hundreds of thousands of unknown sample compositions from a variety of sources over the span of just a few years [27–34]. Over 60,000 measurements were carried out during the last 4 years by NIRS at our NMR and NIR microspectroscopy (UIUC) facility on a wide selection of grains, developmental soybean cultivars, and foods. To develop NIRS calibrations, a set of reference values of standard samples composition are ﬁrst determined by wet chemistry=analytical methods. Speciﬁc, major components selected for such determinations were protein, oil, moisture, total sugars, and ﬁber. The speed of the NIRS technology has been combined with the accuracy of several high-resolution NMR techniques to develop rapid, inexpensive, and precise NIRS methods for food and grain composition analysis [11–13,23–26].

12.2 NIR AND NMR PRINCIPLES AND TECHNIQUES 12.2.1 NIR PRINCIPLES Molecules have many quantized vibration and rotation states that can also occur in combination. Thus, a molecule with an electric dipole moment can undergo transitions between various vibrorotational states if electromagnetic radiation with the appropriate frequency is absorbed by the molecule. The vibration energy levels between different vibration and rotation states are quantized and the energy levels can be approximated by the following equation: sﬃﬃﬃﬃ 1 h k En ¼ n þ 2 2p m

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 249 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

249

where n is the vibrational quantum number h is the Planck’s constant k is a bond’s force constant m is the reduced mass of the vibrating nucleus The transitions can occur with high probability for Dn ¼ 1, whereas such absorption processes are called fundamental absorption when Dn ¼ 1 or overtone when Dn ¼ 2, 3, . . . , m, respectively. The electromagnetic radiation absorbed for vibration state changes is in the infrared region, while NIR uses absorption bands whose absorption is mainly due to overtones. NIR instruments operate with electromagnetic radiation wavelengths between 700 nm and 2500 nm, much shorter than that of NMR wavelengths, which are from 1 to 1000 m. Different molecules have different constants of k and m, and, thus, different NIR absorption frequencies (bands) in their spectrum. Oil has NIR absorption bands at 890, 1162, 1720, 1760, 2308, and 2340 nm for the –CH2 groups. Proteins have NIR absorption bands at 1040, 1210, 1496, 2050, and 2140–2180 nm for NH. Water=moisture has NIR absorption bands at 960, 1150, 1405, and 1905–2000 nm region for the uncharged hydroxyl group, –O–H. Sugars and carbohydrates have NIR absorption bands at 2060–2150 nm for the carbonyl group, –C¼O. This makes possible both the qualitative and quantitative analysis by NIR spectroscopy; the Lambert–Beer law states that the quantity of electromagnetic energy absorbed by any sample is proportional to the amounts of constituents present in the sample, and this allows quantitative analyses to be made if the absorption process can be separated from other processes such as scattering and specular (mirror-like) reﬂection processes. Just like visible light, the near-infrared radiation can be reﬂected, transmitted, and absorbed. NIR can be used in either the reﬂectance or the transmission mode. As shown earlier, different constituents have absorption bands at different wavelengths [19–21].

12.2.2 NIR TECHNIQUES

AND INSTRUMENTS

NIR reﬂectance instruments use electronic detectors to measure the intensity of NIR radiation that is reﬂected from the sample at several different wavelengths. The actual constituent contents can be calculated based on the above calibration equation given the reﬂectance at selected key wavelengths. On the other hand, NIR transmission instruments measure the intensity of NIR radiation that is transmitted through a sample at several different wavelengths. A calibration equation is then established to relate (log of reﬂectance) values at selected key wavelengths to constituent fractional content values by comparison with wet chemistry analysis obtained for a standard set of samples by a primary, or reference, method. Since NIR transmission instruments measure the NIR electromagnetic radiation that is actually transmitted through a sample, it is important that the path length be kept constant and also selected for a very high signal-to-noise ratio. NIR instruments can also be classiﬁed as discrete-region=ﬁlter systems or continuous spectrum detection systems based on their wavelength separation mechanism. Discrete ﬁlter instruments select wavelengths by passing white light—which is produced, for example, by a tungsten–halogen bulb—through a ﬁlter, which allows only a speciﬁc wavelength, narrow region to pass through it. The discrete ﬁlter instruments do not collect data at all wavelengths, but rather only around the wavelengths of interest. The most important advantage of a discrete ﬁlter instrument is the reproducibility of its narrow wavelength ranges. The primary limitation of a ﬁlter-based instrument is that NIR absorption data are collected at only a few speciﬁed narrow-range wavelengths, and therefore the initial wavelength range selection might be difﬁcult if the sample matrix is unknown. Filter-based, discrete wavelength instruments also tend to be slow if they are not combined with simultaneous diode-array (DA) detection for several wavelength ranges. Another limitation of ﬁlter-based instruments is their rather limited spectral resolution; for very broad NIR absorption bands, this may not be a problem especially if the selected ﬁlters satisfy the Nyquist spectral sampling criterion [69].

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 250 14.7.2008 10:33am Compositor Name: BMani

250

Handbook of Food Analysis Instruments

Continuous NIR spectroscopy instruments allow the collection of absorption data for very large numbers of wavelengths and, thus, can be used to select=ﬁnd wavelengths of interest for unknown matrices to develop a precise calibration method. Continuous spectra systems can actually be divided into three subgroups: the moving grating=scanning systems; stationary grating systems, which would include both diode array and AOTF systems; and Fourier-transform NIR spectrometers. Scanning NIR instruments usually use a moving grating to collect data on all wavelengths and it is difﬁcult to obtain reproducible scans and the wavelength accuracy is affected. The stationary grating systems usually use parallel processing diode array to collect data on all wavelengths and wavelength reproducibility and accuracy are signiﬁcantly improved. Another advantage is that the scanning speed is signiﬁcantly improved. A moving grating system usually takes about 30 s to perform a scan, while the diode-array based stationary grating system is claimed to be able to collect hundreds of spectra in 1 s [57]. Two recent NIR grain-tester instruments, models Zeltex ZX800 and Zeltex ZX50, were tested, together with a diode array, more sophisticated model, the Perten DA7000. Zeltex ZX800 and ZX50 are ﬁlter-based transmission instruments, with 13 and 14 ﬁlters, respectively; the wavelengths range for both instruments is from 893 to 1045 nm. The DA7000 diode-array model is based on a stationary grating instrument that allows the entire spectral range from visible 500 nm to NIR 2500 nm to be acquired simultaneously in less than one second by employing a large array of silicon and InGaAs diodes. The ZX800 and ZX50 are commercially available from Zeltex Co., (USA) with preliminary calibrations of protein, oil, and moisture for soybean, wheat, or corn seeds. A total sugar calibration was also established to be possible with these two instruments [17]. The existing calibrations on ZX800 and ZX50 are mainly for regular yellow-coated seeds, and calibrations for seeds with other colors are very difﬁcult and inaccurate. Previous research has shown that NIR reﬂectance and transmission instruments have similar accuracy and reproducibility [22,46,69]. However, the DA7000 is a new type of reﬂectance instrument, and given the fact that 86% of the hull consists of carbohydrates, with only 9% protein and 1% oil [40], the analysis results from ZX800 and DA7000 were compared to investigate the difference between transmission and reﬂectance instrument accuracy as well as overall performance, especially when large sample sizes from different seed varieties=cultivars are available for measurement. The DA7000 instrument is distributed commercially in the USA by Perten Instrument North America Inc., but has no calibrations available for soybeans or corn seeds with the exception of those developed in our laboratory for these grains, as well as for green soybeans and 2- or 3-soybean seeds. The total sugar calibration for ZX800 and ZX50 as well as the calibrations for protein, oil, moisture, and total sugars for the DA7000 model were based on our high-resolution NMR reference methods [1–9,11–13], as well as the standard primary methods described in Appendix A. Three different FT-NIR spectrometer models were extensively tested and their performance for food and grain analysis was compared [69]: a PerkinElmer SpectrumOne-NTS instrument, and the latest (in years 2005=2006) Bruker and Nicolet Technologies, NIR-model spectrometers. The SpectrumOne-NTS model, also commercially available in the United States, appears to have both hardware advantages of a well-designed integrating sphere (NIRA), sample compartment, and lower cost, as well as the ability to transfer calibrations to similarly equipped models. Furthermore, the latter model is readily interfaced with a room temperature, very high sensitivity Sb-detector for FT-NIR chemical=hyperspectral, or microspectroscopic imaging capable of ~1 mm resolution with ~10 pg sensitivity. On the other hand, the other two NIR instrument models tested have more versatile, ﬂexible, and much faster calibration software than the PerkinElmer model tested.

12.2.3 NIR CALIBRATIONS NIR instruments determine protein and other components by measuring the absorbance-log (1=R) values, which must be then related to the fractional content of the component as determined by some other method called reference or standard method. The process of establishing such a relationship by

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 251 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

251

using a set of samples of known composition is called a calibration. The relationship between the absorbance and chemical composition value is usually expressed as an approximation and always involves some forms of regression equations. According to the Beer–Lambert law, the absorbance at a selected wavelength is proportional to the concentration of the component for solution of one component: A ¼ log10

1 ¼ kcl R

where k is the molecular absorption constant l is the path length of the NIR radiation passing through the sample c is the concentration of the absorbing component If a sample contains several components, then the absorption at any wavelength will be the sum of the contributions from all components in the sample: Al1 ¼ x11 c1 þ x12 c2 þ þ x1n cn Al2 ¼ x21 c1 þ x22 c2 þ þ x2n cn .. . Aln ¼ xn1 c1 þ xn2 c2 þ þ xnn cn The measured absorbance is usually referred to as the apparent absorbance, and it can be signiﬁcantly affected by specular reﬂection and light scattering, even for thin samples. Therefore, to obtain reliable NIR quantitation, spectral pre-processing and several intensity corrections are always required. Thus, spectral variations between soybean samples can be caused not only by chemical composition differences but also by spurious effects that do not monitor chemical composition, such as specular reﬂection, multiple scattering effects, and internal reﬂection. Our research group has found that the NIR methods currently employed in industry for spectral pre-processing, including both corrections for multiple light scattering and specular reﬂection effects, are in need of substantial improvements to produce high accuracy, robust, and stable calibrations for rapid composition analyses of seeds. Our NIR calibrations were established mainly for foods and whole kernel seeds, but calibrations for ground samples were also successful. Recently, protein content determinations were reported for single wheat kernels with both transmission [46,49,65] and reﬂectance instruments [51,57]. Oil determinations for single corn kernels were also reported with a transmission NIR instrument [48]. Wheat single-seed studies were also reported by NIR reﬂectance spectroscopy using the DA7000 instrument [57]. Calibrations for protein, oil, moisture, and total sugars of single soybean seeds were obtained in our laboratory both with the DA7000 and the SpectrumOne-NTS instruments. The following are the principal NIR calibration steps: . . . .

.

Generate or select a suitable set of standard samples of known composition. Obtain raw FT-NIR data. Correct NIR data for scattering. Use Lambert–Beer law computations in conjunction with iterated data regression by PLS-1 or 2; also check up on speciﬁc PLS-1 software packages for precision and correct computation through numerical simulations for ideal testing, synthetic=numerical data (Figure 12.1). Examine the calibration’s linear correlations, composition predictions, and validate calibration with a wide range of unknown samples.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 252 14.7.2008 10:33am Compositor Name: BMani

Absorbance

252

0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 –0.1 700

Handbook of Food Analysis Instruments

C1

900

1100

1300

1500

1700

1500

1700

Absorbance

Absorbance

Wavelength, nm 1.1 0.9 0.7 0.5 0.3 0.1 –0.1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 –0.1 700

C2

C3

900

1100 1300 Wavelength, nm

C3 0 100 10 90 20 80 30 70 40 60 50 50 60 40 70 30 80 20 90 10 0 100 C1 0 10 20 30 40 50 60 70 80 90 100 C2

FIGURE 12.1 Computer simulation study: Testing the PLS-1 regression calibration algorithm with the ideal 3-component composition matrix shown at right.

12.2.3.1

Regression Methods

There are several regression methods that have been tested with NIR calibrations. Most widely used regression methods include partial least square (PLS) [29,49], principal component regression (PCR) or principal component analysis (PCA) [29,44,46], and multiple linear regression (MLR) [44,46]. The most widely used spectra pretreatment method is multiple scattering correction (MSC) [29]. DA7000 has built-in software able to carry our PLS, PCR, and PCA regressions and it can also enable or disable MSC. The PLS, PCR, and PCA regression methods were all tried on DA7000, with or without MSC correction (Figures 12.2 and 12.3). The best combination with the highest correlation coefﬁcient is adopted.

12.3 INTRODUCING NMR TECHNIQUES FOR DEVELOPING NOVEL REFERENCE METHODS OF FOOD AND GRAIN COMPOSITION ANALYSIS 12.3.1 PROTEIN ANALYSIS The atomic composition of proteins consists of hydrogen, carbon, nitrogen, oxygen, and sulfur. There are 20 a-amino acid naturally occurring as building blocks of protein. The linkage between amino acid residues in a protein are peptide bonds. Nitrogen is the most distinguishing element present in proteins. However, nitrogen content in various food proteins ranges from about 13% to 19% due to the variation in the speciﬁc amino acid composition of proteins. Protein analysis is complicated by the fact that some food components possess similar physicochemical properties. Nonprotein nitrogen could originate from free amino acids, small peptides, nucleic acids, phospholipids, amino sugars, porphyrin, some vitamins, alkaloids, uric acid, urea, ammonium ions, and perhaps other organic contaminants such as herbicides or fertilizers. Therefore, the total organic

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 253 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

253

(A)

(B)

FIGURE 12.2 (See color insert following page 240.) SpectrumOne NTS spectra of bulk soybean samples, before (A) and after (B) multiple scattering correction (MSC). Note the very signiﬁcant consistency of data obtained after applying MSC.

nitrogen in foods would represent nitrogen primarily from proteins and to a lesser extent from all organic nitrogen-containing nonprotein substances. Numerous methods have been developed to measure protein content. The basic principles of these methods include the determinations of nitrogen, peptide bonds, aromatic acids, UV absorption of proteins, free amino groups, light scattering properties, and dye-binding capacity. In addition to factors such as sensitivity, accuracy, precision, speed, and cost of analysis, what is actually being measured must be considered in the selection of an appropriate method for a particular application. Standard reference methods for protein analysis are summarized in Appendix 1. 12.3.1.1

Nuclear Magnetic Resonance Spectroscopic Analysis Methods for Protein Determination

Several NMR spectroscopy techniques can also be used as primary, or reference, methods for protein content determination. The earlier, solid state NMRS methods [1,4–6,24,50–52,54–66] can be used in principle to determine protein content nondestructively, but they have a relatively low sensitivity and may suffer from interference from other components present; thus, it has been previously suggested that the carboxyl peak at ~172–181 ppm in CP-MASS 13C spectra of various proteins—including food proteins—can be used for determination of protein contents [1,4,6]. On the other hand, protein content determinations are far more accurate either in protein gels [11,12] or protein solutions [2,3,5,37].

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 254 14.7.2008 10:33am Compositor Name: BMani

254

Handbook of Food Analysis Instruments 1.4

Absorbance

1.2 1.0 0.8 0.6 0.4 0.2 0 12,000

11,000

10,000

9,000

7,000

8,000

6,000

5,000

4,000

–1

(A)

Wavenumber, cm 0.8

Absorbance

0.7 0.6 0.5 0.4 0.3 0.2 0.1 0 12,000

11,000

10,000

(B)

9,000

8,000

7,000

6,000

5,000

4,000

Wavenumber, cm–1

FIGURE 12.3 SpectrumOne–NTS FT-NIR spectra of soy protein isolates (SPI) in H2O, before (A) and after (B) multiple scattering correction (MSC).

12.3.2 OIL DETERMINATION Compared with the protein determination, oil determination is relatively straightforward. Both oil and fats belong to lipids, which by deﬁnition, is a group of substances generally soluble in organic solvent and insoluble in water. Oil refers to the liquid lipids at room temperature while fat refers to the solid lipids at room temperature. NMR analysis of total oil and fats is possible by both lowand high-resolution NMR techniques; however, HR-NMR also allows the rapid identiﬁcation of all individual molecules present in oil [12,38]. Nondestructive NMR analyses are possible without any extraction or solubilization of oils in organic solvents, even in samples as small as early developing soybean embryos of about 200 mm in size [12]. Microscopic imaging of oil and moisture distribution in mature wheat grains has been carried out as early as 1979 at ~50 mm resolution by 1 H (N) MRI microscopy on intact, hydrated wheat grains as reported in Ref. [5]; the distinction between germ oil and surrounding moisture in the intact wheat grain was then made by enhanced T1-contrast imaging.

12.3.3 MOISTURE DETERMINATIONS Moisture determination is also one of the more important and frequent analyses for a food product as the quantitation of component fractions is ﬁnally expressed on a dry basis. It is also one of the more difﬁcult to obtain accurate and precise data if special precautions were not taken before and during calibrations, such as tight sealing of all standards with as little as possible headspace, especially for small samples. Generally, water is present in food in several forms—trapped or multilayer water, absorbed or weakly bound water, and tightly bound water [5,7–9]. Only excess water—as in blood, cooked-juicy meet, liquid milk, or fruit juices—can be classiﬁed as free water; often the latter term is too freely used where it does not apply as the only fraction with a water activity of 1.00 is the

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 255 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

255

excess water. Furthermore, various water fractions found in hydrated solid foods or gels exchange with each other either rapidly or slowly, thus contributing to many incorrect ideas, or so-called water-binding theories about the presence of free versus bound water in foods. Low-ﬁeld= low-resolution pulsed NMR [66] is one of the methods partially accepted by AOCS for moisture determination in grains as it may have several drawbacks especially in the low moisture content=low water activity samples. On the other hand, high-resolution or high-ﬁeld NMRS or selective NMR relaxation techniques [5] provide superior means for moisture or liquid water determination, simultaneously with the determination of all major components in foods or grains. The major disadvantage is cost and need of adequately trained personnel to carry out the NMR measurements as well as the correct data analysis.

12.3.4 NUCLEAR MAGNETIC RESONANCE SPECTROSCOPIC ANALYSIS TOTAL SUGARS AND FIBERS

OF

FOOD CARBOHYDRATES:

HR-NMR is an extremely powerful tool for both qualitative and quantitative analysis of food carbohydrates [5,6]; it can be nondestructive, very rapid, and can also be used to determine the absolute structure=conformation of a carbohydrate in either solutions or the solid state [5,8]. The most widely used HR-NMR techniques for quantitative carbohydrate analysis are based on the fact that the area under the peaks is proportional to the number of protons or carbon atoms present in the sample, depending on which speciﬁc nucleus is observed, i.e., 1H or 13C, in this case. 12.3.4.1

NMR Techniques for Sugar Determination

It has been shown in many studies that high-resolution, 1H-decoupled=13C NMR determinations of sugars in tropical root crops give similar results and have similar reproducibility to HPLC [60]. According to Tamate and Bradbury [60], in 13C NMR spectroscopy, the sugars normally present in foods, e.g., glucose, fructose, sucrose, maltose, and rafﬁnose, exhibit speciﬁc or ﬁngerprint resonances that allow one to distinguish them from each other. Dioxan and (CH3)3SiCD2CD2CO2Na may be used as internal references for relative intensity measurements and for chemical shift, respectively. The quantity of each sugar was then determined by measurement of the ratios of the peak heights of the sugar resonance compared with peak height of the internal standard. Meredith et al. [42] also used 13C NMR for study of saccharides of developing wheat grains. Besides the peak area in HR-NMR spectroscopy, the spin–spin relaxation times (T2) in 1H NMR [7] have also been used for sugar content determination, as for example in intact fruit [17]. 12.3.4.2

Fiber Determination by HR-NMR

Dietary ﬁber generally stands for lignin plus plant polysaccharides that cannot be digested by human enzymes. The major components of dietary ﬁber include cellulose, hemicelluloses, pectins, hydrocolloids, and lignin. Dietary ﬁber can be determined either gravimetrically or chemically. In gravimetrical method, the digestible carbohydrates, lipids, and proteins are selectively solubilized by chemicals or enzymes. The remaining undigestible materials are then collected by ﬁltration and the ﬁber residue is quantitated gravimetrically. In chemical method, the digestible carbohydrates are removed by enzymatic digestion, ﬁber components are then hydrolyzed by acid, and the monosaccharides produced are measured. The ﬁber content can be calculated from the total monosaccharides in the acid hydrolysate. Fiber as well as starch determinations by solid-state HR-NMR have also been reported [1,4,5–8].

12.4 DEVELOPMENT OF DIODE-ARRAY CALIBRATIONS OR DETERMINATION OF PROTEIN, OIL, SUGARS, FIBER, AND MOISTURE CONTENTS IN SOYBEAN SEEDS Conventional continuous spectra instruments use a moving grating to allow data collection at different wavelengths, but they suffer from the intrinsic reproducibility problem because of the

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 256 14.7.2008 10:33am Compositor Name: BMani

256

Handbook of Food Analysis Instruments 3.5 Brown Black Yellow Green

3.0

Absorbance

2.5 2.0 1.5 1.0 0.5 0.0 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 Wavelength, nm

FIGURE 12.4 Comparison of NIR and visible diode-array spectra of regular, yellow coat with other color coat whole soybean seeds showing the marked effects of the coat color, especially for black- and green-coat soybean seeds extending well beyond 1100 nm in the NIR range. Such effects are responsible for the large errors in many NIR calibrations of either black- or green-coat whole soybean seeds that did not correct for such colored coat absorption bands before the regression analysis (i.e., during NIR spectral data pre-processing).

use of the moving grating. Recently, Perten Instruments Inc. introduced a new generation of continuous-scanning spectrometer model DA7000, which uses a diode-array based stationary diffraction grating. Because DA7000’s operation does not involve either a moving grating or a monochromator, it completely overcomes the reproducibility problems encountered with moving grating spectrometers and is also much less affected by environmental conditions. To facilitate the plant breeding for improved soybean varieties with high protein, high oil, and low sugar for food applications, the major components, that is, protein, oil, moisture, and sugars all need to be determined. Reliable and accurate NIRS calibrations for the determination of these major soybean components were developed on both the DA7000, DA-instrument, and the Spectrum One—NTS FT-NIR reﬂectance spectrometer for regular sample size analysis (20 g or less can be used) [69]. Although spectra were collected with the DA7000 spectrometer both in the visible and NIR ranges, from 400 to 1700 nm, soybean seed coat color seems to signiﬁcantly affect also the NIR reﬂectance (Figure 12.4), for example depending on the maturity of the soybean seeds with coat colors varying from green to yellow. To minimize the coat color effects, only the absorbance data between 780 and 1680 nm were employed for NIRS calibration developments on green(ish) colored soybeans. Additional spectrum baseline corrections were also carried out with the spline function as described in Section 12.2.

12.5 SINGLE-SEED SOYBEAN COMPOSITION DETERMINATION BY NIRS Acquisition with a DA7000 spectrometer is very fast: this instrument can collect up to 600 raw spectra per second [57]; however, the DA7000 spectral averaging is much slower and limited by its A=D converter and its Windows-based PC. Single-seed measurements, however, require a suitable experimental setup for both higher NIR sensitivity and reproducibility. Our research group at UIUC has, therefore, investigated its suitability for single-seed analysis in relation to experimental plant breeding programs carried out by other collaborating groups (as speciﬁed in the Acknowledgments). If successful, such attempts would have deﬁnite practical advantages because it has

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 257 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

257

been demonstrated that the breeding efﬁciency will signiﬁcantly improve if the breeding selection can be performed on a single seed or kernel basis [10,57]. Nondestructive determination of major seed constituents thus becomes unavoidable if single-seed or single-kernel breeding techniques must be utilized. NIRS becomes one of the most prominent solutions for the grain single seed or oil seed composition determination. Previous research has indicated the potential of NIRS instrumentation for the measurements of moisture in soybeans [39], oil in corn [48] or soybeans [10,11,18], and protein in wheat [26]. All the early work only brieﬂy considered here used NIR transmittance instruments (see for example Figure 12.5), and it is only in the past 8 years that sufﬁciently sensitive NIR reﬂectance instruments could also be used on wheat single kernels for individual seed color classiﬁcations [63]. The possibility of determining the major soybean seed components was investigated in singleseed measurements (for seeds ranging from ~50 to ~130 mg in weight), and then the calibration results were compared with those obtained for bulk samples obtained with the same instrument that was already illustrated in Section 12.4 and in Figures 12.6 and 12.7. Here we shall focus on single soybean seed NIRS calibrations, data analysis, and results. The soybean standard samples utilized for our calibration developments were drawn from the USDA National Soybean Collection at UIUC (Urbana, Illinois), which holds more than 18,000 soybean accessions, one of the largest—if not the largest—soybean germplasm collections in the world. Additional standards were obtained from specially grown developmental lines for either high protein or high oil [36,45] by the research group of Dr. R.L. Nelson in the Crop Science Department at UIUC (see also Acknowledgments).

12.5.1 CALIBRATION PROCEDURES Calibration developments were performed by employing a PLSplus=IQ program in GRAMS=32 (Graphic Relation Array Management System) [29] software package from the Galactic Industries Corporation (Salem, New Hampshire). The regression of calibration models was performed by partial least squares, Type 1 (PLS-1). In PLS-1 regressions, the spectra are ﬁrst decomposed into principal components weighted by each constituent individually and then the regression is performed on the scores of each sample NIR spectrum, so that each constituent is optimized individually. Cross-validation was used to optimize the calibration models. The number of factors was determined based on the criterion of the lowest standard error of cross-validation (SECV). 12.5.1.1

Spectral Pre-Processing

For single-seed calibration, before PLS-1 processing, all the spectra were mean-centered ﬁrst, and then pretreated with or without mathematical transformations. The effects of light scattering were investigated by the application of MSC method. The baseline effects were investigated by the application of the ﬁrst-order and second-order derivatives; both were calculated by the Savitsky– Golay algorithm with 5 data points (SG1–5). There were four combinations of pre-processing involved: (1) no pre-processing, (2) multiple scattering correlation, (3) MSC using the ﬁrst-order derivative (MSC-SG1), and (4) MSC with the second-order derivative (MSC-SG2). For two-seed calibrations and three-seed calibrations, only the MSC with the SG1–5 pre-processing procedure was applied before the PLS-1 regression, as the single-seed calibration experiment showed semiempirically that the MSC with SG1–5 was the best pre-processing choice.

12.5.2 SINGLE SOYBEAN SEED CALIBRATION RESULTS Using MSC and the ﬁrst derivative for the DA7000 acquired NIR spectra, the correlation coefﬁcient for the protein calibration plot was 97.7% and the standard error of cross-validation was 0.92% (which is very close to the laboratory standard error of determination of proteins in single soybean seed samples). Oil, moisture, sugars, and ﬁber had correlation coefﬁcients of 97.2%, 98%, 97.7%, and 97.9%, respectively. The SCV values thus obtained were 0.48%, 0.320%, 0.18%, and 0.66% for oil, moisture,

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 258 14.7.2008 10:33am Compositor Name: BMani

258

Handbook of Food Analysis Instruments ZX50 Spectrum 0.9 0.8 0.7

Absorbance

0.6 0.5 0.4 0.3 0.2 0.1 0 890 900 910 920 930 940 950 960 970 980 990 1000 1010 1020 1030 1040 1050 (A)

Wavelength, nm ZX50 Correlation chart

1000.00 Protein Moisture Oil Sugars

500.00

0.00 K value

880

900

920

940

960

980

1000

1020

1040

1060

–500.00

–1000.00

–1500.00 (B)

Wavelength, nm

FIGURE 12.5 (A) Whole soybean seed composition, nondestructive analysis by NIR. (B) Comparison of the correlation curves for major soybean components with the factory-selected center positions of the ZX50 narrowband (DA) ﬁlters.

sugars, and ﬁber, respectively. Such prediction errors are fractionally larger for oil and ﬁber than the laboratory determination standard errors for single-seed samples with the latter, in their turn, being also larger than the laboratory determination errors for bulk samples by a factor of ~1.4, especially for oil

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 259 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

259

1.5

Absorbance

1.0

Protein correlation Pure protein

0.5 0.0 750

850

950

1050 1150 1250 1350 1450 1550 1650

–0.5 –1.0 –1.5 (A)

Wavelength, nm 1.2 Oil correlation 1.0

Pure oil

Absorbance

0.8 0.6 0.4 0.2 0.0 750 –0.2

850

950

1050 1150 1250 1350 1450 1550 1650

–0.4 –0.6 (B)

Wavelength, nm

FIGURE 12.6 (A) NIR spectrum of puriﬁed SPI protein superimposed on the soybean seed protein correlation curve that shows several sharp and large correlation peaks. (B) NIR spectrum of pure oil superimposed on the soybean seed oil correlation curve with several, sharp and large correlation peaks. (continued )

and ﬁber, but still practically useful for the initial screening and soybean seed line selection of desirable traits such as soy protein and oil for large numbers of single-seed candidates. A single soybean seed calibration plot, i.e., the predicted protein content versus the reference (laboratory) protein content is shown in Figure 12.8. The largest predicted protein content variation between the eight spectra of each single soybean seed is ~1%, which is equivalent to the standard error or cross-validation. It shows that there are signiﬁcant spectral variations among the consecutive eight DA-7000 scans. On the other hand, both the NIRS data and the results obtained with the FT-NIR spectrometer, SpectrumOne-NTS model for single soybean seeds, or even half-seeds, were net superior to the results obtained with the DA7000 for regular color soybean coats; for half-seeds, coat color did not affect at all the accuracy of the NIRS predictions obtained with the speciﬁed FT-NIR spectrometer model (data not shown); in the latter case the SECV’s were very close in value with the laboratory determination standard errors for the single seed (ﬁve) major constituents listed above. This also allowed an extensive, 3-year NIRS study at UIUC of single-seed composition harvested from single soybean plants of both known pedigree and bulk composition that were all grown under well-deﬁned environmental conditions on single lots [18,45,69].

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 260 14.7.2008 10:33am Compositor Name: BMani

260

Handbook of Food Analysis Instruments 2.5 Moisture correlation Pure water

Absorbance

2.0

1.5

1.0

0.5

0.0 750

850

950

1050 1150 1250 1350 1450 1550 1650

–0.5 Wavelength, nm (C) 1.5 Sugar correlation Pure sucrose

Absorbance

1.0

0.5 a 0.0 750

850

950

1050 1150 1250 1350 1450 1550 1650

–0.5

–1.0 (D)

Wavelength, nm

FIGURE 12.6 (continued) (C) NIR spectrum of liquid water superimposed on the soybean moisture correlation curve with several, sharp and large correlation peaks. (D) NIR sucrose spectrum superimposed on the soybean seed total sugar correlation curve that shows several sharp and large correlation peaks.

Bulk soybean sample NIRS analyses are reported next for a much larger collection of soybeans [45,69] than the one discussed in this section (see Figures 12.9 through 12.13).

12.6 DA-NIR ANALYSIS OF THE USDA-UIUC GERMPLASM SOYBEAN COLLECTION: A VALIDATION EXAMPLE OF DA-NIR CALIBRATIONS The results presented in this section when taken together with our recently published data analysis [10,11–13,36,45,69] that were obtained by combining NIRS measurements with selective crossbreeding experiments (see relevant examples in Figure 12.9; Figures 12.11 through 12.13) indicate that soybean PIs have the genetic potential for improving seed yield of US soybean cultivars, as well as increase the protein content, signiﬁcantly above the current, commercial range of the soybeans marketed in the United States.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 261 14.7.2008 10:33am Compositor Name: BMani

261

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis 2.2 Sucrose Oil Water Fiber Protein WholeSoy

2.0 1.8 1.6 Absorbance

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 400

500

600

700

800

900 1000 1100 1200 1300 1400 1500 1600 1700 Wavelength, nm

FIGURE 12.7 (See color insert following page 240.) NIR spectra of the major components present in soybean seeds compared with the NIR spectrum of the whole soybean seed.

12.7 CONCLUSIONS AND DISCUSSION High-accuracy protein, oil, and total sugar statistical distributions, as well as their correlations, are reported for soybeans of different genetic origin. Relevant NIRS calibration and measurement examples of 3 years harvests of developmental cultivars labeled MAPIII (1997, 1998, 1999)

y = 0.9699x + 1.1994 2 = 0.9766 R 52 54

50 Predicted protein , %

48 46 44 42 40 38 36 34 32 30 30

32

34

36

38

40

42

44

46

48

50

52

54

56

Reference protein , %

FIGURE 12.8 NIR predicted protein content versus reference (laboratory) protein content for single soybean seeds. Calibration conditions: multiple scattering corrections with ﬁrst-order derivative and a 7-factor PLS-1 regression.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 262 14.7.2008 10:33am Compositor Name: BMani

262

Handbook of Food Analysis Instruments 24.00 22.00

Oil, wt%

20.00 18.00 16.00

y = –0.4328x + 40.05 R 2 = 0.88

14.00 12.00 10.00 35.00

37.00

39.00

41.00

43.00

45.00

47.00

49.00

51.00

53.00

55.00

Protein, wt%

FIGURE 12.9 Protein-oil correlation plot for the Map 3-97 Developmental Soybean Group of 5,000 soybean lines.

illustrate the high protein–oil (inverse) correlation for soybeans of similar genetic origin=source. For soybeans within the same genetic line, the effects of harvest year, planting conditions, and environment are also discussed. The data analyses of single plant harvests for three consecutive years are also presented as good examples of the validation for these novel NIRS calibrations.

28.00 26.00 24.00

Oil, w t%

22.00 20.00 18.00 16.00 y = –0.4713x + 41.795 R 2 = 0.67

14.00 12.00 10.00 30.00

35.00

40.00

45.00

50.00

55.00

Protein, wt%

FIGURE 12.10 NIR analysis of the USDA-UIUC soybean germplasm collection: Germplasm 990289 protein–oil correlation. Baianu, I.C., You, T., Costescu, D.M., Prisecaru, V., Lozano, P., and Nelson, R.L., Analysis of Oil Seeds., AOCS Public: Champaign, IL, 2004a.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 263 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

263

Protein distribution of F3 Wms 82(3) ⫻ PI 82.278 100 Average: 47.0% 90 80

# Samples

70 60 50 40 30 20 10 0 <44.5 44.5– 45.0– 45.5– 46.0– 46.5– 47.0– 47.5– 48.0– 48.5– 49.0– 49.5– 50.0– 50.5– >51.0 45.0 45.5 46.0 46.5 47.0 47.5 48.0 48.5 49.0 49.5 50.0 50.5 51.0 (A)

Protein (%)

Oil distribution of F3 Wms 82(3) ⫻ PI 82.278 100 Average: 18.3% 90 80

# Samples

70 60 50 40 30 20 10 0 <16.0 16.0– 16.3– 16.6– 16.9– 17.2– 17.5– 17.8– 18.1– 18.4– 18.7– 19.0– 19.3– 19.6– >19.9 16.3 16.6 16.9 17.2 17.5 17.8 18.1 18.4 18.7 19.0 19.3 19.6 19.9 (B)

Oil (%)

FIGURE 12.11 Protein and oil content distributions for developmental soybean cultivars with very high protein content. (A) Protein content distribution for the F3 Wms 82(3) PI 82.278 soybean lines of selected pedigrees. (B) Oil distribution for the F3 Wms 82(3) PI 82.278 soybean lines of selected pedigrees.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 264 14.7.2008 10:33am Compositor Name: BMani

264

Handbook of Food Analysis Instruments Protein distribution of F3 Wms 82(3) ⫻ HHP 100 Average: 46.3% 90 80

# Samples

70 60 50 40 30 20 10

(A)

.0 0 – 43 43. .5 5 – 44 44. .0 0 – 44 44. .5 5 – 45 45. .0 0 – 45 45. .5 5 – 46 46. .0 0 – 46 46. .5 5 – 47 47. .0 0 – 47 47. .5 5 – 48 48. .0 0 – 48 48. .5 5 – 49 49. .0 0 – 49 49. .5 5 – 50 50. .0 0 – 50 50. .5 5 – 51 51. .0 0 –5 1. 5 >5 1. 5

3. 43

2. <4

42

.5

–4

5

0

Protein (%) Oil distribution of F3 Wms 82(3) ⫻ HHP 100 90

Average: 18.3%

80 70

# Samples

60 50 40 30 20 10 0

(B)

<16.0 16.0–16.3– 16.6– 16.9– 17.2– 17.5– 17.8– 18.1– 18.4– 18.7– 19.0– 19.3– 19.6– >19.9 16.3 16.6 16.9 17.2 17.5 17.8 18.1 18.4 18.7 19.0 19.3 19.6 19.9 Oil (%)

FIGURE 12.12 Protein and oil content distributions for developmental soybean cultivars with high protein content. (A) Protein content distribution for the F3 Wms 82(3) HHP soybean lines of selected pedigrees. (B) Oil content distribution for the F3 Wms 82(3) HHP soybean lines of selected pedigrees.

Novel soybean lines developed at UIUC [36,45] have the genetic potential for improving seed yield of US soybean cultivars, as well as concurrently increase signiﬁcantly the protein content, well above the average values of all commercial soybeans marketed in the United States. Single soybean seed NIR reﬂectance spectra (400 to 1700 nm) were collected on a diode-array reﬂectance instrument (DA7000 model) for composition analysis. The experimental design and

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 265 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

265

Protein distribution of F3 Wms82(3) ⫻ L69–183 130 120 110

Average: 45.2%

100 # Samples

90 80 70 60 50 40 30 20 10 0 <42.5 42.5– 43.0– 43.5– 44.0– 44.5– 45.0– 45.5– 46.0– 46.5– 47.0– 47.5– >48.0 43.0 43.5 44.0 44.5 45.0 45.5 46.0 46.5 47.0 47.5 48.0 (A)

Protein (%) Oil distribution of F3 Wms82(3) ⫻ L69–183

100 Average: 18.3% 90 80

# Samples

70 60 50 40 30 20 10 0 <16.0 16.0– 16.3– 16.6– 16.9– 17.2– 17.5– 17.8– 18.1– 18.4– 18.7– 19.0– 19.3– 19.6– >19.9 16.3 16.6 16.9 17.2 17.5 17.8 18.1 18.4 18.7 19.0 19.3 19.6 19.9 (B)

Oil (%)

FIGURE 12.13 (A) Protein content distribution for the F3 Wms 82(3) L69–183 soybean lines of selected pedigrees. (B) Oil content distribution for the F3 Wms 82(3) L69–183 soybean lines of selected pedigrees.

calibrations for single soybean seeds are here reported for the ﬁrst time. Single soybean seed calibrations based on the PLS-1 regression model were developed for determination of protein, oil, moisture, total sugars, and ﬁber content in a single soybean seed with the PLSplus=IQ software for both DA and high-resolution FT-NIR spectrometers. Light scattering variations were effectively corrected by the MSC algorithms and the baseline variations were effectively removed either by means of the ﬁrst-order derivative or using a spline-function, baseline correction [69]. Our single soybean seed results are then compared with large size, multiple soybean sample calibrations. NIR

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 266 14.7.2008 10:33am Compositor Name: BMani

266

Handbook of Food Analysis Instruments

reference values were measured on both large sample sizes and single soybean seeds by HR-NMR methods that can be used to accurately measure soybean composition; this established the use of HR-NMR as the preferred reference method for single-seed NIRS calibrations. The successful single-seed measurements by high-sensitivity=high-resolution FT-NIR have encouraged us to investigate other biotechnology applications such as the detection of minor ingredients, and nutraceuticals, for example, isoﬂavones [68,69] that are also found usually at much less than the 0.5% level in soybean seeds. Our research group recently extended both the FT-NIR and high-resolution, high-sensitivity FT-NIR chemical=hyperspectral imaging to studies of soybean embryogenic cells in culture in parallel with HR-NMR quantitative calibrations for oil in tiny (~200 mm) soybean embryos [12,25]; thus, nanoliter levels of soybean oil could be measured with acceptable signal-to-noise by 500 MHz 1H-NMR [12,25,38]. On the other hand, the FT-NIR chemical imaging based on the SpectrumOne–NTS, equipped also with an additional highsensitivity Sb detector for in vivo microspectroscopy, has surprisingly extended the sensitivity of such FT-NIR microspectrometers to the 10 pg level. Therefore, our research group at UIUC was indeed encouraged to investigate other FT-NIR microspectroscopy applications to pharmaceuticals, pharmacogenomics, medical biotechnology, and medical diagnosis that require exquisite resolution, sensitivity, and reproducibility as in the case of FT-NIR microspectroscopic detection and early diagnosis [13–16] of single cancer cells in human subjects, which is potentially very important for pharmacogenomics [16]. By comparison, earlier 1H (N) MRI imaging methods allowed only the diagnostic identiﬁcation of 1 mm malignant tumors in laboratory rats (Richard Magin, I.C. Baianu, and Paul Lauterbur [former Nobel laureate], 1992, unpublished results). Even at high magnetic ﬁelds, the resolution of 1H (N) MRI microscopy seems to be limited to ~10 mm (private communication from P. Mansﬁeld), which is an order of magnitude worse than the spatial resolution recently obtained by FT-NIR chemical imaging microspectroscopy in our laboratory under highresolution (1 cm1), hyperspectral conditions [12,13].

ACKNOWLEDGMENTS The research support by grants to ICB from Renessen Co. and C-FAR for the work reported here is gratefully acknowledged. The authors thank Perten Instruments, Inc. for the extended loan of their DA-7000 NIRS instrument, which made possible our NIR bulk calibrations for colored coat soybeans. The apt demonstrations and installation of two SpectrumOne—NTS, FT-NIR and midIR, spectrometers by Dr. Steve Bouffard, representing PerkinElmer in the Midwest, are also gratefully acknowledged. Last but not least, the author acknowledges the close and fruitful collaboration with Professor Randall L. Nelson of the Crop Science Department and the National Soybean Research Center at UIUC, who worked in parallel with the authors on numerous soybean accessions and new soybean lines that Professor Nelson developed through cross-breeding experiments for soybean quality improvements; his expert advice on advanced soybean genetics and crop science matters is also gratefully acknowledged.

REFERENCES 1. Baianu, I.C. and Forster, H. 1980. Cross-Polarization, high-ﬁeld Carbon-13 NMR techniques for studying physicochemical properties of wheat grains, ﬂour, starch, gluten and wheat protein powders. J. Appl. Biochem., 2:347–354. 2. Baianu, I.C. 1981. Carbon-13 and Proton NMR studies of wheat proteins. Spectral assignments for Flanders Gliadins in solution. J. Sci. Food Agric., 32:309–311. 3. Baianu, I.C., Johnson, L.F., and Waddell, D.K. 1982. High-resolution proton, Carbon-13 and Nitrogen-15 NMR studies of wheat proteins at high magnetic ﬁelds: spectral assignments, changes with concentration and heating treatments of Flinor Gliadins in solution. J. Sci. Food Agric., 33:373–383.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 267 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

267

4. Baianu, I.C. 1983. High-Field, High-Resolution NMR Studies of Cereal Proteins and Foods. 186th ACS National Meeting, Agric. Food Chemistry Div., Symposium on Applications of Solid-State and New Solution NMR Spectroscopy Methods to Agricultural Problems, Washington, DC, August 28 to September 2, 1983. 5. Baianu, I.C., et al. 1990. Multinuclear spin relaxation and high-resolution nuclear magnetic resonance studies of food proteins, agriculturally important materials and related systems. Basic Life Sci., 56:361–389. (Review) 6. Baianu, I.C. 1993. Solid-State NMR Analysis of Soybeans and Foods. Value Added Soybean Summit Meeting, Washington, DC. 7. Baianu, I.C. and Kumosinski, T.F. 1992. Physical Chemistry of Food Processes. 2nd edn. Vol. 1, Van Nostrand-Reinhold: New York. 8. Baianu, I.C., Pessen, H., and Kumosinski, T.F. Eds. 1994. Physical Chemistry of Food Processes: Principles, Techniques and Applications. Vol. 2, Chapman and Hall: London, New York, and Melbourne. 9. Baianu, I.C. 1994. Thermodynamic linkage of speciﬁc ion binding and hydration properties of soy glycinins and conglycinins determined by NMR techniques. Biophys. J., 66:246. 10. Baianu, I.C., You, T., Guo, J., and Nelson, R.L. 2002. Calibration of Dual Diode Array and Fourier Transform NIR Spectrometers for Composition Analysis of Single Soybean Seeds in Genetic Selection, Cross-Breeding Experiments., Proceedings of the 9th Biennial Conference of the Cellular and Molecular Biology of the Soybean, August 11–14, p.508. Soy2002 Platform presentation. University of Illinois at Urbana-Champaign: Urbana, IL, USA. 11. Baianu, I.C., You, T., Costescu, D.M., Prisecaru, V., Lozano, P., and Nelson, R.L. 2004a. High-resolution NMR and Near Infrared Determination of Soybean Oil, Protein and Amino Acid Residues in Soybean Seeds. In Analysis of Oil Seeds. D. Luthria, Ed., AOCS Publs.: Champaign, IL, pp.193–240. 12. Baianu, I.C., Costescu, D.M., You., T., Prisecaru,V., Lozano, P., Hofmann, N.E., and Korban, S.S. 2004b. Near Infrared Microspectroscopy, Fluorescence Microspectroscopy, Infrared Chemical Imaging and Highresolution NMR Analysis of Soybean Seeds, Somatic Embryos and Single Cancer Cells. In Analysis of Oil Seeds. D. Luthria, Ed., AOCS Publs.: Champaign, IL, pp. 241–273. 13. Baianu, I.C., Korban, S.S., Costescu, D., You, T., Lozano, P., and Hofmann, N.E. 2004c. Fourier Transform Near Infrared Microspectroscopy, Infrared Chemical Imaging, High-Resolution Nuclear Magnetic Resonance and Fluorescence Microspectroscopy, Detection of Single Cancer Cells and Single Viral Particles [Single Cancer Cells from human tumors are being detected and imaged by Fourier Transform Infrared (FT-IR), Fourier Transform Near Infrared (FT-NIR) Hyperspectral Imaging and Fluorescence Correlation Microspectroscopy]. CERN Preprint EXT-2004–069; Urbana, IL. 61801, USA: ICB, 01 May 2004, 21 pp. 14. Baianu, I.C., Prisecaru, V.I., Lozano, P., and Lin, H.C. 2005a. Novel Techniques and Their Wide Applications to Health Foods, Medical and Agricultural Biotechnology in Relation to Policy Making on Genetically Modiﬁed Crops and Foods. CERN preprint EXT-2004–066; Urbana, IL: ICB, 01 Jul. 2004 and Dec. 2005., 38 pp; [Selected applications of novel techniques in Agricultural Biotechnology, Health Food formulations and Medical Biotechnology are being reviewed with the aim of unraveling future developments and policy changes that are likely to open new markets for Biotechnology]. 15. Baianu, I.C., Lozano, P.R., Prisecaru, V., and Lin, H.C. 2005b. Applications of novel techniques to health foods, medical and agricultural biotechnology. Quant. Biol. arXiv, http:==arxiv.org=abs=q-bio=0406047. 2007-Update. 16. Baianu, I.C. 2007. Translational oncogenomics, human interactomics and pharmacogenomics. Invited review (in press), Intl. J. Translational Oncogenomics., 70: xl–x72. 17. Baianu, I.C., Cho, S.I., Stroshine, R.L., and Kuntz, G.W. 1993. Non-destructive sugar content measurements of intact fruit using spin-spin relaxation time (T2) measurements by pulsed 1H magnetic resonance. Trans. ASAE., 36(4):1217–1221. 18. Baianu, I.C., Guo, J., You, T., and Nelson, R.L. 2007. Moisture content calibration development for reﬂectance near infrared spectroscopy and analytical surveys of soybean composition, in submission. J. Sci. Food. Agric. 19. Ben-Gera, I. and Norris, K.H. 1968. Determination of moisture content in soybeans by Direct spectrophotometry. Isr. J. Agric. Res., 18:125–132. 20. Brim, C.A. and Burton, J.W. 1979. Recurrent selection in soybean II. Selection for increased percent protein in seeds. Crop Sci., 19:494–498.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 268 14.7.2008 10:33am Compositor Name: BMani

268

Handbook of Food Analysis Instruments

21. Burton, J.W. 1985. Breeding Soybeans for Improved Protein Quantity and Quality. In World Soybean Research Conference Proceedings. R. Shibles, Ed., WestView Press: Boulder, CO, pp. 361–367. 22. Chang, S.K.C. 1994. Chapter 14. In Introduction to the Chemical Analysis of Foods. S.S. Neilsen, Ed., Jones and Bartlett: Boston, MA. 23. Cho, R.K., Iwamoto, M., and Saito, K. 1987. Determination of 7S and 11S globulins in ground whole soybeans by near infrared reﬂectance spectroscopic analysis, Nippon Shokuhin Kogyo Gakkaishi, 34, 666–672. 24. Coles, B.A. 1980. Protein determination by nuclear magnetic resonance. J. Am. Oil Chem. Soc., 57:202–204. 25. Costescu, D.M., Baianu, I.C., and You, T. 2002. Novel Techniques for FT-NIR Microspectroscopy and Chemical Imaging Analysis of Soybean Seeds and Embryos. The 9th Biennial Conference of the Cellular and Molecular Biology of the Soybean, Aug 11–14 (Abstr.). University of Illinois at Urbana-Champaign: Urbana, IL, USA. 26. Delwiche, S.R. 1995. Single wheat kernel analysis by near-infrared transmittance: Protein content. Cereal Chem., 72:11–16. 27. Diers, B.W., Keim, P., Fehr, W.R., and Shoemaker, R.C. 1992. RFLP analysis of soybean seed protein and oil content. Theor. Appl. Gen., 83:608–612. 28. Fraenkel-Conrat, H. and Cooper, M. 1944. J. Biol. Chem., 154:239–246. 29. Galactic Industries Corporation, 1996. GRAMS=32 User’s Guide: PLS plus=IQ for GRAMS=32 and GRAMS=386, chapters 1–4. 30. Gerhardt, B. and Beevers, H. 1968. Inﬂuence of sucrose on protein determination by the Lowry procedure. Anal. Biochem., 24:337–338. 31. Hymowitz, T., et al. 1974. Estimation of protein and oil concentration in corn, soybean, and oat seed by near infrared light reﬂectance. Crop Sci., 14:713–715. 32. Guo, J., You, T., Baianu, I.C., and Nelson, R.L. 2007. Development of calibration techniques for reﬂectance near infrared spectroscopy measurements on black and brown soybean seeds–predicted composition comparisons (in submission). 33. Guo, J. and Ion Baianu, C. 2002. Determination of Soy and Other Health Foods Composition by Fourier Transform Near Infrared Reﬂectance Spectroscopy. In Proceedings of the 9th Biennial Conference of the Cellular and Molecular Biology of the Soybean., p.506. University of Illinois at Urbana-Champaign: Urbana, IL, USA. 34. Guo, J. and Baianu, I.C. 2002. Rapid Determinations of Soybean Isoﬂavones, Soy and Other Health Foods Composition by Fourier Transform Near Infrared Reﬂectance Spectroscopy. In Proceedings of the China and International Soy Conference and Exhibition 2002 (CISCE 2002), November 6–9, pp. 391–392. CISCE: Bejing. 35. Hinton, R.H., Burge, M.L.E., and Hartman, G.C. 1969. Sucrose Interference in the Assay of Enzymes and Protein. Anal. Biochem., 29:248–256. 36. Kabelka, E.A., Diers, B.W., Fehr, W.R., LeRoy, A.R., Baianu, I.C., You, T., Neece, D.J., and Nelson, R.L. 2004. Putative alleles for increased yield from soybean plant introductions., Crop Sci., 44:784–791. 37. Kakalis, L.T., Kumosinski, T.F., and Baianu, I.C. 1992. The effect of protein aggregation on the sorption of water vapor by proteins: A thermodynamic linkage study. J. Agric. Food Chem., 40:2063–2071. 38. Klempir, J. 1998. MS Thesis, Dept. of Nuclear Engineering, University of Illinois (advisor I.C. Baianu). 39. Lamb, D. and Hurburgh, C.R. 1991. Moisture determination in single soybean seeds by near-infrared transmittance. Trans. ASAE., 34:2123–2129. 40. Liu, K. 1997. Soybeans: Chemistry, Technology, and Utilization. Chapman & Hall: New York, pp.72–76. 41. Lo, C.H. and Stelson, H. 1972. Interference by polysucrose in protein determination by the lowry method. Anal. Biochem., 42:331–336. 42. Meredith, P., et al. 1980. Saccharides of developing wheat grain determined by Carbon-13 NMR spectroscopy. Staerke, (Starch) 32(4):198–205. 43. Messina, M.J. 1997. Soy foods: Their Role in Disease Prevention and Treatment. In Soybeans Chemistry, Technology, and Utilization. K. Liu, Ed., Chapman & Hall: New York. 44. Minson, D.J., Tuckett, P.G., and Law, D.P. 1985. A Comparison of Three Methods of Increasing the Precision of Regression Used for Estimating Nutritive Value of Forages by Near Infrared Reﬂectance. NIR84 Proceedings of 1st International Symposium Near Infrared Reﬂectance Spectroscopy, RACI, Cereal Chem. Div. Melbourne, Australia, pp. 138–148.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 269 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

269

45. Nelson, R., Baianu, I.C., and You, T. 2002. Genetic differences among sources of high protein in soybean seeds. Agronomy Abstr., ARS, Dec.20, http:==www.ars.usda.gov=research=publications=publications.htm? 46. Nielsen, S.S. 1994. Introduction to the Chemical Analysis of Foods. Jones and Bartlett Publishers: Boston, MA, pp. 139–159. 47. Ohnishi, S.T. and Barr, J.K. 1978. A simpliﬁed method of quantitating proteins using the biuret and phenol reagents. Anal. Biochem., 86:193–200. 48. Orman, B.A. and Schumann, R.A. 1992. Nondestructive single-kernel oil determination. J. Am. Oil Chem. Soc., 69:1036–1038. 49. Osborne, B.G., Fearn, T., and Hindle, P.H. 1993. Practical NIR Spectroscopy with Applications in Food and Beverage Analysis. Longman Group UK Ltd.: New York, pp. 99–117. 50. O’-Donnell, D.J., Ackerman, J.J.H., and Maciel, G.E. 1981. Comparative study of whole seed protein and starch content via cross polarisation-magic angle spinning carbon-13 nuclear magnetic resonance spectroscopy. J.A.F. Chem., 29(3):514–518. 51. Pazdernik, D.L., Arthur, S.K., and Orf, J.H. 1997. Analysis of amino and fatty acids composition in soybean seeds using near infrared reﬂectance spectroscopy. Agron. J., 89:679–685. 52. Pomeranz, Y. and Meloan, C.E. 1994. Food Analysis. 3rd ed., Chapman & Hall: New York. 53. Rutar, V. 1987. NMR Studies of Intact Seeds. Chapter 4. In NMR in Agriculture. P. Pfeffer, Ed., CRS Publishers: Boca Raton, FL, pp. 101–107. 54. Rutar, V. and Blinc, R. 1980. Nondestructive determination of protein content of viable seeds by proton enhanced Carbon-13 NMR. Zeitschrift fur Naturforschung Teil C Biochemie Biophysik Biologie Virologie, 35(1–2):12–15. 55. Schoﬁeld, J.D. and Baianu, I.C. 1982. High-resolution, solid-state CP-MAS Carbon-13 NMR studies of wheat gluten, glutenins and gliadins. Cereal Chem., 59(4):240–245. 56. Schuel, H. and Schuel, R. 1967. Automatic determination of protein in the presence of sucrose. Anal. Biochem., 20:86–93. 57. Shadow, W. 1998. Rapid Analysis for the Food Industry Using Near-Infrared Sepctroscopy. Perten Instruments, Inc., 1–17. 58. Silvela, L., Rodgers, R., Barrera, A., and Alexander, D.E. 1989. Theor. Appl. Genet., 78:298. 59. Smith, P.K. et al. 1985. Measurement of protein using bicinchoninic acid. Anal. Biochem., 150:76–85. 60. Tamate, J. and Bradbury, J.H. 1985. Determination of sugars in tropical root crops using 13C NMR spectroscopy: comparison with the HPLC method. J. Sci. Food Agric., 36:1291–1302. 61. Tiwari, P.N. and Burk, W. 1980. Seed oil determination by pulsed NMR without weighing and drying seeds. J. AOCS. 57(3):119–121. 62. USDA National Agriculture Statistical Service (NASS). 1998. USDA-NASS Agricultural Statistics 1998. http:==www.usda.gov=nass=pubs=agr98=acro98.htm. 63. Wang, D., Dowell, F.E., and Lacey, R.E. 1999. Cereal Chem., 76:30–33. 64. Wehling, R.L. 1994. Infrared spectroscopy. Chapter 24. In Introduction to the Chemical Analysis of Foods. S.S. Nielson, Ed., Jones and Bartlett Publishers: Boston, MA, pp. 342–351. 65. Williams, P. and Norris, K. 1987. Near-Infrared Technology in the Agricultural and Food Industries. AACC Inc.: St.Paul, MN. 66. Wright, R.G., Milward, R.C., and Coles, B.A. 1980. Rapid protein analysis by low-resolution pulsed NMR. Food Technol., 34(12):47–52. 67. You, T., Guo, J., Baianu, I.C., and Nelson, R.L. 2002. Rapid Determination of Protein, Oil, Moisture, and Isoﬂavone Contents of Single Soybean Seeds by Fourier Transform Near Infrared Reﬂectance Spectroscopy. In Proceedings of the China and International Soy Conference and Exhibition (CISCE 2002), November 6–9, pp. 414–415. CISCE: Bejing. 68. You, T., Guo, J., Baianu, I.C., and Nelson, R.L. 2002. Determination of Isoﬂavone Contents in Bulk for Selected Soybean Lines by Fourier Transform Near Infrared Reﬂectance Spectroscopy. In Proceedings of the 9th Biennial Conference of the Cellular and Molecular Biology of the Soybean, August 11–14, 2002, p.505. CISCE: Bejing. 69. You, T. 2005. PhD Thesis, University of Illinois, Urbana (PhD Advisor: I.C. Baianu).

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 270 14.7.2008 10:33am Compositor Name: BMani

Appendix A: Standard Reference Methods for Component Analysis 12.A.1 PROTEIN ANALYSIS METHODS Some of the most employed standard methods of analysis for total protein determination are summarized in the following subsections.

12.A.1.1 KJELDAHL METHOD This is one of the oldest methods for the determination of organic nitrogen in grains developed by the Danish investigator Kjeldahl in 1883. The basic process involves ﬁrst digestion of the sample in heated sulfuric acid to oxidize the carbon and hydrogen and the protein nitrogen is reduced and transformed into ammonium sulfate. Then concentrated sodium hydroxide is added into the digest and heated so that the liberated ammonia is driven off into a known volume of a standard acid solution. The unreacted acid is then determined by titration of alkaline. The total organic nitrogen from the sample can then be calculated and converted into a percentage of protein. Complete conversion of organic nitrogen into ammonia during digestion is the prerequisite for accurate protein determination. Digestion temperature has been found to affect the conversion speed and generally digestion temperatures of 3708C–4108C are considered to be the best. Besides temperature, the application of catalysts also signiﬁcantly affects the digestion speed. Mercury, copper, and selenium are the most widely used catalysts. Mercury is superior to copper but it has to be precipitated by sodium thiosulfate after digestion to decompose the mercury–ammonia complex formed during digestion. Selenium has an even more rapid effect than mercury, but the loss of nitrogen can readily occur if too much selenium is used or if the digestion temperature was not carefully controlled.

12.A.1.2 ULTRAVIOLET 280 NM ABSORPTION METHODS Most proteins exhibit a distinct ultraviolet absorption maximum at UV 280 nm, primarily due to tryptophan and tyrosine residues in the protein. The content of these amino acids in proteins from some sources varies within a reasonably narrow range; the absorbance at 280 nm could be used as a rapid and fairly sensitive test for protein concentration estimate. The advantage is that it is nondestructive; the samples can be used for other analyses after protein determination and there is no interference from ammonium sulfate and other buffer salts. However, the aromatic amino acid contents in various proteins differ considerably and the solution must be clear and colorless. Nucleic acids are also found to absorb at 280 nm.

12.A.1.3 BIURET METHOD The biuret method also has a long history. It was ﬁrst proposed by Riegler in 1914 based on the observation that substances containing two or more peptide bonds form a purple complex with

270

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 271 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

271

copper salts in alkaline solutions. It is simple, rapid, and inexpensive. It is believed that the biuret method of protein determination is more accurate than the Kjeldahl because the biuret procedure involves a reaction between the peptide linkage, while the Kjeldahl procedure measures the total nitrogen and does not distinguish between protein and nonprotein nitrogen. In practice, the biuret procedure involves two major steps. A 5 mL biuret reagent (the reagent includes Cu2SO4; NaOH; and potassium sodium tartrate, which is used to stabilize the cupric ion in the alkaline solution) is mixed with a 1 mL portion of protein solution (of 1–10 mg=mL concentration). After standing at room temperature for 15 or 30 min, the absorbance is read at 540 nm against a reagent blank (if the reaction mixture is not clear, ﬁltration or centrifugation before reading absorbance is required). A standard curve of concentration versus absorbance is obtained from measurements of UV absorption using bovine serum albumin (BSA). Major advantages of the biuret method over Kjeldahl are that the former does not have to assume nitrogen-ratio conversion factors based on a different primary method like the Kjeldahl, it is less expensive than the Kjeldahl, and also it is both rapid and simple. Furthermore, there are very few interferences from substances other than protein and the biuret method does not detect nonpeptide or nonprotein nitrogen like Kjeldahl. A disadvantage of the biuret procedure is that it is not an absolute method as the UV absorbance must be standardized against known protein or against other sensitive methods. High concentrations of ammonium salts do interfere with the biuret reaction and also absorbance may vary for very different proteins. Opalescence may also occur in the ﬁnal solution if high levels of lipid or carbohydrate are present.

12.A.1.4 LOWRY METHOD (PHENOL REAGENT) The Lowry method is one of the most widely used methods for the high-sensitivity determination of protein concentration in solutions. It combines the biuret reaction with the reduction of the Folin– Ciocalteu phenol reagent (phosphomolybdic–phosphotungstic acid) by the aromatic amino acids tyrosine and tryptophan residues of proteins. The bluish color developed at 750 nm can be detected spectrophotometrically even at very low protein concentrations, whereas the visible absorption at 500 nm can be utilized at high protein concentrations for protein determination. In practice, the Lowry procedure involves the following seven steps: 1. Proteins that are selected for determination are solubilized and then diluted to an appropriate range of concentrations (e.g., 20–100 mg=mL). 2. K-Na tartrate–Na2CO3 solution is added to the protein solutions after cooling and incubation of the latter at room temperature for 10 min. 3. CuSO4–K-Na tartrate–NaOH solution is added after cooling and incubated at room temperature for 10 min. 4. Freshly prepared Folin–Ciocalteu reagent is added, then mixed, and incubated at 508C for 10 min. 5. Visible absorbance is read on a digital spectrophotometer at 650 nm. 6. Standard curve is carefully determined with highly puriﬁed BSA standards of accurately known protein concentration. 7. Unknown protein concentrations are determined by comparison with the standard curve, preferably in the linear region of the curve. The Lowry method is highly valued and thus widely employed by biochemists because it is—in a certain sense—a primary method, and also because of its sensitivity as it is 10–20 times more sensitive than ultraviolet absorbance methods; thus, it is up to 100 times more sensitive than the biuret method. Moreover, the Lowry method is less affected by the turbidity of the sample than the biuret; however, the speciﬁc absorption may vary for different proteins to a greater extent than in the biuret method, and the absorption is not strictly proportional to protein

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 272 14.7.2008 10:33am Compositor Name: BMani

272

Handbook of Food Analysis Instruments

concentration throughout the entire concentration range. Moreover, the procedure is sensitive to high concentrations of sucrose and ammonium sulfate, and also suffers from interference from the presence of lipids, phosphate buffers, monosaccharides, and hexoamines. The sucrose interferes with the Lowry protein determination results in two ways. In the absence of protein, sucrose by itself in solution can lead to color development with the Lowry reagents. In the presence of protein, the sucrose can interfere with the color development by the protein in solution [30,35]. According to Gerhardt and Beevers [30], as the sucrose concentration in the sample increases—in the absence of protein—there is a linear increase in absorbance (at 540 nm). By adding sucrose to the protein solution to a concentration of 30% (w=w) sucrose, the absorbance from protein was found to be less than two-thirds of the true protein value. For the same sucrose concentration, the interference effects increase at the higher protein concentrations; sucrose reacts with the copper ions to form a stable complex, and thus makes its complexed copper unavailable to the protein. Therefore, increasing the concentration of copper ions in the Lowry reagent to compensate for the sucrose binding effect reduces the interference effect of sucrose [56].

12.A.1.5 MODIFIED BIURET–LOWRY METHODS: OHNISHI–BARR AND SIGMA CHEMICAL CO. METHODS As stated before, the biuret method has a closer to linear correlation between protein concentration and absorbance; the reagent is relatively stable but the method sensitivity to the lower protein concentrations is low. On the other hand, the Lowry method has very good sensitivity, but the stability of the combined reagents in time is relatively poor, and it is therefore necessary to prepare fresh reagent solutions for each series of protein measurements. Furthermore, as discussed above, the linearity of protein absorbance with protein content is not as good as in the biuret method, especially at the higher protein concentrations. Ohnishi and Barr made a modiﬁcation of the Lowry method by using biuret and phenol reagent in the Lowry procedure and thus combine the advantage of biuret method into the Lowry method [47]; it is currently the basis Sigma Chemical Co. (St. Louis, Missouri) micro-protein determination procedure No. 690 (see for example our results are shown in Figure 12.A.1). In the Ohnishi and Barr modiﬁcation, the classical biuret reagent is diluted eight times with a 2.3% Na2CO3 solution. The diluted biuret reagent has the following

1.2 1.0

Absorbance

0.8 0.6 0.4 0.9 mg/mL ADM SPI 1.0 mg/mL BSA

0.2 0.0 0

5

10

15

20

25

30

35

40

45

Time, min

FIGURE 12.A.1 Effect of protein type and dispersibility on the Folin–Ciocalteu reaction (2.0 mg=mL ADM SPI dissolved in NaOH and KCl pH 11.3). (From You, T., PhD Thesis, University of Illinois, Urbana, 2005, PhD Advisor: I.C. Baianu.)

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 273 14.7.2008 10:33am Compositor Name: BMani

273

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

composition: 0.018% CuSO45H2O, 0.075% tartrate, 0.375% NaOH, 2.01% Na2CO3, and 0.0125% KI, which is rather similar to that of the classical Lowry reagent (0.01% CuSO45H2O, 0.02% tartrate, 0.4% NaOH, 2.0% Na2CO3). The assay procedure involves two steps: ﬁrst mix the sample with diluted biuret reagent and let it stand for 10 min at room temperature, then add the phenol reagent, mix thoroughly, and measure the absorbance at a selected, ﬁxed wavelength between 550 and 750 nm against a blank after 30 min of standing. In standard tests with BSA, Ohnishi and Barr [47] have shown that the modiﬁed method provides a more stable color development than the classical Lowry method. At 258C, the absorbance reached a plateau 20 min after the addition of the phenol reagent, maintained this value for 2 h, and only then slowly decreased (after 6 h it decreased only by 1.7% from the plateau value). The temperature appeared to affect both the absorbance and the stability of color development. Although there was no noted difference between 208C and 258C, on reaching 308C the absorbance began to decrease about 40 min after the addition of the phenol reagent; the maximum absorbance at 308C was ~13% less than that measured at 208C–258C. The pH was also found to affect the stability of color development (Figure 12.A.2) and the optimum pH value reported was about 10.9. Strictly speaking, this modiﬁed method is still more like a Lowry method rather than a biuret method. Although some advantages of the biuret method are combined with the advantages of the Lowry method, the intrinsic nonlinearity of the classical Lowry method seems difﬁcult to eliminate; thus, the reported BSA standard curve with concentration between 0 and 600 mg=mL clearly shows a nonlinear dependence on protein concentration; besides this nonlinearity, the intrinsic interference from high sucrose concentration or other substances may also restrict its usage in special cases. 12.A.1.5.1

Detailed Sigma Chemical Co. Procedure

Instrument: Spectronic 21D from Spectronic instruments. Wavelength: 600 nm. (According to Sigma Chemical Co. [Sigma Diagnostics Procedure No. 690]) 1. Standard solution preparation; 2. Label two or more small test tubes: Blank, Test1, Test2;

1.0 0.9 0.8

Absorbance

0.7 0.6 0.5 0.4 PH7 Native PH7 Denatured PH12 Native PH12 Denatured

0.3 0.2 0.1 0.0 0

5

10

15

20 Time, min

25

30

35

40

FIGURE 12.A.2 Effect of pH and denaturation. (From You, T., PhD Thesis, University of Illinois, Urbana, 2005, PhD Advisor: I.C. Baianu.)

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 274 14.7.2008 10:33am Compositor Name: BMani

274

Handbook of Food Analysis Instruments

3. To Blank, add 0.2 mL sodium chloride solution; 4. To Test, add 0.2 mL test sample solution freshly prepared; 5. Add to each, 2.2 mL biuret reagent, Catalog No. 690–1. Mix well and allow to stand at room temperature for 10 min; 6. Add 0.1 mL of the Folin–Ciocalteu phenol reagent, Sigma Co. Catalog No. 690–2. Mix each tube well immediately after addition. Allow to stand at room temperature for 30 min; 7. Transfer contents of tubes to curvets and read absorbance using Blank as reference at the same wavelength and on the same instrument used to prepare calibration curve. Complete reading within 30 min; 8. Determine the protein concentration of the diluted test sample from calibration curve. Multiply by the dilution factor.

12.A.1.6 OTHER METHODS Besides the above listed methods, there are quite a few other methods reported for protein determination. For example, dye-binding methods are based on the fact that under speciﬁed conditions, proteins bind quantitatively with certain organic dyes including disulfonic anionic dye, orange G, etc. Dye-binding to protein can then be used to determine the total fractions of acidic and basic groups of proteins. These methods are rapid, inexpensive, and accurate because they do not measure=are not affected by nonprotein nitrogen content. Their disadvantage is the low sensitivity. Because proteins with different basic or acidic amino acid contents have different dye-binding capacities [28], the use of the Fraenkel-Conrat and Cooper method for total protein content is quite limited.

12.A.2 OIL DETERMINATION BY STANDARD REFERENCE METHODS 12.A.2.1 SOLVENT EXTRACTION METHODS Since oil stands for all the substances in a food which is soluble in organic solvent and insoluble in water, the total oil content in a food can be determined by organic solvent extraction. It is the most widely used oil determination method. On the basis of the extraction operation, the organic solvent extraction method can be further categorized as continuous solvent extraction methods, semicontinuous solvent extraction methods, or discontinuous solvent extraction methods. The Soxhlet method is the typical semicontinuous extraction method and is the most widely used for agricultural products such as grains and oilseeds. Most AOAC (Association of Ofﬁcial Analytical Chemists International) and AOCS (American Oil Chemists Society) standard methods for oil determination are based on the Soxhlet method. The AOCS ofﬁcial method Ac 3–44 is the standard oil determination method for soybean oil determination. It uses petroleum ether as the solvent to extract oil from ground soybean in a Butt-type extraction apparatus, which operates based on the same idea of Soxhlet extraction apparatus. The basic operation involves the following steps: Weigh 2 g of the ground sample into a ﬁlter paper and enclose in a second ﬁlter paper and the second paper is left open at the top like a thimble. Place the sample in the Butt tube and extract with petroleum ether for 5 h. Evaporate the petroleum ether on a steam bath or in a water bath. Weigh the mass of the oil extracted and the oil content can be calculated as mass of oil 100% mass of sample If the moisture in the sample is too high (>10%), the sample may need a drying pretreatment. The particle size of the ground soybean powder also affects the oil extraction: the ﬁner the grind, the more accurate and rapid the oil extraction is.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 275 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

275

12.A.2.2 OTHER OIL DETERMINATION METHODS Besides the solvent extraction methods, there are few other methods that can be used for oil determination. The Babcock method uses H2SO4 to digest protein and releases the fat and thus can be used in milk fat determination. NMR methods are already discussed in the main text.

12.A.3 MOISTURE DETERMINATIONS Moisture determination is also one of the more important and frequent analyses for a food product as component fractions are ﬁnally expressed on a dry basis.

12.A.3.1 OVEN DRYING METHODS The most widely used water determinations for foods, grains, and oilseeds are oven drying methods. The sample is heated under speciﬁed conditions and the loss of weight is used to calculate the moisture content of the sample. The drying condition including the type and conditions of the oven, the time, and temperature of drying signiﬁcantly affect the results. Whereas the boiling point of pure, liquid water is 1008C under ordinary atmospheric pressure, the bound water often has a higher boiling temperature, and therefore the oven drying temperature is often selected higher than 1008C for ordinary pressure oven drying to remove moisture almost completely from the samples. However, if the temperature is too high, decomposition of other food constituents may occur and affect the accuracy of the results. For example, carbohydrates and proteins may decompose if the temperature is too high. As an example, the ASAE standard method (ASAE S352.2) for soybean moisture determination requires 15 g of ungrounded soybean seeds being dried at 1038C for 72 h. The moisture is calculated as %Moisture ¼

wt H2 O in Sample 100% wt Wet Sample

Samples can also be dried under reduced pressure (25–100 mm Hg) to achieve a faster removal of water or to lower down the temperature required for drying.

12.A.3.2 KARL FISCHER TITRATION The Karl Fischer titration method is a chemical analysis method. It is quite rapid and sensitive and it can be used for determination of water in many low-moisture food products; it is based on several reactions involving the reduction of iodine by SO2 in the presence of water, but does not work very well for either corn or soybeans. For each mole of water, 1 mole of iodine, 1 mole of SO2, 3 moles of pyridine, and 1 mole of methanol are reacted. In practice, both iodine and SO2 are added to the sample in a closed chamber, which is isolated from the surrounding atmosphere. The iodine in excess that has not reacted with the water can be determined visually.

12.A.4 TOTAL SUGAR DETERMINATION Sugar is a general term used for both monosaccharides and oligosaccharides. Most determination methods for sugars are performed in solution. Therefore, the determination of monosaccharides and oligosaccharides in solid samples involves the extraction of sugars and thus their solubilization.

12.A.4.1 EXTRACTION

OF

MONOSACCHARIDES

AND

OLIGOSACCHARIDES

The standard AOAC method for the simultaneous extraction of monosaccharides and oligosaccharides from foods or grains—without polysaccharides—(by the AOAC methods 922.02, 925.05) is

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 276 14.7.2008 10:33am Compositor Name: BMani

276

Handbook of Food Analysis Instruments

accomplished by utilizing 80% alcohol. The alcohol extraction requires pretreatment with CaCO3 to neutralize acids, and then mixing with water so that the ﬁnal alcohol concentration is actually 80%. The food or grain needs to be ﬁrst ﬁnely ground or chopped, and only later mixed with an 80% alcohol solution. The liquid mixture is then heated near to the boiling point on a hot plate or steam bath for 30 min. The resulting alcohol solution is then passed through ﬁlter paper on an extraction thimble with retention of the ﬁltrate. The insoluble residue is then extracted again for 1 h with 80% alcohol. If the second ﬁltrate is highly colored, then the extraction is repeated. Finally, the solid residue is dried and ground so that all particles pass through a 1 mm sieve. This material is transferred to an extraction thimble and extracted for 12 h in a Soxhlet apparatus with an 80% alcohol solution. All alcohol ﬁltrates are then combined, and ﬁnally diluted to a known volume with 80% alcohol for the total sugar determination. If the starch determination is not required or involved, the extraction can be further simpliﬁed. Samples to be analyzed are mixed with 80% alcohol and then boiled on a steam bath or hot plate for 1 h. Then the solution is decanted into a volumetric ﬂask, the solid residue is mixed with 80% alcohol in a high-speed blender and the blended material is boiled for 30 min. The resulting mixture is then transferred to a volumetric ﬂask, diluted, and ﬁltered. During alcohol extraction, not only carbohydrates but also some other substances such as lipids, pigments, free amino and organic acids may also be extracted. For accurate total sugar determination, such compounds need to be removed by treatment with either lead acetate or ion-exchange resins.

12.A.4.2 HIGH PERFORMANCE LIQUID CHROMATOGRAPHY High performance liquid chromatography (HPLC) is considered one of the most reproducible, accurate, and sensitive methods for carbohydrates in food currently. Most standard carbohydrate determination methods are based on HPLC. There are three major reasons for the high popularity of HPLC in carbohydrate analysis: shorter analysis time compared with conventional liquid chromatography techniques; high column efﬁciency; and wide applicability to separating monosaccharides, oligosaccharides, as well as polysaccharides.

12.A.4.3 POLARIMETRY Most carbohydrate molecules contain one or more asymmetric carbon atoms and thus have the ability to rotate the plane of polarization of polarized light. Polarimetry measures the rotary power exerted by a compound in solution on the polarization plane of incident light. The polarimeter usually consists of a monochromatic light source, a polarizer for converting the light into planepolarized light, a sample tube of known length, and a detector for measuring the extent of rotation of the plane polarized light. The optical activity of sugars depends on the light wavelength, temperature, sugar concentration, and also the amount of sample. In practice, the wavelength is ﬁxed at 589.3 nm (the sodium D line). The tube=cell length is usually 10 cm and the temperature is normal (208C). The observed angular rotation of the light polarization plane is then proportional to the concentration based on the following equation: a ¼ a20 D l

c 100

where a is the observed angular rotation of light polarization plane l is the length of the sample tube c is the concentration of solution in g=100 mL a20 D is the speciﬁc rotation at the sodium D line at 208C

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 277 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

277

Different sugars generally have different speciﬁc rotations. Thus, the sucrose speciﬁc rotation is 66.462, whereas the one for glucose is 52.50. The polarimetric method requires that the sugar conformers are at equilibrium; it usually takes several hours to establish such a conformer equilibrium. The method usually works only for single sugar solution except in the case of sucrose, where one takes a reading at the beginning and then a second reading after sucrose hydrolysis by acid. The difference between these two readings is then used to calculate the percentage of sucrose: %sucrose ¼

(ab )D (aa )D Q

where ab and aa denote the plane rotation before and after inversion Q is a constant of 88.658 (66.58 to 22.158) This method may, however, fail if other sugars are also hydrolyzed in the process.

12.A.4.4 REFRACTIVE INDEX MEASUREMENTS When electromagnetic radiation passes from one medium into another, it can change direction, thus exhibiting the well-known phenomenon of refraction. The refraction index (RI) is deﬁned as the ratio of the sine of the incidence angle to the sine of the refraction angle of light. RI depends on concentration, temperature, and wavelength of the light. The standard RI value is measured at 208C using monochromatic sodium light (at ~589 nm). RI is performed with a refractometer that has been calibrated, for example, with sucrose; the results obtained are expressed as percent sugar wt=wt.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 278 14.7.2008 10:33am Compositor Name: BMani

Appendix B: A Literature Review of Soybean Chemistry 12.B.1 SOYBEAN CHEMISTRY Soybean originated in China about 4000–5000 years ago and belongs to the family Leguminosae, subfamily Papilionoidae, and genus Glycine [41,69]. Unlike most mature seeds made of three basic parts, soybean seeds are made of only two basic parts: the seed coat and the embryo. It is essentially devoid of endosperm. The seed coat is marked with a hilum. The large, well-developed embryo contains four major parts: the cotyledons, which function as food reserve structures, the radicle, the hypocotyl, and the epicotyl. In general, the seed coat accounts for about 8% of the total seed mass, while the cotyledons account for about 90%, and the hypocotyl accounts for 2% of the total seed mass. On average, the soybean seeds are composed of 40% protein, 20% oil, 35% carbohydrates, and 5% ash, as well as some minor substances such as vitamins. The actual composition varies signiﬁcantly and it depends on many factors including varieties, growing season, geographic location, and environmental stress. For example, among the lines in the US germplasm collection, the protein varies from about 30% to over 50% and the oil varies from about 12% to almost 30%.

12.B.1.1 PROTEIN The fraction and classiﬁcation of soy protein are rather complex. On the basis of biological function, it can be categorized as either metabolic protein or storage protein and the majority of soy protein is storage protein. On the basis of solubility, it can be divided into the water-soluble albumin and the salt-soluble globulin. Most soy protein is globulin and the globulins can be further divided into two distinct types: the glycinin and the conglycinin. Compared with conglycinin, the glycinin has larger molecular size, less solubility in salt solutions, and higher thermal stability. The problem of classiﬁcation based on solubility is that it is possible that the same polypeptides might be extracted into more than one solubility class due to their association with other proteins. Classiﬁcation based on approximate sedimentation coefﬁcients using ultracentrifugation is more precise. Under appropriate buffer conditions (usually at pH 7.6, 0.5 M ionic-strength buffer), soy protein exhibits four fractions after ultracentrifugation, designated as 2, 7, 11, 15S, respectively, where S stands for the Svedberg unit, computed as the rate of sedimentation per unit ﬁeld of centrifugal strength. 11S and 15S fractions are puriﬁed proteins. 11S is thought to be the same as glycinin and accounts for at least one-third of the extractable protein. 15S fraction is thought to be a polymer of glycinin and accounts for about 10% of extractable protein. 2S and 7S are considered to be heterogeneous. The 2S fraction accounts for about 20% of the extractable protein and the 7S fraction accounts for an additional onethird of the extractable protein and consists of conglycinin and some enzymes. 12.B.1.1.1

Isolated Soy Protein

Most commercial processes used today for soy protein isolates production generally follow the traditional process developed in the 1950s. First, oil is removed from the dehulled and ﬂaked beans 278

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 279 14.7.2008 10:33am Compositor Name: BMani

High-Resolution Near-Infrared and Nuclear Magnetic Resonance Analysis

279

by hexane extraction. Next, the soluble proteins and carbohydrates are extracted from the defatted ﬂour with water or dilute alkali (pH 7–10). The insoluble residue is removed by centrifugation. The major protein fractions are then precipitated at pH 4.5, collected by centrifugation, and washed to remove residual soluble carbohydrates, salts, and pH 4.5 soluble (whey) proteins. Typical chemical components of isolated proteins on a dry basis are at least 90% (N 6.25) protein, 4.5% ash, 0.5% ether-extractable fat, and 0.3% carbohydrates. Approximate amounts of each protein fraction in the aqueous extract are 22% 2S, 37% 7S, 31% 11S, and 11% 15S [41,69].

12.B.1.2 LIPIDS The soybean is still categorized as an oilseed instead of a protein seed, although its protein content is twice as high as oil. The main reason is that the oil has been fully and wider utilized long before soybean proteins were utilized in foods. Lipids are extracted from soybean by organic solvents such as hexane and the product is classiﬁed as crude oil. Major components of crude oil are triglycerides. Minor components include phospholipids, unsaponiﬁable material, free fatty acids, and trace metals. Characteristic of soybean oil is that it contains about 53% linoleic acid and 8% linolenic acid and is an excellent source of essential fatty acids.

12.B.1.3 CARBOHYDRATES Carbohydrates are the second largest component in soybeans, but the economical value of soy carbohydrates is considered much less important than soy protein and oil. It is believed that mature soybean seeds contain only trace amount of monosaccharides such as glucose, arabinose, while the di- and oligosaccharides may account for about up to 10%. Sucrose varies in the range of 2.5%–8.2%; rafﬁnose, 0.1%–0.9%; and stachyose, 1.4%–4.1%. Those di- and oligosaccharides are nonreducing sugars and are mostly soluble in water. Rafﬁnose and stachyose have received more attention mainly because their presence has been linked to ﬂatulence and abdominal discomfort associated with human consumption of soybeans and soy products. The insoluble carbohydrates in soybeans include cellulose, hemicellulose, pectin, and trace amounts of starch. The seed coat contains about 86% complex carbohydrates, among which 30% are believed to be pectins, 50% hemicellulose, and 20% cellulose. These complex carbohydrates are also referred to as diet ﬁber.

12.B.1.4 MINOR COMPONENTS

OF

SOYBEANS

Minerals, vitamins, phytates, and isoﬂavones are the minor components of soybeans.

Otles/Handbook of Food Analysis Instruments 45660_C012 Final Proof page 280 14.7.2008 10:33am Compositor Name: BMani

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 281 17.7.2008 5:38pm Compositor Name: VAmoudavally

Magnetic Resonance 13 Nuclear Spectroscopy in Food Analysis Francesco Capozzi and Mauro A. Cremonini CONTENTS 13.1

Theoretical Origin of Magnetization, Its Perturbation by Radiofrequency, and Recovery of the Equilibrium State ............................................................................. 282 13.2 NMR Experiment on the Simple Molecules of Water: FID and Its Fourier Transformation ................................................................................................................... 283 13.3 Theory of Relaxation and Differences between Spin–Spin and Spin–Lattice Interactions ......................................................................................................................... 284 13.4 NMR Instrumentations ...................................................................................................... 285 13.5 Low Resolution 1H-NMR ................................................................................................. 287 13.5.1 Measurement of the Longitudinal Relaxation Time T1 ....................................... 287 13.5.1.1 Pulse Sequence .................................................................................... 287 13.5.1.2 Applications in Food Science ............................................................. 290 13.5.2 Measurement of the Transverse Relaxation Time T2 .......................................... 292 13.5.2.1 Pulse Sequence ................................................................................... 292 13.5.2.2 Applications in Food Science ............................................................. 292 13.6 High-Resolution NMR ...................................................................................................... 294 13.6.1 One-Dimensional 1H-NMR Spectroscopy ........................................................... 294 13.6.2 Solution-State NMR Spectroscopy ...................................................................... 295 13.6.2.1 Acquisition of the One-Dimensional 1H-NMR Spectrum .................. 295 13.6.2.2 Applications in Food Science ............................................................. 297 13.6.3 Heteronuclear Spectroscopy................................................................................. 300 13.6.3.1 Pulse Sequences and Instrumental Requirements ............................... 300 13.6.3.2 Applications in Food Science ............................................................. 300 13.6.4 Solid-State NMR Spectroscopy: CP-MAS and HR-MAS .................................. 302 13.6.4.1 Theory of the Cross Polarization and of the Magic Angle Spinning ........ 302 13.6.4.2 Applications in Food Science ............................................................. 303 13.6.5 Two-Dimensional 1H-NMR Spectroscopy .......................................................... 304 13.6.5.1 Theory of the COSY, NOESY, TOCSY, and ROESY Spectra ......... 304 13.6.5.2 Applications in Food Science ............................................................. 306 13.6.6 Heteronuclear 2D Spectroscopy .......................................................................... 307 13.6.6.1 Pulse Sequence of HSQC, HMQC, HMBC, and INADEQUATE Spectra ................................................................................................ 307 13.6.6.2 Applications in Food Science ............................................................. 309 13.6.7 Pulse Field Gradient Nuclear Magnetic Resonance Spectroscopy ...................... 310 13.6.7.1 Pulse Sequence .................................................................................... 310 13.6.7.2 Applications in Food Science ............................................................. 311 References ..................................................................................................................................... 311

281

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 282 17.7.2008 5:38pm Compositor Name: VAmoudavally

282

Handbook of Food Analysis Instruments

This chapter deals with the food science applications of nuclear magnetic resonance (NMR), a technique very well described in several introductory and advanced books [1–5]. Although the authors do not give an in-depth treatment of the physical phenomenon of NMR, a simple theoretical description is offered at the beginning of each section to better understand the underlying principles and help choosing the right values for critical parameter when setting up an NMR experiment. This chapter has been organized by grouping the most useful NMR experiments, and the derived information that each different kind of NMR spectrometer is able to give, by type of instrument. In other words, in the following chapter the authors cover a number of NMR applications to food science from the instrumental point of view rather than by considering each foodstuff individually.

13.1 THEORETICAL ORIGIN OF MAGNETIZATION, ITS PERTURBATION BY RADIOFREQUENCY, AND RECOVERY OF THE EQUILIBRIUM STATE Nuclear magnetic resonance is a multifaceted technique whose specialties comprise liquid- and solid-state spectroscopy, imaging, and relaxometry. Although apparently different to the uninitiated, all NMR specialties share the same underlying phenomenon: absorption of energy may take place when a sample containing atoms that possess a not null nuclear magnetic moment is placed in a magnetic ﬁeld and irradiated by an electromagnetic wave of suitable frequency. Microscopically, the whole effect is made possible because certain nuclei possess angular moment or spin I (Table 13.1). If an external magnetic ﬁeld B0 is applied, magnetic moments align in 2I þ 1 quantum mechanically allowed orientations. In the case of I ¼ 1=2 (typical of 1H, 13C, 15N, 19F, and 31P, just to name a few common nuclei), magnetic moments can align in only two ways, corresponding (e.g., for hydrogen) to the parallel and antiparallel orientation with respect to B0. While in the absence of B0 all orientations have the same

TABLE 13.1 Magnetic Properties of Some Nuclei Nuclide 1

H H 12 C 13 C 14 N 15 N 16 O 17 O 19 F 23 Na 31 P 32 S 33 S 34 S 35 Cl 37 Cl 2

I

Natural Abundance (%)

Larmor Frequency (Relative to 1H)

Receptivity (Relative to 1H)

1=2 1 0 1=2 1 1=2 0 5=2 1=2 3=2 1=2 0 3=2 0 3=2 3=2

99.99 0.01 98.9 1.1 99.63 0.34 99.76 0.04 100.00 100.00 100.00 94.93 0.76 4.29 75.77 24.23

100.0 15.35 — 25.14 7.23 10.14 — 13.56 94.09 26.45 40.48 — 7.68 — 9.80 8.16

1.00 1.11 106 — 1.70 104 1.00 103 3.84 106 — 1.11 105 0.834 9.27 102 6.65 102 — 1.72 105 — 3.58 103 6.59 104

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 283 17.7.2008 5:38pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

283

energy, the presence of the magnetic ﬁeld removes the degeneracy and two energy levels arise, their energy difference being DE ¼ hg(1 s)B0 =2p

(13:1)

where h is the Planck constant g is the magnetogyric ratio typical of each nucleus s is the so-called shielding constant (vide infra) Alignment of the magnetic moments in an external magnetic ﬁeld does not imply that they are locked in an up or a down orientation with respect to B0 (Figure 13.1a). Each nucleus is actually in a superposition of up (a) and down (b) states and therefore may form any angle with the direction of B0. The tilted momenta precess around B0 with frequency v0 ¼ gB0 (called Larmor frequency) and random phase. Thus, a macroscopic magnetization (M0) arises in the sample along B0; no magnetization is present in the plane transverse to B0 because of phase randomness (Figure 13.1b). Application of a radiofrequency excitation pulse at the Larmor frequency v0 fulﬁlls the so-called resonance condition DE ¼ hv0=2p and causes the magnetic moments (hence the sample magnetization) to be rotated by a certain angle (Figure 13.1c). At the end of the pulse, two components of the excited nuclear magnetization exist, viz. a longitudinal (on Z-axis) and a transverse (on the XY-plane) magnetizations.

13.2 NMR EXPERIMENT ON THE SIMPLE MOLECULES OF WATER: FID AND ITS FOURIER TRANSFORMATION To better describe the result of the NMR experiment, the molecule of water is chosen as the simplest case of a substance, almost always present in foodstuffs, able to originate the NMR signal, the latter providing the information on the physical–chemical state of the molecule. After the excitation pulse (Figure 13.1c), all nuclear momenta associated to the water protons will be brought onto the XY-plane at the same time, where they precess and relax back to equilibrium, with their resonance frequency and with a time constant named spin–lattice, or longitudinal, relaxation time (T1). The receiver coil, whose major axis is placed on the XY-plane, is thus in condition to reveal a decaying oscillating magnetization usually referred to with the name of FID (free induction decay). The latter is described as a graph, in the time domain, reporting the intensity of the oscillating current, detected in the XY-plane, as a function of the acquisition time (Figure 13.1d). The signal disappears from the detection plane with an exponential decay having a time constant named spin–spin, or transverse, relaxation time (T2). For a better interpretation, the signal is easily transformed, by means of the mathematical procedure of Fourier transformation or FT, in a Lorentzian peak located in a frequency domain graph, centered at the resonating frequency for the observed nuclei, whose line width is proportional to T21 (Figure 13.1e). The NMR signal is described by three parameters, namely the relaxation times T1 and T2, and the resonance frequency. The three parameters are related to the magnetic properties of the originating nuclei and depend on the physical–chemical state of the molecules to which the nuclei belong. This chapter describes the information that each parameter is able to give on foodstuffs and the best experimental conditions to be adopted for their acquisition.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 284 17.7.2008 5:38pm Compositor Name: VAmoudavally

284

Handbook of Food Analysis Instruments

B0

B0 (a) z

z M0

x (b)

y

M0

z

z

z M

y x

M

y

y x

x

RC B0

(c)

100 80

0.5

Signal (AU)

Signal (AU)

1.0

0.0 –0.5

60 40 20 0

–1.0 0.0 (d)

y

x

B0

0.5

1.0 1.5 Time (s)

2.0

30 (e)

20 10 0 –10 –20 –30 Frequency (Hz)

FIGURE 13.1 Nuclear spin and sample magnetization. (a) Spin vectors with and without B0 ﬁeld applied; (b) sample magnetization as macroscopic resultant on the z-axis and on the xy-plane; (c) effect of the RF excitation pulse on the sample magnetization; (d) oscillating FID (free induction decay) relative to a single signal, e.g., water; (e) Lorentzian water signal obtained by Fourier transformation (FT) of its FID.

13.3 THEORY OF RELAXATION AND DIFFERENCES BETWEEN SPIN–SPIN AND SPIN–LATTICE INTERACTIONS T1 and T2 relaxation times describe the experimental behavior of an ensemble of nuclides during the NMR experiment. The necessity for a system to regain its original state after being subjected to a perturbation (a radiofrequency pulse) is of course due to the second law of thermodynamics, but the

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 285 17.7.2008 5:38pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

285

kinetic of the relaxation processes is governed by the physics of the sample. In case of spin 1=2 nuclides, two energy levels are present at thermal equilibrium, the population of the upper level being smaller than that of the lower level by the Boltzmann factor. A radiofrequency pulse perturbs the population of each level. For example, a 908 pulse equalizes spin populations on both levels, thus nulling the macroscopic magnetization on the direction parallel to the external magnetic ﬁeld. At the same time, phase coherence is achieved among spin precessions and ‘‘visible’’ magnetization appears on the XY-plane. Populations and coherences relax back to the initial state as soon as the perturbation ends, owing to random ﬂuctuation of magnetic ﬁelds at the sites of the nuclear spins caused by thermal motion of the molecules. Motion is best expressed in terms of correlation time (tc), i.e., the time interval in which it is reasonable to expect the magnetic ﬁeld at the nuclei to be constant during their reorientation. It is useful to consider the random magnetic ﬁeld ﬂuctuations as composed by a certain distribution of frequency components. When the frequency of a component matches the resonance conditions, it triggers some transitions between the levels per unit time, whose number depends on the component intensity. The more the characteristic motion, taking place in the system, matches the resonance frequency of the nucleus under study, the faster the T1 relaxation. At the same time, the random distribution of effective magnetic ﬁelds produces a random distribution of resonance frequencies throughout the sample, leading eventually to complete coherence dephasing by T2 relaxation. According to the Bloembergen–Purcell–Pound (BPP) model, the dependence of the two relaxation processes on the correlation time (tc) is different [7]. In particular, low molecular weight substances in the liquid state (very low tc) show an equal transverse and longitudinal relaxation times (the so-called extreme-narrowing limit), whereas in highly viscous liquids or with high molecular weight compounds (which tumble much slower in solution) T1 tends to be higher than T2 (see Figure 4 in Ref. [8]). With solids, where tc tends to inﬁnity, T1 may be even several seconds long, whereas T2 is of the order of microseconds. For example, protons in ice at 25 MHz relax with T1 ¼ 39 s and T2 8 ms. It is worth noting that although the T2 relaxation always starts after a coherent transverse magnetization (also known as single-quantum coherence) has been created by means of a radiofrequency pulse, T1 describes the onset of equilibrium after a perturbation of either the level population or the levels separation is induced on the system. Therefore, a sudden change of the external magnetic ﬁeld will be followed by a redistribution of the spin population over the new levels and this process will complete in a time 3–5 T1 at the new B0. The BPP model shows that the relaxation times also depend on the resonance frequency v0, which in turn depends on the magnetic ﬁeld. This dependence is at the basis of another NMR technique called ﬁeld cycling relaxometry, by which the T1 of a sample is measured at several magnetic ﬁelds and tc is found by ﬁtting the obtained data to the mathematical expression of the BPP model [9] or other empirical modiﬁcations of it [10].

13.4 NMR INSTRUMENTATIONS A typical NMR instrument is shown in Figure 13.2. It comprises four basic parts: (1) magnet, (2) radiofrequency unit, (3) control and signal recording unit, and (4) visualization unit. The magnet provides the magnetic ﬁeld, thus deﬁning the resonance frequency for each NMRactive nucleus and its sensitivity (the signal-to-noise ratio, S=N), the latter being proportional to ﬃﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ p B30 , g5 and the number of the nuclear spins being observed, and the square root of the total acquisition time. For this reason, a twofold diluted sample requires a fourfold experimental time to achieve the same S=N as its original solution. Other factors affecting the S=N are the probe ﬁlling factor (e.g., the fraction of the coil detection volume ﬁlled with sample), and various other probe and receiver parameters that are approximately equivalent for equipment built in the same period of time. For present instruments, 10 mg can be easily observed in the 13C spectra, although 1 mg is sufﬁcient (1 mmol or 5 mM). The requirements

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 286 17.7.2008 5:38pm Compositor Name: VAmoudavally

286

Handbook of Food Analysis Instruments

2

1 Magnet RF booster RF gate Sample λ/4 Preamp

RF generator

Ref 0⬚

RF receiver

Ref 90⬚

Audio receivers and filters A

3

FIGURE 13.2

Signal

Digital control data acquisition interfaces

B

4

Block scheme of a typical NMR instrument. (Courtesy of S. Sykora, Ebyte, Italy.)

are much less restrictive for 1H-NMR spectra, since 0.1 mg can be easily observed and 15 mg are often enough (30 nmol or 0.05 mM). When sample amount is limited, use of Shigemi tubes may help. The sample is dissolved in the smallest amount of solvent for increasing its concentration (down to 20 mL) and inserted into a small inner microtube sandwiched between two inserts made of special glass, matched to the magnetic susceptibility of the solvent. A further three- to fourfold gain in sensitivity can be obtained by cryogenically cooling (~20 K) the RF coils and the ﬁrst-stage receiver electronics, thus reducing thermal noise. Magnets are built differently depending on the NMR technique for which they are used. In spectroscopy, where spectral resolution matters, nowadays they are of the superconducting kind. They can reach homogeneity up to tens of hertz and stability up to 10 Hz=h. Although magnetic resonance imaging (MRI) places lesser constraints on the magnet, homogeneity of some tens or even a few hundreds of hertz is still required. Both in spectroscopy and imaging, a so-called shimming-unit placed within the bore of the magnet produces an additional magnetic ﬁeld that can be manually or automatically set so as to improve the magnetic ﬁeld homogeneity over the sample. For example, resolutions as high as 0.3 Hz, or less, are obtained with vertical magnets. In addition, an electronic device, known as lock, counteracts the small, but not negligible magnetic ﬁeld ﬂuctuations by slightly changing the operating frequency of the spectrometers so that the position of the lines on the spectrum do not change with time. Quite every type of magnet can be used for relaxometry and actually magnets of any kind have been reported in the literature along the years, from permanent magnets [11–15] to inside-out magnets [16] and Halbach magnets [17] and from electromagnets [9,18] to the Earth itself [19]. In contrast to high-ﬁeld NMR, no spectral information can be obtained in low-ﬁeld NMR unless special sequences are employed [20]. Rather, the behavior of the NMR response in the time domain is studied. For this reason, low-ﬁeld NMR is often also known as time domain NMR (TD-NMR).

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 287 17.7.2008 5:38pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

287

The information available from TD-NMR experiments include proton spin relaxation times and, with slightly more sophisticated approaches, also diffusion coefﬁcients. Most of the probeheads, i.e., the electronic devices deputed to detect the NMR signal, are designed for acquiring the signal of the whole sample. The detection coil is wound around a cylinder hosting the samples (or ﬂanks it by two sides, if saddle coils are used), and the whole device is completely inserted into the bore of the magnet that must be large enough to accommodate the whole sample in its interior. For this reason, traditionally, TD-NMR is carried out on small samples. Thus, it is not possible to apply this method in a nondestructive fashion to entire (packaged) foods. The sample size can be increased by using bigger diameter low-ﬁeld NMR setups, which are typically used for MRI. However, MRI machines are expensive, heavy, and sensitive pieces of equipment, which excludes routine application in most quality control or food veriﬁcation applications. To overcome such a limitation, other single-sided sensors (NMR-MOUSE) designed for surface acquisitions (see Figure 1 in Ref. [21]) have been used for applications with relatively slowdiffusing target materials (e.g., food materials with high-fat content). Yet, studies of materials with higher self-diffusion coefﬁcients are hampered by large magnetic ﬁeld gradients (about 10 T=m or even higher), which are typical for most single-sided NMR setups. By varying the distance between the two magnets (see Figure 1 in Ref. [21]), it is possible to obtain homogeneous ﬁeld regions with a gradient of about 1 T=m at distances of several millimeters from the magnets. A further development of surface magnets is represented by semi single-sided NMR setup, where the magnet system and the surface coil can excite a much bigger part of the sample volume than in the single-sided apparatus [21]. Block 2 of the NMR instruments (Figure 13.2) generates all the radiofrequencies that must be delivered (after ampliﬁcation) to the probe. Because of the high power emitted by the transmitter booster (several watts to tens of watts), the sensitive receiver circuitry must be blanked during pulses and reopened in the shortest possible time after transmitter (TX) power has dropped under an acceptable threshold. The signal received by the pick up coil is preampliﬁed, ﬁltered, and electronically mixed with a reference frequency corresponding to the center of the spectrum; this has the effect of producing two signals (the sum and the difference of the original signals) of which only the latter is retained. With a simpliﬁcation, one would say that the latter signal contains the signal frequencies observed from a rotating frame precessing on the XY-plane at the same frequency of the reference so that, e.g., a signal with the same resonance frequency than the reference appears as ﬁxed (except for the intensity loss owing to T2). The allowed signal (now an audio signal) is eventually digitalized by dedicated circuitry (block 3 in Figure 13.2) and presented to the user in the form of an FID.

13.5 LOW RESOLUTION 1H-NMR 13.5.1 MEASUREMENT OF 13.5.1.1

THE

LONGITUDINAL RELAXATION TIME T1

Pulse Sequence

The collection of radio wave pulses and time delays that are able to produce a certain effect over the macroscopic magnetization is called a pulse sequence. The longitudinal relaxation time can be measured by the inversion-recovery (IR, 1808–t–908) pulse sequence shown in Figure 13.3a. The ﬁrst pulse (a 1808 pulse) rotates the macroscopic magnetization from þZ to Z, where it starts relaxing back to þZ with the T1 time constant. After a variable delay t, the value of the residual magnetization on the Z-axis is transferred to the XY-plane by a 908 pulse, where it is detected and recorded. The experiment is repeated for several values of t, from very short values to about 5 T1 s (usually sampled in a logarithmic fashion), each time collecting the FID in a separate memory space of the computer. The intensity of the magnetization, in the IR experiment, increases exponentially with increasing t, by passing from I0, when t is null, to þI0, when t is inﬁnite (Figure 13.4a, ﬁlled circles).

