Thermal technologies in food processing
Related titles from Woodhead’s food science, technology and nutrition list: Extrusion: Technologies and applications (ISBN: 1 85573 559 8) Extrusion is a major technology for a wide range of food products. The first part of this book introduces this technology, examining extruders and their use in thermal transitions of raw materials into functional forms for the manufacture of particular foodstuffs. The second part of the book looks at the application of extrusion in specific product groups. It will be invaluable reference for improving efficiency and product quality in this area of the food industry. Frying (ISBN: 1 85573 556 3) Edited by one of the world’s leading authorities in the field, this collection covers the key issues in fried product quality. Part I covers general issues such as the market for fried foods, regulation and health issues. Part II covers issues on raw materials such as composition of frying oils and factors affecting the quality of oils. Part III concentrates on processing issues such as the frying process and packaging, all of which affect the flavour, texture and colour of fried foods. Food processing technology: Principles and practice (ISBN: 1 85573 533 4) The first edition Food processing technology was quickly adopted as the standard text by many food science and technology courses. The publication of this completely revised new edition is set to confirm the position of this textbook as the best singlevolume introduction to food manufacturing technologies available. New chapters include computer control of processing, novel ‘minimal’ technologies including processing using high pressures or pulsed electric fields, ohmic heating and an extended chapter on modified atmosphere packaging. Details of these books and a complete list of Woodhead’s food science, technology and nutrition titles can be obtained by: • visiting our web site at www.woodhead-publishing.com • contacting Customer services (email:
[email protected]; fax: +44 (0) 1223 893694; tel.: +44 (0) 1223 891358 ext. 30; address: Woodhead Publishing Limited, Abington Hall, Abington, Cambridge CB1 6AH, England)
If you would like to receive information on forthcoming titles in this area, please send your address details to: Francis Dodds (address, tel. and fax as above; e-mail:
[email protected]). Please confirm which subject areas you are interested in.
Thermal technologies in food processing Edited by Philip Richardson
Published by Woodhead Publishing Limited, Abington Hall, Abington Cambridge CB1 6AH, England www.woodhead-publishing.com Published in North and South America by CRC Press LLC, 2000 Corporate Blvd, NW Boca Raton FL 33431, USA First published 2001, Woodhead Publishing Limited and CRC Press LLC ß 2001, Woodhead Publishing Limited The authors have asserted their moral rights. This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. Reasonable efforts have been made to publish reliable data and information, but the authors and the publishers cannot assume responsibility for the validity of all materials. Neither the authors nor the publishers, nor anyone else associated with this publication, shall be liable for any loss, damage or liability directly or indirectly caused or alleged to be caused by this book. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming and recording, or by any information storage or retrieval system, without permission in writing from the publishers. The consent of Woodhead Publishing Limited and CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from Woodhead Publishing Limited or CRC Press LLC for such copying. Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress. Woodhead Publishing Limited ISBN 1 85573 558 X CRC Press ISBN 0-8493-1216-7 CRC Press order number: WP1216 Cover design by The ColourStudio Project managed by Macfarlane Production Services, Markyate, Hertfordshire (
[email protected]) Typeset by MHL Typesetting Limited, Coventry, Warwickshire Printed by TJ International, Padstow, Cornwall, England
Contents
List of contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . List of notation 1
ix xiii
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . P.S. Richardson, Campden & Chorleywood Food Research Association, Chipping Campden
1
Part I Conventional technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5
2
3
Retort technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . N.S. May, Campden & Chorleywood Food Research Association, Chipping Campden 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 The basic retort cycle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Selection of container . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Selection of a retort . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 The influence of heating medium on retort performance . . . . . 2.6 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7 Sources of further information and advice . . . . . . . . . . . . . . . . . . . 2.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Continuous heat processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . S.P. Emond, Campden & Chorleywood Food Research Association, Chipping Campden 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Indirect heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Direct heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7 7 7 11 11 13 25 27 27 29 29 30 39
vi
Contents 3.4 3.5 3.6 3.7
Holding section . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sources of further information and advice . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
42 44 45 48
Part II Measurement and control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
49
4
5
6
7
Pressure and temperature measurement in food process control P.G. Berrie, Endress+Hauser Process Solutions AG, Reinach 4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2 Pressure measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3 Temperature measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 General instrument design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
51
Validation of heat processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . G.S. Tucker, Campden & Chorleywood Food Research Association, Chipping Campden 5.1 Introduction: the need for better measurement and control . . 5.2 Validation methods: objectives and principles . . . . . . . . . . . . . . . 5.3 Temperature distribution testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 Heat penetration testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 Microbiological spore methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 Biochemical time and temperature integrators . . . . . . . . . . . . . . . 5.7 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.8 Sources of further information and advice . . . . . . . . . . . . . . . . . . . 5.9 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
75
Modelling and simulation of thermal processes . . . . . . . . . . . . . . . . B.M. Nicolaı¨, P. Verboven and N. Scheerlinck, Katholieke Universiteit, Leuven 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 Modelling of conduction heat transfer: the Fourier equation . 6.3 The Navier–Stokes equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Numerical methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.5 Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.7 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
91
Modelling particular thermal technologies . . . . . . . . . . . . . . . . . . . . . . S. Bakalis, P.W. Cox and P.J. Fryer, University of Birmingham 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 Processing of packed and solid foods . . . . . . . . . . . . . . . . . . . . . . . . 7.3 Continuous heating and cooling processes . . . . . . . . . . . . . . . . . . . 7.4 Heat generation methods: ohmic and microwave heating . . . . 7.5 Developments in the field . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
51 52 57 68 73
75 76 79 81 83 85 87 88 89
91 92 93 97 104 108 109 109 113 113 116 122 127 131 133
Contents 8
vii
Thermal processing and food quality: analysis and control A. Arnoldi, University of Milan 8.1 Introduction: the importance of thermal processing . . . . . . . . . . 8.2 The importance of the Maillard reaction . . . . . . . . . . . . . . . . . . . . . 8.3 Thermal processing and food safety . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Thermal processing and nutritional quality . . . . . . . . . . . . . . . . . . . 8.5 Thermal processing, food flavour and colour . . . . . . . . . . . . . . . . 8.6 Maillard reaction and lipid oxidation . . . . . . . . . . . . . . . . . . . . . . . . 8.7 Controlling factors in the Maillard reaction . . . . . . . . . . . . . . . . . . 8.8 Methods of measurement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.9 Application to the processing of particular foods . . . . . . . . . . . . 8.10 Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.11 Sources of further information and advice . . . . . . . . . . . . . . . . . . . 8.12 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
138 139 142 143 145 148 149 150 151 153 154 154
Part III New thermal technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
161
9
Radio frequency heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A.T. Rowley, EA Technology Ltd, Chester 9.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.2 Basic principles of RF heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.3 Application to food processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.4 Advantages and disadvantages of RF heating . . . . . . . . . . . . . . . . 9.5 RF heating technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.6 Case studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.7 Future trends in RF heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.8 Sources of further information and advice . . . . . . . . . . . . . . . . . . . 9.9 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9.10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