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 288 17.7.2008 5:39pm Compositor Name: VAmoudavally

288

Handbook of Food Analysis Instruments

t

t

FID

n ∆ ∆ (a)

(b)

e

t

t

t

t

t

n (c)

(d)

P1 X

FID

FID 1

DEC

H

(e)

(f)

t1

t1

SL

FID (g)

FID

(h)

t1

13C

MIX

1

1

4JCC

4JCC

t1

FID

FID 1H

(i)

(j)

1H

(k)

13C

1H

FID t1 1

1

21JCH

21JCH

FID t1

DEC

13C

13C

1

1

(l)

21JCH

1H

(m)

DEC

22,3JCH

FID t1 1

1

41JCH 41JCH

1 41J

CH

1

DEC

41J

CH

FIGURE 13.3 Schematic representation of the pulse sequences employed in the NMR experiments described within the chapter. The abscissa represents the experimental time for a single scan. The symbols adopted in the ﬁgure are largely accepted in NMR spectroscopy: black and white rectangles represent 908 and 1808 pulses, respectively; empty yellow triangles stand for spin echo detection; the FID yellow triangle is the acquisition time; the DEC rectangles correspond to the selected nuclear decoupling time; t is a ﬁxed delay between pulses; D is the transversal refocusing time; X is the heteronuclear transmitter channel; t1 is the evolution time

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 289 17.7.2008 5:39pm Compositor Name: VAmoudavally

289

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

1.0 60 Signal Np–1(AU)

Signal (AU)

0.5

0.0

50 40 30 20

–0.5 10 0

–1.0 0 (a)

1

2 t (s)

3

4

10–4

5 (b)

10–3

10–2 T2 (s)

10–1

100

FIGURE 13.4 IR and SR proﬁles, expressed as signal integrals or ﬁrst point of FIDs: (a) lines represent the ﬁtting curves to Equation 13.2a (IR) or Equation 13.2b (SR), whereas circles represent the experimental data in the IR (ﬁlled) and SR (open) experiments; (b) T2 relaxogram in turkey meat.

A slight modiﬁcation of the IR sequence is represented by the saturation-recovery (SR) sequence sketched in Figure 13.3b, where a 1808 pulse, ﬂanked by two very short time delays, generates the transverse refocusing of the magnetization (spin echo, see Section 13.5.2.1). The latter may be required for measurements of longitudinal relaxation time in the presence of a very inhomogeneous magnetic ﬁeld such, e.g., that induced by the magnet of a portable NMRMOUSE device. The intensity of the magnetization, in the SR experiment, increases exponentially with increasing t, by passing from 0, when t is null, to þI0, when t is inﬁnite (Figure 13.4a, open circles). In high-resolution NMR, where each chemically inequivalent nucleus yields a spectral line after FT, the longitudinal relaxation times of the nuclei are obtained by ﬁtting each line integral (I) to Equation 13.2a (for IR) or to Equation 13.2b (for SR), as a function of the recovery time (t): I(t) ¼ I0 [1 2a exp (t=T1 )]

(13:2a)

I(t) ¼ I0 [1 a exp (t=T1 )]

(13:2b)

A three-parameter ﬁt yields I0: the intensity of the line at complete signal recovery, the longitudinal relaxation time T1, and a, a parameter that takes into account a possible slight deviation from perfect signal inversion. This method is clearly inapplicable in TD-NMR at low ﬁelds where all spectral lines collapse. T1 of TD-NMR FIDs is therefore obtained by ﬁtting the intensity of each FID (as represented by the value of the ﬁrst point of the FID) to Equation 13.3:

3 necessary to build up a 2D (two-dimensional) spectrum; SL is the spin lock time; MIX is the mixing time for magnetization transfer; J is the scalar coupling constant. (a) IR pulse sequence used for T1 measurements; (b) echo detected SR sequence used for measurements of T1 in highly inhomogeneous magnetic ﬁelds; (c) the Hahn spin echo pulse sequence adopted for the assessment of bulk T2 relaxation time; (d) the CPMG (Carr– Purcell–Meiboom–Gill) pulse sequence adopted for the measurement of transverse relaxation time (T2); (e) one-pulse sequence; (f) pulse sequence for 1D (one-dimensional) heteronuclear {1H}X-NMR spectroscopy; (g) COSY pulse sequence; (h) TOCSY (total correlation spectroscopy) pulse sequence; (i) NOESY pulse sequence; (j) INADEQUATE pulse sequence; (k) HMQC (heteronuclear multiple quantum coherence) pulse sequence; (l) HMBC (heteronuclear multiple bond coherence) pulse sequence; and (m) HSQC (heteronuclear single quantum coherence) pulse sequence.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 290 17.7.2008 5:39pm Compositor Name: VAmoudavally

290

Handbook of Food Analysis Instruments

I¼

X

I0i [1 2a exp (t=T1i )]

(13:3)

i

where the index i runs over all the chemically distinct species contained in the sample (e.g., water and fat, or extra- and intracellular water). However, cases exist where a model based on a discrete summation of components as in Equation 13.3 is not appropriate for the relaxation behavior supposed for the sample (e.g., when the sample is characterized by a distribution of compartments, a distribution of relaxation times may be obtained) and the summation in Equation 13.3 is extended to an inﬁnite number of components in what is called a continuous model. Although ﬁtting relaxation data to a continuous model is an ill-deﬁned mathematical procedure [22], specialized algorithms have been developed [23] that are able to perform the ﬁt and produce a relaxogram, i.e., a plot of I0i versus T1i. 13.5.1.2

Applications in Food Science

13.5.1.2.1 TD-NMR Both longitudinal and transverse relaxation times of protons in neat water are equal (~3 s) and are always equal or longer than those measured in aqueous solutions. The relaxation times, indeed, are decreased by fast chemical exchange with other protons experiencing mechanisms that shorten their 1 ) are additive quantities that relaxation times. The effect is quantiﬁable, since relaxation rates (T1,2 sum up all contributions deriving from different phenomena that contribute to the mechanisms of relaxation. The simplest case is one where only one solute is present, which affects the observable relaxation rate according to Equation 13.4: 1 þ T1i1 fi T11 ¼ (1 fi ) T1w

(13:4)

where 1 is the proton relaxation rate of pure water T1w T1i1 is the proton relaxation rate of exchanging solutes fi is the molar fraction of solute exchangeable protons The measurement of the relaxivity enhancement has been applied to determine the oxygen content in superoxygenated water contained in closed bottles, by using the NMR-MOUSE device and the echo-detected SR sequence [21]. The echo time (2t in Figure 13.3b) was set to 300–500 ms, for this experiment. The longitudinal relaxation rate, measured 2 mm below the bottle surface, is related to the oxygen concentration cox according to the following relationship: 1 þ rox cox T11 ¼ T1w

(13:5)

where T1w is the relaxation time of the oxygen-free water rox is the relaxivity of oxygen in water (0.48 mM1s1) [24] Equation 13.5 is equivalent to Equation 13.4, since the molar fraction of oxygen is much lower than that of water. Dissolved oxygen is not the only possible solute that may lead to increased longitudinal relaxation rates in beverages, but dissolved paramagnetic metal ions also lead to a reduction of T1 [25]. In drinking waters, concentrations of 1 mg=L of highly relaxing ions (e.g., Fe3þ or Mn2þ) or several milligrams per liter of moderately relaxing ions (e.g., Cu2þ) affect the proton relaxation times. However, such concentrations can be excluded both for health regulations and for taste reasons, except in some highly mineralized waters used for therapeutic purposes. This is different in pineapple

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 291 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

291

and blueberry juices, as well as in wines and some tea extracts, which are known to exhibit very short relaxation times, mainly because of manganese complexed by organic ligands [26–28]. The spin–lattice relaxation time (T1) has also been measured on hen eggs to evaluate the change of quality during its ﬁrst few days of storage [29]. The gradual degradation of the gel-like network formed by the ovomucin–lysozyme complex, accounting for the progressive decrease in viscoelasticity observed during the storage [30], was found to correlate with water proton T1 measured at low ﬁeld. Another explanation for the fast increase of the longitudinal relaxation rate of water proton in shell egg albumen was found in the concomitant increase of iron ions concentration in albumen, because of its migration from yolk, where it is more abundant [31]. Whatever is the cause of its change, this NMR parameter is related to the quality of egg, thus inspiring its industrial exploitation, once the reliability and the sensitivity of the technique have reached a satisfactory improvement. 13.5.1.2.2 Relaxometry A further advance in the NMR diagnostic power can be found in the choice of an optimal magnetic ﬁeld at which the relaxation rate is measured. In fact, the BPP model (see Section 13.3) accounts for a dependence of the nuclear relaxation rates on the magnetic ﬁeld strength. Also the presence of paramagnetic species affects, at a different extent, the water proton relaxation rates (see Figure 1 in Ref. [32]) depending on the nuclear Larmor frequency v0, according to the Solomon–Bloembergen– Morgan (SBM) theory [33,34]. The dispersion of the longitudinal relaxation rate (T11 ) observed with increasing strength of B0 is hence explained by taking into account both the diamagnetic (BPP) and paramagnetic (SBM) effects. Such a dispersion curve can be measured using special NMR instruments known as ﬁeldcycling relaxometers [35,36]. In the case of the egg albumen, it has been shown that the inﬂexion point in the proﬁle occurs at about 1 MHz, i.e., at a much lower frequency than the one usually found in benchtop TD-NMR spectrometers, and even more distant from the frequency range in which the main relaxation changes are detected in eggs upon aging (10 kHz–1 MHz). With this in mind, an external unit was developed for use with a commercial spectrometer that permits T1 measurements at 700 kHz [18]. This frequency was sufﬁciently low as to sense the relaxation changes in the samples and far enough from 1 MHz, where small magnetic ﬁeld instabilities may provoke a large variation in the measured T1. This instrument may be useful for a nondestructive analysis of hen shell eggs, because longitudinal relaxation time gives an indication of freshness [37]. 13.5.1.2.3 Protonless TD-NMR Other nuclei, besides water protons, can be exploited for food analysis by TD-NMR. 23Na-NMR has been widely used, e.g., in the determination of salt mobility in food or model systems [38,39] and in the determination of the perception of saltiness in salty gum solutions [40]. However, for lowsensitivity heteronuclei, a dedicated low-ﬁeld relaxometer may be simply too expensive and too rarely used to deserve purchase. Classical NMR spectrometers may be used for the purpose of measuring the heteronuclear relaxation times, although with some difﬁculties arising with quadrupolar nuclei (e.g., 23Na and 35Cl) that show very short T2 reﬂecting into broad spectral lines. In a T1 relaxation study conducted on canned crab meat, NMR spectra were obtained as a function of recovery time (t, see Equation 13.2a), which varied from 0.20 to 40 ms for 23Na and from 0.20 to 20 ms for 35Cl [41]. Sodium-23 results clearly showed that strong heating of crab meat (containing 300–600 mg Naþ and 450–750 mg Cl per 100 g) decreased T1, thus indicating decreased sodium mobility. Results were interpreted in terms of loss of free water and increase in water bound to the macromolecular matrix, which were thought to be the cause for a change in the meat texture. The autoclaving process greatly reduced the mobility of both sodium and chloride, an effect which has been associated to the unvalued taste of canned crab that the heating gives to the meat. T1 values of 35 Cl increased relevantly during storage, especially between days 2 and 10. As for 23Na, an increase in T1 indicated a greater mobility of the Cl ions, thus implying that the Cl ions underwent a transition from a bound to a more free chemical state. A decreased binding of the Cl ion resulted in saltiness increase, which was also comparable to the sensory experience.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 292 17.7.2008 5:39pm Compositor Name: VAmoudavally

292

Handbook of Food Analysis Instruments

13.5.2 MEASUREMENT OF 13.5.2.1

THE

TRANSVERSE RELAXATION TIME T2

Pulse Sequence

Transverse relaxation time is related to the dumping of the NMR signal on the xy-plane: the faster the signal decays to zero, the shorter T2. Signal dumping, however, is not just caused by spontaneous loss of coherence among excited spins. In fact, ﬁeld inhomogeneity induces a spread of resonance frequencies over the sample that sum up and produce a further dumping, thus distorting the signal. Although methods exist for disposing of the unwanted distortion from a recorded FID through the use of suitable reference samples [42], the most useful approach is that of recording an FID free from distortions through the CPMG (Carr–Purcell–Meiboom–Gill) sequence [43]. This is an extension of the Hahn spin-echo method (Figure 13.3c) by which spin precession taking place for a time t=2 in the xy-plane is reversed by a 1808 pulse so that phase coherence is achieved again after another t=2 s. Since signal intensity at point e in Figure 13.3c depends only on T2 and not on ﬁeld homogeneity, running a number of experiments with variable t results in a number of echoes whose intensity decays with T2. The value of T2 is found by ﬁtting the echo intensity to Equation 13.6: I(t) ¼ I0 exp (t=T2 )

(13:6)

where I(t) is the echo intensity at time t I0 is the FID intensity at the beginning of the FID The Hahn echo approach, however, has at least two drawbacks. First, only one echo is obtained from each experiment making the overall experiment time quite long. For example, if 10 equispaced points are needed for reconstructing the T2 decay of a sample having T1 of 0.5 s and T2 of about 0.2 s, and each experiment requires 16 transients, then at least 16 number of transients

[10 5 0:5 55 (5 0:2=10)] total time for T1 relaxation

total time between first pulse and echo

(13:7)

or 312 s are needed. Second, if the sample is liquid or semisolid self diffusion inﬂuences each echo differently, thus biasing the results [44]. Both problems are much alleviated with the CPMG sequence, where the signal is continuously refocused and echoes appear between each couple of refocusing 1808 pulses (Figure 13.3d). Here, several hundreds or even thousands of echoes are recorded in the same transient and experimental time is much shorter. For example, if t is set to 1 ms it is possible to sample the whole T2 relaxation decay of the sample described above, with 1000 echoes and 16 transients, in only 16 (5 0.5 1000 0.001) ¼ 24 s. Furthermore, as echo time is ﬁxed, self diffusion affects all echo intensities in the same way. Care, however, must be taken because the T2’s measured via CPMG may be affected by chemical exchange [45], although this does not appear to be a problem if the experiments are carried out with typical low-ﬁeld spectrometers working at 20 MHz or less. 13.5.2.2

Applications in Food Science

13.5.2.2.1 Water Holding Capacity in Meat and in Fresh Cheese Time domain nuclear magnetic resonance at low resolution has been proposed by some authors as an alternative technique to investigate water holding capacity (WHC) in pork [46–53] and poultry [54] meats. The measure of the transverse relaxation time on meat can give relevant information

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 293 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

293

about the water dynamics. Moreover, the multicomponent T2 relaxation behavior observed in meat highlights its water compartmentalization and, consequently, the WHC. Several papers have described the relaxation data of meat as the summation of a discrete number of exponential functions (the discrete model) corresponding to different types of water [12,46,55]. Brown et al. [56] applied the UPEN (uniform penalty) continuous ﬁtting algorithm [23] to the analysis of pork T2 signals and detected two broad populations of water that did not show any discontinuity in a T2 relaxogram (Figure 13.4b). Also, the performances of the discrete model [57], in the prediction of pork WHC, were compared with those of a different method of continuous ﬁtting and found the latter method to be more precise [58]. However, the shape of continuous T2 relaxograms depends strongly on the type of regularization algorithm used during the calculation, so that the same T2 relaxation decay may give rise to signiﬁcantly different relaxograms [59]. For this reason, an alternative approach for the prediction of WHC was derived by applying chemometrics directly on the raw time-domain NMR signal to avoid any bias in data processing [56,60]. A possible source of error in the acquisition of NMR data of meat is the sample preparation, as the orientation of the muscle ﬁbers in the magnetic ﬁeld may affect the results. In fact, because of diffusion anisotropy, water diffuses preferentially along a ﬁber than in a direction perpendicular to the ﬁber axis. For this reason, samples must be carefully prepared, by also making use of some homemade tools designed to prepare meat rods that ﬁt exactly into a 10 mm diameter NMR tube, without disruption of the meat structure (Figure 13.5). The T2 measurements were also employed in the characterization of WHC in fresh cheese samples, to ﬁnd the best technological treatment that avoids an undesirable phase separation of the serum caused by a contraction of the gel-matrix (syneresis). The dependence of T2, as a function of duration and temperature of the shearing treatment applied to fresh cheese, shows that the water proton T2 relaxation time reaches its minimum, i.e., the highest viscosity and the best water-holding capacity, after 10 min of shearing at 808C [8].

(a)

(b)

(c)

(d)

FIGURE 13.5 (See color insert following page 240.) Meat sample preparation for TD-NMR: the homemade tool permits to excise meat rods (a) without disrupting its inner ﬁbers; the tool ﬁts the 10 mm inner diameter NMR tube (b) and can extrude the rod directly in the bottom of the tube (c); the rod is about 10 mm high (d) and occupies the volume within the probe receiver coils.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 294 17.7.2008 5:39pm Compositor Name: VAmoudavally

294

Handbook of Food Analysis Instruments

13.5.2.2.2 Solid–Liquid Transition in Solutions The T2 relaxation was also applied to the study of molecular mobility in frozen sugar solutions, to ﬁnd relationships between the ice-recrystallization rates and the glass transition occurring at the Tg temperature [61]. Below this temperature, kinetically controlled changes (e.g., phase separation, crystallization, chemical and enzymatic reactions, texture modiﬁcations) become so slow that on a practical time scale they can be considered completely stopped. Thus, Tg is an important reference temperature affecting the quality of many foods and is the object of many thermodynamic studies [62–66]. It has been demonstrated that the relaxation time does not depend on the solute molecular weight for a given value of T – Tg (see Figure 3 in Ref. [61]). The number of components observed in the relaxation decay depends on the position of the sample on the state diagram. Two relaxation components are expected above the freezing point, one attributable to the protons of the solute and the other to the water protons. Below the freezing point, ice forms, and this introduces a third component into the NMR signal. However, the spin–lattice relaxation time of the ice phase is considerably longer than that of the other two components. Therefore, by using optimized recycle delays, in the so-called T1 weighted T2 measurements, the contribution from the ice phase can be suppressed so much that it does not signiﬁcantly contribute to the observed relaxation decays, whilst the other signals are unaffected. In this way, a strong correlation has been found between the NMR parameter and the recrystallization rate of sugar solutions, making the technique attractive for application in food technology [67].

13.6 HIGH-RESOLUTION NMR 13.6.1 ONE-DIMENSIONAL 1H-NMR SPECTROSCOPY The parameter s in Equation 13.1 is at the basis of the great success of NMR spectroscopy. In fact, atomic nuclei in molecules are surrounded by electrons that react to B0 and give rise to additional induced magnetic ﬁelds that may sum, or oppose, B0 depending on the chemical environment in which the nuclei are placed. Nuclei with the same g may thus sense a slightly different effective magnetic ﬁeld Beff ¼ (1 s) B0 if located in molecular sites where the chemical properties are different. Therefore, NMR-active atoms placed in different molecular locations attain resonance at a slightly different frequency as compared to ‘‘naked’’ nuclei of the same kind, and produce as many lines as many nonequivalent nuclei of the same type are present within the molecule. However, a problem arises because the resonance frequency depends linearly on the applied B0. For this reason, a chemical shift (d) scale in units of part per million (ppm) is used in which the position of each line is given by difference (measured in hertz) from the resonance line of an internal standard divided by the operating frequency of the spectrometer expressed in megahertz. The chemical shift scale is independent of the external magnetic ﬁeld and it is thus convenient for comparing data obtained with different spectrometers. The reference chemical shift standard (d ¼ 0 ppm) for 1H- and 13 C-NMR spectroscopies is tetramethysilane (TMS), (CH3)4Si. For other nuclei the reader is referred to the data collected in Table 13.2 [6,68]. Besides electrons, also NMR-active nuclei may mutually inﬂuence each other’s resonance frequencies through a mechanism called coupling. Two general mechanisms are known for coupling, viz. direct dipolar coupling, whereby the magnetic moments interact through space, and J-coupling or scalar coupling, whereby the nuclei interact via chemical bond electrons. J-coupling is the only mechanism active in solution, where dipolar coupling averages to zero. The main effect of J-coupling is the characteristic splitting of the NMR lines according to the number and the spin of the nearby nuclei. Brieﬂy, each resonance is split in 2I þ 1 lines by each of the interacting nuclei of spin I. The distance among the lines is termed coupling constant, proportional to the interaction energy, and ranges from the order of 100–200 Hz in heteronuclear 13C-1H 1J to 0.1–20 Hz in the homonuclear case. For example, the 13C line of a CH fragment is split into a doublet, while it

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 295 17.7.2008 5:39pm Compositor Name: VAmoudavally

295

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

TABLE 13.2 Spectral Window of Most Investigated Nuclei in Food Science Nucleus

Spin

Chemical Shift Range (ppm)

1

1=2 1=2 1=2 5=2 1=2 3=2 1=2 7=2 5=2

12 to 240 to 1200 to 1400 to 100 to 10 to 230 to 40 to 100 to

H C 15 N 17 O 19 F 23 Na 31 P 43 Ca 67 Zn 13

1 10 500 100 300 60 200 140 2700

Internal Reference Standard SiMe4 SiMe4 MeNO2 H2O CFCl3 1 M NaCl in H2O H3PO4 CaCl2 ZnClO4

becomes a quartet in a methyl, because it is repeatedly split by each equivalent proton. This simple rule holds only if the frequency difference among the coupled spins is much bigger than the value of the coupling constant [69]. Direct dipolar coupling takes place because the magnetic moment of a nucleus at one position adds to or subtracts from the ﬁeld experienced by another nearby nucleus. The resonance frequency of the second nucleus thus changes, depending on the orientation of the ﬁrst. In the most common case encountered in food science applications, a less abundant nucleus like 13C is coupled to the highly abundant proton. The interaction energy is proportional to the reciprocal third power of the internuclear distance, and to the term (3 cos2 u 1), where the angle u measures the orientation of the vector connecting the interacting nuclear spins with respect to the direction of the external magnetic ﬁeld. As no chemical bonds are involved, direct dipolar coupling is effective also between nuclei belonging to different molecules, provided they are not too far away from one another, because of the r3 dependence. In solids, where every possible orientation of the vector joining interacting nuclei exists, the line shape of signals takes the form of a so-called Pake-doublet. In liquids, where molecules experience a rapid random reorientation, the term (3 cos2 u 1) averages to zero and no dipolar coupling is detected. However, if complete motion randomness cannot be achieved, e.g., when molecules tumble in small compartments, like metabolites in muscle myoﬁbrils, residual dipolar coupling appears [70,71], whose values resemble that of indirect J-coupling.

13.6.2 SOLUTION-STATE NMR SPECTROSCOPY 13.6.2.1

Acquisition of the One-Dimensional 1H-NMR Spectrum

13.6.2.1.1 Tuning and Shimming Two operations must be performed, before FID recording, to acquire spectra optimized for the highest sensitivity. The ﬁrst one, i.e., tuning=matching, is necessary to change the impedance of the coils to prevent signal losses owing to bad RF power transfer from the probehead to the sample and vice versa. The second one, i.e., shimming, is the operation that improves the magnetic ﬁeld homogeneity, by aligning the force lines in that part of the sample comprised within the receiver coil, resulting in sharper and more symmetric peaks. The effects of both operations on the quality of

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 296 17.7.2008 5:39pm Compositor Name: VAmoudavally

296

Handbook of Food Analysis Instruments

a b c

ppm 3.90

3.80

3.70

3.60

3.50

3.40

3.30

3.20

FIGURE 13.6 1H-NMR spectra, recorded on the same sample of grape juice at 400 MHz in different instrumental conditions: (a) optimal conditions; (b) poor shimming; (c) poor tuning=matching.

the NMR spectra are shown in Figure 13.6, where three 1H-NMR spectra, recorded on the same sample of grape juice at 400 MHz but in different instrumental conditions, are reported. 13.6.2.1.2 FID Acquisition: Pulse Sequence and Solvent Suppressions The simplest pulse sequence generating a typical one-dimensional (1D) 1H-NMR spectrum is shown in Figure 13.3e. Here, P1 is the length of the excitation pulse at a given power and AT is the acquisition time during which the transverse magnetization is collected by a suitable pick-up coil tuned at the resonance frequency, ampliﬁed, and sent to the receiver unit. Note that a radiofrequency pulse of P1 seconds in length will actually excite all nuclei whose resonance frequency lies within 1=P1 Hz from the nominal pulse sequence. For example, a typical pulse length of the order of 5–10 ms (covering 400 and 200 kHz, respectively) is able to excite the whole 1H or 13C spectral range at once (although for high-end spectrometer working at 17.6 T or higher magnetic ﬁeld more sophisticated techniques have been devised). 13.6.2.1.3 Line Broadening, Phase, and Baseline Corrections As a result of phase difference between the sample NMR signal and the reference frequency, the spectrum obtained after FT of the FID may not show lines with pure phases (Figure 13.7). This is usually not a problem because a simple (manual or automatic) phase correction procedure is available in all spectrometers. There are two different types of phase corrections. In the ﬁrst one (known as zero order), the phase of all signals is changed regardless of their position in the spectrum in such a way as to get rid of the line dispersive wings. In the second one (ﬁrst order), phase correction is proportional to the distance between a line and the central frequency of the spectrum. First-order phase distortions affect a spectrum when an unbalanced delay is present between the actual start of the FID and the moment at which FID sampling begins. For example, if acquisition is delayed by Dt then two lines, one exactly at the center of the spectrum and another at f1 Hz farther, will show, respectively, a ﬁrst-order phase shift of 0 and 2pf1Dt rad, because the latter has been precessed by 2pf1Dt rad during the delay (recall the simple explanation given in Section 13.4). Another source of spectral problems is baseline distortion. This can arise from a number of different causes, some of which are easily diagnosed and remedied while others are more difﬁcult to solve. For example, a common error made by NMR beginners is that of using too high a gain, thus

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 297 17.7.2008 5:39pm Compositor Name: VAmoudavally

297

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

ppm (f1) 5.00

4.50

4.00

3.50

ppm (f1) 5.00

4.50

4.00

3.50

(b)

(a)

FIGURE 13.7 Effect of phase errors on the NMR spectrum of a tomato extract is shown (a) before and (b) after the application of the correction.

saturating the dynamic range of the analog-to-digital converter (ADC) and clipping the FID. Spectral lines obtained after FT in this condition show typical sinc wiggles, which severely distort the spectrum and lead to rolling baseline. Setting the gain at a lower value immediately solves the problem. Audio ﬁlters and the so-called probe acoustic ringing (the latter taking place especially at heteronuclear resonance frequencies) are also well known cause of baseline distortion, because of corruption of the ﬁrst few points of the FID. In fortunate cases, ﬂat baselines can still be obtained by submitting the FID to a procedure called backward linear prediction through which the damaged points are mathematically reconstructed using the clean part of the FID. Noise does not actually distort the spectrum, per se. However, when dilution is high and acquisition time is limited, spectral lines may be simply too weak to be clearly detected. In these cases, a line broadening procedure is customary employed by which the FID is point-by-point multiplied by a decaying function of suitable shape (usually an exponential or Gaussian function). This has the effect of convoluting the corresponding spectrum with the FT of the decaying function thus effectively reducing the noise, although at the price of broadening of the spectral lines. In other cases, spectral signals are not particularly weak, but obscured by some other huge signals like those coming from nondeuterated solvents; this forces the user to set gain at too low a value to achieve a proper sampling of the desired signals thus obtaining poor spectra because of the so-called sampling noise. Fortunately, solvent suppression sequences exist by which the unwanted lines are saturated by long (up to 1–2 s) low-power radiofrequency pulses, or dephased by proper application of selective pulses and pulsed ﬁeld gradients, before the actual start of acquisition [72]. 13.6.2.2

Applications in Food Science

13.6.2.2.1 Chemical Composition of Extracts and Solutions One-dimensional 1H-NMR spectra provide plenty of information on the chemical composition of a solution. Each molecule, containing hydrogen atoms, gives as much signals as the number of chemically and magnetically inequivalent hydrogens. Since, practically, all molecules contain hydrogen atoms, the spectra of solutions extracted from foodstuffs are characterized by a crowded envelope of signals, whose overlap tends to decrease as the magnetic ﬁeld strength increases. In the literature, the applications of 1D 1H-NMR spectroscopy to food science may be grouped according to two different strategies: (i) individual signal assignment to the corresponding molecular species and (ii) exploitation of the spectra as a whole metabolic proﬁle to be analyzed by chemometrics.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 298 17.7.2008 5:39pm Compositor Name: VAmoudavally

298

Handbook of Food Analysis Instruments

In both cases, the technique takes advantage of the constancy of the chemical shift (d) of each signal belonging to a certain molecule, provided that the same molecular environment (pH, ionic strength, polarity, etc.) is always ensured. Thus, the presence of a molecule in the analyzed extract is conﬁrmed by the individuation of all its signals at the expected chemical shift and with their correct multiplicity. For this reason, literature chemical shift data are used as guidance for assignments [73]. According to Croasmun and Carlson [74], the 1H resonance of the methyl in position 18 of sterols always appears in the 0.6–0.7 ppm narrow range, well-separated from all of the other resonances; therefore, it is possible to get information about the sterolic composition in food by observing this useful small spectral region [75]. The molar ratio of a given metabolite is calculated using the corresponding area integral, provided that a sufﬁcient relaxation delay is applied to the pulse sequence to allow the complete magnetization recovery between scans, avoiding the partial saturation of long relaxing nuclei. The extraction is required for solids, usually consisting in a one-step procedure to obtain both the organic and the aqueous extracts. For example, a few grams of foodstuff dispersed in the smallest possible volume (few tens of milliliters) of a mixture chloroform=methanol (1:1, v=v) are grounded and homogenized with a ceramic pestle, and then water is added. The homogenate is centrifuged and a three-phase system is obtained (solid-organic-aqueous): the two liquid phases (chloroform and water=methanol) are separated and dried, and their residues are dissolved in CDCl3 (0.6 mL) and D2O (0.6 mL), and placed into 5 mm tubes. Both organic and aqueous solutions are separately investigated by NMR, requiring about half an hour for the acquisition of useful spectra on each of them. It is worth noting here that most of the time is spent for the experimental set up, being about half an hour the time required for tuning, matching, and shimming. These steps require more time for the ﬁrst sample of the same nature, e.g., aqueous or organic, whereas are almost immediate for the successive ones. Only a few minutes (e.g., <5 min) are required for the FID acquisition, thus making it convenient to run NMR experiments by grouping, in succession, samples of the same nature, e.g., solvent or ionic strength. The sample preparation step is absent for nonviscous homogeneous liquids, or consists in just a dilution with an appropriate deuterated solvent for viscous ﬂuids. For instance, sample preparation is immediate in the analysis of edible oils: oil (0.2 g) mixed with 400 mL of deuterated chloroform, containing traces of TMS as internal reference for chemical shift, is introduced into a 5 mm diameter NMR tube. The experiments have a duration of 3.5 min, since 32 scans sufﬁce for the acquisition of spectra, with the optimal relaxation delay of 3 s which follows the 2 s acquisition time (usually, 8 dummy scans precede the acquisition with a recycle time of about 5 s, given by the sum of the relaxation delay and the acquisition time). A single NMR spectrum provides information otherwise gained with cumbersome and timeconsuming methods: the unsaturation degree of butterfat [76], the proportion of v-3 fatty acids in ﬁsh oils [77–80], the unsaturated fatty acid composition of vegetable oils [81], and the iodine value of the same oils [82,83], as well as the proportion of several acyl groups in oils extracted from seeds are meaningful example of NMR spectroscopy application. All these studies are based on the assignment of 1H-NMR spectral signals to different kinds of protons and on the subsequent treatment of the values obtained from the integration of these signals [84]. The solvent affects signal position in a spectrum. For example, differently from what happens in neat chloroform, a mixture of deuterated chloroform and dimethylsulfoxide causes the fatty acids methyl protons in the range 0.8–1.0 ppm to overlap, although other diagnostic signals are better resolved in different spectral regions [85]. To avoid solvent effects, it would be desirable to analyze the sample as it is. For example, it has been possible to obtain good spectra for a sample containing 600 mL of milk and 50 mL of D2O (for lock and shim) [86]. The huge water signal (ranging from 4.5 to 4.9 ppm) has been successfully suppressed by presaturation. Owing to high molecular weight of the molecules and aggregates

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 299 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

299

contained in milk, spectra displayed broad lines and overlapping peaks. In these cases, it is useful to divide the spectra into main regions, where it is possible to allocate classes of substances, instead of single species, characterizing the foodstuff under investigation. Therefore, the 1H-NMR spectrum of milk has been divided into ﬁve main regions: a high-ﬁeld region between 0.6 and 2.2 ppm assigned to the signals of fats acyl chains; a mid-high-ﬁeld region between 2.2 and 3.1 ppm, containing weak signals because of trace compounds of milk; a mid-low-ﬁeld region between 3.1 and 5.5 ppm with signals assigned to lactose, glycerol (at 3.98 and 4.18 ppm), and the protons bound to the carbons forming the double bond of unsaturated milk fats (5.20 ppm); the low-ﬁeld region from 5.5 to 9.0 ppm contains very small signals; and the most downﬁeld being those owing to the amide protons of proteins. In the 1H-NMR spectrum, the signals of lactose are narrow and intense, whereas the signals of milk fats are broad, because lactose is soluble, but milk fats consist of a variety of acyl chains and are suspended in milk as fat globules. The different NMR characteristics presented by species with different molecular weights are exploited to distinguish some fast diffusible molecules (e.g., monosaccharides) from their slow high molecular weight congeners, such as polysaccharides. The latter may not be easily detectable in the 1 H spectrum, because of the signals overlapping or of their low concentration, and are observable only after a signiﬁcant vertical expansion. However, the presence of polysaccharides can be evidenced by means of 1H-diffusion-ﬁlter-edited experiments [87], corresponding to the 1D version of the DOSY (diffusion-ordered spectroscopy) experiment, discussed later in Section 13.6.5. In these experiments, the application of strong magnetic ﬁeld gradients allows the contribution of low molecular weight metabolites to be suppressed: the resulting 1H spectra show only the resonances owing to compounds with a high molecular weight [75]. To acquire such a kind of spectra, the spectrometer must be equipped with a multinuclear z-gradient probehead capable of producing gradients in the z-direction with strength of some tens of 55 G=cm. 13.6.2.2.2 NMR Spectroscopy of Paramagnetic Species in Foodstuffs Foodstuffs may contain molecules bound to paramagnetic ions, i.e., bearing unpaired electrons, such as the transition metal ions—iron, copper, or manganese. The presence of paramagnetic species, naturally present or added during food processing, determines the appearance of some diagnostic signals in the spectra of such complicated systems, providing extra information otherwise not obtainable. Some other paramagnetic metal ions, such as some lanthanides which can substitute the diamagnetic calcium ions, can also be added to samples submitted to NMR investigation to simplify the spectra. In fact, protons experiencing the presence of unpaired electrons undergo large modiﬁcations in the chemical shift (known as the isotropic shift) [88,89], pushing their corresponding signals well outside of the 0–10 ppm range typical of diamagnetic substances. The other effect is on the nuclear relaxation times that are dramatically shortened by the paramagnetic effect, resulting in T1 and T2 relaxation times on the order of milliseconds. The paramagnetic signals may be enhanced with respect to the remaining diamagnetic signals present in the spectra, by using very short relaxation delays (e.g., 100 ms) in the pulse sequences. In these conditions, after tens of dummy scans, the diamagnetic signals are partially saturated whilst paramagnetic ones are unaffected, since they are completely recovered during the short relaxation delay. The fast relaxation of paramagnetic signals was exploited in the study of turkey meat extracts (Figure 13.8) [31]. NMR spectra are able to provide the oxidation state of myoglobin (Mb) in meat extracts. Only the Oxy-Mb is diamagnetic, with low-spin Fe2þ, whereas the other two forms have unpaired electrons on the iron ion: high-spin Fe2þ (Deoxy-Mb) and high-spin Fe3þ (Met-Mb). The latter gives an adduct with cyanide which possesses low-spin Fe3þ, characterized by an 1H-NMR spectrum with the sharpest paramagnetic signals (Figure 13.8b). The quantitative determination of Met-Mb, responsible for meat darkening, is thus easily performed by NMR, which is, different from photometric methods, also able to discriminate among other myoglobin species.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 300 17.7.2008 5:39pm Compositor Name: VAmoudavally

300

Handbook of Food Analysis Instruments

90

80

70

60

25

20

15

ppm

90

80

70

60

50

40

30

20

ppm

0

⫺10

⫺20

⫺30

⫺40

ppm

(a)

(b)

⫺5 50

40

30

20

ppm

0

⫺10

⫺20

⫺30

⫺10 ppm ⫺40

ppm

FIGURE 13.8 1H-NMR spectra of aqueous meat extracts from turkey thigh: (a) eluate from the cationic exchange column; (b) sample (a) with addition of 5 mM sodium cyanide.

13.6.3 HETERONUCLEAR SPECTROSCOPY 13.6.3.1

Pulse Sequences and Instrumental Requirements

Basically, all nuclei exploited in food science, like 1H, 13C, 19F, and 31P, can be excited using the simple one-pulse sequence shown in Figure 13.3e. For nuclei other than 1H, the so-called proton decoupling, or irradiation, is usually applied during the analysis so as to avoid JXH scalar coupling development and spectra complication. Thus, the pulse sequence needs two channels for its application (Figure 13.3f), i.e., one for the 1H-decoupling, the other for the observed X-heteronucleus. Several decoupling techniques, based on composite rectangular pulses, have been proposed such as MLEV [87,90], WALTZ [91], GARP [92], and other ones [93]. WALTZ sequence, with decoupling pulses of 35 ms and a decoupler power of 2 W, gives the best results for simple heteronuclear experiment, since this pulse sequence ensures a homogeneous decoupling along the entire spectral range. However, as a side effect to decoupling, and owing to a type of cross relaxation called NOE (nuclear Overhauser effect), proton irradiation causes an enhancement in lines corresponding to protonated nuclei. In 13C spectroscopy, this effect is particularly evident because enhancements up to 300% can be reached. NOE enhancements, together with long T1 relaxation (tens of seconds to 1 min), are the main causes for the usual noncorrespondence between signal integrals and number of equivalent carbons. In fact, if the actual integrals were to be measured, an interpulse delay of several minutes (5 T1) should be applied, which would make the total analysis too lengthy (days). 13.6.3.2

Applications in Food Science

13.6.3.2.1 13C-NMR Spectroscopy 13 C spectroscopy does not only provide additional information to that already obtained by 1H-NMR, but also provides complementary information without which unambiguous interpretation of spectra may be often difﬁcult. An emblematic case is related to the characterization of the acyl positional

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 301 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

301

distribution of triglycerides in different vegetable oils whereas, because of the strong overlap of the signals, 1H-NMR spectrum cannot provide this type of information. The linoleic=oleic ratios in the inequivalent positions sn 1, 3 and sn 2 of the glycerol molecule, as measured in the carbonyl region of 13C spectrum, are characteristic for vegetable oils, being different for olive, hazelnut, soy, and peanut oils [94]. The sample preparation is very simple, although differences in solvent polarity and concentration may cause mismatching with literature data [95]. Oils (5–100 mL) are placed into 5 mm tube and diluted with deuterated chloroform (700 mL) containing a small trace of TMS. After running the spectrum, small amounts of different known glycerol tri-esters are usually progressively added to the oil sample and the spectrum is acquired again. The assignment is performed by comparing the relative intensities before and after this addition [95]. In fact, the value of 13C chemical shifts is concentration dependent, although it ranges only from 173.241 to 173.278 ppm for the resonance of 1,3-carbonyl group of oleic acid, and with variations of the order of less than 0.001 ppm in the inner parts of fatty chains. It is worth noting that the major observed differences are in the carbonyl region, which is the region proposed by many authors to perform the acyl distribution measurement [96,97]. The major disadvantage of high-ﬁeld 13C compared to 1H-NMR spectroscopy is the loss in the intensity of the NMR signals owing to the low sensitivity of the observed nuclide; as a consequence, a 13C spectrum requires much more analytical time and shows only signals of the major components of olive oil, i.e., glycerol esters. To observe the 13C-methyl signal of linolenic acid in extra virgin olive oils, and to achieve a reasonable S=N, an acquisition of 2 days is necessary at 400 MHz. On the contrary, proton-NMR spectrum allows obtaining information about minor components, i.e., aldehydes, terpenes, and sterols, in reasonable times. Also for the 13C-NMR spectra of whole milk, the assignment of NMR signals was tentatively performed by referring to published data of chemical shifts for most compounds of milk (lactose and fats), and conﬁrmed by addition of authentic compounds to whole milk, followed by inspection of the different spectra [86]. 13.6.3.2.2 31P-NMR Spectroscopy Quantiﬁcation of phospholipids (PLs) mixtures by phosphorus nuclear magnetic resonance (31P NMR) has some advantages over HPLC (high-performance liquid chromatography), despite the disadvantage that the former is much less sensitive than the latter technique. First of all, 31 P-NMR quantiﬁcation of PLs can be carried out using the total lipid fraction of a sample obtained, e.g., by Folch extraction [98]. In fact, only compounds that bear at least one phosphorus atom are detectable in the 31P-NMR spectrum of the total lipids extract, so that no signals from other lipids appear. Secondly, the integrated area of each peak is directly proportional to the concentration of the respective PL, provided that the 31P-NMR spectrum has been recorded under proper conditions (e.g., nuclear relaxation completely allowed). The choice for the proper solvent is of great importance when analyzing PL mixtures by 31P NMR [99]. A good solvent should spread the phosphorus PL resonances over the broadest possible range while keeping line widths at a minimum. As this is generally not achievable with neat organic solvents, owing to the formation of PL aggregates, more complex systems based on the use of detergents [100] or mixtures of solvents [101] have been devised over the years. Phosphorus-31 spectroscopy is also used to study phosphoproteins. Proton-decoupled 31P-NMR spectra at pH 8.0 have been acquired on 30 mg=mL egg white proteins, dissolved in a 0.7% urea and 0.7% SDS D2O solution, to assess the effects of their phosphorylation performed by dry-heating in the presence of phosphate [102]. The peaks at about 4.7 ppm (d ¼ 0.00 ppm for 85% H3PO4), identiﬁed as phosphoesters of serine, are present in both native and dry-heated proteins. The peaks at 6.6 and 5.3 ppm, related to polyphosphates, are only present in dry-heated proteins submitted to phosphorylation, and can be monitored to follow the phosphorylation step that is performed to improve technological properties (e.g., water solubility, emulsifying activity, foaming and gelforming properties) of food proteins.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 302 17.7.2008 5:39pm Compositor Name: VAmoudavally

302

Handbook of Food Analysis Instruments

Phosphorus-31 may also offer a promising prospect to the analysis of dispersion of skimmed milk powder (SMP) [103]. The stability of the micelle, maintained partly by a mineral composite known as colloidal calcium phosphate (CCP), is disrupted by EDTA, breaking down casein micelles into sub-micelles or even individual casein protein components, as emerging from the appearance of nine extra peaks, predominantly upﬁeld from the principal inorganic phosphate peak, observed by 31 P-NMR spectroscopy. Thus, mobile phospho-serine residues from different forms of casein are revealed and such a speciation may be examined to evaluate different milks. 13.6.3.2.3 Site-Speciﬁc Natural Isotopic Fractionation by Nuclear Magnetic Resonance This type of spectroscopy, developed by G.J. Martin in 1980s [104–106], is based upon the natural isotopic distribution of atoms inside molecules. The technique is able to determine the 2H=1H isotopic ratio of the different sites of a molecule. This information, supported by a robust database of geographical isotopic radio distribution, permits to individuate the botanic and geographic origin of natural substances such as sugars, ethanol, aromas, glycerol, fatty acids, etc. The main application of this technique has been so far the wine sector (EC Reg. no. 2676=90), but there are many other food or natural products, such as orange juices [107–109], honey [110–112], and olive oil [113,114] that are subjects of SNIF-NMR (site-speciﬁc natural isotopic fractionation by NMR). A set of 1H-, 2H-, and 13C-NMR spectra, measured on triglycerides extracted from adipose tissue of fresh meat, provides also the isotopic composition of fatty acids resulting from the feeding diet of the animals. Triglycerides are dissolved in CDCl3, to record 1H- and 13C-NMR spectra, or in CHCl3 for 2H-NMR studies. In the latter case, C6F6 is added to provide the lock signal, since deuterium is under observation. The method is based on the integration of signals belonging to deuterium nuclides measured in speciﬁc molecular positions, and on the evaluation of their ratios with the area of the signals corresponding to the hydrogen isotope found on the same molecular positions. A combination of such ratios (e.g., [D=H]d:2, [D=H]d:0.9, etc.) were sufﬁcient to discriminate the production site of the animals [115]. Discriminatory statistical analysis performed on site-speciﬁc 13C=12C ratios, at natural abundance, permits to unambiguously distinguish molecules of different origins, either natural or synthetic. Major sources of inaccuracy in quantitative 13C-NMR analysis are the relaxation times, the decoupling pulse and power, and the NOEs. As far as the ﬁrst point is concerned, the repetition time for accumulative scans must be long enough to permit the complete recovery of the nuclear magnetization before starting a new scan. A major reduction in the experimental time results from a preliminary study of the relaxation times and variance analysis, to sacriﬁce only those signals that are not informative or not so important for the quantitative analysis. An accurate approach for the choice of the correct recycle time is presented in a study on vanillin, an important additive in the food and ﬂavor industry for its organoleptic qualities. It presents several technical difﬁculties for 13 C-NMR spectroscopy, as it has long relaxation times (up to 19 s), a wide range of chemical shifts in carbon (135 ppm) and in proton (6 ppm), and a wide range of the NOE factor. Isotopic deviations from statistical distribution are quite origin speciﬁc and the biggest differences between vanillin of different origins are in sites 1 and 8. By omitting sites 2–4, the maximum T1 value becomes that of site 8 (2.95 s) and the recycle delay becomes 21 s instead of 131 s [116]. Under these conditions, measurements are no longer quantitative for sites 2–4. However, the remaining ﬁve sites are sufﬁcient to discriminate different origins.

13.6.4 SOLID-STATE NMR SPECTROSCOPY: CP-MAS AND HR-MAS 13.6.4.1

Theory of the Cross Polarization and of the Magic Angle Spinning

Proton spectra obtained from solid samples suffer from poor resolution as the magnetic susceptibility gradients, caused by the heterogeneous nature of the sample (e.g., muscle tissue), result in the broadening of the proton resonances [117]. In addition, residual dipolar coupling and chemical shift

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 303 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

303

anisotropy (CSA) interactions, originating from incomplete motional averaging of molecules, can cause line broadening. The dependence of the dipolar coupling and CSA on (3 cos2 u 1), where u is the angle between the internuclear vector and B0, can be exploited for obtaining liquid-like 13C spectra from solids. In the so-called magic angle spinning (MAS) technique, the sample is rapidly spun (3–10 kHz) at the magic angle (u ¼ 54.748), i.e., the angle at which the above expression is null. Proton decoupling during acquisition provides a further mean for improving line narrowing. Even so, the 13 C-NMR intensity of a non-13C-enriched sample may be too small to allow for clear signal detection. To increase its signal, the technique of the cross polarization (CP) is employed, whereby the intensity of the lower sensitivity nucleus is enhanced by transfer of polarization from more abundant high sensitivity nucleus, provided that the two nuclei are dipolarly coupled. Technically, CP is achieved by contemporary irradiation of both nuclei at the respective resonance frequency in such a way as to fulﬁll the Hartman–Hahn condition [118]. Solid-state spectra of protons are unfortunately not as easily obtained. As proton–proton dipolar coupling is of the order of 100 kHz, their lines are still too broad to be useful even at the highest sample spinning frequencies nowadays reachable (50 kHz or more). However, in semisolid materials, where dipolar broadening is already partially averaged out by residual molecular motion, liquid-like 1H spectra can be obtained by the HR-MAS technique [119,120]. HR-MAS (High Resolution Magic Angle Spinning) can be considered a hybrid between solution-state and solid NMR because the sample is placed at the magic angle and spun at 3–15 kHz, but all the common 1D and 2D experiments are carried out as in normal liquid-state NMR. Furthermore, no CP is necessary when looking at low sensitivity nuclei like 13C or 15N. 13.6.4.2

Applications in Food Science

The measure of pH in muscle tissue is a difﬁcult task to accomplish without disrupting the cellular structure. MAS-NMR provides a less invasive pH measurements by its calculation (Equation 13.7) from the chemical shift of the histidine C-4 proton of the dipeptide carnosine (b-alanyl histidine), as reported by Pan et al. [121]. pH ¼ pKa þ log where pKa dH dA d

dH d d dA

(13:7)

(set to 7.10) is the acid dissociation constant of carnosine (set to 7.25 ppm) is the chemical shifts of the acid form (set to 6.92 ppm) is the chemical shifts of the alkaline form is the observed chemical shift

Actually, the conditions experienced by the sample during the MAS-NMR measurement do not correspond to the situation in the intact muscle, since in the present conditions the high spinning rate may cause disintegration of the sample structure. Therefore, development of slow-spinning techniques could be attractive [122]. Still in intact muscle ﬁbers, the degradations (dephosphorylation) of phosphocreatine and ATP (Adenosine Triphosphate), and the subsequent formation of Pi, are measured by 31P-MAS-NMR to follow the postmortem transformation of muscle in meat [123]. Such a biochemical transformation has also been followed by measuring, in the solid-state NMR spectra, the areas of 13C resonance signals from saturated and unsaturated fatty acids, lactate, and glycogen. Since the fast rotor spinning (e.g., 4.5 kHz) may be responsible for sample overheating, measures are usually performed in prechilled (at about 308C) rotors, where the actual sample temperature is determined, in separate experiments, by using 207Pb-MAS-NMR of Pb(NO3)2 as an NMR thermometer [124].

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 304 17.7.2008 5:39pm Compositor Name: VAmoudavally

304

Handbook of Food Analysis Instruments

HR-MAS-proton NMR spectra are recorded using dedicated 1H=13C probeheads, where the samples are spun at 5 kHz. The technique, in combination with common statistical tools (principal component analysis and discriminant analysis), provides a quick tool, despite its limited sensitivity, for characterization of the age of Parmigiano Reggiano cheese [125]. In this way, selected free amino acids and other low molecular weight metabolites were found to be among the most relevant compounds characterizing the ripening of Parmigiano Reggiano cheese. The same approach has also been adopted to discriminate European Emmental cheeses according to their geographical origin with 89.5% correct reclassiﬁcation, although these countries are geographical neighbors. It is possible to point out several spectral domains, suitable for this discrimination, at 2.75–2.80 ppm (unsaturated fatty acids and aspartic acid signals), 5.45–5.50 ppm (oleﬁnic protons), 3.90–3.95 ppm (serine signals), and 2.80–2.85 ppm (unsaturated fatty acids and asparagine signals). The spectral domain at 5.30–5.35 ppm (corresponding to unsaturated fatty acids signals) can be used as a marker of Swiss Emmental cheese [126].

13.6.5 TWO-DIMENSIONAL 1H-NMR SPECTROSCOPY 13.6.5.1

Theory of the COSY, NOESY, TOCSY, and ROESY Spectra

The simple representation of a spectrum as a collection of resonance frequencies running along a single axis changed after 1976, when two-dimensional (2D)-NMR spectroscopy was introduced by Ernst and coworkers [127]. 2D-NMR builds on the notion that special NMR sequences can be designed that induce transfer of magnetization between two scalarly or dipolarly coupled nuclei or between nuclei which experience chemical exchange. As a result of this transfer, signal intensities, or phases, of the 1D spectrum change. The trick played in 2D spectroscopy is that of recording a number of FIDs (usually a few hundreds), while ﬁne tuning the timing of the magnetization transfer in such a way as to modulate the lines of the spectra obtained after the ﬁrst FT, along the acquisition time t2 (t2-domain), with the frequencies of the coupled nuclei. A second FT along the direction t1 of the new oscillation (the so-called interferogram) yields the peaks in the second dimension. In the resulting 2D spectrum, peaks appear on the diagonal, at the same position they occupy in the original 1D spectrum; out-of-diagonal peaks (cross-peaks) indicate signals corresponding to coupled protons. A typical 2D-NMR spectrum is thus characterized by two frequency axes representing the chemical shift of the interacting nuclei and by peaks (usually reported as contour plot) connecting the signals of these nuclei. Correlation spectroscopy (COSY) and total correlation spectroscopy (TOCSY) spectra provide information on internuclear scalar connectivities, i.e., those acting through bonds. In the ﬁrst experiment, only short-range couplings are detected, i.e., 1,2JHH, thus highlighting spin patterns among geminal or vicinal hydrogen atoms, while the second experiment is able to detect long range connectivities, thus offering 2D maps connecting the whole spin patterns. In COSY spectra, the minimal cross-peak intensity between signals of vicinal hydrogens is observed for dihedral angle approaching 908, for which null-scalar interaction energy is expected, according to the cosine Karplus relationship [128]. At variance with the above experiments, where information are obtained about the bonds’ connectivities within a molecule, the nuclear Overhauser effect spectroscopy (NOESY) and rotating frame nuclear Overhauser effect spectroscopy (ROESY) experiments aim at obtaining information on spatial proximity of nuclei. In NOESY, the intensity of each cross-peak is proportional to the inverse of the sixth power of the distance between each couple of nuclei [129]. The absolute intensity of the NOESY cross-peaks also depends, with a useful oversimpliﬁcation, on the molecular weight of the molecule under observation. If all peaks of the spectrum diagonal are phased with negative intensities, cross-peaks are positive for small solutes, negative for high molecular weight solutes (such as proteins or polysaccharides), and close to null for midmolecular-weight samples (in the range 800–1200 Da, also depending on solvent viscosity and

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 305 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

305

spectrometer frequency). In cases of null NOESY cross-peaks, information can still be obtained by resorting to ROESY, whose cross-peaks intensities are always positive regardless of the molecular weight. ROESY sequence, in its basic form is not different from TOCSY, except for a lower spin lock (SL) power used during the mixing time, but more reﬁned versions of this sequence have been designed [130]. The last few years have witnessed the renaissance of a special kind of 2D spectrum (called J-resolved spectroscopy) in which one axis represents the chemical shift of a nucleus and the other shows the shape of the multiplets that every coupled signal of the spectrum is able to yield. J-resolved spectroscopy has proven particularly useful for analyzing complex systems like vegetable extracts [131,132] and for metabolomic studies in general [133]. 13.6.5.1.1 Pulse Sequences Two-dimensional NMR sequences are characterized by four time periods termed preparation, evolution, mixing, and detection. In the ﬁrst period, the system is excited and the energy levels of at least one nucleus are perturbed. The following period is one that is incremented in successive experiments and gives rise to one of the two time domains that form the 2D data set. The mixing time may or may not be explicitly present in the sequence. Sometimes, it is simply represented by a single pulse (like in COSY, Figure 13.3g) and sometimes, it is ﬁlled by a complex pulse sequence known as a spin lock (like in TOCSY). Finally, in the detection period, the FID is obtained, which contributes the time domain in which the ﬁrst FT is applied. The COSY pulse sequence is the archetypical homonuclear 2D experiments that paved the way for the diffusion of 2D-NMR in the scientiﬁc world [127]. In its basic form, the COSY sequence features only two pulses, the ﬁrst acting as excitation and the second as mixing pulse. The evolution time (t1) increases by a constant amount (few microseconds) along the FID series. The intensity of cross-peaks in the COSY spectrum is proportional to the H–H coupling constants between neighboring hydrogen nuclei. The coupling also bears important structural information. For instance, neighboring diaxial hydrogen atoms in pyranosides can be clearly identiﬁed by large coupling constants (7–9 Hz), which contrast with smaller coupling constants (2–4 Hz) of protons in diequatorial or axial-equatorial conﬁgurations. A useful modiﬁcation of the COSY sequence is the double quantum (DQ) ﬁltered COSY. In DQ-COSY, all singlets such as those yielded by many solvents (e.g., water, chloroform, or methanol—whose signals are usually much more intense than the solute) disappear from the FID thus allowing for a better use of the dynamic range of the receiver. In TOCSY (Figure 13.3h), the mixing time contains a long (tens of milliseconds) mid-power pulse or, equivalently, a long series of short high-power pulses, whose effect is that of locking transverse magnetization of the excited nuclei over one axis of the rotating frame (whence the name of spin lock). While TOCSY spectra with short mixing SL times (10–20 ms) resemble COSY, they show all possible connectivities within a spin system when mixing time is lengthened to about 80–100 ms. For example, in the TOCSY spectrum of sucrose, cross-peaks will appear among all peaks of glucose and all peaks of fructose, but not between any two signals of protons belonging to different monosaccharidic unities. In NOESY (Figure 13.3i), the mixing time separates the second 908 pulse and the third 908, which is the detection pulse. Its length is proportional to the internuclear distance range selected to be detected, since longer mixing times are able to detect NOEs among nuclei at longer distances. Cross-peaks are usually observed in NOESY spectra between proton pairs that are close in space, typically less than 5 Å. Care must be taken to avoid the so-called spin diffusion: very long mixing times (hundreds of milliseconds) produce cross-peaks even between nuclei far away, since magnetization may be transferred through a third interposed nucleus. Several modiﬁcations of the basic NOESY pulse scheme have been proposed to remove such undesired spin diffusion effects [134].

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 306 17.7.2008 5:39pm Compositor Name: VAmoudavally

306

13.6.5.2

Handbook of Food Analysis Instruments

Applications in Food Science

The main use of homonuclear 2D NMR in food science is related to the structure elucidation of unknown substances, e.g., the acetophenone glycosides from thyme [135], the carotenoids in tomato juice [136], the apocarotenoids derived from oxidative cleavage of C-40 capsanthin in paprika [137], the oxidative theaﬂavin dimers extracted from the black tea [138], the radical termination products of the turmeric pigment curcumin [139], etc. The chemical characterization also includes the assessment of their stereochemistry or conformational state adopted in different chemical environments [140–144]. There are a paramount number of examples of application of homonuclear 2D spectroscopy to food science studies in literature, and it is not easy to select representative case studies. However, some examples will be offered to give a hint on the potentialities offered by the technique. For example, the NOESY cross peaks showed that C-20 of novel spirostanols and furostanols extracted from the seeds of allium tuberosum is in the R-conﬁguration, which is a rare occurrence for natural steroidal saponins [140]. Homonuclear 2D NMR techniques also served to identify lycopene isomers, carotene, and phytoene in the NMR proﬁle of lipids extracted from tomato juice [136]. The oleﬁnic region, related to the conjugated double bonds (5.8–7 ppm), is the key area for the identiﬁcation of most of these carotenoids by NMR. The comparison between ROESY and TOCSY spectra resulted in the elucidation of (5Z)-lycopene by means of an ROE correlation between the –CH2– group (at 2.23 ppm) in position 4 and the –CH ¼ group (at 6.48 ppm) in position 7, a cross peak is only compatible with the Z-conﬁguration of the double bond in position 5 of this carotenoid. NOESY techniques showed proximity between hydrogens from the cinnamic acid acylating group and the C-4 of the pelargonidin of red radish (Raphanus sativus) anthocyanins [141]. This ﬁnding is consistent with the ﬂexible saccharide chains acting as linkers, allowing the folding of the acyl aromatic rings over the planar pyrylium ring. This structural conformation prevents the addition of nucleophiles, especially water, to the C-2 and C-4 positions of the anthocyanin, protecting the chromophore against hydration and diminishing the formation of the colorless pseudobase. The stereochemistry of C-2, C-3, and C-4 of the bayogenin, a triterpene saponin from aerial parts of Medicago arabica, was conﬁrmed by 2D-NOESY experiments [143]. A cross-peak between H-2 ((d ¼ 4.56 ppm) and H-3 (d ¼ 4.41 ppm) was observed, conﬁrming the presence of the 2b,3b-dihydroxyoleanane skeleton; additionally, the absence of correlation between H-3 (d ¼ 4.41 ppm) and CH3-24 indicated the b-conﬁguration of the 23-hydroxymethyl group. Another generic use of 2D spectroscopy exploits the dramatic increase of spectral resolution gained by dispersing cross-correlation peaks on a bidimensional plot, thus limiting the occurrence of signal overlaps, generally observed in monodimensional spectra. In this way, the chemical composition of tomato juices from two cultivars was determined by a combination of J-resolved, COSY, TOCSY, and other 2D-NMR experiments [131]. In this study, the 2D sequences were used to assign each spin system and to separate the components of the complex patterns found in the 1Doverlapped proton spectra. Different conclusions must be sketched out, when analyzing the 2D maps, by considering the nature of the sample, i.e., whether it consists of a solution of a neat substance or it is a rough mixture of molecules extracted from foodstuffs. In the ﬁrst case, the volume of cross-peaks is directly proportional to the interaction energy between the two connected nuclei, while in the second instance, the volume is also proportional to the concentration of the substance to which the two interconnected nuclei belong. By taking into account such a distinction, the absence of a diagnostic cross-peak in the solution of a pure substance means that either the correspondent interconnected nuclei are no more bound to each other, or the molecule underwent a conformational change, e.g., because of pH, temperature, and matrix effects, leading to a structure where the cross-related nuclei have minor interaction energy. In mixtures, the volume integrals of cross-peaks are also proportional to the concentration of the corresponding substance, thus revealing the quantitative composition of a

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 307 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

307

mixture, once the inﬂuence of structural changes is excluded by the fact that the integral ratios of all signals belonging to same molecule remain constant. A recent application of COSY to automatically and rapidly identifying and quantifying metabolites in deﬁned mixtures of amino acids as well as in real complex biological samples has been reported [145]. Given a database of 2D-COSY spectra for the metabolites of interest, the method provides a list sorted by a heuristic likelihood of the metabolites present in the sample analyzed by COSY. At last, 2D-NOESY spectra also provide chemical exchange correlations, a valuable information about the structural transformations, including evidence for the relationship between various equilibrium forms. For instance, negative exchange cross-peaks are observed in the NOESY-NMR spectra of anthocyanin, between the two epimeric hemiacetals and the corresponding ﬂavylium cation, indicating that hemiacetals are in equilibrium with each other through the corresponding intermediate cation [146].

13.6.6 HETERONUCLEAR 2D SPECTROSCOPY 13.6.6.1

Pulse Sequence of HSQC, HMQC, HMBC, and INADEQUATE Spectra

Are all homonuclear sequences suitable for all homonuclear cases? Unfortunately, they are not. The homonuclear sequences described above can be applied when at least two nuclei of the same nature, belonging to the same molecule or supramolecular species, are scalarly or spatially coupled. In natural abundance 13C spectroscopy, e.g., because of scarcity of the 13C isotope (P ¼ 0.011%), one has a probability equal to N P of ﬁnding at least one 13C nuclide in a molecule containing N carbons, but only a probability of (N 1) P2 of having two vicinal 13C nuclides in a network of N interconnected 13C atoms. In sucrose, these probabilities amount, respectively, to 13.2% and 0.12%. A 13C homonuclear-COSY spectrum of a glucose sample would only yield a 2D map showing 12 peaks on a diagonal, the desired cross-peaks due to J-coupling between vicinal carbons being more than 100 times smaller. Nevertheless, information about 13C–13C connectivity can still be obtained by applying the INADEQUATE sequence reported in Figure 13.3j, which bears some resemblance to DQ-COSY. As a result of DQ excitation, occurring during t1, the signal recorded during detection will be 100 times less intense than one obtained from a normal 1D 13C-NMR spectrum. Moreover, as DQ evolves during t1, an INADEQUATE spectrum will be characterized by a normal horizontal axis showing carbon chemical shifts and a vertical axis displaying the frequencies of the DQ terms. Connectivities between two connected carbons will thus be seen on the spectrum as peaks sharing the same DQ frequency (Figure 13.9). Recently, however, a simple mathematical transformation has been proposed for displaying INADEQUATE data set in the form of a ﬁctitious 13C–13C COSY spectrum [147]. Heteronuclear connectivities (typically between protons and carbons) are nowadays routine tasks. The three most useful experiments in this respect are HMQC (heteronuclear multiple quantum coherence), HSQC (heteronuclear single quantum coherence), and HMBC (heteronuclear multiple bond coherence). The ﬁrst and second experiments exploit scalar coupling connections between carbons and directly bound hydrogens; the third sequence allows detection of long-range connections via 2JCH and 3JCH. The preparation part of an HMQC sequence (Figure 13.3k) comprises the ﬁrst 1H pulse and the subsequent 1=2JCH delay, followed by the ﬁrst 13C pulse. Now, 13C nuclei precess in the xy-plane with the sum (MQ) of the carbon and proton resonance frequencies. The application of a 1808 proton pulse reverses the precession of the proton part of the MQ, thus retaining only 13C precession during t1. MQ is retransformed into an anti-phase term by the second carbon pulse and, after another 1=2JCH delay, it is recorded during detection. After FT in both dimensions, an HMQC spectrum will show proton and carbon frequencies, each on a separate axis, and peaks will be detected between each (nonquaternary) carbon and its attached protons.13C decoupling is usually applied during

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 308 17.7.2008 5:39pm Compositor Name: VAmoudavally

308

Handbook of Food Analysis Instruments

200

180

160

140

120 100 ppm

80

60

40

20

FIGURE 13.9 INADEQUATE spectrum, recorded at 200 MHz, of a lipid component extracted and puriﬁed from avocado pear. The signals are assigned to each sequential backbone 13C nucleus by following straight lines in the 2D spectrum. The top trace shows the 1D 13C spectrum of the sample.

detection so as to obtain proton signals as singlet (and not doublet owing to coupling with the attached 13C) and achieve a higher S=N. In HMBC (Figure 13.3l), insertion of a second proton pulse and time delay after the ﬁrst 1=2JCH delay effectively ﬁlter out most of the magnetization owing to C-H fragments, since the experiment is designed to select only cross-peaks derived from H–C–C–H or longer fragments. As the values for 1JCH and 2,3JCH are on the average 150 and 8 Hz, respectively, only 3.3 ms are needed for producing an anti-phase term because of vicinal coupling and more than 60 ms for the one term yielded by long-range C-H coupling. Therefore, after 3.3 ms the undesired term is transformed into MQ terms and abandoned. Magnetization owing to long-range coupling then follows the pathway of transformations described for HMQC and is eventually detected as an anti-phase term (i.e., without the ﬁnal 1=2JCH delay to avoid T2 losses). Carbon decoupling is usually not used so as to be able to recognize peaks produced by genuine long-range coupling from those produced by 1JCH, which may have escaped ﬁltering. The HSQC sequence (Figure 13.3m) is only apparently more complex than its HMQC counterpart. The sequence features a building block known under the name ‘‘INEPT,’’ the mixing pulses

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 309 17.7.2008 5:39pm Compositor Name: VAmoudavally

309

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

(i.e., the second 908 proton pulse and the ﬁrst 908 carbon pulse), the evolution time t1 (with a proton 1808 pulse in the middle), and a couple of 908 pulses. At the end, a reverse-INEPT allows detection of proton coherent magnetization. Choosing between HMQC and HSQC requires consideration of a number of points. First, HMQC should allow higher resolution in F1 (the frequency domain obtained by FT of t1) but, unfortunately, line shapes in F1 are not singlet, and practically, all HMQC peaks will be broader than HSQC. On the other side, because of the higher number of pulses, HSQC is usually more sensitive to radiofrequency inhomogeneity. Overall, one should use HSQC for crowded spectra and HMQC for noncrowded and more dilute spectra. 13.6.6.2

Applications in Food Science

Usually, heteronuclear 2D spectroscopy is adopted in food science to univocally identify individual molecular species that are present in complex mixtures, owing to much less overlap than in the 1D 1 H-NMR spectra. Many saccharidic compounds are clearly evidenced by the 2D-HSQC map, even though their concentrations are quite low [75]. Figure 13.10 shows many cross-peaks inside the 1H anomeric region (4.4–5.7 ppm) and the corresponding 13C resonances (90–110 ppm). By looking at the HSQC spectrum of whole milk, and through consultation of the databases, the 1 H resonance at 2.68 ppm, showing a cross-peak with the 13C resonance at 25.52 ppm, is assigned to the methylene group next to the double bond of polyunsaturated acyl chains. This signal was concluded to be caused by the linoleate and linolenate chains, because they are the main polyunsaturated fatty acids in milk [86]. HMBC spectra of whole milk could not be obtained, since signals are broadened by long correlation times of fat globules and protein complexes. However, HMBC spectra are useful to assign minor signals of milk samples obtained by removing the fatty acid layer by centrifugation and by denaturating the proteins by addition of twofold volumes of ethanol to skim milk [86]. For the 2DNMR experiments, the total measurement time required for each spectrum is approximately 7–10 h. In the 1H-13C HMBC spectrum of skim milk, the 1H signals at 2.40 and 2.54 ppm show a correlation

ppm 13C

90 b a

c

95 d 100

105

5.5

5.0

4.5

1H

ppm

FIGURE 13.10 Anomeric spectral region of the 1H-13C HSQC map of the aqueous trufﬂe extract. Each crosspeak corresponds to different sugars, the most intense ones being assigned 75 to (a) uridine-50 -(diphospho-Nacetylglucosamine), (b) b-glucose, (c) D-trehalose, and (d) a-glucose. The top trace shows the same region in the 1D 1H spectrum of the extract.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 310 17.7.2008 5:39pm Compositor Name: VAmoudavally

310

Handbook of Food Analysis Instruments

with the 13C signals at 45.15, 77.35, 179.97, and 182.50 ppm [86]. These data indicate that all these signals arise from ionized citric acid, as judged by the chemical structure of citric acid. In the same way, other signals have been assigned to creatine. The 1H resonance at 2.88 ppm was correlated to the 13C signals at 55 and 158 ppm in the 1H-13C HMBC spectrum and to the 15N signal at 78 ppm in the 1H-15N HMBC spectrum [86]. Another 1H signal at 3.79 ppm was correlated to the 13C signals at 38, 158, and 176 ppm in the same spectrum. The chemical structure of creatine is compatible with chemical shift and the cross-peaks pattern present in the spectrum. The noisy and insensitive 2D-INADEQUATE experiment is the main method suitable to obtain a full 13C spectral assignment of puriﬁed molecules necessary to their unambiguous identiﬁcation. Such a spectrum has been successfully employed to identify an unknown lipid component extracted from avocado pears [148]. Figure 13.9 shows the 2D map with 13C–13C scalar correlation peaks useful to trace all the carbon molecular backbone [148]. However, because of the intrinsic scarce sensitivity of the technique, large amount of pure sample was required (100 mg at 200 MHz). It is worth noting here that suitable selective experiments (e.g., 1D-INADEQUATE) are derived from 2D analogues, allowing the identiﬁcation of connectivities between selected 13C resonances, thus concentrating all the experimental time to solve only the ambiguous assignments. This kind of experiment has been successfully applied to obtain the full assignment of signals in the triolein, a necessary step for the acyl positional distribution of glycerol tri-esters in vegetable oils [95].

13.6.7 PULSE FIELD GRADIENT NUCLEAR MAGNETIC RESONANCE SPECTROSCOPY 13.6.7.1

Pulse Sequence

Although perfect refocusing of magnetization defocused by the effect of magnetic ﬁeld inhomogeneity was claimed when describing the spin echo sequence in Section 13.5.2.1, this condition is hardly met in liquids. In fact, essential condition for perfect refocusing is that all nuclei in the sample retain the resonance frequency during the whole course of the experiment. In liquids however, random motions take place at the molecular level and the probability that a nucleus moves Dx in a time Dt is related to its self-diffusion coefﬁcient D. If the effective B0 at the new nucleus position is different from the one it sensed when the ﬁrst 908 pulse was delivered to the sample, there will be a resonance frequency change during the experiment time. In such a situation, the intensity of the echo at time t will be exponentially lower than expected because some nuclei will be out of phase at the end of time t, according to Equation 13.18: I=I0 / exp [Dg2 g2 g2 (D d=3)]

(13:18)

where I=I0 is the signal attenuation D is the self-diffusion coefﬁcient in m2=s g is the gyromagnetic ratio of the observed nucleus (rad=s T ) g is the gradient strength (T ) d is the gradient length (s) The NMR-diffusion time scale is deﬁned by a diffusion delay D, which generally ranges from milliseconds to fractions of a second, so that translational motion of some hundred micrometers is probed [149]. This phenomenon is exploited in a branch of NMR known under the name of pulse ﬁeld gradient NMR (PFG-NMR). PFG-NMR combines NMR information about components of a mixture with

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 311 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

311

their diffusion coefﬁcients. The data processing used in DOSY-NMR generates a 2D plot with a chemical shift scale in one dimension and diffusion coefﬁcient values in the second dimension. The latter is built up by generating a number (i.e., 30–50) of experiments recorded with gradient intensities linearly sampled from 5% to 95% of the maximum obtainable intensity (47.5 G=cm). In the STEDOSY experiment, a stimulated echo (STE) packet is added to the pulse sequence to avoid T2 relaxation effects by storing the magnetization along the Z-axis during evolution, so that relaxation depends primarily on T1, which is usually much longer than T2 for macromolecules [150]. 13.6.7.2

Applications in Food Science

Application of STE-DOSY spectra was used to give information both on diffusion in food matrices and on interactions between ﬂavor molecules and food matrix ingredients [149]. Flavor diffusivity is described as the diffusive mass transfer, and translational motion is the most fundamental form of transport, closely related to molecular size (Stokes–Einstein equation). This diffusive process is called self-diffusion and reﬂects the random translational motion of molecules driven by internal kinetic energy. Microscopic displacements, which range from 1012 to 104 m, covered by a solute per second, are directly estimated using the DOSY method. The self-diffusion coefﬁcient of ethyl butyrate was dramatically decreased in the model fruit preparation by 85%, mainly because of high sucrose content [149]. Macromolecules, such as starch and carrageenans, have a signiﬁcant but much less important effect on the diffusion of ethyl butyrate. No signiﬁcant difference between the chemical shifts of aroma peaks in the different media was experienced. Moreover, in the 35% sucrose solution, the self-diffusion coefﬁcients of ﬁve aroma compounds (ethyl acetate, ethyl hexanoate, linalool, hexanal, and ethyl butyrate) were all signiﬁcantly lower (by about 70%) than those in D2O solution. Thus, there is no speciﬁc diffusion process dependent on the nature of the aroma compound in the presence of sucrose molecules, conﬁrming that no speciﬁc molecular interaction occurs. In fact, aroma self-diffusion is highly related to the mobility of water molecules. As stated by the Stokes–Einstein equation, the threefold decrease in the self-diffusion coefﬁcient from D2O solution to 35% sucrose solution could be partly explained by the increase in medium viscosity from 0.98 to 4.12 mPa=s. DOSY experiments are able to follow enzymatic degradation of polysaccharides by neatly distinguishing among signals derived from mono-, oligo-, or polysaccharides. The technique has been successfully used to exclude signiﬁcant contribution of b-reducing units of low molecular weight maltodextrins to the anomeric signal of maltose, even if they are still embedded in the polysaccharidic matrix. Actually, because of restricted diffusion, no well-resolved DOSY maps are obtainable by HR-MAS in soft matter; in fact, restricted diffusion provides apparent diffusion coefﬁcients, whose values are determined by the size and the shape of the barrier rather than by the shape of the diffusing molecules. Thus, kinetic release of oligosaccharides in hydrated ﬂours is monitored by DOSY experiments in suspensions only by applying modiﬁed sequence [151]. The comparison between the diffusion coefﬁcients of xylose (7 1010 m=s2) with that one of an unknown compound (7 1011 m=s2) that is supposed to be a xylan, together with other considerations emerging from signal assignment and integral measurements of the anomeric 1H resonances, permitted to identify this polysaccharide as lentinan [152].

REFERENCES 1. Abragam, A., The Principles of Nuclear Magnetism, 3rd ed., Clarendon Press: Oxford, United Kingdom, 1983. 2. Kowalewski, J. and Maler, L., Nuclear Spin Relaxation in Liquids: Theory, Experiments, and Applications, Vol. 2, Taylor & Francis: New York, 2006. 3. Berger, S. and Braun, S., 200 and More NMR Experiments: A Practical Course, 1st ed., Wiley-VCH: Weinheim, Germany, 2004. 4. Gullion, T., Introduction to High-Resolution Solid-State NMR, Elsevier: Oxford, United Kingdom, 2007.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 312 17.7.2008 5:39pm Compositor Name: VAmoudavally

312

Handbook of Food Analysis Instruments

5. Bax, A., Two-Dimensional Nuclear Magnetic Resonance in Liquids, Springer-Verlag: Berlin, Germany, 2001. 6. Harris, R.K., Becker, E.D., Cabral de Menezes, S., Goodfellow, R., and Grangere, P., Provisional recommendations for NMR nomenclature: Nuclear spin properties and conventions for chemical shifts, Pure Appl. Chem., 73(11), 1795–1818, 2001. 7. Bloembergen, N., Purcell, E.M., and Pound, R.V., Relaxation effects in nuclear magnetic resonance absorption, Phys. Rev., 73, 679–712, 1948. 8. Hinrichs, R., Götz, J., Noll, M., Wolfschoon, A., Eibel, H., and Weisser, H., Characterisation of the water-holding capacity of fresh cheese samples by means of low resolution nuclear magnetic resonance, Food Res. Int., 37(7), 667–676, 2004. 9. Kimmich, R. and Anoardo, E., Field-cycling NMR relaxometry, Progr. Nucl. Magn. Reson. Spectrosc., 44(3–4), 257–320, 2004. 10. Halle, B., Johannesson, H., and Venu, K., Model-free analysis of stretched relaxation dispersions, J. Magn. Reson., 135(1), 1–13, 1998. 11. Brosio, E., Conti, F., Di Nola, A., and Sykora, S., A pulsed low resolution NMR study on crystallization and melting processes of cocoa butter, J. Am. Oil Chem. Soc., 57(2), 78–82, 1980. 12. Fjelkner-Modig, S. and Tornberg, E., Water distribution in porcine M. longissimus dorsi in relation to sensory properties, Meat Sci., 17(3), 213–231, 1986. 13. Tellier, C., Mariette, F., Guillement, J.P., and Marchal, P., Evolution of water proton nuclear magnetic relaxation during milk coagulation and syneresis: Structural implications, J. Agric. Food Chem., 41(12), 2259–2266, 1993. 14. Lambelet, P., Renevey, F., Kaabi, C., and Raemy, A., Low-ﬁeld nuclear magnetic resonance relaxation study of stored or processed cod, J. Agric. Food Chem., 43(6), 1462–1466, 1995. 15. Tellier, C., Guillou-Charpin, M., Grenier, P., and Le Botlan, D., Monitoring alcoholic fermentation by low-resolution pulsed nuclear magnetic resonance, J. Agric. Food Chem., 37(4), 988–991, 1989. 16. Eidmann, G., Savelsberg, R., Bluemler, P., and Bluemich, B., The NMR MOUSE, a mobile universal surface explorer, J. Magn. Reson. A, 122(1), 104–109, 1996. 17. Hills, B.P., Wright, K.M., and Gillies, D.G., A low-ﬁeld, low-cost Halbach magnet array for open-access NMR, J. Magn. Reson., 175(2), 336–339, 2005. 18. Capozzi, F., Cremonini, M.A., Luchinat, C., Placucci, G., and Vignali, C., A low frequency 1H-NMR external unit for the analysis of large foodstuff samples, J. Magn. Reson., 138(2), 277–280, 1999. 19. Appelt, S., Kuehn, H., Haesing, F.W., and Bluemich, B., Chemical analysis by ultrahigh-resolution nuclear magnetic resonance in the Earth’s magnetic ﬁeld, Nat. Phys., 2(2), 105–109, 2006. 20. Meriles, C.A., Sakellariou, D., Heise, H., Moule, A.J., and Pines, A., Approach to high-resolution ex situ NMR spectroscopy, Science, 293(5527), 82–85, 2001. 21. Stork, H., Gadke, A., and Nestle, N., Single-sided and semisingle-sided NMR sensors for highly diffusive samples: Application to bottled beverages, J. Agric. Food Chem., 54(15), 5247–5252, 2006. 22. Venkataramanan, L., Yi-Qiao, S., and Hurlimann, M.D., Solving Fredholm integrals of the ﬁrst kind with tensor product structure in 2 and 2.5 dimensions, IEEE Trans. Acoust. Speech, 50(5), 1017–1026, 2002. 23. Borgia, G.C., Brown, R.J., and Fantazzini, P., Uniform-penalty inversion of multiexponential decay data, J. Magn. Reson., 132(1), 65–77, 1998. 24. Chiarotti, G., Cristiani, G., and Biulotto, L., Proton relaxation in pure liquids and in liquids containing paramagnetic gases, Nuovo Cimento, 1, 863–873, 1955. 25. Bertini, I. and Luchinat, C., NMR of Paramagnetic Molecules in Biological Systems, Vol. 3, Benjamin= Cummings Publ.: Menlo Park, CA, 1986. 26. Pennington, J.A.T., Schoen, S.A., Salmon, G.D., Young, B., Johnson, R.D., and Marts, R.W., Composition of core foods of the U.S. food supply, 1982–1991. II. Calcium, magnesium, iron, and zinc, J. Food Composition Anal., 8(2), 129–169, 1995. 27. Pennington, J.A.T., Schoen, S.A., Salmon, G.D., Young, B., Johnson, R.D., and Marts, R.W., Composition of core foods of the U.S. food supply, 1982–1991. III. Copper, manganese, selenium, and iodine, J. Food Composition Anal., 8(2), 171–217, 1995. 28. Cabrera-Vique, C., Teissedre, P.L., Cabanis, M.T., and Cabanis, J.C., Manganese determination in grapes and wines from different regions of France, Am. J. Enol. Vitic., 51(2), 103–107, 2000.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 313 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

313

29. Laghi, L., Cremonini, M.A., Placucci, G., Sykora, S., Wright, K., and Hills, B., A proton NMR relaxation study of hen egg quality, Magn. Reson. Imaging, 23(3), 501–510, 2005. 30. Goldsmith, S.M. and Toledo, R.T., Studies on egg albumin gelation using nuclear magnetic resonance, J. Food Sci., 50(1), 59–62, 1985. 31. Alessandri, S., Capozzi, F., Cremonini, M.A., Luchinat, C., Placucci, G., Savorani, F., and Turano, M., Paramagnetic challenges in NMR measurements of foods, in Magnetic Resonance in Food Science: The Multivariate Challenge, Engelsen, S.B., Belton, P.S., Jakobsen, H.J. (Eds.), Royal Society of Chemistry: Cambridge, United Kingdom, 2005, pp. 113–123. 32. Banci, L., Bertini, I., Capozzi, F., and Luchinat, C., The electron nucleus coupling—A breakthrough in the investigation of paramagnetic metalloproteins, Int. J. Quantum Chem., 42(5), 1383–1396, 1992. 33. Bertini, I., Luchinat, C., and Parigi, G., Solution NMR of Paramagnetic Molecules: Applications to Metallobiomolecules and Models, Vol. 2, Elsevier: Amsterdam, the Netherlands, 2001. 34. Bertini, I., Capozzi, F., and Luchinat, C., Water proton relaxation rate enhancements as a function of magnetic ﬁeld strength and nature and size of paramagnetic solutes, Magn. Reson. Imaging, 9(5), 849–853, 1991. 35. Ferrante, G. and Sykora, S., Technical aspects of fast ﬁeld cycling, Adv. Inorg. Chem., 57, 405–470, 2005. 36. Bertini, I., Capozzi, F., Luchinat, C., and Xia, Z., Nuclear and electron relaxation of hexaaquairon(3þ), J. Phys. Chem.—US, 97(6), 1134–1137, 1993. 37. Capozzi, F., Luchinat, C., Bettini, A., Meluzzi, A., and Franchini, A., Shell egg freshness evaluated by nuclear magnetic resonance spectroscopy of albumen, Kijowski, J., Pikul, J. (Eds.), Proceedings 7th WPS A European Fedaration Symposium on the Quality of Eggs and Egg Products, Graf-Com: Poznan, Poland, 1997, pp. 198–207. 38. Lioutas, T.S., Baianu, I.C., Bechtel, P.J., and Steinberg, M.P., Oxygen-17 and sodium-23 nuclear magnetic resonance studies of myoﬁbrillar protein interactions with water and electrolytes in relation to sorption isotherms, J. Agric. Food Chem., 36(3), 437–444, 1988. 39. Fanni, J., Canet, D., Elbayed, K., and Hardy, J., Proton and sodium-22 NMR relaxation studies of the sodium chloride=b-lactoglobulin system equilibrated at various water activities, J. Food Sci., 54(4), 909–916, 1989. 40. Rosett, T.R., Shirley, L., Schmidt, S.J., and Klein, B.P., Sodium ion binding, as measured by sodium-23 nuclear magnetic resonance spectroscopy, inﬂuences the perception of saltiness in gum solutions, J. Food Sci., 59(1), 206–210, 1994. 41. Nagata, T., Chuda, Y., Yan, X., Suzuki, M., and Kawasaki, K., The state analysis of NaCl in snow crab (Chionoecetes japonicus) meat examined by 23Na and 35Cl nuclear magnetic resonance (NMR) spectroscopy, J. Sci. Food Agric., 80(8), 1151–1154, 2000. 42. Cremonini, M.A., Venturi, L., Sykora, S., Cavani, C., and Placucci, G., Reference convolution of lowresolution FIDs in the presence of poor magnetic ﬁeld homogeneity, Magn. Reson. Imaging, 25, 555, 2007. 43. Meiboom, S. and Gill, D., Modiﬁed spin-echo method for measuring nuclear relaxation times, Rev. Sci. Instrum., 29, 688–691, 1958. 44. Carr, H.Y. and Purcell, E.M., Effects of diffusion on free precession in nuclear magnetic resonance experiments, Phys. Rev., 94, 630–638, 1954. 45. Hills, B.P., Takacs, S.F., and Belton, P.S., The effects of proteins on the proton NMR transverse relaxation times of water. I. Native bovine serum albumin, Mol. Phys., 67(4), 903–918, 1989. 46. Renou, J.P., NMR studies in meat, Annu. Rep. NMR Spectrosc., 31, 313–344, 1995. 47. Mortensen, M., Andersen, H.J., Engelsen, S.B., and Bertram, H.C., Effect of freezing temperature, thawing and cooking rate on water distribution in two pork qualities, Meat Sci., 72(1), 34–42, 2006. 48. Micklander, E., Bertram, H.C., Marnø, H., Bak, L.S., Andersen, H.J., Engelsen, S.B., and Nørgaard, L., Early post-mortem discrimination of water-holding capacity in pig longissimus muscle using new ultrasound method, Lebensmittel-Wiss. Technol., 38(5), 437–445, 2005. 49. Bertram, H.C., Schafer, A., Rosenvold, K., and Andersen, H.J., Physical changes of signiﬁcance for early post mortem water distribution in porcine M. longissimus, Meat Sci., 66(4), 915–924, 2004. 50. Bertram, H.C., Engelsen, S.B., Busk, H., Karlsson, A.H., and Andersen, H.J., Water properties during cooking of pork studied by low-ﬁeld NMR relaxation: Effects of curing and the RN-gene, Meat Sci., 66 (2), 437–446, 2004. 51. Bertram, H.C., Kristensen, M., and Andersen, H.J., Functionality of myoﬁbrillar proteins as affected by pH, ionic strength and heat treatment—A low-ﬁeld NMR study, Meat Sci., 68(2), 249–256, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 314 17.7.2008 5:39pm Compositor Name: VAmoudavally

314

Handbook of Food Analysis Instruments

52. Lee, S.-M. and Rho, J.-H., Water holding capacity and NMR relaxation pattern of meats thawed by different methods, Food Sci. Biotechnol., 12(4), 321–325, 2003. 53. Bertram, H.C., Andersen, H.J., Karlsson, A.H., Horn, P., Hedegaard, J., Norgaard, L., and Engelsen, S.B., Prediction of technological quality (cooking loss and napole yield) of pork based on fresh meat characteristics, Meat Sci., 65(2), 707–712, 2003. 54. Bianchi, M., Capozzi, F., Cremonini, M.A., Laghi, L., Petracci, M., Placucci, G., and Cavani, C., Inﬂuence of the season on the relationships between NMR transverse relaxation data and water-holding capacity of turkey breast meat, J. Sci. Food Agric., 84(12), 1535–1540, 2004. 55. Brondum, J., Munck, L., Henckel, P., Karlsson, A., Tornberg, E., and Engelsen, S.B., Prediction of waterholding capacity and composition of porcine meat by comparative spectroscopy, Meat Sci., 55(2), 2000. 56. Brown, R.J.S., Capozzi, F., Cavani, C., Cremonini, M.A., Petracci, M., and Placucci, G., Relationships between 1H NMR relaxation data and some technological parameters of meat: A chemometric approach, J. Magn. Reson., 147(1), 89–94, 2000. 57. Bertram, H.C., Karlsson, A.H., Rasmussen, M., Pedersen, O.D., Donstrup, S., and Andersen, H.J., Origin of multiexponential T2 relaxation in muscle myowater, J. Agric. Food Chem., 49(6), 3092– 3100, 2001. 58. Bertram, H.C., Dønstrup, S., Karlsson, A.H., and Andersen, H.J., Continuous distribution analysis of T2 relaxation in meat—An approach in the determination of water-holding capacity, Meat Sci., 60(3), 279– 285, 2002. 59. Brown, R.J.S., Information available and unavailable from multiexponential relaxation data, J. Magn. Reson., 82, 539–561, 1989. 60. Pedersen, H.T., Bro, R., and Engelsen, S.B., Towards rapid and unique curve resolution of low-ﬁeld NMR relaxation data: Trilinear SLICING versus two-dimensional curve ﬁtting, J. Magn. Reson., 157(1), 141–155, 2002. 61. Ablett, S., Clarke, C.J., Izzard, M.J., and Martin, D.R., Relationship between ice recrystallization rates and the glass transition in frozen sugar solutions, J. Sci. Food Agric., 82(15), 1855–1859, 2002. 62. Chung, H.-J. and Lim, S.-T., Physical aging of glassy normal and waxy rice starches: Effect of aging temperature on glass transition and enthalpy relaxation, Carbohydr. Polym., 53(2), 205–211, 2003. 63. Furuki, T., Effect of molecular structure on thermodynamic properties of carbohydrates. A calorimetric study of aqueous di- and oligosaccharides at subzero temperatures, Carbohydr. Res., 337(5), 441–450, 2002. 64. Adhikari, B., Howes, T., Bhandari, B.R., and Truong, V., Stickiness in foods: A review of mechanisms and test methods, Int. J. Food Prop., 4(1), 1–33, 2001. 65. Tolstoguzov, V.B., The importance of glassy biopolymer components in food, Nahrung, 44(2), 76–84, 2000. 66. Sartor, G. and Johari, G.P., Structural relaxation of a vitriﬁed high-protein food, beef, and the phase transformations of its water content, J. Phys. Chem.—US, 100(24), 10450–10463, 1996. 67. Roos, Y.H., Karel, M., and Kokini, J.L., Glass transitions in low moisture and frozen foods: Effects on shelf life and quality, Food Technol., 50(11), 95–108, 1996. 68. Bruno, T.J. and Svoronos, P.D.N., Handbook of Basic Tables for Chemical Analysis, 2nd ed., CRC Press: Boca Raton, FL, 2003. 69. Silverstein, R.M., Webster, F.X., and Kiemle, D., Spectrometric Identiﬁcation of Organic Compounds, 7th ed., John Wiley & Sons, Ltd.: New York, 2005. 70. Hu, J., Xia, Y., Feng, W., Xuan, Y., Shen, Y., Haacke, E.M., and Jiang, Q., Orientational dependence of trimethyl ammonium signal in human muscles by 1H magnetic resonance spectroscopic imaging, Magn. Reson. Imaging, 23(1), 97–104, 2005. 71. Vermathen, P., Boesch, C., and Kreis, R., Mapping ﬁber orientation in human muscle by proton MR spectroscopic imaging, Magn. Reson. Med., 49(3), 424–432, 2003. 72. Hwang, T.L. and Shaka, A.J., Water suppression that works. Excitation sculpting using arbitrary waveforms and pulsed ﬁeld gradients, J. Magn. Reson. A, 112(2), 275–279, 1995. 73. Fan, T.W., Metabolite proﬁling by one- and two-dimensional NMR analysis of complex mixtures, Progr. Nucl. Magn. Reson. Spectrosc., 28(2), 161–219, 1996. 74. Croasmun, W.R. and Carlson, R.M.K., Steroid structural analysis by two-dimensional NMR, in TwoDimensional NMR Spectroscopy: Application for Chemists and Biochemists, Croasmun, W.R., Carlson, R.M.K. (Eds.), 2nd ed., VCH Publ.: New York, 1994, pp. 785–832.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 315 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

315

75. Mannina, L., Cristinzio, M., Sobolev, A.P., Ragni, P., and Segre, A., High-ﬁeld nuclear magnetic resonance (NMR) study of trufﬂes (Tuber aestivum vittadini), J. Agric. Food Chem., 52(26), 7988–7996, 2004. 76. Nielsen, L.V., Studies on the relation between unsaturation and iodine value of butterfat by high resolution nuclear magnetic resonance (NMR), Milchwiss., 31(10), 598–602, 1976. 77. Aursand, M. and Grasdalen, H., Interpretation of the carbon-13-NMR spectra of omega-3 fatty acids and lipid extracted from the white muscle of Atlantic salmon (Salmo salar), Chem. Phys. Lipids, 62(3), 239–251, 1992. 78. Aursand, M., Rainuzzo, J.R., and Grasdalen, H., Quantitative high-resolution carbon-13 and proton nuclear magnetic resonance of v3 fatty acids from white muscle of Atlantic salmon (Salmo salar), J. Am. Oil Chem. Soc., 70(10), 971–981, 1993. 79. Sacchi, R., Medina, I., Aubourg, S.P., Addeo, F., and Paolillo, L., Proton nuclear magnetic resonance rapid and structure-speciﬁc determination of v-3 polyunsaturated fatty acids in ﬁsh lipids, J. Am. Oil Chem. Soc., 70(3), 225–228, 1993. 80. Igarashi, T., Aursand, M., Hirata, Y., Gribbestad, I.S., Wada, S., and Nonaka, M., Nondestructive quantitative determination of docosahexaenoic acid and n-3 fatty acids in ﬁsh oils by high-resolution 1H nuclear magnetic resonance spectroscopy, J. Am. Oil Chem. Soc., 77(7), 737–748, 2000. 81. Miyake, Y., Yokomizo, K., and Matsuzaki, N., Determination of unsaturated fatty acid composition by high-resolution nuclear magnetic resonance spectroscopy, J. Am. Oil Chem. Soc., 75(9), 1091–1094, 1998. 82. Miyake, Y., Yokomizo, K., and Matsuzaki, N., Rapid determination of iodine value by 1H nuclear magnetic resonance spectroscopy, J. Am. Oil Chem. Soc., 75(1), 15–19, 1998. 83. Miyake, Y., Yokomizi, K., and Matsuzaki, N., Rapid determination of iodine value by 1H nuclear magnetic resonance spectroscopy [Erratum to document cited in CA128:153349], J. Am. Oil Chem. Soc., 75(2), 321, 1998. 84. Guillen, M.D. and Ruiz, A., Edible oils: Discrimination by 1H nuclear magnetic resonance, J. Sci. Food Agric., 83(4), 338–346, 2003. 85. Segre, A. and Mannina, L., H-NMR study of edible oils, Recent Res. Dev. Oil Chem., 1, 297–308, 1997. 86. Hu, F., Furihata, K., Ito-Ishida, M., Kaminogawa, S., and Tanokura, M., Nondestructive observation of bovine milk by NMR spectroscopy: Analysis of existing states of compounds and detection of new compounds, J. Agric. Food Chem., 52(16), 4969–4974, 2004. 87. Jacobs, J.W.M., Van Os, J.W.M., and Veeman, W.S., Broadband heteronuclear decoupling, J. Magn. Reson., 51(1), 56–66, 1983. 88. Allegrozzi, M., Bertini, I., Janik, M.B.L., Lee, Y.M., Liu, G., and Luchinat, C., Lanthanide-induced pseudocontact shifts for solution structure reﬁnements of macromolecules in shells up to 40 Å from the metal ion, J. Am. Chem. Soc., 122(17), 4154–4161, 2000. 89. Bertini, I. and Luchinat, C., The hyperﬁne shift, Coord. Chem. Rev., 150, 29–75, 1996. 90. Fujiwara, T. and Nagayama, K., Composite inversion pulses with frequency switching and their application to broadband decoupling, J. Magn. Reson., 77(1), 53–63, 1988. 91. Shaka, A.J., Keeler, J., Frenkiel, T., and Freeman, R., An improved sequence for broadband decoupling: WALTZ-16, J. Magn. Reson., 52(2), 335–338, 1983. 92. Shaka, A.J., Barker, P.B., and Freeman, R., Computer-optimized decoupling scheme for wideband applications and low-level operation, J. Magn. Reson., 64(3), 547–552, 1985. 93. Shaka, A.J. and Keeler, J., Broadband spin decoupling in isotropic liquids, Progr. Nucl. Magn. Reson. Spectrosc., 19(1), 47–129, 1987. 94. Sacchi, R., Addeo, F., Giudicianni, I., and Paolillo, L., Analysis of the positional distribution of fatty acids in olive oil triacylglycerols by high resolution 13C-NMR of the carbonyl region, Ital. J. Food Sci., 4 (2), 117–123, 1992. 95. Mannina, L., Luchinat, C., Emanuele, M.C., and Segre, A., Acyl positional distribution of glycerol triesters in vegetable oils: A 13C NMR study, Chem. Phys. Lipids, 103(1–2), 47–55, 1999. 96. Ng, S., Analysis of positional distribution of fatty acids in palmoil by 13C NMR spectroscopy, Lipids, 20 (11), 778–782, 1985. 97. Wollenberg, K.F., Quantitative high resolution carbon-13 nuclear magnetic resonance of the oleﬁnic and carbonyl carbons of edible vegetable oils, J. Am. Oil Chem. Soc., 67(8), 487–494, 1990. 98. Folch, J., Lees, M., and Sloane Stanley, G.H., A simple method for the isolation and puriﬁcation of total lipides from animal tissues, J. Biol. Chem., 226(1), 497–509, 1957.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 316 17.7.2008 5:39pm Compositor Name: VAmoudavally

316

Handbook of Food Analysis Instruments

99. Cremonini, M.A., Laghi, L., and Placucci, G., Investigation of commercial lecithin by 31P NMR in a ternary CUBO solvent, J. Sci. Food Agric., 84(8), 786–790, 2004. 100. Pearce, J.M. and Komoroski, R.A., Resolution of phospholipid molecular species by 31P NMR, Magn. Reson. Med., 29(6), 724–731, 1993. 101. Meneses, P. and Glonek, T., High resolution phosphorus-31 NMR of extracted phospholipids, J. Lipid Res., 29(5), 679–689, 1988. 102. Li, C.P., Salvador, A.S., Ibrahim, H.R., Sugimoto, Y., and Aoki, T., Phosphorylation of egg white proteins by dry-heating in the presence of phosphate, J. Agric. Food Chem., 51(23), 6808–6815, 2003. 103. Hubbard, C.D., Caswell, D., Ludemann, H.D., and Arnold, M., Characterisation of pressure-treated skimmed milk powder dispersions: Application of NMR spectroscopy, J. Sci. Food Agric., 82(10), 1107–1114, 2002. 104. Martin, G.J., Martin, M.L., Mabon, F., and Michon, M.J., A new method for the identiﬁcation of the origin of ethanols in grain and fruit spirits: High-ﬁeld quantitative deuterium nuclear magnetic resonance at the natural abundance level, J. Agric. Food Chem., 31(2), 311–315, 1983. 105. Martin, G.J., Benbernou, M., and Lantier, F., Application of site-speciﬁc natural isotope fractionation (SNIF-NMR) of hydrogen to the characterization of European beers, J. Inst. Brewing, 91(4), 242–249, 1985. 106. Martin, G.J., Guillou, C., Martin, M.L., Cabanis, M.T., Tep, Y., and Aerny, J., Natural factors of isotype fractionation and the characterization of wines, J. Agric. Food Chem., 36(2), 316–322, 1988. 107. Zhang, B.-L., Billault, I., Li, X., Mabon, F., Remaud, G., and Martin, M.L., Hydrogen isotopic proﬁle in the characterization of sugars. Inﬂuence of the metabolic pathway, J. Agric. Food Chem., 50(6), 1574–1580, 2002. 108. Pupin, A.M., Dennis, M.J., Parker, I., Kelly, S., Bigwood, T., and Toledo, M.C.F., Use of isotopic analyses to determine the authenticity of Brazilian orange juice (Citrus sinensis), J. Agric. Food Chem., 46(4), 1369–1373, 1998. 109. Martin, G.G., Hanote, V., Lees, M., and Martin, Y.L., Interpretation of combined 2H SNIF=NMR and 13C SIRA=MS analyses of fruit juices to detect added sugar, J. AOAC Int., 79(1), 62–72, 1996. 110. Cotte, J.F., Casabianca, H., Lheritier, J., Perrucchietti, C., Sanglar, C., Waton, H., and Grenier-Loustalot, M.F., Study and validity of 13C stable carbon isotopic ratio analysis by mass spectrometry and 2H sitespeciﬁc natural isotopic fractionation by nuclear magnetic resonance isotopic measurements to characterize and control the authenticity of honey, Anal. Chim. Acta, 582(1), 125–136, 2007. 111. Giraudon, S., Danzart, M., and Merle, M.H., Deuterium nuclear magnetic resonance spectroscopy and stable carbon isotope ratio analysis=mass spectrometry of certain monoﬂoral honeys, J. AOAC Int., 83(6), 1401–1409, 2000. 112. Lindner, P., Bermann, E., and Gamarnik, B., Characterization of citrus honey by deuterium NMR, J. Agric. Food Chem., 44(1), 139–140, 1996. 113. Royer, A., Naulet, N., Mabon, F., Lees, M., and Martin, G.J., Stable isotope characterization of olive oils. II. Deuterium distribution in fatty acids studied by nuclear magnetic resonance, J. Am. Oil Chem. Soc., 76 (3), 365–373, 1999. 114. Ogrinc, N., Kosir, I.J., Spangenberg, J.E., and Kidric, J., The application of NMR and MS methods for detection of adulteration of wine, fruit juices, and olive oil. A review, Anal. Bioanal. Chem., 376(4), 424–430, 2003. 115. Renou, J.P., Bielicki, G., Deponge, C., Gachon, P., Micol, D., and Ritz, P., Characterization of animal products according to geographic origin and feeding diet using nuclear magnetic resonance and isotope ratio mass spectrometry. Part II: Beef meat, Food Chem., 86(2), 251–256, 2004. 116. Tenailleau, E., Lancelin, P., Robins, R.J., and Akoka, S., NMR approach to the quantiﬁcation of nonstatistical 13C distribution in natural products: Vanillin, Anal. Chem., 76(13), 3818–3825, 2004. 117. Laws, D.D., Bitter Hans-Marcus, L., and Jerschow, A., Solid-state NMR spectroscopic methods in chemistry, Angew. Chem. Int. Ed. Engl., 41(17), 3096–3129, 2002. 118. Hartmann, S.R. and Hahn, E.L., Nuclear double resonance in the rotating frame, Phys. Rev., 128(5), 2042, 1962. 119. Maas, W.E., Laukien, F.H., and Cory, D.G., Gradient, high resolution, magic angle sample spinning NMR, J. Am. Chem. Soc., 118(51), 13085–13086, 1996. 120. Maas, W.E., Bielecki, A., Ziliox, M., Laukien, F.H., and Cory, D.G., Magnetic ﬁeld gradients in solid state magic angle spinning NMR, J. Magn. Reson., 141(1), 29–33, 1999.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 317 17.7.2008 5:39pm Compositor Name: VAmoudavally

Nuclear Magnetic Resonance Spectroscopy in Food Analysis

317

121. Pan, J.W., Hamm, J.R., Rothman, D.L., and Shulman, R.G., Intracellular pH in human skeletal muscle by 1H NMR, Proc. Natl. Acad. Sci. USA, 85(21), 7836–7839, 1988. 122. Wind, R.A., Hu, J.Z., and Rommereim, D.N., High-resolution (1)H NMR spectroscopy in organs and tissues using slow magic angle spinning, Magn. Reson. Med., 46(2), 213–218, 2001. 123. Bertram, H.C., Whittaker, A.K., Andersen, H.J., and Karlsson, A.H., The use of simultaneous 1H & 31P magic angle spinning nuclear magnetic resonance measurements to characterize energy metabolism during the conversion of muscle to meat, Int. J. Food Sci. Tech., 39(6), 661–670, 2004. 124. Bielecki, A. and Burum, D.P., Temperature dependence of 207Pb MAS spectra of solid lead nitrate. An accurate, sensitive thermometer for variable-temperature MAS, J. Magn. Reson. A, 116(2), 215–220, 1995. 125. Shintu, L. and Caldarelli, S., High-resolution MAS NMR and chemometrics: Characterization of the ripening of Parmigiano Reggiano cheese, J. Agric. Food Chem., 53(10), 4026–4031, 2005. 126. Shintu, L. and Caldarelli, S., Toward the determination of the geographical origin of Emmental(er) cheese via high resolution MAS NMR: A preliminary investigation, J. Agric. Food Chem., 54(12), 4148–4154, 2006. 127. Aue, W.P., Bartholdi, E., and Ernst, R.R., Two-dimensional spectroscopy. Application to nuclear magnetic resonance, J. Chem. Phys., 64(5), 2229–2246, 1976. 128. Contreras, R.H. and Peralta, J.E., Angular dependence of spin–spin coupling constants, Progr. Nucl. Magn. Reson. Spectrosc., 37(4), 321–425, 2000. 129. Neuhaus, D. and Williamson, M.P., The Nuclear Overhauser Effect in Structural and Conformational Analysis, VCH Publ.: New York, 1989. 130. Hwang, T.L. and Shaka, A.J., Reliable two-dimensional rotating-frame cross-relaxation measurements in coupled spin systems, J. Magn. Reson. B, 102(2), 155–165, 1993. 131. Sobolev, A.P., Segre, A., and Lamanna, R., Proton high-ﬁeld NMR study of tomato juice, Magn. Reson. Chem., 41(4), 237–245, 2003. 132. Khatib, A., Wilson, E.G., Kim, H.K., Lefeber, A.W.M., Erkelens, C., Choi, Y.H., and Verpoorte, R., Application of two-dimensional J-resolved nuclear magnetic resonance spectroscopy to differentiation of beer, Anal. Chim. Acta, 559(2), 264–270, 2006. 133. Lindon, J.C., Nicholson, J.K., Holmes, E., and Everett, J.R., Metabonomics: Metabolic processes studied by NMR spectroscopy of bioﬂuids, Concept Magn. Reson., 12(5), 289–320, 2000. 134. Boulat, B., Experimental control of spin diffusion in liquid state NMR: A comparison of methods, J. Magn. Reson., 139(2), 354–363, 1999. 135. Wang, M., Kikuzaki, H., Lin, C.C., Kahyaoglu, A., Huang, M.T., Nakatani, N., and Ho, C.T., Acetophenone glycosides from thyme (Thymus vulgaris L.), J. Agric. Food Chem., 47(5), 1911–1914, 1999. 136. Tiziani, S., Schwartz, S.J., and Vodovotz, Y., Proﬁling of carotenoids in tomato juice by one- and twodimensional NMR, J. Agric. Food Chem., 54(16), 6094–6100, 2006. 137. Maoka, T., Fujiwara, Y., Hashimoto, K., and Akimoto, N., Isolation of a series of apocarotenoids from the fruits of the red paprika Capsicum annuum L., J. Agric. Food Chem., 49(3), 1601–1606, 2001. 138. Tanaka, T., Inoue, K., Betsumiya, Y., Mine, C., and Kouno, I., Two types of oxidative dimerization of the black tea polyphenol theaﬂavin, J. Agric. Food Chem., 49(12), 5785–5789, 2001. 139. Masuda, T., Toi, Y., Bando, H., Maekawa, T., Takeda, Y., and Yamaguchi, H., Structural identiﬁcation of new curcumin dimers and their contribution to the antioxidant mechanism of curcumin, J. Agric. Food Chem., 50(9), 2524–2530, 2002. 140. Sang, S., Mao, S., Lao, A., Chen, Z., and Ho, C.T., Four new steroidal saponins from the seeds of Allium tuberosum, J. Agric. Food Chem., 49(3), 1475–1478, 2001. 141. Giusti, M.M., Ghanadan, H., and Wrolstad, R.E., Elucidation of the structure and conformation of red radish (Raphanus sativus) anthocyanins using one- and two-dimensional nuclear magnetic resonance techniques, J. Agric. Food Chem., 46(12), 4858–4863, 1998. 142. Capozzi, F., Cremonini, M.A., Russo, N., Tocci, E., and Uccella, N., Oleuropein biomimetic conformations by magnetic resonance experiments and molecular mechanics and dynamics, in Magnetic Resonance in Food Science: A View to the Future, Webb, G.A., Belton, P.S., Gil, A.M., and Delgadillo, I. (Eds.), Royal Society of Chemistry: Cambridge, United Kingdom, 2001, pp. 129–135. 143. Bialy, Z., Jurzysta, M., Mella, M., and Tava, A., Triterpene saponins from aerial parts of Medicago arabica L., J. Agric. Food Chem., 52(5), 1095–1099, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C013 Final Proof page 318 17.7.2008 5:39pm Compositor Name: VAmoudavally

318

Handbook of Food Analysis Instruments

144. Lakenbrink, C., Engelhardt, U.H., and Wray, V., Identiﬁcation of two novel proanthocyanidins in green tea, J. Agric. Food Chem., 47(11), 4621–4624, 1999. 145. Xi, Y., de Ropp, J., Viant, M., Woodruff, D., and Yu, P., Automated screening for metabolites in complex mixtures using 2D COSY NMR spectroscopy, Metabolomics, 2(4), 221–233, 2006. 146. Jordheim, M., Fossen, T., and Andersen, O.M., Characterization of hemiacetal forms of anthocyanidin 3-O-glycopyranosides, J. Agric. Food Chem., 54(25), 9340–9346, 2006. 147. Zhang, F., Trbovic, N., Wang, J., and Bruschweiler, R., Double-quantum biased covariance spectroscopy: Application to the 2D INADEQUATE experiment, J. Magn. Reson., 174(2), 219–222, 2005. 148. Frega, N., Bocci, F., Capozzi, F., Luchinat, C., Capella, P., and Lercker, C., A new lipid component identiﬁed in avocado pear by GC-MS and NMR spectroscopy, Chem. Phys. Lipids, 60(2), 133–142, 1991. 149. Savary, G., Guichard, E., Doublier, J.L., Cayot, N., and Moreau, C., Inﬂuence of ingredients on the selfdiffusion of aroma compounds in a model fruit preparation: A nuclear magnetic resonance-diffusionordered spectroscopy investigation, J. Agric. Food Chem., 54(3), 665–671, 2006. 150. Stilbs, P., Fourier transform pulsed-gradient spin-echo studies of molecular diffusion, Progr. Nucl. Magn. Reson. Spectrosc., 19(1), 1–45, 1987. 151. Amato, M.E., Ansanelli, G., Fisichella, S., Lamanna, R., Scarlata, G., Sobolev, A.P., and Segre, A., Wheat ﬂour enzymatic amylolysis monitored by in situ 1H NMR spectroscopy, J. Agric. Food Chem., 52 (4), 823–831, 2004. 152. Tomati, U., Belardinelli, M., Galli, E., Iori, V., Capitani, D., Mannina, L., Viel, S., and Segre, A., NMR characterization of the polysaccharidic fraction from Lentinula edodes grown on olive mill waste waters, Carbohydr. Res., 339(6), 1129–1134, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 319 16.7.2008 11:01pm Compositor Name: MSubramanian

Absorption, 14 Atomic Atomic Emission, and Inductively Coupled Plasma Spectroscopies in Food Analysis John R. Dean and Renli Ma CONTENTS 14.1 14.2

Introduction ........................................................................................................................ 320 Atomic Absorption Spectroscopy ...................................................................................... 320 14.2.1 Absorption of Radiation....................................................................................... 321 14.2.2 Hollow Cathode Lamp ......................................................................................... 321 14.2.3 Atomization Cells................................................................................................. 321 14.2.3.1 Flames.................................................................................................. 321 14.2.3.2 Graphite Furnace ................................................................................. 322 14.2.3.3 Hydride Generation ............................................................................. 322 14.2.3.4 Cold Vapor .......................................................................................... 322 14.2.4 Wavelength Selection and Detection ................................................................... 322 14.2.5 Background Correction Methods ......................................................................... 323 14.2.5.1 Continuum Source............................................................................... 323 14.2.5.2 Smith–Hieftje....................................................................................... 323 14.2.5.3 Zeeman Effect ..................................................................................... 323 14.2.6 Interferences ......................................................................................................... 324 14.2.6.1 Flame Atomization Interferences ........................................................ 324 14.2.6.2 Graphite Furnace Interferences ........................................................... 324 14.3 Atomic Emission Spectroscopy ......................................................................................... 325 14.3.1 Inductively Coupled Plasma-Atomic Emission Spectroscopy (ICP-AES) .......... 325 14.3.2 Inductively Coupled Plasma ................................................................................ 325 14.3.3 Sample Introduction ............................................................................................. 325 14.3.3.1 Nebulizers............................................................................................ 325 14.3.3.2 Spray Chambers and Desolvation Systems......................................... 326 14.3.3.3 Hyphenated Techniques ...................................................................... 326 14.3.3.4 Hydride Generation ............................................................................. 327 14.3.3.5 Cold Vapor Generation ....................................................................... 327 14.3.4 Spectrometers ....................................................................................................... 327 14.3.4.1 Sequential ............................................................................................ 327 14.3.4.2 Simultaneous ....................................................................................... 327

319

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 320 16.7.2008 11:01pm Compositor Name: MSubramanian

320

Handbook of Food Analysis Instruments

14.3.5

Detectors............................................................................................................... 328 14.3.5.1 Photomultiplier Tube........................................................................... 328 14.3.5.2 Charge Transfer Devices ..................................................................... 328 14.3.6 Interferences ......................................................................................................... 328 14.3.6.1 Spectral Interferences .......................................................................... 328 14.3.6.2 Matrix Interferences ............................................................................ 328 14.4 Comparison of Analytical Techniques .............................................................................. 328 14.4.1 Detection Limits................................................................................................... 328 14.4.2 Analytical Working Range................................................................................... 329 14.4.3 Sample Throughput.............................................................................................. 329 14.4.4 Purchase and Operating Costs ............................................................................. 331 14.4.5 Summary of Comparison Criteria ........................................................................ 331 14.5 Applications in Food Analysis .......................................................................................... 334 14.5.1 Beverages ............................................................................................................. 336 14.5.1.1 Soft Drinks .......................................................................................... 336 14.5.1.2 Alcoholic Drinks ................................................................................. 336 14.5.2 Milk, Infant Formula, and Diary Products........................................................... 337 14.5.3 Cereal, Flour, Rice, and Legumes........................................................................ 339 14.5.4 Fruit and Vegetable.............................................................................................. 340 14.5.5 Meat and Meat Products ...................................................................................... 341 14.5.6 Seafood................................................................................................................. 341 14.5.7 Other and Mixed Types ....................................................................................... 342 References ..................................................................................................................................... 344

14.1 INTRODUCTION The application of atomic spectroscopic techniques for the analysis of major, minor, and trace elements in foodstuffs is outlined. Initial discussion focuses on the range of techniques that can be applied to food analysis. In particular, the instrumentation required to perform atomic absorption and atomic emission spectroscopy (AES) is discussed. In the case of atomic absorption spectroscopy (AAS), the different atom cells available are highlighted including ﬂame, graphite furnace, hydride generation, and cold vapor techniques. For AES the focus is on the instrumentation associated with the inductively coupled plasma (ICP). The main focus of the chapter, however, is on the use of these techniques for the analysis of metals and metalloids in beverages; milk, infant formula, and diary products; cereal, ﬂour, rice, and legumes; fruit and vegetables; meat and meat products; and seafood.

14.2 ATOMIC ABSORPTION SPECTROSCOPY AAS has been used extensively for food analysis due to the simplicity of the technique and its low capital cost. The main components of an AAS instrument are a radiation source (hollow cathode lamp [HCL]), an atomization cell (ﬂame or graphite furnace), a method of wavelength selection (monochromator), and detection (photomultiplier tube [PMT]). A narrow line emission of a selected metal is generated by the HCL. The sample is introduced into the hot environment of a ﬂame (or graphite furnace) where metal atoms are liberated. This hot environment causes broadening of the metal’s absorption line. By utilizing the narrowness of the emission line from the HCL, together with the broad absorption line generated in the ﬂame (or graphite furnace) means that the monochromator only has to isolate the line of interest from other lines emitted by the radiation source. It is this unique feature of AAS that gives it such a high degree of selectivity, the process usually being referred to as the ‘‘lock and key’’ effect.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 321 16.7.2008 11:01pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

14.2.1 ABSORPTION

OF

321

RADIATION

When a beam of light passes through the ﬂame (or graphite furnace) some of the light will be absorbed by the atoms present. The transmittance, T, of the atom cell is deﬁned as that fraction of the light that is allowed to pass through unaffected while the absorbance, A, is deﬁned as log (1=T ). In addition, the absorption of radiation follows the Beer–Lambert law such that, the absorbance can be expressed as the following: A ¼ log10 T ¼ log10 I=Io ¼ abc

(14:1)

where a is the absorptivity (a constant) b the path length of the atom cell c is the concentration of atoms in the atom cell

14.2.2 HOLLOW CATHODE LAMP The HCL is an emission source, in that it emits radiation characteristic of a particular metal (or metalloid). The HCL consists of a hollow cylindrical cathode lined with the metal(s) of interest and a tungsten anode. The cathode and anode are located in a cylindrical glass envelope ﬁtted with a silica window. The HCL is ﬁlled with an inert gas, typically argon or neon, under vacuum at 1–5 Torr (100–200 Pa). In order to initiate the HCL a voltage (100–400 V) is applied across the cathode and anode. This corresponds to a current of between 2 and 30 mA. This passage of electric current causes ionization of the ﬁll gas (for example, Ar þ e ¼ Arþ þ 2e). The positive ﬁll gas ions (Arþ) are then attracted to the cathode (negative charged). The impact of Arþ is sufﬁcient to cause the metal atoms on the surface of the cathode to be liberated, i.e., sputtered.

14.2.3 ATOMIZATION CELLS The atomization cell is required to produce ground state atoms. Several types of atomization cells are used but the most common is the ﬂame. If sensitivity is an issue, then the graphite furnace can be used. Additionally, for elements that form hydrides, the technique of hydride generation can be used while for mercury a cold vapor technique is available. 14.2.3.1

Flames

Two ﬂames are commonly encountered in ﬂame AAS (FAAS), i.e., the air–acetylene ﬂame and the nitrous oxide–acetylene ﬂame. In each case, the ﬂame is located in a slot burner positioned in the light path of the HCL. The choice of ﬂame is either the air–acetylene ﬂame (slot length 100 mm, temperature 2500 K) or the nitrous oxide–acetylene ﬂame (slot length 50 mm, temperature 3150 K). The latter is used for the more refractory elements, e.g., Al. The temperatures of the ﬂame can be altered by changing the ﬂame composition between, fuel-rich, a stoichiometric or a fuel-lean ﬂame. The stoichiometric ﬂame usually provides the highest temperature. Aqueous samples are introduced into ﬂames by the use of a nebulizer=expansion chamber arrangement. The pneumatic concentric nebulizer consists of a concentric stainless steel tube in which is located a Pt=Ir capillary tube. Aqueous sample is drawn up through the capillary tube of the nebulizer by the action of the oxidant gas (e.g., air) escaping through the exit oriﬁce that exists between the outside of the capillary tube and the inside of the stainless steel concentric tube. The action of the escaping air and sample is sufﬁcient to shatter the aqueous sample into a coarse aerosol (Venturi effect). The expansion chamber has dual functions: the ﬁrst is to convert the coarse aerosol into a ﬁner aerosol for transport to the burner for atomization and allow residual aerosol particles to condense

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 322 16.7.2008 11:01pm Compositor Name: MSubramanian

322

Handbook of Food Analysis Instruments

and go to waste; and, secondly, to allow pre-mixing of the oxidant (e.g., air) and fuel (acetylene) gases safely prior to introduction into the laminar ﬂow burner. 14.2.3.2

Graphite Furnace

One way to increase the sensitivity is to introduce a discrete amount of sample into a graphite furnace (GFAAS) atomizer. This device replaces the ﬂame=burner arrangement in the AAS instrument. Essentially a small discrete sample (5–100 mL) is introduced onto the inner surface of a graphite tube through a small opening. Note that often a graphite platform is inserted inside the graphite tube to assist the process. The graphite tube (3–5 cm long with a diameter of 3–8 mm) is arranged so that light from the HCL passes directly through its length. The graphite tube is then heated by the passage of an electric current. Careful control of the heating cycle allows various stages to be incorporated that allow drying of the sample (1108C for 30 s), removal of the sample matrix in the ashing stage (3508C and 12008C for approximately 45 s), and ﬁnally to atomization (20008C and 30008C for 2–3 s) of the metal. An additional heating cycle may be introduced for cleaning of residual material from the surface of the graphite tube. Problems can arise during the ashing stage if the organic matrix components of the sample are not removed without any loss of the metal. Remedies for retention of the metal during this stage are discussed in Section 14.2.6.2 (matrix modiﬁcation). The absorbance signal is measured during the atomization stage. It is common practice for an internal gas ﬂow of an inert gas (N2 or Ar) to ﬂow during the drying and ashing stages to remove any unwanted material. 14.2.3.3

Hydride Generation

Hydride generation (HGAAS) is a form of sample introduction exclusively reserved for a limited number of elements capable of forming volatile hydrides (e.g., As, Bi, Sb, Se, Sn). An acidiﬁed sample solution is reacted with a solution of sodium tetraborohydride liberating the gaseous hydride. For example, the production of arsine (AsH3) can be expressed as þ 3BH 4 þ 3H þ 4H3 AsO3 ! 3H3 BO3 þ 4AsH3 þ 3H2 O

(14:2)

Unfortunately when the basic borohydride is added to an acidic solution, excess hydrogen is liberated: þ BH 4 þ 3H2 O þ H ! H3 BO3 þ 4H2

(14:3)

A gas–liquid separator is required prior to introduction of the hydride into the atom cell (either an electrically heated cell or ﬂame heated quartz tube) for atomization. 14.2.3.4

Cold Vapor

Cold vapor (CVAAS) generation is exclusively reserved for the element mercury. The mercury present in the sample is reduced (for example, using tin (II) chloride) to elemental mercury. Sn2þ þ Hg2þ ! Sn4þ þ Hg0

(14:4)

The generated mercury vapor is then transported to the atom cell (a long-path glass absorption cell) located in the path of the HCL. Note that no ﬂame or other heating device is required. Mercury is monitored at 253.7 nm.

14.2.4 WAVELENGTH SELECTION

AND

DETECTION

The optical arrangement of the spectrometer used in AAS is the Czerny-Turner conﬁguration (see also Section 14.3.4). The Czerny-Turner monochromator for AAS has a focal length of 0.25–0.5 m,

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 323 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

323

a diffraction grating containing only 600 lines=mm and a resolution of 0.2–0.02 nm. In a double beam instrument, light from the HCL is split by a mirrored chopper that allows half the light to pass through the atomizer while half the light is diverted. Then the two beams are recombined by a halfsilvered mirror prior to the light passing to the monochromator. The double beam instrument allows correction of HCL ﬂuctuations caused by warm-up, drift, and source noise leading to improved precision in the absorbance measurement. Light is detected by a PMT (see also Section 14.3.4). The PMT is a device that converts incident light to electrons (electric current). Essentially, light strikes a photosensitive material that converts the light into an electron via the photoelectric effect. This generated electron is then focused and multiplied by a series of dynodes prior to collection at the anode. The multiplied electrons (or electrical current) are then measured.

14.2.5 BACKGROUND CORRECTION METHODS AAS is prone to interference from molecular absorbance and scatter. However, these interferences can be overcome by the use of background correction methods. 14.2.5.1

Continuum Source

This type of background correction uses a continuum source, e.g., D2 lamp. In the atomization cell (e.g., ﬂame), absorption is possible from both atomic species (signal) and molecular species (unwanted signal). By measuring the absorption that occurs from the radiation source (HCL) and comparing it with the absorbance that occurs from the continuum source (i.e., D2 lamp) a corrected absorption signal can be obtained. This is because atomic species absorb the speciﬁc radiation associated with the HCL, whereas the absorption of radiation by the continuum source by the same atomic species will be negligible. 14.2.5.2

Smith–Hieftje

In Smith–Hieftje background correction, a single HCL is used which is capable of operating at high and low (or normal) current. When operating a HCL at a high lamp current self-reversal is induced in the emission proﬁle of the radiation source; it is this feature which is exploited in Smith–Hieftje background correction. Essentially, the HCL operating at normal (low) current absorbs radiation from both the (wanted) atomic and (unwanted) molecular species. If the current in the HCL is then increased, self-reversal occurs effectively splitting the proﬁle into two. The incomplete self-reversal in the HCL leads to incomplete resolution of the atomic line proﬁle from molecular interferents. As the atomic species will no longer occur at exactly the same wavelength (as is the case with the HCL at normal current), they are not observed whereas the broadband molecular absorption is observed. By then subtracting the two signals allows a corrected atomic absorption signal to be monitored. 14.2.5.3

Zeeman Effect

This method also utilizes a line-splitting approach as its method of correction. In this case, however, the atomic spectral line is split by the application of a strong magnetic ﬁeld. The magnetic ﬁeld is normally applied to the atomization cell, i.e., the graphite furnace. When an atom in a magnetic ﬁeld is observed with polarized light, the absorption line is divided into two components symmetrically displaced around the normal position. For most metals therefore, the central s component occurs at the absorption wavelength of the metal of interest while no atomic absorption occurs at the two p components. However, any molecular absorption present will be absorbed by the two p components. By operating with the magnet ‘‘off ’’ atomic and molecular absorption are both observed, when the magnet is switched ‘‘on’’ only background is observed, it is then possible to obtain a corrected atomic absorption signal.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 324 16.7.2008 11:02pm Compositor Name: MSubramanian

324

Handbook of Food Analysis Instruments

14.2.6 INTERFERENCES Interferences in AAS can occur as a result of physical, spectral, chemical, and ionization problems. Speciﬁc interferences can be identiﬁed that occur in FAAS and GFAAS. 14.2.6.1

Flame Atomization Interferences

The dominant interferences that occur in FAAS are chemical, ionization, physical, and spectral. 14.2.6.1.1 Chemical Chemical interferences arise when the metal to be determined forms a thermally stable compound with molecular or ionic species present in the sample solution. The presence of phosphate, silicate, or aluminate in the sample solution can cause suppression of the alkaline earth metal (e.g., Ca) absorption signal in the air–acetylene ﬂame. Increasing the amount of phosphate on the Ca absorption signal at 422.7 nm causes signal depression. This signal depression is due to the formation of a thermally stable compound in the ﬂame, i.e., calcium pyrophosphate. 14.2.6.1.2 Ionization Ionization interferences are vapor-phase interferences that occur in the ﬂame. This problem is most severe for alkali (e.g., Na) and alkaline earth metals (e.g., Ca). The low ionization potential of these metals can lead to ionization in the relatively hot environment of the ﬂame, e.g., 2500 K in an air– acetylene ﬂame. If ionization occurs, then no atoms will be present, therefore, no absorption signal is detected. Na ! Naþ þ e

(14:5)

This can be prevented by the addition of an ionization suppressor (or buffer). An ionization buffer would be another alkali metal, e.g., Cs. In this situation, addition of excess Cs will lead to ionization of Cs in the ﬂame in preference to the metal, i.e., Na. The addition of excess Cs therefore suppresses the ionization of Na. This process is known as the ‘‘mass action’’ effect. 14.2.6.1.3 Physical Physical interferences in FAAS are associated with the transport of the sample solution to the ﬂame. Hence physical interferences are related to the properties of the sample solution, e.g., viscosity, and its conversion into an aerosol within the spray chamber. The formation of the aerosol is dependent upon the surface tension, density, and viscosity of the sample solution. This type of interference can be controlled by matrix matching of sample and standard calibration solutions. An alternative approach is to use the method of standard additions. 14.2.6.1.4 Spectral The occurrence of spectral line overlap is not very common in AAS due to the high selectivity of the technique. An example is the resonance line of Cu (324.754 nm) which has a line coincidence from Eu (324.753 nm). In reality unless the Eu is in 1000X excess (compared with Cu) no signiﬁcant interference occurs. Along with atomic spectral overlap, molecular band absorption can occur. Examples include calcium hydroxide with an absorption band on the Ba wavelength at 553.55 nm and Pb at 217.0 nm which has molecular absorption from NaCl. Molecular absorption is not usually a problem since it can be corrected using background correction techniques (Section 14.2.5). 14.2.6.2

Graphite Furnace Interferences

The main interferences in GFAAS are those due to background absorption and light scattering effects. The latter is due to mist and smoke from particles that form at the cooler ends of the graphite tube leading to the occurrence of molecular absorption. Background absorption is particularly troublesome from alkaline or alkaline earth halides, e.g., NaCl. It is possible to correct for molecular, absorption effects using background correction methods (Section 14.2.5).

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 325 16.7.2008 11:02pm Compositor Name: MSubramanian

325

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

The formation of volatile compounds that are lost prior to atomization is a major problem in GFAAS. Volatile compound formation can be prevented by the use of matrix modiﬁcation. In matrix modiﬁcation, the addition of suitable compounds allows volatile compounds to be retained until the atomization step has been reached. For example, in the determination of Pb the addition of ammonium nitrate eliminates interferences due to sodium chloride according to the equation: NH4 NO3 decomposes at 483 K

þ

NaCl ¼ m:p: 1079 K b:p: 1691 K

NaNO3 decomposes at 653 K

þ

NH4 Cl sublimes 618 K b:p: 798 K

(14:6)

The products formed (sodium nitrate and ammonium chloride) are removed by decomposition or sublimation below 700 K. Also, excess ammonium nitrate is also easily removed during the ashing stage (Section 14.2.3).

14.3 ATOMIC EMISSION SPECTROSCOPY The instrumentation for AES consists of an atom cell, spectrometer=detector, and readout device. Flame emission spectroscopy (FES) or ﬂame photometry (FP) is still used to measure a narrow range of metals, e.g., sodium and potassium, that have low excitation potentials (for example, sodium 5.14 eV and potassium 4.34 eV) and hence ability to be easily ionized. The ﬂame is normally air–propane, air–butane, or air–natural gas while the spectrometer is an interference ﬁlter. The cool nature of the ﬂame prevents other metals from being excited thus the technique is relatively free of interferences. Also as a monochromator is not necessary, the use of an interference ﬁlter allows light emission to be viewed by the detector (photodiode or photoemissive detector). A ﬂame photometer is therefore a simple, robust, and inexpensive instrument for the determination of potassium or sodium in food samples.

14.3.1 INDUCTIVELY COUPLED PLASMA-ATOMIC EMISSION SPECTROSCOPY (ICP-AES) The main use of AES has been the link to an ICP source.

14.3.2 INDUCTIVELY COUPLED PLASMA The ICP is formed within the conﬁnes of three concentric glass tubes (or plasma torch). Argon gas is introduced through each concentric glass tube. The intermediate (plasma) and external (coolant) tubes having tangentially arranged entry points. The sample aerosol is introduced through the inner tube which consists of a capillary tube. Located around the outer glass tube is a coil of copper tubing through which cooling water is re-circulated. The power input to the plasma torch is achieved through a copper (load or induction) coil. Typical power inputs range from 0.5 to 1.5 kW at a frequency of 27 or 40 MHz. Emitted radiation is viewed either laterally (side-on) or axially (end-on).

14.3.3 SAMPLE INTRODUCTION For aqueous samples, the combination of a nebulizer and spray chamber is the most common approach for sample introduction, but results in <2% of the sample reaching the ICP. While alternative approaches for the introduction of liquid samples are available, the use of a nebulizer= spray chamber is still the most common method. 14.3.3.1

Nebulizers

The pneumatic concentric glass nebulizer is the most common nebulizer used. Argon gas introduced in the sidearm exits at the nozzle resulting in the formation of a region of low pressure. The aqueous

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 326 16.7.2008 11:02pm Compositor Name: MSubramanian

326

Handbook of Food Analysis Instruments

sample is then drawn up through the capillary tube and exits through the nozzle resulting in the formation of a coarse aerosol. In a cross-ﬂow nebulizer, the liquid sample and argon gas interact at perpendiculars to one another. Through one capillary tube, the argon carrier gas ﬂows while through the other capillary the aqueous sample is pumped. At the exit point the interaction of the escaping argon gas and the aqueous sample produce a coarse aerosol. Modiﬁcations in this design have been used for sample solutions with a high content of dissolved solids, i.e., V-groove Babington-type or Burgener nebulizers. Ultrasonic nebulizers pump the liquid sample solution onto a vibrating piezoelectric transducer (200 kHz and 10 MHz). This vibrating action is sufﬁcient to transform the sample into a coarse aerosol. The aerosol is then transported by the argon carrier gas through a heated tube and then a condenser to remove solvent. In this case, therefore, the aerosol is desolvated and reaches the plasma as a dry aerosol. The major advantage of the ultrasonic nebulizer is its increased transport efﬁciency (approximately 10%). 14.3.3.2

Spray Chambers and Desolvation Systems

The introduction of coarse aerosols directly into the plasma would have deleterious effects that include extinguishing or induced cooling of the plasma. The latter would result in severe matrix interferences. The inclusion of a spray chamber leads to the production of a ﬁner aerosol (<10 mm) which is more appropriate for the plasma source. Several spray chamber designs are available and include the following: . . .

double-pass or Scott type cyclonic single-pass, direct, or cylindrical type

The most common type of spray chamber is the double-pass which comprises two concentric tubes, an inlet for the nebulizer, an exit for the ﬁner aerosol, and a waste drain. Interactions between the nebulizer generated coarse aerosol and the internal surfaces of the spray chamber lead to the production of a ﬁner aerosol (excess liquid goes to waste). In a cyclonic spray chamber the aerosol is introduced tangentially to induce swirling. Initially the aerosol swirls downwards close to the spray chamber wall, while at the bottom of the spray chamber a second inner spiral carries the aerosol to the exit point. These combined processes lead to a reduction in aerosol particle size. In a single-pass (direct or cylindrical) spray chamber an impact bead is required for aerosol production. 14.3.3.3

Hyphenated Techniques

The linking of chromatographic separation with an ICP source can be useful for either obtaining elemental species information (so-called speciation studies) or the removal of potential matrix interferences. The type of chromatography used largely depends on the nature of the species to be separated but it is common to ﬁnd linkages based on high-performance liquid chromatography (HPLC) or gas chromatography (GC). The interfacing of HPLC with an ICP is relatively simple as HPLC ﬂow rates (typically in the range 1 mL min1) are compatible with the aspiration rates of a nebulizer=spray chamber sample introduction system as used for ICP analysis. No instrumental modiﬁcations are therefore required to connect an HPLC to an ICP system apart from a short length of PTFE or PEEK tubing. The interfacing of GC with an ICP is more complicated. As GC separates volatile compounds it is essential that the interface maintains the compounds in the vapor state. This can be done using a heated transfer line (2008C–2508C) thus allowing the carrier gas (argon, helium, or nitrogen) and volatile compounds to be directly transported to the plasma torch.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 327 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

14.3.3.4

327

Hydride Generation

Hydride generation is a gas-phase sample introduction technique that allows elements that are capable of forming volatile hydrides (at ambient temperature) to be introduced in the plasma torch. Typical elements for which this approach is possible are arsenic, antimony, bismuth, selenium, tellurium, and tin. Under acid conditions and in the presence of sodium tetraborohydride (a reducing agent), covalent hydrides are formed, e.g., AsH3, SbH3, BiH3, H2Se, H2Te, and SnH4. The principles of hydride generation are as follows: 1. 2. 3. 4.

Chemical generation of the hydrides Collection and preconcentration of the evolved hydrides (if necessary) Transport of the hydrides and gaseous by-products to the plasma torch Atomization of the hydrides followed by either atomization=emission

An equation to describe the chemical generation of the arsine hydride (AsH3) is shown in Equation 14.2. A gas–liquid separator is required to introduce the hydride into the plasma torch. 14.3.3.5

Cold Vapor Generation

Cold vapor generation is only applicable to elemental mercury. Mercury containing a sample is reduced, using tin(II) chloride, to elemental mercury (Equation 14.4). The generated mercury vapor is transported to the plasma torch by an argon carrier gas.

14.3.4 SPECTROMETERS The role of the spectrometer is to separate the emitted light into its component wavelengths prior to detection. This can be done by measuring one wavelength, corresponding to one element at a time, or by measuring multiwavelengths at the same time (or multielement detection). The typical wavelength coverage of a spectrometer for AES is between 167 nm (Al) and 852 nm (Cs). 14.3.4.1

Sequential

A sequential spectrometer consists of entrance and exit optics, a diffraction grating, and a single detector. The most common arrangement is the Czerny-Turner conﬁguration. This spectrometer has the advantage of ﬂexibility such that selection of the desired wavelength is achieved by rotation of the grating within its spectrometer mounting. 14.3.4.2

Simultaneous

A major advantage of ICP-AES is the ability to perform simultaneous multielement analysis. Traditionally, the polychromator of choice was the Paschen-Runge mounting. In this system, the grating, entrance slit, and multiple exit slits are ﬁxed around the periphery of a Rowland Circle. However, this system has largely been superseded by the Echelle spectrometer. The major component difference in the Echelle spectrometer is the diffraction grating. The grating, with 300 lines or grooves per millimeter, utilizes spectral order for maximum wavelength coverage. In a conventional spectrometer, the resolution (R) of the diffraction grating is directly related to the groove density (n) and the spectral order (M) (R ¼ m.N). In the Echelle conﬁguration, however, resolution is improved by increasing the Blaze angle and spectral order. To prevent overlapping of spectral orders, a secondary dispersion is required. This is achieved with a prism to create a two-dimensional spectral ‘‘map.’’ The spectral ‘‘map’’ generated is arranged into spectral order vertically and wavelength horizontally.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 328 16.7.2008 11:02pm Compositor Name: MSubramanian

328

Handbook of Food Analysis Instruments

14.3.5 DETECTORS After wavelength separation has been achieved, it is necessary to detect the signal with either a PMT or charge transfer device (CTD). 14.3.5.1

Photomultiplier Tube

The PMT is a device that converts incident light into an electric current (Section 14.2.4). 14.3.5.2

Charge Transfer Devices

CTDs offer high sensitivity and wide wavelength coverage using one device. A CTD is an array of closely spaced metal–insulator–semiconductor diodes formed on a wafer of semiconductor material. CTDs are available in two forms, either the charge coupled device (CCD) or the charge injection device (CID). In each case, incident light is converted to a signal.

14.3.6 INTERFERENCES Interferences for AES can be classiﬁed as those due to spectral overlap and matrix effects. 14.3.6.1

Spectral Interferences

To alleviate a spectral interference, either increase the resolution of the spectrometer or select an alternative spectral emission line. Three types of spectral overlap can be identiﬁed and these are: . .

.

Direct wavelength coincidence from an interfering emission line, e.g., Cd 228.802 nm and As 228.812 nm; Zn 213.856 nm and Ni at 213.858 nm Partial overlapping of the emission line of interest from an interfering line in close proximity. Elimination is usually only possible by an improvement in spectrometer resolution Presence of an elevated or depressed background continuum. It is possible to correct by measurement of the background on either side of the wavelength of interest

14.3.6.2

Matrix Interferences

These interferences are associated with either the sample introduction process or the ICP. In the case of the former, a nebulizer can be affected by the dissolved solids content of the aqueous sample which in turn affects the sample uptake rate. In the case of the latter, the presence of easily ionizable elements (EIE’s), e.g., alkali metals, within the plasma source can lead to some signal suppression or enhancement.

14.4 COMPARISON OF ANALYTICAL TECHNIQUES In comparing atomic spectroscopic techniques it is necessary to identify the key parameters to assess. In this case the major parameters are . . . .

Detection limits Analytical working range Sample throughput Purchase and operating costs

14.4.1 DETECTION LIMITS The detection limit of an individual analytical procedure is the lowest amount of metal or metalloid in a sample which can be detected, but not necessarily quantiﬁed, as an exact value. The limit of

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 329 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

329

detection expressed as the concentration or the quantity is derived from the smallest measure, xL that can be detected with reasonable certainty for a given procedure. The value xL is given by the equation: xL ¼ xbl þ ksbl

(14:7)

where xbl is the mean of the blank measures sbl is the standard deviation of the blank measures k is a numerical factor chosen according to the conﬁdence level required For many purposes, the limit of detection is taken to be 3sbl or 3 ‘‘the signal-to-noise ratio,’’ assuming a zero blank. An extensive list of detection limits for the atomic spectroscopic techniques in this chapter are given in Table 14.1. In general terms, however, it is noted that the lower (better) detection limits are attained using graphite furnace AAS. For speciﬁc elements that form hydrides or mercury, particularly low detection limits can be obtained using HGAAS or CVAAS, respectively. The least sensitive of the techniques discussed is ﬂame AAS.

14.4.2 ANALYTICAL WORKING RANGE The analytical working range is the concentration range over which the analytical working calibration plot remains linear. This has the considerable advantage of allowing samples with different metal concentrations to be determined without further dilution or preconcentration, thereby saving analysis time. The analytical working range is expressed in terms of order of magnitude of signal intensity, one order representing a factor of 10. In terms of the atomic spectroscopic techniques discussed, those based on AAS have the smallest analytical working ranges (GFAAS < Hydride generation AAS < Flame AAS) while the ICP-AES offers the largest analytical range with the possibility of analyzing samples over six orders of magnitude.

14.4.3 SAMPLE THROUGHPUT Sample throughput is the number of samples that can be analyzed or metals (metalloids) that can be determined in a speciﬁc time interval. .

.

.

.

FAAS: FAAS has a relatively high sample throughput, but for a limited number of metals (metalloids). The determination of a single metal=metalloid takes the order of 3–10 s. However, if a different metal=metalloid is required to be determined, it requires the HCL to be changed as well as optical parameters, i.e., select a different wavelength, and perhaps ﬂame gas composition. It is normal therefore to refer to FAAS as a single-element technique. GFAAS: As described above for FAAS, GFAAS is also a single-element technique. As the graphite furnace is required to be heated through a cycle to remove solvent and matrix components prior to atomization means that this technique has a low sample throughput. The determination of a single metal=metalloid takes of the order of 2–3 min. HGAAS and CVAAS: As described above for FAAS, HGAAS and cold vapor AAS are also single-element techniques. In both cases as the process involves the chemical generation of volatile components the technique has a low sample throughput. The determination of a single metal=metalloid takes of the order of 1–2 min. ICP-AES: This technique can be operated as a simultaneous multielement technique (with the appropriate spectrometer) with high sample throughput. The technique can process more than 40 metals=metalloids per minute in a sample.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 330 16.7.2008 11:02pm Compositor Name: MSubramanian

330

Handbook of Food Analysis Instruments

TABLE 14.1 Detection Limitsa for Elements Using Atomic Spectroscopic Techniques (mg=L) Element

FAASb

Ag Al As Au B Ba Be Bi Ca Cd Ce Co Cr Cs Cu Dy Er Eu Fe Ga Gd Ge Hf Hg Ho In Ir K La Li Lu Mg Mn Mo Na Nb Nd Ni Os P Pb Pd Pr Pt Rb Re Rh Ru

1.5 45 150 9 1000 15 1.5 30 1.5 0.8 9 3 15 1.5 50 60 30 5 75 1800 300 300 300 60 30 900 3 3000 0.8 1000 0.15 1.5 45 0.3 1500 1500 6 75000 15 30 7500 60 3 750 6 100

HGAASb

CVAASc

0.03

0.03

GFAASd

ICP-AESe

0.005 0.1 0.05 0.15 20 0.35 0.008 0.05 0.01 0.002

0.6 1 2 1 1 0.03 0.09 1 0.05 0.1 1.5 0.2 0.2

0.15 0.004 0.014

0.06

0.009

0.6

3.0 0.005 0.06 0.004 0.005 0.03 0.005

0.07 130 0.05 0.09 2.0 0.03

1.0

0.4 0.5 0.5 0.2 0.1 1.5 0.9 1 0.5 1 0.4 1 1 1 0.4 0.3 0.1 0.04 0.1 0.5 0.5 1 2 0.5 6 4 1 2 2 1 5 0.5 5 1

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 331 16.7.2008 11:02pm Compositor Name: MSubramanian

331

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

TABLE 14.1 (continued) Detection Limitsa for Elements Using Atomic Spectroscopic Techniques (mg=L) Element S Sb Sc Se Si Sm Sn Sr Ta Tb Te Th Ti Tl Tm U V W Y Yb Zn Zr

FAASb

HGAASb

CVAASc

GFAASd

45 30 100 90 3000 150 3 1500 900 30

0.15

0.05

0.03

0.05 1.0 0.1 0.025

0.03

75 15 15 15000 60 1500 75 8 1.5 450

0.1 0.35 0.1

0.1

0.02

ICP-AESe 10 2 0.1 4 10 2 2 0.05 1 2 2 2 0.4 2 0.6 10 0.5 1 0.2 0.1 0.2 0.5

Source: Adapted from Guide to Inorganic Analysis, Perkin-Elmer, Inc., Connecticut, USA, 2004. a

b c

d e

14.4.4 PURCHASE

Determined using elemental standards in dilute aqueous solution. Detection limits based on a 98% conﬁdence level (3 standard deviations). Parameters optimized for each individual element. Obtained using a ﬂow injection system with amalgamation accessory (detection limit 0.2 mg=L without amalgamation). 50 mL sample volumes used. Simultaneous multielement conditions used with axial viewing of the plasma and a cyclonic spray chamber and concentric nebulizer.

AND

OPERATING COSTS

Capital equipment purchase of atomic spectroscopic techniques is directly related to their ability to measure elements. Single-element techniques, such as FAAS, are the lowest cost whereas multielement techniques, such as ICP-AES are more costly. The inclusion of speciﬁc features, e.g., hydride generation or cold vapor accessories, increases the cost of AAS. In contrast the operating costs are the inverse. The high consumption of argon, necessary for the operation of an ICP, increases the operating cost of an ICP-based system over that of FAAS which requires smaller quantities of acetylene (and air).

14.4.5 SUMMARY

OF

COMPARISON CRITERIA

A summary of all the main criteria for selection of an atomic spectroscopic technique are given in Table 14.2. As indicated above the choice of analytical technique is both metal=metalloid and

Sub-ppb Single element Approx. 2–3 min= element=sample Low ppm range 0.5–5 1–10 Very few Many Very few <50 0.2–1 mL Intermediate cost (~£20,000–40,000)

0.1–1 1–2

Few Many Some >60 4–8 mL=min Low cost (~£10,000–15,000)

GFAAS

High ppb Single element Approx. 3–10 s= element=sample Mid ppm range

FAAS

Very few Some Very few <10 2–10 mL Intermediate cost (~£20,000–30,000)

0.5–5 1–10

Sub-ppb Single element Approx. 1–2 min= element=sample Low ppm range

HGAAS

Very few Some Very few 1 2–10 mL Intermediate cost (~£20,000–30,000)

0.5–5 1–10

Sub-ppb Single element (Hg) Approx. 1–2 min= element=sample Low ppm range

CVAAS

Some Very few Some >60 1–2 mL=min High cost (approx. £70 K); depends whether simultaneous (higher cost) or sequential

0.1–2 1–5

High ppm range

Sub-ppb–ppm Multielement Approx. 1–5 min=sample

ICP-AES

332

Dynamic range Precision (%) Short term Long term Interferences Spectral Chemical Physical No. of elements applicable to Sample volume required Capital cost of instrument

Detection limits Analytical capability Sample throughput

Criterion

TABLE 14.2 Comparison of the Performance of FAAS, GFAAS, HGAAS, CVAAS, and ICP-AES

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 332 16.7.2008 11:02pm Compositor Name: MSubramanian

Handbook of Food Analysis Instruments

User friendly

Autosampler can be easily added Low

Ease of use of instrument

Possibilities for automation; unattended operation Cost per sample (overall)

Autosampler can be easily added Medium

Requires gas supply of nitrogen; purchase hollow cathode lamps for elements to be analyzed. In addition, graphite tubes are required on a frequent basis Requires some expertise to develop methods and initial training on the use of graphite furnace

Autosampler can be easily added Medium

Requires some expertise to develop methods and initial training on the use of hydride generation system

Requires reagents for hydride generation; gas supply of nitrogen; purchase hollow cathode lamps for elements to be analyzed

Autosampler can be easily added Medium

Requires some expertise to develop methods and initial training on the use of cold vapour generation system

Requires reagents for cold vapour generation; gas supply of nitrogen; purchase hollow cathode lamps for elements to be analyzed

Operation is relatively straight forward based on a nebulizer=spray chamber. Obviously complexity is added with other sample introduction devices. Software skills required for operation Autosampler can be easily added Medium

Requires large quantities of argon gas. Periodic replacement of ICP torch

Note: CVAAS, cold vapor atomic absorption spectroscopy; FAAS, ﬂame atomic absorption spectroscopy; GFAAS, graphite furnace atomic absorption spectroscopy; HGAAS, hydride generation atomic absorption spectroscopy; ICP-AES, inductively coupled plasma-atomic absorption spectroscopy.

Requires gas supply of acetylene; purchase hollow cathode lamps for elements to be analyzed

Instrument operating costs (excluding normal supply of electricity and water, as required)

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 333 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies 333

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 334 16.7.2008 11:02pm Compositor Name: MSubramanian

334

Handbook of Food Analysis Instruments

sample throughput focused in a food science laboratory. It is also important to remember that the food sample will require extensive preparation prior to analysis by the selected atomic spectroscopic technique. See Chapter 4 for details of microwave sample preparation methods for food analysis.

14.5 APPLICATIONS IN FOOD ANALYSIS There are a large number of reports in literature on major, minor, and trace element analyses in foodstuffs by atomic spectroscopy. In general, many works involve different forms of AAS to cover a range of metals at various concentration levels from high ppm (such as Ca, K, Mg, and Na) by FAAS to low ppb (such as As, Cd, Co, Cu, Fe, Hg, Mn, Ni, Pb, Se, and Zn) by GFAAS or HGAAS=CVAAS. Increasingly, ICP spectroscopy is favored for multielement analysis with wide linear working range over six orders of magnitude. Some elements either essential (e.g., Ca, Cu, Fe, and Zn) or toxic (e.g., As, Cd, Hg, and Pb) attract more attention than the others. Table 14.3 indicates the EC regulatory limits of some important metals in foodstuffs [1]. Foods and beverages are analyzed for metal content either in the context of (a) nutrition or food safety or (b) for the purpose of categorization or authentication. The latter is usually achieved in conjunction with statistical evaluation, i.e., multivariate analysis with pattern recognition. Sample preparation steps are commonly required which vary from simple dilution, total digestion or preconcentration for various types of foodstuffs. Microwave-assisted destruction procedures are frequently used (Chapter 4). In vitro gastrointestinal digestion, simulating digestive actions in the stomach and intestines, is increasingly performed to assess the bioaccessibility of trace metals in food [2]. Analytical quality assurance is vital for achieving meaningful data. The instrumental parameters and experimental procedures should be optimized to maximize sensitivity and to eliminate matrix effects and interferences. Accuracy must be veriﬁed with the analysis of certiﬁed reference materials

TABLE 14.3 Maximum Levels of Pb, Cd, and Hg in Foodstuffs Metal Pb

Food Product Cows’ milk, infant formulae and follow-on formulae Meat of bovine animals, sheep, pig, and poultry excluding offal Edible offal of cattle, sheep, pig, and poultry Muscle meat of ﬁsh excluding ﬁsh species listed below Muscle meat of wedge sole, eel, spotted sea-bass, horse mackerel or scad, grey mullet, common two-banded sea-bream, grunt, European pilchard or sardine Crustaceans excluding brown meat of crab Bivalve molluscs Cephalopods (without viscera) Cereals (including buckwheat), legumes, and pulses Vegetables including peeled potatoes and excluding Brassica, leafy vegetables, fresh herbs, and all fungi Brassica, leafy vegetables, and all cultivated fungi Fruit excluding berries and small fruits Berries and small fruits Fats and oils, including milk fat Fruit juices, concentrated fruit juices (for direct consumption), and fruit nectars Wines (including sparkling wines and excluding liqueur wines), aromatized wines, aromatized wine-based drinks and aromatized wine-product cocktails and ciders, perry, and fruit wines.

Maximum Permissible Level 0.02 0.1 0.5 0.2 0.4 0.5 1.0 1.0 0.2 0.1 0.3 0.1 0.2 0.1 0.05 0.2

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 335 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

335

TABLE 14.3 (continued) Maximum Levels of Pb, Cd, and Hg in Foodstuffs Metal Cd

Hg

Food Product Meat of bovine animals, sheep, pig, and poultry excluding offal Horsemeat Liver of cattle, sheep, pig, and poultry Kidney of cattle, sheep, pig, and poultry Muscle meat of ﬁsh excluding ﬁsh species listed below Muscle meat of wedge sole, eel, European anchovy, louvar or luvar, horse mackerel or scad, grey mullet, common two-banded sea-bream, European pilchard or sardine Crustaceans excluding brown meat of crab Bivalve molluscs Cephalopods (without viscera) Cereals excluding bran, germ wheat grain, and rice Bran, germ wheat grain, and rice Soybeans Vegetables and fruits excluding leafy=stem=root vegetables, fresh herbs, all fungi and potatoes Leafy vegetables, fresh herbs, celeriac, and all cultivated fungi Stem and root vegetables and peeled potatoes excluding celeriac Fishery products, except those listed below Anglerﬁsh, Atlantic catﬁsh, bass, blue ling, bonito, eel, halibut, little tuna, marlin, pike, plain bonito, Portuguese dogﬁsh, rays, redﬁsh, sail ﬁsh, scabbard ﬁsh, shark, snake mackerel, sturgeon, swordﬁsh, and tuna

Maximum Permissible Level 0.05 0.2 0.5 1.0 0.05 0.1 0.5 1.0 1.0 0.1 0.2 0.2 0.05 0.2 0.1 0.5 1.0

Source: Adapted from Byrne, D., Ofﬁcial J. Eur. Commun. L77, 16.3, 2001. Note: Maximum levels are given in milligram per kilogram wet weight.

and the quantitative recovery of spiked analytes at concentration levels representative of real samples, and precision expressed in relative standard deviation (RSD) be assessed accordingly. In a typical method validation report by Haouet et al., method precision, accuracy, limit of detection (LOD), and limit of quantiﬁcation (LOQ) were evaluated for Hg determination in ﬁshery products by CVAAS [3]. Very recently, Jorhem et al. evaluated the data published during the last 10 years for Cd, Cr, and Pb in milk and muscle tissue from domestic animals and ﬁsh [4]. Results often varied with several orders of magnitude, which was not normally the case of real occurrence but linked to poor analytical quality control. A model for grading the quality of analytical publications was tested in 105 surveys. It was apparent that as the description of the quality procedures increased, the span of results in a given study decreased. This correlation was statistically signiﬁcant (p < .05) for Cd in meat and ﬁsh muscle and for Pb in milk. There was no visible trend towards improvement in quality procedure description over the last decade. The publications were then ranked according to their quality to indicate the level of conﬁdence a reader could have in the results. Papers published in journals specialized in food and analytical chemistry generally received higher ranking than papers published in other ﬁelds of expertise, e.g., toxicology or environment. In this section, analysis of various types of foodstuffs by AAS, AES, and ICP-AES are discussed based on reports published since 2001. Works attempted for many samples or metals and=or involving method development or evaluation are preferred and studies based on total elemental measurements are selected. Elemental speciation analyses usually by hyphenated techniques, such as liquid chromatography coupled with ICP-mass spectrometry are discussed in Chapter 10.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 336 16.7.2008 11:02pm Compositor Name: MSubramanian

336

Handbook of Food Analysis Instruments

14.5.1 BEVERAGES 14.5.1.1

Soft Drinks

The concentrations of major (Ca, K, Mg, Na, S, and P) and minor (Cr, Cu, Fe, Mn, Mo, Se, and Zn) elements in chocolate-ﬂavored beverages from Sao Paulo State (Brazil) were determined by Pedro et al. [5]. Forty-four samples from 13 different brands were analyzed by ICP-AES after digestion in microwave oven. Signiﬁcant differences in the contents of both major and minor elements were found between and within the different brands. Beverages with added powdered milk were good sources of Ca, K, Mg, S, and P; whereas all samples tested were poor sources of Na. Similarly, bottled Brazilian coconut waters were investigated for major (Ca, Mg) and minor (Cu, Fe, Mn, and Zn) constituents using ICP-AES without a mineralization step, since limited information was presented by the manufacturers (only the Ca, Fe, and Na content) [6]. The concentration ranges (mg L1) in different brands were: Ca, 178–232; Mg, 87–129; Mn, 1.8–3.8; Zn, 0.2–0.36; Fe, 0.08–0.18; and Cu, 0.09–0.19. On the basis of the consumption of 300 mL of the beverage, the nutritional contribution was estimated at 6%, 8%, and 56% of the reference daily intakes (RDI) for minerals for Ca, Mg, and Mn, respectively. Samples (248 in duplicate) of cocoa, tea infusions, soft drinks, and fruit juices were examined by GFAAS for their Al content [7]. The Al concentration in commercial tea infusions ranged from 0.9 to 3.3 mg L1 (1.80 0.65 mg L1), whereas in laboratory-prepared samples it was 2.7 0.93 mg L1. In soft drinks, the concentrations of Al were lower, ranging from 9.1 to 179 mg L1; the highest values were observed in samples of orange squash (114 56 mg L1). Apricot juice showed the highest Al level (602 190 mg L1), being statistically different from that of pear (259 102 mg L1), but not different from that of peach juice (486 269 mg L1). Toxicologically, the amount of Al deriving from the consumption of these products was far below the acceptable daily intake of 1 mg kg1 body weight indicated by the Food and Agriculture Organization (FAO) and World Health Organization (WHO), and it was a very low percentage of the normal Al dietary intake. The Al concentration in different brands of fruit juices and tea beverages packaged in different containers (Al cans, glass and PET bottles, and Tetra Brik) was measured by GFAAS [8]. The results varied signiﬁcantly with the brand of fruit juices or tea beverages and the type of packaging containers. The samples of fruit juices or tea beverages packaged in Al cans had signiﬁcantly higher Al content than the same brand bottled in others containers, indicating that some Al was taken up by the fruit juices or tea beverages in Al cans during the storage time. There was no signiﬁcant difference in Al content between the samples packed in glass and PET bottles and Tetra Brik packaging. The mean Al concentration of all fruit juice samples investigated was 0.44 mg L1 and that of tea beverage 2.09 mg L1. The evaluated daily intake of Al (0.22 mg per day) through the consumption of fruit juices was practically negligible in relation to both the total dietary intake and the tolerable daily intake. Tea beverages could be a signiﬁcant source of dietary Al intake in relation to the contribution from others beverages. The evaluated daily Al intake (1.04 mg per day) through consumption of tea beverages was still signiﬁcantly below the calculated tolerable daily intake (TDI) value, calculated from FAO=WHO report. The extractability of Cu, Fe, Mn, and Zn was studied in 30 tea samples of different origins imported to the Czech Republic (13 green, 13 black, 2 semi-fermented, and 1 white tea) [9]. The easily hot-water soluble proportions in 5 min, 60 min, and 24 h infusions were 30% 16% for Cu, 26% 10% for Zn, 18% 10% for Mn, and 1.5% 0.8% for Fe, relative to the respective total contents in tea leaves. Tea infusion could be an important dietary source of Mn as the total contents of Mn were much higher compared to the total contents of Cu, Fe, and Zn, and varied between 511–2220 mg kg1. 14.5.1.2

Alcoholic Drinks

Ten samples of white wine and 10 samples of red wine available from the supermarkets in the province of Mendoza in Argentina were analyzed for Al, Cd, Ca, Cr, Cu, Fe, Pb, and Zn by GFAAS

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 337 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

337

and ICP-AES equipped with ultrasonic nebulization (USN) after mineralization [10]. The concentrations were between 17.0–18.0, 1.0–4.7, 10000–15000, <0.2–6.25, 23.0–28.0, 480–790, 50–90, and 24–130 mg L1, respectively, compared well with those reported for similar wines from some other parts of the world. Seven elements were determined in 14 white and 33 red wines from 9 Uniguayan wineries by FAAS (Cu, Fe, Mn, and Zn) and GFAAS (Cd, Cr, and Pb) [11]. The concentration ranges (micrograms per liter) were 2–3 (Cd), 4–52 (Cr), 6–57 (Pb), 34–650 (Cu), 730–4600 (Fe), 740–2200 (Mn), and 490–2200 (Zn), all within the typical range for wines from around the world and none above the limits established by the Ofﬁce International de la Vigne et du Vin (OIV). Earlier, Larcher and Nicolini surveyed 22 elements (Ag, Al, B, Ba, Ca, Cd, Co, Cr, Cu, Fe, K, Li, Mg, Mn, Na, Ni, Pb, Rb, Sn, Sr, V, and Zn) in 60 wines from Trentino (Italy) using ICP-AES [12]. Red wines had signiﬁcantly higher levels of B, Ba, Fe, K, Li, Mg, Ni, Pb, Rb, and Sr than white wines; and Ca was higher in white wines. The low concentrations of Ca, Mg, Fe, Pb, and Zn might be attributable to the improvement in winery equipment in the last decades. Copper was at times in excess of technological needs, probably due to residues from spray treatments of the grapes. Various types of beers were analyzed for Cu, Fe, and Mn by direct ICP-AES after degassing (to remove CO2 without dilution) in comparison with GFAAS and ICP-AES analysis after decomposition [13]. The LODs were 1.1, 1.1, and 0.3 mg L1 for Cu, Fe and Mn, respectively. Matrixmatched (40% v=v ethanol–water media) calibration was performed for the determination of Al, Cd, and Pb by GFAAS without the real need of chemical modiﬁers in 53 Brazilian sugarcane spirit, cachaca, from 53 different Jequitinhonha High Valley producers in the Minas Gerais State [14]. The Al results varied from nondetectable (<2) to 22.4 mg L1; the Cd values from <0.07 to 0.7 mg L1 and the Pb data from <0.6 to 526.0 mg L1. To investigate which metals constitute diagnostic parameters that establish authenticity of the traditional Cypriot spirit, ‘‘zivania,’’ 68 beverages with an alcoholic degree ranging between 40 and 55 from different countries were analyzed for 16 most abundant metals using ICP-AES [15]. The results were analyzed statistically using two different types of methods: canonical discriminant analysis and classiﬁcation binary trees. Contents of Mg, Zn, and Cu were found as promising distinctive parameters capable of differentiating zivania from other spirits similar in alcoholic degree. The differentiation in those metals between the alcoholic beverages examined might be related to the unique geological and climatic conditions on the island of Cyprus. Meanwhile, concentrations of Cu, Fe, and Zn were determined in ‘‘aguardiente de Cocuy de Penca’’ (Cocuy de Penca Firewater), a spirituous beverage very popular in the north-western region of Venezuela, by FAAS, because their presence could be traced back to the (illegal) manufacturing process [16]. Linear and quadratic discriminant analysis (QDA) and artiﬁcial neural networks (ANNs) trained with the back-propagation algorithm were employed to distinguish such beverages based on the concentrations of these elements in the ﬁnal product and to assess the geographic location of the manufacturers (Lara or Falcon states). Linear discriminant analysis (LDA) performed poorly with an overall rate just above 50%. QDA showed a slightly better yet unsatisfactory overall performance rate over 70%. Of the studied ANNs, comprising a linear function (L) in the input layer, a sigmoid function (S) in the hidden layer(s), and a hyperbolic tangent function (T) in the output layer, the ratio 3L:5S:7S:4T gave the highest rate, 97%, a superb performance for classiﬁcation purposes.

14.5.2 MILK, INFANT FORMULA, AND DIARY PRODUCTS The concentrations of Cd, Cu, Fe, Mn, Pb, and Zn in colostrum and transitory human milk were studied by AAS in conjunction with various factors inﬂuencing the concentrations, i.e., diet, supplementation, place of residence, smoking, as well as socioeconomic, and somatometric characteristics [17]. Colostrum samples from 180 healthy lactating women were collected on the third day postpartum and a second milk sample was collected from 95 subjects 14 days later. The mean values ( standard deviation) of colostrum samples were: Zn, 4905 1725; Fe, 544 348; Cu, 381 132; Mn, 4.79 3.23; Cd, 0.19 0.15; and Pb, 0.4 0.6 mg L1. All metals with the

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 338 16.7.2008 11:02pm Compositor Name: MSubramanian

338

Handbook of Food Analysis Instruments

exception of Cu were found in lower concentrations in transitory samples. The weekly intake of Cd and Pb was found below the Maximum Tolerable Weekly Intakes for infants established by WHO or NRC. Dietary habits played a role in metal levels in human milk as the logistic regression models revealed. The results also revealed higher Pb concentration in samples from urban areas and effect of smoking on Cu level. Toxic elements in ewe milk were investigated by GFAAS (Cd and Pb) and HGAAS (As and Hg) during the summer on pasture [18]. Milk sampling from ewes of Merinolandschaf breed (n ¼ 10) was carried out on the 2nd, 10th, 30th, and 60th lactation day. The concentrations were very low and varied in dependence on lactation stage. In colostrum (2nd lactation day) Cd and Pb concentrations (0.011 and 0.035 mg kg1) were signiﬁcantly higher (p < .01) whereas As concentration (0.011 mg kg1) was lower in comparison with milk on the 10th (Cd: 0.004; Pb: 0.022; As: 0.025 mg kg1), 30th (Cd: 0.005; Pb: 0.024; As: 0.028 mg kg1), and 60th (Cd: 0.006; Pb: 0.026; As: 0.029 mg kg1) lactation day. No signiﬁcant differences (p > .05) were found in milk Hg concentration in relation to lactation stage (from 0.021 to 0.026 mg kg1). The concentrations of Al, Ca, Cd, Cu, Fe, K, Mg, Mn, Na, P, Pb, Se, and Zn in 105 different infant formulae (starter, follow-up, premature, specialized and soya formulae) marketed in Spain were determined by FAAS, GFAAS, and ICP-AES after acid microwave decomposition [19]. Pattern recognition methods including hierarchical cluster analysis (HCA) and principal component analysis (PCA) as unsupervised exploratory techniques, and LDA were applied to characterize, classify, and distinguish different types of infant formulae. The HCA results showed that elemental content data were adequate to obtain the infant formula differentiation. PCA permitted the reduction of 13 variables to 4 principal components accounting for 61.9% of the total variability. This fourfactor model interpreted the correlations of these studied elements reasonably well. The obtained element associations might be attributed to the composition of matrix ingredients, the contamination during elaboration, the additives and mineral supplements added, and the present tendency of standardization in the manufacture of infant formulae. The application of LDA correctly assigned 77.1% of infant formulae with three clearly differentiated and two overlapped groups. An earlier and smaller study involved chemometric tools to observe differences between infant formulas fortiﬁed with inorganic salts (Cu, Fe, and Zn sulphates) or with organic and inorganic salts (Cu and Fe gluconates, or Zn and Fe lactates and Zn oxide) [20]. Thirty-ﬁve infant formula samples from different manufacturers were tested by AAS for Fe, Cu, and Zn. PCA achieved a reduction from 9 variables to 3 (accounting for 80.8% of the total variability), and some differences between the two groups were observed. Cluster analysis gave similar results as PCA. LDA allowed the classiﬁcation of infant formulas in two categories: the ﬁrst formed by samples fortiﬁed with inorganic salts (category A) and the second one by samples fortiﬁed with organic and inorganic salts (category B). The percentages of samples correctly classiﬁed were 96.1 and 100.0 for the categories A and B, respectively. The application of soft independent modeling of class analogy (SIMCA) achieved 87.5% and 12.5% correct assignment for the categories A and B, respectively, poorer results because of the small sample number used, mainly in category B. An in vitro method simulating newborn digestion was developed to study Fe and Zn bioavailability from human milk and cow’s milk-based infant formulas [21]. Enzyme treatment was conducted in two stages involving (1) pepsin at pH 5.0 (for newborn babies) followed by (2) pancreatin at neutral pH, where the incubation times were kept short to mimic the fast transit in the infant’s gastrointestinal tract. Analytes were determined in the fractions obtained after centrifugation by FAAS using a high performance nebulizer. The results were compared to those obtained by performing gastric digestion at pH 2.0 for an adult, using various incubators to treat the sample and centrifugation or ultracentrifugation to separate soluble fractions. No differences in Fe bioavailability from breast milk and infant formulas at different pHs could be detected due to the variability of the infant formulas analyzed. However, Zn bioavailability from breast milk samples was higher than those obtained from infant formulas at the newborn gastric pH. In a later study on Cu with GFAAS, percentage of Cu solubility in the ﬁrst stage was larger from breast milk than infant

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 339 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

339

formulas (65.3% 14.0% vs 40.0% 13.9% at pH 2.0; 61.2% 16.5% vs 26.6% 10.3% at pH 5.0), but the soluble content was larger from infant formulas for both pHs (245.3 82.1 vs 113.0 103.4 ng mL1 at pH 2.0; 169.3 76.9 vs 75.3 21.9 ng mL1 at pH 5.0) [22]. A dynamic continuous-ﬂow dialysis method with online GFAAS was developed to study the Fe bioavailability in milk [23]. The method was based on a simulated gastric digestion in a batch system followed by a continuous-ﬂow intestinal digestion, which was performed in a dialysis bag placed inside a channel containing a ﬂowing stream of dialyzing solution (NaHCO3). The system was applied to powdered cow milk, cereal milk, and two brands of soymilk. Iron dialyzability was found to be 1.7%, 20.4%, 24.9%, and 37.7%, respectively.

14.5.3 CEREAL, FLOUR, RICE, AND LEGUMES Four digestion procedures were investigated for analysis of legumes by ICP-AES, including wet digestion with HNO3=H2SO4 and HNO3=H2SO4=H2O2 and dry ashing with Mg(NO3)2 and Mg(NO3)2=HNO3 [24]. The method of standard additions was preferred for all analytes (Al, Cd, Cr, Cu, Fe, Mg, Mn, Pb, and Zn) with quantiﬁcation limits lower than 2.5 mg g1. Acceptable results were obtained for certiﬁed reference materials from all procedures and the wet digestion method with HNO3=H2SO4=H2O2 provided better recovery. The digestion procedures were also evaluated for the analysis of cereals and cereal ﬂours [25]. Essential (Cr, Cu, Fe, Mg, Mn, and Zn), nonessential (Ag, Al, Ba, Bi, In, and Ga), and toxic (Cd and Pb) minor and trace elements were determined by ICP-AES using standard addition method. In contrast to legumes samples, the use of H2O2 for wet digestion or HNO3 for dry ashing was not necessary. Linear regression analysis and Student’s paired t-test were applied to evaluate the signiﬁcant differences between different procedures and type of samples. Conditions for microwave-assisted wet digestion were optimized for Cu, Fe, Mn, and Zn determination in three Turkish legumes (kidney bean, lentil, and chickpea) by FAAS [26]. Concentrations (milligram per kilogram) were in the ranges of 4–13 (Cu), 46–94 (Fe), 10–20 (Mn), and 19–26 (Zn), respectively. Closed-vessel microwave oven digestion was also used for the determination of Fe, Mn, and Zn in beans, rice, chickpeas, and lentils by FAAS [27]. Samples were effectively dissolved by 4 mL HNO3 and 3 mL H2O2 (30%) in approximately 15 min. The results (microgram per gram) obtained for each sample were: beans (Fe: 79.81, Mn: 15.33, and Zn: 30.77), rice (Fe: 30.96, Mn: 8.28, and Zn: 13.97), chickpeas (Fe: 66.25, Mn: 31.90, and Zn: 29.94), and lentils (Fe: 117.68, Mn: 18.18 and Zn: 44.69), all in very good agreement with the wet digestion and ashing methods. Forty-nine Uruguayan rice samples were digested by dry ashing for the determination of As, Cd, Cr, and Pb by GFAAS; while Ca, Co, Cu, Fe, K, Mg, Mo, Mn, Na, Ni, and Zn were determined by FAAS; and Hg by CVAAS using microwave-assisted decomposition [28]. The following concentration ranges (milligram per kilogram) were obtained for Ca (91–150), Cd (0.002–0.004), Co (0.041–0.060), Cu (1.33–180), Fe (4.41–7.15), K (1670–2170), Mg (450–1210), Mo (0.52– 0.97), Mn (5.45–25.4), Na (9.5–25.0), Ni (0.53–0.72), and Zn (5.86–12.6), all fall within the typical range of rice grown around the world. Potassium was the most abundant mineral, followed by Mg and Ca. Among minor elements, the concentrations of Cu, Fe, Mo, Mn, Na, and Zn in rice were outstanding. It was also found that the milling process affected highly the K, Mg, Mn, Na, and Zn concentrations, while it had little inﬂuence on Ca, Co, Cu, and Fe. In contrast, there was a loss of Ca, Fe, and Mn during the parboiling process. All rice samples tested showed lower levels of As, Cd, Cr, Hg, and Pb in comparison to the maximum limit permitted by government organizations. The levels of Al, Cd, Cr, Cu, Fe, Ni, Pb, and Zn in 40 different legumes and 56 different nuts widely consumed in Spain were evaluated using GFAAS after mineralization with HNO3 and V2O5 [29]. In legumes, the levels (microgram per gram) ranged between 2.7–45.8 Al, nondetectable-0.018 Cd, 0.05–0.60 Cr, 1.5–5.0 Cu, 18.8–82.4 Fe, 0.02–0.35 Ni, 0.32–0.70 Pb, and 32.6–70.2 Zn. In nuts, the levels (microgram per gram) ranged between 1.2–20.1 Al, nondetectable0.018 Cd, 0.25–1.05 Cr, 4.0–25.6 Cu, 7.3–75.6 Fe, 0.10–0.64 Ni, 0.14–0.39 Pb, and 25.6–69.0 Zn.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 340 16.7.2008 11:02pm Compositor Name: MSubramanian

340

Handbook of Food Analysis Instruments

A direct statistical correlation between Cu–Cr, Zn–Al, and Cr–Ni ( p < .05) and Al–Pb ( p < .001) was found. The levels of Cd and Pb (regulated toxic elements) and Cr, Fe, and Ni (markers of metal release from equipment) were investigated by ICP-AES along the durum wheat processing chain, from grain to the ﬁnal product [30]. Durum wheat grain, semolina, and pasta were sampled at an industrial plant for milling and pasta making. Samples were taken at different stages along processing in order to elucidate the inﬂuence of each stage on the element content. Milling was the key process inﬂuencing the concentrations of the studied elements and reduced the metal levels according to a deﬁnite element-speciﬁc pattern. Purity of the water used for grain tempering and dough preparation, element deposition from plant air, and metal release from equipment were identiﬁed as critical issues in contamination control during processing. A simple relationship could be established between the original concentrations in durum wheat grain and those in pasta for Cd and Fe, while for Pb, Cr, and Ni, a greater uncertainty in the estimation of the levels in the ﬁnal product was observed. Food based on cereal grains and soybean and commonly consumed in China was analyzed for Ca, Fe, and Zn using AAS [31]. The Ca contents (milligram per kilogram) were between 20.8 for ground corn and 7606.7 for diced fried soybean curd. The lowest values of Fe and Zn were 0.4 mg kg1 for Panjin pearl rice cooked with discarding extra water and 0.8 mg kg1 for potato and bean starches, while the highest values were observed in dried stick-shaped soybean milk ﬁlm. Although many foods were relatively rich in Ca, Fe, and Zn, many also contained a higher level of phytate impairing their bioavailability depending upon food processing and cooking methods. Of the 60 food samples, 34 foods had a phytate=Ca molar ratio > 0.24, 53 foods had a phytate=Fe molar ratio > 1, 31 foods had a phytate=Zn molar ratio > 15, and only 7 foods had a phytate calcium=zinc > 200. A ﬂow injection-hydride generation (FI-HG) GFAAS method was applied to the determination of Se in cereals and bakery products [32]. The samples were dissolved in a mixture of HNO3 and H2O2 using microwave-assisted digestion. The decomposition of H2Se generated from the sample solutions and the trapping of elemental Se were performed on an Ir-pretreated graphite platform. The overall efﬁciency of hydride generation, transport, and trapping was 86%. The upper limit of the linear dynamic range of calibration was 10 mg L1, corresponding to 0.5 mg g1 in solid samples, and the LOD was 0.06 mg L1 or 3 ng g1 for solid samples. The Se content in bakery products made of common cereals ranged from 7.7 to 68 ng g1 (wet weight) in 18 samples, whereas the Se content of the corresponding cereals was found to be below 100 ng g1 (wet weight). The Se level of cereals grown on soils treated with Se-doped fertilizers ranged from 128 to 1046 ng g1 (wet weight), and it depended linearly on the Se concentration of the corresponding foliar fertilizer.

14.5.4 FRUIT

AND

VEGETABLE

Trace element proﬁling (Ca, Cd, Cr, Cu, Fe, K, Mg, Mn, Na, Ni, P, V, and Zn) of strawberry, blueberry, and pear using ICP-AES was performed to classify the geographic growing origin [33]. Each fruit was collected from two growing regions: Oregon vs Mexico, Chile and Argentina, respectively. Various modeling approaches (LDA, QDA, neural networks, and hierarchical tree model) provided successful classiﬁcation ranging from 70% to 100% depending on commodity and model. Effects of Oregon subregional and variety classiﬁcation were investigated with similar success rates. The levels of Cu and Zn were determined by GFAAS in 66 white and red grapes of the most widely consumed varieties, as well as 60 grape juice samples (39 from white varieties and 21 from red ones) chosen from the main commercial brands in Spain, after digestion with HNO3–H2O2 for grapes and with HNO3 for grape juice [34]. The mean Zn contents obtained, 0.462 mg kg1 in grapes and 0.460 mg L1 in grape juice were lower than those provided by most of the commonly

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 341 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

341

used food composition tables. The mean Cu contents were 0.515 mg kg1 in grapes and 0.063 mg L1 in grape juice. On the basis of the ofﬁcial data on consumption of grapes and grape juice in Spain, the contribution of both products to the recommended daily intake of zinc (15 and 12 mg per day for healthy adult men and women, respectively) was estimated to be approximately 0.1%, whereas it was 0.25% of the established daily intake for Cu (1.5–3 mg per day in adults). Leaves of lettuce and spinach, and roots of radish and carrot were subjected to in vitro gastrointestinal extraction to assess metal bioavailability [35]. Relative to other vegetables investigated, spinach accumulated a high content of Mn and Zn determined by FAAS. The greatest extent of metal releasing was found from lettuce (Mn, 63.7%; Zn, 45.2%) and radish (Mn, 45.8%).

14.5.5 MEAT

AND

MEAT PRODUCTS

Three sample mineralization procedures were tested for the determination of Cu, Fe, Se, and Zn in chicken meat and feed samples [36]. A closed-vessel microwave mineralization procedure was selected for chicken meat samples to avoid Se volatilization losses. For feed samples, HF was required to dissolve siliceous particles which led to lower Zn recovery. Chicken, pork, beef, lamb, and turkey samples (both meat and meat products) collected in the island of Tenerife (Spain) were analyzed for Cd and Pb by GFAAS to assess the percentage contribution of the two metals to provisional tolerable weekly intake (PTWI) [37]. Mean concentrations (microgram per kilogram) of Cd and Pb were 6.94 and 1.68 in chicken meat, 5.00 and 5.49 in pork meat, 1.91 and 1.90 in beef meat, and 1.35 and 1.22 in lamb meat samples, respectively. Lead was below the detection limit in turkey samples and mean cadmium concentration was 5.49 mg kg1. Mean concentrations (mg kg1) of Cd and Pb in chicken meat product samples were 3.16 and 4.15, 4.89 and 6.50 in pork meat product, 6.72 and 4.76 in beef meat product, and 9.12 and 5.98 in turkey meat product samples, respectively. Statistically signiﬁcant differences were found for Pb content in meats between the chicken and pork groups and the turkey and beef groups, whereas for Cd concentrations in meats, signiﬁcant differences were observed between the turkey and chicken, beef and lamb groups. In meat products, no clear differences were observed for both metals between the various groups.

14.5.6 SEAFOOD A comprehensive study on canned ﬁsh was conducted by Ikem and Egiebor [38]. The concentrations of Hg and 13 other trace metals in 104 samples purchased within the states of Georgia and Alabama (United States of America) were determined using CVAAS and ICP-AES. The ranges obtained in milligram per kilogram (wet weight) were as follows: Hg (0.02–0.74), Ag (0.0–0.20), As (0.0–1.72), Cd (0.0–0.05), Co (0.0–0.10), Cr (0.0–0.30), Cu (0.01–5.33), Fe (0.01–88.4), Mn (0.01–2.55), Ni (0.0–0.78), Pb (0.0–0.03), Sn (0.04–28.7), V (0.0–0.31), and Zn (0.14–97.8). Three tuna samples had Hg level above the European dietary limit of 0.5 mg kg1. The mean Hg concentrations in the tuna (285 mg kg1) and sardine (107 mg kg1) brands were higher than the averages posted by the pink salmon (36.1 mg kg1), red salmon (32.8 mg kg1), and mackerel (36.4 mg kg1) brands. Two tuna samples and a sardine sample exceeded the Australian permissible limit of 1 mg kg1 inorganic As (wet weight). Two samples (brand 15: herring) had Zn levels exceeding the FAO recommended limit of 40 mg kg1 and two pink salmons also exceeded the Brazilian regulatory limit of 0.1 mg kg1 for Cr. One tuna sample had a Cd level slightly exceeding the Codex Committee on Food Additives and Contaminants draft guideline of 0.50 mg kg1. However, the concentrations of Cd, Cu, and Pb were below the corresponding MAFF guidelines of 1.0, 30, and 2.0 mg kg1, respectively. Also, based on the United States Environmental Protection Agency (US EPA) health criteria for carcinogens, there were no health risks with respect to Cr, Cu, Pb, and Zn concentrations in canned ﬁshes analyzed. The result of the one-way analysis of variance (ANOVA) conducted on the data suggested that signiﬁcant variations (p < .05) existed

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 342 16.7.2008 11:02pm Compositor Name: MSubramanian

342

Handbook of Food Analysis Instruments

in the concentrations of Hg, As, Co, Cr, Cu, Fe, Mn, Sn, V, and Zn across the various ﬁsh species and canned ﬁsh brands analyzed. The estimated weekly intakes of Hg, As, Cd, Cu, Fe, Pb, Sn, and Zn for a 60 kg adult consuming 350 g of ﬁsh per week were below the respective PTWIs in microgram per kilogram body weight for Hg, 5; Cd, 7; As, 15; Pb, 25; Cu, 3500; Fe, 5600; Zn, 7000; and Sn, 14000. Seven slurry preparation procedures were evaluated for the simultaneous determination of Hg and Se by vapor generation-ICP-AES, using (1) aqua regia and sonication; (2) HCl, sonication and heating at 908C; (3) tetramethylammonium hydroxide (TMAH); (4) H2O2, sonication, addition of HCl, and heating at 908C; (5) K2S2O8, sonication, addition of HCl and heating at 908C (6) TMAH, sonication, ozonation, addition of HCl and heating at 908C; (7) Triton X-100, sonication, ozonation, addition of HCl, and heating at 908C [39]. For certiﬁed lobster material, while all procedures produced Hg results in agreement with the certiﬁed value, the Se results only by procedures 4 and 5 were in agreement with the certiﬁed value. Procedure 5 was adopted for the analysis of 6 oyster samples grown in the Santa Catarina Island coast with LODs of 0.08 mg g1 for Hg and 0.10 mg g1 for Se, for a sample mass of 20 mg in a ﬁnal volume of 15 mL. A simple and rapid method was developed for the determination of As, Cd, Pb, and Se in dogﬁsh muscle by GFAAS [40]. The sample was dissolved in TMAH after 10 min heating in a water bath at 608C. For As and Se measurement, Pd and Mg nitrates were added in solution as chemical modiﬁer; whereas for Cd and Pb, a mixture of Ir and Rh (250 mg each) was used as permanent modiﬁer. The LODs in dry sample were: As 0.4 mg g1, Se 0.6 mg g1, Cd 0.005 mg g1, and Pb 0.04 mg g1, respectively.

14.5.7 OTHER AND MIXED TYPES A total of 166 white sugar samples representative of the production of four Serbian sugar beet reﬁneries were investigated in 2003 for the content of Cu, Fe, and Zn [41]. After wet digestion, the concentrations were determined by FAAS. The mean values were 0.06, 0.37, and 0.02 mg kg1 for Cu, Fe, and Zn, respectively, signiﬁcantly different from the average content of European sugar factories (0.09, 0.28, and 0.07 mg kg1). According to the European Union standards, 76% of all investigated Serbian samples belonged to the second sugar quality category. Direct determination of Cr, Cu, and Ni in honey by GFAAS was developed using experimental design as an optimization tool [42]. Honey was diluted in water, hydrogen peroxide or=and nitric acid. Triton X-100 was added to minimize the matrix effect and the viscosity of the sample, and Pd used as chemical modiﬁer in all cases. For samples from Galicia (north-western Spain) the concentrations were in the range of (5.75 0.64) to (26.4 0.38) ng g1 of Cr, (79 7.8) to (2049 80) ng g1 of Cu, and (12.6 1.36) to (172 6.88) ng g1 of Ni. The continuous-ﬂow dialysis method described earlier for studying Fe in milk samples by GFAAS [23] was also linked to ICP-AES simultaneous multielement measurement to determine the bioavailability of ﬁve essential elements (Ca, Mg, P, Fe, and Zn using Y and Sc as internal standards) for various kinds of foods, i.e., infant formula, milk powder, kale, mungbean, chicken meat, jasmine rice, and Acacia pennata [43]. All studied elements were rapidly dialyzed in the ﬁrst 30 min of simulated intestinal digestion. A study was performed to monitor the exposure of the Korean population to As, Cd, Hg, and Pb from typical diets and to estimate the intake [44]. Foods were prepared for consumption (tableready) according to representative recipes and cooking methods and analyzed by ICP-AES (As, Cd, and Pb) and CV-gold amalgamation-AAS (Hg). The dietary intake of each element was estimated based on the mean food intake of the population. The contribution of foods to total intake was more inﬂuenced by the amount of food consumed. Although seaweeds and ﬁsh had the highest metal content, cooked rice was the most important contributor to Hg intake, and vegetables the most important contributor of Pb. Nevertheless, the estimated dietary intakes of As (38.5 mg per person per day), Cd (14.3 mg per person per day), Pb (24.4 mg per person per day), and

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 343 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

343

Hg (1.61 mg per person per day) from the 116 foods tested were well within the safe limits (under 30% of PTWIs). The dietary intake of As, Cd, Hg, and Pb was also studied among young German children with different food sources, consumption of own grown foodstuffs, and of products from the supermarket [45]. The study area comprised an industrialized and a rural area of West Germany. A total of 588 duplicate portions were collected daily, according to the WHO guideline, from 84 individuals between May and September 1998. Determination of As, Cd, Hg, and Pb was performed by AAS following high-pressure digestion of lyophilized samples. Geometric mean weekly intake (microgram per kilogram body weight) was as follows: As 1.4, Cd 2.3, Hg 0.16, and Pb 5.3. Geometric mean intake corresponded to the percentage of the PTWI as follows: As 9.7%, Cd 32%, Hg 3.3%, and Pb 21%. As and Hg intake were mainly inﬂuenced by ﬁsh consumption. The amount of cereals and bakery wares mainly determined the Cd and Pb intake. Children living in the industrialized area with a substantial food consumption of own grown vegetables or products from domestic animal products had no increased dietary intake of the metals. In a daily dietary intake study, the Zn content of 300 food and 79 beverage samples was determined using FAAS [46]. Mean Zn concentrations varied from 0.02 mg mL1 in fresh water to 71.0 mg g1 (fresh weight) in pork liver. The daily dietary intake of Zn for inhabitants of southeastern Spain was estimated to be 10.1 mg (5.5, 4.0, 0.5, and 0.1 mg Zn per day per person from foods of animal and vegetable origin, drinks, and other foods, respectively). Zinc levels found in high protein foods (meat, ﬁsh, milk products, eggs, dry fruits, cereals, and legumes) were signiﬁcantly higher than those found in food with a low protein content (vegetables, fruits, and drinks) (p < .001). A signiﬁcant linear correlation between Zn levels and the corresponding protein content of cereals, legumes, and dry fruits was found (r ¼ .754, p < .005). Zinc concentrations in milk samples were signiﬁcantly modiﬁed by the thermal treatment (p < .001), and the skimming (p < .05) and calcium enrichment processes (p < .001). Shellﬁsh Zn levels were also signiﬁcantly higher than those measured in ﬁsh (p < .05). Mean Zn concentrations found in cheese were statistically higher than those determined in the remaining milk products (p < .001). Zinc levels measured in distilled beverages were also statistically lower than those found in fermented ones (p < .001). The Cu contents of typical Brazilian foods were determined by AAS with wet oxidation [47]. Samples were bought in retail stores in cities of the southeast region of Brazil. The highest content was found in beef liver (60.6 mg kg1 fresh product) and lowest content found in milk and in ﬁsh ﬁllet (below 0.1 mg kg1 fresh product). Crude beans, Nescau, and whole wheat had Cu contents from 4.4 to 10.4 mg kg1 of fresh food. Other foods, such as fruits, vegetables, grain products, baked products, roots, and meat products had Cu contents varying from 0.2 to 4.1 mg kg1. The Cd levels in a range of food and drink (420 samples) commonly consumed in the Canary Islands (Spain) were determined by AAS to monitor the dietary intake [48]. The measured Cd concentrations combined with the food consumption data resulted in a total Cd intake in the Canary Islands of 0.16 mg kg1 of body weight per day, which was well below the respective PTWI of Cd of 1 mg kg1 of body weight per day determined by the FAO=WHO. The Al content of selected foods and food products in the United States which contained Al as an approved food additive was determined using GFAAS after acid or base digestion [49]. The results ranged between 1–27,000 mg kg1. Intake of Al from the labeled serving size of each food product was calculated. Cheese in a serving of frozen pizzas had up to 14 mg of Al, from basic sodium aluminium phosphate; whereas the same amount of cheese in a ready-to-eat restaurant pizza provided 0.03–0.09 mg. Many single serving packets of nondairy creamer had 50–600 mg kg1 Al as sodium aluminosilicate, providing up to 1.5 mg Al per serving. Many single serving packets of salt also had sodium aluminosilicate as an additive, but the Al content was less than in singleserving nondairy creamer packets. Acidic sodium aluminium phosphate was present in many food products, pancakes, and wafﬂes. Baking powder, some pancake=wafﬂe mixes and frozen products, and ready-to-eat pancakes provided the most Al of the foods tested, up to 180 mg per serving. Many products provide a signiﬁcant amount of Al compared to the typical intake of 3–12 mg per day reported from dietary Al studies conducted in many countries.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 344 16.7.2008 11:02pm Compositor Name: MSubramanian

344

Handbook of Food Analysis Instruments

The As concentrations in a range of foodstuffs, including vegetables, rice, and ﬁsh imported into the United Kingdom from Bangladesh, were surveyed using GFAAS [50]. The mean and range in all vegetables were 54.5 and 5–540 mg kg1, respectively. The highest values found were for the skin of Arum tuber, 540 mg kg1, followed by Arum Stein, 168 mg kg1, and Amaranthus, 160 mg kg1. Among the other samples, freshwater ﬁsh contained As levels between 97 and 1318 mg kg1. In contrast, the As content of the vegetables from the United Kingdom was approximately 2- to 3-fold lower. The As concentration in different foods from south-east Spain was determined by HG-AAS (with standard addition) after mineralization with an HNO3–HClO4 mixture [51]. The highest levels were found in seafood, cereals, meat, and meat by-products. Furthermore, the As level in meat was signiﬁcantly higher than that in sausages (p < .05). In cereals, As concentrations in corn and white rice samples were signiﬁcantly higher (p < .01) than those in wheat by-products. Mean concentrations in cheese were statistically lower than those in other dairy products (p < .01). The estimated daily intake of As in Spanish diet was 221 mg per day.

REFERENCES 1. Byrne, D., Setting maximum levels for certain contaminants in foodstuffs (Commission Regulation No. 466=2001), Off. J. Euro. Communities L77, 16.3.2001. 2. Intawongse, M. and Dean, J.R., In-vitro testing for assessing oral bioaccessibility of trace metals in soil and food samples, Trends Anal. Chem. 25, 876, 2006. 3. Haouet, M.N. et al., Estimation of uncertainty for the determination of mercury in food by CVAAS, Accred. Qual. Assur. 11, 17, 2006. 4. Jorhem, L. et al., Evaluation of measurement data for Cd, Cr and Pb in certain uncontaminated foodstuffs published in surveys: Analytical quality vs. uncertainty of measurements, Accred. Qual. Assur. 10, 647, 2006. 5. Pedro, N.A.R., de Oliveira, E., and Cadore, S., Study of the mineral content of chocolate ﬂavoured beverages, Food Chem. 95, 94, 2006. 6. de Sousa, R.A. et al., Determination of metals in bottled coconut water using an inductively coupled plasma optical emission spectrometer, J. Food Composition Anal. 18, 399, 2005. 7. Sepe, A. et al., Evaluation of aluminium concentrations in samples of chocolate and beverages by electrothermal atomic absorption spectrometry, Food Addit. Contam. 18, 788, 2001. 8. Seruga, M. and Laslavic, B., Aluminium content of fruit juices and tea beverages, Deutsche LebensmittelRundschau 102, 435, 2006. 9. Street, R. et al., The status of micronutrients (Cu, Fe, Mn, Zn) in tea and tea infusions in selected samples imported to the Czech Republic, Czech J. Food Sci. 24, 62, 2006. 10. Lara, R. et al., Trace element determination of Argentine wines using ETAAS and USN-ICP-OES, Food Chem. Toxicol. 43, 293, 2005. 11. Huguet, M.E.R., Monitoring of Cd, Cr, Cu, Fe, Mn, Pb and Zn in ﬁne Uruguayan wines by atomic absorption spectroscopy, At. Spectros. 25, 177, 2004. 12. Larcher, R. and Nicolini, G., Survey of 22 mineral elements in wines from Trentino (Italy) using ICP-OES, Italian J. Food Sci. 13, 233, 2001. 13. Asfaw, A. and Wibetoe, G., Direct analysis of beer by ICP-AES: A very simple method for the determination of Cu, Mn and Fe, Microchim. Acta 152, 61, 2005. 14. Canuto, M.H. et al., Determination of Al, Cd, and Pb in Brazilian sugar cane spirit, cachaca by ETAAS using matrix-matched calibration, At. Spectros. 25, 140, 2004. 15. Kokkinofta, R. et al., Authenticity of the traditional Cypriot spirit ‘‘zivania’’ on the basis of metal content using a combination of coupled plasma spectroscopy and statistical analysis, J. Agric. Food Chem. 51, 6233, 2003. 16. Hernandez-Caraballo, E.A. et al., Classiﬁcation of Venezuelan spirituous beverages by means of discriminant analysis and artiﬁcial neural networks based on their Zn, Cu and Fe concentrations, Talanta 60, 1259, 2003. 17. Leotsinidis, M., Alexopoulos, A., and Kostopoulou-Farri, E., Toxic and essential trace elements in human milk from Greek lactating women: Association with dietary habits and other factors, Chemosphere 61, 238, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 345 16.7.2008 11:02pm Compositor Name: MSubramanian

Atomic Absorption, Atomic Emission and Plasma Spectroscopies

345

18. Antunovic, Z. et al., Concentrations of selected toxic elements (cadmium, lead, mercury and arsenic) in ewe milk in dependence on lactation stage, Czech J Animal Sci. 50, 369, 2005. 19. Sola-Larranaga, C. and Navarro-Blasco, I., Preliminary chemometric study of minerals and trace elements in Spanish infant formulae, Anal. Chim. Acta 555, 354, 2006. 20. Dominguez, R. et al., Fe, Cu and Zn distribution in different components of commercial infant formulas, Eur. Food Res. Technol. 221, 529, 2005. 21. Bermejo, P., Iron and zinc in hydrolised fractions of human milk and infant formulas using an in vitro method, Food Chem. 77, 361, 2002. 22. Pena, E. et al., Enzymolysis approach to compare Cu availability from human milk and infant formulas, J Agric. Food Chem. 52, 4887, 2004. 23. Promchan, J. and Shiowatana, J., A dynamic continuous-ﬂow dialysis system with on-line electrothermal atomic-absorption spectrometric and pH measurements for in-vitro determination of iron bioavailability by simulated gastrointestinal digestion, Anal. Bioanal. Chem. 382, 1360, 2005. 24. Momen, A.A. et al., Investigation of four digestion procedures for multi-element determination of toxic and nutrient elements in legumes by inductively coupled plasma-optical emission spectrometry, Anal. Chim. Acta 565, 81, 2006. 25. Momen, A.A. et al., Development and validation of routine analysis methods for the determination of essential, nonessential, and toxic minor and trace elements in cereal and cereal ﬂour samples by inductively coupled plasma-atomic emission spectrometry, J. AOAC Int. 88, 1797, 2005. 26. Erdogan, S., Erdemoglu, S.B., and Kaya, S., Optimisation of microwave digestion for determination of Fe, Zn, Mn and Cu in various legumes by ﬂame atomic absorption spectrometry, J. Sci. Food Agric. 86, 226, 2006. 27. Anzano, J.M. and Ruiz-Gil, M., Comparison of microwave acid digestion with the wet digestion and ashing methods for the determination of Fe, Mn, and Zn in food samples by ﬂame AAS, At. Spectros. 26, 28, 2005. 28. Rivero-Huguet, M. et al., Concentrations of As, Ca, Cd, Co, Cr, Cu, Fe, Hg, K, Mg, Mn, Mo, Na, Ni, Pb, and Zn in Uruguayan rice determined by atomic absorption spectrometry, At. Spectros. 27, 48, 2006. 29. Cabrera, C. et al., Mineral content in legumes and nuts: Contribution to the Spanish dietary intake, Sci. Total Environ. 308, 1, 2003. 30. Cubadda, F., Raggi, A., and Marconi, E., Effects of processing on ﬁve selected metals in the durum wheat food chain, Microchem. J. 79, 97, 2005. 31. Ma, G. et al., Phytate, calcium, iron, and zinc contents and their molar ratios in foods commonly consumed in China, J. Agric. Food Chem. 53, 10285, 2005. 32. Ajtony, Z. et al., Determination of total selenium content in cereals and bakery products by ﬂow injection hydride generation graphite furnace atomic absorption spectrometry applying in-situ trapping on iridiumtreated graphite platforms, Microchim. Acta 150, 1, 2005. 33. Perez, A.L., Smith, B.W., and Anderson, K.A., Stable isotope and trace element proﬁling combined with classiﬁcation models to differentiate geographic growing origin for three fruits: Effects of subregion and variety, J Agric. Food Chem. 54, 4506, 2006. 34. Olalla, M. et al., Nutritional study of copper and zinc in grapes and commercial grape juices from Spain, J. Agric. Food Chem. 52, 2715, 2004. 35. Intawongse, M. and Dean, J.R., Uptake of heavy metals by vegetable plants grown on contaminated soil and their bioavailability in the human gastrointestinal tract, Food Additives Contam. 23, 36, 2006. 36. Bou, R. et al., Validation of mineralisation procedures for the determination of selenium, zinc, iron and copper in chicken meat and feed samples by ICP-AES and ICP-MS, J. Anal. At. Spectrom. 19, 1361, 2004. 37. Gonzalez-Weller, D. et al., Lead and cadmium in meat and meat products consumed by the population in Tenerife Island, Spain, Food Addit. Contam. 23, 757, 2006. 38. Ikem, A. and Egiebor, N.O., Assessment of trace elements in canned ﬁshes (mackerel, tuna, salmon, sardines and herrings) marketed in Georgia and Alabama (United States of America), J. Food Comp. Anal. 18, 771, 2005. 39. dos Santos, E.J. et al., Evaluation of slurry preparation procedures for the simultaneous determination of Hg and Se in biological samples by axial view ICPOES using on-line chemical vapor generation, Anal. Chim. Acta 548, 166, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C014 Final Proof page 346 16.7.2008 11:02pm Compositor Name: MSubramanian

346

Handbook of Food Analysis Instruments

40. Giacomelli, M.B.O. et al., Determination of As, Cd, Pb and Se in DORM-1 dogﬁsh muscle reference material using alkaline solubilization and electrothermal atomic absorption spectrometry with Ir þ Rh as permanent modiﬁers or Pd þ Mg in solution, Spectrochim. Acta 57B, 2151, 2002. 41. Skrbic, B. and Gyura, J., Survey on some contaminants in white sugar from Serbian sugar beet reﬁneries, Food Addit. Contam. 23, 31, 2006. 42. Garcia, J.C.R. et al., Direct and combined methods for the determination of chromium, copper, and nickel in honey by electrothermal atomic absorption spectroscopy, J. Agric. Food Chem. 53, 6616, 2005. 43. Judprasong, K. et al., A continuous-ﬂow dialysis system with inductively coupled plasma optical emission spectrometry for in vitro estimation of bioavailability, J. Anal. At. Spectrom. 20, 1191, 2005. 44. Lee, H.S. et al., Dietary exposure of the Korean population to arsenic, cadmium, lead and mercury, J. Food Comp. Anal. 19, S31, 2006. 45. Wilhelm, M. et al., Consumption of homegrown products does not increase dietary intake of arsenic, cadmium, lead, and mercury by young children living in an industrialized area of Germany, Sci. Total Environ. 343, 61, 2005. 46. Terres, C. et al., Zinc levels in foods from southeastern Spain: Relationship to daily dietary intake, Food Addit. Contam. 18, 687, 2001. 47. Ferreira, K.S., Gomes, J.C., and Chaves, J.B.P., Copper content of commonly consumed food in Brazil, Food Chem. 92, 29, 2005. 48. Rubio, C. et al., Cadmium dietary intake in the Canary Islands, Spain, Environ. Res. 100, 123, 2006. 49. Saiyed, S.M. and Yokel, R.A., Aluminium content of some foods and food products in the USA, with aluminium food additives, Food Addit. Contam. 22, 234, 2005. 50. Al Rmalli, S.W. et al., A survey of arsenic in foodstuffs on sale in the United Kingdom and imported from Bangladesh, Sci. Total Environ. 337, 23, 2005. 51. Delgado-Andrade, C. et al., Determination of total arsenic levels by hydride generation atomic absorption spectrometry in foods from south-east Spain: estimation of daily dietary intake, Food Addit. Contam. 20, 923, 2003.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 347 14.7.2008 11:34am Compositor Name: BMani

15 Autoﬂuorescence Spectroscopy in Food Analysis Charlotte Møller Andersen, Jens Petter Wold, and Søren Balling Engelsen CONTENTS 15.1 15.2 15.3

Introduction ........................................................................................................................ 347 Basic Deﬁnitions ................................................................................................................ 348 Factors Affecting Fluorescence ......................................................................................... 350 15.3.1 Quenching ........................................................................................................... 350 15.3.2 Inner Filter ........................................................................................................... 350 15.3.3 Concentration ...................................................................................................... 350 15.3.4 Molecular Environment ...................................................................................... 351 15.3.5 Scatter .................................................................................................................. 352 15.4 Fluorescence Spectrophotometer ....................................................................................... 353 15.4.1 Sampling Geometry ............................................................................................. 353 15.4.2 Analysis of Fluorescence Data ............................................................................ 354 15.5 Fluorophores in Food ........................................................................................................ 354 15.6 Applications of Fluorescence to Food Analysis ................................................................ 354 15.6.1 Direct Fluorescence Measurements ..................................................................... 355 15.6.2 Indirect Fluorescence Measurements .................................................................. 356 15.6.3 Fluorescence and Food Authenticity ................................................................... 357 15.6.4 Fluorescence and Process Analytical Technology .............................................. 358 15.7 Future Trends ..................................................................................................................... 359 15.7.1 Process Analytical Technology Monitoring by Fluorescence Spectroscopy ...... 359 15.7.2 Fluorescence Imaging .......................................................................................... 360 References ..................................................................................................................................... 361

15.1 INTRODUCTION Fluorescence spectroscopy is becoming a more and more popular instrumental technique for providing direct and indirect exploratory information about chemical and physical properties of food products (Novales et al. 1996; Munck et al. 1998; Zandomenghi et al. 2005; Christensen et al. 2006). The primary reasons are the high speciﬁcity, the high sensitivity, and that several substances inherent to food systems exhibit intrinsic ﬂuorescence, such as proteins, vitamins, secondary metabolites, pigments, toxins, and ﬂavoring compounds. Fluorescence spectroscopy analysis often involves the use of an extrinsic probe developed for speciﬁc analyses or requires a sample preparation step before the ﬂuorescence measurement to reduce the effect of other inﬂuencing parameters. However, for food analysis, the intrinsic ﬂuorescence of the intact food, called autoﬂuorescence, can be measured. The use of autoﬂuorescence increases the speed of analysis considerably, facilitates nondestructive analysis, and makes ﬂuorescence a potential method for online or at-line applications. Furthermore, direct measurements on 347

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 348 14.7.2008 11:34am Compositor Name: BMani

348

Handbook of Food Analysis Instruments

foods enhance the scientiﬁc exploitation of the measurements, allowing exploratory studies of the more complex relationships within the sample. This chapter provides an overview of the ﬂuorescence spectroscopic techniques relevant for applications within food research and industry. Firstly, the fundamentals of ﬂuorescence spectroscopy are described with the focus on parameters that are important when measuring intact food samples. Secondly, a short section on ﬂuorescence instrumentation is given together with a description of sampling geometry and an introduction to ﬂuorescence data analysis. Thirdly, the ﬂuorophores present in foods are explained and exempliﬁed by applications within food research.

15.2 BASIC DEFINITIONS Fluorescence is caused by an immediate emission of light by molecules subsequent to absorption of UV or visible light (Lackowicz 1999). Only light in these wavelength regions possesses enough energy to excite electrons to a singlet state. The ﬂuorescence phenomenon is a three-stage process. After excitation (1) to an electronic singlet state, the ﬂuorophore in this excited state undergoes internal conversion (2) to the lowest vibrational level in the excited state (Figure 15.1), losing part of the energy as heat. Finally the ﬂuorophore returns to the ground state by emission (3) of longer wavelength light. The emission takes place 1011 to 107 s, after the excitation and it is of

S2

Excitation

Emission S1

Rayleigh SV

Raman scatter

Internal conversion S0

FIGURE 15.1 Jablonski diagram: S0 denotes the electronic ground level; S1 and S2 are electronically excited singlet states and Sv is a virtual electronic state. The horizontal lines indicate the vibrational levels within each electronic state and the vertical arrows illustrate the various forms of electronic transitions and emission of light.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 349 14.7.2008 11:34am Compositor Name: BMani

349

Autoﬂuorescence Spectroscopy in Food Analysis

lower energy than the excitation light, meaning that it has longer wavelengths. The difference between excitation and emission wavelengths is called the Stokes shift. Each electronic state of the ﬂuorophore has several associated vibrational levels and when it returns to the ground state it can reach any of the vibrational levels, resulting in emission of even lower energy and increased Stokes shift. Each molecule or substructure able to emit ﬂuorescence, called a ﬂuorophore, can only be excited by light of speciﬁc wavelengths and will only emit light at characteristic higher wavelengths when the molecule returns to its more stable ground state. Thus, all ﬂuorescent molecules have unique ﬂuorescence properties. In the ideal case, excitation and emission spectra are identical and can be considered as mirror images of each other (Figure 15.2a) due to the same electronic transitions taking place. Moreover, the visual appearance of the emission spectrum does not depend on the excitation wavelength, i.e., if the excitation light is at a wavelength different from the absorption maximum, the molecule will absorb less radiation and the emission will be of lower intensity. This process will result in broad emission=excitation peaks when measuring emission spectra for a number of neighboring excitation wavelengths (Figure 15.2b). The use of both excitation and emission spectra gives ﬂuorescence spectroscopy a high degree of selectivity. Two molecules with similar excitation spectra may have different emission properties. Fluorescence is sometimes called luminescence, which covers the emission of light from molecules in an electronically excited state. Phosphorescence is a similar, but slower luminescence phenomenon. The excited molecule goes through an intermediate excited triplet state from where light is emitted by returning to the ground state. The instrumentation and spectral region for measurement of phosphorescence are almost identical to ﬂuorescence spectroscopy, but will not be further pursued here. Only a few molecules that absorb UV or visible light are able to emit ﬂuorescence. A ﬂuorescent molecule is normally an aromatic compound or a molecule with a highly conjugated structure. In addition, low internal motional freedom is required. The less the ﬂexibility, the higher is the possibility for being a strong ﬂuorescent molecule. The indole ring of tryptophan is one of the strongest and most well-characterized ﬂuorophores omnipresent in food systems. It has excitation maximum at 280 nm and emission maximum at 350 nm. Other important food ﬂuorophores are NADH, FAD, riboﬂavin, and chlorophyll. Compared with other optical spectroscopies, ﬂuorescence spectroscopy is a factor of 100 more sensitive (down to ppb levels) due to highly efﬁcient detectors and light sources and the low background from interfering signals.