138
163 163 163 166 167 169 173 175 176 177 177
10 Microwave processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . M. Regier and H. Schubert, University of Karlsruhe 10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.2 Physical principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.3 Microwave applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.4 Modelling and verification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.5 Summary and outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.6 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
178
11 Infrared heating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . C. Skjo¨ldebrand, ABB Automation Systems (formerly Swedish Institute of Food Research (SIK)), Tumba 11.1 Introduction; principle and uses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Theories and infrared properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.3 Technologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.4 Equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
208
178 180 194 199 203 204
208 210 215 217
viii
Contents
11.5 11.6 11.7
Applications: case studies and modelling . . . . . . . . . . . . . . . . . . . . Future trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
222 225 227
12 Instant and high-heat infusion J. Andersen, APV Systems, Silkeborg 12.1 Instant infusion: an introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.2 Instant infusion in detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.3 Advantages and disadvantages of instant infusion . . . . . . . . . . . 12.4 High-heat infusion: an introduction . . . . . . . . . . . . . . . . . . . . . . . . . . 12.5 The problem of heat resistant spores (HRS) . . . . . . . . . . . . . . . . . 12.6 High-heat infusion in detail . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12.7 Advantages and disadvantages of high-heat infusion . . . . . . . . 12.8 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
229
13 Ohmic heating R. Ruan, X. Ye, P. Chen and C.J. Doona, University of Minnesota and I. Taub, US Army Natick Soldier Center 13.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.2 Ohmic heating process and equipment . . . . . . . . . . . . . . . . . . . . . . . 13.3 Monitoring and modeling of ohmic heating . . . . . . . . . . . . . . . . . 13.4 Major challenges and needs for future research and development . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13.5 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
241
14 Combined high pressure thermal treatment of foods L. Ludikhuyze, A. Van Loey, Indrawati and M. Hendrickx, Katholieke Universiteit, Leuven 14.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.2 Effect of high pressure on micro-organisms . . . . . . . . . . . . . . . . . 14.3 Effect of high pressure on food quality related enzymes . . . . 14.4 Effect of high pressure on food structure and texture . . . . . . . . 14.5 Effect of high pressure on sensorial and nutritional properties of food . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.6 The use of integrated kinetic information in process design and optimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.7 High pressure processing technology and products . . . . . . . . . . 14.8 Conclusive remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.9 Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14.10 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
266
Index
229 230 232 234 234 235 238 239
241 245 247 260 264
266 267 268 271 273 275 276 278 278 278 285
Contributors
Chapter 1
Chapter 3
Professor Philip Richardson Head – Process and Product Development Department Campden & Chorleywood Food RA Chipping Campden Glos GL55 6LD England
Mrs Sue Emond Campden & Chorleywood Food RA Chipping Campden Glos GL55 6LD England
Tel: +44 (0)1386 842036 Fax: +44 (0)1386 842100 E-mail:
[email protected]
Tel: +44 (0)1386 842105 Fax: +44 (0)1386 842100 E-mail:
[email protected]
Chapter 4 Chapter 2 Mr Nick May Department of Process and Product Development Campden & Chorleywood Food RA Chipping Campden Glos GL55 6LD England Tel: +44 (0)1386 842031 Fax: +44 (0)1386 842100 E-mail:
[email protected]
Dr Peter Berrie Endress+Process Solutions AG Christoph-Merian-Ring 23 CH 4153 Reinach BL1 Switzerland Tel: +41 (0) 61 715 7340 Fax: +41 (0) 61 715 7301 E-mail:
[email protected]
x
Contributors
Chapter 5
Chapter 8
Mr Gary Tucker Department of Process and Product Development Campden & Chorleywood Food RA Chipping Campden Glos GL55 6LD England
Professor Anna Arnoldi DISMA University of Milan Via Celoria 2 IT-20133 Milan Italy
Tel: +44 (0)1386 842035 Fax: +44 (0)1386 842100 E-mail:
[email protected]
Tel: +39 02 2362721 Fax: +39 02 70633062 E-mail:
[email protected]
Chapter 9 Chapter 6 Professor Bart Nicolaı¨ Department of Agro-Engineering and Economics Katholieke Universiteit Leuven de Croylaan 42 B-3001 Heverlee Belgium Tel: +32 16 322375 Fax: +32 16 322955 E-mail:
[email protected]
Chapter 7 Dr P. W. Cox, Dr S. Bakalis and Professor P. J. Fryer Food Processing Group Centre for Formulation Engineering School of Chemical Engineering University of Birmingham Edgbaston Birmingham B15 2TT England Tel: +44 (0)121 414 5310 E-mail:
[email protected]
Dr Andrew T. Rowley EA Technology Limited Capenhurst Chester CH1 6ES England Tel: +44 (0)151 347 2392 Fax: +44 (0)151 347 2560 E-mail:
[email protected]
Chapter 10 Mr Marc Regier and Professor Helmar Schubert Institut fu¨r Lebensmittelverfahrenstechnik Universita¨t Karlsruhe Kaiserstr. 12, Geb 30.44 D-76128 Karlsruhe Germany Fax: +49 (0)721/694320 E-mail:
[email protected]
Contributors xi
Chapter 11 Dr Christina Skjo¨ldebrand ABB SE 147 80 Tumba Sweden Tel: +46 08 530 66704 Fax: +46 08 530 66110 Business e-mail:
[email protected] Home e-mail:
[email protected]
Chapter 12 Dr Joergen Andersen APV Systems Pasteursvej 8600 Silkeborg Denmark E-mail: joergen.andersen@invensys
Chapter 13 Dr Roger Ruan, Mr Xiaofei Ye, Dr Paul L. Chen and Mr C.J. Doona Department of Bioscience and Agricultural Engineering University of Minnesota 1390 Eckles Avenue St Paul MN 55108 USA
Tel: +1 612 625 1710 Fax: +1 612 624 3005 E-mail:
[email protected] Dr Irwin A. Taub Sustainability Directorate US Army Natick Soldier Center Natick MA 01560 USA
Chapter 14 Professor Dr Ir M. Hendrickx Department of Food and Microbial Technology Katholieke Universiteit Leuven Kardinaal Mecierlaan 92 BE-3001 Leuven (Heverlee) Belgium Tel: +32 16 321 585 Fax: +32 16 321997 E-mail:
[email protected]
This page intentionally left blank
Notation
a A A a, b Ap Bi hd= ! B B C C
surface area surface area coefficient matrix constants particle surface area Biot number magnetic flux density flux density capacity matrix concentration of bacteria in liquid food
C1, C2 C c c, c0 ca cP ! D D D Dref Dp De d ! E E
constants turbulence constant heat capacity (Chapter 6) velocity of light, in vacuum (Chapter 10) specific moisture capacity of vapour phase specific heat capacity (constant pressure) electric flux density flux density decimal reduction time reference decimal reduction time field penetration depth diffusion coefficient diameter electric field electric field strength
m2 m2 kg s
1
m2 VSm 2 J ºC 1 number of bacteria m J kg
1
ºC
J kg 1 K VSm 2 VSm 2 min min m m2 s 1 m Vm
1
3
1
1
xiv
Notation
F ^ F f f fi Fi Fo G SR=V g g, h H H ! H H h h
1
h hevap hi Ii i ! j K K k k k ! k kt L M M, Ml m, n, o, p. q, r m N n ni n? ! P p Q
F-value: integrated lethality (Chapter 7) volumetric average thermal load vector frequency volumetric body force Cartesian component of particle force Fourier number shape factor of the container acceleration due to gravity (Chapter 7) constant (Chapter 10) static enthalpy (Chapter 6) specific enthalpy (Chapter 7) magnetic field (Chapter 10) magnetic field (Chapter 7) surface heat transfer coefficient (Chapter 6) convective heat transfer coefficient (Chapter 7) Planck’s constant evaporation heat density enthalpy of phase i mass sink or source density of phase i imaginary unit electric current density turbulent kinetic energy stiffness matrix electrical conductivity thermal conductivity (Chapter 7) Boltzmann’s constant (Chapter 10) wave vector reaction rate constant at temperature T characteristic length molecular weight (Chapter 6) moisture content, liquid moisture content (Chapter 10) exponents (Chapter 6) constant (Chapter 10) shape function constant concentration of ion i outward normal to surface polarisation pressure power dissipated per unit volume (Chapter 7)
min sm 3 W GHz Nm 3 N ms
2
J kg Jm
3
Am Wm Wm
1
1 2 2
ºC K
m2 s 2 W ºC 1 Sm 1 W m 1 ºC s 1 m kg mol
Pa Wm
3
1
1 1
1
Notation Q Q Q(R) Q Qem qR R R r r r <
x S S T Tref Tm t t U u u uj V V v ! x xi, x, y, z x, z Z Zq zi Greek symbols
M p = cp 1 E
volumetric heat generation (Chapter 6) rate of heat (Chapter 11) radiant flux emitted per unit area unit increment of wavelength source term vector electromagnetic heat production density radiative power flux density universal gas constant (Chapter 6) smallest characteristic dimension of a geometry (Chapter 7) residual (Chapter 6) radial position of a can (Chapter 7) reflected waves (Chapter 11) real part of x heat transfer surface source of quantity temperature reference temperature temperature of the heating medium time (Chapters 6, 7 10) transmitted waves (Chapter 11) velocity vector nodal temperature vector velocity in the vertical direction Cartesian velocity component volume voltage (Chapter 7) velocity in the radial direction local vector Cartesian coordinate (Chapter 6) distance (Chapter 7) Z-value: slope of the lethality or cooking curve quality factor valence of ion i object domain boundary of object diffusivity of quantity absorbed waves mass diffusivity pressure diffusivity thermal diffusivity dielectric susceptibility electric field attenuation length
xv
3
Wm Js 1
W m 2 m units s 1 1
J mol m
1
K
m m2 units m ºC or K K K s ms
1
ms ms m3 V ms
1
3
s
1
1
m m K K
kg m
m2 s
1
1
s
1
1
xvi
Notation
p
T=Tref # 0 i S L e !
pressure gradient coefficient/power attenuation length thermal gradient coefficient turbulent energy dissipation rate (Chapter 6) emissivity (Chapter 11) permittivity (Chapter 7) relative permittivity (Chapter 10) convergence error dielectric constant (Chapter 7) dielectric loss factor (Chapter 7) dielectric constant of vacuum ratio of vapour flow to total moisture flow ratio of vapour diffusion to total moisture diffusion (Chapter 6) dimensionless axial length (Chapter 7) dynamic viscosity dimensionless temperature (Chapter 7) dimensionless temperature damping constant wavelength (Chapter 10) thermal conductivity (Chapter 7) relative permeability (Chapter 10) apparent viscosity (Chapter 7) permeability of vacuum mobility of ion i transported quantity per unit mass density (charge, mass) Stefan-Boltzmann constant (Chapters 6 and 11) conductivity (Chapter 10) electrical conductivity (Chapter 7) electrical conductivity (solid) electrical conductivity (liquid) electrical conductivity (effective) circular frequency
Subscripts 0 1
initial condition ambient condition
T 0 c 0 00 0 V
00
i
m2 s Fm
3 1
kg m
1
s
1
Wm
1
K
1
Pa s units kg 1 kg m 3 Wm 2K Wm Wm Wm Wm
4
1
1
1
1
K K 1 K 1 K
1 1
1 Introduction P.S. Richardson, Campden & Chorleywood Food Research Association, Chipping Campden
Thermal technologies have been at the core of food preservation and production for many years. Temperature is well known to be one of the major preservation mechanisms that can be applied by the food processor to render food products commercially sterile – that is, free from pathogenic and spoilage microorganisms likely to grow during the normal distribution and shelf-life of the product. Another key attribute of the application of heat in food manufacture is to bring about a modification of the texture and taste of foods, so facilitating the wide variety of foods now available to the consumer. In 1999, in the UK alone, there were in excess of 7000 new product launches, many of them requiring the application of a thermal technology either to cook them before chilled or frozen distribution or to render them commercially sterile for ambient distribution. The focus of this book is on technologies that have traditionally been used for food preservation and manufacture, and also those that are emerging and becoming more widely applied in the food industry. It is important to consider both the use of heat as a preservation mechanism (e.g. canning, UHT, etc.) and also as a tool to bring about textural and structural changes in products (e.g. frying, baking, sous vide) to meet the ever increasing demands of the consumer. The first two parts of the book focus on traditional in-container and UHTstyle technologies and on measurement and control issues. These chapters pick up such issues as techniques for validating the efficacy of a thermal process, and trends in current technologies which are largely driven by the demand for new products in different packaging formats that offer added convenience to the consumer. Changes in packaging have seen the concepts and benefits of traditional canning being achieved with the use of flexible trays and pouches and also glass containers. Trends in steriliser manufacture, offering a less hostile heating environment and
2
Thermal technologies in food processing
also permitting the application of overpressure around such containers, have been necessary to facilitate the use of such packaging media. Similarly the benefits of aseptic packaging have been recognised by many manufacturers who have been developing a number of new formats – pots, trays, bottles – for these products beyond the scope of the original paperboard containers. The availability of such packaging opportunities has created demand for products of more challenging rheological behaviour that may contain differing degrees of particulate material and hence the need for new designs of heat exchanger. Whilst the primary concern of food manufacturers is the production of safe foods, there is little market for low quality foods no matter how safe they are. Indeed, with the widespread availability of chilled foods, those involved in the delivery of thermally processed foods have a very significant challenge to meet in preparing foods that are as close to fresh as possible. Thermally processed foods do, however, have the real benefit of much longer shelf-lives than are currently possible using chilled food technologies. The science that describes the impact of heat on the quality attributes of foods is complex and is overviewed as part of this book. Thermally induced changes in foods are also required for the successful manufacture of many products (e.g. meat products or sauces). This is a positive benefit of the application of thermal technologies in food manufacture that can often be overlooked by the more usual trend to reduce the degree of heat treatment in foods. One of the keys to reducing the thermal damage to products during manufacture is the shortening of the heating, and for that matter cooling, times associated with the process. The third part of the book focuses on recent developments and advances in thermal technology that aim to do this. As an example the development of ohmic and microwave technologies have offered this benefit. Neither have found widespread application in food manufacturing for a variety of reasons, both technical and commercial, but are finding use in niche areas of food manufacturing offering measurable benefits over traditionally applied techniques. Microwaves have been applied very successfully to the tempering or partial defrosting of foods offering much more flexibility than traditional methods. However they have been less successful in the areas of pasteurisation and sterilisation where the uneven heating associated with microwave techniques has made the assurance of safety through process validation and control much more difficult than for traditional approaches. An extension of microwave technology is the application of radio frequency (RF) heating in food manufacturing. This can offer a more controlled heating within food materials with enhanced heating rates over those that than can be achieved with traditional conduction and convection heating technologies. The interaction of radio frequency electromagnetic radiation with the water deep inside food products has led to a widely used application of RF heating to offer fine control over final moisture content to ensure the production of high-quality baked goods (e.g. biscuits and crisps). As a technology RF drying offers significant advantages in flexibility and control over traditional hot air approaches leading to improved quality and consistency of products.
Introduction
3
Similarly infrared heating technologies have been developed to take advantage of the much higher rates of heat transfer that can be delivered by radiation as opposed to convection. This technology has offered exciting opportunities to those involved in continuous cooking operations where the balance of infrared and convective heating can be used to deliver new ranges of exciting products. It is clear that the current trend is to minimise the thermal impact of food processes. The heat process is, however, still an essential element and likely to remain so. This has led to a growth in products that have been specially formulated to require a reduced, pasteurisation heat treatment to render them commercially sterile. In combination with a reduced heat treatment, some of these products achieve microbiological stability through their pH or may require chilled distribution and storage during their extended shelf-life. Such products, whilst being very different in style from traditional canned products, do however still require appropriate advanced thermal technologies at the heart of their manufacture to cook and preserve them. Non-thermal technologies are emerging and, although not the subject of this publication, are worthy of mention in that they may, in the future, offer a viable alternative to in-pack pasteurisation or sterilisation of foods. Examples of such technologies are those based on the application of ultrahigh pressure or the application of pulsed electric fields to a narrow range of products. To date these have found very limited application in the global food industry with their use being largely limited to products that require pasteurisation rather than sterilisation. As with thermal processing, achieving commercial sterility is not always the only goal. There is a growing research and development interest in the application of these ‘new’ technologies to offer manufacturers a different range of raw materials or allow the development of new flavours and textures in existing materials. Current research thinking is not excluding thermal technologies from the food manufacturing arena. Indeed the wish is to harness the benefits of appropriate reduced heat treatments in combination with some of the emerging approaches to deliver a whole new raft of products for the consumer. Looking forward, thermal technologies have a strong role to play in current and future food manufacturing. The challenge is to the food and equipment manufacturers to apply these technologies intelligently and safely adding benefit and value to the products of tomorrow being demanded by consumers in the global food market.
This page intentionally left blank
Part I Conventional technologies
This page intentionally left blank
2 Retort technology N.S. May, Campden & Chorleywood Food Research Association, Chipping Campden
2.1
Introduction
Heat sterilisation of food in containers is an old technology largely attributed to the work of Nicholas Appert in the 1800s. From Appert’s work a substantial industry has developed. For example, the estimated sales of canned products in Europe are 26,000 million/year. The application of heat processes to containers requires not only the ability to heat and cool the container contents efficiently but also the ability to do so while minimising the stresses imposed upon the container. This controlled application of heat and pressure is the function of the modern sterilising retort. When faced with plans to launch a new in-container sterilised or pasteurised food product, various pieces of information must be gathered before decisions on the design of an installation can be made. This chapter aims to give the background required to understand these decisions. The reader will be guided through some of the diverse and ingenious ways that retort systems have evolved to maximise process efficiency. These increases in efficiency have taken place simultaneously while additional challenges have been presented by the development of less rigid packing types.
2.2
The basic retort cycle
The layout of a typical vertical saturated steam retort is given in Fig. 2.1. Understanding the key features of this traditional retort system will give the required insight to the features of more modern systems.
8
Thermal technologies in food processing
Fig. 2.1
Schematic of a batch steam retort.