500 400 300 200 100 0 250

(a)

Stokes shift

Fluorescence intensity

Fluorescence intensity

600

500 400 300 200 100 0 450 400

Em

iss

350

ion

300

350 400 450 Wavelength (nm)

300

(nm

)

500

250

260

270 280

290 300

340 330

)

n (nm

tatio

Exci

320 310

(b)

FIGURE 15.2 Fluorescence properties of tryptophan measured in a concentration of 105 M. (a) Emission spectrum (solid) with excitation set to 280 nm and excitation spectrum (dotted) with emission set to 350 nm, the Stokes shift is illustrated and (b) ﬂuorescence landscape.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 350 14.7.2008 11:34am Compositor Name: BMani

350

Handbook of Food Analysis Instruments

15.3 FACTORS AFFECTING FLUORESCENCE The ﬂuorescence properties of a molecule depend strongly on the sample matrix. The concentration, turbidity, local molecular environment, pH, and temperature inﬂuence the measurement due to phenomena such as quenching, inner-ﬁlter effects, scatter, etc. Some of these will be explained in the following subsections.

15.3.1 QUENCHING The ﬂuorescence intensity may decrease as an effect of intra- or intermolecular interactions either within the ﬂuorescent molecule itself, in the solvent, or with other molecules in the sample matrix. This phenomenon is called ﬂuorescence quenching. Quenching can be either static or dynamic. Static quenching takes place when the ﬂuorescent molecule forms a nonﬂuorescent complex with the quencher molecule, which inhibits the formation of an excited state. While static quenching does not depend on diffusion or molecular collisions, dynamic quenching depends on such molecular movements. It is caused by deactivation of the excited state by contact with another molecule in the sample matrix called the quencher. Oxygen, heavy metals, halide ions, organic and inorganic nitro compounds as well as intramolecular interactions are known to induce dynamic quenching. The molecule returns to the ground state without ﬂuorescence emission and without chemical modiﬁcation (Lackowicz 1999). Higher temperatures may also lead to dynamic quenching, as the increased molecular velocities give rise to more molecular collisions.

15.3.2 INNER FILTER Another type of mechanism that will reduce the measured ﬂuorescence intensity is the so-called inner-ﬁlter effects, which are due to absorption of excitation or emitted light by either the ﬂuorophore itself or by chromophores in the sample matrix. Inner ﬁlter occurs when a chromophore reabsorbs the emitted ﬂuorescence or when a nonﬂuorescent chromophore absorbs parts of the excitation light. The result is decreased ﬂuorescence intensity or distortion of bandshapes. Concentration quenching and inner-ﬁlter effects can normally be handled by diluting the sample or by reducing the optical path length. However, diluting the sample reduces the concentration of other relevant ﬂuorophores and changes the molecular interactions of the intact food matrix.

15.3.3 CONCENTRATION Like absorption spectroscopy, ﬂuorescence spectroscopy depends linearly on the concentration, but it has to be within a certain concentration range in order to avoid artifacts such as quenching and inner-ﬁlter effects. For transparent solutions with absorbance below 0.05, the ﬂuorescence intensity is linearly related to the concentration of the ﬂuorophore (Lackowicz 1999). In this case, the ﬂuorescence intensity, If, depends on the intensity of the incident light, I0, the molar absorptivity, «, the ﬂuorescence quantum yield, w, the optical length, l, and the molar concentration of the ﬂuorophore, c: If ¼ 2:3wf I0 «cl

(15:1)

The quantum yield is deﬁned as the number of emitted photons relative to the number of absorbed photons. Together with the molar absorptivity, it is a measure of the effectiveness of the ﬂuorophore. Figure 15.3 illustrates how the ﬂuorescence intensity of a pure tryptophan solution depends on the concentration. At low concentrations the ﬂuorescence intensity increases with concentration. At concentrations above 104 M, the ﬂuorescence intensity starts to decline due to collision quenching and inner-ﬁlter effects. Not until the tryptophan concentration reaches 105 M does the absorbance

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 351 14.7.2008 11:34am Compositor Name: BMani

351

Autoﬂuorescence Spectroscopy in Food Analysis 2

400 1

Absorbance

Fluorescence intensity

600

200

0

⫺8

⫺7

⫺6 ⫺5 ⫺4 Concentration, log(M )

⫺3

⫺2

0

FIGURE 15.3 Fluorescence intensity of tryptophan measured with excitation at 280 nm and emission at 357 nm versus the concentration (solid) and the corresponding absorbance measured at 280 nm (dotted). The horizontal line shows absorbances below 0.05 and the gray area indicates the concentration levels where the absorbance is below this level.

level reach 0.05 (horizontal line in Figure 15.3). Only at concentration levels below this threshold is Equation 15.1 valid.

15.3.4 MOLECULAR ENVIRONMENT In addition to concentration, ﬂuorescence intensity depends on the molecular environments such as polarity, pH, temperature, etc. For example, tryptophan residues buried in the interior of a protein exhibit different ﬂuorescence than tryptophan exposed to the hydrophilic solvent on the surface. Figure 15.4 shows an example of emission spectra measured on the milk proteins: a-lactalbumin,

Fluorescence intensity

1.0

0.8

0.6

0.4

0.2

0

300

350 400 Emission (nm)

450

FIGURE 15.4 Fluorescence emission spectra of a-lactalbumin (dotted), b-lactoglobulin (dashed), and casein (solid) dissolved in water. Excitation was set to 282 nm.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 352 14.7.2008 11:34am Compositor Name: BMani

352

Handbook of Food Analysis Instruments

b-lactoglobulin, and casein. Casein, which is the least structured protein, has emission maximum at the highest wavelength. In casein, tryptophan is more exposed to the solvent, which in this case is water. A similar change in emission spectra is observed when comparing a ﬂuorophore dissolved in a hydrophilic solvent, such as water, and the same ﬂuorophore dissolved in a hydrophobic solvent, i.e., an organic solvent. An increase in hydrophobicity of the ﬂuorophore’s environment entails relaxation of the excited state to a lower vibrational level before emitting ﬂuorescence. This results in emission of lower energy than obtained from more polar solvents. These differences make it possible to study conformational protein changes as well as denaturation and interaction of proteins with other food components.

15.3.5 SCATTER Scatter of light is a phenomenon that may disturb the measured ﬂuorescence signal. Light is primarily caused by small particles in the samples and is observed by the ﬂuorescence detector as false emitted light. Light scatter disturbs the ﬂuorescence measurements when the Stokes shift is small and the signal overlaps with the analyte emission. There are two main scatter phenomena: elastic Rayleigh scatter and inelastic Raman scatter. Elastic Rayleigh scatter occurs at the same wavelength as the excitation wavelength (Figure 15.5). In Rayleigh scatter there is no electronic interaction between the light and matter, for which reason the Rayleigh scatter will be of the same energy and found at the same wavelengths. The intensity of the Rayleigh scatter varies with the fourth power of the wavelength. Thus, the effect can be minimized using higher excitation wavelengths. Second-order Rayleigh scatter appears at twice the excitation wavelength. Inelastic Raman scatter is much weaker than Rayleigh scatter and is often not observed if the ﬂuorescence intensity is high. Raman scatter appears at longer wavelengths than the excitation wavelength with a constant distance in frequency (not wavelength) to the Rayleigh scatter peak. It is caused by vibrational interactions with the solvent. For water, the Raman scatter peak appears at 3600 cm1 lower wave number than the excitation light (Lackowicz 1999). This corresponds to the Raman shift of the O-H stretching vibrations observed in the mid-infrared region or in dedicated Raman spectrometers. Raman scatter is important because it may overlap with the ﬂuorescence signal and especially at low ﬂuorescence intensities, it can make up a large part of the measured signal. Correction of Raman scatter can be performed by subtracting the ﬂuorescence signal of the pure solvent from the measured sample signal (Jiji and Booksh 2000).

30 25 yleig

h

gt

en

el

Ex ci

400 350 Emis sion 300 wave lengt h

ta tio

450

FIGURE 15.5

r Ra

n

500

w av

5 0

320 310 300 290 280 270 260 250 240 230 orde

Rama

n

10

1st-

15

h

2nd-order Rayleigh

20

250

Appearance of Rayleigh and Raman scatter in a ﬂuorescence landscape.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 353 14.7.2008 11:34am Compositor Name: BMani

353

Autoﬂuorescence Spectroscopy in Food Analysis

15.4 FLUORESCENCE SPECTROPHOTOMETER In principle, a ﬂuorescence spectrometer is simple, requiring only a broadband light source for the near ultraviolet and the visible region of the electromagnetic spectrum (e.g., xenon lamp), two wavelength selectors, one for selecting the excitation wavelength and one for selecting the emission wavelength, and last but not least, a sensitive but color-blind detector such as a photomultiplier tube to record the signal (Figure 15.6). The optical components and sensors can be quite affordable, since most of the measurements are done in the near UV and visible region. Compared with traditional ﬂuorescence spectroscopy, autoﬂuorescence offers some challenges, because the intrinsic ﬂuorescence in many cases is of very low intensity and thus demands a powerful excitation source. This problem becomes more pronounced if collection of narrow spectral bands is required, giving a low spectral throughput. The wavelength selectors (monochromators) are the heart of the ﬂuorescence spectrometer. In the simplest case these are simply ﬁlters, but normally in a scanning instrument these are gratings followed by slits. In a scanning instrument, slit widths of the monochromators can be adjusted and optimized for the current application. Larger slit widths give higher spectral throughput (increased signals) and therefore higher signal-to-noise ratios. On the other hand, higher spectral resolution is obtained using smaller slit widths, but at the expense of light intensity. In most applications, the resolution is not very important, since emission spectra of most ﬂuorophores are rather broad. The correct signal is obtained from a spectroﬂuorometer with a light source that yields a constant photon output at all wavelengths, monochromators that pass photons at all wavelengths with equal efﬁciency, and a detector measuring photons of all wavelengths with similar intensity. However, such a system does not exist. All ﬂuorescence instruments will vary somewhat from the ideal situation. Thus, to compare measurements taken over a long period or on different instruments, standardization must be performed. This is done by correcting the measured data according to a well-deﬁned standard measurement (Costa et al. 1982; Melhuish 1982).

15.4.1 SAMPLING GEOMETRY In an ideal measurement situation, the sample is transparent, the concentration of the ﬂuorophores are at an appropriate level, the signals of all intrinsic ﬂuorophores remain independent of each other and inﬂuencing parameters such as quenching and inner-ﬁlter effects render insigniﬁcant. In this case, ﬂuorescence is measured using a 908 geometry, collecting the ﬂuorescence orthogonally to the

Light source Photomultiplier Emission slit exit Mirror Excitation slit entry

Excitation slit exit Grating Excitation monochromator

Mirror

Emission slit entry

Grating

Emission monochromator Sample compartment

FIGURE 15.6

Schematic overview of a ﬂuorescence spectrophotometer.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 354 14.7.2008 11:34am Compositor Name: BMani

354

Handbook of Food Analysis Instruments

excitation light. However, this ideal measurement situation is rarely the case when measuring intact food samples. Food samples are often opaque, turbid multiphase systems where only a tiny part of the light can pass through to cause excitation. In such situations, front-face ﬂuorescence spectroscopy can be applied. By changing the angle between the sample and the excitation beam to approximately 308, only ﬂuorescence emitted from the surface is measured. A linear relationship between the ﬂuorescence intensity and the analyte concentration is obtained, since most of the incident light is absorbed near the surface of the sample. Thus, it is important that the properties one wants to measure are represented in the outer layer of the sample.

15.4.2 ANALYSIS

OF

FLUORESCENCE DATA

Even today, most ﬂuorescence measurements are collected and analyzed as univariate measurements. This method of analysis is reasonable for chemically prepared samples with only one ﬂuorophore and when the samples are not inﬂuenced by interferences or difﬁculties. However, this approach is not useful when measuring intact food samples, for which reason excitation or emission spectra (or a whole ﬂuorescence landscapes) are often measured. A ﬂuorescence landscape is obtained when an emission spectrum is measured for a number of (regularly spaced) excitation wavelengths. In the case of measurement of entire emission spectra of landscapes, multivariate chemometric methods have proven effective in analyzing the complex measurements containing a large number of variables. Principal component analysis (PCA) or partial least squares (PLS) regression are commonly used for interpretation, outlier detection, quantiﬁcation, etc. (Nørgaard 1995a; Martens and Næs 1989). When measuring ﬂuorescence landscapes of several samples analysis by multi-way, chemometric methods make unique decomposition of the data into pure ﬂuorescent components possible (Munck et al. 1998, Christensen et al. 2006). The PARAFAC (PARAllel FACtor analysis) algorithm is well suited for analyzing the underlying phenomena in ﬂuorescence data, as it gives estimates of the excitation and emission proﬁles of each inherent ﬂuorophore as well as their relative concentrations. An advantage of three-way decomposition over the two-way decomposition is that the analysis and calibration are possible irrespective of the presence of new and unknown interferences (Booksh and Kowalski 1994).

15.5 FLUOROPHORES IN FOOD Food contains a number of intrinsic ﬂuorescent compounds that are important for the product quality such as technological properties, nutritional value, and overall composition. The compounds that are most often used within food research are shown in Figure 15.7.

15.6 APPLICATIONS OF FLUORESCENCE TO FOOD ANALYSIS The number of application studies of autoﬂuorescence and chemometrics in analysis of intact foods has increased during the past decade. This has taken place together with the introduction of ﬂuorescence spectroscopy for fast, nondestructive measurements. A classic example of a ﬂuorescence application is the detection of aﬂatoxins in ﬁgs (Steiner et al. 1988). The aﬂatoxins are strongly ﬂuorescent compounds and exhibit strong, bright greenish-yellow ﬂuorescence emission with excitation at 365 nm. This is used by some green grocers who have blue light in their display shelves. Fluorescence has been applied on various kinds of food from ﬁsh and meat to fruit and vegetables. However, it is almost the same ﬂuorophores that make up the ﬂuorescence emission in the different food items. Protein ﬂuorescence is one of the most frequently studied ﬂuorophores in food due to the strong emission properties of tryptophan. Tryptophan ﬂuorescence has been used as an indicator of protein structure in dairy products, evaluated by minor shift in the emission

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 355 14.7.2008 11:34am Compositor Name: BMani

355

Autoﬂuorescence Spectroscopy in Food Analysis 650

Excitation (nm)

600

Amino acids

550 Chlorophyll Porphyrin

500 450 400 350 300 250

300 350 400 450 500 550 600 650 700 Emission (nm)

NADH

ATP Vitamins (A, B2, B6, E)

FIGURE 15.7 (See color insert following page 240.) Fluorescence landscape of a food sample and emission maxima of the most relevant food ﬂuorophores.

maximum (Herbert et al. 2000; Karoui et al. 2003). Furthermore, it has been correlated to the texture of meat emulsions and sausages and to meat tenderness (Frencia et al. 2003; Allais et al. 2004). Many other food products exhibit tryptophan ﬂuorescence, including wheat, beer, and sugar. For a long time riboﬂavin has been known as the initiator of photooxidation in dairy products and several applications on the subject have been published (Becker et al. 2003; Christensen et al. 2005). It has shown potential in evaluating wheat ﬂour reﬁnement and milling efﬁciency and has been identiﬁed in beer. Chlorophyll, or rather the porphyrin, is another family of strong and omnipresent ﬂuorophores. It is present in a large variety of food products including plants, plant oils, meat, and ﬁsh oils (Engelsen 1997; Pedersen et al. 2003; Guimet et al. 2004; Zandomenghi et al. 2005). Fluorescence of chlorophyll has revealed possibilities to study quality changes of apples and papaya (Song et al. 1997; Bron et al. 2004) and recently, ﬂuorescence of chlorophyll and porphyrins has shown photoinitiation abilities in dairy products (Wold et al. 2006a,b). Connective tissue in meat and ﬁsh as well as bone and cartilage possesses ﬂuorescence properties differing from the meat and can thus be used for quality control (Jensen et al. 1986). Other less-studied food ﬂuorophores are Maillard reaction products, which have been applied to determine heat treatment in dairy products, NADH used for estimating freshness of meat and ﬁsh, ferulic acids in wheat, and iso-a-acids in beer. When measuring a food sample, the ﬂuorescence signals arising from a number of ﬂuorophores make the measurement complex. Ideally, the signal contributions of the different ﬂuorophores are additive, meaning that the measured signal of a given sample can be expressed as the sum of ﬂuorescence contributions of all inherent ﬂuorophores. However, interferences such as reabsorption, various forms of light scattering, unknown ﬂuorophores, and quenching may inﬂuence the ﬂuorescence spectra. An example of such interferences is the presence of inner-ﬁlter effects in undiluted oils (Sikorska et al. 2004). Tocopherol could hardly be detected in the neat oils. In the diluted oils, a clear signal originating from tocopherol could be identiﬁed. However, the emission intensities from polyphenolic and thermal-induced compounds decreased considerably upon dilution. In the following subsections, four speciﬁc examples of food ﬂuorescence applications are described in more detail.

15.6.1 DIRECT FLUORESCENCE MEASUREMENTS Relatively few studies of food product have shown the direct measurement of the ﬂuorophore concentration from ﬂuorescence spectroscopic data. Riboﬂavin is a highly ﬂuorescent molecule with

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 356 14.7.2008 11:34am Compositor Name: BMani

356

Handbook of Food Analysis Instruments

Predicted riboflavin content (ppm)

1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.2

0.4

0.6 0.8 1.0 1.2 1.4 1.6 Measured riboflavin content (ppm)

1.8

FIGURE 15.8 Predicted riboﬂavin content versus the chemically measured values. The predictions were obtained by PLS calibration using scores obtained from a PARAFAC model as X-variables.

excitation maxima at 270, 370, and 450 nm and emission maximum in the range 525–531 nm (Fox and Thayer 1998). The content was determined chemically in plain yoghurt exposed to light for up to 35 days and subsequently predicted from ﬂuorescence measurements by chemometrics (Christensen et al. 2005). Squared correlation coefﬁcients between measured and predicted riboﬂavin were 0.97 and the error was estimated to 7% of the mean riboﬂavin content in the samples (Figure 15.8).

15.6.2 INDIRECT FLUORESCENCE MEASUREMENTS Lipid oxidation is an important area within food research, since exposure of products such as cheese, meat, and other lipid-containing products to light and oxygen for a long period leads to offﬂavor formation, discoloration, nutrient loss, and formation of toxic compounds, which rapidly impair product quality. Several methods are applied for measuring lipid oxidation in foods, including peroxide value, formation of thiobarbituric reactive substances, iodine value, volatile compounds by gas chromatography (GC), and radicals by electron spin resonance (ESR). However, these determinations are all rather time consuming and strongly invasive. Fluorescence measurements have shown potential to predict some of these oxidation parameters. For example, ﬂuorescence measurements of frying oil during production of spring rolls were correlated to the deterioration (Engelsen 1997). In this industrial study, one frying oil batch was used for 20 days after which the plant was cleaned and restarted with fresh plant oil. Daily oil samples collected during the 4 weeks of use were analyzed with traditional chemical oxidation measurements and ﬂuorescence spectroscopy. Fluorescence landscapes of a fresh frying oil and frying oil that was discarded showed that the two-component system with emission maxima at 475 and 660 nm (chlorophyll from the plant oil) degenerated to a one-component system with an emission maximum at approximately 585 nm (Figure 15.9). Five emission spectra with excitation set to 395, 420, 440, 500, and 530 nm were measured and used for predicting chemically measured lipid oxidation parameters such as triglycerides, anisidine value, iodine value, and free fatty acids. These spectral settings were chosen based on the excitation and emission maxima of the ﬂuorescence landscapes (Figure 15.9). Squared correlation coefﬁcients between reference measurements and predicted values varied between 0.91 and 0.97 and the errors varied from 2% to 30%. The best

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 357 14.7.2008 11:34am Compositor Name: BMani

357

Autoﬂuorescence Spectroscopy in Food Analysis

300

400

Emission (nm)

Emission (nm)

300

500 600

(a)

500 600 700

700 800

400

250 300 350 400 450 500 Excitation (nm)

800 200 250 300 350 400 450 500 (b) Excitation (nm)

FIGURE 15.9 Fluorescence landscapes of frying oils: (a) fresh oil and (b) the oil that was discarded after 4 weeks of use.

prediction was obtained for triglyceride, but even for prediction of anisidine value, the ﬂuorescence measurements gave much better results than those obtained using vibrational spectroscopy. Another study measured lipid oxidation in chicken and turkey meat. The meat was stored at 258C for up to 2 years. Fluorescence emission spectra showed increased ﬂuorescence intensity between 400 and 600 nm and high correlations to chemically determined 2-thiobarbituric acid (TBARS) (Wold and Mielnik 2000). It is important to consider that in lipid oxidation studies of food, none of the oxidation parameters predicted by ﬂuorescence are actually ﬂuorescent and that the predictive ability is due to indirect relationships. The formation or disappearance of the lipid oxidation parameters are correlated to formation or disappearance of ﬂuorescent compounds or light-absorbing species inﬂuencing the ﬂuorescence measurements or to changes in physical phases, e.g., protein denaturation and viscosity. An advantage is that emission from several compounds related to the different stages of the oxidation process can be obtained in one measurement, making it possible to simultaneously follow a number of parameters in the product and relate these to the various reference measurements.

15.6.3 FLUORESCENCE

AND

FOOD AUTHENTICITY

Food authenticity is another issue that is becoming increasingly important because the food purchased must match the description. For instance, the content should be equal to what is stated in the declaration, and the geographical, vegetable, or animal origin should be as described. The high sensitivity of ﬂuorescence spectroscopy makes it possible to detect compounds in ppb levels. It has been reported that even concentrations down to ppt have been measured (Li et al. 2003). Thus, it has shown potential for detecting adulteration, fraud, etc. A few studies have been conducted on food items such as fruit juice (Seiden et al. 1996), vegetable oil (Guimet et al. 2004; Zandomenghi et al. 2005), honey (Rouff et al. 2005), and milk (Karoui et al. 2004a,b). Olive oil can be classiﬁed into several quality categories and thus it runs the risk of adulteration. By chemometric modeling of ﬂuorescence landscapes, Guimet et al. (2004) found signiﬁcant differences between two qualities of olive oil and ascribed these to changes in chlorophyll and oxidation products as determined by emission peaks at approximately 670 and 400–500 nm. These results were supported by Zandomenghi et al. (2005) who measured emission spectra of olive oil that varied in origin, cultivar, and age. They found not only chlorophyll ﬂuorescence to be the main contributor to the variations but also pheophytins with emission properties similar to chlorophyll, and compounds denoted as antioxidants gave well-deﬁned differences among the oil samples. The antioxidants had emission maximum between 300 and 350 nm when excitation took place at 280 nm. Fluorescence has also shown the ability to classify several types of edible oils including soybean, sunﬂower, rapeseed, peanut, olive, grape seed, linseed, and corn oils (Zandomenghi et al. 2005).

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 358 14.7.2008 11:34am Compositor Name: BMani

358

Handbook of Food Analysis Instruments

Autoﬂuorescence landscapes with excitation wavelengths of 250–450 nm and emission recorded up to 700 nm showed good discrimination between the oil classes.

15.6.4 FLUORESCENCE

AND

PROCESS ANALYTICAL TECHNOLOGY

Concentration (a.u.)

Process analytical technology (PAT) is a new concept introduced within the pharmaceutical industry, but the principle has been in focus in the food industry for a number of years. Basically, PAT refers to measuring the relevant information at the relevant time during the production. It can be the incoming raw material, some process streams or process parameters or the ﬁnished product. The goal of PAT is through massive internal quality control by remote spectroscopic monitoring to ensure ﬁnal product quality and to improve production efﬁciency. One example of a food PAT application is the monitoring of the beet sugar production by ﬂuorescence spectroscopy. The sugar, sucrose, is not itself ﬂuorescent; however, impurities in both crystalline sugar and sugar solutions make up a strong ﬂuorescence emission. This was recognized already in the 1940s where ﬂuorescence occurring upon illumination with ultraviolet light was applied for quality inspection and determination of impurities (Nørgaard 1995b). More dedicated experiments showed that ﬂuorescence of sugar stems from tyrosine, tryptophan, a polyphenolic compound as well as colorant polymers formed in Maillard reactions during the sugar processing (Baunsgaard et al. 2000). To analyze the sugar production, a number of emission spectra of sugar dissolved in water were measured every 8 h during the sugar campaign that lasted for about 3 months (Munck et al. 1998). Chemometric data analysis revealed the presence of four ﬂuorescent compounds in the samples, which all displayed a weekly periodicity with a tendency of increased concentrations during weekends in the beginning of the campaign (Figure 15.10). This increase was suggested to be due to longer storage time in the weekends, resulting in increased temperature and microbiological and enzymatic activity. The weekly variation leveled off during the sugar campaign when the outdoor temperature decreased. In the middle of November, the concentration of one of the compounds started to increase. This occurred at the same time that outdoor temperatures fell below 08C. The ﬂuorescent compounds were found to correlate with several process parameters such as lime salts in thin juice and pH in thick juice as well as the ash content and the color of the ﬁnished product. Another example shows ﬂuorescence measurements of sugar samples taken during the ﬁrst 3 days of the sugar campaign. A PCA showed a large variation in the scores in the beginning of the production. After a certain time, the production reached equilibrium state and the scores were found in a closed group (Figure 15.11). However, there was segregation in two clusters in the area of balanced samples (Munck et al. 1998) due to a change in the process conditions.

Time

FIGURE 15.10 Relative concentrations of the four ﬂuorophores in sugar versus the time. For visualization these are separated on the concentration scale.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 359 14.7.2008 11:34am Compositor Name: BMani

359

Autoﬂuorescence Spectroscopy in Food Analysis

50 40 30 20 10 0

200

⫺10 ⫺20 ⫺30 ⫺200 ⫺100 0

100

100 200 300 400 500 600

PC2

0

⫺100

⫺200

⫺300

0

500

1000

1500

PC1

FIGURE 15.11 Score plot of PC1 versus PC2 for a PCA made on ﬂuorescence spectra of sugar samples taken during the ﬁrst 3 days of the campaign. The area with balanced samples is enlarged on the right-hand side.

15.7 FUTURE TRENDS Autoﬂuorescence of intact food systems contains valuable information on various quality parameters and has a considerable potential both within research and as industrial online or at-line applications. The fact that each ﬂuorophore is characterized by both excitation and emission parameters enables the resolution and identiﬁcation of ﬂuorescent molecules as well as the differentiation between substitutions and conformations of the same molecule. Fluorescence is trace substance sensitive, but it is also very sensitive to artifacts such as quenching, scatter, and innerﬁlter effects. However, these artifacts are also parameters that make ﬂuorescence unique and by careful development of the measurement method and modeling of the data, these can provide an extremely sensitive probe for variations in the food item under investigation. Multiway analysis of ﬂuorescence landscapes enables identiﬁcation of the underlying structure, providing estimates of the excitation and emission spectra of the pure ﬂuorophores (Christensen et al. 2006). This opens up for new applications of complex chemical systems such as food samples. Multiway chemometrics has, for example, been applied to ﬂuorescence landscapes of ﬁsh oil (Pedersen et al. 2003). The PARAFAC decomposition revealed four ﬂuorophores present in the oil samples, of which one was assigned to chlorophyll. The obtained complex ﬂuorescence ﬁngerprints of the ﬁsh oils were shown to correlate (indirectly) to dioxin content in the ﬁsh oil, and the method was suggested as a screening method for dioxin contamination.

15.7.1 PROCESS ANALYTICAL TECHNOLOGY MONITORING

BY

FLUORESCENCE SPECTROSCOPY

In the PAT guidance to the pharmaceutical industry, it is stated that ‘‘For certain applications, sensor-based measurements can provide a useful process signature that may be related to the

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 360 14.7.2008 11:34am Compositor Name: BMani

360

Handbook of Food Analysis Instruments

underlying process steps or transformations.’’ In many cases, ﬂuorescence spectra of food can be considered a pin-code of the food items that contain a wealth of information related to the raw materials and the process transformations. For this reason, ﬂuorescence spectroscopy has a tremendous unexploited potential in PAT, as it is possible to measure compounds related to the process conditions fast and nondestructively. The multidimensionality, high sensitivity, and selectivity combined with multivariate or multiway data analysis provide valuable information about quality parameters relevant within the food industry. Wavelengths of visual illumination can be transmitted over long distances with almost no loss. This enhances the possibility for industrial online applications, since measurements can be taken at several points along the process line with only one spectrophotometer. The most important problems to be solved with regard to using ﬂuorescence in relation to PAT are to make applicable standardization techniques, since the ﬂuorescence signal is not measured relative to an incident light beam. The high sensitivity also induces a challenge. Small changes in raw materials or changes taking place during production may inﬂuence the ﬂuorescence intensity signiﬁcantly. Handling these problems requires thorough planned multivariate recalibration procedures. In the full implementation of ﬂuorescence sensors in the food industry, the probe must be placed directly on the process line, in the batch reactor, or diverted side streams may pass through the ﬂuorescence instrument. This requires some effort in developing new probes.

15.7.2 FLUORESCENCE IMAGING Last but not least, spatially resolved ﬂuorescence will have a great impact on food science and industry. Fluorescence imaging enables visualization of spatial distribution of chemical compounds, structural changes, and spatial progression of chemical processes. Such images can give unique insight into complex food systems, and a better fundamental and practical understanding of important properties. Figure 15.12 shows how the degree of photooxidation in a piece of Swisslike cheese can be imaged. The cheese was stored under light in commercial packaging. The autoﬂuorescence from a selected subsection of the cheese was imaged and illustrates how the colored part of the packaging ﬁlm protects the product against photooxidation. In the same way, it is possible to image the photodegradation of riboﬂavin, porphyrins, and vitamins as a result of storage conditions, packaging materials, etc. Another illustrative ﬂuorescence image (Figure 15.13) shows how oxidation (stable ﬂuorescent oxidation products) is distributed over a cross-section of a turkey burger stored in high oxygen atmosphere (Veberg et al. 2006). Images like this captured over time 100 90 80 70 60 50 40 30 20 10 (a)

(b)

FIGURE 15.12 (See color insert following page 240.) (a) Subsection of a commercially packed Swiss cheese and (b) ﬂuorescence image of the degree of photooxidation after storage under standard commercial illumination.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 361 14.7.2008 11:34am Compositor Name: BMani

361

2.5 cm

Autoﬂuorescence Spectroscopy in Food Analysis

FIGURE 15.13

Fluorescence image of cross-section of turkey burger. Bright areas indicate lipid oxidation.

reveal when and where the oxidation starts, and how fast and deep it progresses into the product. As in the previous example, this technique can be used to evaluate different storage regimes and packaging concepts. It has been used very effectively to detect pro-oxidative ﬂaws in packaging because the exact spatial origin of oxidation was detected. The images shown here were from rather broadband and dominating ﬂuorescent phenomena. Often the sample to be imaged is a complex mixture of ﬂuorophores resulting in overlapped spectra. In those cases, it is difﬁcult to visualize the desired chemistry in one single image. Then multispectral imaging is required. Multispectral imaging is a scientiﬁc ﬁeld in rapid development and is presently receiving great attention. Triggered by spectral imaging instrumentation, increased computer power, and sophisticated mathematical=statistical algorithms, effective collection and analysis of such images have been established. Few commercial spectral imaging systems are tailor-made for autoﬂuorescence, but those working in the visible region can be modiﬁed for this purpose. Multispectral images allow speciﬁc chemical imaging, even when the spectral proﬁles are overlapping. With a spectrum in each pixel, established multivariate techniques can be applied at pixel spectrum level, including spectral curve resolution. Timlin et al. (2005) have demonstrated this powerful approach for robust analysis of microarrays. They used a hyperspectral microarray ﬂuorescence scanner and were able to identify and quantify emission spectra and concentration from all emission sources. Pure microarray images of the two desired analytes as well as three contaminants could then be produced and analyzed separately. This approach can also be used on complex food systems, and will probably be established in near future. Three-way ﬂuorescence data (excitation emission matrices) are, as mentioned above, an excellent basis for curve resolution, and technically it is fully possible to collect ﬂuorescence landscapes in each pixel. The opportunities for advanced ﬂuorescence imaging are driven by technology and needs appearing within certain ﬁelds of analysis. These techniques will certainly ﬁnd areas of application within food science and process control.

REFERENCES Allais, I. et al., A rapid method based on front-face ﬂuorescence spectroscopy for the monitoring of the texture of meat emulsions and frankfurters, Meat Sci., 67, 219, 2004. Baunsgaard, D. et al., Multi-way chemometrics for mathematical separation of ﬂuorescent colorants and colour precursors from spectroﬂuorimetry of beet sugar and beet sugar thick juice as validated by HPLC analysis, Food Chem., 70, 113, 2000. Becker, E.M. et al., Front-face ﬂuorescence spectroscopy and chemometrics in analysis of yogurt: Rapid analysis of riboﬂavin, J. Dairy Sci., 86, 2508, 2003. Booksh, K.S. and Kowalski, B.R., Theory of analytical chemistry, Anal. Chem., 66, 782, 1994. Bron, I.U. et al., Chlorophyll ﬂuorescence as a tool to evaluate the ripening of ‘‘Golden’’ papaya fruit, Postharvest Biol. Technol., 33, 163, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 362 14.7.2008 11:34am Compositor Name: BMani

362

Handbook of Food Analysis Instruments

Christensen, J., Becker, E.M., and Frederiksen, C.S., Fluorescence spectroscopy and PARAFAC analysis in yogurt, Chemom. Intell. Lab. Syst., 75, 201, 2005. Christensen, J. et al., Multivariate autoﬂuorescence of intact food systems, Chem. Rev. 106, 1979, 2006. Costa, L.F., Mielenz, K.D., and Grum, F., Correction of emission spectra, in Optical Radiation Measurements, Measurements of Photoluminiscence, volume 3, Mielenz, K.D., Ed., Academic Press, New York, 1982, Chap. 4. Engelsen, S.B., Explorative spectrometric evaluations of frying oil deterioration, JAOCS, 74, 1495, 1997. Fox, J.B. and Thayer, D.W., Radical oxidation of riboﬂavin, Int. J. Vit. Res., 68, 174, 1998. Frencia, J.P., Thomas, E., and Dufour, E., Measurement of meat tenderness using front-face ﬂuorescence spectroscopy, Sci. Aliments, 23, 142, 2003. Guimet, F. et al., Application of unfold principal component analysis and parallel factor analysis to the exploratory analysis of olive oils by means of excitation-emission matrix ﬂuorescence spectroscopy, Anal. Chim. Acta, 515, 74, 2004. Herbert, S. et al., Monitoring the identity and the structure of soft cheese by ﬂuorescence spectroscopy, Lait, 80, 621, 2000. Jensen, S.A., Reenberg, S., and Munck, L., Method for quality control of products from ﬁsh, cattle, swine and poultry, U.S. Patent US4631413, 1986. Jiji, R.D. and Booksh, K.S., Mitigation of Rayleigh and Raman spectral interferences in multi-way calibration of excitation-emission matrix ﬂuorescence spectra, Anal. Chem., 72, 718, 2000. Karoui, R., Mazerolles, G., and Dufour, E., Spectroscopic techniques coupled with chemometric tools for structure and texture determinations in dairy products, Int. Dairy J., 13, 607, 2003. Karoui, R. et al., Determining the geographical origin of Emmental cheeses produced during winter and summer using a technique based on the concatenation of MIR and ﬂuorescence spectroscopic data, Eur. Food Res. Technol., 219, 184, 2004a. Karoui, R. et al., Fluorescence and infrared spectroscopies: A tool for the determination of the geographical origin of Emmental cheeses manufactured during summer, Lait, 84, 359, 2004b. Lackowicz, J.R., Principles of Fluorescence Spectroscopy, 2nd ed., Kluwer Academic=Plenum Publishers, New York, 1999. Li, J.-S. et al., Spectroﬂuorimetric determination of total amount of nitrite and nitrate in biological sample with a new ﬂuorescent probe 1,3,5,7-tetramethyl-8-(30 40 -diaminophenyl)-diﬂuoroboradiaza-s-indacene, Talanta, 61, 797, 2003. Martens, H. and Næs, T., Multivariate Calibration, John Wiley & Sons, England, 1989. Melhuish, W.H., Correction of excitation spectra, in Optical Radiation Measurements, Measurements of Photoluminiscence, volume 3, Mielenz, K.D., Ed., Academic Press, New York, 1982, Chap. 3. Munck, L. et al., Chemometrics in food science—a demonstration of the feasibility of a highly exploratory, inductive evaluation strategy of fundamental scientiﬁc signiﬁcance, Chemom. Intell. Lab. Syst., 44, 31, 1998. Novales, B. et al., Multispectral ﬂuorescence imaging for the identiﬁcation of food products, J. Sci. Food Agric., 71, 376, 1996. Nørgaard, L., A multivariate chemometric approach to ﬂuorescence spectroscopy, Talanta, 42, 1305, 1995a. Nørgaard, L., Classiﬁcation and prediction of quality and process parameters of thick juice and beet sugar by ﬂuorescence spectroscopy and chemometrics, Zuckerindustrie, 120, 970, 1995b. Pedersen, D.K., Munck, L., and Engelsen, S.B., Screening for dioxin contamination in ﬁsh oil by PARAFAC and N-PLSR analysis of ﬂuorescence landscapes, J. Chemom., 16, 451, 2003. Rouff, K. et al., Authentication of the botanical origin of honey by front-face ﬂuorescence spectroscopy. A preliminary study, J. Agric. Food Chem., 53, 1343, 2005. Seiden, P. et al., Exploring ﬂuorescence spectra of apple juice and their connection to quality parameters by chemometrics, J. Agric. Food Chem., 44, 3202, 1996. Sikorska, E. et al., Characterization of edible oils using total luminiscence spectroscopy, J. Fluoresc., 14, 25, 2004. Song, J. et al., Changes in chlorophyll ﬂuorescence of apple fruit during maturation, ripening, and senescence, HortScience, 32, 891, 1997. Steiner, W.E., Rieker, R.H., and Battaglia, R., Aﬂatoxin contamination on dried ﬁgs: Distribution and association with ﬂuorescence, J. Agric. Food Chem., 36, 88, 1988.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 363 14.7.2008 11:34am Compositor Name: BMani

Autoﬂuorescence Spectroscopy in Food Analysis

363

Timlin, J.A. et al., Hyperspectral microarray scanning: Impact on the accuracy and reliability of gene expression data, BMC Genomics, 6, 72, 2005. Veberg, A. et al., Mapping of lipid oxidation and porphyrins in high oxygen modiﬁed atmosphere and vacuum packed minced turkey and pork meat by ﬂuorescence spectra and images, Meat Sci., 73, 511, 2006. Wold, J.P. and Mielnik, M., Non-destructive assessment of lipid oxidation in minced poultry meat by autoﬂuorescence spectroscopy, J. Food Sci., 65, 87, 2000. Wold, J.P. et al., Inﬂuence of storage time and color of light on photooxidation in cheese: A study based on sensory analysis and ﬂuorescence spectroscopy, Int. Dairy J., 16, 1218, 2006a. Wold, J.P. et al., Active photosensitizers in butter detected by ﬂuorescence spectroscopy and multivariate curve resolution, J. Agric. Food Chem., 54, 10197, 2006b. Zandomenghi, M., Carbonaro, L., and Capparata, C., Fluorescence of vegetable oils: Olive oils, J. Agric. Food Chem., 53, 759, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C015 Final Proof page 364 14.7.2008 11:34am Compositor Name: BMani

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 365 16.7.2008 11:04pm Compositor Name: MSubramanian

Nose Technology 16 Electronic in Food Analysis Figen Korel and Murat Ö. Balaban CONTENTS 16.1 16.2 16.3

Introduction ........................................................................................................................ 365 Electronic Nose Technology ............................................................................................. 366 Sensors Types Used in Electronic Nose ........................................................................... 367 16.3.1 Conducting Polymer Sensors .............................................................................. 367 16.3.2 Metal Oxide Semiconductor Sensors .................................................................. 367 16.3.3 Metal Oxide Semiconducting Field Effect Transistors ....................................... 367 16.3.4 Surface Acoustic Wave Sensors .......................................................................... 368 16.3.5 Bulk Acoustic Wave Devices .............................................................................. 368 16.3.6 Electrochemical Sensors ...................................................................................... 368 16.3.7 Smell-Seeing Sensors .......................................................................................... 368 16.3.8 Gas Chromatography and Mass Spectrometry-Based Sensors ........................... 368 16.4 Commercially Available Electronic Noses ....................................................................... 369 16.5 Data Analysis Methods ..................................................................................................... 369 16.6 Food Analysis Using Electronic Nose .............................................................................. 370 16.6.1 Milk and Dairy Products ..................................................................................... 370 16.6.2 Meat, Poultry, and Seafood Products .................................................................. 371 16.6.3 Fruits and Vegetables .......................................................................................... 372 16.6.4 Alcoholic Drinks ................................................................................................. 372 16.6.5 Olive Oils and Other Edible Oils ........................................................................ 373 16.6.6 Other Food Applications ..................................................................................... 373 16.7 Conclusion and Future Trends .......................................................................................... 374 References ..................................................................................................................................... 374

16.1 INTRODUCTION The composition and properties of raw materials and processed foods are important for food quality and safety. Food quality depends on color, aroma, ﬂavor, texture, nutrition, and microbial content [1]. Aroma and appearance are the most important parameters the consumer perceives initially in food. Aroma is analyzed traditionally using sensory or gas chromatographic methods. However, these are time consuming, expensive, and require sample preparation. There is growing interest in new rapid methods. Electronic noses (e-noses) offer several advantages such as high sensitivity, quickness, simplicity, lower cost, nondestructive operation, and little or no sample preparation compared to traditional methods. The e-nose concept as a chemical array sensor system for odor classiﬁcation was presented for the ﬁrst time by Persaud and Dodd [2] and e-noses started to appear in the market in the early 1990s. During the last two decades, the uses, capabilities, and applications of e-noses have increased in food science. Unlike traditional analytical methods, e-nose technology does not provide information on the nature of the analyzed product; it only gives a digital ﬁngerprint 365

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 366 16.7.2008 11:04pm Compositor Name: MSubramanian

366

Handbook of Food Analysis Instruments

which can be analyzed chemometrically [3]. E-noses have been used mostly in the assessment of food properties, detection of adulteration, prediction of sensory properties, and classiﬁcation of different food matrices. Whether the e-nose could be used in measurement of quality for speciﬁc food product, the type of e-nose to be selected, and the suitable data analysis method are general considerations that need to be evaluated before applying them in the food area [4]. This chapter reviews the technology, available sensors used in the instrument, commercial e-noses, and data analysis techniques regarding responses taken from the e-nose. Practical applications in foods are discussed in the following sections. Future trends in e-nose technology are described.

16.2 ELECTRONIC NOSE TECHNOLOGY Development of sensing techniques=methods to ensure food quality and safety is a priority to beneﬁt the consumer. The advancements and growth in electronics and sensor technologies show promises for the development of rapid and nondestructive sensors for determining food quality and safety. Such sensors can be used to alert consumers of potential health risks due to the consumption of unsafe food products [5]. A mixture of volatiles (odor) is recognized as a whole instead of its single components in the human olfactory system [6]. An e-nose is called an artiﬁcial nose since it simulates the human olfactory system. It is ‘‘an instrument which comprises an array of electronic chemical sensors with partial speciﬁcity and an appropriate pattern recognition system capable of recognizing simple or complex odors’’ [7]. The e-nose is comprised of: (1) a sampling system, (2) an array of gas sensors with different selectivities, (3) signal processing and conditioning (data acquisition system), and (4) an appropriate pattern recognition algorithm to recognize simple or complex odors [7,8]. The sampling system provides a stable and reproducible headspace gas sampling environment by eliminating any undesirable factors inﬂuencing sensor responses [9]. In most of the commercial e-noses there are two separate chambers for the sampling system: a sample chamber and a sensor chamber. Since the responses are inﬂuenced by the changes in temperature and humidity, these two parameters are controlled in both chambers. The volatiles in the sample headspace are conveyed to sensor chamber through gas ﬂow. Inert gas is applied to both chambers after readings to clean any possible odor leftover from the previous sample or possible contamination [4]. Most of the hand held instruments do not have a sample chamber. This may cause a problem in obtaining repeatable measurements since the possible artifacts coming from the environment could affect the responses. It will be beneﬁcial to design a sample chamber for these instruments. Nonabsorbent and inert materials need to be selected to build sample chambers to avoid any leftover odors from previous samples [9]. The size of the sample chamber is also critical and should be proportional to the sample size. If a small sample is measured, the volume of the sample chamber can be small. However, if the sample size is large such as a tomato, an orange, or any seafood, the sample chamber size should be large enough to hold the sample. The type and number of sensors are important. Some samples need to be heated in order to have their aromas volatilizing into the headspace. However, in other samples there may be very volatile components at room temperature and they do not require heating. An important design parameter of the data acquisition system of an e-nose is how it acquires time-dependent sensor signals. The sensors’ responses can be steady state (static), or transient (dynamic). It would be ideal if the data acquisition system of the instrument records the responses from the sensors over time and the user could access these data. More information is contained in the time-series data compared to the ‘‘static’’ data [4]. The e-nose differs from human noses in its type and number of sensors and signal processing methods [10]. The instrument needs to be trained to discriminate odors, achieved by correlating e-nose responses with sensory analysis, chromatography or wet chemistry, or by calibrating with known samples [4].

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 367 16.7.2008 11:04pm Compositor Name: MSubramanian

Electronic Nose Technology in Food Analysis

367

16.3 SENSORS TYPES USED IN ELECTRONIC NOSE The e-nose can have an array of nonspeciﬁc sensors. Each sensor gives a different response to each of the components of the headspace mixture. The magnitude of the responses of each sensor depends on the sensor material. A gas mixture is identiﬁed by the pattern generated by the array of sensors. The gas mixture is identiﬁed by an appropriate pattern recognition algorithm. Different sensing elements can be used in an e-nose [5]. The selection of sensors is of great importance. A unique ‘‘ﬁngerprint’’ of the product could be obtained using different sensors with overlapping responses to the range of compounds within the samples [11]. The change of response pattern of sensors with time could be the major problem, and sensor drift must be minimized by periodically measuring standard solutions or a calibration substance [11]. Conducting polymer (CP) sensors, metal oxide semiconductor (MOS) sensors, metal oxide semiconducting ﬁeld effect transistors (MOSFETs), surface acoustic wave (SAW) gas sensors, quartz crystal microbalance (QCM) devices, electrochemical sensors, smell-seeing sensors, and gas chromatography (GC) and mass spectrometry (MS) based sensors are discussed below.

16.3.1 CONDUCTING POLYMER SENSORS Early approaches used sensor ﬁlms consisting of conducting organic polymer phases such as polypyrrole, polyacetylene, or polythiophene and their derivatives. Current sensor ﬁlms consist of an electrical conductor dispersed into the swellable organic insulator. The volatiles partition into the polymer ﬁlm and cause swelling. This increases the electrical resistance of the ﬁlm since swelling decreases the number of connected pathways of the conducting component of the composite material [5,12–14]. Detector ﬁlms can be made from conducting polymer composites. The conductive and insulating materials are both organic polymers. In polymer-conductor composites the conductive phase is an inorganic conductor such as carbon black, Au, Ag, etc., and the insulating phase is a swellable organic material [15]. The gas-polymer partition coefﬁcient determines the sensitivity of an individual array element to a particular odorant. These sensors can be used in a wide range of operating conditions [16] and have good sensitivity (between 0.1 and 100 ppm), stability, short response time, and short recovery time [12,17–19], but they are sensitive to moisture [16]. CP sensors are considered to be ‘‘cold sensors’’ since they operate at lower temperatures (such as room temperatures) than ‘‘hot sensors’’ like metal oxide sensors [20].

16.3.2 METAL OXIDE SEMICONDUCTOR SENSORS These sensors have high chemical stability, low aging, high sensitivity, easy manufacturing, and low cost [21,22]. MOS sensors are based on tin dioxide (SnO2) thin ﬁlms doped with metals (such as Cr and In) to alter the response characteristics of the semiconductor. ‘‘Sensor features are based on charge transfer reactions occurring during catalytic reactions of molecules at the surface at high temperature, which cause the change of electrical resistance in the sensor’’ [8]. MOS sensors operate between 3008C and 5508C to avoid interference from water and to aid rapid response and recovery times [12]. They are sensitive to combustible materials, such as alcohols, but less sensitive at detecting nitrogen- and sulfur-based odors.

16.3.3 METAL OXIDE SEMICONDUCTING FIELD EFFECT TRANSISTORS MOSFETs are similar to MOS sensors, but the output signal derives from the change in potential when the volatile compounds react at the catalytic surface. They have lower operating temperatures (1008C–2008C) compared to MOS sensors [23].

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 368 16.7.2008 11:04pm Compositor Name: MSubramanian

368

Handbook of Food Analysis Instruments

16.3.4 SURFACE ACOUSTIC WAVE SENSORS SAW sensors are interdigitated electrode arrays on a piezoelectric crystal, usually quartz, and a thin surface coating of an absorbing material made of polymers, lipids, Langmuir-Blodgett ﬁlms, and self-assembled monolayers. A synchronous mechanical surface wave is created by the radiofrequency excitation of the electrode pair and it is propagated on the surface of the piezoelectric substrate. This is recorded either by another electrode pair on a SAW delay line or by the same pair after reﬂection on a SAW resonator device [24]. The changes in vibrational resonant frequency of piezoelectric quartz oscillators that result from changes in mass on the oscillator’s surface are measured in these devices [23]. A surface wave with a frequency between 100 MHz and 1 Ghz is generated. System sensitivity is controlled by frequency [25,26]. This frequency change is the response of the sensor to the volatiles. SAW sensors have high sensitivities and fast response times. They have good reproducibility; however, they are sensitive to humidity and temperature [12,23]. They are also more selective, which means a larger number of SAW sensors are required to cover all volatiles that may exist in a food.

16.3.5 BULK ACOUSTIC WAVE DEVICES Bulk acoustic wave device is also called as QCM or thickness shear mode (TSM) device. It is the simplest type of piezoelectric sensor. It contains single quartz crystal (around 1 cm in diameter) with electrodes (usually gold) evaporated onto the two large faces. The resonant frequency of the device is in between 5 and 20 MHz [12]. It measures the mass of molecules absorbed on the sensor surface. It operates at room temperature and has high stability over time [23]. The limitations of this device are similar to those of SAW sensors, namely sensitivity to humidity and temperature [12].

16.3.6 ELECTROCHEMICAL SENSORS Electrochemical sensors contain electrodes and an electrolyte. The volatile will be oxidized or reduced at the working electrode. The output signal is the measured voltage between the electrodes generated by the reactions. They have long-term stability and linear dependence on gas concentrations. They are not sensitive to humidity [23].

16.3.7 SMELL-SEEING SENSORS Smell-seeing sensor detects odors using the color change that occurs in gas-sensitive metalloporphyrins or other chemo-responsive dyes immobilized on reverse phase silica gel. Some odor sensors are not able to detect some of the toxic vapors; however, these compounds could easily bind to metalloporphyrins and color change occurs. These sensors are not sensitive to humidity and provide visual odor identiﬁcation [27,28].

16.3.8 GAS CHROMATOGRAPHY

AND

MASS SPECTROMETRY-BASED SENSORS

Recently, e-noses based on GC or MS have been developed as an alternative to conventional sensorbased instruments. These are sometimes referred to as new-generation e-noses. For the GC-based sensor system, volatile compounds are separated by a fast GC column (usually 1–3 m) and detected by either SAW or ﬂame ionization detectors. The responses of these detectors at different times are recognized as ‘‘sensor arrays.’’ The limitation for the GC=SAW-based sensor system is that the single sensor may not respond to some of the important volatile compounds. For the MS-based system, volatile compounds are introduced into MS without prior separation, and ion fragments are considered as sensor arrays. The proﬁle signal is extracted and analyzed. This system has a very fast response, but contains less information. It is cheaper and easier than regular GC=MS analysis [29].

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 369 16.7.2008 11:04pm Compositor Name: MSubramanian

369

Electronic Nose Technology in Food Analysis

16.4 COMMERCIALLY AVAILABLE ELECTRONIC NOSES Like most new technologies, the e-nose ﬁeld experienced an initial ﬂurry of activity in the commercial e-noses. The market then consolidated, as the technology started to ‘‘mature.’’ Several e-noses are available in the market. Table 16.1 lists the major companies selling e-noses for food quality applications.

16.5 DATA ANALYSIS METHODS The signals obtained from the e-nose sensors need to be evaluated using appropriate pattern recognition techniques. There are two basic approaches: multivariate data analysis and artiﬁcial neural networks. Currently, the data analysis methods for e-nose responses are only used to classify the unknown sample into predetermined classes, or to determine patterns [4]. The most commonly used pattern recognition techniques include principal component analysis (PCA), linear discriminant analysis (LDA), partial least squares (PLS), cluster analysis (CA), canonical correlation analysis (CCA), and fuzzy logic or artiﬁcial neural network analysis (ANN) [20,23].

TABLE 16.1 Commercially Available E-Noses for Food Quality Applications Commercial E-Noses Pen, i-Pen Fox and Gemini Kronos Promethus Heracles BH 114 ST 214

Company Name Airsense Analytics, Germany Alpha MOS, France

Chemsensing

Bloodhound Sensors Limited Scensive Technologies Limited, United Kingdom Chemsensing Inc., USA

zNose

Estcal, USA

MOSES II

Lennartz Electronic GmbH, Germany Bodvaki-Maritech, Kópavogur, Iceland Smart Nose, Switzerland Smiths Detection, USA

FreshSense SMart Nose Cyranose320 EnQbe LibraNose

Tor Vergata (University of Rome and CNR), Italy

Sensor Types=Numbers MOS=10 sensors MOS=6–24 sensors MS MOS and MS=18 sensors Ultrafast GC=FID Conducting polymer=14 sensors

Smell-sensing sensor=36 sensors Ultrafast GC=SAW detector QCM=8 sensors SnO2=8 sensors Electrochemical=4 gas sensors MS Conducting polymer=32 sensors TSM=8 sensors QCM=8 sensors

Web Site http:==www.airsense.com=english= index_e.html http:==www.alpha-mos.com=

http:==www.scensive.com

http:==www.chemsensing.com=index. html http:==www.estcal.com http:==www.lennartz-electronic.de= Pages=MOSES=MOSES_home_e.html — http:==smartnose.com= http:==www.smithsdetection.com http:==www.cnr.it=istituti= Focusperistituto_eng.html?cds ¼ 057

Source: Adapted from Guide to Inorganic Analysis, Perkin-Elmer, Inc., 2004. Note: FID, ﬂame ionization detector; GC, gas chromatography; MOS, metal oxide semiconductor; MS, mass spectrometry; QCM, quartz crystal microbalance; SAW, surface acoustic wave; SnO2, tin dioxide; TSM, thickness shear mode.

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 370 16.7.2008 11:04pm Compositor Name: MSubramanian

370

Handbook of Food Analysis Instruments

The most widely used multivariate data analysis techniques are the PCA and LDA. PCA, a projection method, allows an easy visualization of all the information contained in a data set. It also helps to ﬁnd in what respect one sample is different from another and which variables contribute most to this difference. It is used to achieve a reduction of dimension and to observe a primary evaluation of the between-class similarities. This technique is used to observe similarities among different samples by reducing the dimension from many variables to two or three principal components while keeping most of the original information in the data set [30]. LDA, a traditional statistical technique, is used for dimensionality reduction. It is a classiﬁcation procedure in which the classes are considered to have normal distribution and equal dispersion (covariance matrix). The goal is to separate the classes by projecting the samples from p-dimensional space onto a ﬁnely orientated line. LDA maximizes the variance between categories and minimizes the variance within categories to optimize the resolution between classes [30]. The artiﬁcial neural network (ANN), a dynamic and self-adapting system, closely resembles the human learning process. It can be deﬁned as a set of simple calculation units (nodes) that start out from a data set and transform it into a set of response values. It has been widely used to handle data from an e-nose [30,31]. The network builds a model based on a set of input data (the training set) with known outputs by adjusting the weights associated with each connection, and output values as close as possible to the real values are generated. ANN architectures include multilayer perceptron (MLP) trained by back-propagation (BP) [32], MLP with genetic algorithms (GA) [33], adaptive logic network [34], radial basis function network [35], Fuzzy ARTMAP [35], self-organizing map network [36], and time-delay neural network [37]. Zhang et al. [38,39] found a classiﬁcation accuracy of 59% using discriminant function analysis (DFA) for discriminating four species. However, the accuracy was improved to 100% by using ANN-based time delay neural network for the classiﬁcation of the same species. Dutta et al. [40] reported a classiﬁcation accuracy of 98% while discriminating six bacterial species responsible for eye infections by using radial basis function neural network. Some e-noses provide their pattern recognition software package [41]. However, if the instrument does not provide software, commercially available software packages such as Matlab, SAS, Statistica, S-Plus, SPSS, Umetrics, Unscrambler, or other neural network packages could be used to analyze the data. An important requirement is the transportability of the data between e-noses. There are efforts to develop methods to assure the comparability of readings of the same sample with different e-noses [42].

16.6 FOOD ANALYSIS USING ELECTRONIC NOSE Substantial research has been conducted to determine the changes in the microbiological and chemical quality of foods by using an e-nose. The metabolic activities of microorganisms in foods produce metabolites in gas, liquid, or solid forms. Sensing the gaseous metabolites (volatile organic compounds) present in the headspace of the food product could determine the quality of that product [5]. Oxidation of lipids causes undesirable chemical changes that may generate products related to rancidity. Rancidity is based on the subjective organoleptic assessment (appraisal) of offﬂavors from foods. It causes ﬂavor, aroma, and taste deterioration leading to consumers’ rejection of the product. E-noses have also been used to detect adulteration. It is not possible to summarize in this chapter all of the studies using an e-nose for food quality evaluation. Selected food applications will be discussed brieﬂy in the following sections.

16.6.1 MILK

AND

DAIRY PRODUCTS

The e-nose has a widespread application in determining the quality and shelf life of milk and dairy products. Labreche et al. [43] determined the shelf life of milk stored at ambient and at refrigeration

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 371 16.7.2008 11:04pm Compositor Name: MSubramanian

Electronic Nose Technology in Food Analysis

371

temperature of 58C for 52 days by using an e-nose consisting of metal oxide sensors. Korel et al. [44] correlated the microbial, sensory, and e-nose data of pasteurized whole milk. Korel and Balaban [45] investigated the odor of whole, reduced fat, and fat free milk samples inoculated with Pseudomonas ﬂuorescens and Bacillus coagulans stored at 1.78C, 7.28C, and 12.88C up to 10 days, and the e-nose results were correlated with microbial counts and sensory panel results. Haugen et al. [46] monitored the growth of three spoilage bacteria, Serratia marcescens, Serratia proteamaculans, and Pseudomonas putida, in milk using a gas sensor array. The results were compared with microbial counts and headspace GC=MS of volatile microbial metabolites. The researchers concluded that gas sensor array system could be used to monitor the growth of individual strains of spoilage bacteria in a mixed culture in milk based on the type and amount of volatiles produced. E-nose was used for the determination of changes of blue cheese’s ﬂavor during maturation (up to 20 weeks after brining). Volatiles were also investigated by GC=MS. The classiﬁcation of predicted volatiles of unknown samples by their ripening stage was successful. They suggested that the application of the rapid and less demanding e-nose was an alternative to GC=MS for this type of investigation [47]. An e-nose with coated quartz microbalance was utilized to monitor the ripening process of Emmental cheeses. The development of the 2-heptanone concentration over time was followed by the instrument to distinguish different stages of the ripening process [48]. The aroma differences among ﬁve samples of strawberry ice cream with different fat contents (0%, 9%, or 18% dairy or vegetable fat) were detected using an e-nose. It was found that the capability of the instrument for detecting the aroma differences was comparable with sensory evaluation and GC analysis [49].

16.6.2 MEAT, POULTRY, AND SEAFOOD PRODUCTS Panigrahi et al. [5] investigated the freshness of beef strip loins stored at 48C and 108C using an e-nose with CP sensors. Classiﬁcation accuracies of 100% were obtained using radial basis function neural networks. García et al. [8] identiﬁed four different hams using an e-nose with SnO2 thin ﬁlms doped with Cr and In metal oxide sensors. The success rate of the classiﬁcation using PCA and probabilistic neuronal network (PNN) was 100%. E-nose was used for separation of ground raw and cooked samples of pork–beef mixtures by composition and freshness [50] and to determine the change of the volatile fraction of cooked chicken meat during storage at 48C [51]. The instrument was also used to determine the quality of modiﬁed atmosphere packaged broiler chicken cuts stored at different temperatures. The results showed that the e-nose was capable of detecting even early spoilage stages [52]. Boothe and Arnold [53] reported that an e-nose could detect changes in chicken meat samples based on storage time and temperature. Changes in volatiles during fermentation of sausages were monitored by e-nose and compared with sensory analysis. Fermentation time could be predicted from the sensor readings, and sensory results were compared with the e-nose sensor readings in the early and ﬁnal stages of the fermentation [54]. Several studies have been reported on the use of e-nose to assess quality of seafood. Luzuriaga and Balaban [55,56] evaluated the e-nose readings of the odor of decomposition of raw and cooked shrimp, and refrigerated salmon ﬁllets. E-nose determined the spoilage level of different seafood products by detecting the main classes of volatile degradation compounds (alcohols, aldehydes, esters, sulfur compounds, and amines) produced during chilled storage [57–59]. Luzuriaga et al. [60] correlated the results of sensory panel evaluations of shrimp with different chemicals (sulﬁtes, phosphates, bleach) with data from an e-nose with polymer sensors. E-nose was used to build a ﬁsh freshness indicator of sardines and it was reported that it was seen as a promising approach to obtain a comparable judgment with trained panels [61]. Samples of cold smoked salmon from different smokehouses in Europe were classiﬁed based on their quality changes using an e-nose. The samples were stored in different packaging (vacuum and MAP) at 58C and 108C for up to four weeks.

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 372 16.7.2008 11:04pm Compositor Name: MSubramanian

372

Handbook of Food Analysis Instruments

The system was found to be ideal for fast quality control related to freshness evaluation of smoked salmon products [62]. E-nose and headspace GC=MS were used to determine the volatile compounds in canned Alaska pink salmon with various grades of watermarking stored for 2 and 9 months. E-nose data analyzed using forward stepwise general discriminant analysis (FSGDA) resulted with 90% and 92.5% correct classiﬁcations for the samples stored at 2 and 9 months, respectively. However, results showed that the watermarking grades studied were not readily distinguishable from each other with the e-nose [63]. Quality changes of aerobically packed cod ﬁllets stored under superchilling and abusive temperature conditions were monitored by the growth of speciﬁc spoilage bacteria, and the production of microbial metabolites was measured by an e-nose along with traditional sensory and chemical analyses [64]. Korel et al. [65,66] combined the data from e-nose and machine vision to evaluate the quality of catﬁsh and tilapia, and reported substantial improvement of discrimination power by the combination of the methods. Newman et al. [67] correlated the odor of tuna in refrigerated storage with microbiological and sensory evaluation data.

16.6.3 FRUITS

AND

VEGETABLES

The quality of fruits and vegetables is mainly related to consumer perception and preference. There is a good correlation between human senses and application of optical, chemical, and tactile sensors. Research is focused on the development of nondestructive techniques for measuring quality attributes of fruits and vegetables. Aroma sensing is a promising method for determining their quality [68]. Many applications of e-nose technology to evaluate the quality of different fruits have been reported for oranges [69], pears [70,71], peaches [72,73], apples [72,74,75], and nectarines [73]. Maul et al. [76] nondestructively predicted whether a green tomato would turn red or not, by using an e-nose with polymer sensors. Moretti et al. [77] successfully screened tomatoes for internal bruising by evaluating their headspace odor with an e-nose. Gómez et al. [68,78] investigated the change in volatile production of ripeness states (unripe, half-ripe, full-ripe, and over-ripe) of tomato and mandarin, using an e-nose with 10 different MOS sensors. The e-nose could differentiate among the ripeness states of tomato with 100% correct classiﬁcation, and among the ripeness states of mandarin with 92% correct classiﬁcation. Shelf life and cultivar effect on tomato aroma proﬁle were investigated using a QCM-based and an MS-based e-nose. A clear distinction between cultivars was obtained by the e-nose based on MS [79]. Farnworth et al. [80] used an e-nose to distinguish between the two orange juices (original and volatile stripped). Aroma compounds from eight varieties of apricot (Prunus armeniaca) were discriminated by an e-nose using PCA [81].

16.6.4 ALCOHOLIC DRINKS An e-nose with SAW sensor was used to correctly identify different wines produced from the same variety of grapes and coming from the same cellar [82]. Similar results were obtained using an e-nose having MOS sensors to classify four types of red wines [83]. An MS-based e-nose was explored to measure sensory attributes in commercial Riesling wines grown in Australia [84]. An e-nose was also successfully used to identify the typical aromatic compounds of white and red wine aromas [85]. An e-nose based on SnO2 multisensor array was utilized for classiﬁcation of 29 typical aromas in white wine. The results indicated that despite the strong inﬂuence of ethanol and other major compounds of wine, the system could still discriminate the aromatic compounds added to the wine with an accuracy of at least 97.2% [86]. Five molecules responsible for off-ﬂavors in wines have been detected with an e-nose having MOS sensors, and principal component analysis showed clear discrimination between the control wine and off-ﬂavor doped wines, even at concentrations below human expert perception threshold [87]. E-nose was also used to successfully characterize eight fruit wines (blueberry, cherry, raspberry, black currant, elderberry, cranberry, apple, and peach) and four grape wines (red, Chardonnay, Riesling, ice wines) which were obtained from a minimum of ﬁve Ontario wineries [88].

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 373 16.7.2008 11:04pm Compositor Name: MSubramanian

Electronic Nose Technology in Food Analysis

16.6.5 OLIVE OILS

AND

373

OTHER EDIBLE OILS

The classiﬁcation of virgin olive oils into three categories (extra virgin, virgin, and lampante) based on current European regulations is performed by sensory analysis which is a lengthy and expensive method and cannot be applied online [89]. High resolution gas chromatography (HRGC) can be used as an alternative to sensory assessment to determine the volatile compounds responsible for virgin olive oil aroma. However, this method is also time consuming and requires trained personnel [90]. Sensor systems could be another alternative which are able to analyze the aroma proﬁle of virgin olive oils faster and simpler, and without any reagents [91]. E-nose and electronic tongue, in combination with chemometrics (multivariate data analysis), have been utilized to verify the geographical origin and the uniqueness of speciﬁc extra virgin olive oils. Neural networks provided satisfactory results, and that indicated the e-nose was an appropriate instrument for the characterization of the oils [92]. E-nose based on CPs was used for discrimination of olive oil quality, variety of olive, and geographical origin [93]. An e-nose based on a solid phase microextraction (SPME) unit coupled with a SAW sensor array has been used to distinguish the aroma of virgin olive oil. The results indicated that nonlampante and lampante virgin olive oils were classiﬁed successfully [94]. Ten different extra virgin olive oil samples were classiﬁed with 78% accuracy, and misclassiﬁcation occurred mostly between similar olive oils. Defective oil samples were separated from defect-free olive oils with 97% accuracy [95]. The vinegary defect in virgin olive oils was detected by an e-nose with MOS sensors [96]. A SAW based e-nose was used to characterize 16 types of vegetable oils, and the score plot of the PCA showed that 97% of the total variance in the data were described by principal component 1 and principal component 2 [97]. Two edible oil blends (groundnut–coconut and rice bran–palmolein oil blends) were analyzed for odor changes during deep fat frying by sensory analysis and e-nose. Samples were drawn after 1, 4, 7, and 10 sessions of frying. E-nose analysis showed good discrimination between the initial and fried oil samples, indicating that the odor pattern of oil blends changed during frying [98]. The storage stability of RBD palmolein was determined using SAW sensor-based e-nose. High correlation was obtained between e-nose measurements, chemical analyses, and sensory evaluation data [99]. Another application of e-nose in food analysis is the detection of adulteration especially in oils. The e-nose was used to detect the maize oil adulteration in camellia seed oil and sesame oil [30,100]. Virgin olive oils adulterated with sunﬂower oil and olive–pomace oil were detected using an e-nose, and promising results were obtained for quantiﬁcation of the percentages of adulteration [101].

16.6.6 OTHER FOOD APPLICATIONS Alasalvar et al. [102] compared e-nose, GC=MS, and sensory evaluation in the classiﬁcation of hazelnuts. E-nose was used to determine the difference between the palm civet coffee (Kopi Luwak) and African civet coffee aroma proﬁles, and the instrument was capable of distinguishing both civet coffees from their controls [103]. Tognon et al. [104] used an e-nose to determine the levels of mycotoxin contamination in durum wheat. E-nose was found to be useful for evaluation of the presence of mycotoxins in wheat and could be used for a preliminary screening of wheat stocks in order to reduce the number of samples requiring chemical analysis. Two e-noses were employed for the early detection and discrimination between bacterial species, fungal spores, and trace amounts of pesticides in potable water. Using PCA, DFA, and CA, it was possible to differentiate between different bacterial and fungal species after incubating at 258C for 24 h, but the pesticides (10 and 100 ppb) could not be effectively discriminated from the controls [105]. Zhang et al. [38,39] applied the time delay neural networks to the identiﬁcation of spice mixture, and compared the e-nose to the GC and sensory methods for accuracy of compositional prediction. GC and e-nose results were accurate and comparable, but panelists did not have a good accuracy in predicting spice mixture compositions.

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 374 16.7.2008 11:04pm Compositor Name: MSubramanian

374

Handbook of Food Analysis Instruments

16.7 CONCLUSION AND FUTURE TRENDS The e-nose offers possibilities for development of quick, accurate, and full scale determination of shelf life, detection of spoilage indicators, rapid identiﬁcation of undesired microorganisms, and rapid measurement of spoilage. However, there are limitations for the selectivity of the sensors to speciﬁc volatiles. Either more sensors are added to a sensor array or narrower ranges of selectivity which are speciﬁc to a target volatile compound are utilized in the instruments. New sensors having fast, accurate, and reproducible responses and lower prices need to be developed. New sensors based on different principles may overcome some of the sensor limitations. With the development of faster and better selectivity sensors, e-nose could be used as a potential tool for online food quality evaluation.

REFERENCES 1. Deshpande, S.S., Biosensors in food quality evaluation, in Gunasekaran, S. (Ed.), Nondestructive Food Evaluation, Marcel Dekker Inc., New York, 2001, pp. 335–377. 2. Persaud, K.C. and Dodd, G.H., Analysis of the mammalian olfactory system using a model nose, Nature, 299, 352–355, 1982. 3. Ballabio, D., Cosio, M.S., Mannino, S., and Todeschini, R., A chemometric approach based on a novel similarity=diversity measure for the characterisation and selection of electronic nose sensors, Anal. Chim. Acta, 578, 170–177, 2006. 4. Zhang, H., Using electronic noses to assess food quality, in Tothill, I.E. (Ed.), Rapid and On-line Instrumentation for Food Quality Assurance, CRC Press, Boca Raton, Florida, 2003, pp. 324–338. 5. Panigrahi, S., Balasubramanian, S., Gu, H., Logue, C., and Marchello, M., Neural-network-integrated e-nose system for identiﬁcation of spoiled beef, LWT, 39, 135–145, 2006. 6. Stussi, E., Cella, S., Serra, G., and Venier, G.S., Fabrication of conducting polymer patterns for gas sensing by a dry technique, Mater. Sci. Eng. C, 4, 27–33, 1996. 7. Gardner, J.W. and Bartlett, P.N., A brief history of electronic noses, Sens. Actuators B, 18–19, 211–220, 1994. 8. García, M., Aleixandre, M., Gutiérrez, J., and Horrillo, M.C., Electronic nose for ham discrimination, Sens. Actuators B, 114, 418–422, 2006. 9. Falcitelli, M., Benassi, A., Di Francesco, F., Domenici, C., Marano, L., and Pioggia, G., Fluid dynamic simulation of a measurement chamber for electronic noses, Sens. Actuators B, 85, 166–174, 2002. 10. Doleman, B.J. and Lewis, N.S., Comparison of odor detection thresholds and odor discriminabilities of a conducting polymer composite electronic nose versus mammalian olfaction, Sens. Actuators B, 72, 41–50, 2001. 11. Kent, M., Oehlenschlager, J., Mierke-Klemeyer, S., Manthey-Karl, M., Knöchel, R., Daschner, F., and Schimmer, O., A new multivariate approach to the problem of ﬁsh quality estimation, Food Chem., 87, 531–535, 2004. 12. Gardner, J.W. and Bartlett, P.N., Electronic Noses, Principles and Applications, Oxford University Press, New York, 1999, 1–245. 13. Shurmer, H.V. and Gardner, J.W., Odor discrimination with an electronic nose, Sens. Actuators B, 8, 1–11, 1992. 14. Schaller, E., Bosset, J.O., and Escher, F., ‘‘Electronic noses’’ and their application to food, Lebensm.Wiss. Technol., 31, 305–316, 1998. 15. Corcoran, P., Electronic odour sensing systems, Electron. Commun. Eng. J., 5, 303–308, 1993. 16. Annor-Frempong, I.E., Nute, G.R., Wood, J.D., Whittington, F.W., and West, A., The measurement of the responses of different odor intensities of ‘‘boar taint’’ using a sensory panel and an electronic nose, Meat Sci., 50, 139–151, 1998. 17. Bartlett, P.N., Elliot, J.M., and Gardner, J.W., Electronic noses and their application in the food industry, Food Technol., 51(12), 44–48, 1997. 18. Bai, H. and Shi, G., Gas sensors based on conducting polymers, Sensors, 7, 267–307, 2007. 19. Matthews, B., Li, J., Sunshine, S., Lerner, L., and Judy, J.W., Effects of electrode conﬁguration on polymer carbon-black composite chemical vapor sensor performance, IEEE Sens. J., 2(3), 160–168, 2002.

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 375 16.7.2008 11:04pm Compositor Name: MSubramanian

Electronic Nose Technology in Food Analysis

375

20. Haugen, J.E. and Kvaal, K., Electronic nose and artiﬁcial neural network, Meat Sci., 49, S273–S286, 1998. 21. Honh, H.K., Shin, H.W., Yun, D.Y., Kim, S.R., Kwon, C.H., Lee, K., and Moriizumi, T., E-nose system with micro-gas sensor array, Sens. Actuators B, 35–36, 338–341, 1996. 22. Egashira, M. and Shimizu, Y., Odour sensing by semiconductor metal oxides, Sens. Actuators B, 13–14, 443–446, 1993. 23. Ólafsdóttir, G., Rapid olfaction arrays for determining ﬁsh quality, in Tothill, I.E. (Ed.), Rapid and On-line Instrumentation for Food Quality Assurance, CRC Press, Boca Raton, Florida, 2003, pp. 339–360. 24. Deshpande, S.S. and Rocco, R.M., Biosensors and their potential use in food quality, Food Technol., 48(6), 146–150, 1994. 25. Ballantine, D.S. and Wohltjen, H., Surface acoustic wave devices for chemical analysis, Anal. Chem., 61, 704A–706A, 1989. 26. Hodgins, D., The electronic nose: Sensor array-based instruments that emulate the human nose, in R. Marsili (Ed.), Techniques for Analyzing Food Arom, Marcel Dekker, New York, 1997, pp. 331–371. 27. Suslick, K.S. and Rakow, N.A., A colorimetric nose: ‘‘Smell-seeing,’’ in Stetter, J.R. and Pensrose, W.R. (Eds.), Artiﬁcial Chemical Sensing: Olfaction and the Electronic Nose, Electrochem. Soc., Pennington, New Jersey, 2001, pp. 8–14. 28. Rakow, N.A. and Suslick, K.S., A colorimetric sensor array for odour visualization, Nature, 406, 710–713, 2000. 29. Dittmann, B., Zimmermann, B., Engelen, C., Jany, G., and Nitz, S., Use of the MS-sensor to discriminate between different dosages of garlic ﬂavoring in tomato sauce, J. Agri. Food Chem., 48, 2887–2897, 2000. 30. Hai, Z. and Wang, J., Detection of adulteration in camellia seed oil and sesame oil using an electronic nose, Eur. J. Lipid Sci. Technol., 108, 116–124, 2006. 31. Shin, H.W., Llobet, E., Gardner, J.W., Hines, E.L., and Dow, C.S., Classiﬁcation of the strain and growth phase of cyanobacteria in potable water using an electronic nose system, IEE Proc. Sci. Meas. Technol., 147, 158–164, 2000. 32. Srivastava, A.K., Detection of volatile organic compounds using SnO2 gas-sensor array and artiﬁcial neural network, Sens. Actuators B, 96, 24–37, 2003. 33. Kermani, B.G., Schiffman, S.S., and Nagle, H.T., Using neural networks and genetic algorithms to enhance performance in an electronic nose, IEEE Trans. Biomed. Eng., 46, 429–439, 1999. 34. Qu, G., Feddes, J.J.R., Armstrong, W.W., Coleman, R.N., and Leonard, J.J., Measuring odor concentration with an electronic nose, Trans. Am. Soc. Agric. Eng., 44, 1807–1812, 2001. 35. Boilot, P., Hines, E.L., Gardner, J.W., Pitt, R., John, S., Mitchell, J., and Morgan, D.W., Classiﬁcation of bacteria responsible for ENT and eye infections using the Cyranose system, IEEE Sens. J., 2, 247–253, 2002. 36. Sinesio, F., Di Natale, C., Quaglia, G.B., Bucarelli, F.M., Moneta, E., Macagnano, A., Paolesse, R., and D’Amico, A., Use of electronic nose and trained sensory panel in the evaluation of tomato quality, J. Sci. Food Agric., 80, 63–71, 2000. 37. Zhang, H., Balaban, M.Ö., and Principe, J.C., Improving pattern recognition of electronic nose data with time-delay neural networks, Sens. Actuators B, 96, 385–389, 2003. 38. Zhang, H., Balaban, M.Ö., Principe, J.C., and Portier, K., Quantiﬁcation of spice mixture compositions by electronic nose: Part I. Experimental design and data analysis using neural networks, J. Food Sci., 70, E253–E258, 2005. 39. Zhang, H., Balaban, M.Ö., Portier, K., and Sims, C.A., Quantiﬁcation of spice mixture compositions by electronic nose: Part II. Comparison with GC and sensory methods, J. Food Sci., 70, E259–264, 2005. 40. Dutta, R., Hines, E.L., Gardner, J.W., and Boilot, P., Bacteria classiﬁcation using Cyranose 320 electronic nose, Biomed. Eng. Online, 1, 4, 2002. 41. Snopok, B.A. and Kruglenko, I.V., Multisensor systems for chemical analysis: State-of-art in electronic nose technology and new trends in machine olfaction, Thin Solid Films, 418, 21–41, 2002. 42. Balaban, M.O., Korel, F., Odabasi, A., and Folkes, G., Transportability of data between electronic noses: Mathematical methods, Sens. Actuators B, 71, 203–211, 2000. 43. Labreche, S., Bazzo, S., Cade, S., and Chanie, E., Shelf life determination by electronic nose: Application to milk, Sens. Actuators B, 106, 199–206, 2005. 44. Korel, F., Luzuriaga, D.A., and Balaban, M.O., Microbial, sensory and electronic nose evaluation of pasteurized whole milk, in Hurst, W.J. (Ed.), Electronic Noses & Sensor Array Based Systems Design and Applications, Technomic Pub. Co., Lancaster, Philadelphia, 1999, pp. 154–161.

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 376 16.7.2008 11:04pm Compositor Name: MSubramanian

376

Handbook of Food Analysis Instruments

45. Korel, F. and Balaban, M.Ö., Microbial and sensory assessment of milk with an electronic nose, J. Food Sci., 67, 758–764, 2002. 46. Haugen, J.E., Rudi, K., Langsrud, S., and Bredholt, S., Application of gas-sensor array technology for detection and monitoring of growth of spoilage bacteria in milk: A model study, Anal. Chim. Acta, 565, 10–16, 2006. 47. Trihaas, J., Van Den Tempel, T., and Nielsen, P.V., Electronic nose technology in quality assessment: Predicting volatile composition of Danish blue cheese during ripening, J. Food Sci., 70, E392–E400, 2005. 48. Bargon, J., Braschob, S., Flörke, J., Herrmann, U., Klein, L., Loergen, J.W., Lopez, M., Maric, S., Parham, A.H., Piacenza, P., Schaefgen, H., Schalley, C.A., Silva, G., Schlupp, M., Schwierz, H., Vögtle, F., and Windscheif, G., Determination of ripening state of Emmental cheese via quartz microbalances, Sens. Actuators B, 95, 6–19, 2003. 49. Miettinen, S.-M., Piironen, V., Tuorila, H., and Hyvönen, L., Electronic and human nose in the detection of aroma differences between strawberry ice cream of varying fat content, J. Food Sci., 67, 425–430, 2002. 50. Turhan, M., Balaban, M.O., Turhan, K.N., and Luzuriaga, D.A., Potential use of electronic nose technique for detection of meat adulteration: Separation of pork-beef mixtures, Fleischwirtschaft Int., 6, 26–28, 1998. 51. Siegmund, B. and Pfannhauser, W., Changes of the volatile fraction of cooked chicken meat during chill storage: Results obtained by the electronic nose in comparison to GC-MS and GC olfactometry, Z. Lebensm. Unters. Forsch A, 208, 336–341, 1999. 52. Rajamäki, T., Alakomi, H.-L., Ritvanen, T., Skyttä, E., Smolander, M., and Ahvenainen, R., Application of an electronic nose for quality assessment of modiﬁed atmosphere packaged poultry meat, Food Control, 17, 5–13, 2006. 53. Boothe, D.D.H. and Arnold, J.W., Electronic nose analysis of volatile compounds from poultry meat samples, fresh and after refrigerated storage, J. Sci. Food Agric., 82, 315–322, 2002. 54. Eklöv, T., Johansson, G., Winquist, F., and Lundström, I., Monitoring sausage fermentation using an electronic nose, J. Sci. Food Agric., 76, 525–532, 1998. 55. Luzuriaga, D.A. and Balaban, M.O., Evaluation of the odor of decomposition in raw and cooked shrimp: Correlation of electronic nose readings, odor sensory evaluation and ammonia levels, in Hurst, W.J. (Ed.), Electronic Noses & Sensor Array Based Systems Design and Applications, Technomic Pub. Co., Lancaster, Pennsylvania, 1999, pp. 177–184. 56. Luzuriaga, D.A. and Balaban, M.O., Electronic nose odor evaluation of salmon ﬁllets stored at different temperatures, in Hurst, W.J. (Ed.), Electronic Noses & Sensor Array Based Systems Design and Applications, Technomic Pub. Co., Lancaster, Pennsylvania, 1999b, pp. 162–169. 57. Olafsdottir, G., Hognadottir, A., Martinsdóttir, E., and Jonsdottir, R., Application of an electronic nose to predict total volatile bases in capelin (Mallotus villosus) for ﬁshmeal production, J. Agric. Food Chem., 48, 2353–2359, 2000. 58. Olafsdottir, G., Xiuchen, I., Lauzon, H., and Jonsdottir, R., Precision and application of electronic nose for freshness monitoring of whole redﬁsh stored in ice and modiﬁed atmosphere bulk storage, J. Aquat. Food Prod. Technol., 11, 229–249, 2002. 59. Di Natale, C., Olafsdottir, G., Einarsson, S., Mantin, A., Martinelli, E., Paolesse, R., Falconi, C., Esaiassen, M., Nilsen, H., Joensen, S., Skjerdal, T., Carlehog, M., Eilertsen, G., Gundersen, B., and Elvevoll, E., Effects of catching methods on quality changes during storage of cod (Gadus morhua), Lebensm. Wiss. Technol., 37, 643–648, 2004. 60. Luzuriaga, D.A., Korel, F., and Balaban, M.O., Odor evaluation of shrimp treated with different chemicals using an electronic nose and a sensory panel, J. Aquat. Food Prod. Technol., 16, 57–75, 2007. 61. Macagnano, A., Careche, M., Herrero, A., Paolesse, R., Martinelli, E., Pennazza, G., Carmona, P., D’Amico, A., and Di Natale, C., A model to predict ﬁsh quality from instrumental features, Sens. Actuators B, 111–112, 293–298, 2005. 62. Olafsdottir, G., Chanie, E., Westad, F., Jonsdottir, R., Thalmann, C.R., Bazzo, S., Labreche, S., Marcq, P., Lundby, F., and Haugen, J.E., Prediction of microbial and sensory quality of cold smoked Atlantic salmon (Salmo salar) by electronic nose, J. Food Sci., 70, S563–S574, 2005. 63. Oliveira, A.C.M., Crapo, C.A., Himelbloom, B., Vorholt, C., and Hoffert, J., Headspace gas chromatography-mass spectrometry and electronic nose analysis of volatile compounds in canned Alaska pink salmon having various grades of watermarking, J. Food Sci., 70, S419–S426, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 377 16.7.2008 11:04pm Compositor Name: MSubramanian

Electronic Nose Technology in Food Analysis

377

64. Olafsdottir, G., Lauzon, H.L., Martinsdóttir, E., Oehlenschläger, J., and Kristbergsson, K., Evaluation of shelf life of superchilled cod (Gadus morhua) ﬁllets and the inﬂuence of temperature ﬂuctuations during storage on microbial and chemical quality indicators, J. Food Sci., 71, S97–S109, 2006. 65. Korel, F., Luzuriaga, D.A., and Balaban, M.Ö., Quality evaluation of raw and cooked catﬁsh (Ictalurus punctatus) using electronic nose and machine vision, J. Aquat. Food Prod. Technol., 10, 3–18, 2001. 66. Korel, F., Luzuriaga, D.A., and Balaban, M.Ö., Objective quality assessment of raw Tilapia (Oreochromis niloticus) ﬁllets using electronic nose and machine vision, J. Food Sci., 66, 1018–1024, 2001. 67. Newman, D.J., Luzuriaga, D.A., and Balaban, M.O., Odor and microbiological evaluation of raw tuna: Correlation of sensory and electronic nose data, in Hurst, W.J. (Ed.), Electronic Noses & Sensor Array Based Systems Design and Applications, Technomic Pub. Co., Lancaster, Pennsylvania, 1999, pp. 170–176. 68. Gómez, A.H., Hu, G., Wang, J., and Pereira, A.G., Evaluation of tomato maturity by electronic nose, Comput. Electron. Agric., 54, 44–52, 2006. 69. Di Natale, C., Macagnano, A., Martinelli, E., Paolesse, R., Proietti, E., and D’Amico, A., The evaluation of quality of post-harvest oranges and apples by means of an electronic nose, Sens. Actuators B, 78, 26–31, 2001. 70. Oshita, S., Shima, K., Haruta, T., Seo, Y., Kagawoe, Y., Nakayama, S., and Kawana, S., Discrimination of odors emanating from ‘‘La France’’ pear by semi-conducting polymer sensors, Comput. Electron. Agric., 26, 209–216, 2000. 71. Correa, E., Barreiro, P., Ruiz-Altisent, M., Lopez, M.L., Miro, J., and Graeli, J., An aroma sensor for assessing peach quality, in Proc. Sixth Int. Symp. Fruits, Nuts, and Vegetables Production Engineering, Potsdam, Germany, 162, 2001. 72. Brezmes, J., Llobet, E., Vilanova, X., Saiz, G., and Correig, X., Fruit ripeness monitoring using an electronic nose, Sens. Actuators B, 69, 223–229, 2000. 73. Di Natale, C., Macagnano, A., Martinelli, E., Proietti, E., Paolesse, R., Castellari, L., Campani, S., and D’Amico, A., Electronic nose based the investigation of the sensorial properties of peaches and nectarines, Sens. Actuators B, 77, 561–566, 2001. 74. Brezmes, J., Llobet, E., Vilanova, X., Orts, J., Saiz, G., and Correig, X., Correlation between electronic nose signals and fruit quality indicators on shelf-life measurements with pinklady apples, Sens. Actuators B, 80, 41–50, 2001. 75. Saevels, S., Lammertyn, J., Berna, A.Z., Veraverbeke, E.A., Di Natale, C., and Nicolaï, B.M., Electronic nose as a non-destructive tool to evaluate the optimal harvest date of apples, Postharvest Biol. Technol., 30, 3–14, 2003. 76. Maul, F., Sargent, S.A., Huber, D.J., Balaban, M.O., Sims, C.A., and Baldwin, E.A., Harvest maturity and storage temperature affect volatile proﬁles of ripe tomato fruits: Electronic nose and gas chromatographic analyses, in Hurst, W.J. (Ed.), Electronic Noses & Sensor Array Based Systems Design and Applications, Technomic Pub. Co., Lancaster, PA, 1999, pp. 1–13. 77. Moretti, C.L., Sargent, S.A., Balaban, M.O., and Puschmann, R., Electronic nose: Non-destructive technology to screen tomato fruit with internal bruising, Horticultura Brasileira, 18, 20–23, 2000. 78. Gómez, A.H., Wang, J., Hu, G., and Pereira, A.G., Electronic nose technique potential monitoring mandarin maturity, Sens. Actuators B, 113, 347–353, 2006. 79. Berna, A.Z., Lammertyn, J., Saevels, S., Di Natale, C., and Nicolaï, B.M., Electronic nose systems to study shelf life and cultivar effect on tomato aroma proﬁle, Sens. Actuators B, 97, 324–333, 2004. 80. Farnworth, E.D., Mckellar, R.C., Chabot, D., Lapointe, S., Chicoine, M., and Knight, K.P., Use of an electronic nose to study the contribution of volatiles to orange juice ﬂavor, J. Food Qual., 25, 569–576, 2002. 81. Solis-Solis, H.M., Calderon-Santoyo, M., Gutierrez-Martinez, P., Schorr-Galindo, S., and RagazzoSanchez, J.A., Discrimination of eight varieties of apricot (Prunus armeniaca) by electronic nose, LLE and SPME using GC-MS and multivariate analysis, Sens. Actuators B, 125, 415–421, 2007. 82. García, M., Fernández, M.J., Fontecha, J.L., Lozano, J., Santos, J.P., Aleixandre, M., Sayago, J., Gutiérrez, J., and Horrillo, M.C., Differentiation of red wines using an electronic nose based on surface acoustic wave devices, Talanta, 68, 1162–1165, 2006. 83. García, M., Aleixandre, M., Gutiérrez, J., and Horrillo, M.C., Electronic nose for wine discrimination, Sens. Actuators B, 113, 911–916, 2006. 84. Cozzolino, D., Smyth, H.E., Lattey, K.A., Cynkar, W., Janik, L., Dambergs, R.G., Francis, I.L., and Gishen, M., Combining mass spectrometry based electronic nose, visible-near infrared spectroscopy and

Otles/Handbook of Food Analysis Instruments 45660_C016 Final Proof page 378 16.7.2008 11:04pm Compositor Name: MSubramanian

378

85. 86. 87.

88.

89. 90. 91. 92.

93.

94.

95.

96. 97. 98. 99. 100. 101.

102.

103. 104.

105.

Handbook of Food Analysis Instruments chemometrics to assess the sensory properties of Australian Riesling wines, Anal. Chim. Acta, 563, 319–324, 2006. Lozano, J., Santos, J.P., Aleixandre, M., Sayago, I., Gutiérrez, J., and Horrillo, M.C., Identiﬁcation of typical wine aromas by means of an electronic nose, IEEE Sens. J., 6, 173–178, 2006. Lozano, J., Santos, J.P., and Horrillo, M.C., Classiﬁcation of white wine aromas with an electronic nose, Talanta, 67, 610–616, 2005. Ragazzo-Sanchez, J.A., Chalier, P., and Ghommidh, C., Coupling gas chromatography and electronic nose for dehydration and desalcoholization of alcoholized beverages: Application of off-ﬂavour detection in wine, Sens. Actuators B, 106, 253–257, 2005. McKellar, R.C., Rupasinghe, H.P.V., Lu, X., and Knight, K.P., The electronic nose as a tool for the classiﬁcation of fruit and grape wines from different Ontario wineries, J. Sci. Food Agric., 85, 2391–2396, 2005. Anon, E.C., Ofﬁcial Journal European Communities, May 15, Regulation 796=2002, 2002. Morales, D.L., Luna, G., and Aparicio, R., Comparative study of virgin olive oil sensory defects, Food Chem., 91, 293–301, 2005. García-González, D.L. and Aparicio, R., Sensors: From biosensors to the electronic nose, Grasas Aceites, 53, 96–114, 2002. Cosio, M.S., Ballabio, D., Benedetti, S., and Gigliotti, C., Geographical origin and authentication of extra virgin olive oils by an electronic nose in combination with artiﬁcial neural networks, Anal. Chim. Acta, 567, 202–210, 2006. Guadarrama, A., Rodríguez-Méndez, M.L., Sanz, C., Ríos, J.L., and de Saja, J.A., Electronic nose based on conducting polymers for the quality control of olive oil aroma—Discrimination of quality, variety of olive and geographic origin, Anal. Chim. Acta, 432, 283–292, 2001. García-González, D.L., Barié, N., Rapp, M., and Aparicio, R., A fuzzy ﬁlter to study the selectivity and sensitivity of a SPME enhanced SAW sensor system characterizing virgin olive oil aroma, Sens. Act. B, 116, 49–54, 2006. Brezmes, J., Cabré, P., Rojo, S., Llobet, E., Vilanova, X., and Correig, X., Discrimination between different samples of olive oil using variable selection techniques and modiﬁed fuzzy artmap neural networks, IEEE Sens. J., 5, 463–470, 2005. García-González, D.L. and Aparicio, R., Detection of vinegary defect in virgin olive oil by metal oxide sensors, J. Agric. Food Chem., 50, 1809–1814, 2002. Gan, H.L., Che Man, Y.B., Tan, C.P., NorAini, I., and Nazimah, S.A.H., Characterization of vegetable oils by surface acoustic wave sensing electronic nose, Food Chem., 89, 507–518, 2005. Raj, P.N., Prakash, M., and Bhat, K.K., Quality assessment of oil blends by electronic nose technique and sensory methods, J. Sens. Stud., 21, 322–332, 2006. Gan, H.L., Tan, C.P., Che Man, Y.B., NorAini, I., and Nazimah, S.A.H., Monitoring the storage stability of RBD palm olein using the electronic nose, Food Chem., 89, 271–282, 2005. Hai, Z. and Wang, J., Electronic nose and data analysis for detection of maize oil adulteration in sesame oil, Sens. Actuators B, 119, 449–455, 2006. Oliveros, M.C.C., Pavón, J.L.P., Pinto, C.G., Laespada, M.E.F., Cordero, B.M., and Forina, M., Electronic nose based on metal oxide semiconductor sensors as a fast alternative for the detection of adulteration of virgin olive oils, Anal. Chim. Acta, 459, 219–228, 2002. Alasalvar, C., Odabasi, A.Z., Demir, N., Balaban, M.O., Shahidi, F., and Cadwallader, K.R., Volatile and ﬂavor of ﬁve Turkish hazelnut varieties as evaluated by descriptive sensory analysis, electronic nose, and dynamic headspace analysis=gas chromatography-mass spectrometry, J. Food Sci., 69, SNQ99–106, 2004. Marcone, M.F., Composition and properties of Indonesian palm civet coffee, Food Res. Int., 37, 901–912, 2004. Tognon, G., Campagnoli, A., Pinotti, L., Dell’Orto, V., and Cheli, F., Implementation of the electronic nose for the identiﬁcation of mycotoxins in durum wheat (Triticum durum), Vet. Res. Commun., 29 (Suppl. 2), 391–393, 2005. Canhoto, O. and Magan, N., Electronic nose technology for the detection of microbial and chemical contamination of potable water, Sens. Actuators B, 106, 3–6, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 379 14.7.2008 8:24pm Compositor Name: BMani

Techniques 17 Electroanalytical and Instrumentation in Food Analysis Rubin Gulaboski and Carlos M. Pereira CONTENTS 17.1 17.2 17.3 17.4

Introduction ........................................................................................................................ 379 Principles of Voltammetric Techniques ............................................................................ 380 Theory and Deﬁnitions in Voltammetric Techniques ....................................................... 381 Instrumentation in Voltammetric Experiments ................................................................. 382 17.4.1 Potentiostat ........................................................................................................... 382 17.4.2 Electrochemical Cell and Electrodes ................................................................... 384 17.5 Short Review of Some Common Voltammetric Techniques Used in Food Analysis ...... 384 17.5.1 Cyclic Voltammetry ............................................................................................. 385 17.5.2 Pulse Voltammetric Techniques .......................................................................... 386 17.5.2.1 Normal Pulse Voltammetry ................................................................ 386 17.5.2.2 Differential Pulse Voltammetry .......................................................... 387 17.5.2.3 Square-Wave Voltammetry ................................................................. 388 17.6 Preconcentration Voltammetric Techniques ...................................................................... 388 17.6.1 Adsorptive Stripping Voltammetry ..................................................................... 389 17.6.2 Anodic Stripping Voltammetry ........................................................................... 389 17.6.3 Cathodic Stripping Voltammetry ........................................................................ 390 17.7 Electrochemical Techniques in Food Analysis ................................................................. 390 17.7.1 Food Colorants ..................................................................................................... 390 17.7.2 Metal Contaminants ............................................................................................. 390 17.7.3 Pesticides and Herbicides .................................................................................... 391 17.7.4 Biosensors ............................................................................................................ 395 17.7.5 Other Applications ............................................................................................... 395 17.7.6 Hyphenated Techniques ....................................................................................... 395 17.8 New Trends ....................................................................................................................... 396 References ..................................................................................................................................... 396

17.1 INTRODUCTION Electrochemical techniques are inevitable tools in almost every chemical and biochemical research laboratory. In addition to their application in fundamental studies of oxidation and reduction processes to unravel reaction mechanisms, these techniques are also used in studying the kinetics and thermodynamics of electron and ion transfer processes [1]. Moreover, electrochemical techniques have also proven to be useful tools for the study of adsorption and crystallization phenomena at electrode surfaces [2]. Among the electrochemical techniques applied in food analysis, the

379

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 380 14.7.2008 8:24pm Compositor Name: BMani

380

Handbook of Food Analysis Instruments

principal ones are polarographic and voltammetric techniques [3]. Their wide application is attributed to the relatively cheap instrumentation, very good sensitivity with wide linear concentration ranges for both inorganic and organic compounds, rapid analysis times (in seconds), and simultaneous determination of several analytes. Currently, polarographic techniques have almost been completely excluded from research laboratories, and are being replaced by the more sophisticated voltammetric techniques [1]. Voltammetry (abbreviation of volt-amper-metry) is a branch of electrochemistry that was developed by the discovery of polarography in 1922 by Jaroslav Heyrovsky (Nobel Prize in 1959). A major breakthrough in voltammetry was made in the early 1960s, when an expanded repertoire of analytical methods was reported, appearing in parallel with the corresponding well-developed theories [1,4]. At the same time, these developments led to enhanced sensitivities obtained with the voltammetric techniques. The excitation signal in all voltammetric techniques is the applied potential difference (or potential as it is commonly referred to), E, between the electrodes, whereas the monitoring output parameter is the resulting current, I, ﬂowing through the electrochemical cell. Electrochemistry provides powerful and versatile tools for food analysis: powerful with regard to detecting very low concentrations, while providing a wide linear relationship between the measured signal (current intensity) and concentration; versatile as a consequence of the possibility of simultaneous analysis, extended to a large number of organic and inorganic compounds that can be found in food. One of the main advantages of electrochemical techniques is usually considered to be the possibility of direct analysis of the sample without tedious and long preparative steps and subsequent separation. Voltammetric measurements can be applied directly to colored systems, in the presence of suspended matter, or even to colloidal systems. When using biological or complex environmental samples, pretreatment is required. However, this process is faster, cheaper, and easier when compared with the standard treatment used to prepare samples to be analyzed by chromatographic techniques [5]. When compared with chromatographic techniques, effect of interference is lower when electrochemical techniques are used [6]. If required, voltammetric methods can also be easily hyphenated with other techniques, e.g., their use as detectors in chromatographic separations [7]. Another advantage of voltammetric techniques in food analysis is the various possible modes and methodologies that can be used for the determination of both electrochemically active and electrochemically inactive compounds [8].

17.2 PRINCIPLES OF VOLTAMMETRIC TECHNIQUES Voltammetric experiments are usually carried out in simple electrochemical cells, similar to that shown in Figure 17.1. A common electrochemical cell consists of a working electrode, a reference electrode, and usually an auxiliary (counter) electrode. The working electrode is an electron conductor at which the reaction or transfer of interest takes place. In all voltammetric experiments, it is necessary to keep one of the electrodes at a constant potential. This electrode, designed to have a constant (reversible) potential, is called a reference electrode. The auxiliary electrode is one at which a counter reaction to that at the working electrode takes place, for the sake of balancing the total charge in the system. In the presence of electroactive species in the electrochemical working cell, the applied potential will provoke a change in the concentration of the monitored electroactive species at the electrode surface by electrochemically reducing or oxidizing them. Changing the concentration of any electroactive participant at the working electrode surface will cause mass transport toward the electrode, and current ﬂow through the electrode is directly proportional to the analyte concentration. This simple dependence between measured current and analyte concentration makes voltammetric techniques to be routinely used for the quantitative determination of a variety of inorganic and organic compounds. Although there are many voltammetric techniques, they are all based on the same electrochemical theory. Summarized here are some of the electrochemical laws common to all voltammetric techniques.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 381 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

381

V

A R C W

Purging gas

FIGURE 17.1 Schematic representation of a common electrochemical cell. R, reference electrode; W, working electrode; C, counter electrode; V, voltmeter; A, amperemeter.

17.3 THEORY AND DEFINITIONS IN VOLTAMMETRIC TECHNIQUES Consider the simplest electrochemical reaction of the type Ox þ ne ! Red

(I)

where Ox refers to the oxidized form of an electroactive substance initially present in the electrode cell, while Red is its reduced form (charges are omitted for the sake of simplicity), at least two wellknown laws can be applied for expressing the interdependence between the applied potential and the surface concentrations of Ox and Red. For a thermodynamically reversible electrochemical reaction (i.e., a fast reaction where equilibrium is always reestablished as changes in electrode potential are made), the following type of Nernst equation always holds: E ¼ Eu þ

RT c(Ox)x¼0 ln nF c(Red)x¼0

(17:1)

where R is the gas constant (8.3144 J=mol K) T is the absolute temperature (K) n is the number of electrons exchanged F is the Faraday constant (96,485 C=mol) Eu is the standard redox potential for the couple Ox=Red For many electrochemical systems, especially for kinetics-controlled equations, the Butler–Volmer equation is often used:

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 382 14.7.2008 8:24pm Compositor Name: BMani

382

Handbook of Food Analysis Instruments

I ¼ ks eaw ½c(Ox)x¼0 ew c(Red)x¼0 nFA

(17:2)

where w ¼ nF(E – Eu)=RT ks is the standard rate constant of electron transfer a is the electron transfer coefﬁcient A is the exposed (active) area of the working electrode This equation is useful for estimating the heterogeneous standard rate constant of electron transfer ks. When performing electrochemical experiments, applying a potential difference between the working and reference electrodes would alter the surface concentrations of both forms of the redox couple. This will cause a mass transfer toward (or from) the working electrode and current will ﬂow through the cell. The current resulting from the electrochemical transformation of the electrochemically active species is known as faradaic current, and it is related to the material ﬂux at the electrode–solution interface, as described by well-known Fick’s laws [1]. The instrumental output that is obtained in the voltammetric techniques is known as a voltammogram, which is a current–potential or I–E curve. The features of the obtained voltammograms depend on the type of the mass transfer phenomena, the number of exchanged electrons, the timescale of the measurement, as well as on the nature of the various coupled chemical reactions and surface phenomena that can happen in the electrochemical cell, and on the voltammetric technique used, in particular [1]. For analytical purposes, one is mainly interested to know the magnitude of the faradaic current of the voltammograms, which is a quantitative measure of concentration and the kinetics of the redox transformation of given electroactive species at the electrode surface. The magnitude of the faradaic current is a function of the analyte concentration, but it is also affected by additional factors, such as the size, shape, and material of the electrode, the solution resistance, the cell volume, and the number of electrons transferred in the electrode reaction [1].