The first action in the cycle of operation of a retort is to load the containers of product; in most systems this involves putting containers into crates. The sides and base of the crates, and any dividers used between layers of containers must be perforated sufficiently to allow heat to penetrate the load. The function of dividers between layers of containers can be to aid heat penetration but also to minimise container damage through abrasion and sometimes electrolytic effects from the crate. Once the required number of crates of product are loaded, the lid (or door) is closed and the steam turned on. After early attempts to bring about product sterilisation at above 100ºC using salt baths the heat preserving industry settled on the use of steam as the heating medium of choice. Steam is an excellent heat medium because of its ability to condense on container surfaces releasing large amounts of latent heat. The main enemy of efficient heating in a closed vessel (e.g. a retort) using saturated steam is the presence of entrapped air, especially that trapped in the small spaces between containers in the load. Even a small quantity of air has a significant effect upon the temperature. For example, at any given location 10% of air by volume will reduce the temperature by about 3ºC, which will have a dramatic effect on product sterilisation. This means that in order to bring about efficient and uniform heating, air must be purged, or as it is referred to in the industry, ‘vented’, from the retort at the start of the process. In the canning industry venting is achieved at the start of the operating cycle by introducing high velocity steam into the retort. This steam is allowed to pass through the vessel and exit through a vent at the opposite side of the machine,
Retort technology
9
i.e. if the retort is in a vertical orientation the steam enters at the bottom and exits at the top. This flow of steam displaces air out through the vent and when the flow rate is sufficient draws air from within the load by a Venturi effect.1 The efficiency of this process is sometimes aided by allowing steam to bypass the main control valve giving more rapid steam flow and a shorter come-up time. In practice, the efficiency of the venting process can be monitored by measurement of the temperatures in various locations throughout the retort, and the time required to remove all the air is determined experimentally. At the end of the venting time the vent valve is closed and the retort pressurised until the desired process temperature is reached. In the US target venting pressures, as well as times, are also commonly specified, though if the vent is sufficiently large in relation to the steam flow from the inlet, these should not be significant. Here it is interesting to contrast canning with other autoclaving industries which prefer to use automatic vent valves which operate throughout the entire cycle by opening on the detection of low temperature air. However, this does not give the high steam velocities that are considered to be necessary to draw air from the centre of crate loads of cans. After venting steam utilisation is reduced, and depending on the heat, absorbing capacity of the load continues to diminish throughout the holding period of the process. This high demand at the start of the process means that in some installations with low boiler capacities only one retort may be started at a time. During the hold phase there must be a means for the water generated as steam condenses to escape from the retort, because immersion of the lowest containers in the vessel could result in under-sterilisation. This is normally achieved either by a small condensate bleed that is permanently open or having the drain cracked slightly open. In either case the absence of water in the base of the retort can be detected by a free flow of steam. On reaching the intended holding temperature the requirements are for a narrow spread of temperatures throughout the vessel and stable control at the intended temperature. For a properly vented steam retort a temperature range of less than 0.5ºC is achievable, and this is a good target for all retort systems, though difficult to achieve in some types. The control of temperature fluctuations is generally by an automatic control system working from feedback from a temperature sensor in the instrument pocket of the retort. Formerly pneumatic controllers were used but electronic systems and/or programmable controllers have largely replaced these. There are minimum standards of instrumentation expected of any retort system which aim to ensure that product receives the correct process, and operator safety. The expected instruments include: 1. 2. 3.
A pressure gauge. A chart recorder to generate a permanent record of the heat process applied. A master temperature indicator (MTI). This is a calibrated thermometer
10
Thermal technologies in food processing
4.
independent of the chart recorder, which can be used to cross-check the chart record so that process safety does not rely on only one instrument. It is usually a mercury in glass or platinum resistance thermometer. A process timer. In simple installations a wall clock, but increasingly commonly the timer is built into the programmable controller. An automatic temperature controller, which may or may not operate from the same sensor as the chart recorder.
5.
To ensure the even application of heat to all containers in the load it is important to ensure that during the holding period the services which are not in use (air and water for cooling) do not leak resulting in low temperature regions. Valve types are selected to give minimum risk of leakage and double valves, with a route to atmosphere, are sometimes used. Once the desired hold period has been completed then the retort is cooled. The start of the cooling phase is most critical for ensuring the continued integrity of the processed containers. During the holding period the pressure inside the containers undergoing sterilisation increases as the contents heat and expand. With cans this internal pressure can be restrained adequately by the rigidity of the container and the external steam pressure, e.g. 1.0 kg/cm2 (15 p.s.i) at 121.1ºC. However, if at the start of cooling the supply of steam is turned off and the pressure allowed to drop quickly, then the internal pressure can be sufficient to cause permanent damage to the pack; in canning this is referred to as ‘peaking’. Although some canning operations producing tough cans do use this approach (air cooling) most operations introduce cold water for cooling. Introduction of water into a retort full of steam causes a very rapid collapse in the residual steam with a consequent loss of pressure. This is overcome by the replacement of the steam pressure with a supply of compressed air at the same, or a slightly higher, pressure, e.g. 0.1 kg/cm2 above the steam pressure. Thus in a short period of time the internal pressure in the containers is switched from being balanced by the external steam pressure to being balanced by an external air pressure, which for both manual and automatic controllers is no mean feat. On switching from heating to cooling the requirement is for rapid and uniform cooling. In practice though the uniformity of cooling rarely matches the uniformity of heating. As cooling progresses the internal temperature inside containers gradually falls and the internal pressure diminishes. When the internal pressure has been balanced by an external air pressure then there is a danger that the external pressure can crush the container; for cans this is called ‘paneling’. It is therefore necessary that a pressure control profile is developed that appropriately matches the internal temperature profile so that neither excessive container swelling nor crushing take place. Once the target cooling temperature has been reached the water supply is switched off and any residual pressure released. The target temperature for product is normally in the region of 40ºC which is a compromise to prevent the growth of heat-loving bacteria that survive the retort process and leave enough residual heat to aid container drying. The drain is then opened and the cooling
Retort technology
11
water released. The door or lid is opened and the crates unloaded. The retort is then ready for the next cycle.
2.3
Selection of container
The selection of a retort system cannot be made independently from the packing medium to be used; so the key features of each of the main packing types used for container sterilisation/pasteurisation are reviewed. Selection of the containers in which to heat process a food is to a large extent a marketing decision, based upon shelf appearance, ease of opening, microwavability and overall product quality. Technologically cans and jars have the advantage that they are well understood. Production lines for new container types, e.g. semi-rigid plastics or pouches, are likely to generate more teething troubles. The transport of plastic containers (trays and pouches) before use tends to be more economic than cans and jars (assuming the food manufacturer does not also make the latter). Trays can be nested, i.e. stacked one within another, for transport, whilst pouches can be transported flat. Both pouches and trays are light in weight for transport. Empty cans and jars are bulky to transport, and glass containers have a relatively poor ratio for container weight to product. Cans have inherent strength to protect them from a degree of mishandling, both before and after filling. Plastics can be prone to puncture by handling equipment and even from sharp food components on filling, e.g. bones. Glass is fragile and presents a hazard if fragments contaminate the product; therefore plants must be operated under close supervision to minimise the potential for product breakage and product contamination. In processing and distribution, cans and jars can be stacked on top of one another with little damage, whilst pouches and trays tend to require outer packaging, e.g. board boxes, to protect them and allow stacking, though they will still only withstand relatively low stacking heights. Filling operations for cans are faster than for glass and plastic containers. Pouches in particular are difficult to open and fill at speed, without contamination of the sealing area. Likewise sealing operations are faster for double seamed cans than heat sealed pouches or jars. The shelf-life of product packed in cans/jars is theoretically longer than pouches/trays if the latter do not contain a complete oxygen barrier. Some plastics and glass are poor ultraviolet barriers so the detrimental effects of storage in light, during distribution and retail, must be considered. Cans and plastics can both contribute to product flavour development; glass is more inert.
2.4
Selection of a retort
As with all aspects of designing a production plant the selection of a retort system will be based ultimately on process economics. The main factors affecting the economics of a retort process will be:
12 • • • •
Thermal technologies in food processing process throughput energy efficiency product wastage retort life.