17.4 INSTRUMENTATION IN VOLTAMMETRIC EXPERIMENTS The modern electroanalytical system for voltammetric measurements is usually composed of three modules: a potentiostat, a personal computer, and an electrochemical cell (Figure 17.2). In some cases, the potentiostat and computer are packed into one part, whereas in most of the electrochemical systems the computer and the converters and microcontrollers are separate, so the potentiostat can operate autonomously.

17.4.1 POTENTIOSTAT The potentiostat is considered the ‘‘heart’’ of every electroanalytical instrumentation. In voltammetric techniques, the task of the potentiostat is to apply an exact potential and to observe

RE

Digital controller

Computer –Parameter entry –Waveform generation –Data display

CE Potentiostat WE Electrochemical cell

FIGURE 17.2

Simpliﬁed scheme of an electrochemical system designed for voltammetric measurements.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 383 14.7.2008 8:24pm Compositor Name: BMani

383

Electroanalytical Techniques and Instrumentation in Food Analysis

the current changes in the system. In all modern electroanalytical instruments, the potentiostat package includes electrometer circuits, various converters and ampliﬁers, as well as microprocessors with internal memory. In older types of instruments, continuous linear change in the potential from one preset value to another was applied. However, all the modern potentiostats designed after 1980s operate in a digital (incremental) fashion. In these, a ‘‘staircase’’ modulated potential is generally used, with the ‘‘steps’’ having constant potential increment. The major beneﬁts of staircase-modulated potentials are seen by the signiﬁcant discrimination over the capacitative (nonfaradaic) currents [1]. Thanks to the digital fabrication of the applied potential, a wealth of pulsed voltammetric techniques have been designed [9], leading to increased sensitivity and much shorter times for performing experiments. The most frequent waveforms in modern potentiostats are linear scan, differential pulse, and square wave. Nowadays, various types of potentiostats can be found at the instrumentation markets, the size, power, and sophistication of which ranges from large researchgrade instruments (10–30 kg with 30 V potential and 1 A–100 nA current ranges) to simple batterypowered units (3–10 kg with 2.5 V potential and 6 mA–50 pA current ranges). Indeed, the type of voltammetric analysis, the required information, and the size of the electrodes are the main factors that will determine the choice of a particular instrument. While cyclic voltammetry can be performed with most of the commercially available potentiostats, quantitative tracing of some analytes requires use of microelectrodes and a sensitive voltammetric technique, which will be more expensive. Currently, several companies manufacture high-quality potentiostats capable of performing various voltammetric analyses. Among them, EG&G Princeton Applied Research, EcoChemie Netherlands, ACM Instruments, Cypress Systems, and Radiometer are the leaders. Along with the instrumentation, the manufacturer usually provides the software for data analysis, as well as the electrochemical cells and the electrodes. The problems with voltammetric procedures usually are related to a part of the system external to the instrument. In case of instrumental troubles, the ﬁrst source of help should be someone with electrochemical experience. An instrument that operates well when it is set up is most likely to do so for many years. The main features of some of the commonly used potentiostats are given in Table 17.1.

TABLE 17.1 Manufacturing Data of Some Most Exploited Electroanalytical Instruments Manufacturer Radiometer Princeton Applied Research EcoChemie Cypress Systems Gamry Instruments HEKA Scribner Associates Incorporated ACM Instruments BioLogic Science Instruments Zahner PalmSens

Potentiostat Model

Maximum Compliance Voltage

Current Range

Web Site

PGZ100 (all-in-one) PARSTAT 2273

30 V 100 V

30 pA–1A 40 pA–2 A

www.radiometer-analytical.com www.princetonappliedresearch.com

MicroAutolab III 66-EI400 Bipotentiostat G 300 Potentiostat=ZRA PG 310 Multichannel microelectrode analyzer 900B Gill-AC-Bi-Stat VSP-Versatile Modular Potentiostat IM6 PalmSens

30 V 10 V 20 V 20 V 10 V

1 10 1 1 30

www.ecochemie.nl www.cypresshome.com www.gamry.com www.heka.com www.scribner.com

15 V 20 V

10 pA–500 mA 1 nA–400 mA

10 V 2 V

nA–250 mA pA–10 mA pA–300 mA nA–2A pA–100 mA

100 nA–3 A 1 nA–10 mA

www.acminstruments.com www.bio-logic.info www.zahner.de www.palmsens.com

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 384 14.7.2008 8:24pm Compositor Name: BMani

384

17.4.2 ELECTROCHEMICAL CELL

Handbook of Food Analysis Instruments AND

ELECTRODES

An electrochemical cell is considered to be a sample holder, in which the corresponding analyte is dissolved in an appropriate solvent, and placed thereafter in an ionic electrolyte, where usually three electrodes (working, reference, and counter) are situated. The cells come commercially in various designs, built mainly of glass, Teﬂon, or polyethylene material. The electrodes of the electrochemical system are usually all submerged in the electrochemical cell, whereas in some systems the reference electrode is placed in a separate compartment to avoid contamination, and it is connected to the cell via an electrolyte bridge. The main criteria for an electrode to be classiﬁed as a reference electrode are to provide a reversible half-reaction with Nernstian behavior, to have a constant potential over time, and to be easy to assemble and maintain [10]. Among the most widespread reference electrodes used for experiments performed in aqueous solutions are the calomel electrode, whose potential is determined by the reaction Hg2Cl2(s) þ 2e ! 2Hg(l) þ 2Cl, and the silver=silver chloride electrode (Ag=AgCl), whose potential is deﬁned by the reaction AgCl(s) þ e ! Ag(s) þ Cl. These electrodes are commercially available in a variety of sizes and shapes. With respect to the counter electrodes, Pt wire, graphite, or a thin piece of gold are the common variants in most of the voltammetric techniques. The task of the counter electrode is to preserve electroneutrality in the system, which is done by the occurrence of an electrochemical reaction (usually Hþ reduction, or water oxidation) counter to that taking place at the working electrode. A wide range of working electrodes are available. The metallic electrodes, such as Hg drops and Pt or Au disks, are among the most frequently used [11]. Mercury is very practical because of its highly negative overpotential for hydrogen ions reduction, and because of its constantly renewed surface [1,9,12]. Most electrochemical studies have been done using mercury as the working electrode, since it is equally adequate for studying the redox processes of inorganic and organic compounds as well as for elucidation of many surface phenomena, such as adsorption and crystallization. The main shortcoming with the mercury electrode is its low potential of oxidation (about þ0.1 to þ0.6 V depending on pH), because of which many compounds cannot be studied in the oxidation mode. Mercury can be used in different measuring modes such as hanging mercury drop electrode (HMDE) [13,14], dropping mercury electrode (DME) [15,16], static mercury drop electrode (SMDE) [17], or even as mercury ﬁlm electrodes [18,19]. Mercury is still being frequently used particularly for the analysis of metal ions, halides, and sulﬁde anions because of the possibility of amalgam formation in the former, and insoluble ﬁlms in the case of the latter. Nowadays, because of mercury toxicity, the use of this electrode material is being restricted and the search is on for new electrode material. In last few decades, along with metallic electrodes, various modiﬁcations of carbon electrodes are also in use. Carbon electrodes are extremely lipophilic, and hence appropriate for studying the electrochemical features of lipophilic organic compounds. Carbon electrodes can also be found in a large variety of forms such as graphite composite electrodes [20], glassy carbon electrodes [21], carbon-paste electrodes [22,23], screen-printed electrodes [24], diamond electrodes [25], and bismuth-coated carbon electrodes [26], among other examples. Gold [27,28], bismuth ﬁlms [29], biosensors [30], and immunosensors [31] are also part of the paraphernalia of electrode materials used with some advantage to solve analytical problems in food analysis.

17.5 SHORT REVIEW OF SOME COMMON VOLTAMMETRIC TECHNIQUES USED IN FOOD ANALYSIS Numerous dynamic methods in electroanalytical chemistry have been developed in recent decades. Among them, the most familiar in food analysis are the cyclic voltammetric and the so-called pulse voltammetric techniques. Here we give a brief summary of some of these techniques.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 385 14.7.2008 8:24pm Compositor Name: BMani

385

Electroanalytical Techniques and Instrumentation in Food Analysis

17.5.1 CYCLIC VOLTAMMETRY Cyclic voltammetry is one of the most exploited techniques in electrochemical studies. Its primary advantage comes from the fact that it gives insight into both the half-reactions taking place at the working electrode, providing at the same time information about the chemical or physical phenomena coupled to the studied electrochemical reaction [1,32,33]. Hence cyclic voltammetry is often considered as electrochemical spectroscopy [32]. Although its usage is relatively minimal in quantitative food analysis, it is important to elaborate on the principles of cyclic voltammetry, since every electroanalytical study almost inevitably commences with this technique. In cyclic voltammetry, starting from an initial potential Ei, a staircase (Figure 17.3a) potential sweep (or linear sweep in older potentiostats) is applied to the working electrode. After reaching a switching potential Ef, the sweep is reversed and the potential returns to its initial value. The main instrumental parameter in the cyclic voltammetry is the scan rate (v ¼ dE=t), since it controls the timescale of the voltammetric experiment. The useful scan rates range from 1 to 1000 mV=s, although scan rates of over 10 V=s are technically achievable. The instrumental output in cyclic voltammetric techniques is a current–potential curve, a cyclic voltammogram (Figure 17.3b). The main features of the cyclic voltammogram are the cathodic and anodic peak potentials, the cathodic and anodic peak currents, and the formal (or half-peak) potential. While the mid-peak potential (deﬁned simply as a median between the cathodic and the anodic peak potentials) provides mainly thermodynamics information, the magnitudes of the peak currents reveal the kinetics involved in the electrochemical reaction. The shape of the cyclic voltammogram gives information about the type of the electrode reaction, the number of electrons involved in the elementary step of electrochemical transformation, as well as about the additional phenomena coupled to the electrochemical reaction of interest, like those for coupled chemical reactions or adsorption and crystallization [1,32]. If the electron transfer process is much faster than the kinetics of the mass transport processes (diffusion), then the electrode reaction is electrochemically reversible. In this case, the peak separation DEp is deﬁned as follows: Ep ¼ jEp,c Ep,a j ¼ 2:303

RT nF

For example, in a simple reversible and diffusion-controlled electrochemical reaction, where one electron is exchanged in an elementary act, the peak separation should be about 59 mV (at 258C) [1].

Ep,a Ip,a

Potential step

Sampling period

Ef

lnorm

E

Ei Ip,c Ep,c ⫺0.2 ⫺0.16 ⫺0.12⫺0.08⫺0.04 0

(a)

Time

(b)

0.04 0.08 0.12 0.16

E/V

FIGURE 17.3 (a) Staircase potential ramp used in cyclic voltammetry and (b) a cyclic voltammogram simulated for one-electron reversible charge transfer: Ep,c, cathodic peak potential; Ep,a, anodic peak potential; Ei, initial potential; Ef, switching potential; Ip,c, cathodic peak current; Ip,a, anodic peak current.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 386 14.7.2008 8:24pm Compositor Name: BMani

386

Handbook of Food Analysis Instruments

Moreover, the peak potential separation should not vary by increasing the scan rate, while both cathodic and anodic peak currents should be a linear function of the square root of the scan rate. Every breach of these criteria means deﬂection of the electrochemical reversibility, caused either by the slow electron transfer (quasi-reversibility or irreversibility) or by additional involvement of the electroactive species in chemical reactions or adsorption phenomena. For an electrochemically reversible reaction, the concentration of the electroactive species is linked to the peak current Ip by the Randles–Sevcik expression [1], and at 258C it reads as follows: pﬃﬃﬃﬃﬃﬃ Ip ¼ 2:69 105 n3=2 Ac0 Dv where A is the active electrode surface area (m2) c0 is the initial concentration of the electroactive species in the solution (mol m3) D is its diffusion coefﬁcient (m2s2) n is the number of the electrons exchanged v is the scan rate (Vs1) This equation enables exploration of cyclic voltammetry for quantitative determination purposes.

17.5.2 PULSE VOLTAMMETRIC TECHNIQUES The invention of pulse voltammetric techniques was motivated by the fact that by changing the potential and measuring the current in a pulsed manner, a signiﬁcant discrimination of the charging (non-faradaic) current can be achieved [1,9]. Applying the potential difference between the working and the reference electrodes in an electrochemical cell is a precondition for initiating an electron exchange between the working electrode and the electroactive species in the cell. However, this change in the potential difference also causes charging and discharging of the electrical double layer at the electrode–electrolyte interface, which initiates a ﬂow of capacitive (charging) current too [1]. This current is undesirable for kinetic and analytical purposes, and efforts are being undertaken to minimize its contribution. The basis of all pulse techniques lies in the difference in the rate of the decay of the capacitive and faradaic currents following the potential steps. While the faradaic current decays with t1=2 for diffusion-controlled electrode reactions, for the same reactions, the capacitive current decays exponentially with time. Accordingly, by sampling the currents at the end of the applied potential pulses, one gets negligible capacitive currents, yet signiﬁcant faradaic currents [34]. In this way, the sensitivity of the voltammetric method will be signiﬁcantly increased, and the measured current will refer almost exclusively to the faradaic reaction of interest. In novel electrochemical instruments, one meets various pulse voltammetric techniques, which differ in the pulse-wave form and the way by which the current is sampled. The most important parameters of all pulse voltammetric techniques are as follows: (1) pulse amplitude, which is the height of the potential pulse, (2) pulse width, which is the duration of the potential pulse, and (3) sampling period, deﬁned as a time at the end of the potential pulse in which current is measured. We refer to some of the most important pulse techniques in this chapter. 17.5.2.1

Normal Pulse Voltammetry

In normal pulse voltammetry (NPV), a series of potential pulses with constant width and permanent increased amplitude are applied, with the potential returning to the initial value after each pulse [9]. The current is measured in a certain period at the end of each pulse, which is enough to diminish signiﬁcantly the charging current component (Figure 17.4a). The duration of the pulses ranges from 1 to 200 ms, while the interval between the pulses is several seconds. The instrumental output

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 387 14.7.2008 8:24pm Compositor Name: BMani

387

Electroanalytical Techniques and Instrumentation in Food Analysis 1.0 0.9 0.8 I (µA)

E

Step potential

Sampling Pulse period width

0.7 0.6 0.5 0.4 0.15

Time

(a)

FIGURE 17.4

0.1

0.05

(b)

0 –0.05 –0.1 –0.15 –0.2 E (V)

(a) Potential form and (b) simulated voltammogram in NPV.

in this technique is an I–E curve (normal pulse voltammogram), with sigmoidal shape (Figure 17.4b) as it is commonly obtained in classical (normal) polarographic techniques [1]. Hence this technique is called ‘‘normal’’ pulse voltammetry. 17.5.2.2

Differential Pulse Voltammetry

The potential form in differential pulse voltammetry (DPV) consists of small pulses of constant amplitude (10–100 mV) superimposed on a staircase-wave form. The current in this technique is measured twice in each pulse period, ﬁrst at the beginning of the applied pulse and second at the ending of the same pulse (Figure 17.5a) [34]. The measured current in the instrumental output, referred to as differential pulse voltammogram (Figure 17.5b), is actually the difference between the currents measured for each single pulse. The current measured thus enables one to obtain much higher sensitivity of DPV with respect to NPV.

Sampling period 2

Step potential Pulse amplitude

Inorm

Sampling period 1 Pulse width

E

(a)

FIGURE 17.5

Time

⫺0.2⫺0.16⫺0.12⫺0.08⫺0.04 0

(b)

E (V)

(a) Potential form and (b) resulting simulated voltammogram in DPV.

0.04 0.08 0.12 0.16

0.2

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 388 14.7.2008 8:24pm Compositor Name: BMani

388

17.5.2.3

Handbook of Food Analysis Instruments

Square-Wave Voltammetry

Square-wave voltammetry (SWV) is the most advanced and the most sophisticated technique in the family of pulse voltammetric techniques [10,35,36]. The potential form in SWV consists of symmetrical square-wave pulses with constant amplitude ESW, which are superimposed on a staircase-wave form (Figure 17.6a). The potential in SWV changes for a constant potential step dE. The current in this technique is measured twice at the end of each half cycle. The currents measured at the end of oxidation half cycles give the oxidative (forward) current component, while the currents measured at reduction half cycles give the reduction (backward) current component (Figure 17.6b). The net current in SWV is obtained as a subtraction between the forward and the backward currents. However, since the reductive currents (by convention) have a negative sign, the net current in SWV is actually a sum of the absolute values of both the current components (Figure 17.6b). This method of measurement makes SWV the most sensitive electroanalytical technique. The net peak current in SWV, as in other pulse voltammetric techniques, is proportional to the analyte concentration, resulting often in detection limits in sub-nanomolar ranges. Besides, SWV provides an insight into both the halfelectrode reactions, thus having a distinct advantage over cyclic voltammetry for studying the mechanisms of electrochemical reactions. SWV is a very fast technique, providing insight into the kinetics of fast electron transfer reactions, and into the kinetics of rapid chemical reactions coupled to the electroactive species. Although the theory of SWV is still developing, many excellent theoretical papers on SWV have appeared in last 30 years, providing criteria for recognition of many complex electrode mechanisms, and providing methods for measuring the kinetics and thermodynamics of various processes encountered in the investigated electrochemical systems [1,9,35]. During the last 25 years, this has led to SWV becoming one of the most explored voltammetric techniques for both quantitative applications and mechanistic studies, as well as for the determination of kinetics and thermodynamics in various electrochemical systems.

17.6 PRECONCENTRATION VOLTAMMETRIC TECHNIQUES The goal of every analytical technique is to provide lower detection limits, good sensitivity, and good selectivity. In all voltammetric techniques, the detection limits depend mainly on the nature of lnet dE lf

Esw

t lnorm

E

lb ⫺0.2⫺0.16⫺0.12⫺0.08⫺0.04 0 0.04 0.08 0.12 0.16 0.2

(a)

Time

(b)

E (V)

FIGURE 17.6 (a) Potential form in SWV: Esw, potential amplitude; dE, potential step; t, duration of a single pulse. The current is sampled twice in each pulse, in the time period between two arrows at the inset; (b) resulting simulated voltammogram in square-wave voltammetry: If, forward current; Ib, backward current; Inet, net current.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 389 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

389

mass transport phenomena, as well as on the nature and features of the electrode and the electroactive species. If mass transport occurs only by diffusion, then the detection limits usually range around micromolar concentrations [1]. However, many organic compounds exhibit surfaceactive properties, which are manifested by their adsorption from the solution to the electrode surface. Many other compounds are capable of dissolving or reacting with the electrode material (usually with the mercury electrode), forming an amalgam or a sparingly soluble mercury complex, respectively. These phenomena are bases for developing preconcentration techniques, such as adsorptive stripping voltammetry (AdSV), the anodic (ASV) and the cathodic (CSV) stripping voltammetric methods, respectively. The mercury electrode is one of the most used working electrode in stripping voltammetric techniques, while examples exist where other metallic modiﬁed electrodes have also been explored [12]. Stripping voltammetric techniques are among the most sensitive for trace analysis. Examples are known where detection limits below picomolar range have been reported by using some of the stripping voltammetric modes [37–39]. To obtain reproducible results, the following important conditions must be taken into account in all stripping techniques: constancy of the electrode surface, invariable rate of stirring, and constant deposition time. We give here a brief summary of the most frequently used preconcentration voltammetric techniques.

17.6.1 ADSORPTIVE STRIPPING VOLTAMMETRY Adsorptive stripping voltammetry (AdSV) is one of the most sensitive voltammetric techniques, which has been successfully applied for the determination of traces of various compounds at subnanomolar levels [1,12,39]. AdSV is based on previous accumulation (by adsorption) of the compound on the working electrode at some adequate constant potential, and consecutive electrolysis (oxidation or reduction) of the adsorbed material. It is well known that many organic and inorganic anions have a tendency of adsorbing at the mercury electrode [39], which is a prerequisite step for their accumulation. Besides the mercury electrode, which is commonly used as a working electrode in adsorptive stripping techniques [39], another type of accumulation can be achieved with a chemical interaction between the analyte and the modiﬁed electrode surfaces of another (nonmercury) type of working electrode [1,12]. Along with the electrode material and the accumulation time, the choice of the accumulation potential is a very important factor that leads to the substantial increase of sensitivity for a particular AdSV. As the working voltammetric technique in AdSV, the DPV and SWV are commonly used [12]. As with other stripping techniques, attention must be paid to sample preparation and avoidance of possible contaminations, since these techniques are extremely sensitive to various impurities.

17.6.2 ANODIC STRIPPING VOLTAMMETRY Anodic stripping voltammetry (ASV) ﬁnds enormous practical use for the determination of trace of various metals [1,12,40]. It is well known that many metals can form amalgams with mercury (i.e., to dissolve in it). By applying sufﬁcient negative potentials for a prolonged electrolysis time, the reduction of the present metal ions from the solution takes place, and subsequent concentration of the metals into the mercury electrode is achieved. After amalgamation, the metals are thereafter stripped off (oxidized) from the mercury electrode, when running the potential in positive (anodic) direction. Their dissolution process is depicted in voltammetric peaks, with positions at the potential scale depending on the nature of the metals. The resulting peak currents of the voltammetric responses are proportional to the metal ion concentrations in the solutions. With ASV, it is possible to determine simultaneously several metal ions (up to six), if their standard redox potential differs for at least 200 mV. This technique also achieves low detection limits, frequently in the nanomolar

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 390 14.7.2008 8:24pm Compositor Name: BMani

390

Handbook of Food Analysis Instruments

range. The main shortcoming of this technique is the possibility of formation of intermetallic compounds, like ZnCu for example, which can lead to misinterpretation of voltammetric responses. Overlapping stripping peaks caused by similarity in oxidation potential or the presence of surfaceactive organic compounds that adsorb on the mercury surface and inhibit metal deposition can also introduce difﬁculties in the implementation of this technique. However, such problems may be avoided by adjusting the deposition potential.

17.6.3 CATHODIC STRIPPING VOLTAMMETRY Cathodic stripping voltammetry (CSV) is a technique of choice when the investigated compounds are capable of reacting with the mercury ions of the mercury electrode, forming sparingly soluble salts at the surface of the working electrode [1,3,12,41–43]. Initially, by applying a positive potential for a certain time, an insoluble ﬁlm, created by chemical reaction between the investigated compound and the mercury ions at the surface of the working (mercury) electrode, will be formed. In the second step, this ﬁlm is stripped off by running the potential in the negative (cathodic) direction. This method has been explored for quantitative determination of many inorganic anions, such as halides, sulﬁdes, selenides, as well as for many biologically active compounds (amino acids, proteins, nucleic acids, etc.) containing groups that can react with the mercury ions [3,12,41–43]. Generally, the compounds containing Cl, F, Br, SH, and SeH groups are potentially cathodicstripping active species [12]. The main advantage of this technique is its employment for the determination of electrochemically inactive compounds, i.e., compounds that do not show electrochemical activity in the given potential range [41].

17.7 ELECTROCHEMICAL TECHNIQUES IN FOOD ANALYSIS A large number of examples of electrochemical methods developed and applied to a vast number of compounds relevant to food industry and food quality can be found in the literature. However, here we select only papers where electrochemical techniques have been directly applied to food analysis. Some valuable review papers can be found on the use of stripping voltammetry [3], biosensors [44–46], and speciation of arsenic [47] in food analysis.

17.7.1 FOOD COLORANTS Electrochemistry of azo compounds was ﬁrst reported by Florence in 1974 [48]. The ﬁrst description of the use of electrochemical techniques for food colorant analysis was published in 1979 by Fogg and Yoo [49]. This work describes the methodology for the determination of tartrazine, amaranth, green S, and sunset yellow FCF analysis in orangeade, limeade, and black currant health drinks by differential pulse polarography (DPP). Strategies for simultaneous determination of colorants in fruit juices can also be found [50]. Several others works dealing with the electrochemical determination of colorants and ﬂavors can be found in the literature and some of them are listed in Table 17.2 [51–57].

17.7.2 METAL CONTAMINANTS Although trace amounts of metal ions in food products play an essential role in metabolic processes, they can easily surpass the toxicity limits and therefore it is of extreme importance to have reliable and fast methods to assure the observation of the established health and risk regulations [58]. Voltammetry and in particular methods that involve stripping stages are well ﬁtted for metal analysis in food as the numerous references found in the literature demonstrate (Table 17.3) [59–78]. Two main procedures are adopted in these analysis: ASV for those metals that can easily form amalgams with mercury, namely Pb, Zn, Cu, Cd, Sb, Tl, and Hg (in gold or carbon electrodes), and cathodic adsorptive stripping voltammetry (CAdSV) for other metal ions such as Al, Cr, Co, Fe, Ni, Ti, or U.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 391 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

391

TABLE 17.2 Overview of the Methods and Electrodes in Voltammetric Determination of Some Colorants in Drinks Colorant

Method

Electrode

Sample

Reference

Allura red

DPP DPP DPP AdSV DPP CSV DPP CSV DPP CSV DPV AdSV DPP DPP CSV AdSV DPP ASSWV DPP DPP AdSV CSV ASV DPP

Hg Hg Hg HMDE Hg HMDE Hg HMDE Hg HMDE NGCE HMDE Hg Hg HMDE HMDE Hg Hg Hg Hg HMDE HMDE HMDE Hg

Sweets and soft drinks Soft drinks Soft drinks Soft drinks Soft drinks Soft drinks Sweets and soft drinks Soft drinks Soft drinks Sweets and juices Soft drinks and liquors Soft drinks Sweets and soft drinks Soft drinks Soft drinks Soft drinks Soft drinks Soft drinks Soft drinks Soft drinks Soft drinks Sweets and soft drinks Soft drinks Soft drinks

[50] [52] [47] [48] [52] [55] [50] [55] [47] [49] [19] [48] [50] [52] [55] [48] [53] [54] [47] [47] [48] [49] [51] [53]

Amaranth Azorubin Brilliant blue FCF Carmoisine Erythrosine Green-S Indigo-carmine Patent Blue V Ponceau 4R

Quinoline yellow Sunset yellow

Sunset yellow FCF Tartrazine

Note: DPP, differential pulse polarography; AdSV, adsorptive stripping voltammetry; HMDE, hanging mercury drop electrode; CSV, cathodic stripping voltammetry; DPV, differential pulse voltammetry; NGCE, nanochannel glassy carbon electrode; ASSWV, anodic stripping square-wave voltammetry; ASV, anodic stripping voltammetry.

17.7.3 PESTICIDES

AND

HERBICIDES

Pesticides and herbicides represent a major concern in food quality assurance. Voltammetry can be a useful tool for their analysis since most of the organic compounds used for pest and herb control have electroactive groups. Some recent reviews can give a good highlight of electrochemical methods for determination of herbicides and pesticides [79–81]. Hance was the pioneer in the use of electroanalytical techniques for pesticide residue analysis [82]. In his work, Hance describes the behavior of 38 herbicides using derivative polarography, showing that 28 of them were electroactive. The well-known inhibition effect of pesticides over enzymes is also explored to build electrochemical sensors for pesticide analysis (see Section 17.7.4). Pesticides have different electroactive groups and some of them require derivatization before determination. Triazines present one or more reduction peaks corresponding to the reduction of mono- and biprotonated forms. For s-triazine, reduction occurs at the –C¼N– bond of the heterocyclic ring [83]. The mechanism of reduction of asymmetrical triazines depends on the structure of the molecule. As an example, guthion is reduced in the –N¼N– bond of the heterocyclic ring [84], and reduction of metamitron involves the functions C¼N and NNH2 that are present in the molecule [85]. Mercury [86] electrodes are commonly used for the electrochemical determination of triazines; however, more recently, biosensors have also been employed in the determination of triazines in food [87].

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 392 14.7.2008 8:24pm Compositor Name: BMani

392

Handbook of Food Analysis Instruments

TABLE 17.3 Overview of the Methods and Electrodes Used in Voltammetric Determination of Some Metals in Food Metal

Sample Treatment

As(III) Se(IV)

Acid digestion Digestion

Se(VI)

Digestion

Hg(II) Cd(II) Cu(II) Ni(II) Cd(II) Pb(II) Fe(II) Sn(IV) Pb(II) Mo(VI) U(VI) Sn(IV) Pb(II) Cu(II) Pb(II) Cu(II) Cd(II) Cu(II) Pb(II) Pb(II) Cu(II) As(V) As(III) As(V) As(III)

— — Digestion — — — Acid digestion — — UV irradiation, dry-ashing — — — — Solubilization Solubilization Solubilization — —

MWAAD MWAAD Acid digestion Acid digestion

Sample Biological material Pig kidney (BCR No. 186 certiﬁed selenium content) Pig kidney (BCR No. 186 certiﬁed selenium content) Alcoholic drinks Meat and egg powder Cow’s liver tissue Canned vegetables Beer Beer SRM-1547 from peach leaves Juice and canned fruit Juice and canned fruit Biological material Sugar Canned fruit drinks and juices Canned fruit drinks and juices Whisky Edible oils Edible oils Edible oils Beer Milk Wine Wine Tobacco leaves, nettles Tobacco leaves, nettles Zinc oxide as food additive used in feed Zinc oxide as food additive used in feed

Method

Electrode

Reference

ASV ASV

Au Au

[57] [58]

ASV

Au

[58]

ASV CAdsSV CAdsSV CAdsSV SSWASV SSWASV OSWV LSASV LSASV CSV CSV ASV ASV ASV ASV ASV ASV ASV LSV ASV ASV CSV CSV CSV CSV

Au HMDE HMDE HMDE MTFE MTFE SGE MTFE MTFE MTFE GCE CCE CCE mPt MTFE MTFE MTFE GCE SMFE MFmE MFmE HMDE HMDE HMDE HMDE

[59] [60] [61] [62] [63] [63] [64] [65] [65] [66] [67] [68] [68] [69] [70] [70] [70] [71] [72] [73] [73] [74] [74] [75] [75]

Note: MWAAD—micro-wave activation and digestion, ASV—anodic stripping voltammetry, CAdSV—catalytic adsorptive voltammetry, SSWASV—surface square-wave anodic stripping voltammetry, OSWV—osteryoung square-wave voltammetry, LSASV—linear scan anodic stripping voltammetry, CSV—cathodic stripping voltammetry, LSV—linear scan voltammetry, HMDE—hanging mercury drop electrode, MTFE—mercury thin-ﬁlm electrode, SGE— spectroscopically pure graphite electrode, GCE—glassy-carbon electrode, CCE—carbon-cloth electrode, SMFE— sintered metal ﬁbres electrode, MFmE—mercury ﬁlm micro electrode.

The redox process of organophosphates presents a well-deﬁned process at neutral, acid or, basic media, which is attributed to the reduction of the C¼C bond or to the reduction of chloride or nitro groups presented in the pesticide structure [88]. The use of gold microelectrodes [89], mercury electrodes [90], and biosensors has been reported for the analysis of organophosphate pesticides [30]. Metabolites of organophosphate pesticides (e.g., p-nitrophenol) can also be used in electrochemical methodologies for the indirect analysis of organophosphate pesticides [91]. The reduction mechanism of pesticides with nitropesticides is associated with the reduction of the nitro group with consequent formation of hydroxylamines or further to the corresponding amines and is currently a well-deﬁned process [92]. As with the triazines, some of these pesticides

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 393 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

393

containing nitro groups can be adsorbed onto the surface of mercury electrodes [93] or as an alternative can be analyzed by using biosensors [94]. Carbamates represent a large number of insecticides and herbicides. Some of the carbamates require a derivatization step before their electrochemical determination or analysis of their residues; however, the use of biosensors usually solves this difﬁculty [30]. Adsorption at gold electrode is also used for the direct analysis of disulﬁram in peas [95]. Another important class of pesticides is the bipyridium pesticides also known as ‘‘viologens’’ Not all the pesticides in this family can be reduced at an electrode surface. From mechanistic studies, it seems that an important condition is the coplanarity of the two heterocyclic nuclei with the electrode surface [96]. Analysis of paraquat in foodstuff using gold microelectrodes has been successfully carried out [27,97]. Organochloride pesticides are part of another important family of pest control chemicals. Several electrochemical studies were consistent in their conclusion of a reaction mechanism involving the removal of one atom of chlorine [17]. The adsorption of organochloride pesticides at the mercury electrode and other metallic electrodes can impair the sensitivity and reproducibility of measurements. The addition of surfactants or the use of micellar systems seems to improve the signal-to-noise ratio [98]. This methodology was used to demonstrate the presence of several organochloride pesticides in apples. Pyrethroids are insecticides originally extracted from natural sources. In the case of deltamethrin, the compound exhibits a single well-deﬁned peak because of the reduction of the –C¼C– moiety at the mercury electrode. The analysis of deltamethrin in vegetables and cereals has been reported [99]. Analysis of other classes of pesticides such as xylylalanine, dicarboximide, azole, anilide, and strobin has also been reported in the literature [30]. Table 17.4 presents a list of works reporting the determination of a large number of pesticides, indicating a large diversity of matrixes [100–110] and electrode materials. Although sulfonylureas TABLE 17.4 Overview of the Methods and Electrodes Used in Voltammetric Determination of Some Pesticides in Food Pesticide Fenhexamid Myclobutanil Propiconazol Paraquat Paraquat Aldicarb Aldicarb Carbaryl Carbaryl Carbaryl Carbaryl Carbaryl Carbaryl Carbaryl Carbaryl Carbaryl Carbaryl Carbofuran Carbofuran Carbofuran Carbofuran

Anilide Azole Azole Bipyridinium Bipyridinium Carbamates Carbamates Carbamate Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates

Electrode

Matrix

Reference

Biosensor Biosensor Biosensor Au UME Au UME Biosensor Biosensor Biosensor Vitreous carbon Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor Vitreous carbon Biosensor Biosensor

Fruits Fruits Fruits Fruits, potatoes, and sugarcane Fruit juice Vegetables Fruits and vegetables Milk Vegetables Kiwi Vegetables Egg Meat Milk Honey Fruits and vegetables Fruits Fruit juice Vegetables Vegetables Fruits, vegetables, and dairy products

[30] [30] [30] [27] [97] [103] [104] [100] [102] [105] [103] [94] [94] [94] [94] [104] [30] [101] [102] [103] [106] (continued )

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 394 14.7.2008 8:24pm Compositor Name: BMani

394

Handbook of Food Analysis Instruments

TABLE 17.4 (continued) Overview of the Methods and Electrodes Used in Voltammetric Determination of Some Pesticides in Food Pesticide Carbofuran Disulﬁram Disulﬁram Ethiofencarb Methomyl Methomyl Pirimicarb Promecarb Propoxur Propoxur Propoxur Thiram Methiocarb Iprodion Dinoseb Methyl parathion Methyl parathion Methyl parathion Methyl parathion Dieldrin Endosulfan sulphate Heptachlor Sulfan a-Endosulfan b-Endosulfan Clorfenvinphos Crotoxyphos Dicrotophos Paraoxon Paraoxon Paraoxon Coumaphos Chloropyrifos-methyl Paraoxon Dichlorvos Dichlorvos Deltamethrin Esfenvalerat Azoxystrobin Desmetryne Simazine Simazine Simazine Simazine Piperonylbutoxid Metalaxyl

Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Carbamates Dicarboximide Nitropesticides Nitropesticides Nitropesticides Nitropesticides Nitropesticides Organochloride Organochloride Organochloride Organochloride Organochloride Organochloride Organophosphate Organophosphate Organophosphate Organophosphate Organophosphate Organophosphate Organophosphate Organophosphate Organophosphate Organophosphate Organophosphate Pyrethroid Pyrethroid Strobin Triazines Triazines Triazines Triazines Triazines Unclassiﬁed Xylylalanine

Electrode

Matrix

Reference

Biosensor Au UME Modiﬁed graphite Biosensor Biosensor Biosensor Biosensor Vitreous carbon Vitreous carbon Biosensor Biosensor Modiﬁed graphite Biosensor Biosensor Mercury ﬁlm Biosensor Biosensor Biosensor Biosensor HMDE HMDE HMDE HMDE HMDE HMDE DME DME DME Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor HMDE Biosensor Biosensor HMDE Biosensor Biosensor Biosensor Biosensor Biosensor Biosensor

Fruits and vegetables Peas Strawberry Fruits Vegetables Fruits and vegetables Fruits Vegetables Vegetables Vegetables Fruits and vegetables Strawberry Fruits Fruits Fruit juice Egg Meat Milk Honey Apples Apples Apples Apples Apples Apples Cereals Cereals Cereals Baby food Fruit juice Kiwi Grape juice Grape juice Milk Wheat Wheat Vegetables and cereals Fruits Fruits Fruit juice Meat Vegetables Milk Fruit juice Fruits Fruits

[104] [95] [107] [30] [103] [104] [30] [102] [102] [103] [104] [107] [30] [30] [93] [94] [94] [94] [94] [98] [98] [98] [98] [98] [98] [90] [90] [90] [30] [101] [105] [108] [108] [100] [109] [110] [99] [30] [30] [86] [87] [87] [87] [87] [30] [30]

Note: UME—ultra micro electrode, HMDE—hanging mercury drop electrode, DME—dropping mercury electrode.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 395 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

395

and phenylureas constitute an important group of pest control chemicals with redox activity and albeit there are some works reported on their electrochemical behavior [111], it was not possible to ﬁnd any paper describing their determination in natural or processed food products. One drawback of biosensor strategy that uses enzyme inhibition by pesticides is the fact that it is not speciﬁc and cannot be used to identify the pesticide or herbicide per se but solely to identify their presence [112]. The advantage of using the biosensor is that it allows a fast and cheap method for pesticide detection, and hence can be used for screening purposes.

17.7.4 BIOSENSORS A general deﬁnition of biosensors includes all devices that incorporate a biological or biologically derived sensing element, which is usually associated with a transducer. The main advantage of the biosensors is to take beneﬁt of thousands of years of evolution that made available biological macromolecules that react speciﬁcally with the target analyte. The primary aim of a biosensor is to produce a signal that can be related to a speciﬁc analyte. Biosensors have proven to be a good analytical tool, well-ﬁtted to cope with the challenges that food analysis gives rise to, and several books and review papers are available on biosensors [113–116]. As biosensors use different transducing modes [117] they can surpass the demanding requirements found when dealing with difﬁcult matrixes and tough standards currently found in food analysis. Two main groups of biosensors can be identiﬁed: (1) those dealing with direct measurement of the analyte or a product of the enzymatic reaction and (2) those dealing with the indirect quantiﬁcation of the analyte evaluating the decrease in the electrochemical signal of the biosensor caused by the poisoning of the sensor element when the analyte is present (e.g., contaminant [29], insecticide [28]). Amperometric biosensors rely on the current intensity generated at the working electrode (either its increase in case of direct measurements, or its decrease for indirect measurements) after the analytes interact with an enzyme. According to the electrochemical process used for transduction, the biosensors are usually classiﬁed into four main groups: oxygen electrodes (when the oxygen consumption during the enzymatic reaction is assessed) [118], peroxide electrode (when the oxygen peroxide formed during the enzymatic reaction is monitored) [119], redox mediators, and NADH electrochemical sensors (electrocatalytic detection based on the recycling of redox mediators, most of them involving the use of NADH in the recycling step) [120,121].

17.7.5 OTHER APPLICATIONS Electrochemical methods also have a signiﬁcant contribution to the determination of other analytes with prime importance for food analysis such as antibiotics (ceftazidime in milk by differential pulse CSV [122]), possible human carcinogens (butylated hydroxyanisole in potato chips [20]), ﬂavors or ﬂavor enhancers (vanillin in vanilla sugar, chocolate biscuits, and chocolate [18], inosinic acid in dehydrated soups [123], diacetyl in beer [124], and a-diketones in brandy, vinegar, wine, and butter samples [125]), antidepressants in herbal medicinal products [126], tert-butylhydroquinone in popcorn [127], hormones (progesterone in milk [128]), and polychlorinated biphenyls (PCBs) in meat and milk [129].

17.7.6 HYPHENATED TECHNIQUES A large percentage of food analysis relies on the measurement of a vast number of chemicals and parameters that make use of chromatographic techniques. Electrochemistry can also play an important role in the detection of the analytes in chromatographic analysis either in high-performance liquid chromatography or in electrophoresis. Electrochemical detectors present detection limits (5–50 ng=mL) very similar to those of ﬂuorescence detectors (5–500 ng=mL) and much better than ultraviolet detectors (1–20 mg=mL) [130]. Examples of the use of electrochemical detectors are the

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 396 14.7.2008 8:24pm Compositor Name: BMani

396

Handbook of Food Analysis Instruments

determination of amino acids and sugars in rice wine [131], heterocyclic aromatic amines in beef extract [132], nitrite [133] and nitrate [125] in meat products, organic acids in red wine by ionexclusion chromatography [134], taurine in plant extracts, milk powder, and health beverages, ﬂavonoids and vitamin E isomers in seeds and nuts [5], isoﬂavones in soybean food [135], vitamin A in cereals [136], alkaloids in foods [117], heterocyclic aromatic amines in food ﬂavors [118], heterocyclic aromatic amines in commercial beef extract [137], mutagenic or carcinogenic compounds in charcoal-grilled meat [138], nitrofuran derivatives in milk [139], ascorbic acid in athletes food, nutritional supplement, infant milk [140], aldehydes in honey, coffee, refreshments, sherry, port, dry fruits, and breakfast cereals [141], nitrite and nitrate in ﬁsh and meat products [142], and coenzymes in ﬁsh, meat, and rye ﬂour [143].

17.8 NEW TRENDS The requirements for highly precise and fast quantiﬁcation of many important compounds by food analysis lead to the rapid development of various electrochemical sensors. Indeed, the future of electrochemical techniques applied in food analysis is to enable simultaneous analysis for several compounds. Demanding electrochemical applications require high-performance instrumentation coupled with high-throughput capabilities. It is recognized that electrochemical sensor development has equipment demands that are sometimes hard to satisfy with conventional potentiostats. To address these applications, various multichannel instruments are starting to appear in the electrochemical instrumentation market. Each channel of these potentiostats can be controlled independently or used in conjunction with other channels (potentiostats) to perform the same experiment on different electrodes. In addition, up to 20 channels can be used with one reference and one counter electrode (the so-called N’stat mode). Each potentiostat is capable of at least 20 V scan ranges, 400 mA current capabilities with 20 V compliance, and a timebase of 200 ms. This is a critical feature for multielectrode potentiostatic and potentiodynamic pitting experiments. Certainly, the successful employment of multichannel potentiostats in food analysis is closely tied with the development of highly speciﬁc sensors. Only in this way can electrochemical techniques compete with other techniques currently used intensively in food analysis, for example, chromatographic techniques. Although important, there has been no research effort in recent years on the application of novel electrode materials and electrochemical techniques into the electrochemical detectors in chromatographic analysis, and this can be a ﬁeld of fruitful research in the near future.

REFERENCES 1. Bard, A.J. and Faulkner, L., Electrochemical Methods, Wiley, New York, 2001. 2. Bagotsky, V.S., Fundamentals of Electrochemistry, 2nd ed., Wiley, New York, 2005. 3. Brainina, Kh.Z., Malakhova, N.A., and Stojko, N.Y., Stripping voltammetry in environmental and food analysis, Fresenius J. Anal. Chem., 368, 307, 2000. 4. Antropov, L.I. and Beknazarov, A., Theoretical Electrochemistry, University Press of the Paciﬁc, Moscow, 2001. 5. Codognoto, L. et al., Electroanalytical and chromatographic determination of pentachlorophenol and related molecules in a contaminated soil: A real case example, Microchem. J., 77, 177, 2004. 6. Tsai, Y.C. et al., Microwave-enhanced anodic stripping detection of lead in a river sediment sample. A mercury-free procedure employing a boron-doped diamond electrode, Electroanalysis, 13, 831, 2001. 7. Delgado-Zamarreño, M.M. et al., Pressurized liquid extraction prior to liquid chromatography with electrochemical detection for the analysis of vitamin E isomers in seeds and nuts, J. Chromatogr. A, 1056, 249, 2004. 8. Xu, Q. et al., Application of a single electrode, modiﬁed with polydiphenylamine and dodecyl sulfate, for the simultaneous amperometric determination of electro-inactive anions and cations in ion chromatography, J. Chromatogr. A, 997, 65, 2003.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 397 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

397

9. Osteryoung, J.G. and Schreiner, M.M., Recent advances in pulse voltammetry, CRC Crit. Rev. Anal. Chem., 19, S1, 1988. 10. Kahlert, H., Reference electrodes, in Electroanalytical Methods, Scholz, F. (Ed.), Springer, Berlin=Heidelberg=New York, 2001, chap. III. 11. Komorsky-Lovric, S., Working electrodes, in Electroanalytical Methods, Scholz, F. (Ed.), Springer, Berlin=Heidelberg=New York, 2001, chap. III. 12. Wang, J., Analytical Electrochemistry, 3rd ed., Wiley-VCH, New York, 2006. 13. Castrillejo, Y. et al., Determination of food additive azo dyes at an HMDE with adsorptive stripping voltammetry, Electroanalysis, 2, 553, 1990. 14. Qiong, L. et al., Determination of trace aluminum in foods by stripping voltammetry, Food Chem., 97, 176, 2006. 15. Inam, R. and Toprak, C., Polarographic determination of some toxic trace elements in ﬁsh muscles, Fresenius Environ. Bull., 14, 489, 2005. 16. Obendorf, D. and Reichart E., Determination of hesperidin by cathodic stripping voltammetry in orange juice and helopyrin, a phytopharmaceutical preparation, Electroanalysis, 7, 1075, 1995. 17. Carrai, P., Nucci, L., and Pergala, F., Polarographic behavior of alachlor application to analytical determination, Anal. Lett., 25, 163, 1992. 18. Babkina, S.S. and Ulakhovich, N.A., Amperometric biosensor based on denatured DNA for the study of heavy metals complexing with DNA and their determination in biological, water and food samples, Bioelectrochemistry, 63, 261, 2004. 19. Lo Coco, F. et al., Determination of lead (II) and cadmium (II) in hard and soft wheat by derivative potentiometric stripping analysis, Anal. Chim. Acta, 409, 93, 2000. 20. Luque, M. et al., Supported liquid membranes for the determination of vanillin in food samples with amperometric detection, Anal. Chim. Acta, 410, 127, 2000. 21. Desimoni, E., Brunetti, B., and Cosio, M.S., Determination of patent blue V (E131) at a Naﬁon-modiﬁed glassy carbon electrode, Electroanalysis, 18, 231, 2006. 22. Jayasri, D. and Narayanan, S.S., Manganese(II) hexacyanoferrate based renewable amperometric sensor for the determination of butylated hydroxyanisole in food products, Food Chem., 101, 607, 2007. 23. Li, Y.H. et al., Determination of trace tin by anodic stripping voltammetry at a carbon paste electrode, Electroanalysis, 18, 976, 2006. 24. Miserere, S. et al., Biocompatible carbon-based screen-printed electrodes for the electrochemical detection of nitric oxide, Electrochem. Commn., 8, 238, 2006. 25. Bouamrane, F. et al., Electrochemical study of diamond thin ﬁlms in neutral and basic solutions of nitrate, J. Electroanal. Chem., 405, 95, 1996. 26. Wang, J. et al., Bismuth-coated carbon electrodes for anodic stripping voltammetry, Anal. Chem., 72, 3218, 2000. 27. De Souza, D., Machado, S.A.S., and Pires, R.C., Multiple square wave voltammetry for analytical determination of paraquat in natural water, food, and beverages using microelectrodes, Talanta, 69, 1200, 2006. 28. Radke, S.A. and Alocilja, E.C., A high density microelectrode array biosensor for detection of E-coli O157: H7, Biosens. Bioelectron., 20, 1662, 2005. 29. Sánchez Arribas, A. et al., Voltammetric detection of the herbicide metamitron at a bismuth ﬁlm electrode in nondeaerated solution, Electroanalysis, 18, 2331, 2007. 30. Schulze, H. et al., Development, validation, and application of an acetylcholinesterase-biosensor test for the direct detection of insecticide residues in infant food, Biosens. Bioelectron., 17, 1095, 2002. 31. Adanyi, N. et al., Development of new immunosensors for determination of contaminants in food, Curr. App. Phys., 6, 279, 2006. 32. Heinze, J., Cyclic voltammetry—‘‘Electrochemical Spectroscopy,’’ Angew. Chem., 23, 831, 1984. 33. Kounaves, S.P., Voltammetric techniques, in Handbook of Instrumental Techniques for Analytical Chemistry, Settle, F. (Ed.), Prentice Hall PTR, 1997, chap. 37. 34. Stojek, Z., Pulse voltammetry, in Electroanalytical Methods, Scholz, F. (Ed.), Springer, Berlin=Heidelberg=New York, 2001, chap. II. 35. Lovric, M., Square-wave voltammetry, in Electroanalytical Methods, Scholz, F. (Ed.), Springer, Berlin=Heidelberg=New York, 2001, chap. II.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 398 14.7.2008 8:24pm Compositor Name: BMani

398

Handbook of Food Analysis Instruments

36. Osteryoung, J.G. and O’Dea, J.J., Square-wave voltammetry, in Electroanalytical Chemistry, Bard, A.J. (Ed.), Vol. 14, Marcel Dekker, New York, 1986. 37. Xu, G. et al., Adsorptive square-wave stripping voltammetry for determination of azobenzene at trace levels, Anal. Chem., 66, 808, 1994. 38. Mirceski, V. et al., Square-wave voltammetry of 5-ﬂuorouracil, J. Electroanal. Chem., 490, 37, 2000. 39. Kalvoda, R. and Kopanica, M., Adsorptive stripping voltammetry in trace analysis, Pure Appl. Chem., 61, 97, 1989. 40. Kounaves, S.P. et al., Square-wave anodic-stripping voltammetry at the mercury ﬁlm electrode-theoretical treatment, Anal. Chem., 59, 386, 1987. 41. Gulaboski, R. et al., Square-wave voltammetry of cathodic stripping reactions-diagnostic criteria, redox kinetic measurements, and analytical applications, Electroanalysis, 16, 832, 2004. 42. Lovric, M., Stripping voltammetry, in Electroanalytical Methods, Scholz, F. (Ed.), Springer, Berlin=Heidelberg=New York, 2001, chap. II. 43. Brainina, K. and Neyman, E., Electroanalytical Stripping Methods, Wiley-VCH, New York, 1994. 44. Palleschi, G., Electrochemical biosensors for food analysis and the food industry, Ital. J. Food Sci., 13, 137, 2001. 45. Mello, L.D. and Kubota, L.T., Review of the use of biosensors as analytical tools in the food and drink industries, Food Chem., 77, 237, 2002. 46. Terry, L.A., White, S.F., and Tigwell, L.J., The application of biosensors to fresh produce and the wider Food industry, J. Agric. Food Chem., 53, 1309, 2005. 47. Cavicchioli, A., La-Scalea, M.A., and Gutz, I.G.R., Analysis and speciation of traces of arsenic in environmental, food and industrial samples by voltammetry: A review, Electroanalysis, 16, 698, 2004. 48. Florence, T.M., Polarography of azo compounds and their metal complexes, J. Electroanal. Chem. Interfacial Electrochem., 52, 115, 1974. 49. Fogg, A.G. and Yoo, K.S., Direct differential-pulse polarographic determination of mixtures of the food colouring matters tartrazine-Sunset Yellow FCF, tartrazine-Green S and amaranth-Green S in soft drinks, Analyst, 104, 723, 1979. 50. Ni, Y., Bai, J., and Jin, L., Simultaneous adsorptive voltammetric analysis of mixed colorants by multivariate calibration approach, Anal. Chim. Acta, 329, 65, 1996. 51. Kapor, M.A. et al., Electroanalysis of food dyes: Determination of indigo-carmine and tartrazine, Ecletica Quim., 26, 53, 2001. 52. Chanlon, S. et al., Determination of carmoisine, allura red and ponceau 4R in sweets and soft drinks by differential pulse polarography, J. Food Composition Anal., 18, 503, 2005. 53. Berzas Nevado, J.J., Rodriguez Flores, J., and Villasenor Llerena, M.J., Adsorptive stripping voltammetry (ASV) of tartrazine at the hanging mercury drop electrode in soft drinks, Fresenius J. Anal. Chem., 357, 989, 1997. 54. Combeau, S., Chatelut, M., and Vittori, O., Identiﬁcation and simultaneous determination of Azorubin, Allura red and Ponceau 4R by differential pulse polarography: Application to soft drinks, Talanta, 56, 115, 2002. 55. Becerro Dominguez, F., Gonzalez Diego, F., and Hernandez Mendez, J., Determination of sunset yellow and tartrazine by differential pulse polarography, Talanta, 37, 655, 1990. 56. Berzas Nevado, J.J., Rodríguez Flores, J., and Villaseñor Llerena, M.J., Square wave adsorptive voltammetric determination of sunset yellow, Talanta, 44, 467, 1997. 57. Florian, M. et al., Determination of brilliant blue FCF in the presence and absence of erythrosine and quinoline yellow food colours by cathodic stripping voltammetry, Food Addit. Contam., 19, 803, 2002. 58. Commission Regulation (EC) No. 466=2001, Off. J. Eur. Communities, L 77=1, 8 March, 2001. 59. Kopanica, M. and Novotny, L., Determination of traces of arsenic(III) by anodic stripping voltammetry in solutions, natural waters and biological material, Anal. Chim. Acta, 368, 211, 1998. 60. Bryce, D.W., Izquierdo, A., and Luque de Castro, M.D., Flow-injection anodic stripping voltammetry at a gold electrode for selenium(IV) determination, Anal. Chim. Acta, 308, 96, 1995. 61. Zakharova, E.A., Pichugina, V.M., and Tolmacheva, T.P., Determination of mercury in water and alcoholic drinks by stripping voltammetry, Zh. Anal. Khim., 51, 918, 1996.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 399 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

399

62. Meryan, V.M., Chugureanu, D.G., and Zayats, G.D., Determination of cadmium in the presence of the 2,20 -dipyridyl2,4-dihydroxybenzoic acid molecular complex and SCN ions by adsorption stripping voltammetry, Zh. Anal. Khim., 53, 48, 1998. 63. Safavi, A. and Shams, E., Determination of trace amounts of copper(II) by adsorptive stripping voltammetry of its complex with pyrogallol red, Anal. Chim. Acta, 385, 265, 1999. 64. Meryan, V.T., Mokanu, R., and Taragan, N.F., Determination of nickel in the presence of eriochrome black T by adsorption stripping voltammetry, Zh. Anal. Khim., 52, 463, 1997. 65. Gutierrez, C.A., Suarez, M.F., and Compton, R.G., Optimization of mercury thin ﬁlm electrodes for sono-ASV studies, Electroanalysis, 11, 16, 1999. 66. Ji, Z. and Guadalupe, A.R., Reusable doped sol-gel graphite electrodes for metal ions determination, Electroanalysis, 11, 167, 1999. 67. Fomintseva, E.E., Zakharova, E.A., and Pikula, N.P., Determination of tin and lead in canned juices and fruit by stripping voltammetry, Zh. Anal. Khim., 52, 590, 1997. 68. Adeloji, S.B.O. and Pablo, F., Adsorptive stripping voltammetric determination of ultratrace concentrations of molybdenum in biological and environmental materials on a glassy carbon mercury ﬁlm electrode, Electroanalysis, 7, 476, 1995. 69. Abo-Maali, N. and Abd El-Hady, D., Square-wave stripping voltammetry of uranium(VI) at the glassy carbon electrode. Application to some industrial samples, Electroanalysis, 11, 201, 1999. 70. Faller, C. et al., Modiﬁed solid electrodes for stripping voltammetric determination of tin, Fresenius J. Anal. Chem., 358, 670, 1997. 71. Matysik, F.-M., Gläser, P., and Werner, G., Analytical possibilities of microelectrode use for stripping voltammetry, Fresenius J. Anal. Chem., 349, 646, 1994. 72. Wahdat, F., Hinkel, S., and Neeb, R., Direct inverse voltammetric determination of Pb, Cu and Cd in some edible oils after solubilisation, Fresenius J. Anal. Chem., 352, 393, 1995. 73. Agra-Gutiérrez, C. et al., Anodic stripping voltammetry of copper at insonated glassy carbon-based electrodes: Application to the determination of copper in beer, Analyst, 124, 1053, 1999. 74. Babkina, S.S. and Ulakhovich, N.A., Amperometric biosensor based on denatured DNA for the study of heavy metals complexing with DNA and their determination in biological, water and food samples, Bioelectrochemistry, 63, 261, 2004. 75. Baldo, M.A., Bragato, C., and Daniele, S., Determination of lead and copper in wine by anodic stripping voltammetry with mercury microelectrodes: Assessment of the inﬂuence of sample pretreatment procedures, Analyst, 122, 1, 1997. 76. Kowalska, J. et al., Voltammetric determination of arsenic in plant material, Electroanalysis, 11, 1301, 1999. 77. Kowalska, J. and Golimowski, J., Voltammetric determination of arsenic in zinc oxide used as a feed additive, Electroanalysis, 10, 857, 1998. 78. Sancho, D. et al., Determination of copper and arsenic in reﬁned beet sugar by stripping voltammetry without sample pretreatment, Analyst, 123, 743, 1998. 79. Aprea, C. et al., Biological monitoring of pesticide exposure: A review of analytical methods, J. Chromatogr., 769, 191, 2002. 80. Garrido, E.M. et al., Electrochemical methods in pesticide control, Anal. Lett., 37, 1755, 2004. 81. Terry, L.A., White, S.F., and Tigwell, L.J., The application of biosensors to fresh produce and the wider food industry, J. Agric. Food Chem., 53, 1309, 2005. 82. Hance, R.J., Polarography of herbicides—A preliminary survey, Pest. Sci., 1, 112, 1970. 83. Ignjatovic, L.M. et al., Polarographic behavior and determination of some s-triazine herbicides, Electroanalysis, 5, 529, 1993. 84. Hernández Méndez, J., Carabias Martínez, R., and Rodríguez Gonzalo, E., Electroanalytical study of the pesticide guthion, J. Electroanal. Chem., 244, 221, 1988. 85. Olmedo, C. et al., Polarographic study of the herbicide metamitron, Electroanalysis, 6, 694, 1994. 86. Pedrero, M. et al., Determination of the herbicide desmetryne in organised media by adsorptive stripping voltammetry, Talanta, 53, 991, 2001. 87. Yulaev, M.F. et al., Development of a potentiometric immunosensor for herbicide simazine and its application for food testing, Sens. Actuators B, 75, 129, 2001. 88. Subbalakskmamma, M. and Reddy, S.J., Electrochemical reduction behavior and analysis of some organophosphorous pesticides, Electroanalysis, 6, 521, 1994.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 400 14.7.2008 8:24pm Compositor Name: BMani

400

Handbook of Food Analysis Instruments

89. De Souza, D. and Machado, S.A.S., Electroanalytical method for determination of the pesticide dichlorvos using gold-disk microelectrodes, Anal. Bioanal. Chem., 382, 1720, 2005. 90. Sreedhar, N.Y. et al., Differential pulse polarographic determination of dicrotophos, crotoxyphos and chlorfenvinphos in grains and soils, Talanta, 44, 1859, 1997. 91. Calvo-Marzal, P. et al., Electroanalytical determination of acid phosphatase activity by monitoring p-nitrophenol, Anal. Chim. Acta, 441, 207, 2001. 92. Southwick, L.M. et al., The polarographic reduction of some dinitroaniline herbicides, Anal. Chim. Acta, 82, 29, 1976. 93. Pedrero, M. et al., Determination of dinoseb by adsorptive stripping voltammetry using a mercury ﬁlm electrode, Fresenius J. Anal. Chem., 349, 546, 1994. 94. Del Carlo, M. et al., Screening of food samples for carbamate and organophosphate pesticides using an electrochemical bioassay, Food Chem., 84, 651, 2004. 95. Aguí, L. et al., Determination of disulﬁram by adsorptive stripping voltammetry at gold disk microelectrodes, Electroanalysis, 14, 486, 2002. 96. Bird, C.L. and Kuhn, A.T., Electrochemistry of the viologens, Chem. Soc. Rev., 10, 49, 1981. 97. De Souza, D. and Machado, S.A.S., Electrochemical detection of the herbicide paraquat in natural water and citric fruit juices using microelectrodes, Anal. Chim. Acta, 546, 85, 2005. 98. Reviejo, A.J. et al., Determination of organochlorine pesticides in apple samples by differential-pulse polarography in emulsiﬁed medium, Anal. Chim. Acta, 264, 141, 1992. 99. Samatha, K. and Sreedhar, N.Y., Polarographic determination of deltamethrin, Talanta, 49, 53, 1999. 100. Zhang, Y. et al., Disposable biosensor test for organophosphate and carbamate insecticides in milk, J. Agric. Food Chem., 53, 5110, 2005. 101. Albareda-Sirvent, M., Merkoçi, A., and Alegret, S., Pesticide determination in tap water and juice samples using disposable amperometric biosensors made using thick-ﬁlm technology, Anal. Chim. Acta, 442, 35, 2001. 102. Olek, M., Blanchard, F., and Sudraud, G., Application de la détection électrochimique au dosage des résidus de quelques insecticides carbamates par chromatographie liquide haute performance, J. Chromatogr. A, 325, 239, 1985. 103. Nunes, G.S. et al., Determination of carbamate residues in crop samples by cholinesterase-based biosensors and chromatographic techniques, Anal. Chim. Acta, 362, 59, 1998. 104. Nunes, G.S. et al., Evaluation of a highly sensitive amperometric biosensor with low cholinesterase charge immobilized on a chemically modiﬁed carbon paste electrode for trace determination of carbamates in fruit, vegetable and water samples, Anal. Chim. Acta, 399, 37, 1999. 105. La Rosa, C. et al., Amperometric ﬂow through biosensor for the determination of pesticides, Anal. Chim. Acta, 308, 129, 1995. 106. Nikolelis, D.P. et al., Flow injection analysis of carbofuran in foods using air stable lipid ﬁlm based acetylcholinesterase biosensor, Anal. Chim. Acta, 537, 169, 2005. 107. Fernández, C., Reviejo, A.J., and Pingarrón, J.M., Development of graphite-poly(tetraﬂuoroethylene) composite electrodes voltammetric determination of the herbicides thiram and disulﬁram, Anal. Chim. Acta, 305, 192, 1995. 108. Ivanov, A. et al., Cholinesterase sensors based on screen-printed electrodes for detection of organophosphorus and carbamic pesticides, Anal. Bioanal. Chem., 377, 624, 2003. 109. Longobardi, F. et al., Use of electrochemical biosensor and gas chromatography for determination of dichlorvos in wheat, J. Agric. Food Chem., 53, 9389, 2005. 110. del Carlo, M. et al., An electrochemical bioassay for dichlorvos analysis in durum wheat samples, J. Food Prot., 69, 1406, 2006. 111. Concialini, V., Lippolis, M.T., and Galletti, G.C., Preliminary studies for the differential-pulse polarographic determination of a new class of herbicides: Sulphonylureas, Analyst, 114, 1617, 1989. 112. Palchetti, I. et al., Determination of anticholinesterase pesticides in real samples using a disposable biosensor, Anal. Chim. Acta, 337, 315, 1997. 113. Wagner, G. and Guibauld, G.G. (Ed.), Food Biosensor Analysis, Marcel-Dekker, New York, 1994. 114. Scott, A.O., Biosensors for Food Analysis, Royal Society of Chemistry, Cambridge, United Kingdom, 1998. 115. Del Carlo, M. et al., Biosensors for food quality assessment, in Food Biotechnology, Shetty, K., Paliyath, G., Pometto, A., and Levin, R. (Eds.), 2nd ed., CRC Press, Taylor & Francis Group, Portland, 2006.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 401 14.7.2008 8:24pm Compositor Name: BMani

Electroanalytical Techniques and Instrumentation in Food Analysis

401

116. Terry, L.A., White, S.F., and Tigwell, L.J., The application of biosensors to fresh produce and the wider food industry, J. Agric. Food Chem., 53, 1309, 2005. 117. Palleschi, G. and Cubadda, R., Electrochemical biosensors for food analysis and the food industry, Ital. J. Food Sci., 13, 137, 2001. 118. Dinçkaya, E. et al., Sulﬁte determination using sulﬁte oxidase biosensor based glassy carbon electrode coated with thin mercury ﬁlm, Food Chem., 101, 1540, 2007. 119. Kelly, S.C. et al., Development of an interferent free amperometric biosensor for determination of L-lysine in food, Anal. Chim. Acta, 412, 111, 2000. 120. Stobiecka, A., Radecka, H., and Radecki, J., Novel voltammetric biosensor for determining acrylamide in food samples, Biosens. Bioelectron., 22, 2165, 2007. 121. Maestre, E., Katakis, I., and Dominguez, E., Amperometric ﬂow-injection determination of sucrose with a mediated tri-enzyme electrode based on sucrose phosphorylase and electrocatalytic oxidation of NADH, Biosens. Bioelectron., 16, 61, 2001. 122. Ferreira, V.S., Zanoni, M.V.B., and Fogg, A.G., Indirect cathodic stripping voltammetric determination of ceftazidime as a mercury salt, Anal. Chim. Acta, 367, 255, 1998. 123. Portela, M.J. et al., Voltammetric method for the determination of the ﬂavor enhancer inosinic acid, Analyst, 119, 2183, 1994. 124. Rodrigues, P.G. et al., Automatic ﬂow system with voltammetric detection for diacetyl monitoring during brewing process, J. Agric. Food Chem., 50, 3647, 2002. 125. Rodrigues, J.A., Barros, A.A., and Rodrigues, P.G., Differential pulse polarographic determination of dicarbonyl compounds in foodstuffs after derivatization with o-phenylenediamine, J. Agric. Food Chem., 47, 3219, 1999. 126. Michelitsch, A. and Wurglics, M., Electrochemical oxidation of hyperforin on glassy carbon electrode and determination in herbal medicinal products, Electroanalysis, 15, 797, 2003. 127. Kaláb, T. and Skládal, P., Evaluation of different mediators for the development of amperometric microbial bioelectrodes, Electroanalysis, 6, 1004, 1994. 128. Xu, Y.F., Velasco-Garcia, M., and Mottram, T.T., Quantitative analysis of the response of an electrochemical biosensor for progesterone in milk, Biosens. Bioelectron., 20, 2061, 2005. 129. Laschi, S. et al., Polychlorinated biphenyls (PCBs) detection in food samples using an electrochemical immunosensor, J. Agric. Food Chem., 51, 1816, 2003. 130. Lin, L.A., Detection of alkaloids in foods with a multi-detector high-performance liquid chromatographic system, J. Chromatogr. A, 632, 69, 1993. 131. Yu, H., Ding, Y.S., and Mou, S.F., Direct and simultaneous determination of amino acids and sugars in rice wine by high-performance anion-exchange chromatography with integrated pulsed amperometric detection, Chromatographia, 57, 721, 2003. 132. Van Dyck, M.M.C., Rollmann, B., and De Meester, C., Quantitative estimation of heterocyclic aromatic amines by ion-exchange chromatography and electrochemical detection, J. Chromatogr. A, 697, 377, 1995. 133. Siu, D.C. and Henshall, A., Ion chromatographic determination of nitrate and nitrite in meat products, J. Chromatogr. A, 804, 157, 1998. 134. Xu, J.M. et al., Determination of electroinactive organic acids in red wine by ion-exclusion chromatography using a poly-o-phenylenediamine ﬁlm modiﬁed electrode, Chromatographia, 57, 751, 2003. 135. Klejdus, B. et al., Determination of isoﬂavones in soybean food and human urine using liquid chromatography with electrochemical detection, J. Chromatogr. B, 806, 101, 2004. 136. Schneiderman, M.A., Sharma, A.K., and Locke, D.C., Determination of vitamin A palmitate in cereal products using supercritical ﬂuid extraction and liquid chromatography with electrochemical detection, J. Chromatogr. A, 765, 215, 1997. 137. Galceran, M.T., Pais, P., and Puignou, L., High-performance liquid chromatographic determination of ten heterocyclic aromatic amines with electrochemical detection, J. Chromatogr. A, 655, 101, 1993. 138. Rivera, L. et al., Solid-phase extraction for the selective isolation of polycyclic aromatic hydrocarbons, azaarenes and heterocyclic aromatic amines in charcoal-grilled meat, J. Chromatogr. A, 731, 85, 1996. 139. Galeano Díaz, T. et al., Determination of nitrofurantoin, furazolidone and furaltadone in milk by high-performance liquid chromatography with electrochemical detection, J. Chromatogr. A, 764, 243, 1997.

Otles/Handbook of Food Analysis Instruments 45660_C017 Final Proof page 402 14.7.2008 8:24pm Compositor Name: BMani

402

Handbook of Food Analysis Instruments

140. Iwase, H., Use of an amino acid in the mobile phase for the determination of ascorbic acid in food by high-performance liquid chromatography with electrochemical detection, J. Chromatogr. A, 881, 317, 2000. 141. Schultheiss, J., Jensen, D., and Galensa, R., Determination of aldehydes in food by high-performance liquid chromatography with biosensor coupling and micromembrane suppressors, J. Chromatogr. A, 880, 233, 2000. 142. Lookabaugh, M. and Krull, I.S., Determination of nitrite and nitrate by reversed-phase high-performance liquid chromatography using on-line post-column photolysis with ultraviolet absorbance and electrochemical detection, J. Chromatogr., 452, 295, 1988. 143. Mattila, P., Lehtonen, M., and Kumpulainen, J., Comparison of in-line connected diode array and electrochemical detectors in the high-performance liquid chromatographic analysis of coenzymes Q9 and Q10 in food materials, J. Agric. Food Chem., 48, 1229, 2000.

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 403 16.7.2008 11:06pm Compositor Name: MSubramanian

Electrophoresis 18 Capillary in Food Analysis Carmen García-Ruiz and Maria Luisa Marina CONTENTS 18.1 18.2 18.3

Introduction ........................................................................................................................ 403 Principles ........................................................................................................................... 404 Theory ................................................................................................................................ 405 18.3.1 Capillary Zone Electrophoresis ......................................................................... 405 18.3.2 Electrokinetic Chromatography ......................................................................... 407 18.4 Instrumentation .................................................................................................................. 407 18.4.1 Separation Capillaries ........................................................................................ 408 18.4.2 High Voltage Power Supplies ........................................................................... 409 18.4.3 Sample Introduction ........................................................................................... 409 18.4.4 Detectors ............................................................................................................ 410 18.4.5 New Analytical Approaches .............................................................................. 411 18.4.6 Two-Dimensional CE ........................................................................................ 411 18.4.7 Microchip Electrophoresis ................................................................................. 411 18.5 Applications of CE in Food Analysis ............................................................................... 411 18.5.1 Additives ............................................................................................................ 412 18.5.2 Amino Acids and Related Compounds ............................................................. 413 18.5.3 Peptides and Proteins ......................................................................................... 413 18.5.4 Phenolic Compounds ......................................................................................... 414 18.5.5 DNAs ................................................................................................................. 415 18.5.6 Carbohydrates .................................................................................................... 416 18.5.7 Vitamins ............................................................................................................. 416 18.5.8 Small Organic and Inorganic Ions ..................................................................... 416 18.5.9 Organic Contaminants ....................................................................................... 417 18.5.10 Enantiomeric Analysis of Chiral Compounds ................................................... 417 18.6 Future Trends ..................................................................................................................... 419 Acknowledgments ........................................................................................................................ 420 References ..................................................................................................................................... 420

18.1 INTRODUCTION The applications of capillary electrophoresis (CE) to food analysis are reviewed in this chapter. However, before discussing its applications to food analysis, sections describing brieﬂy the principles and theory of CE and its instrumentation are discussed as detailed descriptions are beyond the scope of this book. CE is a relatively modern separation technique characterized by high efﬁciency, low consumption of samples and reagents (and considered a clean analytical technique), and high versatility. This separation technique has shown excellent potential in the analysis of foods, the development of more applications being expected in the near future. 403

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 404 16.7.2008 11:06pm Compositor Name: MSubramanian

404

Handbook of Food Analysis Instruments

18.2 PRINCIPLES Separation by CE is based on two electrokinetic phenomena: electrophoresis and electroosmosis [1–4]. Electrophoresis is the migration of a charged solute under the presence of an electric ﬁeld. The solute is accelerated by electrostatic forces and is broken down by friction with the surrounding medium. Under steady-state conditions, the two opposite forces balance each other and a ﬁnal electrophoretic velocity, nep, is reached, which remains constant at a constant electric ﬁeld strength, E (in a capillary ﬁlled with a homogeneous running buffer, E is calculated as the voltage applied divided by the total length of the capillary) [2]. nep ¼

qE ¼ mep E 6phr

(18:1)

where q is the charge r is the ratio of the charged solute h is the viscosity of the surrounding medium mep is the electrophoretic mobility of the charged solute in the separation medium considered Electroosmosis is the movement of liquid inside a capillary when an electric ﬁeld is applied, producing an electroosmotic ﬂow (EOF). It is produced because at pH > 2 the surface of fused silica capillaries is negatively charged. There will be an excess of positive counterions in the zone forming the boundary layer. This zone of positive charge adjacent to the capillary wall will be accelerated by the electric ﬁeld and will also experience friction by the medium next to this layer. Therefore, a steady-state constant velocity due to EOF, neo, is reached in the liquid outside the electrical double layer (Helmholtz–Smoluchowski equation) [2]: neo ¼

«zE ¼ meo E 4ph

(18:2)

where « is the electric permittivity of the surrounding medium z is the electrokinetic potential at the surface of the charged wall meo is the electroosmotic mobility As a consequence, the observed velocity (apparent velocity) of a solute, ns, will be due to the sum of both velocities: ns ¼ nep þ neo ¼ (mep þ meo )E ¼ ma E

(18:3)

where ma is the apparent mobility, that is, the sum of the electrophoretic and the electroosmotic mobilities. Separations based on electrokinetic phenomena can also be described as chromatographic processes when an additive with different velocity from that of the solutes (pseudostationary phase) is present in the running buffer or when capillary columns with true stationary phases are employed for solute separations. As a consequence, depending on the separation principle, different CE separation modes (or working modes) exist: capillary zone electrophoresis (CZE), capillary gel electrophoresis (CGE), capillary isoelectric focusing (CIEF), and capillary isotachophoresis (CITP) based only on electrokinetic principles, and electrokinetic chromatography (EKC) and capillary electrochromatography (CEC) based on the combination of electrokinetic and chromatographic principles [1]. It is important to remark that the aqueous buffer employed in CE can be replaced by

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 405 16.7.2008 11:06pm Compositor Name: MSubramanian

405

Capillary Electrophoresis in Food Analysis

TABLE 18.1 Basic and Speciﬁc Separation Principles and Main Application Fields of Each Separation Mode by CE Basic Separation Principle Electrokinetic

Electrokinetic þ chromatographic

Electrokinetic or electrokinetic þ chromatographic

Speciﬁc Separation Principle Free mobility in aqueous solution Size and charge Isoelectric point Limits of mobility Distribution in a pseudostationary phase Distribution in a stationary phase Mobility in nonaqueous solution

Separation Mode CZE CGE CIEF CITP EKC CEC NACE

Main Application Field Food, environmental, pharmaceutical, and clinical Proteomic and genomic Proteomic Proteomic and environmental Food, environmental, pharmaceutical, and clinical Environmental and pharmaceutical Pharmaceutical

Source: Adapted from Marina, M.L., Ríos, A., and Valcárcel, M. (Eds.), Analysis and Detection by Capillary Electrophoresis, CAC Series, Elsevier, Amsterdam, the Netherlands, 2005.

an organic solvent (mainly formaldehyde or methanol) containing an electrolyte, giving rise to a working mode called nonaqueous capillary electrophoresis (NACE). Separation methods by NACE can be based only on electrokinetic principles or on combination of electrokinetic and chromatographic principles depending on the additives used in the separation media. Table 18.1 shows separation principles of different separation modes in CE. Contrary to chromatographic techniques where a change in the separation principle implies a change in the separation column, in CE it is possible to modify the separation principle by changing the running buffer and additives in the same separation capillary (except in CEC separations where the capillary has to be changed).

18.3 THEORY Because CZE and EKC are the most widely employed separation modes in CE and the main separation modes used in food analysis (see Table 18.1), their theory is brieﬂy described herein. In addition, Table 18.2 shows the fundamental and practical equations used for the determination of analytical parameters such as the theoretical plate number and resolution valid for both separation modes [5].

18.3.1 CAPILLARY ZONE ELECTROPHORESIS Capillary zone electrophoresis (CZE) is the easiest and most widely employed working mode in CE. In CZE the separation capillary is ﬁlled with a background electrolyte (BGE), usually a buffer solution at certain pH, and solutes are separated based on electrophoresis and electroosmosis phenomena, that is, according to their charge-to-mass ratio. This technique is only applicable to ionic or ionizable compounds soluble in aqueous buffers. Figure 18.1a shows how solutes (at velocity ns) are separated by CZE according to the combination of their own electrophoretic (nep) and electroosmosis (neo) velocities when the normal polarity mode is employed (cathode is placed at the detection side). Thus, because of the velocity of cations toward the cathode, they elute ﬁrst (ns > neo), then neutral molecules move through the separation capillary at the same velocity as

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 406 16.7.2008 11:06pm Compositor Name: MSubramanian

406

Handbook of Food Analysis Instruments

TABLE 18.2 Fundamental and Practical Equations for the Determination of Theoretical Plate Number and Resolution in CE Analytical Parameters

Fundamental Equations ma Vl ma El ¼ 2DL 2D

Theoretical plate number (N)

N¼

Resolution (Rs)

Rs ¼ 0:177 (mep2 mep1 ) with m ¼

mep2 þ mep1 2

V D(m meo )

Practical Equations ti 2 N ¼ 5:54 w1=2

1=2

Rs ¼ 1:18

ti2 ti1 w1=2(1) þ w1=2(2)

Note: D, diffusion coefﬁcient of the solute; ti, migration time of the solute; w1=2, peak width at half height; l, effective length of the capillary (from the injection of the detection point); L, total length of the capillary; ma, apparent mobility; meo, electroosmosis mobility; mep, electrophoretic mobility.

the EOF (ns ¼ neo) (not separated), and ﬁnally, anions will reach the detector if their velocity, which is toward the anode, is lower than the EOF velocity to the cathode (ns < neo). Therefore, it is obvious that the electroosmotic velocity is a very important parameter in CZE that has to be controlled to achieve a certain separation. The change of BGE pH is the most effective way to control it. However, the nature and concentration of BGE, the modiﬁcation of the capillary wall, the use of additives, and the voltage applied also inﬂuence the EOF [5].

Detector

ns neo

ns

+

neo nep

Anode

(a)

ns neo nep

– Cathode

CZE Detector

+

PS

ns

neo

ns neo

PS

nep

ns neo

Anode

(b)

FIGURE 18.1

–

nep

EKC

Separation principle in (a) CZE and (b) EKC.

Cathode

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 407 16.7.2008 11:06pm Compositor Name: MSubramanian

407

Capillary Electrophoresis in Food Analysis

18.3.2 ELECTROKINETIC CHROMATOGRAPHY In electrokinetic chromatography (EKC), the most versatile CE working mode, one or several pseudostationary phases (PS, also called pseudophases or separation carriers) are added to the BGE, obtaining a separation buffer at certain pH capable of solute separation. A large number of pseudostationary phases have been used: anionic, cationic or nonionic surfactants, microdroplets in oil-in-water or water-in-oil microemulsions, polymerized micelles or polymers, charged dendrimers, soluble linear polymers, and cyclodextrins or metal complexes. Contrary to CZE, EKC can be applied to separate neutral as well as charged solutes [2]. Figure 18.1b shows how the separation of a neutral solute is possible in EKC using a negatively charged PS in the running buffer. The EOF velocity should be greater than the PS velocity toward the anode (ns > nPS) to enable the PS to reach the detector (that is placed in the cathode capillary extreme in the normal polarity mode). Separation of several neutral solutes will occur due to the different partitioning of analytes between the PS and the surrounding medium according to a distribution coefﬁcient (P) [2]: P ¼ CPS =CS

(18:4)

which is a ratio between the molar concentrations of a solute in the PS (CPS) and surrounding phase (CS). This distribution coefﬁcient can be related with the retention factor (k) by means of the phase ratio (volume of PS (VPS) divided by volume of aqueous phase (VAQ) [2]): k ¼ P(VPS =VAQ )

(18:5)

and k can be easily calculated from the migration times of the solute (tr), EOF marker (t0), and PS marker (tPS) as [6]: tr t0 k¼ tr t0 1 tPS

(18:6)

When the EOF is null (t0 ! 1), the retention factor can be rewritten as [6]: k¼

tPS tr tPS

(18:7)

18.4 INSTRUMENTATION The basic instrumentation in CE (see Figure 18.2a) consists of a separation capillary (mostly a fused silica capillary) with both ends immersed in vials ﬁlled with the BGE (also called running buffer or separation buffer). An electrode is placed in each vial, immersed in the BGE, and connected to a high voltage power supply. The introduction of the sample into the separation capillary is made by replacing the buffer vial placed in the inlet position for a sample vial and using electrokinetic, siphoning, or overpressure methods of injection. Then, the separated bands of analytes are detected, usually by an on-capillary UV-Vis spectrophotometric detector. The detector signal is usually registered and processed by a suitable PC software. The plot of the variation of the detector signal as a function of time is called an electropherogram. In the 1980s, the ﬁrst compact CE instrument was introduced in the market, and since then CE instruments have been improved to satisfy the requirements demanded nowadays. As an example, the scheme of a modern commercial CE instrument is shown in Figure 18.2b. It can be observed that the separation capillary is placed in a cassette holder in a thermostatized compartment.

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 408 16.7.2008 11:06pm Compositor Name: MSubramanian

408

Handbook of Food Analysis Instruments

Capillary thermostatting Diodearray detector

HV

Separation capillary Detector

Computer

Vial carousel (thermostatted) Buffer replenishment

BGEinlet Sample electrodes

BGEoutlet

High voltage power supply (a)

(b)

FIGURE 18.2 (a) Basic scheme of the instrumentation in CE and (b) scheme of a modern instrument used for routine analysis. (From Agilent Technologies. With permission.)

In addition, sample and running buffer vials are located in an autosampler that allows working in an automatic way (sequence mode). The most widely used CE instruments are from Agilent Technologies [7], Beckman Coulter [8], and Prince Technologies [9] although there are also other manufacturers [2,10]. In fact, it is interesting to remark that a growing number of manufacturers are marketing CE as a targeted solution to a problem rather than as a technique, for instance, DNA sequencers have never been marketed as CE instruments have been [10].

18.4.1 SEPARATION CAPILLARIES Separation capillaries employed in most applications of CE are fused silica capillaries with circular cross section and are externally covered with a polyimide layer [11]. Usually, an inner diameter of 50 or 75 mm and an outer diameter of 365 mm are employed. The total length varies from several centimeters up to 1 m (mostly 40–60 cm are used). Since the polyimide ﬁlm on the fused silica capillaries is not transparent to UV-Vis radiation, a small segment should be removed to make a detection window. Usually, the capillary detection window is made by burning off with a gas torch or lighter and removing the polyimide section by cleaning with a wet paper. It can also be removed by using a heated drop of concentrated sulfuric acid or even mechanically using a razor blade [2]. It is important to consider that the detection window obtained is very fragile and has to be manipulated with special care. There are commercially available capillaries with UV-transparent coatings or rectangular cross sections [11], but they are scarcely employed. Sometimes coated fused silica capillaries are used to avoid undesired effects such as the adsorption of analytes to the capillary walls. These coatings can be mobile (i.e., adding cellulose

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 409 16.7.2008 11:06pm Compositor Name: MSubramanian

409

Capillary Electrophoresis in Food Analysis

derivatives to the running buffer) or permanent (i.e., polyacrylamide or polyethyleneglycol capillaries) [11], and are especially useful for compounds with positive net charge such as proteins.

18.4.2 HIGH VOLTAGE POWER SUPPLIES CE allows high electric ﬁelds (~100 kV=m) to be applied. The maximum voltage in commercial CE instruments does not exceed 30 kV [2,10], and the electric current is usually limited to 300 mA. In addition, instrument versatility increases when the high voltage power supply allows working in the normal (þkV) and reverse (kV) polarity modes.

18.4.3 SAMPLE INTRODUCTION Very small sample volumes, in the nanoliter range, should be injected in CE to obtain high resolution because band broadening depends on the length of the sample injected. There are two ways for the introduction of nanoliter volumes in CE: hydrodynamic and electrokinetic sample injections. Due to the importance of on-line sample preconcentration techniques for the detection of trace levels of analytes in real samples, they are also brieﬂy described under sample introduction techniques. In hydrodynamic sample injection, the inlet end of the capillary is immersed in the sample vial and a pressure difference (which can be achieved by overpressure, vacuum, or siphoning) is applied between the two capillary ends. The amount of analyte introduced into the capillary [2], Qinj, at a constant pressure difference, DP, between the inlet and outlet capillary ends, is directly proportional to the injection time, tinj, and to the molar concentration of the analyte in the sample, ci: Qinj ¼ DP

ptinj ci dc4 128hL

(18:8)

where L is the total length of the capillary with inner diameter dc and h the sample solution viscosity. Hydrodynamic injection is universal because it does not discriminate between components with different electrophoretic mobilities having the sample plug the same composition as the sample contained in the injection vial. In electrokinetic sample injection, the inlet end of the capillary is immersed in the sample vial and a voltage difference is applied between the two capillary ends for a time (tinj) (usually 10–15 kV for ~10 s). In this case, the amount of analyte injected [2], Qinj, depends on the voltage applied during the injection (Vi) and on the apparent mobility (ma): Qinj ¼ ma

ptinj ci dc2 Vi 4 L

(18:9)

where ma can be calculated from the migration time of a given analyte (ti), the applied voltage (Vi), and the effective (l) and total (L) capillary lengths as ma ¼

Ll ti V i

(18:10)

As a consequence, this injection is selective because injected amounts of faster-moving ions are higher than those of ions with lower mobilities. Sample stacking and sweeping on-line preconcentration techniques can be considered as sample introduction techniques based on electrokinetic principles [1]. They allow an increase of sample loading, minimizing the impact of the sample zone length on the band broadening and increasing detection sensitivity.

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 410 16.7.2008 11:06pm Compositor Name: MSubramanian

410

Handbook of Food Analysis Instruments

Sample stacking is based on the injection of a sample zone prepared in a matrix with a higher resistance, that is, a lower conductivity, than the running buffer. Thus, when a voltage difference is applied between the ends of the capillary, sample ions acquire higher electrophoretic mobilities in the sample plug than in the buffer region. In this way, sample ions reduce their velocity and are focused in a narrow band between sample and buffer regions. In sample stacking, also called the normal stacking mode (NSM), the sample is dissolved in water, diluted buffer, or organic solvents, and then injected hydrodynamically. Usually, an hydrodynamic injection is limited to 3%–4% of the capillary length to not lose efﬁciency, but under the sample stacking mode it is possible to ﬁll between 10% and 20% of the capillary length without loss of efﬁciency. Even ﬁlling the total capillary length is possible if a step to remove the sample matrix by applying reverse polarity is made previously to apply the normal separation voltage, in which case the reverse electrode polarity stacking mode is used (REPSM). If reverse migrating micelles are used to stack the sample analytes, then the stacking mode is called stacking with reverse migrating micelles (SRMM). If a water plug is also included, the introduction of the sample is by stacking with reverse migrating micelles and a water plug (SRW). However, if sample analytes are preconcentrated during an electrokinetic injection, then ﬁeld-ampliﬁed sample stacking (FASS), also named as ﬁeld-enhanced sample injection (FESI), or ﬁeld-ampliﬁed sample injection (FASI) occurs [1,12]. Sweeping requires sample dissolution in a buffer (mostly phosphate buffer) without the pseudostationary phase (i.e., micelles, cyclodextrins, polymers, dendrimers) used for the separation and hydrodynamic injection in a long capillary length depending on the interaction between the solutes and the pseudostationary phase. Quantitatively, the resulting length of the swept zones (lsweep) can be approximated by: lsweep ¼ linj (1=1 þ k)

(18:11)

where linj is the length of the injected sample zone. Then, sweeping is basically dependent on the retention factor and the length of the initial zone, which suggests narrower zones for high k analytes [1,13,14].

18.4.4 DETECTORS On-line detection in CE can be accomplished directly on the separation capillary (on-column detection), or using a detector connected to the end of the capillary (end-column detection). Oncolumn detection is used for UV-Vis spectrophotometric, ﬂuorescence, and nuclear magnetic resonance (NMR) detectors whereas end-column detection is used for atomic and molecular mass spectrometry (ICP-MS and ESI-MS), electrochemical, and Raman spectroscopy detectors [1]. With few exceptions, only UV-Vis absorption or ﬂuorescence detectors are installed in compact commercial instruments [2,10]. In addition to the commercial availability of UV-Vis absorption detectors, the simplicity, versatility, relatively low cost, and nondestructive characteristics (which enable tandem connection with end-column detectors) have made UV-Vis spectrophotometric detection the most widely used in CE. In fact, diode-array UV-Vis detectors are usually employed, having advantages over conventional UV-Vis absorption detection by providing peak purity and some structural information on the basis of the compound spectra obtained. The main drawback of this detection technique is its limited sensitivity due to the small inner diameter of the separation capillaries (usually 50 or 75 mm), which limits the volume of sample that can be injected as well as the optical path length. To improve sensitivity with UV-Vis spectrophotometric detectors, capillaries with extended optical path length [7] can be used. In addition, alternative detection systems [1], especially laser-induced ﬂuorescence detectors (LIF), can be employed to increase sensitivity achieved by UV-Vis spectrophotometric detection, although in this case, a derivatization step is usually needed because only few compounds possess

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 411 16.7.2008 11:06pm Compositor Name: MSubramanian

Capillary Electrophoresis in Food Analysis

411

native ﬂuorescence. It is important to consider that the use of on-line sample preconcentration techniques (stacking or sweeping) can be enough to obtain good sensitivities, in addition to selectivity, in CE with UV-Vis detection.

18.4.5 NEW ANALYTICAL APPROACHES The use of existing sample screening systems previously to CE and the automation of sample treatment and calibration in CE instruments are interesting practical approaches required nowadays in routine analysis [15]. Continuous ﬂow systems (CFS) have been used for the integration of the sample treatment unit into commercial CE instruments, thereby playing two roles: working as a unit for the automatic sample preparation and=or functioning as an interface between the sample treatment unit and the CE equipment. Such arrangements can be used for sample clean-up, preconcentration, derivatization, or other more complex sample treatments. On the other hand, CFS can also be used for the automation of calibration in CE, allowing the preparation of standard solutions and their introduction in the CE autosampler.

18.4.6 TWO-DIMENSIONAL CE In two-dimensional CE, samples from the ﬁrst dimension can be directly transferred to the second dimension (comprehensive 2D-CE) [16], or a fraction leaving from the ﬁrst dimension can be introduced and analyzed in the second one (heart-cutting 2D-CE) [17]. In the latter case, the sample transfer has been mainly made on-line, where the capillaries are directly connected through a ‘‘tee’’ interface, which permits the introduction of the desirable electrophoretic buffer in the second capillary and also has the electrode to apply the high voltage [17].

18.4.7 MICROCHIP ELECTROPHORESIS Miniaturization is a general trend in all scientiﬁc and technical ﬁelds that has also affected CE. Miniaturization of CE has been performed using microchips, also called chips or microﬂuidic devices. Electrophoretic microchips are planar structures with channels of diameter comparable to those of standard separation capillaries in CE but much shorter (~50 mm width ~20 mm depth and ~3 cm length), allowing high-speed separations (in seconds) and high efﬁciency due to the high ﬁeld strength per unit length of separation channel. Microchannels are made by photolithography and etching of silica or quartz structures, and by laser ablation or casting of polymeric structures. Because polymeric=plastic microchips can be manufactured at low cost, they can be disposed of after use. Moreover, parallel separation channels can be fabricated in a single microchip, enabling high-throughput analyses. In addition to these speciﬁc advantages of microchips, a low consumption of samples and reagents as well as minimal waste generation are advantages shared with CE systems [18]. Fluid manipulation and sample introduction on electrophoretic microchips are made by electrokinetic phenomena by applying voltage. Separated zones are mainly detected using LIF or electrochemical detectors.

18.5 APPLICATIONS OF CE IN FOOD ANALYSIS In food analysis quality and safety control there are two main aspects demanded by consumers [19]. For this reason, the development of analytical methodologies solving different problems such as detection of adulterations [20], control of food contamination, study of effect of processing, determination of chemical composition of foods, etc. is needed. Food analysis involves the determination of very diverse food components ranging from small ions to macromolecules like proteins. In addition, these compounds can occur in very complex sample matrices and=or at

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 412 16.7.2008 11:06pm Compositor Name: MSubramanian

412

Handbook of Food Analysis Instruments

trace levels requiring selective and sensitive analytical methods for their determination. These requirements can be achieved in CE by using off-line and on-line sample treatment techniques (solid-phase extraction, liquid–liquid extraction, microdialysis, etc.), on-column preconcentration techniques based on electrophoretic principles (stacking, sweeping), and alternative detection systems (spectroscopic, spectrometric, and electrochemical) to the widely used UV-Vis absorption detection. Reviews of applications of CE in food analysis have been available in the last few years [21–24]. In this section, an overview of the different applications of CE in food analysis including a more detailed description of the most recent ones is given, with the objective to be representative of the ﬁeld rather than being comprehensive.

18.5.1 ADDITIVES Food additives refer to any substance added to food to fulﬁll a particular function [23]. Commonly used food additives include colorants, antioxidants, sweeteners, and preservatives. To regulate and control the use of synthetic food additives there are legislative requirements. Among the many analytical procedures developed, CE can play an important role to determine and analyze food additives. Most applications performed by CE have recently been reviewed [23,24], the analysis of colorants (mainly) and preservatives by CE being the most usual applications in this ﬁeld. Colorants are usually added to various commercial food products to increase their attractiveness and to achieve the desired color. Colorants permitted as food additives can be natural or synthetic. The allowable amount of synthetic colorants is strictly limited because of their potential toxicity. Synthetic colorants have been analyzed in different samples such as ice creams, candies, and beverages [23,24]. In fact, the determination of groups of colorants, mainly synthetic, has been achieved in milk and alcoholic beverages by cyclodextrin-modiﬁed EKC using solid-phase extraction (SPE) for sample pretreatment [24]. However, recently, a microemulsion EKC procedure has been shown to be a good alternative to analyze food dyes since the sample preparation step can be skipped. In this work, eight food colorants (tartrazine, fast green FCF, brilliant blue FCF, allura red AC, indigo carmine, sunset yellow FCF, new coccine, and carminic acid) in three different ﬂavored soft drinks and three different ﬂavored popsicles have been analyzed [25]. The analysis of preservatives in wine, meat products, and vegetables has been performed by rapid CZE methods [23]. Recently, the analysis of commonly used preservatives (benzoate, sorbate) and vitamin C by both conventional CE and microchip electrophoresis with capacitively coupled contactless conductivity detection has been performed. Although the optimized working conditions suitable for CE were not adequate for microchips, which required a further optimization of separation conditions, qualitative and quantitative determination of these food additives was made in food samples such as soft drinks and vitamin C tablets [26]. Other applications of CE to analyze food additives include the determination of antioxidants and sweeteners. A group of eight antioxidants (protocatechuic acid, salicylic acid, p-hydroxybenzoic acid, vanillic acid, syringic acid, p-coumaric acid, ferulic acid, and sinapic acid) have been separated by CE using hexadimethrine bromide to make the direction of the EOF the same as the electrophoretic mobility of the ions (phenomenon called co-electroosmosis). This method was applied to quantitate antioxidants present in different real samples such as cereals, wine, and beer. These additives are signiﬁcant since they prolong the shelf-life of foods by protecting against deterioration caused by oxidation (i.e., fat rancidity and color changes) and also contribute to the sensory characteristics of foods [27]. On the other hand, since the amount of sweeteners in foodstuffs is regulated, establishing maximum usable doses depending on the product, the development of analytical methodologies to separate and determine them in foods is necessary. CE has recently been used to determine the artiﬁcial sweetener sucralose, which is approximately 600 times sweeter than sugar. In this work, the CE method, optimized chemometrically, underwent a complete in-house validation protocol for the qualiﬁcation and quantitation of sucralose in various foodstuffs (carbonated and alcoholic beverages, yogurts, and hard-boiled candy) [28].

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 413 16.7.2008 11:06pm Compositor Name: MSubramanian

Capillary Electrophoresis in Food Analysis

18.5.2 AMINO ACIDS

AND

413

RELATED COMPOUNDS

Amino acids found in food comprise the 20 protein amino acids, which are indicators of protein composition, and other nonprotein amino acids, which may be added to increase the functional properties of foodstuffs (i.e., carnitine is a nonprotein amino acid added to processed foods to favor the transformation of fat into energy). Since most amino acids do not possess a strong chromophore, they are frequently derivatized to achieve sensitive detection. Thus, they can be derivatized with adequate probes to enhance the sensitivity obtained by UV-Vis or ﬂuorescence detection. The larger molecules resulting from derivatization provide better sensitivity with ESI-MS (larger molecules can be ionized with a higher yield and higher molecular masses [>150 m=z] usually produce lower MS background noise) [29]. Although derivatization introduces an additional analytical step, derivatization also facilitates the chiral separation of amino acids. In fact, chiral separation of amino acids by CE with cyclodextrins is only possible after derivatization. Nevertheless, underivatized amino acids have also been analyzed by CE. The analysis of amino acids by CE including its application to food analysis has been reviewed in several reviews from 1999 to 2006 [30–34]. Most applications were focused on the determination of protein amino acids in protein hydrolysates, fodder and raw materials, and mainly beverages [24,30–34]. Only very few applications on nonprotein amino acid analysis have been reported in food samples, an example of which is the determination of levodopa in beans [24]. Nevertheless, the most recent applications of CE to the analytical determination of amino acids by CE in foods include the analysis of carnitine and domoic acid (two nonprotein amino acids) in food supplements by CE with direct UV detection [35,36]. Amino acids can be subjected to bacterial=enzymatic degradation in some foods, leading to the formation of biogenic amines, excessive consumption of which can be toxic. Biogenic amines present in foods, primarily as a consequence of microbial amino acid decarboxylation, have been analyzed by CE as derivatized and underivatized compounds [23,24]. In a recent application, biogenic amines, trimethylamine, putrescine, cadaverine, spermine, tryptamine, spermidine, phenylethylamine, and tyramine, have been determined after their derivatization in solid food samples (ﬁsh, meat, and sausages) using a fully automated method based on pervaporation coupled on-line with CE and indirect UV detection [37].

18.5.3 PEPTIDES

AND

PROTEINS

Food proteins are vital nutrients that provide the amino acid building blocks from which the human body synthesizes its own proteins. The fragmentation of proteins into two or more amino acids joined by peptidic links forms peptides. Because of their functional properties, proteins are being used as food ingredients to modify and stabilize processed foods. In addition, it is interesting to study proteolysis products, peptides, to obtain information about degradation, treatment (in cured products proteolysis occurs), organoleptic properties, and of ﬁnal products [23]. These are some of the reasons for the analysis of proteins and peptides as a key area in food analysis. The two main application areas of CE in food proteins and peptides are the analysis of proteins from milk and other dairy products and the analysis of proteins from cereals, mainly wheat proteins [23,24]. Most of these applications are collected in recent reviews [23,24,38]. More recently, the analysis of whey proteins by CE with on-capillary derivatization and laser-induced ﬂuorescence detection has been performed [39,40]. This methodology allowed enough sensitivity for the analysis of trace amounts of b-lactoglobulin in commercial hypoallergenic formulas, the total content of this allergenic protein being useful to estimate the potential allergenicity of these formulas [40]. On the other hand, CE has allowed the characterization of fresh cheeses with respect to the manufacturing process and the source of milk by using the casein fraction of the end product [41]. In addition, a CE method with UV detection at 254 nm has been recently developed for the quantitation of soybean proteins in commercial gluten-free dietary products (bread and biscuits) elaborated with soybean protein and soybean ﬂour and rice ﬂour. The identiﬁcation of soybean proteins in a control sample (a rice

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 414 16.7.2008 11:06pm Compositor Name: MSubramanian

414

Handbook of Food Analysis Instruments 10

8

mAU (254 nm)

mAU (254 nm)

6 s + 2% SPI 4 2

M2

s

M1

0

0 0

(a)

5 M3

5

10 15 t (min)

20

25

0 (b)

5

10 t (min)

15

20

FIGURE 18.3 Electropherograms corresponding (a) to a sample (commercial product S: rice biscuit, 200 mg=mL) nonspiked and spiked with soybean protein isolate (SPI) and (b) to three commercial gluten-free dietary products (biscuits M1, M2, and M3) containing soybean protein and soybean ﬂour and rice ﬂour (at 200 mg=mL each). CE conditions: uncoated capillary with 75 mm ID (375 mm OD) and 58.5 cm total length and 50 cm effective length; separation buffer, 80 mM borate buffer at pH 8.5 containing 20% ACN; run voltage, 15 kV; temperature, 258C; injection by pressure of 50 mbar during 4 s of sample solution followed by 50 mbar during 4 s of running buffer. UV detection at 200 5 nm. Arrows mark the peak corresponding to soybean proteins. (From García-Ruiz, C. et al., Electrophoresis, 27, 452, 2006. With permission.)

biscuit) spiked with soybean proteins (see Figure 18.3a) as well as the electropherograms corresponding to three different soybean–rice biscuits analyzed, indicating the peak due to soybean proteins considered for quantitation (Figure 18.3b), illustrate the analysis of proteins in food samples [42]. Other applications have also been developed [24]. Thus, the differentiation of lentil cultivars from false lentil species (i.e., single-ﬂour vetch and common vetch) [43] as well as the differentiation of soybean from other seeds also commercialized as soybeans [44] based on CE patterns of the extracts have recently been reported.

18.5.4 PHENOLIC COMPOUNDS Phenolic compounds naturally occur in a variety of plant-derived foods and beverages and constitute a large group of secondary plant metabolites ranging from simple compounds to more complex oligomeric and highly polymerized structures. Their analysis in foods is important due to their contribution to the color, taste, and ﬂavor characteristics of foods [23]. In addition, the favorable biological properties of these compounds (namely as antioxidative, anti-inﬂammatory, antihistaminic, and=or antitumorigenic, and as free-radical scavengers and protectors against cardiovascular diseases) make them very important components in human diets. Analysis of polyphenolic compounds is one of the main applications of CE in food analysis. Thus, different polyphenolic compounds have been analyzed in grapes and wines, plants (including spices and medicinal herbs), different teas, olive oils, berries, fruits, soy, algae, and chocolate [24]. The monitoring of anthocyanins (plant polyphenols widely studied due to their positive effects on health) in wine and wine must samples has been performed by CE with MS detection after optimization of the MS setup. The study of the fragmentation of common anthocyanins was made in detail, with focus on the fragmentation of the anthocyanidin skeleton. The optimized method was employed for the monitoring of changes in anthocyanin proﬁle in red wines as well as the process of release of anthocyanins to wine must [45]. On the other hand, phenolic compounds in virgin olive oil have been analyzed by CE with UV detection (see the identiﬁcation of eight phenolic compounds in Figure 18.4). Sample extraction was

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 415 16.7.2008 11:06pm Compositor Name: MSubramanian

415

Capillary Electrophoresis in Food Analysis 6, 7

60,000

4

Absorbance (µAU)

40,000

Phenolic acid zone 1 20,000 2 5 8 3 0 10

15

20

25

30

35 40 Time (min)

45

50

55

60

FIGURE 18.4 Electropherogram of the polyphenolic fraction obtained from an extra-virgin olive oil sample (Picual variety). CE conditions: uncoated capillary with 75 mm ID (375 mm OD) and 110 cm total length and 100 cm effective length; separation buffer, 30 mM borate buffer at pH 9.30; run voltage, 25 kV; temperature, 258C; injection by pressure of 0.5 psi (~34 mbar) during 8 s. UV detection at 214 nm. Identiﬁcation of peaks: 1, tyrosol; 2, (þ)-pinoresinol; 3, 1-(þ)-acetoxypinoresinol; 4, deacetoxy oleuropein aglycon; 5, ligstroside aglycon; 6, hydroxytyrosol; 7, oleuropein aglycon; 8, elenolic acid. (From Gómez Caravaca, A.M. et al., Electrophoresis, 26, 3538, 2005. With permission.)

necessary to obtain the polyphenolic fraction of virgin olive oil obtaining good results when solidphase and liquid–liquid extractions were performed [46]. In addition, some polyphenols (rutin, chlorogenic acid, and quercetin) have been determined as pharmacologically active ingredients in sweet potato (Ipomoea batatas L) by CE with electrochemical detection. In this work, the method developed was appropriate for the analysis of different parts of the sample as well as of the sample under different cooking conditions [47].