The rate of production required will play a part in the decision on whether batch or continuous retorts are selected, the latter offering higher rates of production. However, where an installation must cope with regular changes in container specification/process time requirements either because the plant has a wide product range or develops many new products, continuous retorts can lack flexibility. Hence in many factories both continuous and batch systems are installed. The life of a retort system is largely dependent upon the material of construction and the treatment chemicals added to cooling water. Many older retorts are constructed of mild steel which is prone to corrosion especially if corrosive chlorine-based compounds are used for disinfection of the cooling water. In the absence of such chemicals the author has observed mild steel retorts in operation fifty years after their construction; however, using untreated water has product safety implications if the water is not from a source of good microbiological quality.2 With the newer generation of retorts with more complex computer control systems it makes little sense to attach such an investment to a vessel that will disintegrate in 5–10 years, so they tend to be constructed from stainless steel. 2.4.1 Temperature requirements The process time/temperature requirements for a particular product can be predicted from a relatively small amount of experimental data. This data for products is usually expressed in terms of a heating factor fh, or thermal diffusivity for the product (formula and container size) in question. This information, which may exist from previous experimental trials or pilot trails upon a new product, is required as data input into finite difference models that allow prediction of product heating throughout a simulated retort heating profile. For example, these programs allow predictions based upon infinite combinations of container dimensions and differential surface heat transfer coefficients, e.g. for prediction of heating in a glass jar with a metal lid. From these predictions it is possible to determine approximate retort programs (times/temperatures) in order to achieve the desired level of microbial destruction. Once a time temperature profile has been resolved the entire installation can be planned based upon the plants throughput requirements. Target microbial destruction levels are usually expressed as an F0, P value or time above a specified temperature. Selection of process targets should be done in consultation with a specialist microbiologist and a specialist in taking temperature measurements in products.3, 4 Retort temperatures for sterilisation processes are generally between 110 and 130ºC, a temperature range that gives
Retort technology
13
acceptable rates of microbial destruction while not presenting excessive risks of product burn. Sometimes temperatures are limited by specific characteristics of the packaging, for example welded seals on pouches soften significantly as the retort temperature increases and printing in lithographed cans will have an upper temperature limit. It is worth noting that containers/closures, which look similar, may be specified for use in pasteurisation or sterilisation processes only. 2.4.2 Pressure requirements Selection of a retort pressure profile is as important to process safety as the retort temperature profile, because although the latter destroys the product’s microbial loading, minimising mechanical stressing of containers is essential to prevent recontamination. Pressure profiles can be determined from experimental trails by two means: measurements of internal/external pressures or by direct measurements of pack distortion. As with measurements of product heating, consistent results from pressure measurements in containers (or deflections) will only be obtained if pack preparation is well controlled. The following factors can contribute to pressure conditions developed inside containers during retort processes: • • • • • •
pack vacuum at sealing the product temperature at sealing the headspace size (the space above product in a container) the process temperature the product formulation the quantity of gas entrapped in the product.
Typical overpressures for different container types are: Plastic trays 2.3–2.9 kg/cm2 (32–40 lb/inch2) with a tolerance of 0.1–0.2 kg/cm2 Pouches 1.0 kg/cm2 (15 lb/inch2) with a tolerance of 0.1–0.3 kg/cm2 Plastic cans 2.9 kg/cm2 (40 lb/inch2) with a tolerance of 0.4–0.5 kg/cm2
2.5
The influence of heating medium on retort performance
2.5.1 Batch retorts The use of batch retorts allows far greater flexibility than continuous retorts, particularly if the batch system has a capability for overpressure throughout the heating/cooling process. The use of overpressure batch retorts, using either full water immersion, raining water or a steam/air mixture for heating, allows independent control of temperature and pressure during the heating process. This contrasts with saturated steam retorts where the process pressure is directly related to the chosen holding temperature. In the overpressure retorts process pressure conditions can be established that minimise pack damage. Therefore, if jars,
14
Thermal technologies in food processing
semi-rigid trays or pouches are to be produced, overpressure retorts are generally used. Selection of specific retorts should also consider the sensitivity of the pack(s) in question to the rapid change in pressure or temperature that are inherent in some designs. For batch retorts, new container designs may be accommodated in existing retort crates or, if necessary, alternative crate designs may be used. The loading of containers into the retort crates should be considered at an early stage of process design because retort performance can be adversely affected if the crate loading is too dense. If the loading pattern needs to be adjusted at a later stage, to achieve acceptable retort temperature distributions, original estimates of retort throughput will be reduced. The support and orientation of containers within crates is also important as it affects retort temperature distribution and product heating. For example, a semirigid tray which heats at different rates through its lid and base, will give a different heat penetration performance if the pack is heated lid up or lid down, because of the insulation effect of the headspace on heat transfer through the lid. Product appearance, as seen by the consumer when the pack is opened, may also play a part in determining the required container orientation during retorting. Some plastic containers require specialised support racks to ensure adequate temperature distribution and pack performance. It should be noted that during retorting, plastic containers can soften and change shape, their orientation and support during processing will affect the extent and nature of this deformation. The plastic containers (trays and pouches) do have one potential advantage for product quality because as they are generally of thinner profile than cans they heat more rapidly (so process safety requirements can be achieved with a minimum degree of overcooking). Preliminary testing to ensure that any proposed combination of crate, container, pack arrangement, and layer divider can achieve acceptable temperature distribution performance is advisable. Purely practical considerations when planning the crate loading operation are the degree of automation achievable, the stability of containers, i.e. will they fall over when the crate is moved, and the potential for puncturing of containers during loading/unloading (a particular problem with pouches). Saturated steam retorts The basic operating principles of a saturated steam retort are covered in Section 2.2. A high number of saturated steam retorts are installed when production requires a large throughput of varying canned products. Installations may be in the order of 40 retorts, normally of the vertical type to save space. In such large installations the practice of venting significant quantities of steam to the atmosphere is problematic as the working environment becomes very hot and humid. On a practical level the presence of steam in the retort room can be controlled by the installation of a vent manifold to carry the steam outside the building. However, careful design is required to ensure that this does not inhibit the flow of venting steam.
Retort technology
15
Steam/air retorts These types of retorts utilise the deliberate mixing of steam and air in the retort vessel to provide an overpressure environment suitable for processing pressure sensitive containers such as jars, pouches or semi-rigid trays. Typical examples of steam/air retorts are those supplied by Lagarde and Panini. The principle of the steam/air retort is totally contrary to the traditional saturated steam retort where air removal is regarded at a necessary precursor to starting the hold period. In a steam/air retort a proportion of air is retained, or introduced, to balance the internal pressures within containers undergoing sterilisation. Clearly this principle will not give uniform sterilisation unless it is ensured that the mixture of steam and air is uniform, as the presence of large pockets of air will result in under-sterilisation. It should also be noted that as steam condenses, the air portion is left potentially forming an insulating layer around the packs to be heated.5 In commercial retorts this mixing and breakdown of the residual air layer is achieved by the use of a large fan which draws the steam-air mixture through the load and recirculates it to the opposite end of the machine (such retorts are almost universally horizontal) (Fig. 2.2). It has been suggested that the mixing could be achieved by continuous venting, but this is not used in commercial systems and is unlikely to be energy efficient. The temperature uniformity of such steam/air systems relies heavily upon the correct operation and maintenance of the fan and circulation systems. Fan failure or damage would be regarded as critical process deviations. There is debate as to whether the achievement of temperature uniformity in steam/air systems is best achieved by a full vent, followed by reintroduction of a portion of air, or simply modifying the ratio of steam to air progressively at the start of the process. It is also worth noting that for both a saturated steam retort and steam/air retorts rotary processing (see page 21 below) may not always be helpful in removing air from the centre of crates. Like saturated steam retorts, steam/air retorts are potentially prone to steam collapse on the introduction of cooling water, which is particularly unhelpful in systems used for pressure sensitive packaging types. Retort manufacturers have addressed this issue by including a precool stage in the operating cycle. This precool stage involves the initiation of the cooling process using condensate collected from the base of the retort vessel and/or introduction of very small amounts of cold water. By this means the steam contents of the retort can be condensed in a relatively slow and controlled manner while the air content is increased. Full water immersion Probably the oldest mechanism for overpressure processing is to process containers under water with an overpressure applied to the free space above the water in the retort vessel (Fig. 2.3). The overpressure in the free space can be varied to give the required control over container deformation whilst the water can also be superheated above 100ºC. Typical machines of the full water immersion principle are those produced by Stock and Lubeca.
Fig. 2.2 Schematic of a steam/air retort.
Fig. 2.3 Schematic of a full water immersion retort.
18
Thermal technologies in food processing
Given that the volume of water inside full water immersion retort systems is large, heating this amount of water could be a very slow process giving a long come-up time to the desired sterilisation temperature. A means of overcoming this problem, which also adds to the energy efficiency of the system, is to have a second vessel in which the water is preheated above the desired sterilisation temperature. When the process is started this water is then pumped or dropped under gravity into the retort vessel containing the product. Although there may be some temperature drop in the water as it is transferred this method significantly reduces the retort come-up time and increases throughput. This approach must be used with care where jars are being processed and a large temperature differential exists between the product and the incoming water, as thermal shock can result in breakage. At the end of the holding period the water can be pumped back to the storage vessel for use on the next batch, thus saving on the energy required to heat water. Cooling is carried out with an external cooling water supply. A side effect of the double vessel water immersion retort process is that the water capacity of the storage vessel is usually matched to the requirement for a fully loaded retort. Therefore if part loads are processed there is insufficient water. To overcome this problem the retort manufacturers supply dummy crates whose function is simply to replace the missing crates of product. In water there is a natural tendency for convection currents to develop which will make the top of the retort hotter than the bottom, and therefore there will be different levels of product sterilisation at each position. This problem is overcome by different means by different retort manufacturers. The more sophisticated systems use an external water circulation loop on the vessel, so that water is pumped from the colder regions of the vessel through a steam injector and back to the retort. Where possible this mixing process is combined with rotary agitation of the load, which further aids the mixing of the water. It is not uncommon to find vertical full water immersion retorts that can also be used in saturated steam mode. In these systems the agitation of the water is sometimes provided by the use of a cross-shaped spreader which directly injects steam into the retort vessel. These are designed to give good mixing of the steam and water, for example by having two live and two dead quadrants to increase mixing.6 In addition, some systems have a small injection of air into the steam supply that bubbles through the water and mixes it. The headspace above the water can either be filled with pressurised air or steam. The use of steam has the potential advantage that process deviations due to abnormally low water level are unlikely to result in gross understerilisation. It is generally recommended that the water level in such systems should be kept at least 10 cm above the topmost containers in the process. Some packs will have a tendency to float during full water immersion processes; where this is not desired a suitable restraint must be put in place, e.g. tops on crates. There are circumstances where this buoyancy effect is beneficial, e.g. semi-rigid trays and pouches soften during processing at high temperatures; without support this can lead to permanent deformation under the effect of gravity. In water immersion this tendency for packs to sag is minimised.
Retort technology
19
Raining water/sprayed water retorts Typical raining water retort designs are those of Barriquand and Prominox. Sprayed water retort systems are made by FMC, Radabe and Surdry. It is difficult to say whether the raining water and sprayed water systems are directly comparable in the physical mechanisms by which they transfer heat to the load. The raining water principle is simple: water is sucked from the trough at the bottom of a horizontal vessel, passed through a heat exchanger, and pumped to the top of the vessel where it is dumped at high velocity onto a sieve plate in the top quadrant of the vessel (Fig. 2.4). The water is then distributed under gravity over the sieve plate and runs down through the holes onto the load below. Sprayed water systems operate in a similar manner except that the water rather than being put onto a sieve plate is put into one or more spreaders which run down the length of the vessel. These spreaders have spray nozzles along their length which spray water into the load from the top (and sometimes sides). A potential weakness of the sprayed water systems compared with raining water is the natural tendency for the water pressure to drop along the length of the spreader meaning that water coverage is less even. The relative roles of the perforations in the ‘stair rod’-forming sieve plate and mist-forming nozzles on the ability of the water to carry heat to the load is unclear. The directional nature of the water input means that both types of systems are believed to be affected by what are referred to as ‘umbrella effects’ where the penetration of heat into some containers is slowed because the water flow impacts on a container above.7 The occurrence of such effects though appears to be highly dependent upon load content and layout. In practice the heat transfer rates in raining water systems seem to be comparable to those from saturated steam. Although the use of a heat exchanger in the recirculation loop is common on this type of system other alternatives are possible such as • a heat exchanger in the water trough • direct steam injection into the water trough • steam injection into the process vessel.
Clearly for any of the raining water or spray water systems correct maintenance of the water circulation system is critical for ensuring uniform sterilisation is achieved. This means that when water contamination is likely due to overspill at container filling, or pack failure during the process, regular cleaning is required. The recirculation system usually includes a filter that must be routinely inspected and emptied. The US Food and Drug Administration (FDA) like to see data proving that the water flow rates inside these retorts are uniform. Ironically one of the niche markets for raining water systems is the industry canning oily fish where these types of retort effectively double as can washers. One of the benefits sometimes claimed for these systems is that the load is heated and cooled in the same water (the external side of the heat exchanger is
Fig. 2.4
Schematic of a raining water retort.
Retort technology
21
switched from steam to cold water). In theory this means that the water should be free from microorganisms and therefore is an excellent means of preventing post process recontamination. However, post process recontamination spoilage incidents have occurred which could be attributable to leakage in the heat exchanger, dead-legs (regions of low flow) in the recirculation loop building up contamination or contamination of the air supply. Crateless retorts These retorts are large vertical vessels with doors at either end. During filling the top door is opened and cans feed under gravity from a conveyor. The cans fall to the bottom of the retort and are cushioned on their way down by water. Once the retort is full the top door is closed and the water is flushed out by steam, and the process hold commences. When the hold is finished, the load is cooled with water, drained and the bottom door is opened. The cans then fall out into a cooling canal. Batch rotary retorts Increased retort throughput can be achieved for products that undergo forced convection by using rotary batch retorts. Such rotary agitation can be applied to any of the heating media described above, e.g. steam, steam/air, raining/sprayed water or full water immersion. Rotary processes, which agitate the product inside the container, are often used in conjunction with higher process temperatures than would be used for the same product in a static process. The higher temperature is less detrimental to product because the movement prevents overheating, particularly at the container wall. Two modes of agitation can be employed; these are referred to as ‘axial’, in which a can spins on its own long axis, or ‘end over end’ in which the can is tipped over. Commercial batch retorts generally use end over end agitation, as this is more suited to easy loading and simple design of the crate system. A very important factor in the effectiveness of agitation in bringing about product heating is the size of the headspace (the free space in the top of the container after filling) which plays a big part in the mixing process. In rotary retorts the product within the crates must be clamped in place preventing damage from the movement during rotation. If a range of container sizes/types is to be put through the same retort, then consideration will be required as to how the clamping mechanism will cope. Some containers are not suitable for rotary processing, as they do not perform well when clamped, e.g. the sealing compound in jar lids can be cut as a combination of the softening in the heat process and the pressure of clamping. Some products are not suitable for rotary processes, for example where the product texture is adversely affected, e.g. cream or soft fruit. For those products that heat rapidly by natural convection, e.g. thin soups or vegetables in brine, the benefits of rotary processes are marginal.
22
Thermal technologies in food processing
2.5.2 Continuous retorts Hydrostatic retorts Hydrostatic retorts (Fig. 2.5) utilise a water lock to transfer conveyed containers into a pressurised process vessel. The incoming water ‘leg’ can contain a controlled temperature gradient for optimal preheating of the incoming containers. Likewise temperature gradients in the cooling ‘leg’ can be optimised. The process chamber is normally filled with saturated steam but steam/air mixtures are sometimes used. Current manufacturers of hydrostatic retort systems include Stork and FMC. The process applied is determined by the temperature in the process chamber (which is limited by the size of the hydrostatic legs) and the speed of the conveyor system. Hydrostatic retorts have historically been used for production of cans, glass bottles and plastic bottles, though this situation is changing (see Section 2.6). The loading mechanisms and orientation of the containers tend to differ, cans are loaded from a conveyor at right angles to the retort conveyor, while bottles are fed into pockets from the direction of travel of the retort conveyor. This configuration avoids pushing the relatively fragile glass/plastic containers against each other. Hydrostatic retorts are designed to work with a limited range of container sizes, depending upon the carrier bar diameter. Some systems have two sets of carrier bars of different diameters on either side of the conveyor chain. Those retorts using the bar mechanism can cope with variations in can heights without much difficulty, e.g. promotional packs, the pocket systems are less flexible.
Fig. 2.5 Schematic of a hydrostatic retort.