18.5.5 DNAS The use of DNA as an analytical target in food analysis is becoming of great importance because this enables to obtain highly speciﬁc biological information. Polymerase chain reaction (PCR) is of great importance for analytical purposes since it allows exponential ampliﬁcation of selected DNA sequences with high degrees of sensitivity and speciﬁcity. PCR-based techniques combined with CE separation have been demonstrated to be powerful analytical methods for the detection of genetically modiﬁed organisms (GMOs), also called transgenic foods, detection of food-borne pathogens, food-spoilage bacteria, and species identiﬁcation [24,48]. Recently, the analysis of DNA segments (50–750 bp) by CE with LIF detection has successfully been used for the detection of genetically modiﬁed organisms (GMOs) in soybean and corn [49].

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 416 16.7.2008 11:06pm Compositor Name: MSubramanian

416

Handbook of Food Analysis Instruments

18.5.6 CARBOHYDRATES Carbohydrates range from discrete molecules of low-molecular weight (mono-, di-, and trisaccharides, deoxysugars, sugar acids, and sugar alcohols) to polymers (polysaccharides). Their qualitative and quantitative distribution in foods (fruits, vegetables, honey, and others) is essential because they are involved in the ﬂavor, maturity, quality, authenticity, and storage conditions (sugars in their raw state diminish rapidly during storage at ambient temperature). As a consequence, the determination of carbohydrates in the food industry is very relevant [50]. In carbohydrate analysis by CE, several aspects have to be taken into account. One is that the pKa of neutral sugars and sugar alcohols are high (between 12 and 14), which necessitates the use of strong alkaline separation media or derivatizing agents producing charged derivatives to be analyzed in CZE or the use of an EKC separation principle [51]. The other is that carbohydrates lacking chromophore=ﬂuorophore groups require the use of indirect optic detection, direct optic detection of the derivatized carbohydrates, or alternative detection systems such as electrochemical or MS [52]. Main applications of CE to the analysis of carbohydrates in food samples are directed toward process control, product development, and quality control of the ﬁnished product such as the determination of authenticity and possible adulteration. These applications are described in detail in three recent review papers [23,24,50].

18.5.7 VITAMINS Vitamins are a group of substances essential for normal metabolism, growth and development, and regulation of cell function [53]. Applications of CE to the analytical determination of vitamins have focused on the determination of water-soluble vitamins [23,24]. Main applications deal with the determination of ﬂavins, which are native ﬂuorescent compounds, by using LIF detection. This detection type allows to achieve sensitivity and selectivity, which are two requirements in the determination of vitamins at trace levels in complex food matrices [23,24]. In addition, other watersoluble vitamins such as vitamins B5, B6, and C have also been determined in different food samples by CE with UV detection [24]. Recently, vitamin C has been determined as a pharmacologically active ingredient in sweet potato using CE with electrochemical detection. The content determined in pulp was higher than in peel whereas the content determined after cooking was lower [47]. Folic acid, which is a water-soluble vitamin of the B-complex group, has been analyzed in apple juices by CE with chemiluminescence detection. This method was based on the enhanced effect of folic acid on the chemiluminescence reaction between luminol and BrO in alkaline aqueous solution. For this, a running buffer of borate at pH 9.4 containing luminol and an oxidizer solution of NaBrO in a carbonate buffer solution at pH 12.0 were used. The determination of folic acid under these conditions was achieved in less than 20 min with a detection limit of 2.0 108 M [54].

18.5.8 SMALL ORGANIC

AND INORGANIC IONS

Small ions of interest determined by CE in food samples comprise mainly small organic or inorganic anions. Thus, low-molecular weight organic acids such as fumaric, malic, oxalic, pyroglutamic, etc., in beer and coffee samples [55,56] and inorganic anions such as ﬂuoride, chloride, bromide, nitrite, nitrate, sulfate, and phosphate in water samples [57] have been analyzed. However, the number of small cations analyzed till now is less; vanadium species in water samples can be cited as an example [58]. The determination of small organic acids in foods is important due to the contribution of these molecules to the organoleptic properties of foods. Inorganic ions in foods are usually determined for quality and safety reasons. There are two main problems that have to be solved when analyzing small ions by CE: (1) poor or null UV absorptivity of many of these species and (2) peak broadening induced by electrophoretic dispersion that in many cases ruins the CE separation. Although research focused on solving

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 417 16.7.2008 11:06pm Compositor Name: MSubramanian

Capillary Electrophoresis in Food Analysis

417

the sensitivity problem using indirect UV detection, alternative detectors to UV detectors, as well as derivatization before the UV detection have been published [24], UV detection at very low wavelengths has also been used. Thus, the determination of organic acids (formic, tartaric, malic, succinic, maleic, glutaric, pyruvic, acetic, lactic, citric, butyric, benzoic, sorbic, ascorbic, and gluconic acids) in several beverage samples has recently been achieved using direct UV detection at 185 nm [59].

18.5.9 ORGANIC CONTAMINANTS Food contaminants are deﬁned in the European Union legislation as any substance not intentionally added to food, which is present in food as a result of the production, manufacture, processing, preparation, treatment, packaging, transport, or holding of such food, or as a result of environmental contamination [60]. Organic contaminants in foods can be chemicals (fertilizers, insecticides, herbicides, fungicides and other pesticides, growth stimulants, antibiotics, and pollutants) or biologicals (biological toxins and microorganisms such as bacteria, fungi, and algae) [61]. Foods must be free of organic contaminants because they can be dangerous for human health even at very low concentrations. To ensure human food safety, maximum residue limits for many of these residues in food products have been deﬁned [24]. Applications developed in food analysis by CE are mainly focused on the determination of pesticide residues and antibiotics. Other applications deal with the analysis of bacterial toxins. Very few research has been devoted to determine food-borne pathogens and toxin-producing microorganisms [61]. The main drawback reported for the CE determination of organic contaminants in food samples is its inadequate sensitivity for trace analysis of organic contaminants. Nevertheless, this limitation has been overcome by using on-line preconcentration techniques (stacking) and=or using speciﬁc detectors such as LIF and MS detectors [61]. In a recent work, CE has been used for analyzing pesticides (ﬂutriafol, cyproconazole I, cyproconazole II, miclobutanil, tebuconazole, acrinathrin, bitertanol, ﬂudioxinil, and pyriproxyfen) in fruits (grape and strawberry) and vegetables (lettuce and tomato) using solid-phase extraction and stir-bar sorptive extraction [62].

18.5.10 ENANTIOMERIC ANALYSIS

OF

CHIRAL COMPOUNDS

Chiral compounds are composed of stereoisomers (compounds made up of the same atoms, bonded by the same sequence of bonds, but possessing different three-dimensional structures), which are non-superposable mirror images. There are three different types of chiral compounds: those with one or more than one asymmetrical center, those with axial chirality, and asymmetrical cyclic compounds [63]. Except when axial chirality exists, where stereoisomers are called atropisomers, the stereoisomers of a chiral compound are named enantiomers. Analysis of the enantiomers present in foods is important because it enables to identify adulterations in food and beverages, to evaluate and identify the age, treatment, and storage effects on foods, to control and monitor fermentation processes and products, to evaluate ﬂavor and fragrance components, to ﬁngerprint complex mixtures, to analyze chiral metabolites of many chiral and prochiral constituents of foods and beverages, and to control those additives used as pure enantiomers [24]. Applications of chiral CE methods in food analysis reported till date are not much and they have been collected in different reviews [24,33,64] and in a book chapter [63]. Food samples such as beverages (fruit juices, nectars and concentrates, beer, wine, shake, and tea drinks), fruits (oranges and strawberries), vegetables (potatoes and onions), yogurts, candies, jams, ice cream, and pickles have been analyzed. Although the chiral separation of amino acids and peptides has been of special interest in food analysis by CE [33,65,66], other compounds such as organic acids, food colorants, ﬂavonoids, and pesticides have also been analyzed in foods [24,63].

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 418 16.7.2008 11:06pm Compositor Name: MSubramanian

418

Handbook of Food Analysis Instruments

In a very recent work, a CE method using LIF detection has been developed to analyze and quantitate the two enantiomers (L- and D-forms) of several chiral amino acids (derivatized with ﬂuorescein isothiocyanate, FITC) in different types of vinegars. The FITC–amino acid derivatives, which were identiﬁed by spiking the samples with the corresponding standard, enabled to characterize the vinegars by the different content of amino acids found in them. Figure 18.5 clearly shows

3000

3000

3

2

2800

A

2800

2600

2600 Zone II

Zone I

Zone III

2200

9

2000

2200

11

2000

3

B

1800 RFU

2400

5 7

1800 RFU

2400

1600

1600

1400

1400 1200

7

5

2

9

1200

11

1000

1000 800

800

C

600

600

3

400

400 200

2

5 7

200

9

11

0 7

8

9

10

11

12

13

14

15

16

17

18

19

0

20

21

Time (min) Zone II

1

2

A

3

4

Zone III

200

B

2

0 300

C

1

8.2

8.3

3

5

4

7

100 100

Time (min)

8.5

0

B

6

10 11

4

5

7

C

6 12.5

A

0 20

8

12.9

13.3

Time (min)

13.7

9

10 11

B

10 0 20

3

50 8.4

9

10

6

200 0

2

8

A

100

1

20

7

5

RFU

300

RFU

RFU

Zone I 40 30 20 10 0 30 20 10 0 30 20 10 0

10 0

9 8

17.2

11 C

10

18.0

18.8

19.6

Time (min)

FIGURE 18.5 FITC–amino acids in vinegars: (a) vinegar 1 (balsamic vinegar), (b) vinegar 5 (sherry vinegar), and (c) vinegar 12 (white wine with extracts of herbs). CE conditions: uncoated capillary with 50 mm ID (375 mm OD) and 57 cm total length and 50 cm effective length; separation buffer, 100 mM borate buffer containing 30 mM sodium dodecyl sulfate and 20 mM b-cyclodextrin (pH 9.7); run voltage, 20 kV; temperature, 258C; injection by pressure of 0.5 psi (~34 mbar) during 5 s. Fluorescence detection: 488 nm (excitation wavelength) and 520 nm (emission wavelength). Identiﬁcation of peaks: 1, D-Arg; 2, L-Arg; 3, L-Pro; 4, D-Pro; 5, g-amino butyric acid; 6, D-Ala; 7, L-Ala; 8, D-Glu; 9, L-Glu; 10, D-Asp; 11, L-Asp. (From Carlavilla, D. et al., Electrophoresis, 27, 2551, 2006. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 419 16.7.2008 11:06pm Compositor Name: MSubramanian

419

Capillary Electrophoresis in Food Analysis

that balsamic vinegar contained the highest levels of L- and D-amino acids in comparison with the other two samples (sherry and white wine with extracts of herb vinegars) [67].

18.6 FUTURE TRENDS Food applications using CE microchips are now emerging. In fact, food samples present complex matrices where selectivity is a very important challenge because the total integration of analytical steps into a microchip format is very difﬁcult [68]. The main advantage of using microchips instead of conventional CE systems is the considerable reduction in analysis time. Thus, the ﬁrst contributions that have recently appeared in the literature are based primarily on fast separations of food analytes of high signiﬁcance [68]. However, not many applications have been reported in the analysis of food samples using microchips. The applications developed have been devoted to the analysis of amino acids in green tea [69], DNA for the analysis of genetically modiﬁed organisms in soybeans [70], polyphenolic compounds in beverages such as wine [71,72], pulp and juices [73], food additives such as antioxidants [74], and preservatives [75] in several food samples, vitamins in tablets [75], and small ions such as inorganic cations and inorganic and organic anions in beverages [76] as well as organic acids in wines [77]. An example showing the possibilities of microchips in the analysis of foods is depicted in Figure 18.6. This ﬁgure displays the ultrafast analysis (~11 s) of genetically modiﬁed organisms (GMOs) in soybeans by microchip capillary gel electrophoresis using programmed ﬁeld strength gradients in a conventional glass double-T microchip. The separation mechanism in microchip capillary gel electrophoresis is based on differences in DNA size as analytes migrate through the pores of the gel-ﬁlled microchip, eluting in ﬁrst place the small DNA molecules (a 100 bp DNA ladder has been used). After PCR ampliﬁcation of samples, a 100 bp DNA fragment representing the CaMV 35S promoter from genetically modiﬁed soybeans and a 250 bp DNA fragment

450 250

100

RFU

Unit: bp

(b) GM-soybean 350 250 (c) Non GM-soybean

Electric field (V/cm, dotted line)

(a) 100 bp DNA ladder

100 200 300 500 400 600 800 1000 1500 2000 3000

0.5 RFU

550

50 µm 20 µm

85 mm 2 16 mm 8 mm

0

5

10

15

20

Migration time (s)

25

3 4

250

(a)

Detection

1

95 mm

30

(b)

FIGURE 18.6 (a) Separation of PCR products, 100 and 250 bp DNA fragments in microchip electrophoresis by programmed ﬁeld strength gradient. (b) Experimental conditions: microchip (Schott Boroﬂoat glass) with channels 50 mm wide and 20 mm deep and reservoirs with 2.0 mm diameter and 1 mm deep. The injection channel length was 8 mm, the separation channel was 85 mm long, and detection was performed at 15 and 20 mm from the injection-T; separation buffer, 1 TBE buffer (0.089 M Tris, 0.089 M borate, and 0.002 M EDTA) at pH 8.3 with 0.5mg=mL ethidium bromide; coating matrix, 0.5% polyvinylpyrrolidone (PVP, Mr 1,000,000); sieving matrix, 0.3% ethylene oxide (PEO, Mr 8,000,000); run voltage, 470.6 V=cm for 9 s, 294.1 V=cm for 1 s, 470.6 kV for 0.5 s, 294.1 V=cm for 1.5 s, and 470.6 V=cm for 20 s; injection by pressure of 0.48 kV during 60 s. Fluorescence detection: 532 nm (excitation wavelength) and 605 nm (emission wavelength). (From Kim, Y.J. et al., J. Chromatogr. A, 1083, 179, 2005. With permission.)

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 420 16.7.2008 11:06pm Compositor Name: MSubramanian

420

Handbook of Food Analysis Instruments

representing the lectin endogenous gene present in genetically modiﬁed and nongenetically modiﬁed soybeans were observed [70]. Moreover, the implementation of hyphenated techniques coupling CE to detection systems providing structural information (i.e., MS) in food analysis is expected as well as other analytical approaches with a high potential in food analysis such as the development of screening systems and the automation of sample treatment and calibration in CE instruments by using CFS units [15]. Finally, although 2D-CE instrumentation has been developed recently and reported works were mainly devoted to instrument development, it is expected that 2D-CE will be applied in the near future in the food analysis ﬁeld, especially for those samples with complex matrices.

ACKNOWLEDGMENTS The authors thank the Comunidad Autónoma de Madrid (Spain) for the research project S-0505=AGR=0312. C. García-Ruiz thanks the Ministry of Science and Technology (Spain) for her research contract with the Ramón y Cajal program (RYC-2003-001).

REFERENCES 1. Marina, M.L., Ríos, A., and Valcárcel, M. (Eds.), Analysis and Detection by Capillary Electrophoresis, CAC, Elsevier, Amsterdam, the Netherlands, 2005. 2. Pyell, U. (Ed.), Electrokinetic Chromatography. Theory, Instrumentation and Applications, Wiley, Chichester, United Kingdom, 2006. 3. Weinberger, R., Practical Capillary Electrophoresis, 2nd ed., Academic Press, San Diego, CA, 2000. 4. Camilleri, P. (Ed.), Capillary Electrophoresis: Theory and Practice, 2nd ed., Springer-Verlag, Heidelberg, Germany, 1998. 5. Heiger, D., High Performance Capillary Electrophoresis, Agilent Technologies, Waldbronn, Germany, 2000. 6. Quirino, J.P. and Terabe, S., J. Chromatogr. A, 856, 456–482, 1999. 7. http:==www.chem.agilent.com. 8. http:==www.beckman.com. 9. http:==www.princetechnologies.nl. 10. Mukhopadhyay, R., Anal. Chem., 78, 2109–2111, 2006. 11. http:==www.polymicro.com=products=products.htm. 12. Kim, J.B. and Terabe, S., J. Pharm. Biomed. Anal., 30, 1625, 2003. 13. Quirino, J.P. and Terabe, S., Science, 282, 465, 1998. 14. Quirino, J.P., Kim, J.-B., and Terabe, S., J. Chromatogr. A, 965, 357, 2002. 15. Castañeda, G., Rodríguez-Flores, J., and Ríos, A., J. Sep. Sci., 28, 915, 2005. 16. Sheng, L. and Pawliszyn, J., Analyst, 127, 1159, 2002. 17. Kvasnicka, F., Jaroš, M., and Gaš, B., J. Chromatogr. A, 916, 131, 2001. 18. Liu, S. and Guttman, A., Trends Anal. Chem., 23, 422, 2004. 19. Müller, A. and Steinhart, H., Food Chemistry. 20. Kvasnicka, F., J. Sep. Sci., 28, 813, 2005. 21. Fraizer, R.A., Ames, J.J., and Nursten, H.E., Electrophoresis, 20, 3156, 1999. 22. Fraizer, R.A., Electrophoresis, 22, 4197, 2001. 23. Frazier, R.A. and Papadopoulou, A., Electrophoresis, 24, 4095, 2003. 24. Cifuentes, A., Electrophoresis, 27, 283, 2006. 25. Huang, H.Y. et al., Electrophoresis, 26, 867, 2005. 26. Law, W.S. et al., Electrophoresis, 26, 4648, 2005. 27. Hernández-Borges, J. et al., Chromatographia, 62, 271, 2005. 28. McCourt, J., Stroka, J., and Anklam, E., Anal. Bioanal. Chem., 382, 1269, 2005. 29. Simó, C. et al., Electrophoresis, 26, 1432, 2005. 30. Smith, J.T., Electrophoresis, 20, 3078, 1999. 31. Prata, C. et al., Electrophoresis, 22, 4129, 2001. 32. Poinsot, V., Bayle, C., and Couderc, F., Electrophoresis, 24, 4047, 2003.

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 421 16.7.2008 11:06pm Compositor Name: MSubramanian

Capillary Electrophoresis in Food Analysis 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63.

64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77.

421

Poinsot, V. et al., Electrophoresis, 27, 176, 2006. Peace, R.W. and Gilani, G.S., J. AOAC Int., 88, 877, 2005. Kvasnicka, F., Ševcík, R., and Voldrich, M., J. Chromatogr. A, 1113, 255, 2006. Prokorátová, V. et al., J. Chromatogr. A, 1081, 60, 2005. Ruiz-Jiménez, J. and Luque de Castro, M.D., J. Chromatogr. A, 1110, 245, 2006. Dolnik, V., Electrophoresis, 27, 126, 2006. Veledo, M.T., de Frutos, M., and Díez-Masa, J.C., J. Sep. Sci., 28, 935, 2005. Veledo, M.T., de Frutos, M., and Díez-Masa, J.C., J. Sep. Sci., 28, 941, 2005. Miralles, B., Ramos, M., and Amigo, L., Milchwissenschaft, 60, 278, 2005. García-Ruiz, C. et al., Electrophoresis, 27, 452, 2006. Piergiovanni, A.R. and Taranto, G., J. Agric. Food Chem., 53, 6593, 2005. García-Ruiz, C., García, M.C., and Marina, M.L., Electrophoresis, 28, 2314, 2007. Bednar, P. et al., J. Sep. Sci., 28, 1291, 2005. Gómez Caravaca, A.M. et al., Electrophoresis, 26, 3538, 2005. Guan, Y. et al., J. Agric. Food Chem., 54, 24, 2006. García-Cañas, V., González, R., and Cifuentes, A., Trends Anal. Chem., 23, 637, 2004. Sanchez, L., González, R., Crego, A.L., and Cifuentes, A., J. Sep. Sci., 30, 579, 2007. Martínez Montero, C. et al., Chromatographia, 59, 15, 2004. Corradini, C. and Cavazza, A., Ital. J. Food Sci., 10, 299, 1998. Campa, C. et al., Electrophoresis, 27, 2027, 2006. http:==www.nlm.nih.gov=medlineplus=ency=article=002399.htm#Deﬁnition. Zhao, S.L. et al., J. Chromatogr. A, 1107, 290, 2006. Cortacer-Ramírez, S. et al., J. Chromatogr. A, 1064, 115, 2005. Galli, V. and Barbas, C., J. Chromatogr. A, 1032, 299, 2004. Negro, A., Paz, E., and Tabanal, B., J. Liq. Chromatogr. Rel. Technol., 26, 709, 2003. Zu-Liang, C. and Ravendra, N., Anal. Bioanal. Chem., 374, 520, 2003. Mato, I. et al., Anal. Chim. Acta, 565, 190, 2006. European Union, Council Directive 92=59=EEC of 29 June 1992 on General Product Safety, Ofﬁcial Journal L228, 11=08=1992, P0024–0032, Brussels, Belgium, 2005. Juan-García, A., Font, G., and Picó, Y., J. Sep. Sci., 28, 793, 2005. Juan-García, A., Picó, Y., and Font, G., J. Chromatogr. A, 1073, 229, 2005. García-Ruiz, C. and Marina, M.L., Chiral analysis by CE, in Marina, M.L., Ríos, A., and Valcárcel, M. (Eds.), Analysis and Detection by Capillary Electrophoresis, CAC, Elsevier, Amsterdam, The Netherland, 2005, Chapter 13. Simó, C., Barbas, C., and Cifuentes, A., Electrophoresis, 24, 2431, 2003. Scriba, G.K.E., Electrophoresis, 24, 4063, 2003. Wan, H. and Blomberg, L.G., J. Chromatogr. A, 875, 43, 2000. Carlavilla, D. et al., Electrophoresis, 27, 2551, 2006. Escarpa, A. et al., Electrophoresis, 28, 1002, 2007. Kato, M. et al., J. Chromatogr. A, 1013, 183, 2003. Kim, Y.J. et al., J. Chromatogr. A, 1083, 179, 2005. Scampicchio, M. et al., J. Chromatogr. A, 1049, 189, 2004. Crevillen, A.G. et al., Anal. Chim. Acta, 562, 137, 2006. Blasco, A.J. et al., Electrophoresis, 26, 4664, 2005. Ding, Y.S., Mora, M.F., and Garcia, C.D., Anal. Chim. Acta, 561, 126, 2006. Law, W.S. et al., Electrophoresis, 26, 4648, 2005. Kuban, P. and Hauser, P.C., Electrophoresis, 26, 3169, 2005. Masar, M. et al., J. Sep. Sci., 28, 905, 2005.

Otles/Handbook of Food Analysis Instruments 45660_C018 Final Proof page 422 16.7.2008 11:06pm Compositor Name: MSubramanian

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 423 17.7.2008 5:08pm Compositor Name: VAmoudavally

Electrophoresis 19 Gel in Food Analysis Reiner Westermeier and Burghardt Scheibe CONTENTS 19.1 19.2

Introduction ........................................................................................................................ 424 Principles ........................................................................................................................... 424 19.2.1 Analytes ............................................................................................................... 425 19.2.1.1 Nucleic Acids ...................................................................................... 425 19.2.1.2 Proteins ................................................................................................ 425 19.2.1.3 Dyes .................................................................................................... 425 19.2.2 Separation ............................................................................................................ 425 19.2.2.1 Gels ..................................................................................................... 425 19.2.2.2 Buffers ................................................................................................. 426 19.2.3 Detection .............................................................................................................. 427 19.2.3.1 Staining ............................................................................................... 427 19.2.3.2 Blotting ................................................................................................ 427 19.2.4 Relative Quantiﬁcation ........................................................................................ 428 19.3 Deﬁnitions .......................................................................................................................... 428 19.3.1 Gel Compositions ................................................................................................ 428 19.3.1.1 Agarose Gels ....................................................................................... 428 19.3.1.2 Polyacrylamide Gels ........................................................................... 428 19.4 Theory of Gel Electrophoresis .......................................................................................... 428 19.4.1 Zone Electrophoretic Techniques in Gels ........................................................... 428 19.4.1.1 Homogeneous and Discontinuous Systems ........................................ 429 19.4.1.2 Native Conditions ............................................................................... 429 19.4.1.3 SDS Electrophoresis............................................................................ 429 19.4.1.4 Immuno Electrophoresis ..................................................................... 429 19.4.2 Isoelectric Focusing ............................................................................................. 430 19.4.3 Two-Dimensional Electrophoresis ....................................................................... 430 19.5 Instrumentation .................................................................................................................. 431 19.5.1 Power Supply ....................................................................................................... 431 19.5.2 Vertical Electrophoresis Systems ........................................................................ 431 19.5.2.1 Gel Rod Apparatus .............................................................................. 431 19.5.2.2 Slab Gel Apparatus ............................................................................. 432 19.5.3 Horizontal Flatbed Electrophoresis Systems ....................................................... 433 19.5.3.1 Submarine Apparatus for DNA .......................................................... 433 19.5.3.2 Standard Size Apparatus for Thin-Layer Electrophoresis .................. 433 19.5.3.3 Mini-Format Automated Electrophoresis System ............................... 433

423

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 424 17.7.2008 5:08pm Compositor Name: VAmoudavally

424

Handbook of Food Analysis Instruments

19.5.4 19.5.5

Systems for High-Resolution Two-Dimensional Electrophoresis ....................... 434 Blotting Systems .................................................................................................. 434 19.5.5.1 Tank Blotters ....................................................................................... 434 19.5.5.2 Semidry Blotters.................................................................................. 435 19.5.6 Automatic Staining Apparatus ............................................................................. 435 19.5.7 Imagers ................................................................................................................. 435 19.5.7.1 Still CCD Cameras .............................................................................. 436 19.5.7.2 White Light Scanners .......................................................................... 436 19.5.7.3 Laser Scanner ...................................................................................... 436 19.6 Applications to Food Analysis .......................................................................................... 436 19.6.1 Analysis of Nucleic Acids ................................................................................... 436 19.6.1.1 Genetic Differentiation ........................................................................ 436 19.6.1.2 Detection of Genetically Modiﬁed Food ............................................ 437 19.6.2 Analysis of Proteins ............................................................................................. 437 19.7 Future Trends ..................................................................................................................... 437 References ..................................................................................................................................... 437

19.1 INTRODUCTION Electrophoretic techniques in gels have been used in food analysis for various tasks, like cultivar and species control of raw material, for developing and monitoring food preservation and technology processes, and uncovering illegal mixtures or addition of certain components in food products. They can be used for both qualitative and quantitative aspects. Depending on the analytical task and the substrate to be analyzed, different separation methods are employed, such as basic or acidic zone electrophoresis, SDS electrophoresis, and isoelectric focusing. Gel electrophoretic techniques do not require a high instrumental effort and they are relatively easy to carry out.

19.2 PRINCIPLES Electrophoresis is based on the migration of charged particles and molecules in an electric ﬁeld. Therefore it can only be applied for the analysis of chargeable components such as nucleic acids, proteins, peptides, and some dyes. The main area of application is the analysis of macromolecules. Different component homologues are separated, because they migrate with different velocities. Their ‘‘electrophoretic mobilities’’ differ due to the size and net charge heterogeneity of the molecules. Generally gel electrophoresis is performed in compact agarose or polyacrylamide gel media serving as anticonvective medium. The buffers are composed of charged compounds providing a certain pH value and buffering power. Because also the charged buffer compounds will migrate in the electric ﬁeld, sufﬁciently sized buffer reservoirs are required on both ends of the gel. The electric ﬁeld is induced through platinum electrode wires which are placed into these buffer tanks. The transport of the electric ﬁeld is carried via the migrating buffer ions, which causes the development of Joule heat. Therefore most systems for gel electrophoresis are equipped with a cooling device. For detection of the separated zones, the gels are submitted to a series of staining or development solutions. In some exceptional cases prestained samples are applied on a gel, which can be detected directly. When highly complex protein mixtures have to be analyzed, two different separation methods are combined to a high resolution two-dimensional electrophoresis. Immuno electrophoresis allows the detection of speciﬁc proteins by forming precipitates with polyclonal antibodies in the gel. With blotting the separated macromolecules can be transferred onto the surface of an immobilizing membrane, where they are easily accessible for probing by DNA hybridization, with speciﬁc ligands or speciﬁc antibodies.

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 425 17.7.2008 5:08pm Compositor Name: VAmoudavally

Gel Electrophoresis in Food Analysis

425

19.2.1 ANALYTES 19.2.1.1

Nucleic Acids

DNA molecules can be ampliﬁed by the PCR process. Due to this unique feature, a low sample amount is never an issue, even traces can be efﬁciently detected. In alkaline environments all DNA molecules are negatively charged with constant charge densities. The migration distances of DNA molecules in agarose or polyacrylamide gels are proportional to their sizes. DNA molecules can be separated under native or denatured conditions in basic gels. 19.2.1.2

Proteins

In food analysis gel, electrophoretic methods are mainly applied for protein separations and detection. Proteins and peptides are amphoteric molecules, therefore they can be analyzed as anions or cations using basic or acidic buffers, respectively. A very high resolution is achieved with isoelectric focusing in agarose or polyacrylamide gels, where proteins and peptides are migrating to their isoelectric points. After separation under native conditions proteins can be detected with bioactivity staining, like zymogram techniques. Because some proteins are hydrophobic, there are sometimes issues with limited solubility. Another challenge is the existence of protein complexes in the sample. Therefore it might be necessary to separate proteins under denaturing conditions in presence of SDS or urea. However, with denaturation most of the species- and cultivar-differentiating features are lost. 19.2.1.3

Dyes

Gel-electrophoretic analysis of dyes is a rather exotic application. However, many dyes show charge and size differences, which are high enough for electrophoretic separations. Some dyes, like antocyans, are amphoteric and can be separated by isoelectric focusing. Because of the relatively low molecular weights of dyes, the best technique is running them in very thin gel layers and drying the gel matrix quickly down for ﬁxation.

19.2.2 SEPARATION 19.2.2.1

Gels

Gel-electrophoretic methods are performed in compact agarose or polyacrylamide gel media, in contrast to granulated gels typically used in chromatography. Mostly polyacrylamide gels are employed, because they have a high mechanical stability, they are clear and transparent, and chemically inert. They can be polymerized in the laboratory from acrylamide monomers and a crosslinker, mostly N,N0 -methylenebisacrylamide, in aqueous solution with a catalyst system via radical initiation. Polymerization is performed in glass rods or glass cassettes in absence of oxygen. The sieving properties of the gel can be modiﬁed by changing the monomer concentration. Typical gel concentrations vary from 4% to 16% acrylamide monomer concentration per gel volume. Slab gels are sometimes copolymerized with a plastic ﬁlm to improve the mechanical stability further. In special cases porosity gradient gels are employed to increase the separation size range and the sharpness of zones. For sample applications, little wells are formed with a mould or a plastic comb. Figure 19.1B and C shows gel casting setups for both vertical and horizontal gel systems. Acidic gels are more complicated to polymerize than basic gels, because the polymerization reaction is pH dependent. Various suppliers offer a wide range of ready-made gels of all sizes and gel and buffer compositions. Polyacrylamide gels can also be prepared with a builtin pH gradient. Agarose gels have larger pore sizes than polyacrylamide gels. They are more opaque and brittle. Agarose gels are prepared by boiling agarose powder in water. The gel is formed by chilling the

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 426 17.7.2008 5:08pm Compositor Name: VAmoudavally

426

Handbook of Food Analysis Instruments

(A)

(B)

(C)

FIGURE 19.1 Gel casting setups. (A) For submarine agarose gels. (B) and (C) For polyacrylamide gels. (B) Gel casting assembly for vertical systems containing glass plates and spacers, clamps and casting stand, and the comb for forming sample wells. (C) Gel casting assembly for horizontal gels with glass plates, one containing a U-shaped gasket, and clamps.

agarose sol. A gel typically contains 0.6% to 1% agarose. The main application ﬁelds of agarose gels are the analysis of large DNA molecules, immuno electrophoresis, and separating proteins with isoelectric focusing under native conditions. Figure 19.1A shows the setup for pouring an agarose gel for submarine electrophoresis. Agarose contains remains of agaropectin, coming from the raw material, red seaweed. These compounds are incorporated in the gel matrix and become negatively charged in basic buffers and carrier ampholytes. Isoelectric focusing requires premium agarose quality with a very low amount of these moieties, optimally with mr ¼ 0. 19.2.2.2

Buffers

The composition of the buffer depends on the substrate to be analyzed and the selected method. For nucleic acid separations the standard buffer consists of Tris-borate-EDTA. Because protein molecules are more complex, a number of different buffers are used in practice: .

. .

For general protein separation, a discontinuous buffer with Tris-chloride in the gel and Tris-glycine in the buffer reservoir with or without presence of the anionic detergent sodium dodecylsulfate (SDS) For the separation of alcohol-soluble proteins (prolamines) an acidic buffer made up from glycine and acetic acid For isoelectric focusing a mixture of carrier ampholytes, which is a heterogeneous composition of amphoteric buffers exhibiting a pH spectrum from 3 to 10

When proteins with poor solubility or very complex protein mixtures are analyzed, high concentrations of urea (like 8 mol=L) are added to the sample during extraction and to the gel. There are ready-to-use gels and buffer kits on the market, which contain alternative buffer compositions. Some of these buffers are modiﬁed to achieve a pH value below 7 to prevent alkaline

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 427 17.7.2008 5:08pm Compositor Name: VAmoudavally

Gel Electrophoresis in Food Analysis

427

hydrolyzation of the matrix to extend the shelf life polyacrylamide gels. In other cases, amphoteric buffers are employed instead of classic buffers, because those allow quicker separations and sharper zones. Small molecules like peptides and dyes are usually separated in simple sodium phosphate buffers.

19.2.3 DETECTION Usually the separated molecules are not visible, and the direct detection with ultraviolet light is not sensitive enough. Therefore the zones are stained in the gel matrix or they are transferred onto an immobilizing membrane for further speciﬁc analysis. 19.2.3.1

Staining

DNA molecules are either stained with intercalating ﬂuorescent dyes like ethidium bromide or with a multistep silver staining procedure. The latter technique shows a higher sensitivity, the bands are directly visible without a UV table or ﬂuorescent scanner; it is much less harmful, but can only be reasonably well performed in polyacrylamide gels. Alternatively, ﬂuorescent labeled primers can be employed for PCR ampliﬁcation, allowing direct detection in the gel. The classical method for protein detection is staining with Coomassie Brilliant Blue. It can be used for quantitative assessments in a limited concentration interval. When higher sensitivity is required, silver staining is usually employed. But silver staining is a multistep procedure and is less appropriate for quantiﬁcation due to its narrow dynamic range. Novel ﬂuorescent dyes have been introduced to the market during the last few years like SYPRO Ruby, Flamingo Pink, and Deep Purple, which show sensitivity close to silver staining and a very wide dynamic range of four orders of magnitude. Most of them are based on heavy metals; solely Deep Purple is a natural product, produced by a fungus, and easy to dispose of. After separation under native conditions, enzymes and enzyme inhibitors can be speciﬁcally detected in the gel with activity staining, the so-called zymogram technique. The gel is placed into substrate containing buffer, the reaction is coupled to a color reaction producing a poorly soluble visible dye product. 19.2.3.2

Blotting

Blotting methods were ﬁrst introduced for hybridization of DNA fragments after agarose gel electrophoresis, known under the name ‘‘Southern Blotting.’’ Soon the technique has been extended to speciﬁcally detect proteins separated by SDS electrophoresis, termed ‘‘Western Blotting.’’ While DNA molecules are usually carried over to the membrane by capillary forces, proteins require more active transfer conditions, like electrophoretic force after zone electrophoresis or pressure after isoelectric focusing. The most frequently used membranes consist either of nitrocellulose or PVDF. Nylon membranes are only used for nucleic acids. Because the membranes bind all macromolecules nonspeciﬁcally, the nonoccupied surface has to be blocked by a substrate, which does not interfere with the subsequent staining process, like bovine serum albumin, Tween 20, skim milk powder solution, or ﬁsh gelatine. For proteins the blotting method works best after SDS gel electrophoresis or denaturing isoelectric focusing, because then the polypeptides are completely unfolded, and the epitopes are readily accessible for the primary antibody. After the primary antibody is attached to the respective antigen zones, the membrane is washed, and subsequently probed with a ‘‘labeled’’ secondary antibody. This label for detection is mostly a conjugate of biotin for ampliﬁcation with streptavidin or an enzyme, like horseradish peroxidase or alkaline phosphatase for very sensitive detection with cascade reactions leading to enhanced chemiluminescence (ECL).

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 428 17.7.2008 5:08pm Compositor Name: VAmoudavally

428

Handbook of Food Analysis Instruments

19.2.4 RELATIVE QUANTIFICATION It has already been mentioned above that gel electrophoresis is often applied for quantitative measurements of additions to food products. This is usually performed via characteristic indicator proteins. Because different proteins show different binding afﬁnities to the dye, it is impossible to perform absolute quantiﬁcation. Relative quantiﬁcation can work very reliably, when sample preparation, sample application, and the entire electrophoresis procedure have been carried out with utmost accuracy. The substance to be measured has to be applied on the same gel in a dilution series. Coomassie staining is adequate, however it must be a methodical version without an alcoholic solution containing background destaining step. Hot staining with acetic acid or colloidal staining works very well. The measurement of the zones has to be performed with a calibrated densitometer or scanner over the entire width of the zones.

19.3 DEFINITIONS 19.3.1 GEL COMPOSITIONS 19.3.1.1

Agarose Gels

Concentration: the pore size is determined by the concentration of agarose: x% (w=v) agarose. Electroendosmosis: mr (should be <0.10) 19.3.1.2

Polyacrylamide Gels

Total acrylamide concentration: T¼

(a þ b) 100 % V

Crosslinking factor: C¼

b 100 % aþb

where a is the mass of acrylamide in grams b is the mass of N,N0 -methylenebisacrylamide in grams V is the volume in milliliters Relative electrophoretic mobility: Migration distance of a zone related to the migration distance of the fast migrating tracking dye, usually Bromophenol Blue. Isoelectric point: The pH value where the positive and negative side charges of a protein, peptide, amino acid, or amphoteric dye are equal, the net charge is zero. O.D.: The unit O.D. for the optical density is mostly used in biology and biochemistry and is deﬁned as follows: 1 O.D. is the amount of the substance, which has an absorption of 1 when dissolved and measured in 1 mL in a cuvette with a thickness of 1 cm.

19.4 THEORY OF GEL ELECTROPHORESIS More detailed descriptions of the theoretical background of gel electrophoresis are found in the monography by Westermeier [1].

19.4.1 ZONE ELECTROPHORETIC TECHNIQUES

IN

GELS

In an electrical ﬁeld charged molecules migrate towards the direction of the electrode with the opposite sign. The gel matrix exhibits a certain retardation effect on the molecules, depending on

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 429 17.7.2008 5:08pm Compositor Name: VAmoudavally

Gel Electrophoresis in Food Analysis

429

the gel pore size. Because of their varying charges and masses, different compounds of a mixture migrate at different velocities and will thus be separated into single zones. Zone electrophoresis techniques are not end point methods, thus the separation has to be stopped at a deﬁned time point: usually when the tracking dye Bromophenol Blue reaches the end of the gel. The result is a band pattern which looks like a bar code. 19.4.1.1

Homogeneous and Discontinuous Systems

The plainest setup employs a homogeneous buffer and gel structure. It can easily be applied on nucleic acids and peptides. However, protein mixtures exhibit a more complicated behavior and require often a more sophisticated system. A critical step is the sample entry of complex protein mixtures into a gel with narrow pore sizes. During a fast transition from the liquid phase into the gel medium protein mixtures experience a sudden increase in concentration, which can lead to partial or partly aggregation, resulting in precipitation on the gel surface. As a remedy for protein separations, very frequently discontinuous electrophoresis techniques are employed. Several discontinuities are combined: . . .

Gel is divided into a large porous stacking gel and tightly porous resolving gel. Gel and the running buffer contain different ions: a highly mobile ion like chloride and a lowly mobile ion like glycine, respectively. Stacking gel buffer has a different pH and ionic strength than the resolving gel buffer.

This setup provides a slow entrance of the proteins into the stacking gel matrix, a preseparation and concentration effect during the migration in the stacking gel (stacking effect), and an efﬁcient separation into sharp zones in the resolving gel. 19.4.1.2

Native Conditions

In native separations both parameters, charge and size, inﬂuence the electrophoretic mobilities. Under native conditions secondary and tertiary structures of the molecules can play an important role. Nucleic acids, peptides, and easily soluble acidic proteins are separated as anions in basic buffers. However, some hydrophobic and basic proteins can only be analyzed in acidic gels, where the proteins are positively charged and migrate towards the cathode. Under native conditions the gel temperature plays an important role, because the pK values of the buffer ions and the analytes vary at different temperatures. 19.4.1.3

SDS Electrophoresis

When protein mixtures are treated with SDS-containing buffers the individual charges and structures of the proteins are canceled, because the SDS cleaves the hydrogen bonds, stretches the polypeptide chains, and covers the polypeptides completely. Usually a reducing agent like dithiothreitol (DTT) is added to cleave the disulﬁde bridges. In this way all proteins have the same stretched shape and they are negatively charged. They migrate all into the anodal direction at a velocity depending only on the length of the polypeptide, which is directly proportional to the molecular weight. Molecular weight standard proteins are run in each gel for the molecular weight determination of sample proteins. 19.4.1.4

Immuno Electrophoresis

In immuno electrophoresis, the sample proteins migrate into an agarose gel which contains a certain concentration of polyclonal, monospeciﬁc antibodies against a certain protein. In order to prevent the antibody from migration, the gel buffer has a pH value as close as possible to the isoelectric point of the antibody, the immunoglobulin G. When the antigen and the antibody reach the

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 430 17.7.2008 5:08pm Compositor Name: VAmoudavally

430

Handbook of Food Analysis Instruments

‘‘equivalence point,’’ a three-dimensional—antigen–antibody–antigen–antibody—precipitate is formed, which can be detected with Coomassie Brilliant Blue staining. The migration distance of the protein to the equivalence point is a measure for its abundance in the sample. Immuno electrophoresis is a very speciﬁc and quantitative method. It does not work with monoclonal antibodies, because they do not form precipitates. Several modiﬁcations including two-dimensional setups are possible. However, because of the high antibody consumption this method has been partly replaced by blotting methods and ELISA tests.

19.4.2 ISOELECTRIC FOCUSING Isoelectric focusing is a special form of electrophoresis and can exclusively be applied on amphoteric compounds, mostly proteins, but also peptides and some dyes. In isoelectric focusing, the electrophoretic migration happens in a gel containing a pH gradient. Depending on the application position within the pH gradient, the molecules bear either a negative or a positive net charge. In the electric ﬁeld they migrate towards the electrode with the opposite sign. When a protein arrives at its isoelectric point, it will stop to migrate because its net charge is zero. This is an end point method. If a protein would diffuse away, it would pick up a charge and migrate back to its isoelectric point. This focusing effect provides very sharp zones with high resolution. The value of the isoelectric point is strongly dependent on the temperature, therefore isoelectric focusing has to be performed under stringent and deﬁned temperature control. Native separations are carried out at 108C, denaturing runs in presence of high molar urea at 208C. There are two ways to establish a pH gradient: Carrier ampholytes: When a heterogeneous mixture of specially synthesized amphoteric buffers with a wide spectrum of isoelectric points, the carrier ampholytes, is exposed to an electric ﬁeld, the carrier ampholytes will migrate according to their net charge. Basic homologues migrate towards the cathode and acidic ones towards the anode. Because carrier ampholytes exhibit a high buffer power at their isoelectric point, a stable pH gradient will be established. No buffer reservoirs are required. To prevent the very acidic and very basic carrier ampholytes from leaving the gel, thick ﬁlter paper strips soaked in electrode solutions are placed between the gel edges and the platinum electrodes. For the anodal side mostly diluted phosphoric acid is used, for the cathode diluted sodium hydroxide solution. Immobilized pH gradient gels are produced by casting a gradient thin layer gel with two monomer solutions, to which different compositions of acidic and basic acrylamide derivatives were added. After the polymerization the buffering groups are ﬁxed. The gels are washed with distilled water, dried down on a ﬁlm support, and cut into narrow strips. These ‘‘IPGstrips’’ are rehydrated in a urea=detergent=reductant mixture before they are used for isoelectric focusing of complex protein mixtures under denaturing conditions in high resolution two-dimensional electrophoresis.

19.4.3 TWO-DIMENSIONAL ELECTROPHORESIS For the separation of highly complex protein mixtures isoelectric, focusing under denaturing conditions is combined with SDS electrophoresis. The ﬁrst dimension is mainly performed with IPGstrips. The protein mixture is either applied by soaking the strip in the sample solution, or by loading it onto the acidic or basic end of the strip with a little sample application cup. When the proteins have been focused at their isoelectric points, the strip is equilibrated in SDS buffer and applied onto the upper edge of an SDS polyacrylamide gel. The resulting pattern shows scattered protein spots related to their isoelectric points and molecular weights. Because the separation occurs according to two completely independent physicochemical parameters, the charge and the size, it exhibits very high resolving power. When large gels are used, this method can resolve several thousand proteins and has therefore become the standard separation method for proteomics.

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 431 17.7.2008 5:08pm Compositor Name: VAmoudavally

Gel Electrophoresis in Food Analysis

431

19.5 INSTRUMENTATION In principle, a gel electrophoresis system consists of three instruments: a power supply, an electrophoresis chamber, and a thermostatic circulator. Some simple chambers, like submarine and mini-gel devices, do not need cooling, because the separation time is short. A few very sophisticated systems have everything built in: the power supply and a Peltier cooling element. The cooling device, either a thermostatic circulator or a Peltier system, has two functions: 1. Dissipation of Joule heat, which is caused by the applied electric power 2. Temperature control, to ensure deﬁned separation conditions When the temperature cannot be dissipated efﬁciently enough, zone electrophoresis results will show a ‘‘smiling effect’’: The heat removal works better at the lateral sides; thus the temperature in the center of the gel is higher, leading to an increased ion mobility. The result is a curved ‘‘smiling’’ front. As a consequence the electric settings need to be reduced, and the separation will be slower. Native electrophoresis is mostly run at 108C, SDS electrophoresis at 158C, for isoelectric focusing see above.

19.5.1 POWER SUPPLY All electrophoresis separations require a DC power supply. For zone electrophoresis it should provide up to 600 V and 200 mA. Ideally it offers a maximum power cutoff to prevent overheating. In most cases discontinuous buffers are employed, which result in a strong variation of electric conductivity, a crossover from high conductivity with high current ﬂow to low conductivity with high electric ﬁeld strength. If the power supply does not have the maximum power setting feature, the electric conditions need to be manually adjusted during the electrophoresis run, alternatively the conditions must be selected at very low settings of maximum voltage or current. This results in long separation times and diffused zones. For isoelectric focusing this maximum power setting feature is an absolute must. For the focusing function lowly charged proteins have to be moved to the isoelectric point, which requires very high electric ﬁeld strength. A power supply for isoelectric focusing should provide up to 3000 V, the current requirements are relatively low. Usually these high-end instruments also contain a multistep programming feature. During blotting the electric ﬁeld is established across a large intersection area, which results in a high current ﬂow. Because the electrode distances are short, the applied voltage is rather low. Power supplies for blotting typically offer up to 200 V and 2 A. There are instruments on the market, which offer enough voltage and current for a wide range of methodical applications. Most power supplies have two or more outlets for running several chambers simultaneously. For such runs the current and power settings have to be adjusted accordingly.

19.5.2 VERTICAL ELECTROPHORESIS SYSTEMS 19.5.2.1

Gel Rod Apparatus

Figure 19.2 shows an apparatus for gel rods in glass tubes. Each sample is separated in an individual gel, which is about 10 cm long. After separation the gel cylinders are removed from the tube with the help of a water-ﬁlled syringe with a long blunt needle by injecting water while carefully removing the gel rod. These systems are more and more replaced by slab gel systems, because their handling, staining, and evaluation are easier and slab gels can be blotted.

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 432 17.7.2008 5:08pm Compositor Name: VAmoudavally

432

Handbook of Food Analysis Instruments

FIGURE 19.2

19.5.2.2

Gel rod apparatus for separating samples in individual gels.

Slab Gel Apparatus

Slab gels are polymerized between two glass plates and two spacers, which form a cassette with a thickness from 0.7, 1, or 1.5 mm. Different apparatus designs are used, which are shown in Figure 19.3. Different formats are used: minigel systems with a gel size of about 8 8 cm for 10 samples per gel, standard sizes are about 14 14 cm for 10 or 15 samples per gel. Minigel apparatus have mostly no cooling device. Some buffer-back systems offer the possibility to cool the ‘‘upper buffer’’ via a heat exchanger. The design type shown in Figure 19.3A and B requires a notched glass plate on one side, which is shown in the assembly in Figure 19.1B. Double chambers with a design shown in Figure 19.3C do not have notched glass plates, but the edges of the glass plates need to be very smooth. This is necessary to ensure good contact with the sealing casket, which separates upper and lower buffer chamber. The lower buffer tank can either be used as a heat sink or is actively cooled with a thermostatic circulator. All these systems can be used for all types of zone electrophoresis. For running two gels in a double chamber the settings for current and power have to be doubled.

(A)

(B)

(C)

FIGURE 19.3 Vertical slab gel systems. (A) Plain system without any heat dissipation. (B) Buffer-back system, which uses the upper buffer as a heat sink. (C) Double chamber with heat exchanger and lower buffer cooling. The lower buffer can either be used as a heat sink or actively cooled with a thermostatic circulator.

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 433 17.7.2008 5:08pm Compositor Name: VAmoudavally

433

Gel Electrophoresis in Food Analysis

(A)

(B)

(C)

FIGURE 19.4 Horizontal ﬂatbed systems. (A) Submarine chamber. (B) Multipurpose apparatus for thin-layer gels with external cooling. (C) Peltier-cooled ﬂatbed system for thin-layer gels.

19.5.3 HORIZONTAL FLATBED ELECTROPHORESIS SYSTEMS 19.5.3.1

Submarine Apparatus for DNA

For separation of DNA fragments a simple plastic apparatus without cooling is mostly sufﬁcient. The system is called submarine chamber, because the gel surface is submerged below the buffer surface (Figure 19.4A). 19.5.3.2

Standard Size Apparatus for Thin-Layer Electrophoresis

Figure 19.4B shows a horizontal ﬂatbed apparatus which can be used for all electrophoresis techniques with the exception of submarine electrophoresis. Gels with thicknesses of 0.5 to 1 mm are placed on the cooling block, which is connected to an external thermostatic circulator. The standard gel size is 25 12 cm, but the cooling plate can accommodate up to 25 20 cm large gels. Usually 25 samples are applied on such a gel, but it is possible to apply more proteins using gels with more narrow sample wells. When high sample throughput is required, the electrode holder plate can be equipped with three electrodes for running a double gel: with a common cathode in the center and two anodes at the lateral sides. For immuno electrophoresis the buffer tanks are used; the lid of the instrument has an arrangement of holes to introduce voltage probes for measuring the electric ﬁeld strength during the run. For zone electrophoresis techniques the buffer tanks are employed with paper wicks connecting the buffer and the gel edges. However, it is easier to use ready-made polyacrylamide buffer strips or thick ﬁlter paper strips soaked with concentrated running buffer. The instrument shown in Figure 19.4C has a built-in Peltier cooling and can be used for all thinlayer techniques which do not require buffer tanks. 19.5.3.3

Mini-Format Automated Electrophoresis System

Sometimes it is not necessary to obtain high resolution. For very quick separations, an automated mini-format system is very useful. The instrument shown in Figure 19.5 is based on the use of 0.5 mm thin ﬁlm-supported ready-made gels of 4 5 cm and buffer strips, and does almost everything automatically: cooling, sample application, different separation steps, and staining the gels. All programs, including cooling temperature, different stages of electric settings, and the staining procedure, are controlled by a central processor. Usually eight samples per gel are run.

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 434 17.7.2008 5:08pm Compositor Name: VAmoudavally

434

Handbook of Food Analysis Instruments

8/1

0 1 2

FIGURE 19.5 Automated minigel system. From left to right: Sample applicator for 8 samples of 1 mL, electrophoresis gel and agarose buffer blocks; electrophoresis module containing central processor, built-in Peltier cooling, power supply; staining module containing heatable staining chamber with turning gel holder, pump, and 10-port valve connecting the chamber to the tubing for the staining and waste bottles.

19.5.4 SYSTEMS

FOR

HIGH-RESOLUTION TWO-DIMENSIONAL ELECTROPHORESIS

Although two-dimensional electrophoresis can be performed in standard systems, which have been described above, higher resolution and reproducibility can be obtained with dedicated systems. For the workﬂow, as well as for optimal pattern evaluation, it is advantageous to run multiple gels in one instrument. Figure 19.6A shows an instrument designed speciﬁcally for isoelectric focusing in immobilized pH gradient strips, which allows high voltages up to 10,000 V, because all modules including cooling and power supply are contained in a closed system. Figure 19.6B shows a multiple SDS gel chamber for up to six gels. Typically high resolution gels are 25 cm wide and 20 cm long.

19.5.5 BLOTTING SYSTEMS 19.5.5.1

Tank Blotters

Blotting tanks are also available in different sizes. The blotting sandwich—ﬁlter paper stack–gel– membrane–ﬁlter paper stack—is assembled in a tray under buffer and then clamped into a gridcassette, which is vertically introduced into the tank. To the two opposite walls of the tank electrode

(A)

(B)

FIGURE 19.6 Instruments for high resolution two-dimensional electrophoresis. (A) Isoelectric focusing chamber with built-in Peltier cooling, electrode plates, and power supply. (B) Multiple SDS gel chamber for up to six gels with lower buffer circulation pump and heat exchanger.

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 435 17.7.2008 5:08pm Compositor Name: VAmoudavally

435

Gel Electrophoresis in Food Analysis

(A)

(B)

FIGURE 19.7 Blotting instruments. (A) Tank blotter. (B) Semidry blotter to be used inside the box of a multipurpose chamber like in Figure 19.4B.

grids are attached, which ensure an even electric ﬁeld. Some tanks are equipped with a heat exchanger for removal of Joule heat. The transfer takes mostly over night. Several blotting transfers can be performed simultaneously. In Figure 19.7A such a blotting tank is displayed. 19.5.5.2

Semidry Blotters

Instead of using a large volume of buffer it is also possible to transfer the proteins with a stack of ﬁlter papers soaked in buffer. The blot sandwich is placed between two electrode plates made from graphite (as shown in Figure 19.7B) or grids of inert metal. Filter papers and membrane are either cut to the same size like the gel, or a thin plastic mask is placed between gel and blot membrane. After the sandwich has been assembled, air bubbles are pressed out with a roller. The transfers are performed within a few hours. It is not recommended to transfer more than one gel at the same time in a stack because the efﬁciency will be considerably reduced. Semidry blotters are available as stand-alone systems or accessories to electrophoresis chambers.

19.5.6 AUTOMATIC STAINING APPARATUS Coomassie Brilliant Blue and ﬂuorescent staining requires only a few steps. However, silver staining is a multistep procedure and becomes only reproducible, when the staining intervals are followed with stringent timing. A staining automat is therefore very useful. An automated staining device is included in the minigel apparatus shown in Figure 19.5. Figure 19.8 shows a gel and blot processor for gels and blot membranes of standard sizes. The staining is carried out in a rocking tray, which is connected to a ten-port valve via silicone tubing. The ﬂuids are delivered and removed with a peristaltic pump.

19.5.7 IMAGERS Electrophoresis gels are often evaluated by visual inspection. However, there are many cases where the image has to be acquired with a camera or scanner: . . . . .

Fluorescence dyes and chemiluminescence detection Quantiﬁcation Digitization for image analysis with computer software Documentation and reports Communication of results

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 436 17.7.2008 5:08pm Compositor Name: VAmoudavally

436

Handbook of Food Analysis Instruments

FIGURE 19.8

19.5.7.1

Automated staining device and blot processor.

Still CCD Cameras

A still camera inside a dark box is the ideal solution for chemiluminescence detection because the signals can be accumulated over a relatively long time period. They are usually used for miniformats. For large formats the resolution is insufﬁcient. 19.5.7.2

White Light Scanners

Different types of white light scanners are on the market: for visible staining a line scanner can be employed, like a desk top scanner. However, it must offer the feature to scan in transmission mode and the possibility to be calibrated. With reﬂection mode scanning quantiﬁcation of protein bands or spots would be impossible. For ﬂuorescent detection a scanner needs a set of excitation and emission ﬁlters. In this way certain ﬂuorophores can selectively be excited with monochromatic light with a deﬁned wavelength, and their emitted light signals can selectively be measured. For the image acquisition a moving change coupled device (CCD) camera can be employed, which scans the pattern ‘‘on the ﬂy.’’ 19.5.7.3

Laser Scanner

The most sophisticated instruments are laser scanners. High-end instruments are equipped with several laser sources with different wavelengths, confocal optics, and photomultipliers for signal acquisition. With such instruments highest sensitivity is coupled with a wide dynamic range, high measuring precision, and resolution.

19.6 APPLICATIONS TO FOOD ANALYSIS Some typical examples of using gel electrophoresis in food analysis have been selected from a very high number of papers.

19.6.1 ANALYSIS 19.6.1.1

OF

NUCLEIC ACIDS

Genetic Differentiation

Gel electrophoresis of DNA fragments is frequently applied for plant cultivar and animal-species differentiation. In most cases the selected DNA fragments are linked to genes with certain functional properties. These DNA fragments are speciﬁcally ampliﬁed using the PCR (polymerase chain reaction) technology with different primer combinations; the separations are performed in

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 437 17.7.2008 5:08pm Compositor Name: VAmoudavally

Gel Electrophoresis in Food Analysis

437

polyacrylamide gels either under denaturing conditions in the presence of high molar urea: for instance AFLP (ampliﬁed fragments length polymorphism), or under native conditions, like for instance RAPD (random ampliﬁed polymorphic DNA), or SSCP (single strand conformation polymorphism). While AFLP and RAPD are mainly applied on plants [2], animal species and subspecies are identiﬁed with SSCP [3]. 19.6.1.2

Detection of Genetically Modiﬁed Food

Another important application is the detection of the use of genetically modiﬁed material for food production by PCR and subsequent submarine gel electrophoresis in agarose gels, as proposed by the FDA (The Food and Drug Administration) [4].

19.6.2 ANALYSIS

OF

PROTEINS

Easy and fast crop cultivar identiﬁcation is performed on the protein level, usually by speciﬁcally extracting the prolamines and separating this protein subset in cathodal direction in thin horizontal acidic polyacrylamide gel layers, as shown by Hsam et al. [5]. Alternatively, all proteins can be separated on SDS polyacrylamide gels, with or without subsequent speciﬁc ligand detection with blotting [6,7]. SDS polyacrylamide gel electrophoresis and blotting were also employed by Scheibe et al. [8,9] for immunodetection of hidden allergens, like hazelnut and almonds in chocolate, as well as industrial enzymes in apple juice. Isoelectric focusing is a particularly sensitive method for detecting animal-species speciﬁc proteins in meat [10] and milk [11] products. The allergens in wheat causing baker’s asthma have been detected with high resolution two-dimensional electrophoresis with subsequent immunoblotting [12]. As already mentioned, two-dimensional electrophoresis is the most employed separation method in proteomics: after image analysis of the protein spots patterns the spots of interest are picked from the gel, the proteins are digested with trypsin, and the peptides are further analyzed with mass spectrometry. Proteomics as a concept for studying the structures, modiﬁcations, and functions of proteins will be increasingly applied to food analysis for improvement of food quality, as predicted, for instance, by Kvasnicka [13] and Carbonaro [14].

19.7 FUTURE TRENDS Fluorescent dyes offer very high sensitivity and the widest dynamic range among all detection systems. These features improve the detection and quantiﬁcation considerably. Another important advantage of ﬂuorophores is the possibility to employ multiplexing. There are ﬂuorescent dyes available, which selectively detect glycosylated or phosphorylated proteins. Furthermore, proteins of different samples can be prelabeled with different ﬂuorophores, combined together, and run on the same gel. This concept is already very successfully applied in proteomics and is called DIGE (difference gel electrophoresis). In this way quantitative differences of protein abundances can be determined with utmost precision. Therefore it can be predicted that ﬂuorescent dyes will more and more replace the classical staining methods Coomassie Blue and silver staining.

REFERENCES 1. Westermeier, R. Electrophoresis in Practice. 4th ed., WILEY-VCH, Weinheim, 2004 [chap. 1]. 2. Thomas, C.M. et al. Identiﬁcation of ampliﬁed restriction fragment polymorphism (AFLP) markers tightly linked to the tomato Cf-9 gene for resistance to Cladosporum fulvum. Plant J. 8, 785, 1995. 3. Rehbein, H. et al. Differentiation of sturgeon caviar by single strand conformation polymorphism (PCRSSCP) analysis. Arch. Lebensmittelhygiene. 50, 13, 1999. 4. FDA. U.S. Food and Drug Administration Center for Food Safety & Applied Nutrition. Foods derived from new plant varieties derived through recombinant DNA technology. 2000. Online: http:==vm.cfsan. fda.gov=~lrd=biocon.html

Otles/Handbook of Food Analysis Instruments 45660_C019 Final Proof page 438 17.7.2008 5:08pm Compositor Name: VAmoudavally

438

Handbook of Food Analysis Instruments

5. Hsam, S.L.K., et al. Identiﬁcation of cultivars of crop species by polyacrylamide electrophoresis. Brauwissenschaft 3, 86, 1993. 6. Weiss, W., Postel, W., and Görg, A. Barley cultivar discrimination: I. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis and glycoprotein blotting. Electrophoresis 12, 323, 1991. 7. Weiss, W., Postel, W., and Görg, A. Qualitative and quantitative changes in barley seed protein patterns during the malting process analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis with respect to malting quality. Electrophoresis 13, 787, 1992. 8. Scheibe, B. et al. Detection of trace amounts of hidden allergens: Hazelnut and almond proteins in chocolate. J. Chromatogr. B. 756, 229, 2001. 9. Scheibe, B. et al. Electrophoretic and immunochemical determination of bioindustrial enzymes in apple juice. Eur. Food Res. Technol. 212, 691, 2001. 10. Skarpeid, H.J., Kvaal, K., and Hildrum, K.I. Identiﬁcation of animal species in ground meat mixtures by multivariate analysis of isoelectric focusing protein proﬁles. Electrophoresis 19, 3103, 1998. 11. Moio, L., Luccia, A.D., and Addeo, F. Fast isoelectric focusing of milk proteins on small ultrathin polyacrylamide gels containing urea. Electrophoresis 10, 533, 1989. 12. Weiss, W., Vogelmeier, C., and Görg, A. Electrophoretic characterization of wheat grain allergens from different cultivars involved in bakers’ asthma. Electrophoresis 14, 805, 1993. 13. Kvasnicka, F. Proteomics: General strategies and application to nutritionally relevant proteins. J. Chromatogr. B. 787, 77, 2003. 14. Carbonaro, M. Proteomics: Present and future in food quality evaluation. Trends Food Sci. Technol. 15, 209, 2004.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 439 16.7.2008 11:09pm Compositor Name: MSubramanian

Immunoassays 20 Multiplexed in Food Analysis Chien-Sheng Chen, Antje J. Baeumner, and Richard A. Durst CONTENTS 20.1 20.2

Introduction ........................................................................................................................ 439 Multiple-Label Immunoassays .......................................................................................... 440 20.2.1 Multiple Radioisotopes ........................................................................................ 440 20.2.2 Multiple Lanthanides ........................................................................................... 440 20.2.3 Multiple Visible and Fluorescent Dyes ............................................................... 440 20.2.4 Multiple Enzymes ................................................................................................ 441 20.2.5 Multiple Metals .................................................................................................... 441 20.2.6 Multiple Nanocrystals .......................................................................................... 441 20.3 Spatially Resolved Multiplexed Immunoassays ................................................................ 442 20.3.1 Membrane Assays ............................................................................................... 442 20.3.2 Capillary Assays .................................................................................................. 442 20.3.3 Microchip Assays ................................................................................................ 443 20.3.3.1 Surface Materials ................................................................................ 443 20.3.3.2 Biomolecule Patterning ....................................................................... 444 20.3.3.3 Assay Platforms .................................................................................. 446 20.3.3.4 Detection Systems .............................................................................. 447 20.3.3.5 Examples of Microchip Assays .......................................................... 448 20.3.4 Microtiter Plate Assays ........................................................................................ 449 20.3.5 Bead Assays ........................................................................................................ 450 20.3.6 Electrode Assays .................................................................................................. 451 20.3.7 Nanowire Assays ................................................................................................. 452 20.4 Multiplexed Immunoassay Challenges .............................................................................. 452 20.5 Conclusions ....................................................................................................................... 453 Acknowledgments......................................................................................................................... 453 References ..................................................................................................................................... 454

20.1 INTRODUCTION In general, immunoassays can be deﬁned as analytical methods in which antibodies are used for the identiﬁcation and quantiﬁcation of deﬁned target analytes. At present, immunoassays are the most commonly used types of ligand-binding assays for the identiﬁcation of a large variety of analytes, such as proteins, peptides, microorganisms, and low molecular weight molecules [1]. Multiplexed immunoassays (simultaneous multianalyte immunoassays), in which several analytes are measured simultaneously in a single assay, present several advantages over single-analyte assays such as reduction of analysis time, work simpliﬁcation, decrease in total sample volume, and lower overall cost per test [2,3]. The ability to simultaneously measure multiple analytes in a single assay holds 439

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 440 16.7.2008 11:09pm Compositor Name: MSubramanian

440

Handbook of Food Analysis Instruments

enormous potential for meeting the growing demands of proteomics, diagnosis, environmental monitoring, food safety, and homeland security. Multiplexed immunoassays are broadly classiﬁed as multiple-label assays and spatially resolved assays. Multiple-label assays are methods in which antibodies or antigens are localized in the same assay zone with distinct labels such as radiolabels, ﬂuorophores, enzymes, metals, and nanocrystals. Spatially resolved assays use discrete analytical zones on a single device without the need for multiple labels. The substrates used in spatially separated assays are usually membranes, capillaries, microchips, microtiter plates, beads (microspheres), and electrodes.

20.2 MULTIPLE-LABEL IMMUNOASSAYS 20.2.1 MULTIPLE RADIOISOTOPES The ﬁrst multiplexed immunoassays employed multiple radioisotopic labels. Dual-analyte immunoassays reported in 1966 used 131I and 125I for the simultaneous detection of human insulin and growth hormone [4]. The discrimination of radioisotopes requires that the energy spectra are distinct. The radioisotopes 125I and 57Co have been combined for the detection of lutropin and follitropin [5], vitamin B12 and folic acid [6], as well as thyrotropin and thyroxin [7]. Because of the concerns over the safety of radioisotopic labels and the disposal of radioactive waste, non-isotopic labels are typically chosen for immunoassays today.

20.2.2 MULTIPLE LANTHANIDES Lanthanides are a series of elements of increasing atomic numbers beginning with lanthanum (57) or cerium (58) and ending with lutetium (71). Lanthanides for time-resolved ﬂuorescence detection were the ﬁrst choice of non-isotopic labels for multiplexed immunoassays because of their characteristic narrow-banded emission lines. Their emissions are clearly distinguishable from one another with respect to both wavelengths and lifetimes [8]. In addition, time-resolved ﬂuorescence is detectable at low levels and enables the development of several highly sensitive immunoassays because of the nearly zero background signal [9]. Europium (Eu3þ) and terbium (Tb3þ) as well as europium and samarium (Sm3þ) pairs have been effectively combined for dual-analyte immunoassays [10–13]. Xu et al. [14] even described a quadruple-label ﬂuorometric immunoassay using Eu3þ, Tb3þ, Sm3þ, and Dy3þ (dysprosium) for a simultaneous detection of thyroid-stimulating hormone, 17-a-hydroxyprogesterone, immunoreactive trypsin, and creatine kinase MM isoenzyme in dried blood spots. The assay was performed with a sandwich format in a microtiter well. The limits of detection (LODs) for thyrotropin, 17-a-hydroxyprogesterone, immunoreactive trypsin, and creatine kinase MM isoenzyme were 0.1 mIU=L, 2 nmol=L, 2 mg=L, and 4 U=L, respectively.

20.2.3 MULTIPLE VISIBLE

AND

FLUORESCENT DYES

Fluorescent dyes are commonly used in single-analyte immunoassays. However, their broad emission spectra with red spectral tails have limited their use in multiple-label assays owing to spectral overlap; therefore, only dual-dye assays were used in multiplexed immunoassays. Green ﬂuorescein and red rhodamine or their derivatives are the two common dyes used for dual-analyte immunoassays [15–17]. Highly colored ﬂuorescent dyes may also be used as visible dyes based on spectral reﬂectance measurements. A multianalyte lateral-ﬂow immunoassay was developed using three preparations of liposomes, each encapsulating a different dye, which were captured in a single zone on a lateral-ﬂow strip, and the reﬂectance spectrum was measured. Partial least square regression was used to develop multivariate calibrations that could successfully deconvolute the overlapping spectra [18].

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 441 16.7.2008 11:09pm Compositor Name: MSubramanian

Multiplexed Immunoassays in Food Analysis

441

20.2.4 MULTIPLE ENZYMES The most commonly used labels for immunoassays are enzymes, of which ELISAs (enzyme-linked immunosorbent assays) are the primary format. Both enzyme-labeled and substrate-labeled ﬂuorescent immunoassays have been used for dual-analyte detection. Alkaline phosphatase, horseradish peroxidase, and b-galactosidase are the three most popular enzymes for immunoassays. The combination of alkaline phosphatase and b-galactosidase was used for the detection of two thyroid hormones [19]. However, enzyme labels usually have different optimum ranges of temperature, pH, and reaction time for their activity. For examples, the optimum pH for alkaline phosphatase, horseradish peroxidase, and b-galactosidase are 8–10, 5–7, and 6–8, respectively [20]. Therefore, identifying a ﬁnal set of assay conditions without sacriﬁcing sensitivity would be a difﬁcult task. To eliminate the difﬁculty, bound enzymes were sequentially reacted with the corresponding substrates [21–24]. To avoid the problem that results from using the same kind of label, an enzyme was also combined with radioisotope in a dual-analyte immunoassay [25].

20.2.5 MULTIPLE METALS Different metals have been used with different transducers for multi-label immunoassays. Salmain et al. [26] used FTIR (Fourier transform infrared) spectroscopy to simultaneously detect two distinct organometallic labels (metal carbonyl complexes). Hayes et al. [27] used differential pulse anodic stripping voltammetry at a hanging mercury drop electrode to detect two metal ion labels (In3þ and Bi3þ). Bordes et al. [28] developed a dual-analyte immunoassay using two cationic metal complexes (cobaltocenium and ferroceneammonium) and a Naﬁon-loaded carbon paste electrode by squarewave voltammetry. Although they later also added pentaammine ruthenium(II) complexes for a simultaneous detection of three-labeled drugs, the immunoassay was not successful due to the instability of the pentaammine ruthenium(II) complexes [29].

20.2.6 MULTIPLE NANOCRYSTALS Inorganic nanocrystals are relatively new classes of labels for immunoassays. Luminescent, colloidal, and semiconductor nanocrystals (quantum dots, QDs) have the potential to overcome the limitations of organic ﬂuorophores. They have broad excitation as well as size-dependent, tunable, narrow-emission spectra that allow the simultaneous excitation of several different-colored QDs at a single wavelength with little spectral emission overlap for multiplex analysis [30–32]. Also, QDs have been reported to be about 20 times brighter and 100 times more photostable in comparison with organic dyes such as rhodamine [33]. QDs enable multiplex assays, requiring only a single excitation source having an excitation wavelength far from that of the QD emission peaks [29–30]. Goldman et al. [32] used four antibody-conjugated QDs of different sizes to demonstrate multiplexed assays for four protein toxins in a single sample. The LODs for the detection of cholera toxin, ricin, shiga-like toxin 1, and staphylococcal enterotoxin B were 10, 30, 300, and 3 ng=mL, respectively. However, emission spectral overlaps were still observed in the four QDs even though they have relative narrow-emission spectra. A linear equation-based algorithm was used to deconvolute the signal from mixed toxin samples for the simultaneous quantitation of all four toxins. Nanocrystals are not only ideal optical labels, but also excellent electrochemical labels for immunoassays. Liu et al. [34] detected four proteins (b-microglobulin, IgG, bovine serum albumin, and C-reactive protein) simultaneously using four kinds of colloidal nanocrystals (ZnS, CdS, PbS, and CuS) yielding distinct voltammetric peaks, the positions and sizes of which reﬂect the identity and concentration of the respective analytes. Nanocrystals were measured at an in situ plated mercury ﬁlm on a glassy carbon electrode by square-wave anodic stripping voltammetry, which combines the ampliﬁcation feature of stripping voltammetry with the speed advantage of square-wave scanning. Femtomole detection limits resulted from the combination of electrochemical stripping transduction and immunomagnetic beads.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 442 16.7.2008 11:09pm Compositor Name: MSubramanian

442

Handbook of Food Analysis Instruments

20.3 SPATIALLY RESOLVED MULTIPLEXED IMMUNOASSAYS 20.3.1 MEMBRANE ASSAYS Membranes have been the standard substrate for western blot assays. Western blotting ﬁrst separates antigens electrophoretically on a gel, and then a membrane (usually nitrocellulose) is placed on the gel and antigen bands are driven onto the membrane using electrophoresis. The membrane, which provides discrete spatially separated test zones, is then incubated with antibodies. Combined with multiple labels (usually enzymes), a multiplexed detection is achieved [35]. Antigens can also be applied directly to the membrane in parallel lines. Ijsselmuiden et al. [36] used this technique to detect antibodies against two treponemal antigens. A disposable membrane test card with multiple test areas processed in an incubation instrument for simultaneous determinations from a single specimen was designed by Donohue et al. [37]. Buechler et al. [38] used the same principle with colloidal gold labels to simultaneously detect seven drugs of abuse in urine. Instead of incubating the whole membrane card with samples and reagents, an alternative assay format is used to apply the sample and reagents at one end of a membrane strip, and the mixture moves along the membrane to multiple analytical zones by capillary migration (Figure 20.1). This type of membrane strip assay is called an immunochromatographic dipstick or a lateral-ﬂow assay. The advantage of this type of assay is simple handling without requiring any washing steps, thereby usually being completed in less than 30 min. The intensities of the color bands, which correlate to the concentrations of the analytes, can be visually estimated or quantiﬁed by a reﬂectometer [39], a microarray scanner [40], or densitometry using specialized computer software [41]. This type of assay has also been applied to the simultaneous detection of drugs of abuse in urine [37]. Several researchers also used this device for multiplexed detection of food contaminants, including pesticide residues [42,43] and mycotoxins [44,45].

20.3.2 CAPILLARY ASSAYS Capillary-based immunoassays employ ﬂow-through optical systems. Capillaries have several advantages over other substrates for assays, such as improved assay kinetics attributed to the higher surface-to-volume ratio and less mass transport of the reactants being required [46]. Also, consumption of reagents is reduced and samples for the immunoassays are small due to the small dimensions of capillaries [47]. Capillaries can be made of plastic [47,48] and fused silica [46,49]. Antibodies have usually been immobilized in the capillary by physical adsorption [46–48]; however, it has also been reported that antibodies were coated onto the silanized inner walls of fused

Analytical zone for analyte 3 Analytical zone for analyte 2 Analytical zone for analyte 1

Mixture of analytes and reagents

FIGURE 20.1 Multiplexed lateral-ﬂow assay format. The test strip has multiple discrete analytical zones. Analytes bind to the reagents in the mixture, which migrates through the test strip to reach the analytical zones by capillary action, thereby developing the color in the bands on the test strip.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 443 16.7.2008 11:09pm Compositor Name: MSubramanian

443

Multiplexed Immunoassays in Food Analysis Excitation light Light sensor

Fluorescence intensity

Analyte bands

Scanned capillary length (mm)

FIGURE 20.2 Optical set-up of a capillary-based multiplexed immunosensor and idealized recording of the ﬂuorescence intensity obtained by scanning the capillary. (Redrawn from Petrou, P.S., Kakabakos, S.E., Christoﬁdis, I., Argitis, P., and Misiakos, K., Biosens. Bioelectron., 17, 261, 2002.)

silica capillaries using a heterobifunctional crosslinker, N-succinimidyl-4-maleimidobutyrate [49]. Platforms of either multiple capillaries coated with different antibodies [46,49] or multiple discrete antibody bands in one capillary [47,48] have been used for the capillary-based multiplexed immunoassays. Petrou et al. [48] used a precision microsyringe to inject 2.5 mL of each antibody solution into the capillary, thereby forming three discrete antibody bands. The multiple ﬂuorescent bands formed in the capillary can be quantiﬁed by scanning the capillary with a light beam of appropriate wavelength for the excitation of the ﬂuorescent label (Figure 20.2). The photons are trapped by internal reﬂectance in the capillary walls and waveguided toward its end. A light sensor detects the waveguided photons at the end of the capillary [48]. Three hormones in human serum, follitropin, human chorionic gonadotropin, and prolactin, could be detected with LODs of 1.3 mg=L, 2.3 IU=L, and 3.6 IU=L, respectively. Although they were adequate for the determination of hormones in human serum samples, they were higher than those determined by microtiter plate assays (0.7 mg=L, 1.3 IU=L, and 2.6 IU=L, respectively) [48].

20.3.3 MICROCHIP ASSAYS Multiplexed immunoassays are currently undergoing intensive development because of the application of protein microarrays (microchip) and proteomics. Microchip technology development has made great advances and has shown the most promising potential for multiplexed immunoassays in terms of high throughput. A large number of studies have been reported on microchip surface materials, biomolecule patterning, assay formats, and detection systems. 20.3.3.1

Surface Materials

It is ideal for microchips to have a surface material providing high binding capacity, high protein density, low nonspeciﬁc background, and high reproducibility. Although PVDF (polyvinylidene ﬂuoride) membranes were commonly used in the early array-based immunoassays [50–52], they did not allow a sufﬁciently high protein density [53] because of the spreading of the spotted materials. Therefore, glass slides have become the preferred solid supports, which are able to provide 1600

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 444 16.7.2008 11:09pm Compositor Name: MSubramanian

444

Handbook of Food Analysis Instruments

spots=cm2 [54]. Glass also has great durability, good optical properties, and compatibility with the platforms already established for DNA microarrays. Since the glass surface is not suitable for passive adsorption of biomolecules, polystyrene [55] and polylysine [56] have been used to modify the glass surface for adsorption. To achieve more speciﬁc and stronger biomolecule attachment, glass slides are also activated with coupling groups, such as aldehyde, epoxy, carboxylic esters, mercaptopropyl trimethoxysilane, or amine, to covalently crosslink to biomolecules [57,58]. Polymer layers (dendrimer and polyethylene glycol) with coupling groups have also been used to coat the glass to reduce nonspeciﬁc binding and to increase the density of accessible functional groups [59]. An alternative is high-porosity gels or membrane-coated surfaces, such as hydrogel [60,61], agarose [62], and nitrocellulose [63]. These surfaces prevent both rapid evaporation and the close contact of the protein with the surface, thereby preserving the three-dimensional structures of immobilized biomolecules. Also, hydrogels were reported to have low intrinsic background ﬂuorescence that further improves the sensitivity of immunoassays [64]. The polyacrylamide-based hydrogel substrate yielded a sixfold higher signal-to-noise ratio than the poly-L-lysine substrate in an antibody microarray proﬁling of human prostate cancer sera [65]. Although it has been reported that lower detection limits were achieved, reproducibility was poorer with the hydrogel slides than with non-gel-coated slides [55]. Gold-coating has also been used to integrate surface plasmon resonance (SPR) [66,67] or mass spectrometry [68] as the detection system. In general, a bifunctional thio-alkylene is usually used to spontaneously form a self-assembled monolayer (SAM), which has an SH-group that reacts with the gold surface while the other free end reacts with the bio-recognition molecules. 20.3.3.2

Biomolecule Patterning

Many biomolecule patterning methods have been used such as microcontact printing (spotting), inkjet printing, DNA-directed immobilization (DDI), and microﬂuidic network (mFN) patterning. 20.3.3.2.1 Microcontact Printing Microcontact printing (spotting) delivers sub-nanoliter sample volumes directly to the surface using tiny pins that deliver a liquid ﬁlm onto the surface. It is particularly convenient and simple, requiring no extraordinary apparatus or skill and yields spots about 150–200 mm in diameter (i.e., 1600 spots=cm2) [54,69]. However, contact-printing robots cannot align their pins to the prefabricated structures and may cause damage to the substrates, especially gel-based surface materials [70]. 20.3.3.2.2 Ink-Jet Printing Ink-jet printers do not contact the printing substrate, thereby avoiding the possible damage caused by the contact [71]. Piezoelectric dispensers are the main type of ink-jet printers used for protein microarrays because they allow the recovery of the portion of the sample that is not dispensed [72] and there is no change in temperature involved in the printing process. These printers are equipped with borosilicate glass capillaries surrounded by a piezoelectric-element collar. The sample is dispensed by the application of a voltage to the piezoelectric collar, typically resulting in the release of a droplet of less than 1 nL [72]. They are generally believed to yield the lowest spot-to-spot variability in the amount of antibodies deposited [73]. However, the ink-jet microarrayer is slow when spotting many different samples and the shearing force during drop formation may cause damage to samples [74]. 20.3.3.2.3 DNA-Directed Immobilization Since DNA patterning has been successfully developed for the fabrication of DNA microarrays, protein chips can be formed on the base of a DNA chip. DNA can be a good chaperone to direct and immobilize protein at the desired spot. By designing the DNA sequence on the microchip and the complementary chaperone DNA tagged to proteins, different proteins can be directed and immobilized to the desired spots by self-assembly. Single-stranded DNA–streptavidin conjugates can be molecular adapters for the tagging of biotinylated proteins with DNA [75–78]. Hence, the setup of

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 445 16.7.2008 11:09pm Compositor Name: MSubramanian

445

Multiplexed Immunoassays in Food Analysis

the microscale ﬂuorescence immunoassay is readily conﬁgurable from the modular reagents used, i.e., covalent DNA–streptavidin conjugates, biotinylated antibodies, and a microarray containing complementary DNA capture oligomers. As an additional advantage of DDI in immunoassay applications, the binding of the target antigen by antibodies can be carried out in a homogeneous solution, instead of by heterogeneous solid-phase immunosorption, and subsequently the immunocomplexes formed are captured at the DNA microarray by nucleic acid hybridization [79,80]. Wacker et al. [78] have compared DDI with direct spotting of antibodies on chemically activated glass slides and with immobilization of biotinylated antibodies on streptavidin-coated slides. Although all three methods had an LOD of 150 pg=mL for the detection of rabbit IgG, DDI had higher ﬂuorescence intensities than streptavidin–biotin attachment and the best spot homogeneity, as well as experimental reproducibility. Also, a 100-fold less antibody is needed for preparing an array by DDI than by direct spotting. However, DDI is more difﬁcult to perform compared with the other methods. 20.3.3.2.4 Microﬂuidic Network Patterning Microﬂuidic network patterning localizes chemical reactions between the biomolecules and the surface, requiring microliters of reagent to cover square millimeter-sized areas. Immunoglobulins patterned on substrates by mFNs remain strictly conﬁned to areas enclosed by the network with submicron resolution and are viable for subsequent use in assays [81]. This approach to localize chemical reactions on surfaces is based on the deﬁnition of open networks of conduits in an elastomeric polymer, poly(dimethylsiloxane) (PDMS) formed by molding the polymer on a lithographically deﬁned master, to form mFNs. When applied to a substrate, the structured elastomer seals the surface by its conformal contact and makes linked, closed capillaries that are ﬁlled with liquid reactants and guides them along these conduits with great ﬁdelity to the pattern deﬁned in the elastomer (Figure 20.3) [82]. This patterning method is generic and suggests a practical way to incorporate biological materials on substrates such as gold, glass, or polystyrene. However, it is difﬁcult to use for high-throughput patterning.

Biomolecule 1 Biomolecule 2 Biomolecule 3

(a) Immobilization

Sample 1

2

3

(b) Sample assays

FIGURE 20.3 Pattern of biomolecules and sample assays using microﬂuidic devices: (a) biomolecules are loaded into horizontally oriented channels in the ﬂow chamber module, (b) samples ﬂow through vertically oriented channels in the sample ﬂow chamber module. (Modiﬁed from Bernard, A., Michel, B., and Delamarche, E., Anal. Chem., 73, 8, 2001.)

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 446 16.7.2008 11:09pm Compositor Name: MSubramanian

446

20.3.3.3

Handbook of Food Analysis Instruments

Assay Platforms

Most of the microchip-based immunoassays use an ELISA-type platform that consists of incubating and washing steps with shaking. They are slower than ﬂow systems and lack the possibility of simultaneous analysis of different samples on one chip because the entire chip usually can only be exposed to one sample at a time. This does not allow a direct, accurate comparison of different samples without deviations in the results caused by interchip variations, which can range between 12% and 60% depending on the coating of the microarray [55]. However, an accurate comparison between samples and standards is important for quantitative assays, which typically require approximately 10 samples for standard curves, negative controls, and several serial sample dilutions. Some other platforms based on ﬂow systems have been reported to improve the assay performances. 20.3.3.3.1 Microﬂuidic Network This platform is based on patterning lines of biomolecules onto a surface by means of a mFN [83–86]. Solutions to be analyzed are delivered by the multiple channels of a second mFN across the pattern of antigens using capillary forces [85], thus allowing a direct and accurate comparison of different samples. Individual assays are conducted using independent channels (Figure 20.3). These assays use a sequential series of samples, reagents, and buffers that are displaced one after the other over the PDMS surface, and as these assays are conducted under microﬂuidic conditions, they are fast, economical in their use of reagents, highly integrated, and yield high-quality signals [84]. After the analysis processing is completed, the ﬂow chamber module can be easily removed to allow for the optical readout. Rowe et al. [86] employed this platform to simultaneously detect staphylococcal enterotoxin B, F1 antigen from Yersinia pestis, and D-dimer in spiked clinical samples in 35 min with LODs of 50, 625, and 500 ng=mL, respectively. Although this platform is fast and sensitive, it is not easy to perform high-throughput assays because the analyte number depends on the number of microﬂuidic channels. 20.3.3.3.2 Multiple Spotting Technology Multiple spotting technology (MIST) transfers individual samples by spotting them on top of the different immobilized biomolecules, thereby allowing a multiplex analysis of different samples on a single chip. In the ﬁrst spotting, biomolecules such as antibodies are transferred to a slide and then the samples and standards, corresponding to the applied antibodies in the ﬁrst spotting, are transferred to the chip by the second spotting for interaction [87,88]. After binding, the slides are washed and scanned in a microarray scanner. This platform does not apply the samples by total incubations, but it allows the transfer of a multitude of different samples to different spots. Therefore, accurate comparisons and much smaller amounts of samples are possible since the volumes of spotted solution are only 0.19 or 0.6 nL depending on the pins [55]. This technique was able to detect down to 400 zmol of analytes. Moreover, the need for extra incubation time for the binding reaction is eliminated, thereby reducing the assay time. However, this platform relies on the precise alignment of spotting. Since those spots are extremely small (150–200 mm in diameter), very careful alignment is required to match the spots. 20.3.3.3.3 Filtration-Based Microchips A ﬁltration-based microchip utilizes a ﬁltration assay with protein microarrays printed on proteinpermeable nitrocellulose ﬁlter membranes. Instead of incubating and shaking for the binding reaction to occur as with the conventional microarray, the sample is ﬁltered through the microarray-containing membrane chip to facilitate the binding between analytes and their corresponding capture molecules [89]. Compared with protein microarrays formed on an impermeable solid surface, such as glass slides, this assay platform overcomes the diffusion limitations and enhances the assay sensitivity and speciﬁcity [89]. However, this platform does not allow for analysis of multiple samples on the same chip.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 447 16.7.2008 11:09pm Compositor Name: MSubramanian

447

Multiplexed Immunoassays in Food Analysis

20.3.3.4

Detection Systems

20.3.3.4.1 Label Detection The confocal laser scanner and the charge-coupled device (CCD) imaging detector are the two main detection systems for the analysis of ﬂuorescence-based microchips [73]. The confocal laser scanner provides more sensitive detection than CCD does, but it is slower than CCD since it scans every spot individually on the microchip. Evanescent wave excitation of a planar waveguide has been integrated with a microchip and a CCD camera to identify signals simultaneously across the entire area of the planar waveguide [83,86,90]. The evanescent wave, an electromagnetic component of the light guided down the microscope slide, is used for excitation and extends out from the surface of the microscope slide into the lower refractive index medium, decaying exponentially with distance from the surface (Figure 20.4). Because an evanescent ﬁeld only extends a few hundred nanometers into the solution, only the surface-bound ﬂuorophores are excited, thereby greatly eliminating nonspeciﬁc signals [91]. The planar waveguide technology was successfully applied to detect 1 pg=mL of interleukin-6 [92]. Another alternative non-ﬂuorescent label is resonance light-scattering colloidal gold particles that scatter light intensely and quantitative readouts can also be obtained with a CCD imaging system [73]. The LOD could be as low as 1 pg=mL depending on the antibody pair. 20.3.3.4.2 Label-Free Detection Because labeling molecules sometimes affect protein activity and are restricted to the available detection channels, label-free detection has advantages as a direct detection approach for antibody microarrays. Also, non-label methods have simpler protocols and can provide real-time measurement. SPR, mass spectrometry (MS), SPR-MS, imaging ellipsometry, atomic force microscopy (AFM), and Kelvin nanoprobe have been reported for label-free detection on microchips. Surface plasmon resonance biosensors use an evanescent ﬁeld to quantify interactions between analytes and surface-immobilized ligands by changes in surface refractive index, thus providing real-time measurement of biomolecular interactions without labeling and washing steps [66,67]. The feature of not requiring washing is especially important for low-afﬁnity antibody–antigen interactions that would not be stable if washed before analysis. Samsonova et al. [93] used the Biacore SPR biosensor to measure the antiparasitic agent ivermectin and were able to detect down to 19.1 ng=mL in bovine liver. Although SPR is a good detection system, it takes longer than ﬂuorescence-label detection for the analysis of multiple spots. Label

Evanescent field

Waveguide layer

Chip

Substrate

FIGURE 20.4 Evanescent wave excitation in a planar waveguide. The evanescent ﬁeld only excites the ﬂuorophores on or near the surface. (Modiﬁed from Joos, T.O., Stoll, D., and Templin, M.F., Curr. Opin. Chem. Biol., 6, 76, 2002.)

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 448 16.7.2008 11:09pm Compositor Name: MSubramanian

448

Handbook of Food Analysis Instruments

Mass spectrometry can be used to determine the structural features of bound proteins. Surfaceenhanced laser desorption=ionization (SELDI)-MS has been used to detect captured proteins in an array on a metal surface. With SELDI, protein arrays act as a surface to which the sample binds uniformly, and the matrix is placed on the microchip after the proteins have been attached. The captured proteins are vaporized using a laser beam, followed by the analysis of the mass spectra to reveal the identities of these proteins [68]. Matrix-assisted laser desorption=ionization coupled to a time-of-ﬂight (MALDI-TOF) mass spectrometer has also been used to sublimate and ionize the samples out of a dry, crystalline matrix via laser pulses for the MS analysis of microchips [94]. An LOD of 20 fmol of antigen=spot was obtained with chips using MALDI-TOF. However, if the matrix material is acidic, the target protein may detach from the surface and thus escape detection [95]. In comparison, the spectra obtained from SELDI are more uniform and reproducible than MALDI-TOF spectra [96]. Because SPR is used for protein quantiﬁcation as well as real-time interaction analysis, and MS is used to determine the structural features of proteins, the combination of SPR and MS offers unique capabilities for complete protein analysis. Since SPR detection is nondestructive, proteins retrieved from the SPR sensing surface can be further analyzed by MS [97]. Ellipsometry is a nondestructive method for determining ﬁlm thickness with a resolution of 0.01 nm or better based on the detection of phase shift during reﬂection of a plane of polarized light [98]. Imaging ellipsometry is an enhancement of standard single-beam ellipsometry that combines the power of ellipsometry with microscopy [99]. High-spatial resolution on the order of micrometers (laterally) and sub-nanometers (vertically) can be achieved in bioafﬁnity-based sensing by an immunosensor based on imaging ellipsometry. Wang and Jin [98] demonstrated that an immunochip based on imaging ellipsometry was able to not only detect multiple analytes simultaneously without any labeling but also monitor multiple interaction processes in real-time and in situ conditions. The AFM method detects changes in surface topography with a force probe to identify proteins captured in an antibody array. The atomically sharp probe is scanned over a surface with feedback mechanisms that enable the piezoelectric scanners to maintain the probe at a constant force (to obtain height information), or height above the sample surface (to obtain force information). The detection system does not measure force or height directly. It senses the deﬂection of the cantilever with the probe at its end. Generally, a light beam is reﬂected from the mirrored surface on the backside of the cantilever onto a position-sensitive photodetector. A small deﬂection of the cantilever will tilt the reﬂected beam and change the position of the beam on the photodetector. This approach relies on the change in the height or force that results from ligand-receptor binding, and therefore, does not require the use of labeled receptors. Height changes of 3–4 nm have been observed as a consequence of adsorption of antigenic IgG to a gold or SiO2 surface, followed by an additional increase upon antibody–antigen binding [1,100]. The Kelvin nanoprobe makes use of the principles of Kelvin physics and the AFM. The nanoprobe measures the current generated when two materials, one subjected to vibration, are connected. When contact occurs, the equilibration of the fermi levels of the two substrates leads to a current [101]. The probe uses an AFM-like tip as one of the materials and can detect both topographic and surface-potential maps of a planar surface. It has been used to detect antibody– antigen interactions in a label-free protocol through measurement of the surface potential of the biochemical pair on indium tin oxide, amine-treated slides, and gold substrates [102,103]. Although there are a variety of label-free techniques for protein microarrays, the novelty of the approach or the special expertise and equipment required appear to have limited the use of these non-ﬂuorescent approaches [104]. 20.3.3.5

Examples of Microchip Assays

Direct [105], sandwich [73,106], and competitive [67,107] multiplexed immunoassays have all been used in microchips. Below are some examples.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 449 16.7.2008 11:09pm Compositor Name: MSubramanian

Multiplexed Immunoassays in Food Analysis

449

Schweitzer et al. [106] used sandwich immunoassays to measure 75 cytokines on two separate arrays using rolling-circle ampliﬁcation (RCA). RCA labels antibodies with DNA. Once the detection antibodies are localized to the antigens on the protein array features, their DNA sequences are extended by a DNA polymerase with circular DNA in situ. Therefore, they form long DNA polymers of deﬁned sequence that are tethered to the detection antibody. After this polymerization step, the extended DNA sequence is allowed to hybridize to ﬂuorescently labeled DNA of complementary sequence. Since the extended DNA polymers are very long, multiple DNA labeled with ﬂuorophores are attached to each detection antibody, thereby amplifying the signal. Hence, detection limits as low as 0.5 pg=mL and a dynamic range of 3 decades were achieved. However, the additional procedures for signal ampliﬁcation increase the assay time and experimental difﬁculties. Sreekumar et al. [108] created antibody arrays with 146 distinct antibodies against proteins involved in stress response, cell cycle progression, and apoptosis on poly-L-lysine-coated or superaldehyde-modiﬁed glass slides. Microcontact printing was used to monitor the alternations of protein levels in LoVo colon carcinoma cells that were treated with ionizing radiation. The reference standards and samples were labeled separately using either Cyanine 5 or Cyanine 3 dyes. The slides were incubated with a labeled protein mixture and washed with buffer. The signals were detected by a confocal microarray scanner. They observed differential expression proﬁles with radiationinduced up-regulation of apoptotic regulators such as p53, DNA fragmentation factors, and tumor necrosis factor-related ligands. Delehanty and Ligler [109] developed an antibody microchip for the rapid detection of proteins and bacterial analytes. A piezoelectric noncontact dispenser was used to immobilize biotinylated capture antibodies on the surface of an avidin-coated glass slide. A microﬂuidic six-channel ﬂow module was used to conduct the assay by sequentially introducing samples, detection antibodies, and wash buffers. The signals in the microchip were measured by a scanning confocal microscope. Assays were able to be completed in 15 min, and cholera toxin, staphylococcal enterotoxin B, ricin, as well as Bacillus globibii were detected at concentrations as low as 8 ng=mL, 4 ng=mL, 10 ng=mL and 6.2 104 CFU=mL, respectively. Taitt et al. [91] patterned biotinylated antibodies in mFNs (three channels) on avidin-coated glass slides and conducted the assay with microﬂuidic devices (three channels), forming a 3 3 array for nine analytes. Evanescent wave excitation of a planar waveguide was used and a CCD camera was employed for visualization and quantiﬁcation of the spots. Staphylococcal enterotoxin B, ricin, cholera toxin, Bacillus anthracis Sterne, B. globigii, Francisella tularensis LVS, Y. pestis F1 antigen, MS2 coliphage, and Salmonella typhimurium were detectable at the concentrations of 100 ng=mL, 200 ng=mL, 100 ng=mL, 1.5 104 CFU=mL, 1 105 CFU=mL, 9 106 CFU=mL, 100 ng=mL, 1 109 PFU=mL and 5 106 CFU=mL, respectively. Knecht et al. [67] reported the simultaneous detection of 10 antibiotics in milk using an automated microarray system. They chose an indirect competitive immunoassay format of immobilizing haptens on glass slides modiﬁed with (3-glycidyloxypropyl) trimethoxysilane by a noncontact piezoelectric arrayer. Antibody binding was detected by a second antibody labeled with horseradish peroxidase generating enhanced chemiluminescence, which was recorded with a CCD camera. An analysis was carried out in milk within 5 min since all liquid handling was fully automated. The detection limits ranged from 0.12 to 32 mg=L. The LODs for all antibiotics except penicillin G were far below the maximum residue limits.

20.3.4 MICROTITER PLATE ASSAYS A typical microtiter plate contains 96 wells and some even have 384 wells; hence it is easy to immobilize biomolecules in separate wells in an array format by adsorption. Several multiplexed immunoassays using microtiter plates have been reported [110–112]. Another very interesting approach for multiplexed immunoassays using microtiter plates is to have a microarray platform at each well, which can greatly increase the assay throughput (Figure 20.5). A microprinter for

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 450 16.7.2008 11:09pm Compositor Name: MSubramanian

450

Handbook of Food Analysis Instruments

FIGURE 20.5 Microarrays on a 96-well microtiter plate. Each of the individual 96 wells contains biomolecules printed on its surface. (Modiﬁed from Rao, R.S., Visuri, S.R., McBride, M.T., Albala, J.S., Matthews, D.L., and Coleman, M.A., J. Proteome Res., 3, 736, 2004.)

microchips is adapted to print microspots by adsorption in the wells of a regular polystyrene microtiter plate [113], a silanized glass plate [114], or an N-hydroxysuccinimide-activated glass plate [115]. More than 100-spot displays in a single well have been reported [73,115]. A CCD camera is generally used to quantitatively image the arrays. Compared to a regular glass slide-based microarray, this microarray provides high sample throughput, thereby allowing a direct, accurate comparison of different samples without deviation of the results by interchip variations. In addition, it also minimizes nonspeciﬁc cross-reactivity between numerous antigen and antibody mixtures, thus maintaining the integrity of the assay [115]. However, it is more difﬁcult to print biomolecules on a microtiter plate than on a glass slide. For combining the advantages of both microtiter plates and glass slides for protein microarrays, Jones et al. [116] fabricated 96 arrays on a glass substrate to match the spacing of a microtiter plate. After printing, the glass substrate was attached to a bottomless 96-well plate using an intervening silicone gasket. Although this technique was used for the study of quantitative protein interaction, it can also be adapted to multiplexed immunoassays.

20.3.5 BEAD ASSAYS Bead-based immunoassays immobilize biomolecules, such as antibodies, onto distinct color-coded polystyrene microspheres embedded with precise ratios of red and infrared ﬂuorescent dyes [117]. Each color-coded bead is identiﬁed by a ﬂuorescent signal measured in a ﬂow cytometer. Fluorescently labeled detection antibodies are employed as sandwich immunoassays to quantify the amount of captured targets on each individual bead. The signal intensities from labeled detection antibodies

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 451 16.7.2008 11:09pm Compositor Name: MSubramanian

451

Multiplexed Immunoassays in Food Analysis

Immunobeads Antigens

+

Labeled antibodies

+

(a)

Lasers (b)

FIGURE 20.6 (See color insert following page 240.) Bead-based multiplexed immunoassays: (a) antibody–antigen complexes are formed after immunobeads incubate with samples. After the addition of labeled antibodies, the sandwich complexes are formed and (b) beads are analyzed by ﬂuorescence ﬂow cytometry. A red laser classiﬁes the bead and a green laser quantiﬁes the antigens.

are measured by a second ﬂuorescent signal measurement. In general, a ﬂow cytometer is comprised of a red and a green laser, which classiﬁes the beads and quantiﬁes the antigens, respectively (Figure 20.6). Only the ﬂuorophores that are bound to the surface of the beads are counted in the ﬂow cytometer, so it is possible to perform the assay without washing steps. While it has been possible to produce up to 100 distinct beads [118], compared to the thousands of spots on a microchip, beadbased immunoassays still provide a lower throughput. Rao et al. [118] used the Luminex bead-based system to simultaneously detect bacterial and viral proteins and compared them with microchip-based multiplexed immunoassays. The beadbased system showed lower LODs of 7.05 104 CFU=mL and 3.51 106 PFU=mL for the detection of B. globigii and the RNA bacteriophage virus (MS2), respectively, while microchips were more amenable to miniaturization. McBride et al. [117] also used the same system to demonstrate the simultaneous detection of four simulants of biological warfare agents, comprising a virus (MS2), protein toxins (ovalbumin), bacterial spores (B. globigii spores), and vegetative cells (Erwinia herbicola) in 1 h with LODs of 4.2 107 PFU=mL, 1 mg=L, 1.5 104 CFU=mL, and 5 104 CFU=mL, respectively.

20.3.6 ELECTRODE ASSAYS Although most multiplexed immunoassays are based on optical biosensors, amperometric multiplexed immunoassays have been developed as well [119–121]. Different antibodies are immobilized on a variety of working electrodes [120,121] or on the surface next to the working electrodes [119] by adsorption [119], photolithography [120], or DDI [121]. Horseradish peroxidase or alkaline phosphatase was labeled to the detection antibody. An electric current was produced, and

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 452 16.7.2008 11:09pm Compositor Name: MSubramanian

452

Handbook of Food Analysis Instruments

thus detected by the electrodes, after the enzyme reacted with the substrate. Simultaneous detection of ﬁve analytes (a1 acid glycoprotein, ricin, M13 phage, B. globigii spores, and ﬂuorescein) based on the redox enzyme (horseradish peroxidase) ampliﬁed electrochemical detection was demonstrated recently by Dill et al. using microelectrode arrays [121]. Horseradish peroxidase catalyzed the oxidation of a substrate, while using peroxide as the electron acceptor, yielding a high-rate enzyme turnover. The LODs were in the attomole range and the dynamic range was 4–5 decades of analyte concentration with an assay volume of 50 mL or less. Because complementary metal oxide semiconductor (CMOS)-based silicon chip was employed, the microelectrode arrays could provide over 1000 electrodes=cm2. However, it seems that only DDI is suitable for use on microelectrodes among the entire antibody patterning methods mentioned above.