Retort technology
23
A small amount of product agitation is imparted during the change of direction on the conveyor of a hydrostatic retort, but this does not dramatically affect product heating. Some hydrostatic retorts are designed with a planetary motion of the carrier bars to enable high temperature short time processing, e.g. for dairy products. Care must be taken in the design of hydrostatic retort installations (and any other continuous retort) to ensure that there is no possibility of the unprocessed containers jumping from the conveyors feeding the retort to those taking product away. With hydrostatic retorts there is also the specific issue that the water used in the preheat leg should not be in direct contact with the outfeed leg because cross-contamination from fill overspill can result. Hydrolock retorts These are retorts that operate on a similar principle to the hydrostatic retort except that the hydrostatic seal contains a mechanical element so that the ‘legs’ do not need to be so high. In fact the machines are like a hydrostatic retort operated in a horizontal orientation. Some of these machines have the unusual feature that during some of the conveyed distance inside the steam chamber the carrier bars become free to roll imparting extra agitation to the product. Reel and spiral retorts Despite its mechanical complexity the reel and spiral retort (Fig. 2.6) is, in fact, an old design dating from the 1930s. Current manufacturers of reel and spiral retorts are FMC, Stork, AMC and Molenaar. The principle of operation is that cans (the system can only be used for cans) are fed from a conveyor system, twisted onto their side, so that they roll freely and then passed through a star valve directly into a pressurised retort vessel. Each can fits into a pocket between the points of the star, which as it turns moves the can from outside to inside the retort. Once inside the retort vessel, which is a horizontal cylinder, the cans are fed onto an internal spiral track in the inner wall of the cylinder and pushed along by blades attached to a reel rotating in the centre of the vessel. The spiral nature of the track means that the cans move from one end of the cylinder to the other, and when they are in the bottom third of the cylinder they rotate freely on the bottom wall. At the end of the cylinder cans are passed out through another star valve and either exit the pressurised environment or are passed to another ‘shell’ (a similar retort vessel). Where such transfer values are used for moving cans between heating shells, the valves are fitted with a steam supply to ensure that cans in the valves are exposed to the intended process temperature and pressure conditions, especially during stoppages. The configuration (number and types of shells) of reel and spiral retorts varies depending on the type of product to be processed. The process duration is determined by the length of the shells, the number of shells, and the rate of rotation of the reel (process time and reel speed cannot be controlled independently). For example, a reel and spiral retort processing a dairy product
Fig. 2.6
Cut-away diagram of an FMC Reel and Spiral retort.
Retort technology
25
might have a preheat shell operating a temperature well below sterilisation temperature, one sterilisation shell and a cooling shell, while a machine for cans of baked beans might have two sterilisation shells and two cooling shells (to allow for the longer hold and cool periods required). The shells are generally operated with steam or water for heating and water for cooling. Reel and spiral retorts rely on the movement of containers rotating on their sides, which limits their use to cans. There is little flexibility in can size that can be processed, with machines being built for one can diameter and a limited range of heights. Agitation is inherent in the reel and spiral design enabling high temperature short time processes so they are not suitable for product sensitive to mechanical action, e.g. strawberries. For many years can rotation rates in reel and spiral retorts have been assumed but more recent electronic can rotation counters have become available. These devices can be put through the retorts to ensure that the theoretical rotation rates are being achieved. Failures in can rotation have been attributed to track wear, bowing of the reel and build up of deposits of lacquer removed from the rims of cans as they rotate. There is evidence to suggest that changes in can specification can alter rotation rates and therefore product heating.8 One of the greatest challenges in recent years for operators of reel and spiral retorts was the introduction of the easy open can end. This was because during the heating processes the internal pressures in the cans causes the ends to dome slightly. With an easy open end this doming can cause the ring pull to stick out to the extent that it can catch on the spiral tracks inside the shell, with the consequence of turning the retort into a very large can opener. However, this difficulty has now been largely overcome.
2.6
Future trends
2.6.1 High temperature short time (HTST) processing Trends in the future will to some extent be extrapolations of patterns over the last few years. Therefore we can expect to see a continued increase in high temperature short time processes, generally aided by rotary processing to minimise product degradation at the container surfaces. The science of incontainer mixing is not as well developed as the technology, and research is under way to optimise process conditions to enhance the mixing processes taking place. This might, for example, be achieved by study of the headspace movement in products or simulants of equivalent rheological properties at the intended process temperature which can determine when the headspace is most effectively passing through the container to bring about mixing.9 If the mixing process is well understood then we can expect benefits in terms of reduced process times and/or improved product quality. An interesting recently patented development by Crown Cork & Seal has been the use of shaking type agitation to increase product heating rates.10 The patent covers ranges of reciprocating acceleration movements, and these were
26
Thermal technologies in food processing
applied in an especially designed shaking retort system. Results seem to indicate that for several traditional canned products heat transfer rates can be greatly improved even in comparison with rotary agitation. The mechanism for this extra efficient agitation is presumably the fact that greater turbulence is introduced into the movement of fluids in the packs compared with rotary motions. One possible drawback of the approach is that the shaking motion may damage tender food components but trials using asparagus have indicated that this is not the case. 2.6.2 Flexibility in packaging formats As the marketplace increasingly demands innovation, flexibility of retort systems, particularly in relation to packaging format, is becoming increasingly important. For example, it seems that there are forces developing which will drive canned food manufacturers, especially in the catering sector, away from cans toward packs that are more space efficient in disposal, recyclable and less likely to contaminate product on opening. It seems that the heat processable pouch is becoming the natural alternative to the catering can. For batch retort systems a change to a new pack format may mean a simple, though not cheap, change of racking system. With batch systems there is an increasing move toward automatic retort loading and unloading which yields long-term benefits in reducing labour costs, as retort operators are commonly supported by a team of loaders. Second, automation of the loading system can be used to control the flow of product through the factory. If implemented correctly this automated control of the loading operation can be one of the mechanisms used to prevent unprocessed product bypassing the heat process, which is one of the major safety risks associated with any such operation. In some installations this control is enhanced by the use of double doors, one at either end of the process vessel, with a wall built to prevent product getting from one room in the factory to another without going through the retort (though this does not necessarily guarantee that a process is applied). However, for the continuous retort systems, especially reel and spiral retorts, changes in packaging format have been more difficult. Stork have introduced their new Vario hydrostatic retort system which uses a cassette which allows a range of packaging formats to be processed through the same retort. These cassettes are the basic unit which is transported through the retort, within which, depending on the internal construction, any pack can be processed, e.g. pouches, glass jars, cans. 2.6.3 Environmental issues It seems likely that the retorting industry, like many others, will in the near future have to take active steps to reduce energy usage and maximise water recovery. Most of the modern overpressure retort systems already incorporate energy and water efficiency features but there may be further developments.
Retort technology
27
2.6.4 Intelligent control One of the most problematic aspects of controlling a retorting operation is dealing with product that has not received the intended thermal process. Such situations are described as process deviations and may result, for example, from failure in services, e.g. boiler breakdown. Historically deviations have been dealt with either during the process from tables prepared from experimental data or from experimental data generated after the problem has occurred. More recently, computational modelling methods have been used to predict off-line the effect of time temperature deviations based on known product heat transfer characteristics. Commercial programs of this type have existed for several years, e.g. CTemp from CCFRA and NumeriCAL developed by Technical. It is a logical step that in order to minimise the product lost from deviations and to minimise safety risks, this kind of mathematical model should be used on-line. Such predictive code has been included by FMC in their Log-TechTM batch retort control systems. The application of such heat transfer models to continuous processes is more complex because the deviation will have a different impact upon containers at different points through their residence time. However, this is a likely development.
2.7
Sources of further information and advice
The most detailed advice on each type of retort system can be obtained from specific retort manufacturers. General guidance on good manufacturing practice is available from research organisations specialising in heat sterilised foods such as: Campden & Chorleywood Food Research Association CCFRA (UK) National Food Processors Association (NFPA) (USA) CTCPA (France) TNO (Netherlands) KIN (Germany). Information on packaging systems can be obtained from packaging suppliers or the Metal Packaging Association.
2.8 1. 2. 3.
References CAMPDEN & CHORLEYWOOD FOOD RESEARCH ASSOCIATION,
‘Canning Retorts and Their Operation’. 1975 Technical Manual No. 31. THORPE R H, ‘Leaker spoilage of foods heat processed in hermetically sealed containers’. In Heat Preserved Foods, eds J.A.G. Rees and J. Bettison, Blackie and Sons Ltd. 1991. CAMPDEN & CHORLEYWOOD FOOD RESEARCH ASSOCIATION, ‘Guidelines for performing heat penetration trials for establishing thermal processes in batch retorts’. 1997 Guideline No. 16.
28 4. 5. 6.
Thermal technologies in food processing CAMPDEN & CHORLEYWOOD FOOD RESEARCH ASSOCIATION,
‘Heat processing of low acid foods: an approach for selection of F0 requirements’. 1998 Review No. 9. TUNG M A, BRITT I J, RAMASWAMY H S, ‘Food sterilisation in steam/air retorts’. Food Technology 1990 December 105. NATIONAL FOOD PROCESSORS ASSOCIATION/THE FOOD PROCESSORS INSTITUTE,
7. 8. 9. 10.