20.3.7 NANOWIRE ASSAYS Another kind of electrical multiplexed biosensors are silicon-nanowire ﬁeld-effect devices [122,123]. Unlike most electrode assays, it is label-free, real time, and antibodies can be spotted on the nanowire surfaces by microchip arrayers. Analytes will cause sensitive changes in conductance when they bind to antibodies on the nanowire surfaces. Zheng et al. [123] have demonstrated the multiplexed detection of cancer markers, prostate-speciﬁc antigen (PSA), PSA-a1-antichymotrypsin, carcinoembryonic antigen, and mucin-1 using the nanowire sensor chip. Protein-sensing experiments were performed in microﬂuidic channels formed by a PDMS polymer sealed to the chip. Concentrations as low as 0.9 pg=mL of PSA can be detected in undiluted serum samples. The authors also compared their work to SPR detection of marker proteins [124–126], which showed the LOD of 10–100 pg=mL. Although nanowire arrays can perform a sensitive, label-free, and real-time detection, the fabrication was complicated, including photolithography and metal deposition steps [123]. Also, the simultaneous real-time measurements were limited to three distinct nanowire sensor devices because of the availability of measurement electronics. Although at least 100 independently addressable sensor elements were available in the arrays and could be used with more sophisticated multiplexing electronics [123], it was still relatively low throughput, compared to thousands of spots on protein microarrays.

20.4 MULTIPLEXED IMMUNOASSAY CHALLENGES Although there are many monoclonal and polyclonal antibodies commercially available, most are very expensive and many of them do not have high enough afﬁnity and speciﬁcity for multiplexed immunoassays. To replace expensive, low-afﬁnity, or low-speciﬁcity antibody, recombinant antibodies have been made by phage display [127,128], ribosomal display [129,130], or yeast display [131,132]. Different antibodies display widely varying performance in immunoassays attributed to varying afﬁnities [111,133]. Phage display of antibody fragments is the most popular method for generating recombinant antibodies and has been successfully used to increase antibody afﬁnity more than 1000-fold [134]. Another great advantage of recombinant antibodies is the relatively small molecular weight, thereby facilitating high-density immobilization on a support surface. Also, antibody engineering for the oriented attachment on the solid surface is also possible [135]. In addition to the challenge of antibody resources, the assay performance is also a big challenge. Different antibodies have different optimum performance conditions such as pH, ionic strength, temperature, and time. Sometimes it is necessary to sacriﬁce individual antibody performance to achieve the overall performance of multiplexed immunoassays. The labeling of antibodies is also a challenge for multiplexed immunoassays, especially on a massively parallel scale. The strategies for conjugating labels to antibodies mostly involve covalent binding, which is usually complicated and time consuming [136–139]. In addition, these methods generally require the use of the amino groups on the antibody. Even the non-covalent biotin-(strept)

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 453 16.7.2008 11:09pm Compositor Name: MSubramanian

Multiplexed Immunoassays in Food Analysis

453

avidin coupling method also involves the biotinylation of antibodies [140]. These approaches, however, are limited because most antibodies contain randomly distributed amino groups, leading to multiple attachment sites. The random nature of this attachment can cause some of the conjugated antibodies to lose antigen-binding activity because of direct chemical modiﬁcation or steric hindrance of the antigen-binding site [141,142]. To have a simple labeling procedure and improve the afﬁnity of labeled antibodies, IgG Fc-binding proteins, like protein A or protein G, can be used to couple to the antibodies. A genetically fused protein A-luciferase has been developed for bioluminescent immunoassays [143]. Protein G also binds to the IgG Fc fragment, and it represents a more general and versatile IgG-binding reagent [144]. It binds a wider range of IgG subclasses and a greater variety of mammalian species with a higher afﬁnity than protein A. For example, protein G has a strong afﬁnity to goat IgG, but protein A barely binds to goat IgG. In addition, protein G is not as pH dependent as protein A when binding to immunoglobulins [144–146]. Protein G-based universal reagents have been developed for time-resolved immunoﬂuorometry [147]. Up to eight europium atoms were tagged on each protein G, but this greatly reduced the IgG-binding ability of protein G. Another universal reagent for immunoassays was recently developed: protein G-liposomal nanovesicles that were generated by coupling a large amount of protein G to each liposome, thereby not only providing great ampliﬁcation, but also retaining the IgG-binding ability of protein G (decreased by only 5.3%) [111,148]. Liposomal nanovesicles, i.e., liposomes, are spherical vesicles consisting of phospholipid bilayers surrounding an aqueous cavity. Because each liposomal nanovesicle can contain up to several million ﬂuorescent dye molecules, thereby providing greatly enhanced signals, liposomal nanovesicles have been successfully used as reporter particles in immunoassays [149–152]. The process of antibody coupling to protein G-liposomal nanovesicles is simple and rapid (30 min or less) without reducing the afﬁnity of the labeled antibody. A universal reagent like protein G-liposomal nanovesicles is desirable for facile labeling of a large number of different antibodies for multiplexed immunoassays.

20.5 CONCLUSIONS Only a few multiple-label immunoassays, such as multiple lanthanides and multiple nanocrystals, can simultaneously detect more than three analytes because of the difﬁculty of discriminating a larger number of distinct simultaneously detectable labels. Spatially resolved immunoassays, however, can easily provide higher throughput, especially in miniaturized multiplexed immunoassays. Because of advances in micro=nanotechnology, antibody microarrays using microchips, beads, microtiter plates, microelectrodes, and nanowires can operate in a high-throughput mode to detect more than 100 analytes simultaneously. However, only microchip-based immunoassays have really demonstrated high throughput. Continued research in nanotechnology is expected to be the key to increase the throughput and reduce the assay time and cost. Recombinant antibody technology will also play a major role in improving the sensitivity and speciﬁcity of the simultaneous multianalyte immunoassays. Thus, future developments in the promising ﬁeld of simultaneous multianalyte immunoassays are linked to the continued progress in nanotechnology and recombinant antibody technology.

ACKNOWLEDGMENTS The authors thank I-Yuan Chiang for assistance in drawing the ﬁgures. This research was supported in part by the Cornell University Agricultural Experiment Station Federal Formula Funds, Project No. NYG 623498, received from Cooperative State Research, Education, and Extension Service, U.S. Department of Agriculture. Any opinions, ﬁndings, conclusions, or recommendations expressed in this publication are those of the authors and do not necessarily reﬂect the view of the U.S. Department of Agriculture.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 454 16.7.2008 11:09pm Compositor Name: MSubramanian

454

Handbook of Food Analysis Instruments

REFERENCES 1. Joos TO, Stoll D, and Templin MF: Miniaturised multiplexed immunoassays, Curr. Opin. Chem. Biol. 2002, 6:76–80. 2. Kricka LJ: Multianalyte testing, Clin. Chem. 1992, 38:327–328. 3. Yvert JP and Delagneau JF: Simultaneous multiple immunoassays: Current data, Ann. Biol. Clin. (Paris) 1993, 51:43–48. 4. Morgan CR: Immunoassay of human insulin and growth hormone simultaneously using I-131 and I-125 tracers, Proc. Soc. Exp. Biol. Med. 1966, 123:230–233. 5. Wians FH, Jr., Dev J, Powell MM, and Heald JI: Evaluation of simultaneous measurement of lutropin and follitropin with the SimulTROPIN radioimmunoassay kit, Clin. Chem. 1986, 32:887–890. 6. Gutcho S and Mansbach L: Simultaneous radioassay of serum vitamin B12 and folic acid, Clin. Chem. 1977, 23:1609–1614. 7. Gow SM, Caldwell G, Toft AD, and Beckett GJ: SimulTRAC simultaneous radioimmunoassay of thyrotropin and free thyroxin evaluated, Clin. Chem. 1986, 32:2191–2194. 8. Nakamura RM, Kasahara Y, and Rechnitz GA: Immunochemical Assays and Biosensor Technology for the 1990s, Christopoulos TK (Ed.), Washington DC: American Society for Microbiology, 1992. 9. Diamandis EP: Immunoassays with time-resolved ﬂuorescence spectroscopy: Principles and applications, Clin. Biochem. 1988, 21:139–150. 10. Hemmila I, Holttinen S, Pettersson K, and Lovgren T: Double-label time-resolved immunoﬂuorometry of lutropin and follitropin in serum, Clin. Chem. 1987, 33:2281–2283. 11. Saarma M, Jarvekulg L, Hemmila I, Siitari H, and Sinijarv R: Simultaneous quantiﬁcation of two plant viruses by double-label time-resolved immunoﬂuorometric assay, J. Virol. Methods 1989, 23:47–54. 12. Bookout JT, Joaquim TR, Magin KM, Rogan GJ, and Lirette RP: Development of a dual-label timeresolved ﬂuorometric immunoassay for the simultaneous detection of two recombinant proteins in potato, J. Agric. Food Chem. 2000, 48:5868–5873. 13. Eriksson S, Vehniainen M, Jansen T, Meretoja V, Saviranta P, Pettersson K, and Lovgren T: Dual-label time-resolved immunoﬂuorometric assay of free and total prostate-speciﬁc antigen based on recombinant Fab fragments, Clin. Chem. 2000, 46:658–666. 14. Xu YY, Pettersson K, Blomberg K, Hemmila I, Mikola H, and Lovgren T: Simultaneous quadruplelabel ﬂuorometric immunoassay of thyroid-stimulating hormone, 17 alpha-hydroxyprogesterone, immunoreactive trypsin, and creatine kinase MM isoenzyme in dried blood spots, Clin. Chem. 1992, 38:2038–2043. 15. Sidki AM, Smith DS, and Landon J: Dual-label ﬂuoroimmunoassay for simultaneous determination of primidone and phenobarbital, Ther. Drug Monitoring 1985, 7:101–107. 16. Pachmann K and Killander D: Simultaneous quantitation of two antigens in mixture in individual cells by microimmunoﬂuorimetry, J. Immunol. Methods 1976, 12:67–79. 17. Gillis TP and Thompson JJ: Double-label ﬂuorescence immunoassay of bacteria, J. Clin. Microbiol. 1978, 8:351–353. 18. Mathai M, Siebert KJ, Siebert STA, Durst RA, and Reeves SG: Development of a signal measurement technique for application in multi-analyte liposome immunomigration assays, Food Agric. Immunol. 1998, 10:195–202. 19. Blake C, Al-Bassam MN, Gould BJ, Marks V, Bridges JW, and Riley C: Simultaneous enzyme immunoassay of two thyroid hormones, Clin. Chem. 1982, 28:1469–1473. 20. Gosling JP: A decade of development in immunoassay methodology, Clin. Chem. 1990, 36:1408–1427. 21. Spencer K, Macri JN, Anderson RW, Aitken DA, Berry E, Crossley JA, Wood PJ, Coombes EJ, Stroud M, Worthington DJ, et al.: Dual analyte immunoassay in neural tube defect and Down’s syndrome screening: Results of a multicentre clinical trial, Ann. Clin. Biochem. 1993, 30(Pt 4):394–401. 22. Macri JN, Spencer K, and Anderson R: Dual Analyte immunoassay—A new approach to neural tube defect and Down’s syndrome screening, Ann. Clin. Biochem. 1992, 29(Pt 4):390–396. 23. Dean KJ, Thompson SG, Burd JF, and Buckler RT: Simultaneous determination of phenytoin and phenobarbital in serum or plasma by substrate-labeled ﬂuorescent immunoassay, Clin. Chem. 1983, 29:1051–1056.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 455 16.7.2008 11:09pm Compositor Name: MSubramanian

Multiplexed Immunoassays in Food Analysis

455

24. Porstmann T, Nugel E, Henklein P, Dopel H, Ronspeck W, Pas P, and von Baehr R: Two-colour combination enzyme-linked immunosorbent assay for the simultaneous detection of HBV and HIV infection, J. Immunol. Methods 1993, 158:95–106. 25. DeMars DD, Katzmann JA, Kimlinger TK, Calore JD, and Tracy RP: Simultaneous measurement of total and IgA-conjugated alpha 1-microglobulin by a combined immunoenzyme=immunoradiometric assay technique, Clin. Chem. 1989, 35:766–772. 26. Salmain M, Vessieres A, Brossier P, and Jaouen G: Use of Fourier transform infrared spectroscopy for the simultaneous quantitative detection of metal carbonyl tracers suitable for multilabel immunoassays, Anal. Biochem. 1993, 208:117–120. 27. Hayes FJ, Halsall HB, and Heineman WR: Simultaneous immunoassay using electrochemical detection of metal ion labels, Anal. Chem. 1994, 66:1860–1865. 28. Bordes AL, Limoges B, Brossier P, and Degrand C: Simultaneous homogeneous immunoassay of phenytoin and phenobarbital using a naﬁon-loaded carbon paste electrode and two redox cationic labels, Anal. Chim. Acta 1997, 356:195–203. 29. Bordes AL, Schollhorn B, Limoges B, and Degrand C: Simultaneous detection of three drugs labeled by cationic metal complexes at a naﬁon-loaded carbon paste electrode, Talanta 1999, 48:201–208. 30. Alivisatos AP: Semiconductor clusters, nanocrystals, and quantum dots, Science 1996, 271:933–937. 31. Mattoussi H, Radzilowski LH, Dabbousi BO, Thomas EL, Bawendi MG, and Rubner MF: Electroluminescence from heterostructures of poly(phenylene vinylene) and inorganic CdSe nanocrystals, J. Appl. Phys. 1998, 83:7965–7974. 32. Goldman ER, Clapp AR, Anderson GP, Uyeda HT, Mauro JM, Medintz IL, and Mattoussi H, Multiplexed toxin analysis using four colors of quantum dot ﬂuororeagents, Anal. Chem. 2004, 76(3): 684–688. 33. Chan WC and Nie S: Quantum dot bioconjugates for ultrasensitive nonisotopic detection, Science 1998, 281:2016–2018. 34. Liu G, Wang J, Kim J, Jan MR, and Collins GE: Electrochemical coding for multiplexed immunoassays of proteins, Anal. Chem. 2004, 76:7126–7130. 35. Lee N, Zhang SQ, and Testa D: A rapid multicolor Western blot, J. Immunol. Methods 1988, 106:27–30. 36. Ijsselmuiden OE, Beelaert G, Schouls LM, Tank B, Stolz E, and van der Groen G: Line immunoassay and enzyme-linked line immunoﬁltration assay for simultaneous detection of antibody to two treponemal antigens, Eur. J. Clin. Microbiol. Infect. Dis. 1989, 8:716–721. 37. Donohue J, Bailey M, Gray R, Holen J, Huang TM, Keevan J, Mattimiro C, Putterman C, Stalder A, and Defreese J: Enzyme immunoassay system for panel testing, Clin. Chem. 1989, 35:1874–1877. 38. Buechler KF, Moi S, Noar B, McGrath D, Villela J, Clancy M, Shenhav A, Colleymore A, Valkirs G, Lee T, et al.: Simultaneous detection of seven drugs of abuse by the Triage panel for drugs of abuse, Clin. Chem. 1992, 38:1678–1684. 39. Wallace JF, Pugia MJ, Lott JA, Luke KE, Shihabi ZK, Sheehan M, and Bucksa JM: Multisite evaluation of a new dipstick for albumin, protein, and creatinine, J. Clin. Lab. Anal. 2001, 15:231–235. 40. Jeong DS and Choi EY: Simultaneous quantitative determination of multiple analytes with ﬂuorescencetagged probes by immunochromatography, Korean J. Biol. Sci. 2003, 7:89–92. 41. Ahn-Yoon S, DeCory TR, Baeumner AJ, and Durst RA: Ganglioside-liposome immunoassay for the ultrasensitive detection of cholera toxin, Anal. Chem. 2003, 75:2256–2261. 42. Gabaldon JA, Cascales JM, Morias S, Maquieira A, and Puchades R: Determination of atrazine and carbaryl pesticide residues in vegetable samples using a multianalyte dipstick immunoassay format, Food Addit. Contam. 2003, 20:707–715. 43. Rubtsova MY, Samsonova JV, Egorov AM, and Schmid RD: Simultaneous determination of several pesticides with chemiluminescent immunoassay on a multi-spot membrane strip, Food Agric. Immunol. 1998, 10:223–235. 44. Schneider E, Usleber E, Martlbauer E, Dietrich R, and Terplan G: Multimycotoxin dipstick enzyme immunoassay applied to wheat, Food Addit. Contam. 1995, 12:387–393. 45. Abouzied MM and Pestka JJ: Simultaneous screening of fumonisin B1, aﬂatoxin B1, and zearalenone by line immunoblot: a computer-assisted multianalyte assay system, J. AOAC Int. 1994, 77:495–501. 46. Koch S, Wolf H, Danapel C, and Feller KA: Optical ﬂow-cell multichannel immunosensor for the detection of biological warfare agents, Biosens. Bioelectron. 2000, 14:779–784.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 456 16.7.2008 11:09pm Compositor Name: MSubramanian

456

Handbook of Food Analysis Instruments

47. Mastichiadis C, Kakabakos SE, Christoﬁdis I, Koupparis MA, Willetts C, and Misiako K: Simultaneous determination of pesticides using a four-band disposable optical capillary immunosensor, Anal. Chem. 2002, 74:6064–6072. 48. Petrou PS, Kakabakos SE, Christoﬁdis I, Argitis P, and Misiakos K: Multi-analyte capillary immunosensor for the determination of hormones in human serum samples, Biosens. Bioelectron. 2002, 17:261–268. 49. Narang U, Gauger PR, Kusterbeck AW, and Ligler FS: Multianalyte detection using a capillary-based ﬂow immunosensor, Anal. Biochem. 1998, 255:13–19. 50. Cahill DJ: Protein and antibody arrays and their medical applications, J. Immunol. Methods 2001, 250:81–91. 51. Lueking A, Horn M, Eickhoff H, Bussow K, Lehrach H, and Walter G: Protein microarrays for gene expression and antibody screening, Anal. Biochem. 1999, 270:103–111. 52. Holt LJ, Bussow K, Walter G, and Tomlinson IM: By-passing selection: Direct screening for antibody– antigen interactions using protein arrays, Nucleic Acids Res. 2000, 28:E72. 53. Templin MF, Stoll D, Schrenk M, Traub PC, Vohringer CF, and Joos TO: Protein microarray technology, Trends Biotechnol. 2002, 20:160–166. 54. MacBeath G and Schreiber SL: Printing proteins as microarrays for high-throughput function determination, Science 2000, 289:1760–1763. 55. Angenendt P, Glokler J, Murphy D, Lehrach H, and Cahill DJ: Toward optimized antibody microarrays: A comparison of current microarray support materials, Anal. Biochem. 2002, 309:253–260. 56. Haab BB, Dunham MJ, and Brown PO: Protein microarrays for highly parallel detection and quantitation of speciﬁc proteins and antibodies in complex solutions, Genome Biol. 2001, 2:RESEARCH0004. 57. Kusnezow W and Hoheisel JD: Solid supports for microarray immunoassays, J. Molecular Recognition 2003, 16:165–176. 58. Kusnezow W, Jacob A, Walijew A, Diehl F, and Hoheisel JD: Antibody microarrays: An evaluation of production parameters, Proteomics 2003, 3:254–264. 59. Angenendt P, Glokler J, Sobek J, Lehrach H, and Cahill DJ: Next generation of protein microarray support materials: Evaluation for protein and antibody microarray applications, J. Chromatogr. A 2003, 1009:97–104. 60. Rubina AY, Dementieva EI, Stomakhin AA, Darii EL, Pan’kov SV, Barsky VE, Ivanov SM, Konovalova EV, and Mirzabekov AD: Hydrogel-based protein microchips: Manufacturing, properties, and applications, Biotechniques 2003, 34:1008–1014, 1016–1020, 1022. 61. Rubina AY, Dyukova VI, Dementieva EI, Stomakhin AA, Nesmeyanov VA, Grishin EV, and Zasedatelev AS: Quantitative immunoassay of biotoxins on hydrogel-based protein microchips, Anal. Biochem. 2005, 340:317–329. 62. Afanassiev V, Hanemann V, and Wolﬂ S: Preparation of DNA and protein micro arrays on glass slides coated with an agarose ﬁlm, Nucleic Acids Res. 2000, 28:E66. 63. Stillman BA and Tonkinson JL: FAST slides: A novel surface for microarrays, Biotechniques 2000, 29:630–635. 64. Wang CC, Huang RP, Sommer M, Lisoukov H, Huang R, Lin Y, Miller T, and Burke J: Array-based multiplexed screening and quantitation of human cytokines and chemokines, J. Proteome Res. 2002, 1:337–343. 65. Miller JC, Zhou H, Kwekel J, Cavallo R, Burke J, Butler EB, Teh BS, and Haab BB: Antibody microarray proﬁling of human prostate cancer sera: Antibody screening and identiﬁcation of potential biomarkers, Proteomics 2003, 3:56–63. 66. Yuk JS and Ha KS: Proteomic applications of surface plasmon resonance biosensors: Analysis of protein arrays, Exp. Molecular Med. 2005, 37:1–10. 67. Knecht BG, Strasser A, Dietrich R, Martlbauer E, Niessner R, and Weller MG: Automated microarray system for the simultaneous detection of antibiotics in milk, Anal. Chem. 2004, 76:646–654. 68. Davies H, Lomas L, and Austen B: Proﬁling of amyloid beta peptide variants using SELDI protein chip arrays, Biotechniques 1999, 27:1258–1261. 69. Bernard A, Fitzli D, Sonderegger P, Delamarche E, Michel B, Bosshard HR, and Biebuyck H: Afﬁnity capture of proteins from solution and their dissociation by contact printing, Nat. Biotechnol. 2001, 19:866–869. 70. Zhu H and Snyder M: Protein chip technology, Curr. Opin. Chem. Biol. 2003, 7:55–63.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 457 16.7.2008 11:09pm Compositor Name: MSubramanian

Multiplexed Immunoassays in Food Analysis

457

71. Roda A, Guardigli M, Russo C, Pasini P, and Baraldini M: Protein microdeposition using a conventional ink-jet printer, Biotechniques 2000, 28:492–496. 72. Delehanty JB: Printing functional protein microarrays using piezoelectric capillaries, Methods Molecular Biol. 2004, 264:135–143. 73. Nielsen UB and Geierstanger BH: Multiplexed sandwich assays in microarray format, J. Immunol. Methods 2004, 290:107–120. 74. Haab BB: Advances in protein microarray technology for protein expression and interaction proﬁling, Curr. Opin. Drug Discovery Dev. 2001, 4:116–123. 75. Niemeyer CM, Boldt L, Ceyhan B, and Blohm D: DNA-directed immobilization: Efﬁcient, reversible, and site-selective surface binding of proteins by means of covalent DNA-streptavidin conjugates, Anal. Biochem. 1999, 268:54–63. 76. Niemeyer CM, Sano T, Smith CL, and Cantor CR: Oligonucleotide-directed self-assembly of proteins: Semisynthetic DNA–streptavidin hybrid molecules as connectors for the generation of macroscopic arrays and the construction of supramolecular bioconjugates, Nucleic Acids Res. 1994, 22:5530–5539. 77. Niemeyer CM: The developments of semisynthetic DNA-protein conjugates, Trends Biotechnol. 2002, 20:395–401. 78. Wacker R, Schroder H, and Niemeyer CM: Performance of antibody microarrays fabricated by either DNA-directed immobilization, direct spotting, or streptavidin-biotin attachment: A comparative study, Anal. Biochem. 2004, 330:281–287. 79. Wacker R and Niemeyer CM: DDI-microFIA—A readily conﬁgurable microarray-ﬂuorescence immunoassay based on DNA-directed immobilization of proteins, Chembiochem 2004, 5:453–459. 80. Niemeyer CM, Wacker R, and Adler M: Combination of DNA-directed immobilization and immunoPCR: Very sensitive antigen detection by means of self-assembled DNA-protein conjugates, Nucleic Acids Res. 2003, 31:e90. 81. Delamarche E, Bernard A, Schmid H, Michel B, and Biebuyck H: Patterned delivery of immunoglobulins to surfaces using microﬂuidic networks, Science 1997, 276:779–781. 82. Delamarche E, Bernard A, Schimid H, Bietsch A, Michel B, and Biebuyck H: Microﬂuidic networks for chemical patterning of substrates: Design and application to bioassays, J. Am. Chem. Soc. 1998, 120: 500–508. 83. Rowe CA, Tender LM, Feldstein MJ, Golden JP, Scruggs SB, MacCraith BD, Cras JJ, and Ligler FS: Array biosensor for simultaneous identiﬁcation of bacterial, viral, and protein analytes, Anal. Chem. 1999, 71:3846–3852. 84. Wolf M, Juncker D, Michel B, Hunziker P, and Delamarche E: Simultaneous detection of C-reactive protein and other cardiac markers in human plasma using micromosaic immunoassays and self-regulating microﬂuidic networks, Biosens. Bioelectron. 2004, 19:1193–1202. 85. Bernard A, Michel B, and Delamarche E: Micromosaic immunoassays, Anal. Chem. 2001, 73:8–12. 86. Rowe CA, Scruggs SB, Feldstein MJ, Golden JP, and Ligler FS: An array immunosensor for simultaneous detection of clinical analytes, Anal. Chem. 1999, 71:433–439. 87. Angenendt P, Wilde J, Kijanka G, Baars S, Cahill DJ, Kreutzberger J, Lehrach H, Konthur Z, and Glokler J: Seeing better through a MIST: Evaluation of monoclonal recombinant antibody fragments on microarrays, Anal. Chem. 2004, 76:2916–2921. 88. Angenendt P, Glokler J, Konthur Z, Lehrach H, and Cahill DJ: 3D protein microarrays: Performing multiplex immunoassays on a single chip, Anal. Chem. 2003, 75:4368–4372. 89. Xu Y and Bao G: A ﬁltration-based protein microarray technique, Anal. Chem. 2003, 75:5345–5351. 90. Rowe-Taitt CA, Hazzard JW, Hoffman KE, Cras JJ, Golden JP, and Ligler FS: Simultaneous detection of six biohazardous agents using a planar waveguide array biosensor, Biosens. Bioelectron. 2000, 15:579–589. 91. Taitt CR, Anderson GP, Lingerfelt BM, Feldstein MJ, and Ligler FS: Nine-analyte detection using an array-based biosensor, Anal. Chem. 2002, 74:6114–6120. 92. Weinberger SR, Morris TS, and Pawlak M: Recent trends in protein biochip technology, Pharmacogenomics 2000, 1:395–416. 93. Samsonova JV, Baxter GA, Crooks SR, Small AE, and Elliott CT: Determination of ivermectin in bovine liver by optical immunobiosensor, Biosens. Bioelectron. 2002, 17:523–529. 94. Borrebaeck CA, Ekstrom S, Hager AC, Nilsson J, Laurell T, and Marko-Varga G: Protein chips based on recombinant antibody fragments: A highly sensitive approach as detected by mass spectrometry, Biotechniques 2001, 30:1126–1130, 1132.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 458 16.7.2008 11:09pm Compositor Name: MSubramanian

458

Handbook of Food Analysis Instruments

95. Williams C and Addona TA: The integration of SPR biosensors with mass spectrometry: Possible applications for proteome analysis, Trends Biotechnol. 2000, 18:45–48. 96. Fung ET, Thulasiraman V, Weinberger SR, and Dalmasso EA: Protein biochips for differential proﬁling, Curr. Opin. Biotechnol. 2001, 12:65–69. 97. Nedelkov D and Nelson RW: Surface plasmon resonance mass spectrometry: Recent progress and outlooks, Trends Biotechnol. 2003, 21:301–305. 98. Wang ZH and Jin G: A label-free multisensing immunosensor based on imaging ellipsometry, Anal. Chem. 2003, 75:6119–6123. 99. Jin G, Tengvall P, Lundstrom I, and Arwin H: A biosensor concept based on imaging ellipsometry for visualization of biomolecular interactions, Anal. Biochem. 1995, 232:69–72. 100. Jones VW, Kenseth JR, Porter MD, Mosher CL, and Henderson E: Microminiaturized immunoassays using atomic force microscopy and compositionally patterned antigen arrays, Anal. Chem. 1998, 70:1233–1241. 101. Angenendt P: Progress in protein and antibody microarray technology, Drug Discovery Today 2005, 10:503–511. 102. Cheran LE, Chacko M, Zhang M, and Thompson M: Protein microarray scanning in label-free format by Kelvin nanoprobe, Analyst 2004, 129:161–168. 103. Thompson M, Cheran LE, Zhang M, Chacko M, Huo H, and Sadeghi S: Label-free detection of nucleic acid and protein microarrays by scanning Kelvin nanoprobe, Biosens. Bioelectron. 2005, 20:1471–1481. 104. Rucker VC, Havenstrite KL, and Herr AE: Antibody microarrays for native toxin detection, Anal. Biochem. 2005, 339:262–270. 105. Espina V, Woodhouse EC, Wulfkuhle J, Asmussen HD, Petricoin EF, III, and Liotta LA: Protein microarray detection strategies: Focus on direct detection technologies, J. Immunol. Methods 2004, 290:121–133. 106. Schweitzer B, Roberts S, Grimwade B, Shao W, Wang M, Fu Q, Shu Q, Laroche I, Zhou Z, Tchernev VT, et al.: Multiplexed protein proﬁling on microarrays by rolling-circle ampliﬁcation, Nat. Biotechnol. 2002, 20:359–365. 107. Barry R, Diggle T, Terrett J, and Soloviev M: Competitive assay formats for high-throughput afﬁnity arrays, J. Biomolecular Screening 2003, 8:257–263. 108. Sreekumar A, Nyati MK, Varambally S, Barrette TR, Ghosh D, Lawrence TS, and Chinnaiyan AM: Proﬁling of cancer cells using protein microarrays: Discovery of novel radiation-regulated proteins, Cancer Res. 2001, 61:7585–7593. 109. Delehanty JB and Ligler FS: A microarray immunoassay for simultaneous detection of proteins and bacteria, Anal. Chem. 2002, 74:5681–5687. 110. Samsonova JV, Rubtsova M, Kiseleva AV, Ezhov AA, and Egorov AM: Chemiluminescent multiassay of pesticides with horseradish peroxidase as a label, Biosens. Bioelectron. 1999, 14:273–281. 111. Chen C-S and Durst RA: Simultaneous detection of Escherichia coli O157:H7, Salmonella spp., and Listeria monocytogenes with an array-based immunosorbent assay using universal protein G-liposomal nanovesicles, Talanta 2006, 69:232–238. 112. Bhand S, Surugiu I, Dzgoev A, Ramanathan K, Sundaram PV, and Danielsson B: Immuno-arrays for multianalyte analysis of chlorotriazines, Talanta 2005, 65:331–336. 113. Moody MD, Van Arsdell SW, Murphy KP, Orencole SF, and Burns C: Array-based ELISAs for highthroughput analysis of human cytokines, Biotechniques 2001, 31:186–190, 192–194. 114. Wiese R, Belosludtsev Y, Powdrill T, Thompson P, and Hogan M: Simultaneous multianalyte ELISA performed on a microarray platform, Clin. Chem. 2001, 47:1451–1457. 115. Mendoza LG, McQuary P, Mongan A, Gangadharan R, Brignac S, and Eggers M: High-throughput microarray-based enzyme-linked immunosorbent assay (ELISA), Biotechniques 1999, 27:778–780, 782–786, 788. 116. Jones RB, Gordus A, Krall JA, and Macbeath G: A quantitative protein interaction network for the ErbB receptors using protein microarrays, Nature 2006, 439:168–174. 117. McBride MT, Gammon S, Pitesky M, O’Brien TW, Smith T, Aldrich J, Langlois RG, Colston B, and Venkateswaran KS: Multiplexed liquid arrays for simultaneous detection of simulants of biological warfare agents, Anal. Chem. 2003, 75:1924–1930. 118. Rao RS, Visuri SR, McBride MT, Albala JS, Matthews DL, and Coleman MA: Comparison of multiplexed techniques for detection of bacterial and viral proteins, J. Proteome Res. 2004, 3:736–742.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 459 16.7.2008 11:09pm Compositor Name: MSubramanian

Multiplexed Immunoassays in Food Analysis

459

119. Ding Y, Zhou L, Halsall HB, and Heineman WR: Feasibility studies of simultaneous multianalyte amperometric immunoassay based on spatial resolution, J. Pharm. Biomed. Anal. 1999, 19:153–161. 120. Pritchard DJ, Morgan H, and Cooper JM: Simultaneous determination of follicle stimulating hormone and luteinising hormone using a multianalyte immunosensor, Anal. Chim. Acta 1995, 310:251–256. 121. Dill K, Montgomery DD, Ghindilis AL, Schwarzkopf KR, Ragsdale SR, and Oleinikov AV: Immunoassays based on electrochemical detection using microelectrode arrays, Biosens. Bioelectron. 2004, 20:736–742. 122. Cui Y, Wei Q, Park H, and Lieber CM: Nanowire nanosensors for highly sensitive and selective detection of biological and chemical species, Science 2001, 293:1289–1292. 123. Zheng G, Patolsky F, Cui Y, Wang WU, and Lieber CM: Multiplexed electrical detection of cancer markers with nanowire sensor arrays, Nat. Biotechnol. 2005, 23:1294–1301. 124. Campagnolo C, Meyers KJ, Ryan T, Atkinson RC, Chen YT, Scanlan MJ, Ritter G, Old LJ, and Batt CA: Real-time, label-free monitoring of tumor antigen and serum antibody interactions, J. Biochem. Biophys. Methods 2004, 61:283–298. 125. Chou SF, Hsu WL, Hwang JM, and Chen CY: Development of an immunosensor for human ferritin, a nonspeciﬁc tumor marker, based on surface plasmon resonance, Biosens. Bioelectron. 2004, 19:999–1005. 126. Miyashita M, Shimada T, Miyagawa H, and Akamatsu M: Surface plasmon resonance-based immunoassay for 17beta-estradiol and its application to the measurement of estrogen receptor-binding activity, Anal. Bioanal. Chem. 2005, 381:667–673. 127. Winter G, Grifﬁths AD, Hawkins RE, and Hoogenboom HR: Making antibodies by phage display technology, Annu. Rev. Immunol. 1994, 12:433–455. 128. Enever C, Tomlinson IM, Lund J, Levens M, and Holliger P: Engineering high afﬁnity superantigens by phage display, J. Mol. Biol. 2005, 347:107–120. 129. Xu L, Aha P, Gu K, Kuimelis RG, Kurz M, Lam T, Lim AC, Liu H, Lohse PA, Sun L, et al.: Directed evolution of high-afﬁnity antibody mimics using mRNA display, Chem. Biol. 2002, 9:933–942. 130. Hanes J, Schafﬁtzel C, Knappik A, and Pluckthun A: Picomolar afﬁnity antibodies from a fully synthetic naive library selected and evolved by ribosome display, Nat. Biotechnol. 2000, 18:1287–1292. 131. Feldhaus MJ and Siegel RW: Yeast display of antibody fragments: A discovery and characterization platform, J. Immunol. Methods 2004, 290:69–80. 132. Colby DW, Kellogg BA, Graff CP, Yeung YA, Swers JS, and Wittrup KD: Engineering antibody afﬁnity by yeast surface display, Methods Enzymol. 2004, 388:348–358. 133. Yu H and Bruno JG: Immunomagnetic-electrochemiluminescent detection of Escherichia coli O157 and Salmonella typhimurium in foods and environmental water samples, Appl. Environ. Microbiol. 1996, 62:587–592. 134. Schier R, McCall A, Adams GP, Marshall KW, Merritt H, Yim M, Crawford RS, Weiner LM, Marks C, and Marks JD: Isolation of picomolar afﬁnity anti-c-erbB-2 single-chain Fv by molecular evolution of the complementarity determining regions in the center of the antibody binding site, J. Molecular Biol. 1996, 263:551–567. 135. Piervincenzi RT, Reichert WM, and Hellinga HW: Genetic engineering of a single-chain antibody fragment for surface immobilization in an optical biosensor, Biosens. Bioelectron. 1998, 13:305–312. 136. Schwendener RA, Trub T, Schott H, Langhals H, Barth RF, Groscurth P, and Hengartner H: Comparative-studies of the preparation of immunoliposomes with the use of 2 bifunctional coupling agents and investigation of in vitro immunoliposome-target cell binding by cytoﬂuorometry and electronmicroscopy, Biochim. Biophys. Acta 1990, 1026:69–79. 137. Loughrey HC, Choi LS, Cullis PR, and Bally MB: Optimized procedures for the coupling of proteins to liposomes, J. Immunol. Methods 1990, 132:25–35. 138. Frost SJ, Firth GB, and Chakraborty J: Antibody-coated liposomes as a particulate solid-phase for immunoassays—Measurement of urinary micro-albumin, J. Immunol. Methods 1990, 134:207–213. 139. Shahinian S and Silvius JR: A novel strategy affords high-yield coupling of antibody Fab’ fragments to liposomes, Biochim. Biophys. Acta 1995, 1239:157–167. 140. Loughrey H, Bally MB, and Cullis PR: A noncovalent method of attaching antibodies to liposomes, Biochim. Biophys. Acta 1987, 901:157–160. 141. Lu B, Smyth MR, and Okennedy R: Oriented immobilization of antibodies and its applications in immunoassays and immunosensors, Analyst 1996, 121:R29–R32.

Otles/Handbook of Food Analysis Instruments 45660_C020 Final Proof page 460 16.7.2008 11:09pm Compositor Name: MSubramanian

460

Handbook of Food Analysis Instruments

142. Peluso P, Wilson DS, Do D, Tran H, Venkatasubbaiah M, Quincy D, Heidecker B, Poindexter K, Tolani N, Phelan M, et al.: Optimizing antibody immobilization strategies for the construction of protein microarrays, Anal. Biochem. 2003, 312:113–124. 143. Kobatake E, Iwai T, Ikariyama Y, and Aizawa M: Bioluminescent immunoassay with a protein A-luciferase fusion protein, Anal. Biochem. 1993, 208:300–305. 144. Bjorck L and Kronvall G: Puriﬁcation and some properties of streptococcal protein-G, protein-A novel Igg-binding reagent, J. Immunol. 1984, 133:969–974. 145. Akerstrom B, Brodin T, Reis K, and Bjorck L: Protein G: A powerful tool for binding and detection of monoclonal and polyclonal antibodies, J. Immunol. 1985, 135:2589–2592. 146. Akerstrom B and Bjorck L: A physicochemical study of protein G, a molecule with unique immunoglobulin G-binding properties, J. Biol. Chem. 1986, 261:10240–10247. 147. Markela E, Stahlberg TH, and Hemmila I: Europium-labelled recombinant protein G. A fast and sensitive universal immunoreagent for time-resolved immunoﬂuorometry, J. Immunol. Methods 1993, 161:1–6. 148. Chen C-S, Baeumner AJ, and Durst RA: Protein G-liposomal nanovesicles as universal reagents for immunoassays, Talanta 2005, 67:205–211. 149. Park S and Durst RA: Immunoliposome sandwich assay for the detection of Escherichia coli O157:H7, Anal. Biochem. 2000, 280:151–158. 150. Rongen HAH, Vanderhorst HM, Hugenholtz GWK, Bult A, Vanbennekom WP, and Vandermeide PH: Development of a liposome immunosorbent-assay for human interferon-gamma, Anal. Chim. Acta 1994, 287:191–199. 151. Ho JA and Durst RA: Detection of fumonisin B1: Comparison of ﬂow-injection liposome immunoanalysis with high-performance liquid chromatography, Anal. Biochem. 2003, 312:7–13. 152. Locascio-Brown L, Plant AL, Horvath V, and Durst RA: Liposome ﬂow injection immunoassay: Implications for sensitivity, dynamic range, and antibody regeneration, Anal. Chem. 1990, 62:2587–2593.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 461 16.7.2008 11:12pm Compositor Name: MSubramanian

Instruments 21 Rheological in Food Analysis Nesli Sozer and Jozef L. Kokini CONTENTS 21.1 21.2

Introduction ........................................................................................................................ 461 Types of Food Material Behavior ..................................................................................... 462 21.2.1 Solid Behavior ..................................................................................................... 462 21.2.1.1 Tests for Solid Materials .................................................................... 463 21.2.2 Tests for Viscous Materials ................................................................................. 466 21.2.2.1 Measurement Techniques for Fluid Behavior .................................... 468 21.2.2.2 Modeling of Fluid Behavior Data ...................................................... 469 21.2.3 Tests for Transient Viscoelasticity ...................................................................... 473 21.2.4 Tests for Viscoelasticity ...................................................................................... 478 21.2.4.1 Linear Viscoelasticity ......................................................................... 479 21.2.4.2 Nonlinear Viscoelasticity ................................................................... 485 21.3 Concluding Remarks ......................................................................................................... 490 References ..................................................................................................................................... 491

21.1 INTRODUCTION Rheology, by name, originated from a combination of two Greek words ‘‘rheo’’ meaning ﬂow and ‘‘logy’’ meaning science. So basically, rheology can be deﬁned as the science of deformation and ﬂow of matters. Food materials can be categorized in three main groups: ﬂuid, visco-elastic, and solid materials. They exhibit ﬂow and deformation under external forces. If the material is an elastic solid, the applied force will not result in a ﬂow but in a deformation; however, if it is a viscous material it will start to ﬂow with the application of a force. To analyze the response of food materials to load deformation, controlled experiments should be carried out. Food materials can be deformed in shear, extension, and bulk compression (Figure 21.1). The tests can be done under steady state or dynamic conditions and ﬁnally the corresponding data can be analyzed by resulting moduli, compliances, and use of various models. The most important and basic concepts of rheology are stress, strain, and strain rate. The rheological response of any material is physically expressed by stress, which is measured by the force concentration on material [1]. Rheological principles and theory can be used as an aid in quality control and design, calculating energy usage, process control, and selection of the most appropriate equipment during production. Rheological parameters such as maximum stress, maximum strain, elastic modulus, compliance, and measure of stiffness contain useful information for the textural characteristics of solid food. Testing is conducted by taking a sample size of food and deforming it in a controlled way, generally with a motor-driven machine such as the texture analyzer or rheometer, and the force is measured as well as the distance moved or displacement of the object within testing time. The force=strain is then usually plotted against the displacement to give a force displacement curve.

461

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 462 16.7.2008 11:12pm Compositor Name: MSubramanian

462

Handbook of Food Analysis Instruments Force

Force

Compression

Shear

Bulk compression

FIGURE 21.1

Types of deformation.

The invention of several types of rheometers and texture analyzers enabled the rheologists to ﬁnd out rheological properties that are independent of size, shape, and measurement techniques. The objective of this chapter is to review measurement techniques and modeling of rheological data of food materials.

21.2 TYPES OF FOOD MATERIAL BEHAVIOR 21.2.1 SOLID BEHAVIOR Some food materials like dried pasta, biscuits, and chips show elastic behavior. A perfectly elastic material is called Hookean solid, which obeys Hooke’s law and is one of the basic equations of (Equation 21.1) rheology. s ¼ Gg_

(21:1)

where s is the shear stress g_ is the shear strain G refers to shear modulus This type of material does not show ﬂow behavior, but it is linearly elastic. Not all solid materials are perfectly elastic, they can also be elastoplastic (e.g., butter, margarine, jellies, etc.) or nonlinear elastic. The main difference of elastoplastic materials rises from the fact that they need a yield stress for permanent deformation.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 463 16.7.2008 11:12pm Compositor Name: MSubramanian

463

Rheological Instruments in Food Analysis

21.2.1.1

Tests for Solid Materials

21.2.1.1.1 Puncture Test The main principle of a puncture test is to measure force while pushing a probe inside the food material. It is a destructive test and one of the important points that should be taken into account is to have a constant penetration depth for all samples. A puncture test is set on ‘‘semi-inﬁnite geometry’’ theory, where the sample size is so much larger than the probe that there is no interfering effect of the geometry of the sample on to the puncture force. Another point that should be taken into account while performing a puncture test is the base support. If a solid base support is going to be used, then the sample must not be so ﬂat that the deformation force turns out to be compression rather than puncture. A support with a hole can also be used. The diameter of this hole should be 1.5–3 times the diameter of the probe so that the test is a puncture test but not a die or bending test [2]. Tsukakoshi et al. [3] carried out a puncture test to evaluate the stress changes during cereal snack puncture. They carried out the measurements with two different equipments: tensipresser, model TTP 50 BXII (Taketomo Electric Inc., Tokyo, Japan) and TA.XT2 texture analyzer (Stable Micro Systems, Godalming, United Kingdom). They made the measurements with the same pin plunger at several speeds 0.1, 0.5, 2, and 8 mm=s. They found that with increase in deformation rate during the puncture test, the number of fractures increased (Figure 21.2). They also concluded that stress reduction did not depend on the test machine. But they also found that large number of small fractures occurred in crackers more in the tensipresser than in the texture analyser with more noisy results. The puncture test is a simple and easy way to test puncture force that gives an idea about the hardness=stiffness of the solid samples. It is also a good and simple way to test food materials with heterogeneous structure (e.g., cereal bars, chocolates with various layers, etc.). 21.2.1.1.2 Bending Test Applying bending test, the material is subjected to both tensile, shear and compression forces. Bending test is mainly applied to solid food materials with a more homogenous shape like cereal=chocolate bars, dry spaghetti, or after cutting the samples to a constant size and shape [4]. The important outcomes of bending test are deﬂection point, Young’s modulus, and ﬂexure stress. Deﬂection point is obtained while subjecting the material to a force after placing it between two supports (Figure 21.3). The Young’s modulus or elastic modulus (E) is the ratio of stress (s) to strain («) where strain is the ratio of change in length or radius to original length (Figure 21.4) [5,6].

102 0.5 mm/s Cumulative number

2 mm/s 101 8 mm/s 100 0.1 mm/s 10–1

10–2 10–3

10–2 10–1 100 Size of force change (N)

101

FIGURE 21.2 Frequency size distributions of crackers during puncture test with speeds of 0.1, 0.5, 2, and 8 mm=s. (Needs copyright from Tsukakoshi, Y., Naito, S., and Ishida, N., J. Texture Stud., 38, 220, 2007.)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 464 16.7.2008 11:12pm Compositor Name: MSubramanian

464

Handbook of Food Analysis Instruments X

d Force

FIGURE 21.3

Three-point bending test.

s «

(21:2)

E¼

Fx3 4adh3

(21:3)

E¼

4Fx3 3dPR4

(21:4)

E¼

The elastic modulus for a rectangular and cylindrical material can be found from Equations 21.3 and 21.4, respectively. Chen et al. [7] carried out acoustic measurements after performing a three-point bending test on various kinds of biscuits with TA.XT plus texture analyzer (Stable Micro Systems, Godalming, United Kingdom) attached to the Bruel & Kjaer force ﬁeld microphone (Bruel & Kjaer, Naerum, Denmark). The biscuits were placed on two supports that were 60 mm apart from each other. In Figure 21.5, the continuous curve represents the force and the bars fracture events. In the ﬁrst regime (a), the probe contacted with biscuits without major deformation, which could be used for elastic modulus evaluation. As the probe went further down they found that the force curve became jagged and many acoustic events took place. 21.2.1.1.3 Tensile and Extensibility Tests Tensile tests are used to determine the mechanical properties such as yield strength, tensile strength, strains at peak and break, Young’s modulus, and total energy to break [8]. During tensile test, the sample fractures instantaneously which is perpendicular to the plane of the applied tension where the maximum force is the tensile strength of the material. Several cracks may occur and spread in any direction [2]. R

h

h

a

FIGURE 21.4 Dimensions of a rectangular and a cylindrical food material. (Requires copyright from Steffe, J.F., Rheological Methods in Food Process Engineering, Freeman Press, MI, 1996.)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 465 16.7.2008 11:12pm Compositor Name: MSubramanian

465

Rheological Instruments in Food Analysis 115

10 (f) 127.904–130.800 (e) 125.200–127.902

9 8

(d) 116.700–119.500 97

7

(g) 137.200–141.900

(c) 106.600–110.598 (b) 93.600–103.998 (a) 74.700–79.598

87

6 5

78

Force (N)

Aux1: Reference AED (dB [SPL])

106

4 3

69

2 59 1 50 25

50

75

100

125

150

175

200

0 225

Time (s)

FIGURE 21.5 Graph showing the force and acoustic level during the breaking of biscuits with test speed of 0.01 mm=s (AED—acoustic envelope detector, SPL—sound pressure level). (Requires permission from Chen, J., Karlsson, C., and Povey, M., J. Texture Stud., 36, 139, 2005.)

Farouk et al. [9] performed tensile and extensibility tests on cooked meat spaghetti strands (approximately, 60–80 cm long and 10 mm wide; Figure 21.6). To assess the mechanical performance of dry spaghetti, tensile testing was also performed by Guinea et al. [10]. The measurement was performed by using a universal testing machine (Instron

FIGURE 21.6 Meat spaghetti sample mounted on the texture analyzer showing a nicked section. (Requires permission from Farouk, M.M., Zhang, S.X., and Waller, J., J. Food Qual., 28, 452, 2005.)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 466 16.7.2008 11:12pm Compositor Name: MSubramanian

466

Handbook of Food Analysis Instruments

TABLE 21.1 Break Load and Break Strain of Laboratory-Produced Fresh Egg Pasta Analogues, Both Before and After Cooking (F100: Wheat Pasta, B40-F60: Mixture of Buckwheat and Wheat with a Ratio of 1:1.5, B60-F40: Mixture of Buckwheat and Wheat with a Ratio of 1.5:1) Raw Pasta

Cooked Pasta

Sample

Break Load (N)

Break Strain (%)

Break Load (N)

Break Strain (%)

F100 B40-F60 B60-F40

1.08 0.06 0.48 0.06a 0.61 0.03a

67.83 6.61 35.85 8.02b 20.31 1.70a

3.24 0.02 1.75 0.24b 1.44 0.22a

84.71 8.48 41.66 5.57b 29.10 3.42b

Source: Requires permission from Alamprese, C., Casiraghi, E., and Pagani, M.A., Eur. Food Res. Technol., 225, 205, 2007. a b

P < 0.01, with reference to F100. P < 0.05.

4111) driven at 1 mm=min elongation rate. As a result they claimed that dry spaghetti ﬁbers mainly exhibit a linear-elastic behavior. Alamprese et al. [11] evaluated the response of both raw and cooked pasta using an Instron universal testing machine 4301 (Instron Ltd., High Wycombe, United Kingdom) supported by Series IX automated material testing software (Instron Co., 1998) with a cross-head speed of 20 mm=min. They calculated the Young’s modulus which is an index of pasta hardness as well as the break load indicating pasta toughness (Table 21.1). They found that as the amount of buckwheat increased, there was a decrease in sample break strain that showed loss of elastic properties.

21.2.2 TESTS

FOR

VISCOUS MATERIALS

Viscous materials can be deformed by shearing. Viscous behavior of a food material shows its ability to resist ﬂowing. If a ﬂuid ﬂows between two parallel plates by applying shearing force (F) per unit area, the relative motion of the upper plate to lower plate causes the material to move with a velocity n (Figure 21.7). If the material is a Newtonian ﬂuid then the shearing stress (F=A) is directly proportional to the mean rate of shear y

F

V

x ∆x

FIGURE 21.7

Simple shear ﬂow.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 467 16.7.2008 11:12pm Compositor Name: MSubramanian

467

Rheological Instruments in Food Analysis

Pseudoplastic with yield point

Shear stress

Bingham Dilatant with yield point

Pseudoplastic Newtonian

Dilatant

Shear rate

FIGURE 21.8

Time-independent ﬂow behavior.

s g_

(21:5)

s ¼ hg_

(21:6)

h¼ or

where h refers to Newtonian viscosity or dynamic viscosity g_ is the shear rate s is the shear stress For non-Newtonian ﬂuids, the relationship between shear stress and shear rate is nonlinear and the viscosity is a function of shear rate. The shear rate dependent viscosity is called apparent viscosity (ha). Figure 21.8 shows various types of ﬂuids that exhibit this type of behavior. Timeindependent non-Newtonian ﬂuids can be summarized as Bingham (plastic), pseudoplastic (shear thinning), dilatant ﬂuid (shear thickening), pseudoplastic and dilatant ﬂuids with a yield stress. Food materials like emulsions, gums, proteins, carbohydrates, and chocolate syrups are either pseudoplastic (shear thinning) or dilatant (shear thickening). Bingham ﬂuids require an initial stress which is called ‘‘yield stress’’ to initiate the ﬂow, and the shear stress versus shear rate plot is linear to it. This type of behavior is quite common in foods like mayonnaise, ketchup, and margarine. For pseudoplastic and dilatant ﬂuids with a yield stress, these types of ﬂuid behavior still require a yield stress, but above the yield stress the relationship between shear stress and shear rate is nonlinear. Time-dependent non-Newtonian ﬂuids can be grouped in four: thixotropic, rheopectic, shear thinning, and shear thickening (Figure 21.9). In thixotropic ﬂuids (some starch-paste gels), the apparent viscosity decreases with time of shearing that is reversible. In rheopectic ﬂuids, the apparent viscosity increases with time of shearing where the change is reversible. However, it is difﬁcult to ﬁnd food materials that exhibit this type of behavior. Some gum solutions and starch pastes are shear-thinning ﬂuids where the apparent viscosity decreases with time with an irreversible change. The opposite of this behavior is shear thickening that can be seen in egg white or heavy cream where during shearing, viscosity increases irreversibly with time.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 468 16.7.2008 11:12pm Compositor Name: MSubramanian

468

Handbook of Food Analysis Instruments

Thixotropic

Shear rate (1/s)

Viscosity (Pa · s)

Rheopectic

Newtonian Slightly thixotropic

Newtonian

Strongly thixotropic (a)

Time (s)

FIGURE 21.9 stress.

21.2.2.1

Rheopectic (b)

Time (s)

Time dependency factors in ﬂuid ﬂow: (a) at constant rate of shear and (b) at constant shear

Measurement Techniques for Fluid Behavior

Generally, rheometer, mixer, or relevant instruments like capillary viscometer and pipe setups can be used to analyze ﬂuid behavior. A capillary viscometer consists of a very small diameter and a cylindrical capillary tube where liquid is forced through the capillary by imposing a pressure drop. The ratio of the very small diameter of the tube and the very large length to diameter minimizes entrance and exit effects and ensures a fully developed velocity proﬁle. Dervisoglu and Kokini [12] developed a capillary rheometer where they used long-entrance region and determined the pressure drop as the difference of two pressure values measured in the fully developed laminar region (Figure 21.10). Bagley’s procedure [13] was used to correct the data for entrance effects. In this method, true shear stress equals to

Air pressure regulating valve Pressure vessel

Jacket

Constant temperature of water circulator

Le

DC power supply

Dual pen Tele recorder thermometer

L Pressure transducers

Temporary sensing probe

FIGURE 21.10 Schematic diagram of the capillary setup. (Requires permission from Dervisoglu, M. and Kokini, J.L., J. Food Sci., 51, 541, 1986.)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 469 16.7.2008 11:12pm Compositor Name: MSubramanian

469

Rheological Instruments in Food Analysis

t¼

DPR 2(L þ eR)

(21:7)

where DP is the total pressure drop L is the length of the capillary R is the radius of the capillary e is the Bagley end correction factor This equation can be rearranged to DP ¼ 2t

L þ 2te R

(21:8)

After plotting DP versus L=R, the true shear stress is estimated from the slope of the line and e is estimated through the value of L=R where DP is zero. For the rheometer which is a rotational type instrument, there are various geometries used for shear stress and shear rate measurements, which are listed in Figure 21.11. 21.2.2.2 21.2.2.2.1

Modeling of Fluid Behavior Data

21.2.2.2.1.1

Time-Independent Fluid Behavior

Power Law (Ostwald–de Wael Model) Equation

In rheology, there are several empirical equations that have no theoretical background, but facilitate to handle rheological data. Power law equation is one of the main equations, which is deﬁned as follows:

g· = a

s=

w a

g· = Shear rate

3T 2pR 3

T: Torque w: Frequency a: Angle

w

g· = r

wR h 3+

3 d In s d In(wR/h)

s=

w L v

sw : Shear stress at the wall v: Average velocity

h

P1

R: Plate radius

P: Pressure

3T 2ΠR 3

g· = D sw =

8v 3 1d In(8v/D) + D 4 4 d In sw

(P1 – P2)D 4L

FIGURE 21.11 Some commonly used geometries for measurements of shear stress and shear rate with appropriate formulations.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 470 16.7.2008 11:12pm Compositor Name: MSubramanian

470

Handbook of Food Analysis Instruments

TABLE 21.2 Power Equation Parameters for Steady Viscosities of Some Foods Food

Flow Behavior Index (n)

Consistency Index (K) (Pa sn)

0.074 0.042 0.378 0.061 0.273 0.107 0.0043 0.174 0.501 0.168

333 417 15.1 776 550 79.4 549 7.6 670 316

Butter, stick, unsalted, Land O’Lakes Butter, whipped, unsalted, Land O’Lakes Cool whip, Birdseye Cream cheese, whipped, Temptee Frosting, canned, Betty Crocker Ketchup, tomato, Heinz Margarine, stick, Parkay Margarine, squeeze, Parkay Marshmallow ﬂuff, Durkee-Mower Peanut butter, creamy, Skippy

Source: Requires permission from Bistany, K.L. and Kokini, J.L., J. Rheology, 27, 608, 1983. Copyright by Society of Rheology.

_ n s ¼ K(g)

(21:9)

or ha ¼

s _ n1 ¼ K(g) g_

(21:10)

where s is the shear stress K is the consistency index (Pa sn) g_ is the shear rate n is the ﬂow index showing how close the type of ﬂow is to Newtonian behavior ha is the apparent viscosity (Pa s) If n > 1, the plot for shear stress versus shear rate is an upward curve (shear-thickening ﬂuid); when 0 < n < 1, the plot for the shear stress versus shear rate is a downward curve (shear-thinning ﬂuid); and when n ¼ 1, there is a linear relationship between shear stress and shear rate (Newtonian ﬂuid). Table 21.2 lists experimentally determined power equation constants for various food samples. 21.2.2.2.1.2

Herschel–Bulkley Model

Non-Newtonian ﬂuids with a yield stress are mostly analyzed by the Herschel–Bulkley model, which is as follows: s ¼ s0 þ K g_ n

(21:11)

or h¼

s s0 ¼ K g_ n1 g

(21:12)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 471 16.7.2008 11:12pm Compositor Name: MSubramanian

471

Rheological Instruments in Food Analysis

TABLE 21.3 Classiﬁcation of Newtonian and Non-Newtonian Fluids Fluid Type

t0

Newtonian Non-Newtonian Shear thinning (pseudoplastic) Shear thickening (dilatant) Bingham Pseudoplastic with yield stress Dilatant with yield stress

0 0 0 >0 >0 >0

n

K

1

K¼h

01 1 01

>0 >0 >0 >0 >0

Source: Requires permission from Barbosa-Canovas, G.V., Kokini, J.L., Ma, L., and Ibarz, A., Adv. Food Nutr. Res., 39, 1, 1996.

where s is the shear stress g_ is the shear rate n is the ﬂow index s0 is the yield stress K is the consistency index ha is the apparent viscosity Table 21.3 summarizes the classiﬁcation and general ﬂow types of ﬂuid foods based on the magnitude of n and s0 [14]. 21.2.2.2.1.3

Casson Model

This model was found to be effective for some foods like chocolate and some other ﬁlled ﬂuids [2]. The equation for this model is pﬃﬃﬃ pﬃﬃﬃﬃ pﬃﬃﬃﬃﬃ s ¼ s0 þ ha g_

(21:13)

where s is the shear stress s0 is the yield stress ha is the apparent viscosity g_ is the shear rate The Casson model has been recommended as a standard method by the International Ofﬁce of Cocoa, Chocolate and Sugar Confectionery to evaluate the rheological behavior of chocolate ﬂuid. Karnjanolarn and McCarthy [15] made a research on different formulations of milk chocolate by the use of a concentric-cylinder rotational viscometer (Bohlin CVO, Malvern Instruments, East Brunswick, NJ). The measurements were carried out in a ramp up=ramp down mode over a shear-rate range of 2–50=s and the data were evaluated with the Casson equation (Table 21.4). 21.2.2.2.2 Time-Dependent Fluid Behavior Several foods exhibit time-dependent ﬂow changes that are summarized as mixtures of xanthan and carob gums [16], mango pulp [17], hot pepper-soybean paste [18], food suspensions [19], and ice cream [20]. Time-dependent ﬂow properties establish relationship between structure and ﬂow. Time-dependent behavior arises from viscoelasticity, thixotropy, or their combination. The typical property of a viscoelastic material to shearing is retarded where it is instantaneous for thixotropic

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 472 16.7.2008 11:12pm Compositor Name: MSubramanian

472

Handbook of Food Analysis Instruments

TABLE 21.4 Casson Yield Stress and Viscosity of Chocolate at 408C for Coarse and Fine Grinds Coarse Grind Emulsiﬁer Content No emulsiﬁer Soy lecithin

PGPR

% w=w

Casson Yield (Pa)

Casson Viscosity (Pa s)

0.0 0.1 0.2 0.3 0.4 0.5 0.1 0.2 0.3 0.4 0.5

21.8 5.93 3.99 3.86 4.19 4.73 7.52 1.81 0.41 0.15 0.00

3.73 2.65 2.06 1.81 1.63 1.49 4.32 4.28 4.27 4.29 4.16

Fine Grind R2

Casson Yield (Pa)

Casson Viscosity (Pa s)

R2

0.992 0.999 0.999 0.998 0.999 0.998 0.999 0.998 0.998 0.998 0.998

34.8 22.3 13.3 10.5 10.4 11.2 11.3 1.15 0.0 0.0 0.0

11.0 4.95 3.33 2.79 2.26 1.97 8.48 8.55 7.74 6.37 5.63

0.995 >0.999 0.999 0.999 0.999 0.999 0.999 0.999 0.998 0.997 0.995

Source: Requires permission from Karnjanolarn, R. and McCarthy, K.L., J. Texture Stud., 37, 674, 2006. Copyright by Blackwell publishing. Note: PGPR—polyglycerolpolyricinoleate.

materials. Time-dependent behavior breaks down the structure. Structural changes can be summarized as follows: disentanglement of polymer molecules in solution, deﬂocculation of globules in emulsion, favorable spatial distribution of particles, and molecular association of biopolymers [21]. Time-dependent behavior of food materials is analyzed by measuring stress of viscosity decay at a constant shear rate. Various mathematical models have been designed to study the time-dependent ﬂow. The earliest model of time-dependent behavior was the Weltman model (Equation 21.14) where a logarithmic decay of shear stress was assumed in the absence of equilibrium conditions [22]. Hahn model (Equation 21.15) is an extension of Weltman model [22] that includes an equilibrium shear-stress term [23]. The constants A and B represent the stress at the beginning of shearing and P represents the rate of structural breakdown [7,8]. Later on Figoni and Shoemaker [24] described the stress decay process as a ﬁrst-order kinetic model with a nonzero equilibrium value. Weltman model: s ¼ A B ln t

(21:14)

logðs se Þ ¼ P at

(21:15)

s ¼ se þ ðsmax se Þekt

(21:16)

Hahn model:

Figoni and Shoemaker model:

where s is the shear stress (Pa) at any time of shearing (t) se is the equilibrium shear stress (Pa) smax is the maximum shear stress (Pa) A, B, P, a, and k are constants

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 473 16.7.2008 11:12pm Compositor Name: MSubramanian

473

Rheological Instruments in Food Analysis 21.2.2.2.2.1

Effects of Temperature and Concentration on Viscous Flow

It is a well known fact that viscosity is strongly affected by temperature. Taking into account that the temperature changes during processing and storage, evaluation of rheological data with temperature change is important. The effect of temperature on viscosity is described by the Arrhenius equation: h ¼ h1 exp (Ea =RT)

(21:17)

where h is the apparent viscosity at a speciﬁc shear rate h1 is the frequency factor Ea is the amount of energy that should be passed before ﬂow starts The temperature and concentration dependence of the consistency coefﬁcient can be ﬁtted by the following equation: ln K ¼ aT þ b ln C

(21:18)

where a and b are constants C is the concentration (%) T is the temperature (8C) Kaya and Sozer [25] determined the ﬂow behavior, temperature, and concentration dependence of the consistency coefﬁcient in pomegranate juice concentrates with various soluble-solid contents. They found that the exponential model describes the dependency of Ea on soluble-solid contents better than the power law model. To modify processing machinery used for pomegranate juice, a simple equation (Equation 21.19) was obtained which described the combined effect of temperature (T) and concentration (C) on the material’s viscosity. h ¼ 1:25 10

21.2.3 TESTS

FOR

11

4398 exp 0:148C þ T

(21:19)

TRANSIENT VISCOELASTICITY

One of the characteristics of viscoelastic foods is that when a shear stress is instantaneously applied on them, the shear stress overshoots at inception of steady-shear ﬂow. These overshoots can range anywhere from 30% to 300% of their steady-state value, depending on the particular shear rate and material used. These stresses are important in terms of startup of processes and properties like spreadability [26]. Kokini and Chou [27] measured the viscous properties of tomato pectins by preparing aqueous pectin solutions of 0.25, 0.4, 0.5, 0.75, 1, and 1.5 kg=m3 (Figure 21.12). First (c1) and second normal-stress coefﬁcients (c2) can be found from the equations below that _ combine speciﬁc stress components sij to the shear rate, g: _ 2 s11 s22 ¼ c1 (g)

(21:20)

_ 2 s22 s33 ¼ c2 (g)

(21:21)

Kokini and Dickie [28] carried out some measurements on several food materials (e.g., margarine, butter, mayonnaise, ketchup, mustard, etc.) using the cone and plate geometry. Typical magnitudes of c of foods are shown in Figure 21.13. In their work they also presented normal-stress growth

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 474 16.7.2008 11:12pm Compositor Name: MSubramanian

474

Handbook of Food Analysis Instruments

103

Specific viscosity

102

101

HBP1 HBP2 HBP3 HBP4 HBP5

100

CLitr NSS NS NS2

10–1

NL NLG

10–2

10–2

10–1

100 C (h )

101

102

FIGURE 21.12 Speciﬁc viscosity versus c(h) for tomato, citrus, and apple pectins in pH 4.6 citrate phosphate buffer (HBP1–HBP5 refer to dextrose equivalent that varies between 70.3 and 43.3). (Requires permission from Kokini, J.L. and Chou, T.C., J. Texture Stud., 24, 117, 1993, Figure 7.)

development of ketchup at 258C (Figure 21.14). The authors found that steady-state values were reached fastest with high shear rates. Dickie and Kokini [29] modeled various types of food materials by using Leider and Bird equation [30] h t i _ n 1 þ (bgt _ 1) exp t uf ¼ m(g) anl 0 1=(n0 n) m l¼ 2m where tuf is the shear stress m and n are limiting viscous power law parameters g_ is the shear rate t is the time a and b are adjustable parameters l is the time constant m0 and n0 are the ﬁrst normal stress power law parameters

(21:22) (21:23)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 475 16.7.2008 11:12pm Compositor Name: MSubramanian

475

Rheological Instruments in Food Analysis

106

Margarine Mayonnaise Apple butter

105

106

Butter Canned frosting

105

Mustard

y (Pa · s2)

Ketchup 104

104

103

103

102

102

101

101

100

100

10–1

10–1

10–2 .1

1

10

10–2 100 .1 Shear rate

1

10

100

(s–1)

FIGURE 21.13 Primary normal-stress coefﬁcient versus shear rate. (Requires permission from Kokini, J.L. and Dickie, A., J. Texture Stud., 25, 539, 1981, Figure 3.)

1.4 100 s1 10 s

⫺1 ⫺1

1.2

1s

t21 t21

∞

1.0 .1 s

⫺1

0.8

Experimental

0.6

Predicted 0.0 0

10

20

30

Time (s)

FIGURE 21.14 Shear stress development for ketchup at 258C. (Requires permission from Kokini, J.L. and Dickie, A., J. Texture Stud., 25, 539, 1981, Figure 4.)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 476 16.7.2008 11:12pm Compositor Name: MSubramanian

476

Handbook of Food Analysis Instruments

If this model is rearranged by using power law equation, then the following equations are obtained: _ n t 12 ¼ m(g)

(21:24)

_ n t12 t 22 ¼ m0 (g)

0

(21:25)

where t12 is shear stress t12 – t22 is the ﬁrst normal-stress difference m, n, m0 , and n0 are the power law parameters g_ is the shear rate Table 21.5 shows rheological parameters of various foods that are deﬁned by normal stress [31]. Since Leider and Bird equation [30] has only one exponential decay term, it appears to be inadequate for viscoelastic ﬂuids having a wide spectrum of relaxation times. Campanella and Peleg [32] presented stress growth and decay data on mayonnaise, modeled the data with modiﬁed Larson’s model [33] and tested the applicability of this model on rheological behavior of mayonnaise. 0 s(t) G(m) sin mp @t ¼ s1 _ m1 p(ng)

1m ngt _

e m

ðt

1

_ þ sm engs dsA

(21:26)

0

s ¼ t t0

TABLE 21.5 Rheological Parameters of Foods

Apple butter Canned frosting Honey Ketchup Marshmallow cream Mayonnaise Mustard Peanut butter Stick butter Stick margarine Squeeze margarine Tub margarine Whipped butter Whipped cream cheese Whipped dessert topping

m(Pa sn)

n

R2

m0 (Pa sn 0 )

n0

R2

l(s)

222.90 355.84 15.39 29.10 563.10 100.13 35.05 501.13 199.28 297.58 8.68 106.68 312.30 422.30 35.98

0.145 0.117 1.989 0.136 0.379 0.131 0.196 0.065 0.085 0.074 0.124 0.077 0.057 0.058 0.120

0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99 0.99

156.03 816.11 — 39.47 184.45 256.40 65.69 3785.00 3403.00 3010.13 15.70 177.20 110.76 363.70 138.00

0.566 0.244 — 0.258 0.127 0.048 0.136 0.175 0.398 0.299 0.168 0.353 0.476 0.418 0.309

0.99 0.99 — 0.99 0.99 0.99 0.99 0.99 0.98 0.99 0.99 0.99 0.99 0.99 0.99

8.21 102 2.90 100 — 4.70 102 1.27 103 2.51 101 2.90 100 1.86 105 1.06 103 1.34 103 9.93 102 5.16 101 1.61 102 8.60 102 3.09 101

Source: Requires permission from Heldman, D.R. and Lund, D.B., in Handbook of Food Engineering, Marcel Dekker, NY, 1992.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 477 16.7.2008 11:12pm Compositor Name: MSubramanian

477

Rheological Instruments in Food Analysis

where s1 is the steady state stress G is the gamma function m and n are constants g_ is the shear rate and the deformation history of the experiment was g¼

1 < t0 < 0 0 < t0 < t

_ gt _ t0 ) g(t

The constant m has a value 0 < m < 1 and it is representative of the relaxation modulus (G) when it is expressed as follows: G ðt t 0 Þ ¼ C ðt t 0 Þ

m

(21:27)

The constant n is obtained from time where maximum shear stress is obtained (tmax): n¼

1 gtmax

(21:28)

The authors demonstrated that mayonnaise exhibits stress overshoot in the order of 70% above the equilibrium level. The experimental data are ﬁtted to Equation 21.26 using the values m ¼ 0.3 and n ¼ 0.07, which is shown in Figure 21.15.

2.0

g·0 = 1.8 s–1 m = 0.3

R (t)

1.5

1.0

Experimental data 0.5

Model (Eq. 20) n = 0.07 Model (Eq. 20) n = 0.115

0 0

40

20

60

Time (s)

FIGURE 21.15 Shear stress overshoot data of commercial mayonnaise at a shear rate of 1.8=s and the ﬁt of Equation 21.26 with n ¼ 0.07 and n ¼ 0.074. (Requires permission from Campanella, O.H. and Peleg, M., J. Rheol., 31, 439, 1987. Copyright by the Society of Rheology.)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 478 16.7.2008 11:12pm Compositor Name: MSubramanian

478

Handbook of Food Analysis Instruments

The main difﬁculty with the new model is that its mathematical derivation is based on the assumption that the damping constant n is shear rate independent, while in reality it is a weak function of the latter especially at low rates. Campanella and Peleg [32] recommended to overcome this problem by using a representative value of n or by incorporating more exact account of the relaxation time spectrum and the rate dependency of n into the model’s basic equation format.

21.2.4 TESTS

FOR

VISCOELASTICITY

Viscoelasticity is a term that deals with solids and liquids and their combined effect. Generally, food materials are viscoelastic in nature and time dependent. Viscoelastic materials are assumed to have memory since the internal stress developed is a function of both deformation and the previous history of deformation. Food polymers, either in solutions, dispersions, melts, or cross linked, generally exhibit both viscous and elastic properties. The ﬂexibility arising from the long chains is responsible for elastic properties [34]. These properties are analyzed by dynamic testing, stress relaxation, creep recovery, shear-stress growth, and normal-stress growth experiments. In process engineering, data on viscoelasticity are collected either in the linear or in the nonlinear region. When materials are tested in the linear range, material functions do not depend on the magnitude of stress, the magnitude of the deforming strain, or the rate of application of strain. If linear, an applied stress will produce a proportional strain response. The linear range of testing is determined from experimental data. Testing can easily enter the nonlinear range by applying excessive strain (usually greater than 1%) or high deformation rates to the sample. For a large deformation compression test before the initiation of relaxation testing, the strain or stress level is always a few folds higher than its linear viscoelastic range, and hence, shows a typical nonlinear decay trend. Nonlinear viscoelasticity is experimentally and theoretically much more complex than linear viscoelasticity. An ideal viscous body cannot maintain any force=stress in the absence of motion, and, thus, reaches the lowest datum level. On the contrary, an ideal elastic solid is able to attain instantly the force=stress that is equal to the same magnitude that it possessed at the beginning of the relaxation test. It is obvious that a viscoelastic material such as food dough shows an intermediate effect between these two extreme cases. In Figure 21.16, the force curve of a nonlinear compression test is shown. At a constant strain, force depends on time alone with usually three zones: the ﬁrst zone shows a high slope, whereas the third zone has the lowest slope and appears to approach a residual (or an equilibrium) value, whereas the second zone is an intermediate of these two zones. The slope of the ﬁrst zone of the curve is independent of the rate of strain when the sample is compressed to a small strain level [35]. Many processes, such as mastication and swallowing, are only accomplished with very large

Force (N)

Compression

Relaxation

Energy dissipated

Energy stored Time (s)

FIGURE 21.16

Nonlinear compression–relaxation diagram of a viscoelastic body.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 479 16.7.2008 11:12pm Compositor Name: MSubramanian

Rheological Instruments in Food Analysis

479

deformations. The importance of large deformation (nonlinear) in food rheology must not be overestimated. Collecting viscoelastic data relevant to this type of problem involves testing in the nonlinear range of behavior. Practically, these data are quite useful, but from a fundamental standpoint they can only be used for comparative purposes because the theoretical complexity of nonlinear viscoelasticity makes it impractical for most applications [5]. Pure elastic behavior is deﬁned such that when a force is applied to a material, it will instantaneously and ﬁnitely deform, and when the force is released, the material will instantaneously return to its original form. Such a material is called a Hookean solid. The amount of deformation is proportional to the magnitude of force. The rheological representation of this type of solid is a spring. The modulus calculated by applying force perpendicular to the area deﬁned by stress is called the modulus of elasticity (E), the modulus calculated by applying a force parallel to the area deﬁned by stress, or shearing force, is called the shear modulus or modulus rigidity (G). If the force is applied from all directions and the change in volume per original volume is obtained, then the bulk modulus (K) can be calculated. Pure viscous ﬂow of liquid means that the liquid begins to ﬂow with the slightest force and that the rate of ﬂow is proportional to the magnitude of force applied. This liquid ﬂows inﬁnitely until force is removed and upon removal of force, it has no ability to regain its original state. Such a material is called a Newtonian liquid. The rheological representation for this type of liquid is a dashpot, which can be thought of as a piston inside a cylinder. When force is applied to the piston, it moves in or out of the cylinder at constant velocity, the rate depending upon the magnitude of force. When force is removed, the piston remains ﬁxed and cannot return to its original position. A material of this nature has a rheological constant called the coefﬁcient of viscosity. If foods were either Hookean solids or Newtonian liquids, determination of their rheological constants would be simple. However, foodstuffs possess rheological properties associated with both elastic solid and viscous ﬂuid. The rheological representation of this type of material is a body incorporating at least one spring (representing the solid character) and at least one dashpot (representing the viscous character). The number of springs and dashpots in the body and the manner in which they are connected are manipulated to represent different types of viscoelastic materials and to demonstrate how they will behave under a stress or strain [36]. 21.2.4.1

Linear Viscoelasticity

21.2.4.1.1 Stress Relaxation Test In stress relaxation test, an instantaneous strain is applied and the force required to maintain the deformation is observed as a function of time. Frequently used mathematical models for stress relaxation are simple Maxwell, generalized Maxwell, and Peleg and Normand models. The Maxwell model, which is the simplest viscoelastic material representation, consists of an elastic (spring) and a viscous (dashpot) element in series (Figure 21.17). Regardless of whether a particular model contains or does not contain a parallel spring, its relaxation curve will be determined by initial force and ﬁxed deformation. In other words, if any particular Maxwellian model or even conventional nonlinear model is let to relax from initial conditions of higher deformation and force, the relaxation curve will always be above the curves that started at initial conditions in which both deformation and force had smaller values. In Maxwell model, applied force (F) is used instead of stress. The instantaneous force can be replaced by any other decaying parameter such as stress or modulus of elasticity [37]. For a simple Maxwell model, at constant strain, the applied force (F) decays from F1 to F(t), after time t F(t) ¼ F1 exp (t=l1 )

(21:29)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 480 16.7.2008 11:12pm Compositor Name: MSubramanian

480

FIGURE 21.17

Handbook of Food Analysis Instruments

Maxwell model.

where F1 is the decay force l1 is the relaxation time Although an exact deﬁnition of relaxation time is difﬁcult to describe, it can be thought of as the time it takes a macromolecule to be stretched out when deformed [38]. The simple Maxwell model is not sufﬁcient enough to describe the behavior of a linear viscoelastic material. For example, if a constant stress is applied to a Maxwell model, the model exhibits only Newtonian ﬂow and not a retarded elastic deformation, which is experimentally observed in a creep or constant stress test. To avoid this problem, an inﬁnite number of Maxwell models are used in parallel and the resulting model is called a generalized Maxwell model [36,39]. Most viscoelastic foods do not follow the simple Maxwell model (Equation 21.29) and it is necessary to use more complex models to describe their stress relaxation curves. The generalized Maxwell model, consisting of a small number of parallel simple elements, is described using the following equation: F(t) ¼ F1 exp (t=l1 ) þ F2 exp (t=l2 ) þ þ Fn exp (t=ln )

(21:30)

where l1, . . . , ln are the relaxation times F1, . . . , Fn are the decay forces F(t) is the instantaneous force in a stress relaxation test The instantaneous force could be replaced by any other decaying parameter such as stress or modulus of elasticity [37]. For most of the foods a Maxwell model with three terms involving six constants is sufﬁcient enough to represent stress relaxation data. Kokini et al. [34] simulated the relaxation modulus using the generalized Maxwell model for wheat ﬂour dough and wheat gluten (Figure 21.18). It is difﬁcult to express biological materials with a ﬁxed number of elements. The general case of deformation mainly consists of three progressive stages in which different kinds of mechanical phenomena may play the dominant role

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 481 16.7.2008 11:12pm Compositor Name: MSubramanian

481

Rheological Instruments in Food Analysis

4 ⫻ 104 3 ⫻ 104

G(t) (dyne/cm2)

2 ⫻ 104 R 2 = 0.9994 104 9 ⫻ 103 8 ⫻ 103 7 ⫻ 103 6 ⫻ 103 5 ⫻ 103 4 ⫻ 103

10% Experimental data 10% Wagner prediction

3 ⫻ 103 10–1

100

101 Time (s)

102

FIGURE 21.18 Simulation of the 55% moisture gluten dough using 12 element generalized Maxwell model with the experimental and Wagner predicted data. (Requires permission from Kokini, J.L., Wang, C.F., Huang, H., and Shrimanker, S., J. Texture Stud., 26, 421, 1995, Figure 2.)

1. Stage in which no permanent physical change occurs and the deformation is dominantly elastic and rate independent. 2. Stage in which some irreversible changes progressively occur. This stage is characterized by viscoelastic behavior and history-dependent phenomena. 3. Failure and post failure stages. These are characterized by an apparent physical rupture of the material and should be discussed in terms of failure phenomena. From a rheological point of view, the material is still viscoelastic but shows considerably different viscoelastic behavior if compared to its pre-failure stages [40]. In many materials of biological origin, difﬁculty of rheological data analysis arises because of heterogeneous and nonuniform internal structure that does not allow many of the simplifying assumptions in existing theories. Furthermore, most biological materials tend to exchange moisture with the environment, a factor that has a signiﬁcant effect on their rheological properties. In such cases, long-term rheological characteristics in the conventional sense, e.g., equilibrium stress in relaxation or strain in creep, either does not exist or must be extremely difﬁcult to determine. Conventional models and methods of rheological characterization, therefore, have only limited applicability when applied to these kinds of materials. To overcome these difﬁculties, they suggest stress relaxation data to be calculated as a normalized stress (or force) and ﬁt it to the following linear equation: s0 t=(s0 s) ¼ k1 þ k2 t

(21:31)

where s0 is the initial stress s is the decreasing stress at time t k1 and k2 are constants Fitting experimental data to Equation 21.31 is a quick and effective way to handle stress relaxation data [5].

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 482 16.7.2008 11:12pm Compositor Name: MSubramanian

482

Strain

Handbook of Food Analysis Instruments

Viscous flow Retarded elastic region

Instantaneous response to deformation

Time (s)

FIGURE 21.19 behavior.

Typical creep curve showing where various elements of the Burgers model describe ﬂow

21.2.4.1.2 Creep Recovery One of the manifestations of viscoelastic materials is that they undergo creep, i.e., continue to deform under constant stress or load. The distinction between constant stress and constant load (force) is necessary, especially for highly deformable foods, because of the progressive change of the specimen’s cross-sectional area. Thus, a constant load (i.e., dead weight) produces a progressively increasing stress in uniaxial tension and decreasing stress in compression [41]. A typical creep curve is shown in Figure 21.19. The output of creep tests is normally in three forms: 1. Strain–time curves under various selected constant loads 2. Recovery curves after removal of the loads 3. Time to failure (if within reasonable experimental duration) under various loads [42] Creep data may be described in terms of a creep compliance function given by Equation 21.32 in terms of shear deformation. D(t) ¼ g(t)=s

(21:32)

To develop a mechanical analogue for creep behavior the starting point is Kelvin model which contains a spring connected in parallel with a hydraulic dashpot. In creep where the material is allowed to ﬂow after being subjected to a constant shear stress (so), the change in stress with time is zero, resulting in the following equation: g ¼ f (t) ¼ so ½1 exp (t=lret )=E

(21:33)

The Kelvin model shows excellent elastic retardation, but is not sufﬁcient enough to model creep in many biological materials. The solution to this problem is to use a Burgers model which is a Kelvin and a Maxwell model placed in series (Figure 21.20). Data following Burgers model (Figure 21.20) show an initial elastic response because of the free spring, and retarded elastic behavior related to the parallel spring–dashpot combination and Newtonian type of ﬂow after long periods of time because of the free dashpot

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 483 16.7.2008 11:12pm Compositor Name: MSubramanian

483

Rheological Instruments in Food Analysis

FIGURE 21.20

Burgers model.

g ¼ f (t) ¼ (so =Eo ) þ (so =E1 )½1 expðt=lret1 Þ þ (so =E2 )½1 expðt=lret2 Þ þ (so t=mo ) (21:34) lret ¼ m=E, the retardation time of the Kelvin portion of the model. The Burgers model can also be expressed in terms of creep compliance by dividing Equation 21.34 by constant stress: g=so ¼ f (t) ¼ (1=Eo ) þ (1=E1 )½1 expðt=lret1 Þ þ (1=E2 )½1 expðt=lret2 Þ þ (t=mo ) (21:35) which gives D ¼ f (t) ¼ Do þ D1 ½1 expðt=lret1 Þ þ D2 ½1 expðt=lret2 Þ þ (t=mo )

(21:36)

where Do is the instantaneous compliance D1 and D2 are retarded compliances lret1 and lret2 are retardation times of the Kelvin component mo is the Newtonian viscosity of the free dashpot At the beginning of creep there is an instantaneous change in compliance because of the spring in the Maxwell portion of the model. Then, the Kelvin component produces an exponential change in compliance related to the retardation time. After sufﬁcient time has passed, the independent dashpot generates a purely viscous response. Data from the linear portion of the creep curve are related to two parameters: the slope is equal to 1=m0; and the intercept, sometimes called the steadystate compliance, is equal to D0 þ D1 (Figure 21.21).

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 484 16.7.2008 11:12pm Compositor Name: MSubramanian

484

Handbook of Food Analysis Instruments

J0

Compliance

Slope = 1/m0

J1

J1 t1/m0 J0

s = s0

m=0

Creep

0

Recovery t1

0

Time

FIGURE 21.21 Compliance and recovery curves showing compliance. (Requires permission from Steffe, J.F., Rheological Methods in Food Process Engineering, Freeman Press, MI, 1996.)

At t ¼ t1 the load is removed and there is an instantaneous change in compliance equal to D0. The free dashpot causes permanent deformation in the material related to a compliance of t1=m0. If a substance obeying Burgers model is tested in the linear viscoelastic region of material behavior, then the values of D0 and D1 determined from the creep curve will be equal to the values from recovery curve [5]. Peleg has also suggested that creep data could be modeled with the following linear equation [41,43]: t=D ¼ k1 þ k2 t

(21:37)

21.2.4.1.3 Oscillatory Measurements Small amplitude oscillatory measurements are also commonly used to characterize linear viscoelastic properties. The following equations can be used for the storage (G0 ) and loss (G00 ) modulus when generalized Maxwell model is used to model the linear viscoelastic behavior: G0 (v) ¼ 00

G (v) ¼

N X Gi (vli )2 2 i¼1 1 þ (vli )

(21:38)

N X Gi (vli )2 2 i¼1 1 þ (vli )

(21:39)

Kokini et al. [44] predicted storage and loss modulus of wheat ﬂour dough (Figure 21.22). They found that even the predicted G0 and G00 showed higher discrepancy from the experimental data at the higher frequency region, it was important to indicate that both G0 and G00 predicted from stress relaxation spectrum did approach the experimental data at longer time scales when frequencies were around 102 rad=s. If the deformation is done by shearing, then the relaxation modulus can be represented as G(t). Both G(t) and G0 (v) are measures of stored elastic energy upon deformation. Theoretically, a dynamic measurement in a frequency domain is qualitatively equivalent to a transient domain by using t ¼ 1=v [45]. Storage modulus (G0 ) and relaxation modulus (G00 ) are

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 485 16.7.2008 11:12pm Compositor Name: MSubramanian

Rheological Instruments in Food Analysis

485

106 Maxwell’s prediction of G⬘ Maxwell’s prediction of G⬙ 105

Experimental G⬘

G⬘ (Pa); G⬙ (Pa)

Experimental G⬙ 104

103

102

101

10–6 10–5 10–4 10–3 10–2 10–1 100 101 102 103 104 Frequency (rad/s)

FIGURE 21.22 Comparison of the experimental and predicted G0 and G00 using the generalized Maxwell model for 18% protein ﬂour dough at 278C. (Requires permission from Kokini, J.L., Dhanasekharan, M., Wang, C.F., and Huang, H., Trends in Food Engineering, Technomic Publishing Company, Inc., PA, 2000.)

mirror images of each other, G(t) ¼ G0 (1=v). This relationship between storage and relaxation modulus can be used to predict the linear relaxation modulus where experimental values are difﬁcult to evaluate because of sensitivity considerations. 21.2.4.2

Nonlinear Viscoelasticity

Food materials exhibit linear viscoelastic behavior under low-strain conditions. For larger or faster deformations, the compliance is a function of stress or strain, time, and deformation type. Considering that the food materials are generally applied to large deformation conditions either during processing or eating, information obtained from the nonlinear measurements are also useful. Linear viscoelastic theory cannot be used to predict nonlinear response [46]. Nonlinear effects include transient ﬂows with large strains or strain rates. The relaxation modulus is a function of both time and strain. The Boltzmann superposition principle was used to develop the linear viscoelastic theory; however, this is not sufﬁcient enough for the nonlinear viscoelastic theory. There exists no generally valid quantitative model for the interpretation of the nonlinear viscoelasticity of polymeric liquids since the response of the material to a stress or strain imposed at time t depends not only on time but also on magnitude of the strain and stress [45]. One of the approaches that can be used to analyze nonlinear viscoelasticity is the formulation of nonlinear constitutive equations that make use of empirical equations. There are two basic means of incorporating nonlinearity to existing models: Jaumann derivative of strain rate that follows the rotation and translation of a material element with respect to time. One application is the Oldroyd model [47]: Dg_ ij Dt ij _ 2 ¼ m gl t ij þ l1 Dt Dt

(21:40)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 486 16.7.2008 11:12pm Compositor Name: MSubramanian

486

Handbook of Food Analysis Instruments

The dynamic mechanical properties are obtained using the following equations: h0 (v) ¼

m(1 þ l1 l2 v2 ) 1 þ l21 v2

(21:41)

vm(l1 l2 ) 1 þ l21 l22

(21:42)

h00 (v) ¼

where l1, l2, and m are constants that reduce to a Newtonian model when l1 ¼ l2. Spriggs [48] introduced an additional parameter, «, and incorporated it into a modiﬁed Maxwell model as follows: t p þ lp F« p« ¼ 2hp g_

(21:43)

The parameters lp and hp can be deﬁned from the molecular theories of Zimm [49] and Rouse [50] as follows: lp la p

(21:44)

ho l p h ¼ a o hp ¼ P 1 p Z(a) lp

(21:45)

and

p¼1

where Z(a) ¼

1 X

pa

(21:46)

p¼1

which is the Reimann zeta function ho is the shear viscosity l, a, and « are constants of the model The generalized Maxwell model incorporated the memory function m(t t0 ) in the following manner: t ij ¼

ðt (X 1 hp 1

p¼1

) (t t) _ 0 ) dt 0 g(t exp lp l2p

(21:47)

Bird and Carreau [51] introduced shear dependence into the memory function by including dependence upon the second invariant of the rate of strain tensor as a means of accounting for nonlinear behavior in an integral model. By the interpretation of this tensor, the memory function becomes as follows: 3 2 tt 0 1 exp X hp l2p 4 5 (21:48) m(t t0 , t 0 ) ¼ 2 1 2 0 p¼1 l2p 1 þ 2 l1p II(t ) The constants l1p and l2p are related with network formation and dissolution occurring at separate rates [52]. II is the invariant of the rate of strain tensor which is equal to y2. The empirical modiﬁcation for the determination of hp, l1p, and l2p are as follows:

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 487 16.7.2008 11:12pm Compositor Name: MSubramanian

487

Rheological Instruments in Food Analysis

l1p hp ¼ ho P 1 l1p

(21:49a)

p¼1

1 þ n1 a1 p þ n1 1 þ n2 a2 ¼ l2 p þ n2

l1p ¼ l1

(21:49b)

l2p

(21:49c)

where n1 ¼ n2 ¼ 1, l1 ¼ l2, and a1 ¼ a2. In Sprigg’s model, n1 ¼ n2 ¼ 0, l1 ¼ l2 ¼ cl, and a1 ¼ a2 ¼ 2. In nonlinear viscoelasticity, viscosity is shear-rate dependent and the ﬁrst normal-stress difference is nonzero. The transient viscosity at start up of steady-shear ﬂow depends on shear rate and the relaxation modulus depends on strain magnitude. Kokini et al. [34] mentioned that in the case of oscillatory shear, the strain amplitude is usually less than 0.05 to determine the storage modulus G0 and the loss modulus G00 . When deformations are more rapid or larger, the linear theory is no longer valid. Therefore, oscillatory properties like G0 and G00 will no longer be useful because they are based on the assumption that the stress is sinusoidal, and when nonlinearities occur this is no longer valid. In the case of nonlinear viscoelastic behavior, there is also no universal theory like the Boltzmann superposition principle to help in describing ﬂows where neither strain nor strain rate is small. Nonlinear constitutive equations that evolve from empirical equations are used to guide in the study of the experimental results. There are two measures of strain that are useful in data analysis: Cauchy tensor, Cij, and Finger tensor, Bij. For simple extension, the tensors are shown as follows: 2

1 n ½g(t2 ) g(t1 ) o 6 Cij (t1 , t2 ) ¼ 4 ½g(t2 ) g(t1 ) 1 þ ½g(t2 ) g(t1 )2 0 0 2n o ½g(t1 ) g(t2 ) 1 þ ½g(t1 ) g(t2 )2 6 Bij (t1 , t2 ) ¼ 4 ½g(t1 ) g(t2 ) 1 0 0

3 0 7 05 1 3 0 7 05 1

(21:50)

(21:51)

and for simpler extension the components become 2

3 expf2½«(t2 ) «(t1 )g 0 0 5 0 0 expf½«(t2 ) «(t1 )g Cij (t1 , t2 ) ¼ 4 0 0 expf½«(t2 ) «(t1 )g 2 3 expf2½«(t1 ) «(t2 )g 0 0 5 Bij (t1 , t2 ) ¼ 4 0 0 expf½«(t1 ) «(t2 )g 0 0 expf½«(t1 ) «(t2 )g

(21:52)

(21:53)

Since the components of the tensors do not equal to zero when a material is in the undeformed state, the Cauchy and Finger strain tensors are generally used as follows: Cauchy strain tensor ¼ Cij dij Finger strain tensor ¼ dij Bij

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 488 16.7.2008 11:12pm Compositor Name: MSubramanian

488

Handbook of Food Analysis Instruments

The Finger tensor has three scalar invariants for a given deformation, which can be calculated as follows: I1 (Bij ) ¼ B11 þ B22 þ B33 I2 (Bij ) ¼ C11 þ C22 þ C33 I3 (Bij ) ¼ 1 Lodge [53] generalized the Boltzmann superposition principle to formulate a theory of nonlinear viscoelasticity. The obtained equation can be used to analyze the rubber-like liquid behavior. Lodge assumed that time dependency in the model is created and lost continuously by the network junctions. The memory function from Lodge’s theory is as follows: N X Gi (t t 0 ) m(t t ) ¼ exp li li i¼1 0

(21:54)

The relaxation modulus corresponding to this memory function is identical to that of the generalized Maxwell model which is as follows: ðt t ij (t) ¼ 1

Gi (t t 0 ) Bij (t, t 0 ) dt 0 li li

(21:55)

The rubber-like liquid theory is applicable in cases where the viscosity and ﬁrst normal-stress difference is independent of shear rate. The equation proposed by Burnston, Kearsley, and Zapas [54] that is known as BKZ equation can be used for food materials that show a viscosity and ﬁrst normal-stress coefﬁcient dependent on shear rate. The BKZ equation can be proposed in the following form: ðt @m @m 2 Cij (t, t 0 ) 2 Bij (t, t 0 ) dt 0 tij ¼ @I1 @I2

(21:56)

1

where m is a time-dependent elastic energy potential function given as follows: m ¼ m(I1 , I2 , t t 0 )

(21:57)

A modiﬁed version of BKZ equation is more practical with the introduction of a relaxation modulus, which is as follows: m(I1 , I2 , t t 0 ) ¼ m(t t 0 )U(I1 , I2 )

(21:58)

Wagner [55] simpliﬁed the equation to the following form: M(t t 0 , I1 , I2 ) ¼ m(t t 0 )h(I1 , I2 )

(21:59)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 489 16.7.2008 11:12pm Compositor Name: MSubramanian

489

Rheological Instruments in Food Analysis

where h(I1, I2) is the damping function. The relaxation modulus is as follows: ðt t ij (t) ¼

m(t t 0 )h(I1 , I2 )Bij (t, t 0 ) dt 0

(21:60)

1

The damping function in Wagner’s equation has to be determined experimentally. Several researchers like Wagner [55], Osaki [56], Zapas [57], and Soskey and Winter [58] proposed damping functions for shear ﬂows, respectively, as follows: h(g) ¼ eng

(21:61)

h(g) ¼ a½exp (n1 g) þ (1a) exp (n2 g)

(21:62)

h(g) ¼

1 1 þ ag2

(21:63)

h(g) ¼

1 1 þ agb

(21:64)

For extensional ﬂows, Meissner [59] used the following equation:

1 h(«) ¼ a e2« þ (1 a)em«

(21:65)

Huang and Kokini [60] obtained dumping functions for hard wheat ﬂour dough. They found that the damping function with two parameters from Soskey and Winter [58] ﬁts the data best (Figures 21.23 and 21.24).

[T12(t, g0)/g0]/[T12(1,0.005)/0.005]

1.0 Experimental h(g0) = t /1 + (ag0b ) where a = 12.4547 b = 0.7761 0.1

0.01

0

1

2

3

4

Step shear strain (g0)

FIGURE 21.23 Steady shear and extensional rheological measurements for 18.8% protein hard wheat ﬂour dough. (Requires permission from Huang, H. and Kokini, J.L., Prog. Trends Rheol. IV, Proc. fourth Eur. Rheol. Conf., Sevilla, Spain, 240, 1994, Figure 2.)

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 490 16.7.2008 11:12pm Compositor Name: MSubramanian

490

Handbook of Food Analysis Instruments

109 Wagner & predictions Experimental 108 g· = 0.0001 h (t,g) (Poise)

107

·

g· = 0.001

106

g· = 0.005 g· = 0.3

105

104

103 100

101

102 Time (s)

103

104

FIGURE 21.24 Wagner constitutive model prediction for the shear-stress growth function of 18.8% protein ﬂour dough. (Requires permission from Huang, H. and Kokini, J.L., Prog. Trends Rheol. IV, Proc. fourth Eur. Rheol. Conf., Sevilla, Spain, 240, 1994, Figure 4.)

21.3 CONCLUDING REMARKS In this chapter, the rheological behavior of food materials has been summarized and demonstrated with examples. The rheological properties of food materials are quite important since they are used to predict the structure. Selected methodologies existed mainly from the works that have been carried out for years in Dr. Kokini’s laboratory. Basic information on instrumentation and modeling of data were also given. A better knowledge on interpretation of rheological data enables one to interpret and regulate food texture and structure. It is hoped that this work will serve as an introduction to the study of rheology of food materials. Food materials are complex in structure, they do not consist of just one type of material, but several types of materials with various properties like ﬂour, fats and oils, protein, water, sugar, etc., which makes the rheological measurements and simulation of data difﬁcult. Therefore, it is necessary to develop better methods and constitutive models that can predict rheological properties. The lists of food materials that have been discussed were selected arbitrarily, but examples were given for each of the solid, viscoelastic, and ﬂuid behaviors. Up to now rheology has been used not only for quality control, but also as an in-line control in plants. By the help of rheological measurements, one can predict how any change in the formulation and process will affect the product. Day by day rheometers are developed to do more accurate and reproducible measurements. Rheological measurements can be used as a good tool to determine various food material properties like glass transition temperatures and polymeric properties, and can simulate several processes like extrusion, mixing, and dough sheeting. However, there is still much to be done in this area of science like developing better and standard methods. Also constitutive models need to be developed especially for nonlinear measurement techniques.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 491 16.7.2008 11:12pm Compositor Name: MSubramanian

Rheological Instruments in Food Analysis

491

REFERENCES 1. Faridi, H. and Faubion, J.M., Dough Rheology and Baked Product Texture, AVI Book Publishers, New York, 4, 1990. 2. Bourne, M.C., Food Texture and Viscosity: Concept and Measurement, Academic Press, London, United Kingdom, 89, 124, 141, 2002. 3. Tsukakoshi, Y., Naito, S., and Ishida, N., Probabilistic characteristics of stress changes during cereal snack puncture, J. Texture Stud., 38: 220, 2007. 4. Deman, J.M., Voisey, P.W., Rasper, V.F., and Stanley, D.W., Rheology and Texture in Food Quality, the AVI Publishing Company, Inc., Westport, CT, 200, 1976. 5. Steffe, J.F., Rheological Methods in Food Process Engineering, Freeman Press, East Lansing MI, 6, 295–297, 304–309, 1996. 6. Barnes, H.A., Hutton, J.F., and Walters, K., An Introduction to Rheology, Elsevier Science Publishers B.V., Amsterdam, the Netherlands, 167, 1989. 7. Chen, J., Karlsson, C., and Povey, M., Acoustic envelope detector for crispness assessment of biscuits, J. Texture Stud., 36: 139, 2005. 8. Domingo, B.J. and Morris, S.A., Mechanical performance studies on extruded cornstarch-based plastic manufactures, J. Appl. Polym. Sci., 71: 2147, 1999. 9. Farouk, M.M., Zhang, S.X., and Waller, J., Meat spaghetti tensile strength and extensibility as indicators of the manufacturing quality of thawed beef, J. Food Qual., 28: 452, 2005. 10. Guinea, G.V., Rojo, F.J., and Elices, M., Brittle failure of dry spaghetti, Eng. Failure Anal., 11: 705, 2004. 11. Alamprese, C., Casiraghi, E., and Pagani, M.A., Development of gluten-free fresh egg pasta analogues containing buckwheat, Eur. Food Res. Technol., 225: 205, 2007. 12. Dervisoglu, M. and Kokini, J.L., The steady rheology and ﬂuid mechanics of four semi solid foods, J. Food Sci., 51: 541, 1986. 13. Bagley, E.B., End corrections in the capillary ﬂow of polyethylene, J. Appl. Phys., 28: 624, 1957. 14. Barbosa-Canovas, G.V., Kokini, J.L., Ma, L., and Ibarz, A., The rheology of semiliquid ﬂuids, Adv. Food Nutr. Res., 39: 1, 1996. 15. Karnjanolarn, R. and McCarthy, K.L., Rheology of different formulations of milk chocolate and the effect on coating thickness, J. Texture Stud., 37: 668, 2006. 16. Cuvelier, G., Tanon, C., and Launay, B., Xanthan-carob mixtures at low concentration: Viscometric study, Food Hydrocolloids, 1(5=6): 583, 1987. 17. Bhattacharya, S., Yield stress and time-dependent rheological properties of mango pulp, J. Food Sci., 64(6): 1029, 1999. 18. Yoo, B., Rheological properties of hot pepper-soybean paste, J. Texture Stud., 32: 307, 2002. 19. Choi, Y.H. and Yoo, B., Characterization of time dependent ﬂow properties of food suspensions, Int. J. Food Sci. Technol., 39: 801, 2004. 20. Kus, S., Altan, A., and Kaya, A., Rheological behavior and time dependent characterization of ice cream mixture with different salep content, J. Texture Stud., 36: 273, 2005. 21. Basu, S., Shiuhare, U.S., and Raghavan, G.S.U., Time dependent rheological characteristics of pineapple jam, Int. J. Food Eng., 3(3): 1, 2007. 22. Weltman, R.N., Breakdown of thixotropic structure as a function of time, J. Appl. Phys., 14: 343, 1943. 23. Hahn, S.L., Ree, T., and Eyring, H., Flow mechanism of thixotropic substances, Ind. Eng. Chem., 51: 856, 1959. 24. Figoni, P.I. and Shoemaker, C.F., Characterization of time dependent ﬂow properties of mayonnaise under steady shear, J. Texture Stud., 14: 431, 1983. 25. Kaya, A. and Sozer, N., Rheological behavior of sour pomegranate juice (Punica granatum L.), Int. J. Food Sci. Technol., 40: 223, 2005. 26. Chang, C.N., Dus, S., and Kokini, J.L., Measurement and interpretation of batter rheological properties, in Batters and Breadings in Food Processing, Kulp, K. and Loewe, R. (Eds.), American Association of Cereal Chemists, Inc., St. Paul MN, 199–226, 1990. 27. Kokini, J.L. and Chou, T.C., The steady rheology and ﬂuid mechanics of four semi solid foods, J. Texture Stud., 24: 117, 1993. 28. Kokini, J.L. and Dickie, A., An attempt to identify and model transient viscoelastic ﬂow in foods, J. Texture Stud., 25: 539, 1981.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 492 16.7.2008 11:12pm Compositor Name: MSubramanian

492

Handbook of Food Analysis Instruments

29. Dickie, A.M. and Kokini, J.L., An improved model for food thickness from non Newtonian ﬂuid mechanics in the mouth, J. Food Sci., 48(1): 57, 1983. 30. Leider, P.J. and Bird, R.B., Squeezing ﬂow between parallel disks-I. Theoretical analysis, Ind. Eng. Chem. Fundam., 13: 4, 336, 1974. 31. Kokini, J.L., Rheological properties of foods, in Handbook of Food Engineering, Heldman, D.R. and Lund, D.B. (Eds.), Marcel Dekker, New York, 19, 1992. 32. Campanella, O.H. and Peleg, M., Analysis of the transient ﬂow of the mayonnaise in a coaxial viscometer, J. Rheol., 31: 439, 1987. 33. Larson, R.G., Nonlinear shear relaxation modulus for a linear low-density polyethylene, J. Rheol., 29: 823–831, 1985. 34. Kokini, J.L., Wang, C.F., Huang, H., and Shrimanker, S., Constitutive models of foods, J. Texture Stud., 26: 421, 1995. 35. Yadav, N., Roopa, B.S., and Bhattacharya, S., Viscoelasticity of a simulated polymer and comparison with chickpea ﬂour doughs, J. Food Process Eng., 29: 234, 2006. 36. Rao, M.V.N. and Skinner, G.E., Rheological properties of solid foods, in Engineering Properties of Foods, Rao, M.A. and Rizvi, S.S.H. (Eds.), Marcel Dekker, New York, 215–226, 1986. 37. Khazaei, J. and Mann, D.D., Effects of temperature and loading characteristics on mechanical and stress-relaxation behavior of sea buckthorn berries. Part 3: Relaxation behavior, CIGR J. Sci. Res. Dev., 6: 1–12, 2004. 38. Cheng, Y., Shimizu, N., and Kimura, T., The viscoelastic properties of soybean curd (tofu) as affected by soymilk concentration and type of coagulant, Int. J. Food Sci. Technol., 40: 385, 2005. 39. Peleg, M. and Normand, M.D., A computer assisted analysis of some theoretical rate effects in mastication and in deformation testing of foods, J. Food Sci., 47: 1573, 1982. 40. Peleg, M. and Calzada, J.F., Stress relaxation of deformed fruits and vegetables, J. Food Sci., 41:1325, 1976. 41. Purkayastha, S., Peleg, M., Johnson, E.A., and Normand, M.D., A computer aided characterization of the compressive creep behavior of potato and cheddar cheese, J. Food Sci., 50: 45, 1985. 42. Peleg, M., A model for creep and early failure, Mater. Sci. Eng., 40: 197, 1979. 43. Peleg, M., Linearization of relaxation and creep curves of solid biological materials, J. Rheol., 24(4): 451, 1980. 44. Kokini, J.L., Dhanasekharan, M., Wang, C.F., and Huang, H., Integral and differential linear and nonlinear constitutive models for the rheology of wheat ﬂour doughs, in Trends in Food Engineering, Lozano, J.E., Anon, C., Panado-Arias, E., and Canovas, G.V.B. (Eds.), Technomic Publishing Company, Inc., Lancaster PA, Chapter 9, 23, 2000. 45. Ferry, J.D., Viscoelastic Properties of Polymers, John Wiley & Sons, New York, 177–223, 1980. 46. Dealy, J.M. and Wissburn, K.F., Melt Rheology and Its Role in Plastics Processing: Theory and Application, VNR, NY, 1990. 47. Oldroyd, J.G., Non-Newtonian effects of steady motion of some idealized elasto-viscous liquids, Proc. R. Soc. Ser. A, 245: 278, 1958. 48. Spriggs, T.W., A four constant model for viscoelastic ﬂuids, Chem. Eng. Sci., 20: 931, 1965. 49. Zimm, B.H., Dynamics of polymer molecules in dilute solution: Viscous ﬂow birefringence and dielectric loss, J. Chem. Phys., 24: 269, 1956. 50. Rouse, J.E., A theory of linear viscoelastic properties of dilute solutions of coiling polymers, J. Chem. Phys., 21: 1272, 1953. 51. Bird, R.B. and Carreau, P.J., A nonlinear viscoelastic model for polymer solutions and melts, Chem. Eng. Sci., 23: 427, 1968. 52. Lodge, A.S., A theory network of constrained elastic recovery in concentrated polymer solutions, Rheol. Acta, 1: 158, 1958. 53. Lodge, A.S., Elastic Liquids, Academic Press, New York, 1964. 54. Bernstein, B., Kearsley, E.A., and Zapas, L.J., Thermodynamics of perfect elastic solids, J. Res. Nat. Bur. Stand, 68B: 103, 1964. 55. Wagner, M.H., Analysis of time dependent non-linear stress growth data for shear and elongational ﬂow of a low density branched polyethylene melt, Rheol. Acta, 15: 2, 136, 1976. 56. Osaki, K., On the damping function of shear relaxation modulus for entangled polymers, Rheol. Acta, 32(5), 429, 1993.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 493 16.7.2008 11:12pm Compositor Name: MSubramanian

Rheological Instruments in Food Analysis

493

57. Zapas, L.J., Viscoelastic behavior under large deformations, J. Res. Nat. Bur. Stand., 70A: 525, 1966. 58. Soskey, P.R. and Winter, H.H., Large step shear strain experiments with parallel-disk rotational rheometers, J. Rheol., 28: 625, 1984. 59. Meissner, J., Dehnungsverhalten von Polyathylen-Schmelzen, Rheol. Acta, 10: 230–239, 1971. 60. Huang, H. and Kokini, J.L., Steady shear and extensional rheological measurements of hard wheat ﬂour doughs and their simulation using Wagner constitutive model, Prog. Trends Rheol. IV, Proc. fourth Eur. Rheol. Conf., September, Sevilla, Spain, 1994.

Otles/Handbook of Food Analysis Instruments 45660_C021 Final Proof page 494 16.7.2008 11:12pm Compositor Name: MSubramanian

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 495 16.7.2008 11:15pm Compositor Name: MSubramanian

Electron and 22 Scanning Transmission Electron Microscopies in Food Analysis José M. Aguilera and Pedro Bouchon CONTENTS 22.1 22.2

Introduction ........................................................................................................................ 495 Principles and Theory ........................................................................................................ 498 22.2.1 Transmission Electron Microscopy ..................................................................... 498 22.2.2 Scanning Electron Microscopy ............................................................................ 500 22.3 Instrumentation .................................................................................................................. 503 22.3.1 Transmission Electron Microscopes .................................................................... 503 22.3.1.1 Conventional Transmission Electron Microscopes ............................ 503 22.3.1.2 Cryogenic Transmission Electron Microscopy .................................. 503 22.3.1.3 Scanning Transmission Electron Microscopy .................................... 503 22.3.2 Scanning Electron Microscopes .......................................................................... 504 22.3.2.1 Conventional Scanning Electron Microscopes ................................... 504 22.3.2.2 Low Vacuum Scanning Electron Microscopes .................................. 504 22.3.2.3 Cryogenic Scanning Electron Microscopy ......................................... 505 22.3.3 Analytical Electron Microscopy Capabilities ...................................................... 505 22.3.3.1 X-Ray Microanalysis .......................................................................... 505 22.3.3.2 Electron Energy Loss Spectrometry ................................................... 506 22.3.3.3 Raman Scanning Electron Microscopy .............................................. 507 22.4 Applications to Food Science ............................................................................................ 507 22.4.1 Transmission Electron Microscopy in Food Science .......................................... 507 22.4.2 Scanning Electron Microscopy in Food Science ................................................ 508 22.5 Future Trends ..................................................................................................................... 510 References ..................................................................................................................................... 510

22.1 INTRODUCTION Food technology may be deﬁned as a controlled attempt to preserve, transform, create, or destroy a structure that has been imparted by nature or processing. In order to achieve this complex goal, it is required to understand the materials (building blocks) and ingredients used in their manufacture [1]. In a ﬁrst attempt, understanding is acquired through empiricism, along with visual inspection coupled with traditional physical probing. However, unaided visual inspection is only adequate when examining gross structural information; observation of ﬁne arrangements requires microscopic assistance. This is because most structural elements that contribute to food identity and 495

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 496 16.7.2008 11:15pm Compositor Name: MSubramanian

496

Handbook of Food Analysis Instruments

quality are below the 100 mm range. Among them we ﬁnd plant cells and cell walls, meat ﬁbers, starch granules, protein bodies, food polymer and fat crystal networks, membranes and interfaces, crystals, oil droplets, emulsions, gas bubbles, and particles of colloidal nature. As a consequence, the relevant scale at which most important transformations occur during processing is beyond the resolution of the naked eye (e.g., ice crystal growth during freezing, starch swelling, internal cracks developed during drying, etc.). Improvements on the quality of existing foods and creation of new products to satisfy growing and demanding consumer’s needs are largely based on interventions at the microscopic level. This product-driven process engineering era, as coined by Aguilera [2], requires building the right (micro)structures and therefore, understanding the functionality of the structural elements prior to and after processing. To do so, integration of basic science, novel laboratory techniques and new approaches, such as advance microstructural analysis, are of paramount importance. Light microscopy (LM), so called because it employs visible light to detect small objects, is probably the most well-known and well-used microscopy tool in food research. As stereomicroscopy it can be used to characterize the surface structure of foods. The compound microscope, in its many versions, is routinely used to examine food sections due to its low price, and relatively ease of use and simple sample preparation. In fact, the advent of the electron lens in biological studies has complemented its use in structural studies. However, new developments, such as the confocal laser scanning microscope and its advantage to make noninvasive ‘‘optical sections’’ to study the interior of foods, show how this technology has still great potential [3]. Basic research on electron beams started early in the 1900s and had an important impact in the way biological microstructure was examined. By means of an electron beam encased in a high vacuum chamber and focused into the specimen by electromagnetic lenses, it was possible to create an image through transmitted electrons. The ﬁrst transmission electron microscope (TEM) was commercialized around 1940 and food scientists soon incorporated this new technique. The new device allowed not only detailed examination of specimens, because of the higher magniﬁcation provided (the upper limit rose from 1,000X to 10,000X) and higher resolution (Table 22.1 compares different microscopy techniques), but also enhanced contrast and sharpness of details (Figure 22.1). Despite these revolutionary improvements, scientists soon realized that the new electron imaging techniques had some limitations: physical problems linked to the high vacuum to be maintained, the electron beam itself, and difﬁculties associated to sample preparation and ultrasectioning. In fact, artifacts are a major drawback. Scientists need to examine the microstructure of interest; however, the prepared specimen does not necessarily match with it. It has been

TABLE 22.1 Comparison of Microscopes

General use Resolution (nm) Magniﬁcation (X) Depth of ﬁeld at 500X (mm) Specimen preparation Specimen thickness

Light Microscopy

Scanning Electron Microscope

Transmission Electron Microscope

Surface and sections 200–500 10–1,500 2 Easy Thick

Surface structure 3–6 20–300,000 1,000 Easy Very thin

Thin sections 0.1–1 200–1,500,000 800 Difﬁcult Reﬂectance

Source: Adapted from Aguilera, J.M. and Stanley, D.W., Microstructural Principles of Food Processing and Engineering, 2nd ed., Aspen, Gaithersburg, 1999.

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 497 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

497

FIGURE 22.1 TEM micrograph of freeze-fractured cheese (mag. 30,000X at an accelerating voltage of 200 keV), showing the high resolving power of this technique. (Courtesy of Jeol.)

emphasized that magniﬁcation by itself does not guarantee enhanced image detail; also, structural organization is frequently more meaningful than observation of microstructural details. The scanning electron microscope (SEM) debuted in 1942 with the ﬁrst commercial instrument built in the 1960s. This new device was welcomed by scientists since it overcame some of the disadvantages of the former equipment, allowing three-dimensional (3D) imaging with a great depth of ﬁeld, 500 times that of a light microscope at the same magniﬁcation. This late development was mainly due to the electronics involved in scanning the beam of electrons across the sample [4]. Also, the SEM allowed surface characterization of external or internal features (by fracture) without the need to prepare ultrathin sections, so scientists rapidly began examining food specimens. Nowadays, this technique constitutes one of the most valuable tools in microstructural food analysis [5–7]. Innovations continue to be made in electron microscopy. In terms of equipment, advances include development of devices that require minimal sample preparation (e.g., environmental SEM, variable pressure SEM), exert minimal sample beam damage (e.g., medium-low voltage TEM, SEM), or allow work under cryogenic conditions (e.g., cryo-TEM, cryo-SEM), and those which combine both principles (e.g., scanning transmission electron microscope or STEM). Also, ancillary methodologies such as backscattered electron analysis and x-ray microanalysis have been developed and perfected. Advances in digital display technology along with image analysis software development have helped to overcome the tediousness of analogue image acquisition, providing analytical capabilities along with high resolution observation.

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 498 16.7.2008 11:15pm Compositor Name: MSubramanian

498

Handbook of Food Analysis Instruments

22.2 PRINCIPLES AND THEORY 22.2.1 TRANSMISSION ELECTRON MICROSCOPY Every electron microscope (scanning or transmission) must have a source of high energy primary electrons, the electron gun, to produce a ﬁne beam of electrons of precisely controlled energy. Several types of electron guns are found in different equipment. Most (traditional) units use thermionic emitters (tungsten or LaB6), but new microscopes are increasingly equipped with ﬁeld emission sources (cold, thermal or Schottky), which provide enhanced performance, reliability, and lifetime [8]. As shown in Figure 22.2, when a primary beam impinges a sample, several interactions occur, which generate a variety of signals that can be captured to generate images. The TEM makes use of transmitted electrons, whereas the SEM mainly captures secondary or backscattered electrons because these vary primarily as a result of differences in topography. Transmission electron microscopes are named after transmission light microscopes since their internal structure is analogous (Figure 22.3). In this case, a beam of electrons instead of light photons are used to form a magniﬁed image of the specimen, providing a thousand-fold increase in resolving power. The voltage applied to achieve the required beam speed is in the 40–300 keV range. The electron beam is focused using magnetic lenses, making use of the fact that electron beams are deﬂected by magnetic ﬁelds, acting as the converging glass lens. Conventional TEM forms the ﬁnal image by focusing transmitted electrons on a ﬂuorescent screen or onto a photographic plate. State-of-the-art equipment are computerized, intuitive, simple to operate (usually based on an MS WindowsTM platform), and the resulting signal is saved as a digital image. Images can be visualized on a computer screen, printed directly or stored without any loss for later

Primary beam Secondary electrons Backscattered electrons X-rays

Transmitted electrons

FIGURE 22.2 Diagram showing the several interactions between electrons and specimen. (From Aguilera, J.M. and Stanley, D.W., Microstructural Principles of Food Processing and Engineering, 2nd ed., Aspen, Gaithersburg, 1999.)

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 499 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

Electron source

499

Light source

Anode Aperture Double lens

Condenser

Detection of secondary and reflected radiation

Specimen

Objective Aperture Intermediate electron lens Projector Electron projector lens Transmitted signal image Final image Electron optical system

Light optical system

FIGURE 22.3 Schematic representation of imaging systems of TEMs (left) and LM (right). (From Kessel, R.G. and Shih, C.Y., Scanning Electron Microscopy in Biology, Springer-Verlag, New York, 1974.)

observation or as a permanent record. Images can be further processed and analyzed to get quantitative information giving an incredible degree of convenience and ﬂexibility. To ensure a clean electron path through the tube a high vacuum is required. The vacuum ranges between 104 and 105 Torr (1 Torr ¼ 1 mm Hg) when using thermionic emitters. Field emission sources, though, require higher vacuums (1010 Torr) or better to operate reliably. To maintain such a high vacuum, specimens need to be extremely dry. In addition, samples must be ultrasectioned to allow electron transmission, and must be strong enough to resist beam damage. These factors severely limit the types of specimens that can be viewed; therefore, a great compromise must be made in order to take advantage of TEM magniﬁcation (up to 1,500,000X). New equipment can provide excellent image contrast even from thick or unstained thin specimens, as will be later explained. Recent studies in relation to the development of an environmental transmission electron microscope (ETEM) make think that such a device will be soon available commercially. Recent research has shown that it is possible to equip a conventional TEM with an environmental cell into which gases can be introduced [10,11]. Preparation of biological samples (food) for TEM observation is rather complicated and difﬁcult compared to those required for other microscopy techniques. Three approaches can be followed. If the sample is thin enough it can be directly mounted. This rarely occurs in food samples and more likely the sample will have to be ultramicrotomed to get a sufﬁciently thin section (<100 nm). Another option is to make a replica of the sample. In this approach, the surface details of a thick sample are reproduced using a shadow-casting technique, followed by carbon coating. In a thin

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 500 16.7.2008 11:15pm Compositor Name: MSubramanian

500

Handbook of Food Analysis Instruments

section, tissue must be ﬁrstly ﬁxed. Fixation procedures can be quite laborious and it is common to use various agents sequentially. The ﬁnal procedure depends on sample structure and composition. Most common ﬁxatives are aldehydes and osmium tetroxide. The latter acts as a protein ﬁxative and also as electron stain, which is a major advantage. Besides, osmium tetroxide is uniquely suitable for lipid ﬁxation. As will be later explained, cryogenic ﬁxation by ultrarapid freezing is also an alternative. Following washing and dehydration, a suitable embedding medium (e.g., epoxy or polyester resins, methacrylates) must be used in order to provide support for sectioning. The obtained contrast between features depends on the differential electron opacity of the several constituents. Substances that combine with speciﬁc components to increase their molecular density through enhanced electron scattering called negative stains (e.g., heavy metals such as osmium, lead, tungsten, uranyl salts) are frequently used.

22.2.2 SCANNING ELECTRON MICROSCOPY Since their invention, SEM devices have been widely used by food scientists, because they overcome many of the limitations imposed by TEM. Sample preparation is easier and introduces fewer artifacts, mainly because no ﬁxation and sectioning are required. A major advantage of this technique is its capability of obtaining a 3D image of the surface of a wide range of materials, with great resolution (1–5 nm) and large depth of ﬁeld (Figure 22.4). The latter is in part responsible for the 3D appearance of the specimen image. Some drawbacks related to sample preparation still remain in traditional equipment. These are mainly due to the high vacuum to be maintained, which demands total dehydration of the sample, to the fact that nonconductive materials must be coated with a thin layer of a conducting metal to avoid sample charge, and to the harm that can be produced by the electron beam itself. Brieﬂy, specimen preparation includes ﬁxing (to halt unwanted structurally degrading reactions, if needed), dehydration, and coating. Dehydration is a critical step and is needed unless the sample is low in moisture content (less than approx. 10% w.b.). This step must be performed with great care to avoid distortion and shrinkage. Air drying is generally undesirable, and the procedures of critical point drying and freeze-drying are the most widely used. Thereafter, the specimen is usually ﬁxed to a metal stub using conducting glue and metal coated using evaporative or sputter techniques. New microscopes, however, permit examining nonconducting specimens without intrusive preparation and the possibility not only to examine moist samples but also to introduce water vapor to avoid dehydration damage during the observation. The basic components of the SEM are the electron gun, the magnetic lens system, the electron collector, the visual and recording cathode ray tubes, and the associated electronics. As explained by Goldstein et al. [8], the earliest recognized work describing the concept of an SEM device is from Knoll [12]. Later von Ardenne [13] added scanning coils to a TEM, enabling STEM. The ﬁrst SEM

FIGURE 22.4 Comparison of LM (A, mag. 65X) and SEM (B, mag. 65X at an accelerating voltage of 15 keV) images of frozen bean curd. These images clearly show the difference in depth of ﬁeld between both techniques. (Courtesy of Jeol.)

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 501 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

501

used to examine thick specimens was described by Zworykin et al. [14]. In SEM, both surface and internal features of a thick specimen can be studied, depending on the preparative technique used. A wide range of magniﬁcations can be used (5–1,000,000X), with a great depth of ﬁeld. New equipment can accommodate samples as large as 30 cm in diameter. As in TEMs, new microscopes are computerized and simple to operate, so that computer storage and processing of images are possible. The primary electron beam is generated by the electron gun by one of the electron sources described earlier (thermionic emitters or ﬁeld emission sources). Generated electrons are drawn to the anode by a voltage difference, known as the accelerating voltage, where they pass through an aperture to produce the electron beam (Figure 22.5). SEMs can operate between 0.1 and 30 keV. The path of the electron beam must be vacuum evacuated before it impacts the specimen to avoid collision with gas molecules. The electron beam is focused by one or two condensers, passes through pairs of electromagnetic scanning coils, deﬂecting the beam, and leaves the ﬁnal lens into the specimen chamber, where it strikes the sample surface obliquely. The degree of obliqueness is termed the tilt. The deﬂection system moves the beam along a line, through discrete locations, and then along the next line below the ﬁrst previous one, until a rectangular raster is generated. The magniﬁcation of the image is the ratio of the length of the raster on the viewing screen to the corresponding length of the raster in the specimen. Increase in magniﬁcation is achieved by decreasing the scanned area. The raster size also depends on the working distance, that is, the distance between the bottom of the ﬁnal objective lens and the sample. In modern SEMs the magniﬁcation is automatically compensated for each working distance to assure that the indicated magniﬁcation is correct [8]. When the electron beam impacts the specimen, many signals are generated, which could be eventually used for imaging purposes. Among them, secondary electrons

Electron gun Electron lens (1st Condenser) Spray aperture Scan coils Magnification control

Scan generator

Final lens aperture

Detector

Amp

Display CRT

Specimen To vacuum pumps

FIGURE 22.5 Schematic diagram of the electron column showing the electron gun, lenses, deﬂection system, and detector. (From Goldstein, J. et al., Scanning Electron Microscopy and X-Ray Microanalysis, 3rd ed., Springer, New York, 2003.)

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 502 16.7.2008 11:15pm Compositor Name: MSubramanian

502

Handbook of Food Analysis Instruments

(SE) and backscattered electrons (BSE) are the two signals more often used to produce SEM images. X-rays, which are also produced by the interaction of primary electrons with the sample, may be detected to get compositional information. Secondary electrons are sample electrons that are ejected from the surface after interaction with the incident beam, which are collected to form an image from the sample topography. The depth to which the primary beam penetrates the specimen and generates secondary emission is a function of the accelerating voltage and the density of the specimen. The brightness of the signal depends on the number of SE that escape and reach the detector. Steep surfaces and edges tend to be brighter than ﬂat surfaces, resulting in images with a three-dimensional appearance. Generated SE are captured by a scintillator–photomultiplier device, conveyed to an ampliﬁer, and then passed onto the screen of a cathode ray tube (CRT), where the resulting signal is rendered into a two-dimensional intensity distribution that images the exterior aspects of the sample. Modern SEMs can view and save images in a digital form. Interestingly, traditional equipment can be upgraded by integrating a digital scanning interface, allowing them to store images in a digital form, and consequently further processed. In addition to SE, BSE can also be collected. When the primary beam impinges the specimen, some electrons (10%–50%) are scattered back out of the sample by elastic collisions with the nuclei of sample atoms. The image formed from their collection is characteristic of the atomic weight of the elements encountered in that volume of the sample. Consequently, it is possible to detect contrast between elements and get chemical information of biological structures, when the average atomic weight of the various regions is different. This can be achieved by determining natural occurring differences or by incorporating speciﬁc heavy metals. Backscattered electrons can be detected by means of nondedicated or dedicated detectors. The standard Everhart–Thornley detector can be used to collect both SE and BSE. This detector is positioned on one side of the sample and collects both SE and BSE when a positive voltage is applied to the collector screen in front of the detector. When a negative voltage is applied, only BSE are captured since low energy SE are repelled. This detector has low collection efﬁciency for BSE due to small acceptance angles. Dedicated detectors are placed above the sample in an annular arrangement, allowing the primary beam to pass through the inner hole, greatly increasing the solid angle of collection and the number of BSE gathered. One of the great advances in SEM has been the development of the variable pressure scanning electron microscopes (VPSEMs) and of the environmental scanning electron microscope (ESEM). These new devices allow examination of wet or dry specimens of almost any nature, overcoming many of the limitations related to sample preparation, making it an ideal tool for biological samples [16]. In fact, a major limitation of conventional SEMs is the requirement for a high vacuum in the specimen chamber. Samples of interest for food scientists usually cannot withstand high vacuum levels or the rigors of preparatory drying without undergoing structural collapse, loosing structural deﬁnition. This means that many specimens of potential interest, such as biological tissues, liquids, gels, emulsions, etc., cannot be observed in their natural state. These two new devices overcome this problem, by allowing a higher pressure to be kept in the sample chamber. This is achieved, thanks to a differential pumping system that maintains the electron gun at high vacuum while the specimen is under a much higher pressure. The ESEM is a trademark of Philips=FEI=Electroscan. In this equipment, the specimen chamber can be at a pressure up to 20 Torr, and has a type of gaseous electron detector that enables SE emitted from the surface of irradiated samples to be collected via an ionizing gas cascade, which ampliﬁes the SE signal [16]. Possible gases to be employed include water vapor, nitrous oxide, carbon dioxide, nitrogen, and helium. Backscattered electron collection and x-ray analysis are also possible with appropriate detectors. Microscopes from other manufacturers are usually referred as VPSEMs. As in the case of ESEMs the specimen environment can contain water vapor or other gases, limiting the upper pressure to 2 Torr. New VPSEM can detect SE and BSE, and often x-rays. As with ESEMs it is possible to perform in situ real-time experiments, where the specimen interacts with the surrounding gaseous atmosphere.

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 503 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

503

22.3 INSTRUMENTATION 22.3.1 TRANSMISSION ELECTRON MICROSCOPES 22.3.1.1

Conventional Transmission Electron Microscopes

Conventional TEMs can operate at different accelerating voltages depending on the applications. Units can be equipped with thermionic emitters or ﬁeld emission sources, which can provide ultrahigh resolution (0.1 nm). Most common equipment operate in the range of 40–120 keV, achieving magniﬁcations up to 1,000,000, with approximately 0.2 nm lattice resolution. Another set of units can operate up to 200 keV, where magniﬁcations up to 1,500,000 can be achieved with around 0.1–0.15 nm lattice resolution. An additional set of units operate in the range 100–300 keV, where similar magniﬁcations can be attained with approximately 0.15–0.2 nm lattice resolution. Some new devices are equipped with an in-column energy ﬁlter (omega ﬁlter), which along with a Koehler illumination system provide excellent image contrast from thick or unstained thin samples. Overall, modern units are compact and include user friendly controls, ﬁle storage, automatic ﬁlament heating, and automatic exposure micrograph photography. Photography does not require a darkroom since they are equipped with LCD monitors and may include extensive functions for image processing and computer analysis. Some equipment include a side-entry goniometer stage, providing ease of use for tilt, rotation, heating and cooling, programmable multi-point settings, all without mechanical drift. With the addition of energy dispersive x-ray analysis (EDS) or electron energy loss spectrometry (EELS), the TEM can also be used as an elemental analysis tool, capable of identifying the elements in areas less than 0.5 mm in diameter. 22.3.1.2

Cryogenic Transmission Electron Microscopy

Biological samples can be considerably damaged during TEM observation. Electron-beam irradiation as well as the high vacuum to be maintained can both affect the specimen. A method for overcoming these difﬁculties, which allows observation at atomic resolution while keeping the sample in a hydrated state, is the ice embedding method described by Adrian et al. [17]. It consists of rapidly freezing the sample in a cryogenic liquid, embedding it directly in a thin ﬁlm of amorphous ice, allowing observation of unstained specimens in hydrated state, close to structure native state (e.g., protein). The frozen sample is then transferred to a cryo-TEM, where it is observed at liquid nitrogen temperature. The temperature can be raised to produce surface ice sublimation and reveal underlying structures. Two types of cryo-TEM have been developed. Dedicated cryo-TEMs are exclusively used for cryogenic observation and are not suited for other uses. In these units the specimen is cryotransferred to the device at liquid nitrogen temperature under vacuum. The specimen stage cools the specimen with super critical ﬂuid helium, temperature at which the observation is carried out. Another possibility is to use a conventional TEM equipped with a cryo-transfer holder, which can transfer the specimen at liquid nitrogen temperature into the TEM. This type of instrument becomes a cryogenic TEM only when cryo-observation is performed. Field emission electron guns best suit cryogenic observation. However, if ultimate resolution is not pursued, good results can be obtained with conventional thermal electron guns. LaB6 are preferred, since hairpin-type tungsten ﬁlaments are not bright enough [18,19]. 22.3.1.3

Scanning Transmission Electron Microscopy

As mentioned earlier, in 1938 von Ardenne [13] added scanning coils to a TEM, setting up the basis for the development of the scanning transmission electron microscope (STEM). As in TEM, incident electrons are transmitted through the specimen, but the point source of electrons scans the specimen, and the transmitted image is produced on a cathode ray tube, as in SEM. Recent advances in electron optics and computer interfacing have resulted in an instrument which is an

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 504 16.7.2008 11:15pm Compositor Name: MSubramanian

504

Handbook of Food Analysis Instruments

alternative to other forms of electron microscopy. At the present time, nondedicated and dedicated STEM equipment exist. In a nondedicated STEM thin enough samples are viewed in a conventional SEM with an attached transmitted electron detector placed below the specimen. The obtained image has a poor resolution compared to the one obtained with a dedicated STEM, where a small-diameter electron beam is scanned across the specimen. In both microscopes, an image can be formed from the secondary electrons generated at the surface. A major advantage of dedicated STEMs is that thicker (100–1000 nm) biological specimens can be observed without loss in resolution compared to the ultrathin sections used in a conventional TEM [1]. This avoids the inclusion of many of the artifacts produced during sample preparation for conventional TEM observation. A dedicated STEM operates in the 80 to 200 keV range, can achieve magniﬁcations up to 2,000,000X with approximately 0.2 nm lattice resolution. Since STEMs provide information on the internal structure of the specimen and the SEM technique provides topographical information, availability of both images can highly enrich the observation.

22.3.2 SCANNING ELECTRON MICROSCOPES 22.3.2.1

Conventional Scanning Electron Microscopes

Over the last decades, SEMs have become an indispensable tool in both advanced research and routine analysis for science and industry. Traditional devices are widely used. They operate under high vacuum conditions and require nonconductive specimen coating. New devices are simple to operate, since interfaces are intuitive, so that optimum setting conditions are easier to determine compared to old devices. They incorporate a 5-axes (x, y, z, tilt, and rotation) motorized specimen stage that facilitates manipulation through software (or mouse) control. Operating voltages can range from 0.1 to 30 keV, magniﬁcations can be up to 300,000, whereas resolution is about 3 nm. Average specimen chambers can accommodate specimens up to 15 cm in diameter, but, some units are meant to handle large specimens up to 17.5 cm tall and 30 cm diameter. New units are compact and include user friendly controls, ﬁle storage, automatic ﬁlament heating, and automatic exposure micrograph photography. Standard automated features include auto focus=auto stigmator, auto gun, and automatic contrast and brightness. As in TEM, they can provide extensive functions for image processing and computer analysis. Some units embed an energy dispersive x-ray (EDX) analyzer, or a wavelength dispersive spectrometer (WDS), which can be used to map chemical elements. 22.3.2.2

Low Vacuum Scanning Electron Microscopes

In this category are grouped both the VPSEM and the ESEM devices that can operate under different pressure ranges (high vacuum, variable pressure, and extended pressure), by incorporating air, water vapor, or other gases. Typically, but not exclusively, VPSEM is limited to a maximum pressure in the specimen chamber of about 2 Torr and may not offer use of the SE detector when there is gas in the chamber, so the analysis is only based on BSE detection. ESEMs can operate with a pressure up to 20 Torr and offer a secondary electron detector for use in the gas. These instruments usually use a thermionic gun, nevertheless, ﬁeld emission guns are now being incorporated. Devices operate in the 0.2–30 keV range, however, they are usually designed to operate optimally in the energy range 10–30 keV rather than at the lower energy of 1–5 keV favored by conventional high-vacuum machines. Besides this, VPSEMs and ESEMs and conventional SEMs are identical [20]. A major advantage of this new technology refers to the possibility to observe biological specimens that otherwise would not be adequately examined because of the loss of original state when dehydrated. Another main advantage arises from the fact that the gaseous environment permits poorly conducting materials to be steadily imaged at different accelerating voltages without the need to coat them with a conductive metal layer. It is known that poorly conducting specimens

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 505 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

505

require metal coating when imaged in SEM. This occurs because the incident beam current that impinges the sample exceeds the number of emitted electrons (SE and BSE), charging the sample negatively. This charge excess affects emitted electron collection, producing instability not only in the collected signal, but also in the primary impinging electron beam, distorting the image. Gas presence in the specimen chamber can counteract this effect, since impinging electrons can ionize the gas (e.g., by breaking up a molecule as in water vapor, H2O ! Hþ þ OH), producing both negative and positive charges, which can change the charge balance of the sample. If a reasonable number of ions is produced, the specimen charge can be neutralized by the ﬂood of cations. All equipment can provide this capability. As in conventional SEM, elemental mapping can be carried out by means of EDX or WDS accessories. Special mention has to be made to new developments in ﬁeld emission SEM. Field emission sources allow ultra high resolution imaging over a wide voltage range. For instance at accelerating voltages of 0.1 and 1 keV, resolution of approximately 4 nm and less than 2 nm, respectively, may be achieved. These units may also comprise variable pressure technology, allowing observation of hydrated nonconducting specimens. 22.3.2.3

Cryogenic Scanning Electron Microscopy

Cryogenic scanning electron microscopy (Cryo-SEM) is another technique that can be utilized to observe wet biological specimens. As in Cryo-TEM, water is immobilized in situ by lowering the temperature to a point where the vapor pressure is reduced and the escape of water vapor and other volatiles is negligible. In order to alter minimally the specimen microstructure, extremely high freezing rates must be achieved. The use of low temperatures to stabilize structures, termed cryogenic preparation, has several advantages most notably that delicate structures in high-moisture biological specimens may be well preserved and disruptive metabolic activities can be stopped. Also, since the sample is viewed directly after freezing, not only dehydration is not required but chemical ﬁxation is also unnecessary. Importantly, problems with low melting components, such as fats, are avoided. The equipment used is often known as low temperature scanning electron microscope (LTSEM), but standard SEMs can be adapted by ﬁtting a cryochamber, which allows specimens to be kept cold during preparation and loading into the imaging chamber. The cryochamber has a dedicated vacuum system, which is separated from that of the microscope. Cryogenic preparation involves quenching the sample in subcooled nitrogen at 2108C or in liquid propane and then transfer to the vacuum chamber, where it can be fractured, etched, and heated to remove surface ice. Subsequently, the sample still in the frozen state, may be coated to eliminate charging. The prepared sample is then directly transferred to the cold stage of the microscope by an exchange air lock. In freeze-fracturing, the sample is frozen and fractured along cleavage planes, usually membranes, to allow observation of internal facets. In foods, this technique can be used when biological membranes are present, but also to examine other structures such as emulsions (e.g., margarines and dressings), starchy products, and cheeses, among many others (Figure 22.6). Freeze-etching is a technique used to improve the image formed by cryopreparation. In this technique, the temperature is raised to produce surface ice sublimation and reveal underlying structures.

22.3.3 ANALYTICAL ELECTRON MICROSCOPY CAPABILITIES 22.3.3.1

X-Ray Microanalysis

X-rays, which are also produced by the interaction of electrons with the sample, may be detected in any electron microscope (SEM, TEM, STEM) equipped for energy-dispersive x-ray (EDX) spectroscopy or wavelength dispersive x-ray (WDS) spectroscopy. The most commonly used technique is EDX analysis, in which the energy levels of x-rays entering the detector are measured. With this system it is possible to simultaneously analyze all detectable elements, thanks to a simple computer

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 506 16.7.2008 11:15pm Compositor Name: MSubramanian

506

Handbook of Food Analysis Instruments

(A)

(B)

FIGURE 22.6 SEM micrograph of a (A) precooked and (B) cooked Chinese noodle, example of a food containing much water (mag. 1,500X at an accelerating voltage of 15 keV). (Courtesy of Jeol.)

interface system. Most systems are capable to measure elements with atomic numbers of 11 or higher. WDS measures the wavelengths of x-rays produced when the electron beam hits the sample. It has the advantage of being able to detect lighter elements, including boron. 22.3.3.2

Electron Energy Loss Spectrometry

In TEM, when the electron beam impinges the specimen, some inelastic interaction occurs between the electrons and the specimen, thus electrons suffer an energy loss, but virtually no change in direction. These electrons can be deﬂected into an electron spectrometer, which is usually ﬁtted below the column of the microscope. A map can be produced from the electrons that have suffered the energy loss or be interpreted in terms of the vibrational spectrum of the interacting species in the

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 507 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

507

sample, a technique known as electron energy loss spectrometry (EELS). STEM in combination with EELS may ﬁnd interesting applications in foods to map water distribution at the submicron scale. 22.3.3.3

Raman Scanning Electron Microscopy

X-ray microanalysis is a powerful technique, however, it only gives information about the elements present in a sample (and a mapping of the element distribution within the sample). Raman spectroscopy is a light scattering technique used to study molecular vibrations in a sample similar to infrared spectroscopy, which is specially suited for organic materials. An SEM unit can be ﬁtted with a structural chemical analyzer (SCA), which provides the interface between the SEM and the Raman spectrometer, allowing for the collection of a Raman spectrum, resulting in a powerful technique for studying the chemical and structural composition of materials in situ without the need to move the sample between instruments. Applications of laser Raman microscopy to foods, with a resolution of ca. 1 mm, are discussed by Celedón and Aguilera [21].

22.4 APPLICATIONS TO FOOD SCIENCE 22.4.1 TRANSMISSION ELECTRON MICROSCOPY

IN

FOOD SCIENCE

Advantages of the TEM in magniﬁcation of details must be counterbalanced against certain drawbacks, most importantly, introduction of structural artifacts during sample preparation. Researchers must be extremely conscious of this fact and particularly alert. In addition, extreme magniﬁcations may, in some cases, even become counterproductive because focus on microstructural organization can be distracted by ultrastructural details. This technique is therefore most applicable in studies of particular food macromolecules, such as proteins, and for tissue foods at the cellular lever (e.g., meat ﬁbers). As a matter of fact, TEMs have been successfully used to reveal quaternary structure of food proteins, complementing the insight provided by complementary techniques, such as amino acid analysis, circular dichroism, and x-ray crystallography, which give information about primary, secondary, and tertiary structures. As explained by Yada et al. [22], although x-ray crystallography has great resolution (0.3 nm), it is not adequate for many food-related proteins, since they do not form suitable crystals. The high resolution that can be achieved by TEMs coupled to modern preparation techniques and image analysis capabilities makes it the tool of choice. In fact, cryopreparation (utilized in cryo-TEM) allows fully hydrated sample observation, avoiding the need to remove water, a preparation step that may well disrupt quaternary structure. In addition, traditional negative staining may be avoided. Improvements in lattice resolution at low-voltage operation minimize protein beam damage. Advances in STEMs technology also minimize specimen beam damage, because of improvements in electron collection systems, and additionally allow a 3D image reconstruction. The reader is referred to the work carried out by Marcone et al. [23] and Yada et al. [22], to ﬁnd additional information related to quaternary-structure food protein characterization. Many applications in the food science ﬁeld related to protein characterization can be found in recent literature, mainly because of these advantages, but also because of the ease of use of new equipment. TEMs are being increasingly used to characterize aggregation of ﬁsh proteins [24,25,26] and casein micelles conformation [27,28,29]. Materials such as fats and fat-containing foods are rather onerous to characterize since their microstructure is quite temperature dependent and also because oil traces within the equipment have to be avoided. Many of these products are emulsions that present their own peculiarities because of the membrane structures and interfaces around emulsion droplets. One approach that helps to overcome these difﬁculties when preparing samples of food emulsions or suspensions for TEM is to use microencapsulation. This technique involves in mobilizing the material by combining it with liquid agarose, allowing it to solidify in tubes. Small pieces are then ﬁxed in

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 508 16.7.2008 11:15pm Compositor Name: MSubramanian

508

Handbook of Food Analysis Instruments

glutaraldehyde, postﬁxed in osmium tetroxide (a step that additionally partially ﬁxes the lipids), embedded in an epoxy resin and ultramicrotomed. Frequently, contrast is enhanced by staining with uranyl acetate and lead citrate. Another possibility is to use cryo-TEM as shown in Figure 22.7. Additional applications of TEM include characterization of samples such as meat emulsions [30], food gels [31,32], and certain microorganisms-containing systems [33,34].

22.4.2 SCANNING ELECTRON MICROSCOPY

IN

FOOD SCIENCE

SEM is widely used in food studies, perhaps even abused. This technique provides adequate magniﬁcations of thick specimens, allowing a 3D-like image. An important problem that must be overcome in food studies is that in almost all cases the material to be examined is moist. Water is not only present in bulk form as a dispersion medium for the various components, but also determines the speciﬁc structures of macromolecules such as proteins and membranes. Consequently, in conventional SEM samples must be adequately dried, except for very few products such as low moisture or powdered foods (ﬂour, sugar, milk powder, etc.). However, overall sample preparation is rather simple compared to the one required for TEM. Major advances in sample preparation (still the main drawback of this technique), as well as the development of new technology along with its ease of use, are some reasons that explain why this technique is available in almost any food science laboratory. The ease of use and wide availability and accessibility of SEM equipment may also be a major problem. SEM micrographs ‘‘decorate’’ research papers published in scientiﬁc journals. Foods are

FIGURE 22.7 TEM micrograph of a cryo-section of margarine, obtained with an accelerating voltage of 200 keV. (Courtesy of Jeol.)

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 509 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

509

largely heterogeneous at the microstructural level, hence, the problem of how representative is a micrograph not only of the observed specimen but also of the entire sample from which it was obtained, becomes a major issue. Low vacuum scanning electron microscopes (VPSEM and ESEM) allow the examination of many food samples in their natural state without the need of dehydration or metal coating, heavily broadening the range of food materials that can be studied [35] (Figure 22.8). Moist vegetable tissue, emulsions, food gels, dispersions, etc., can be imaged with minimal intrusion. Interestingly these new technologies allow food processing miniaturization. In fact, processes such as dehydration, hydration, freezing, freeze drying, crystallization, and melting can be viewed in real time by

FIGURE 22.8

VPSEM micrograph of a sesame seed (at an accelerating voltage of 15 keV). (Courtesy of Jeol.)

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 510 16.7.2008 11:15pm Compositor Name: MSubramanian

510

Handbook of Food Analysis Instruments

altering the chamber conditions. The dynamic mechanical behavior of dry or moist food materials can be studied by using a tensile stage [36]. In addition, x-ray analysis can also be possible with the appropriate detector.

22.5 FUTURE TRENDS The availability of electron microscopy hardware is driven by advances in biological and materials sciences, and recently by developments in nanotechnology. The food scientist must be alert to progress in equipment and applications coming from these areas. But it would be illusory to assume that the use of more powerful and sophisticated microscopes will lead directly to a better understanding of the role of food microstructure. Although the proverb says that ‘‘ . . . an image is worth a thousand words,’’ the human visual system is not well suited to make objective and quantitative determinations of features we see in images. In fact, what is needed is to understand the relationship between the microstructure and the desirable properties we expect to derive from foods. Modern electron microscope units are compact, easy to operate, include user friendly controls, allow recovery of ancillary information, and are equipped with digital imaging acquisition systems and processing software. The option to vary gas chamber conditions coupled to the possibility to include physical probes, in conjunction with analytical electron microscopy capabilities (x-ray microanalysis, Raman probes, etc.), opens a promising area of miniaturization of experiments, meaning direct intervention on a sample inside the microscope (e.g., shearing, drying or heating, etc.). For electron microscopy to become an engineering tool, quantitative data must be derived from microstructural information. Quantitative food microscopy will be a subject of much interest in years to come and already signiﬁcant advances are being made in algorithms and software for data processing for speciﬁc applications [37,38]. Eventually, robust models will be made that link speciﬁc microstructural features with properties that are sometimes elusive and difﬁcult to measure in foods (e.g., texture, nutrient bioavailability, ﬂavor release, etc.).

REFERENCES 1. Aguilera, J.M. and Stanley, D.W., Microstructural Principles of Food Processing and Engineering, 2nd ed., Aspen, Gaithersburg, 1999, chap. 1. 2. Aguilera, J.M., Food product engineering: Building the right structures, J. Sci. Food Agr., 86, 1147, 2006. 3. Bouchon, P. et al., Oil distribution in fried potatoes monitored by infrared microspectroscopy, J. Food Sci., 66, 918, 2001. 4. McMullan, D., A history of the scanning electron microscope, 1928–1965, Adv. Imag. Electron Phys., 133, 523, 2004. 5. Brooker, B.E., Food Technology International Europe 1988, Stearling, London, 1988, p. 289. 6. Kalab, M., Electron microscopy of foods, in Peleg, M. and Bagley, E.B. (Eds.), Physical Properties of Foods, AVI, Westport, 1983, p. 43. 7. Heertje, I. and Pâques, M., Advances in electron in electron microscopy, in Dickinson, E. (Ed.), New Physico-chemical Techniques for the Characterization of Complex Food Systems, Blackie Academic & Professional, Glasgow, 1995, chap. 1. 8. Goldstein, J. et al., Scanning Electron Microscopy and X-Ray Microanalysis, 3rd ed., Springer, New York, 2003, chap. 1. 9. Kessel, R.G. and Shih, C.Y., Scanning Electron Microscopy in Biology, Springer-Verlag, New York, 1974, chap. 2. 10. Wise, M.E. et al., Phase transitions of single salt particles studied using a transmission electron microscope with an environmental cell, Aerosol Sci. Tech., 39, 849, 2005. 11. Sharma, R., Design and applications of environmental cell transmission electron microscope for in situ observations of gas-solid reactions, Microsc. Microanal., 7, 494, 2001. 12. Knoll, M., Charging potential and secondary emission of bodies under electron irradiation, Z. Tech. Phys., 11, 467, 1935.

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 511 16.7.2008 11:15pm Compositor Name: MSubramanian

Scanning Electron and Transmission Electron Microscopies in Food Analysis

511

13. von Ardenne, M., The scanning electron microscope, Z. Tech. Phys., 19, 407, 1938. 14. Zworykin, V.K., Hiller, J., and Snyder, R.L., A scanning electron microscope, ASTM Bull., 117, 15, 1942. 15. Goldstein, J. et al., Scanning Electron Microscopy and X-Ray Microanalysis, 3rd ed., Springer, New York, 2003, chap. 2. 16. Danilatos, G.D., Introduction to the ESEM instrument, Microsc. Res. Tech., 25, 354, 1993. 17. Adrian, M. et al., Cryo-electron microscopy of viruses, Nature, 308, 32, 1984. 18. Chiu, W. et al., High resolution electron cryo-microscopy of macromolecular assemblies, Trends Cell Biol., 9, 154, 1999. 19. Pamela, A. et al., Mechanisms of scaffolding-directed virus assembly suggested by comparison of scaffolding-containing and scaffolding-lacking P22 procapsids, Biophys. J., 76, 3267, 1999. 20. Goldstein, J. et al., Scanning Electron Microscopy and X-Ray Microanalysis, 3rd ed., Springer, New York, 2003, chap. 5. 21. Celedón, A. and Aguilera, J.M., Applications of a Raman microprobe to study food microstructure, Food Sci. Technol. Intl., 8, 101, 2002. 22. Yada, R.Y. et al., Visions in the moist: The Zeitgeist of food protein imaging by electron microscopy, Trends Food Sci. Tech., 6, 265, 1995. 23. Marcone, et al., Quaternary structure and model for the oligomeric seed globulin from Amaranthus hypochondriacus K343, J. Agr. Food Chem., 42, 2675, 1994. 24. Hsu, K.C. et al., Changes in conformation and in sulfhydryl groups of actomyosin of tilapia (Orechromis niloticus) on hydrostatic pressure treatment, Food Chem., 103, 560, 2007. 25. Badii, F., Zhdan, P., and Howell, N.K., Elucidation of protein aggregation in frozen cod and haddock by transmission electron microscopy=immunocytochemistry, light microscopy and atomic force microscopy, J. Sci. Food Agr., 84, 1919, 2004. 26. Tironi, et al., Malonaldehyde-induced microstructural modiﬁcations in myoﬁbrillar proteins of sea salmon (Pseudopercis semifasciata), J. Food Sci., 69, C519, 2004. 27. Lencki, R.W., Evidence for ﬁbril-like structure in bovine casein micelles, J. Dairy Sci., 90, 75, 2007. 28. Portnaya, et al., Micellization of bovine beta-casein studied by isothermal titration microcalorimetry and cryogenic transmission electron microscopy, J. Agr. Food Chem., 54, 5555, 2006. 29. Karlsson, et al., Relationship between physical properties of casein micelles and rheology of skim milk concentrate, J. Dairy Sci., 88, 3784, 2005. 30. Barbut, S., Microstructure of white and dark turkey meat batters as affected by pH, Brit. Poultry Sci., 38, 175, 1997. 31. Lofgren, C. and Hermansson, A.H., Synergistic rheological behaviour of mixed HM=LM pectin gels, Food Hydrocolloid, 21, 480, 2007. 32. Boye, J.I. et al., Molecular and microstructural studies of thermal denaturation and gelation of betalactoglobulins A and B, J. Agr. Food Chem., 45, 1608, 1997. 33. Hajmeer, et al., Impact of sodium chloride on Escherichia coli O157:H7 and Staphylococcus aureus analysed using transmission electron microscopy, Food Microbiol., 23, 446, 2006. 34. Rasooli, I., Rezaei, M.B., and Allameh, A., Growth inhibition and morphological alterations of Aspergillus niger by essential oils from Thymus eriocalyx and Thymus x-porlock, Food Control, 17, 359, 2006. 35. Stokes, D.J., Thiel, B.L., and Donald, A.M., Direct observation of water-oil emulsion systems in the liquid state by environmental scanning electron microscopy, Langmuir, 14, 4402, 1998. 36. Thiel, B.L. and Donald, A.M., In situ mechanical testing of fully hydrated carrots (Daucus carota) in the environmental SEM, Ann. Bot-London, 82, 727, 1995. 37. Russ, J.C., Image Analysis of Food Microstructure, CRC Press, Boca Raton, 2004, chap. 3. 38. Aguilera, J.M. and Germain, J.C., Advances in image analysis for the study of food microstructure, in McClements, D.J. (Ed.), Understanding and Controlling the Microstructure of Complex Foods, Woodhead Publishing, Cambridge, 2007 (in press).

Otles/Handbook of Food Analysis Instruments 45660_C022 Final Proof page 512 16.7.2008 11:15pm Compositor Name: MSubramanian

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 513 15.7.2008 8:20am Compositor Name: BMani

Index A Absorption spectrophotometers, 234 Absorption=transmission spectrometry, 236 Accelerated solvent extraction, 98 Accuracy, 4 Acquisition mode, 210 ADC, see Analog-to-digital converter Adsorptive stripping voltammetry, 389 AdSV, see Adsorptive stripping voltammetry AED, see Atomic-emission detector AES, see Atomic emission spectroscopy Aﬂatoxins, 354 AFM method, surface topography, 448 Agarose gels, 425–426 (all-E)-Astaxanthin, effects of microwaves, 70 Allergens, 437 Amino acids, 163 analysis by CE, 413 Analog-to-digital converter, 297 Analtech semiautomatic TLC sample streaker, 146 Analyte–antibody interaction, 13 Anions, 162 ANN, see Artiﬁcial neural network Anodic stripping voltammetry, 389–390 shortcoming of, 390 for trace analysis, 389 Anthocyanins, monitoring by CE, 414 Antibody arrays, 449 Antioxidant extraction technique, 93–95 Antioxidants, separation using CE, 412 APCI, see Atmospheric pressure chemical ionization APMALDI, see Atmospheric pressure MALDI Aquaculture and MAP technique, 65–66 Arbitrary zero mark, 110 Aroma sensing, 372 Artiﬁcial neural network, 370 ASE, see Accelerated solvent extraction ASV, see Anodic stripping voltammetry Atmospheric pressure chemical ionization, 200, 204 Atmospheric pressure MALDI, 206 Atomic absorption spectroscopy atomization cell, type of cold vapor, 322 ﬂames, 321–322 graphite furnace, 322 hydride generation, 322 background correction methods, type of atomic spectral line, splitting, 323 continuum source, 323 Smith–Hieftje, 323 Czerny-Turner monochromator, 322 ﬂame atomization interferences chemical, ionization, physical, and spectral, 324 graphite furnace interferences, 324–325 HCL, 321

main components of, 320 radiation, absorption of, 321 wavelength selection and detection, 322–323 Atomic-emission detector, 136–137 Atomic emission spectroscopy charge transfer devices, 328 inductively coupled plasma (see Inductively coupled plasma-atomic emission spectroscopy) sample introduction cold vapor generation, 327 hydride generation, principles of, 327 nebulizers, 325–326 spray chambers and desolvation systems, 326 spectral interferences, 328 spectrometer sequential, 327 Atomic spectroscopic techniques, parameters of analytical working range, 329 detection limit, for elements, 328–331 performance, comparison of, 332–333 purchase and operating costs, 331 sample throughput, 329 selection criteria, 331 Autoinjectors, 169 Automated electrophoresis system, 433–434 Automated microarray system, 449 Automated solid-phase extraction approach, 123 Automatic sampling, 5 Automatic staining apparatus, 435

B Backscattered electrons, 502 Band-broadening effects, 170 Bead-based immunoassays antibodies immobilization, 450–451 bacterial and viral proteins detection, 451 Beer–Lambert law, 170–171, 321 Bending test, solid food materials, 463–464 Betaxanthins, 236 Bingham ﬂuids, 467 BioArena bioautography system, 151 Biosensor, 395 Bisphenol A diglycidyl ether (BADGE), analysis of, 225 Bisphenol F diglycidyl ether (BFDGE), analysis of, 225 Bligh and Dyer method, for extraction of lipids from ﬁsh, 65–66 Bloembergen–Purcell–Pound model, 285 Blotting, 427 Blotting systems, 434–435 Blotting tanks, see Tank blotters BPP, see Bloembergen-Purcell-Pound model Bremen processes, 32 BSE, see Backscattered electrons Buffers, 426–427

513

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 514 15.7.2008 8:20am Compositor Name: BMani

514 Bulk acoustic wave device, 368 Burgener nebulizers, 326 Butler–Volmer equation, 381–382

C Calibration curve, 2 Calibration technique, 2 Camag DigiStore 2 documentation system, 154 Camag Linomat 5 sample streaker, 146–148 Camag TLC Scanner 3, 153–154 Camag twin-trough chamber, 148 Capillary-based immunoassays, 442–443 Capillary electrochromatography, 404 Capillary electrophoresis, 403 basic instrumentation in components of, 407 continuous ﬂow systems, 411 detectors, 410–411 high voltage power supplies, 409 sample introduction, 409–410 separation capillaries, 408–409 food analysis by additives, 412 amino acids, 413 carbohydrates and vitamins, 416 chiral compounds, 417–419 DNA, as analytical target, 415 organic contaminants, 417 peptides and proteins, 413–414 phenolic compounds, 414–415 small organic and inorganic ions, 416–417 microchips, 419 miniaturization of, 411 separation principle and separation modes, 404–405 theoretical plate number and resolution in, 406 two-dimensional, 411 Capillary ﬂow development, in CPLC, 148–150 Capillary GC column, 126 Capillary gel electrophoresis, 404 Capillary isoelectric focusing, 404 Capillary isotachophoresis, 404 Capillary zone electrophoresis, 239, 404–406 electroosmotic velocity, 406 ionizable compound separation, 405 Capsaicinoids, extraction of, 68 Carbamates, 393 Carbohydrate analysis, by CE, 416 Carbon-based solid phase extraction tubes, 12 Carbon dioxide, 33 commercial applications, 39–45 density of, 35 phase diagram, 29–30 physical properties of, 35 solubility of naphthalene in supercritical, 30–31 Carbon–hydrogen–nitrogen standard, 2 Carbonyl iron powder, 66 Carbowax (CW)-DVB, 15 Carr–Purcell–Meiboom–Gill, 292 Casson model, 471 Cathodic stripping voltammetry, 390 Cations, 162

Handbook of Food Analysis Instruments CCP, see Colloidal calcium phosphate CE, see Capillary electrophoresis CEC, see Capillary electrochromatography CFS, see Continuous ﬂow systems CGE, see Capillary gel electrophoresis Chelating moieties, 42 Chemical ionization, 202–203 Chiral compounds, 417 Chromatographic dispersion effects, 105 ChromXtractor, 155 CI, see Chemical ionization CIEF, see Capillary isoelectric focusing CIELAB system, 237 CIE system, 236–237 CIP, see Carbonyl iron powder CITP, see Capillary isotachophoresis Classical calibration, 2, 250, 257 Clavenger technique, 66 Closed-vessel applications, 59–60 CMC color tolerancing scale, 237 COC, see Cold on-column Cold on-column, 120, 122 Cold on-column injection, 122–123 Colloidal calcium phosphate, 302 Colorants, 412 Color imaging analysis, 238 Colorimetric evaluations, 235–237 Colorimetry, 231, 233 Colour and light biology of, 233 chemistry of, 232–233 physics of, 230–231 Colour measurement, 233–239 color description systems and notions, 236–237 as food quality indicators, 239–242 imaging analysis, 238 reference methods, 238–239 Columns, of IC, 169 Compound microscope, 496 Comprehensive two-dimensional gas chromatography, 130 Computer-controlled pumping, 168 Computer vision systems, 238 Concentration of sample, method of developing, 8 Conducting polymer sensors, 367 Conductivity detectors, 172–173 Constant ﬂow pumps, 167 Constant pressure pumps, 167 Continuous-ﬂow systems, 88, 411 Convenience, 5 Cooling device, 431 Corticosteroids, 14 Cosolvents, 27, 32 Coulombic interactions, 163 and temperature, 166 Coulometry, 177 Counter electrodes, 384 Counterion concentration, 166 Coupled supercritical ﬂuid extraction–supercritical ﬂuid chromatography, 27 CPMG, see Carr-Purcell-Meiboom-Gill Creep recovery, 482–484 Critical point, 26

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 515 15.7.2008 8:20am Compositor Name: BMani

515

Index Critical pressure (Pc), 26 Critical temperature (Tc), 26 Crop cultivar identiﬁcation, on protein level, 437 Crosslinking factor, 428 Cross-reactivity, in compounds, 110 Cryogenic scanning electron microscopy, 505 Cryogenic transmission electron microscopy, 503 C18 silica, 11 CSV, see Cathodic stripping voltammetry CVS, see Computer vision systems Cyclic voltammetry, 385–386 peak potential separation, 385–386 principles of, 385 CycloGraph with RAVE, 152 CZE, see Capillary zone electrophoresis

D Dairy products, shelf life determination of, 371 DAPI, see Desorption Atmospheric Pressure Chemical Ionization DART, see Direct analysis in real time Data acquisition system, 366 Data analysis measurement techniques for, 2–3 technique selection, criteria for accuracy, 3 convenience, 5 detection limit, 4 precision, 4 quantiﬁcation limit, 4 range, 4 robustness, 4 selectivity, 3 speed, 5 DC amperometry, 177–178 2D-CE, see Two-dimensional CE Deﬂection system, 501 Densitometers, 234 Densitometry, 155 Derivatization, 17–18, 153 Desaga AS-30 semiautomated gas-stream spray-on applicator, 148 Desorption Atmospheric Pressure Chemical Ionization, 206 Detectability (sensitivity), of instrument, 210–211 Detection limit, 4, 328–329, 389, 395, 449 Detection systems, ﬂuorescence-based microchips, 447–448 Detectors, of IC absorbance-based detectors, 170–172 electrochemical amperometric=coulometric detection, 177–178 conductivity detectors, 173–177 potentiometric detectors, 173 pulsed amperometric detection, 180–182 pulsed electrochemical detection, 178–180 mass spectrometric detection, 172 refractive index (RI) detectors, 170 Diagnostic ultrasound, 86 Differential pulse voltammetry, 387 Difﬁcult matrix introduction, 125

Diode-array (DA) detection, 249 soybean seeds, coat color, 255–256 Direct analysis in real time, 200, 206–207 Direct-MS, ion sources for, 205–207 Direct sample introduction, 125 Dispersions, in supercritical ﬂuids, 32 Distribution=partition coefﬁcient (KD), 8 Divinylbenzene (DVB)-PDMS, 15 DMI, see Difﬁcult matrix introduction DNA, 425 capillary electrophoresis of, 415 gel electrophoresis of, 436–437 DNA-directed immobilization, 444–445 DPV, see Differential pulse voltammetry Dry spaghetti, tensile testing, 465–466 DSI, see Direct sample introduction Dual-analyte immunoassays, 440 Dual-phase stir bars, 18 Dyes gel-electrophoretic analysis of, 425 in immunoassays, 440 Dynamic techniques, 60 Dzido-Soczewinski (DS)-II chamber, 148–150

E ECD, see Electron capture detector Echelle spectrometer, 327 Edible oil blends, quality assessment of, 373 Efﬁciency, of any extraction, 9 EKC, see Electrokinetic chromatography ELCD, see Electrolytic conductivity detector Electroactive species, 380 Electrochemical cells electrodes, 380 as sample holder, 384 schematic representation of, 381 Electrochemical detectors applications, 396 detection limits, 395 Electrochemical sensors, 368 Electrochemical techniques, in food analysis, 379–380 Electrode assays, 451–452 Electrode criteria for classifying, 384 Electrokinetic chromatography, 404, 407 pseudostationary phases, 407 separation principle in, 406 Electrokinetic phenomena, 404 Electrokinetic sample injection, 409 Electrolytic conductivity detector, 136 Electron capture detector, 133, 136 Electron energy loss spectrometry, 506–507 Electroneutrality, 163 Electronic nose commercial available, 369 data analysis methods for pattern recognition techniques, 369 PCA, LDA, and ANN, 370 food analysis using alcoholic drinks, 372 fruits and vegetables, 372 hazelnuts, 373

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 516 15.7.2008 8:20am Compositor Name: BMani

516 meat, poultry, and seafood products, 371–372 milk and dairy products, 370–371 olive oils and edible oils, 373 gas sensors used in bulk acoustic wave device, 368 CP sensors, 367 electrochemical sensors, 368 GC=MS-based, 368 MOS sensors, 367 and samples, 366 SAW sensors, 368 sampling system, 366 Electronic nose technology advantages, 365 Electron impact (EI) process, 201 Electron spin resonance, 356 Electroosmosis, 404 Electrophoresis, 404 basic principle of, 424 Electrophoretic microchips, 411 Electrospray ionization, 200–204 ELISA-type platforms, 446 Ellipsometry, 448 Eluent concentrations, 168 Enantiomeric analysis, in foods importance of, 417 using CE method, 417–418 E-nose, see Electronic nose Environmental applications, of SFE, 42 Environmental scanning electron microscope, 502 EOF, see Electroosmotic ﬂow EPA Pollution Prevention Act, in USA, 34 Errors, in sample preparation, 2 ESEM, see Environmental scanning electron microscope ESI, see Electrospray ionization ESR, see Electron spin resonance Essential oils, extraction of, 68 Ethanol, 27 Ethylenediaminetetraacetic acid (EDTA) standard, 2 Evanescent wave, excitation in planar waveguide, 447 Extensibility tests, meat spaghetti sample, 465 Extraction techniques centrifugation of samples before extraction, 9 by direct immersion of ﬁber into liquid samples, 16 extracted ion chromatogram, 18 ﬂavor compounds, from rancid corn oil, 16 headspace single-drop microextraction, 10 liquid-phase microtechnique, 9–10 solid-phase, 10–14 solid-phase microtechnique, 14–17 stir bar sorptive, 17–18 theory, 8–9 times of, 58 Extraction, with supercritical ﬂuids, 33–36 of soluble species (solutes) from solid matrices, 36 thermodynamic parameters, 36

F FAAS, see Flame atomic-absorption spectrometer FAIMS, see High-ﬁeld asymmetric waveform ion mobility spectrometry Faradaic current, 382

Handbook of Food Analysis Instruments Fast gas chromatography, 126 Fast temperature programming, 128 Fatty acids, 17 FES, see Flame emission spectroscopy FID, see Free induction decay; Flame ionization detector Field cycling relaxometry, 285 Filtration-based microchip, 446 Flame atomic-absorption spectrometer, 88 Flame emission spectroscopy, 325 Flame ionization detector, 133, 135 Flame photometric detector, 136 Flame photometry, 172, 325 Flavonoids extraction, from Acanthopanax senticosus leaves, 68 Flavor analysis, 17 Flavor extraction technique, 91–92 Florisil, 11 Flowing eluent wall (FEW) technique, 151 Fluorescence deﬁnition of, 348–349 factors affecting concentration, 350–351 inner-ﬁlter, 350 light scatter, 352 molecular environments, 351–352 quenching, 350 ﬁsh oils, 359 in food, 231, 354–355 imaging degree of photooxidation, cheese, 360 lipid oxidation, 361 turkey burger, 361 intensity inner-ﬁlter effects, 350 molecular environments, 351–352 of tryptophan, 351 landscapes, oils, 357 quenching, 350 Rayleigh and Raman scatter, 352 spectrometer ﬂuorescence data, analysis of, 354 optical components and sensors, 353 sampling geometry, 353 spectrometry, 236 spectroscopy, 347 concentration, 350 PAT, 360 three-stage process, 348 Fluorescence emission spectra, 351 Fluorescence spectrophotometer, 353–354 Fluorescent dyes, 437 in immunoassays, 440 Fluoroquinolones, 14 Focused microwave-assisted Soxhlet extraction (FMASE), 65–66, 70 Folic acid, 416 Folin–Ciocalteu reaction, 272 Food additives, 412 Food analysis; see also Capillary electrophoresis; Electronic nose; Voltammetric techniques carbohydrate using CE, 416 colour measurement, 233–242 as food quality indicators, 239–242

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 517 15.7.2008 8:20am Compositor Name: BMani

517

Index DNA as analytical target, 415 electrochemical methods in biosensors, 395 colorants and ﬂavors, 390–391 electrochemical sensors, 396 hyphenated techniques, 395–396 metal contaminants, 390, 392 pesticides and herbicides, 391–395 using atomic spectroscopy alcoholic drinks, 336–337 As, Pb, Cd, and Hg, level of, 334–335, 343 cereal, ﬂour, rice, and legumes, 339–340 Fe in milk samples, 342 fruit and vegetable, 340–341 meat and meat products, 341 metal content, 34 milk, infant formula, and diary products, 337–339 seafood, 341–342 soft drinks, 336 using CE enantiomeric analysis of chiral compounds, 417–419 soybean proteins, 413–414 whey proteins, 413 using ﬂuorescence emission adulteration detecting, 357–358 ﬂuorophore concentration from, 355–356 lipid oxidation, 356–357 PAT application, 358 of sugar samples, 359 using gel electrophoresis nucleic acids, 425, 436–437 proteins, 425, 437 using SEM dry or moist food materials, 509–510 sample preparation, 508 using TEM food proteins, quaternary structure, 507 sample characterization, 508 using voltammetric techniques advantages of, 380 cyclic voltammetry, 385–386 pulse voltammetric techniques, 386–388 Food applications; see also Microwave-assisted processes of MAP, 69–78 of microextraction techniques, 19 with PLC, 155–156 of SFE, 42–43 with UAE, 91–96 Food and Agriculture Organization (FAO), 336 Food colorants, electrochemical determination of, 390–391 Food contaminants analysis, 110 Food decay processes, 229 Food ﬂuorophores, 349, 355 Food materials categorization of, 461 deformation, 462 mayonnaise, rheological behavior of, 476–477 solid behavior, 462 tests for, 462–466 tests for transient viscoelasticity normal-stress coefﬁcients, 473, 475 using Leider and Bird equation, 474, 476

tests for viscoelasticity creep, 482–484 linear range of, 478 nonlinear, 485–490 nonlinear compression test, 478–479 oscillatory measurements, 484–485 stress relaxation test, 479–481 viscous behavior measurement techniques for, 468–473 non-Newtonian ﬂuids, 467 shear ﬂow, 466–467 Food quality, 365 Food science applications 13 C-NMR spectroscopy, 300–301 1D 1H-NMR spectroscopy, 297 extracts and solutions, chemical composition of, 297–299 pH measurements, muscle tissue, 303–304 31 P-NMR spectroscopy, 301–302 solid–liquid transition, 294 STE-DOSY spectra, 311 TD-NMR proton relaxation time, 290–291 water holding capacity, meat and cheese, 292–293 Foodstuffs, Pb, Cd, and Hg, levels of, 334–335 Food technology, 495 Four ﬂuorescent compounds, sugar, 357 Fourier transform, 248 Fourier transform infrared spectroscopy, 4 Fourier transform ion cyclotron resonance (FT-ICR), 201, 217, 219 FP, see Flame photometry FPD, see Flame photometric detector Fragrance extraction technique, 91–92 Free induction decay, 283 Fruits quality assessment, of orange juices, 372 FT, see Fourier transform FT-ICR analyzer, 201, 217–218 FT-NIR microspectrometers, 266 FT-NIR spectrometer models, 250 Full width at half maximum (FWHM), deﬁnition, 210 Furosine detection, 67 Fused silica capillaries, 408–409

G Gas chromatography advantages, 133–134 characterization of stationary phases used in, 127 detectors applicable for food components determination, 135 food analysis applications, 140–142 ion sources for, 201–203 matrix effects, 137–140 peak comparison of thiabendazole and procymidone, 130 quantiﬁcation strategies and assessment, 139 sample detection, 134–137 sample introduction strategy cold on-column injection, 122–123 direct sample introduction=difﬁcult matrix introduction, 125

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 518 15.7.2008 8:20am Compositor Name: BMani

518 programmable temperature vaporization (PTV) injection, 123–124 solid-phase microextraction (SPME), 125–126 split=splitless injection, 120–122 sample separation, 126–134 solvents used in, 122 Gas chromatography-mass spectrometry, 60, 88 Gas-phase extraction techniques, 60–61 GC, see Gas chromatography GC–FID chromatograms, of fatty acid methyl esters, 129 GCGC system, 131 advantages, 133–134 optimization of operation conditions and instrumental requirements in, 131–133 GC-MS, see Gas chromatography-mass spectrometry GC=MS-based sensor system blue cheeses ﬂavor, 371 operation mechanism, 368 Gel electrophoresis analytes, 425 blotting methods, 427 buffer, 426–427 gels, 425–426 instrumentation automatic staining apparatus, 435 blotting systems, 434–435 high-resolution two-dimensional electrophoresis, 434 horizontal ﬂatbed electrophoresis systems, 433–434 imagers, 435–436 power supply, 431 vertical electrophoresis systems, 431–432 operating mechanism of, 424 relative quantiﬁcation, 428 staining method, protein detection, 427 theory of isoelectric focusing, 430 two-dimensional electrophoresis, 430 zone electrophoresis techniques, 428–430 Gel rod apparatus, 431 Gels agarose, 425–426 compositions of, 428 polyacrylamide, 425 Genetically modiﬁed material gel electrophoresis, 437 Ginseng saponins, extraction of, 67 Ginsenosides extraction, 67–68 Gold-coating, 444 Gradient elution, 168 Graphite furnace, 322 Green colorants, absorption rate, 233

H 1 H anomeric region, cross-peak, 309 HCL, see Hollow cathode lamp Headspace, of a sample, 10 Headspace sampling, 16 Herschel–Bulkley model, 470–471 Heteronuclear multiple bond coherence, 288–289 Heteronuclear multiple quantum coherence, 288–289, 307

Handbook of Food Analysis Instruments Heteronuclear single quantum coherence, 288–289 HF-LPME, see Hollow ﬁber protected liquid-phase microextraction High-ﬁeld asymmetric waveform ion mobility spectrometry, 220 High performance layer chromatography, 145–146, 238–239, 276, 301, 326 High performance liquid chromatography, 105, 145, 276, 301, 326, 395 High performance thin layer chromatography, 145–146 High-pressure mixing, 168 Hildebrand solubility parameter, to density, 37 HMBC, see Heteronuclear multiple bond coherence HMQC, see Heteronuclear multiple quantum coherence 1 H-NMR spectra grape juice, 296 meat extracts from turkey thigh, 300 Hollow cathode lamp, 320–321 Hollow ﬁber protected liquid-phase microextraction, 9–10 Hookean solid, 462 Horizontal ﬂatbed electrophoresis systems, 433–434 Hot spot theory, 87 HPLC, see High performance liquid chromatography HPLC-MS-MS stable isotope method, 107 HPTLC, see High performance thin layer chromatography HR-NMR techniques CP-MAS and HR-MAS, solid-state NMR spectroscopy cheeses, 304 pH measurements, muscle tissue, 303 theory of, 302–303 heteronuclear 2D spectroscopy HSQC, HMQC, HMBC, and INADEQUATE spectra, pulse sequence of, 307–309 milk, 309–310 heteronuclear spectroscopy 13 C-NMR spectroscopy, vegetable oils, 300–301 natural isotopic fractionation, 302 31 P-NMR Spectroscopy, 301–302 pulse sequences and instrumental requirements, 300 one-dimensional 1H-NMR spectroscopy, 288, 294–296 extracts and solutions, chemical composition of, 297–299 PFG-NMR spectroscopy food matrices, diffusion in, 311 pulse sequence, 310–311 quantitative carbohydrate analysis, 255 solution-state NMR spectroscopy baseline distortion, 296 FID acquisition, 296 foodstuffs, paramagnetic species in, 299–300 phase errors, effect of, 297 tuning and shimming, 295–296 two-dimensional 1H-NMR spectroscopy COSY, NOESY, TOCSY, and ROESY spectra, theory of, 304–305 homonuclear 2D NMR techniques, tomato juice, 306–307 HSQC, see Heteronuclear single quantum coherence HS sampling technique, 60 Hunter Lab system, 237 Hybrid instrument analyzers, 218–219 Hydride generation, 322 Hydrodynamically controlled system, 177

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 519 15.7.2008 8:20am Compositor Name: BMani

519

Index Hydrodynamic injection, 409 Hydrogels, 444 Hydrogen bonding, 131 Hydrolysis, 66, 69 Hydrophobicity, 8

I ICP, see Inductively coupled plasma ICP-AES, see Inductively coupled plasma-atomic emission spectroscopy ICP spectroscopy, multielement analysis, 334 IC pumps, 167 Ideal gas law, 32 Imagers, electrophoresis gels, 435–436 Imaging mass spectrometry, 220–221 Immunoafﬁnity columns, 110 Immunoafﬁnity solid-phase extraction, 110 Immunoassays, 439; see also Multiplexed immunoassays Immunochromatographic dipstick, 442 Immunoelectrophoresis, 429–430 antibody migration, 429–430 buffer tanks, 433 principle, 424 Immunosorbents, 13–14 IMS, see Imaging mass spectrometry INADEQUATE spectrum, 308 Inductively coupled plasma, 320 Inductively coupled plasma-atomic emission spectroscopy, 325 analytical working range, 329 legumes, analysis of, 339 performance, comparison of, 332–333 simultaneous multielement analysis, 327, 329 with ultrasonic nebulization (USN), 337 Inductively coupled plasma (ICP)-MS, ion sources for, 205 Injection techniques, 120–124 Ink-jet printers, 444 Interchannel cross-talk, 109 Internal standard, 3 Internal surface reversed phase, 13 In-tube SPME method, 17 Inverse calibration technique, 2 Ion chromatography food applications amines and organic bases, 189 carbohydrates and oligosaccharides, 189–193 concentration of ions, 185–188 detection of organic acids, 188 future prospects, 193–194 instrumentation columns, 169 connective tubing and ﬁttings, 183 data collection and output device, 182–183 detectors, 170–182 mobile phase reservoir, 167–168 online reagent generation, 168 post-column apparatus, 169–170 post-detection eluent processing, 183 related separation techniques, 183–184 sample introduction device, 168–169 solvent delivery system, 167–168

theory ion-exchange phase, 163–165 mobile phase vs. stationary phase, 165–166 separation principles, 163 Ion-exchange process, 164 Ion-exchange resins, 164 Ionic compound, 162 Ionizability (pKa), 8 Ionization energy, 201 Ions, 162 Ion trap mass spectrometers, 106 Isoelectric focusing, power supply for, 431 Isoelectric point, 428 Isotherms, of gas, 28–29 Isotopic peaks, 198–199 ISRP, see Internal surface reversed phase

J Jablonski diagram, 348

K Kelvin nanoprobe, 448 Kinetics-driven technology, 61 Kjeldahl method, 89

L Label-free detection, 447 Lambert–Beer law, 249 Lanthanides, 440 Laser scanner, 436 Lateral-ﬂow assay, see Immunochromatographic dipstick L*C*H* polar coordinate, 237 LDA, see Linear discriminant analysis Light microscopy, 496 Light scatter Rayleigh scatter and inelastic Raman scatter, 352 Linear discriminant analysis, 337, 369–370 Linear dynamic range, 211 Linear quadrupole ion trap (2D) mass analyzer, 213 Liquid chromatography, 163 ion sources for, 203–205 Liquid chromatography=ultraviolet (LC=UV) chromatogram, of acrylamide, 12 Liquid-phase extraction techniques, 58–60 Liquid-phase microextraction, 8–10 LM, see Light microscopy Lorentzian water signal FID, Fourier transformation, 248 Low-pressure mixing, 168 Low resolution 1H-NMR longitudinal relaxation time, measurement of IR and SR proﬁles, 289 pulse sequence, 287–290 TD-NMR, 290–291 transverse relaxation time, measurement of pulse sequence, 292 solid–liquid transition, 294 water holding capacity in meat and cheese, 292–293

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 520 15.7.2008 8:20am Compositor Name: BMani

520 Low vacuum scanning electron microscopes, 504–505 LPME, see Liquid-phase microextraction Luminescence, 349; see also Fluorescence

M MAD, see Microwave-accelerated distillation MAE, see Microwave-assisted extraction Magnetic resonance imaging, 286 Magnetic sector mass analyzer, 215–217 Maillard reaction products, 69 MALDI, see Matrix-assisted laser desorption ionization MAP, see Microwave-assisted processes MASD, see Microwave-accelerated steam distillation method Mass accuracy, 210 Mass analyzers detectors, 219–220 in food applications acrylamide precursors, 224–225 brominated ﬂame retardants (BFRs), 223–224 chloropropanols, 225 detection of natural substances, 221–222 epoxy compounds, 225 mutagenic and carcinogenic substances, 225 mycotoxins, 224 packaging migrants, 225 pesticides, 222–223 phthalate and adipate esters, 225 polychlorinated biphenyls (PCBs), 223 polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDDs=PCDFs), 223 polycyclic aromatic hydrocarbons (PAHs), 224 process of contaminants, 224–225 vetenary drug residues, 224 fourier transform ion cyclotron resonance (FT-ICR), 217–218 general speciﬁcations and features, 207–211 hybrid instruments, 218–219 linear quadrupole ion trap (2D), 213 magnetic sector, 215–217 orbitrap analyzer, 218 quadrupole, 211–212 quadrupole ion trap (3D), 213 time-of-ﬂight (TOF), 213–215 Mass range, 207 Mass resolution=mass resolving power, 207–210 Mass spectrometer (MS) detector, 137 Mass spectrometric detection, 106–109 Mass spectrometry, 3, 239 Mass spectrum, 198–199 Matrix-assisted laser desorption ionization, 200, 205–206, 221 Matrix-assisted laser desorption=ionization coupled to time-of-ﬂight (MALDI-TOF) mass spectrometer, 448 Matrix effects, in GC analysis, 137–140 Matrix-induced chromatographic response enhancement, 137 Maximal spectral acquisition speed, 210 Mayonnaise, rheological behavior of, 476–478 Membrane assays, 442

Handbook of Food Analysis Instruments Membrane-coated surfaces, 444 Mercury, as working electrode, 384 Mercury electrode, 389 Metal contaminants, in food products voltammetric determination of, 390, 392 Metal extraction, from food samples, 92–94 Metallic electrodes, 384 Metal oxide semiconducting ﬁeld effect transistors, 367 Metal oxide semiconductor sensors, for wines classiﬁcation, 372 Microchip assays assay platforms, 446 biomolecule patterning methods, 444–445 detection systems label detection, 447 label-free detection, 447–448 examples of, 448–449 surface materials glass slides, 443–444 PVDF, 443 Microchip electrophoresis, 411 Microcontact printing, 444 Microﬂuidic network, 446 and patterning, 445 Microtiter plate assays, multiplexed immunoassays using, 449–450 Microwave-accelerated distillation technique, 66 Microwave-accelerated steam distillation method, 68 Microwave-assisted extraction, 57–61, 65, 96, 98–99 food application, 71–77 Microwave-assisted processes, 57 application to food analysis bakery products, 69–70 cereals and oilseeds, 67 dairy and egg products, 66 food ingredients, 68–69 food safety, 70–78 fruits and vegetables, 66–67 herbs and spices, 67–68 meat, poultry and ﬁsh, 65–66 instrumentation, 61–64 theoretical considerations, 58–61 Microwave-assisted Soxhlet extraction, 65–67, 70 Microwave distillation-SPME, 65 Microwave hydrodistillation, 68 Microwaves, 58 Milk, shelf life determination of, 370–371 MIP, see Moleculary imprinted polymers MIST, see Multiple spotting technology MLR, see Multiple linear regression Mobile phase reservoir, of IC, 167–168 Modulation criterion, 132 Moisture determination Karl Fischer titration method, 275 oven drying methods, 275 Molecular ion, 198, 201 Molecular weight, 8 Moleculary imprinted polymers, 14, 110–111 Monolithic columns and multiwalled carbon nanotubes, 12 MOSFETs, see Metal oxide semiconducting ﬁeld effect transistors MOS sensors, see Metal oxide semiconductor sensors

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 521 15.7.2008 8:20am Compositor Name: BMani

521

Index MRI, see Magnetic resonance imaging MRM, see Multiple reaction monitoring MS-based system, 368 MSC, see Multiple scattering correction MS detectors, 201 Multimode cavities, 61 Multiple-label immunoassay, 440–441 Multiple linear regression, 252 Multiple reaction monitoring, 211 Multiple scattering correction, 252–253 Multiple spotting technology, 446 Multiplexed immunoassays advantages, 439 challenges for antibody resources, 452 labeling of antibodies, 452–453 classiﬁcation, 440 enzymes used in, 441 ﬂuorescent dyes, 440 Lanthanides and radioisotopes, 440 metals and nanocrystals used for, 441 spatially resolved (see Spatially resolved multiplexed immunoassay) Multiresidue analysis, of foods and drinking water, 106 Multivariate data analysis techniques, 370 MWCNTs, see Monolithic columns and multiwalled carbon nanotubes Mycotoxins, analysis of, 13–14

N NACE, see Nonaqueous capillary electrophoresis Nanocrystals, for immunoassays, 441 Nanowire assays, 452 Naphthalene solubility, in SCCO2, 31–32 National Institute of Standards and Technology, 2 Near-infrared spectroscopy, 248 single-seed analysis, soybean, 257 Nebulizer, 325–326 Nernst distribution law, 8 Neurotransmitters, 177 Neutral loss scan, 211 NIR instruments basic parts, 285 block scheme of, 286 calibration regression methods, 252 steps of, 251 grain-tester, 250 principles of, 248–249 reﬂectance and transmission, 249 spectroscopy, 250 NIRS, see Near-infrared spectroscopy NIST, see National Institute of Standards and Technology Nitrogen-phosphorus detector, 136 NMR, see Nuclear magnetic resonance NMR-MOUSE device, 289 NMR spectroscopy techniques food carbohydrates, sugars and ﬁbers, 255 protein determination, 253–254 N-octanol=water partition coefﬁcient Kow, 8 NOESY, see Nuclear Overhauser effect spectroscopy

Nonaqueous capillary electrophoresis, 405 Nonfaradaic currents, 383, 386 Non-Newtonian ﬂuids classiﬁcation of, 471 Herschel–Bulkley model, 470–471 time-dependent, 467 Non-suppressed-ion chromatography, 174 Normal pulse voltammetry, 386–387 Normal stacking mode, 410 NPD, see Nitrogen-phosphorus detector NPV, see Normal pulse voltammetry NSM, see Normal stacking mode Nuclear magnetic resonance food and grain composition analysis moisture determination, 254–255 oil determination, 254 proteins, 252–254 magnetic ﬁeld, 282 water molecule, 283 Nuclear Overhauser effect spectroscopy, 304

O Ochratoxin, 111 Oil determination Babcock method, 275 solvent extraction methods, 274 Oil extraction, 96 Olive oils, 373 Online degassers, of IC, 167 On-line detection, in CE, 410–411 Online reagent generation, of IC, 168 Online sample preparation, 111–112 Open-vessel heating, 60 OPLC, see Overpressured layer chromatography Optical reﬂection techniques, 234 Optical spectrometry, 235–236 Orbitrap analyzer, 218 Organic contaminants, CE determination of, 417 Organic-in-carbon dioxide dispersions, 32 Organic solvents, 27, 33 Organometallic labels, for multi-label immunoassays, 441 Organophosphate pesticides adsorption of, 393 redox process of, 392 Ostwald–de Wael model, 469–470 Overpressured layer chromatography, 150–152 Oxygen electrodes, 395

P PA (polyacrylate), 15 Paprika extraction, 64 PARAFAC decomposition, 359 Partial least squares, 252, 354 PAT, see Process analytical technology PCA, see Principal component analysis PCR, see Principal component regression PCR products, separation, 419 PDMS (polydimethylsiloxane), 15 PED, see Pulsed electrochemical detection

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 522 15.7.2008 8:20am Compositor Name: BMani

522 Peroxide electrodes, 395 Pesticides analysis by CE, 417 electrochemical determination of, 391, 393–394 PFPD, see Pulsed ﬂame photometric detector PH and ion exchange system, 166 Pharmaceutical and toxicological applications, of SFE, 41 Phase diagram, for a typical real gas, 28 Phenolic compounds, analysis by CE, 414 Photo-ionization detector, 136 Photomultiplier tube, 320, 328 Photon multiplier, 220 PID, see Photo-ionization detector Pigment extraction, from paprika, 67 Pigments, in food, 233 p–p interactions, 131 Plant breeding, 256 PLC, see Preparative layer chromatography PLS, see Partial least squares PLS-1 regression calibration algorithm, 252 PMT, see Photomultiplier tube Polarity, 8 Polyacrylamide gels, 425, 428 Polymeric resins, 11 Polyphenolic compounds analysis, 414–415 Post-column suppression, beneﬁts, 175 Potassium iodide–ethanol system, 31 Potentiostat, 382–383 Power ultrasounds, 87 Precision, 4 Pre-column derivatization, 169 Preconcentration voltammetric techniques adsorptive stripping voltammetry, 389 anodic stripping voltammetry, 389–390 cathodic stripping voltammetry, 390 Precursor ion scan, 211 Predicted riboﬂavin content, 356 Preparative layer chromatography application in food analysis, 155–156 detection of zones, 152–155 development of layer, 148–152 future prospects, 156–157 layers for, 146 recovery of separated zones from layer, 155 sample application, 146–148 Preservative analysis, 412 Pressure–temperature phase diagram, 30 Principal component analysis, 252, 354, 370 Principal component regression, 252 Process analytical technology, 358 Programmable temperature vaporization, 123–124 Protein analysis methods biuret method, 270–271 dye-binding methods, 274 Folin–Ciocalteu reaction, 272 Kjeldahl method, 270 Lowry method, 271–272 Ohnishi-Barr and Sigma Chemical Co. methods, 272–274 ultraviolet 280 nm absorption methods, 270 Provisional tolerable weekly intake, 341 Pseudostationary phases, 407, 410 PTV, see Programmable temperature vaporization

Handbook of Food Analysis Instruments PTV solvent vent injection, 123–124 PTV splitless injection, 123–124 PTWI, see Provisional tolerable weekly intake Pulsed electrochemical detection, 162, 178 Pulsed ﬂame photometric detector, 136 Pulsed splitless injection, 122 Pulse sequence, 287, 292, 305, 310 Pulse voltammetric techniques basic principle of, 386 differential pulse voltammetry, 387 normal pulse voltammetry, 386–387 parameters of, 386 square-wave voltammetry, 388 Pulse voltammetry, 386 Puncture test, 463 Purge-and-trap technique, 60

Q QDA, see Quadratic discriminant analysis QDs, see Quantum dots Quadratic discriminant analysis, 337 Quadrupole, 211–212 Quadrupole ion trap (3D) mass analyzer, 213 Quantiﬁcation limit, 4 Quantum dots, 441

R Radioactive PLC zones, 155 Radioautography, 155 Radioisotopic labels, 440 Raman scanning electron microscopy, 507 Raman spectroscopy, 239 Range, 4 RDI, see Reference daily intakes Recombinant antibodies, 452 Reduced pressure (Pr), 26 Reduced temperature (Tr), 26 Reference daily intakes, 336 Reference electrode, 380, 384 Reﬂectance, 153 Reﬂection spectrometry, 234, 236 Regression, 252 Relative electrophoretic mobility, 428 Relative quantiﬁcation, 428 Relaxation, theory of, 284–285 Residuum oil supercritical extraction process, 39 Resin-based phases, 163 Restricted Access Media (RAM) sorbents, 12 Retention time (tR), in GC analysis, 127–128 Rheological parameters of foods, 476 Rheology deﬁnition, 461 Robustness, 4 ROESY, see Rotating frame nuclear Overhauser effect spectroscopy ROSE process, see Residuum oil supercritical extraction process Rotating frame nuclear Overhauser effect spectroscopy, 304 Rotational energy, of a motor, 167 RPC separations, 152

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 523 15.7.2008 8:20am Compositor Name: BMani

Index

S Sacriﬁcial electrode, 178 Sample introduction device, of IC, 168–169 Sample magnetization, nuclear spin and, 248 Sample preparation and injection, hydrodynamic and electrokinetic, 409 for SEM observation, 500 for TEM observation, 499–500 Sample stacking, 409–410 Sandwich immunoassays, 449 SAW sensors, see Surface acoustic wave sensors SAX, see Strong anion exchangers SBSE, see Stir bar sorptive extraction Scanning electron microscopes, 504–505 Scanning electron microscopy, 68, 497, 500–505, 508–509 basic components of, 500 cryogenic, 505 primary electron beam, 501 secondary electrons, 502 Scanning transmission electron microscopy, 503–504 SCE, see Strong cations exchangers SDME, see Single-drop microextraction SDS electrophoresis, 429 SDS polyacrylamide gels, 437 SE, see Secondary electrons Seafood quality assessment, 371–372 Sealable microwave-transparent container, 60 Secondary electrons, 502 Selected ion monitoring, 210 Selected reaction monitoring, 106, 109, 211 Selectivity, 3 SEM, see Scanning electron microscopy Semidry blotter, 435 Sensor types, in E-nose, 367–368 Separation capillaries, 407–409 Separation, in liquid chromatography, 163 SFC, see Supercritical ﬂuid chromatography SFE, see Supercritical ﬂuid extraction SFME, see Solvent-free microwave extraction Sigma Chemical Co. procedure, effect of pH and denaturation, 27, 273 Signal-to-noise ratio (S=N), 4 Silica-based columns, 169 Silica-based phases, 163 Silicon-nanowire ﬁeld-effect devices, 452 SIM, see Selected ion monitoring Single-drop microextraction, 9–10 Single point calibration, 3 Skimmed milk powder, 302 Slab gel apparatus, 432 Slab gels, 425 Slit-scanning densitometer, 153 Small organic and inorganic ions, in foods analysis by CE, 416–417 SMB interface, see Supersonic molecular beam interface Smell-seeing sensor, 368 Smith-Hieftje, 323 SMP, see Skimmed milk powder Sodium chloride, 162 Solid-phase dynamic extraction method, 17 Solid-phase extraction, 8, 10–14, 65, 112–114

523 Solid-phase microextraction, 8, 14–17, 60, 65, 69, 125–126 Solomon–Bloembergen–Morgan (SBM) theory, 291 Solubility (hydrophobicity), 8 in supercritical ﬂuids, 32–33 Solute–ﬂuid binary diffusion coefﬁcients, 34 Solvent delivery systems, 105, 167–168 Solvent-free microwave extraction, for essential oils, 66 Solvent residues, in pharmaceutical and food products, 33 Sorbents in cartridge, 11 RAM, 12–13 for solid-phase extraction, 113–114 in SPE, 12 Southern Blotting, 427 Soxhlet extraction, 33, 36, 58, 65, 69, 86, 89, 97 Soybean carbohydrates, 279 lipids, 279 minor components of, 279 moisture correlation curve, 260 proteins classiﬁcation, 278–279 quantitation by CE, 413–414 Soybean seeds DA spectra, coat color, 255–256 NIRS calibration procedures, 257 liquid water, 260 nondestructive analysis, 258 puriﬁed SPI protein superimposed, 259 protein content distribution, 263–265 protein-oil correlation plot, 262 USDA-UIUC germplasm DA-NIR analysis of, 260–261 NIR analysis of, 262 Soy protein isolates, 254 Spatially resolved multiplexed immunoassay, 442–451 bead assays, 450–451 capillary assays, 442–443 electrode assays, 451–452 membrane assays, 442 microchip assays, 443–449 microtiter plate assays, 449–450 nanowire assays, 452 SPDE method, see Solid-phase dynamic extraction method SPE, see Solid-phase extraction Spectral acquisition speed, 210 Spectral window, nuclei, 295 Spectroﬂuorometer, 353 SpectrumOne NTS FT-NIR spectra, soy protein isolates, 254 spectra, bulk soybean samples, 253 Speed, 5 SPI, see Soy protein isolates Split=splitless injection, 120–121 SPME, see Solid-phase microextraction Square-wave voltammetry, 388 Staining method, 427 Staircase potential ramp, 385 Standard addition, 3 Static techniques, 60 STEM, see Scanning transmission electron microscopy

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 524 15.7.2008 8:20am Compositor Name: BMani

524 Still CCD cameras, 436 Stir bar sorptive extraction, 9, 17–18 Stokes shift, 349 Stress relaxation test, 479–481 Stripping voltammetric techniques, 389 Strong anion exchangers, 165 Strong cations exchangers, 165 Submarine chamber, 433 Sugar determination extraction of monosaccharides and oligosaccharides, 275–276 high performance liquid chromatography, 276 polarimetry, 276–277 refraction index (RI) measurements, 277 Supercritical carbon dioxide (SCCO2), 27 European extraction plants utilizing, 39–41 Supercritical ﬂuid, 26 applications and commercial processes of chromatography, 43–45 environmental, 42 food, 42–43 pharmaceutical, 41 background and historical perspective, 27 basic properties and fundamentals of dispersions in, 32 phase behavior, 29–31 phase transitions, 28–29 solubility in, 32–33 solvent strength of a ﬂuid, 31–32 extraction, 27, 33–36 advantages, 35, 50–51 experimental considerations, 37–38 instrumentation, 45–51 mechanism, 36 theory, 37 trends and future prospects, 51–53 Supercritical ﬂuid chromatography, 26, 43–51 in food industry, 26 advantage, 43–44 commercial application, 43–45 Supercritical ﬂuid extraction, 98 Supersonic molecular beam interface, 220 Suppressed-ion chromatography, 174–175 Surface acoustic wave sensors functionality of, 368 virgin olive oils, aroma classiﬁcation, 373 Surface-enhanced laser desorption=ionization (SELDI)-MS, 448 Surface plasmon resonance biosensors, 447 Surfactant design, 32 Sweeteners in foodstuffs, 412 SWV, see Square-wave voltammetry Syringe-based or displacement pumps, 167

Handbook of Food Analysis Instruments Technique selection, criteria for accuracy, 3 convenience, 5 detection limit, 4 precision, 4 quantiﬁcation limit, 4 range, 4 robustness, 4 selectivity, 3 speed, 5 TEM, see Transmission electron microscopy Tensile and extensibility test, 464 Tensile tests, food material, 464–466 Texturized vegetable products, 248 Thermal conductivity detector, 136 Thermal modulation, in capillary GC, 132 Thin layer chromatography, 145–150, 155–156 Thin-layer electrophoresis, Peltier-cooled ﬂatbed system for, 433 Thixotropic ﬂuids, 467 Three-point bending test, 464; see also Bending test Time-dependent ﬂuid behavior, 471 effects of temperature and concentration on, 473 mathematical models, 472 Time domain NMR longitudinal and transverse relaxation times, 290–291 meat sample preparation for, 293 protonless, 291 Time-independent ﬂuid behavior, 469–471 Time-of-ﬂight, 213, 222 Time-of-ﬂight (TOF) mass analyzer, 213–215 TLC, see Thin layer chromatography TOCSY spectra, see Total correlation spectroscopy spectra Tolerable daily intake, 336 Total acrylamide concentration, 428 Total correlation spectroscopy spectra, 304 Traceability, deﬁned, 1 Transient viscoelasticity, 473 Transmission electron microscopes, instrumentation, 503–504 Transmission electron microscopy, 496–500, 502–503, 507–508 principles and theory, 498–500 Triazines mechanism of reduction of asymmetrical, 391–392 Tristimulus colorimeter, 235 Tryptophan, ﬂuorescence properties of, 349 TVP, see Texturized Vegetable Products Two-dimensional CE, 411 Two-dimensional gas chromatography (2D-GC) separation, 130

U T Tandem MS function, 211 Tank blotters, 434–435 Tannin, standard determination in, 3 TCD, see Thermal conductivity detector TDI, see Tolerable daily intake TD-NMR, see Time domain NMR

UAE, see Ultrasound-assisted extraction U-column MAP system, 69 Ultra-performance liquid chromatography, 105 Ultrasonic cleaning bath, 88 Ultrasonic imaging, 87 Ultrasonic nebulization, 337 Ultrasonic nebulizers pump, 326 Ultrasonic spectroscopy, 87

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 525 15.7.2008 8:20am Compositor Name: BMani

525

Index Ultrasound-assisted extraction, 85–99 Ultrasound cavitation phenomena, in solvent, 89 Ultrasound extraction technology basic principles, 86–89 environmental impact, 99 food application antioxidants, 93–96 ﬂavors and fragrances, 91 metals, 91–93 oil and fat, 96 future trends, 99 parameters and mechanism, 89–90 vs. traditional technique accelerated solvent extraction (ASE), 98 microwave-assisted extraction (MAE), 98–99 Soxhlet technique, 97 supercritical ﬂuid extraction (SFE), 98 Unsuppressed-ion chromatography, 174 UPLC, see Ultra-performance liquid chromatographs US leaching technique, 93 USN, see Ultrasonic nebulization UV absorbance methodology, 171 UV-Vis detectors, 410

V Valve-based sample introduction system, 169 Vanillin concentrations, extraction of, 91 Variable pressure scanning electron microscopes, 502 Vegetables quality assessment tomatoes, 372 Versatility, 211 Vertical electrophoresis systems, 431–432 Vinegars, FITC–amino acids in, 418–419 Virgin olive oils aroma classiﬁcation, 373 phenolic compounds analysis, 414–415 Viscoelasticity linear, testing of creep, 482–484 oscillatory measurements, 484–485 stress relaxation test, 479–481 nonlinear compression test, 478 nonlinear, testing of, 485–490 tests for transient, 473 Leider and Bird equation, 474–477 shear stress, 477 Viscoelasticity test, 478–490 Viscous materials testing ﬂuid behavior apparatus for, 468–469 modeling, 469–473 shear ﬂow, 466–467

Visible=near-infrared reﬂection (VNIR) spectrophotometers, 236 Vitamins, analytical determination using CE, 416 Volatile aroma compounds, 17 Volatility, 8 Voltammetric resolution, of complex mixtures, 179–180 Voltammetric techniques Butler–Volmer equation, 381–382 electroanalytical system for electrochemical cell and electrodes, 384 potentiostat, 382–383 electrochemical reaction and Nernst equation, 381 food analysis using advantages of, 380 cyclic voltammetry, 385–386 metals contaminants, 390, 392 pulse voltammetric techniques, 386–388 preconcentration (see Preconcentration voltammetric techniques) principles of, 380 Voltammetry, 380–384 VPSEMs, see Variable pressure scanning electron microscopes

W Water holding capacity, 292 Water inonization, 162 Weak acids and bases, 162 Weak ion exchangers, 165 Weibull-Berntrop extraction method, 66 Western blot assays, 442 WHC, see Water holding capacity White light scanners, 436 WinCATS software, 147, 153 Wine matrix, selective heating of, 60–61 Wrap-around modulation, 133

X X-ray microanalysis, 505–506

Z Zeeman effect, 323 Zone detection, with PLC, 152–155 Zone electrophoretic techniques charge mobility, 428 homogeneous buffer and gel structure, 429 immuno electrophoresis, 429–430 native conditions, 429 Zwitterionic compound, 163

Otles/Handbook of Food Analysis Instruments 45660_C023 Final Proof page 526 15.7.2008 8:20am Compositor Name: BMani

Otles/Handbook of Food Analysis Instruments 45660_CI Final Proof page 1 15.7.2008 8:20am Compositor Name: BMani

Pc P

T6 T5 T4 T3 T1 Vc

Tc T2

V

Liquid

Liquid and vapor

Gas

SCF

FIGURE 3.1 Phase diagram for a typical real gas.

(a)

FIGURE 4.1 Typical commercial MAP extraction apparatus suitable for food analysis laboratories. (a) shows a focused mono-mode open-vessel apparatus for multi-step fast extraction. (continued )

Otles/Handbook of Food Analysis Instruments 45660_CI Final Proof page 2 15.7.2008 8:20am Compositor Name: BMani

(b)

FIGURE 4.1 (continued)

(b) shows a multi-mode closed-vessel multi-sample apparatus.

Chlorophyll a Chlorophyll b Carotenoids

400 Violet

500 Indigo

Blue

600 Green

Yellow Orange

700 Red

A: Absorbed colors

Yellow Orange Red Violet Indigo Blue Green T: Colors complementory to a, seen after transmission or reflection of light

FIGURE 11.2 Chemical identiﬁcations by complementary display of color absorption and transmission. The absorption spectra of chlorophyll a, chlorophyll b, and carotenoids are plotted over the wavelength scale. Below the wavelength scale: A—the absorbed colors (not perceivable by the eye) are named and T—the complemetary colors are named which are transmitted or reﬂected and are perceived by the eye.

Otles/Handbook of Food Analysis Instruments 45660_CI Final Proof page 3 15.7.2008 8:20am Compositor Name: BMani

(A)

(B)

FIGURE 12.2 SpectrumOne NTS spectra of bulk soybean samples, before (A) and after (B) multiple scattering correction (MSC). Note the very signiﬁcant consistency of data obtained after applying MSC.

Otles/Handbook of Food Analysis Instruments 45660_CI Final Proof page 4 15.7.2008 8:20am Compositor Name: BMani

2.2 Sucrose Oil Water Fiber Protein WholeSoy

2.0 1.8 1.6 Absorbance

1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 400

500

600

700

800

900 1000 1100 1200 1300 1400 1500 1600 1700 Wavelength, nm

FIGURE 12.7 NIR spectra of the major components present in soybean seeds compared with the NIR spectrum of the whole soybean seed.

(a)

(b)

(c)

(d)

FIGURE 13.5 Meat sample preparation for TD-NMR: the homemade tool permits to excise meat rods (a) without disrupting its inner ﬁbers; the tool ﬁts the 10 mm inner diameter NMR tube (b) and can extrude the rod directly in the bottom of the tube (c); the rod is about 10 mm high (d) and occupies the volume within the probe receiver coils. (Courtesy of E. Tettamanti, University of Teramo, Italy.)

Otles/Handbook of Food Analysis Instruments 45660_CI Final Proof page 5 15.7.2008 8:20am Compositor Name: BMani

650

Excitation (nm)

600

Amino acids

550 Chlorophyll Porphyrin

500 450 400 350 300 250

300 350 400 450 500 550 600 650 700 Emission (nm)

NADH

ATP Vitamins (A, B2, B6, E)

FIGURE 15.7 Fluorescence landscape of a food sample and emission maxima of the most relevant food ﬂuorophores.

100 90 80 70 60 50 40 30 20 10 (a)

(b)

FIGURE 15.12 (a) Subsection of a commercially packed Swiss cheese and (b) ﬂuorescence image of the degree of photooxidation after storage under standard commercial illumination.

Otles/Handbook of Food Analysis Instruments 45660_CI Final Proof page 6 15.7.2008 8:20am Compositor Name: BMani

Immunobeads Antigens

+

Labeled antibodies

+

(a)

Lasers (b)

FIGURE 20.6 Bead-based multiplexed immunoassays: (a) antibody–antigen complexes are formed after immunobeads incubate with samples. After the addition of labeled antibodies, the sandwich complexes are formed and (b) beads are analyzed by ﬂuorescence ﬂow cytometry. A red laser classiﬁes the bead and a green laser quantiﬁes the antigens. (Modiﬁed from Rao, R.S., Visuri, S.R., McBride, M.T., Albala, J.S., Matthews, D.L., and Coleman, M.A., J. Proteome Res., 3, 736, 2004.)