‘Canned Foods – Principles of thermal process control, acidification and container closure evaluation’ 6th edition, Washington DC. CAMPBELL S, RAMASWAMY H S, ‘Distribution of heat transfer rate and lethality in a single basket water cascade retort’. Journal of Food Process Engineering 1992 (15) 31. ZAMAN S, ROTSTEIN E, VALENTAS K J, ‘Can material influence on the performance of rotating cookers’. Journal of Food Science 1991 (6) 1781. EISNER M, Introduction into the technique and technology of rotary sterilisation. Private Authors Edition, Milwaulkee, Wisconsin, USA 1988. THOMPSON S, ‘Future trends in thermally processing foods’. CCFRA seminar: Heat processed foods: new insights and approaches, Nov. 1999.
3 Continuous heat processing S.P. Emond, Campden & Chorleywood Food Research Association, Chipping Campden
3.1
Introduction
An improvement in quality was one of the main driving forces behind the development of continuous heat processes. Liquid and semi-liquid products such as milks, juices and sauces suffered from overprocessing in the traditional low temperature–long time of in-container or batch processing. Caramelised flavours, poor colour retention and a lack of a reproducible product were all problems associated with products processed by batch methods. Improving quality whilst maintaining product safety was the main aim for those developing continuous processing approaches. The achievement of safe products by thermal processing is based upon the theory behind the destruction of microorganisms. Products must be heated to a set temperature for a set time in order to achieve a commercially sterile product. For continuous heat processing, also called continuous flow processing, the product is thermally processed before being placed into an appropriate container, on a continuous basis through a heat exchange plant. Heat exchange apparatus will be used for both the heating and cooling (if required) phase of the process. In a continuous system the foods under consideration are liquid or semi-liquid products, which may be pumped through a system, heated and cooled whilst continuously flowing down the processing line. A wide range of products are processed by this method, either as the main process to achieve a safe product (as in Ultra Heat Treated or Ultra High Temperature (UHT) processing) or as a step within a further process. Continuous heat processing is not a new technology and several good texts exist which give background to the developments through the years in this area.
30
Thermal technologies in food processing
This chapter will review the current status within processing equipment and highlight the developments which have been made in order to take into account the processor’s requirements for safe, reproducible systems which will ensure a safe product, maintain product quality and also maximise process efficiency. The three main types of process that are suitable for continuous flow processing are, aseptic systems (high and low acid), hot fill systems and pasteurisation processes. Aseptically packed products are processed at temperatures that will render the product commercially sterile. High acid products such as juices can be processed at pasteurisation temperatures to destroy the microorganisms that can cause the spoilage of the product; these are then rapidly cooled (to reduce losses of volatiles within the product) and filled into a pre-sterilised pack under sterile conditions. Low acid products will undergo the same principle, however the temperatures employed are much higher to ensure no survival of pathogenic bacteria. The temperatures used within a UHT system for low acid products are usually in the range 125ºC to 145ºC, so allowing for much shorter holding times and promoting a higher quality product. Continuous flow processing systems can also be used in hot fill processes for high acid products that would otherwise lose product quality through slow cooling methods. High acid sauces, pure´es and chutneys can benefit from a continuous process by heating the product to pasteurisation temperatures and then filling directly into suitable containers, using the heat of the product to decontaminate the packaging. This method allows for a much quicker throughput than a typical batch process would offer. The final heating method for this type of system is pasteurisation of low acid products that will then be cooled and held under chilled conditions (e.g. pasteurised milk, juices and soups). This processing step extends the shelf-life and ensures a safe product. The product must be chilled to maintain its safety and quality throughout the shelf-life. Shelf-lives of up to ten days can be achieved for some products. There are two main options open to a food manufacturer considering a continuous heat process, to process by indirect method, or by direct method. Indirect heating involves a heat transfer surface between the product and the heating media. Direct heating occurs where the product and heating media are in direct contact. Figure 3.1 illustrates the main options open to a processor. Another established continuous heating method available is Ohmic heating. This will be discussed in Chapter 12.
3.2
Indirect heating
Indirect heating methods rely on having a heat transfer surface between the product and the heating media. There are three main types of indirect heating system: plate heat exchangers, tubular heat exchangers and scraped surface heat exchangers. Each system has benefits and drawbacks depending on the product requiring processing and each has been adapted since the advent of these
Continuous heat processing
Fig. 3.1
31
Methods of continuous heat processing.
systems to process a wider range of products and enabling the systems to compete directly. 3.2.1 Plate heat exchangers Plate heat exchangers are a well-established method for processing homogenous products of low viscosity, making them ideal for use within dairies. Plate heat exchangers consist of a series of plates connected on a frame. The product and heating (or cooling media) flow in alternate channels in thin layers to provide good heat transfer conditions (Fig. 3.2). The plates are sealed by elastic sealing gaskets cemented into a perforated groove. Generally the plates are of polished stainless steel of 0.5–1.25 mm in thickness separated by 3–6 mm. The surface of the plates is usually corrugated in order to increase the area available for heat transfer and enhance the turbulence present in the system, resulting in a high thermal efficiency. Thermal regeneration can lower energy costs substantially. The narrow gaps mean that the units are best suited to low viscosity homogenous products. Attempts to process particulate products (e.g. fruit juice cells) may result in blocked channels and eventually blown plates due to the pressure imbalance between product and media sides of the plates. For this reason only products with less than 10% cell content are normally recommended when processing with plate heat exchangers. The design of the plates can vary from supplier to supplier, each having different designs to maximise process efficiency and ensure product safety. Plates can be product specific. The corrugations that are present in plates are usually of a chevron or herringbone design in order to develop a turbulent flow through the plate as it passes through the plate pack (so increasing heat transfer). As the plates are assembled, the herringbone pattern is usually alternated, with the chevrons going up on one plate and down on the next, creating the channels through which the product can flow. The thickness of the plates will vary from
32
Thermal technologies in food processing
Fig. 3.2
Flow through a plate heat exchanger (courtesy of Tetra Pak).
type to type, depending on the working pressure expected from the plates, but these can also be of thin gauge material which will ensure a high heat transfer. The design of the pattern on the plates in most cases allows for support of the whole plate pack, the plates are touching a designated point to ensure that the strength within the system is maintained but also the ease of cleanability is taken into account. The plates may also have larger spaces between, which will allow small particles to be processed in the system. A further development in this area has been with the double-separation plate, which is designed for highest security, stopping contamination between the heating media and the processed product. The design is similar to a traditional plate heat exchanger but plate pairs are mounted together and are welded at the
Continuous heat processing
33
product ports. As the pairs are mounted together, this forms the channels through which the product flows and these are sealed together by an elastomeric gasket. Because of this design, if the system does suffer a gasket failure, the leak will be detected externally and action may be taken. This high level of security makes it ideal for high security operations such as in the pharmaceutical industry. The design used on the plates may also take into account the level of fouling that will occur throughout the process. This is usually the limiting factor in production times and the longer a plant can run for, the less downtime costs there are for the process. The plates should also be designed to be cleaned in place, however, in some cases or during planned maintenance, the plates may have to be dismantled from the frame and cleaned and serviced by hand. The application of the heat exchanger will determine the type of frame that should be used to hold the plates together. In the food industry (for food applications) the frame would normally be hygienically designed either being of solid stainless steel or completely clad in stainless steel. In industrial applications a mild steel frame would be sufficient for the heating of cleaning chemicals or media heating. The plates hang on the frame and are held together by a series of compression bolts which are tightened depending on plate size, thickness and number of plates. In some cases the heat exchangers will be modular, therefore allowing for easy extension of the plate pack, or changing for new applications. The advent of the hanging frame has enabled servicing and inspection, a much easier process than with the early frames where the plates had to be removed one by one. The gaskets that are used to hold the plates together can be either cemented into place, or can be clipped into place. For ease of service, the clip on gaskets ensure that downtime is kept to a minimum whilst still ensuring that the hygienic barrier is maintained. The clips usually work by having two prongs which sit in the gap between two plates; in this way, in combination with the hanging plates, any changes in gasket can be carried out in situ. Plate heat exchangers were traditionally used for pasteurisation processes and have been adapted to withstand the higher temperatures and pressures required for UHT processes. The main difficulty with plate heat exchangers was their tendency to foul followed by inefficient cleaning in place. A build-up of such debris in the streams in the system may ultimately lead to the product being understerilised leading to product spoilage or an unsterile product. The manufacturers of such systems are designing the plates in such a way as to make them suitable for cleaning in place. Producers using these systems should have planned preventative maintenance schemes in place to ensure that there is scheduled servicing and cleaning of the machines before this situation occurs. One of the main advantages of plate heat exchangers is in the regeneration of energy used in the system. Product will pass through three sections in a plate heat exchanger. The first section will be a regeneration section where the incoming product will be heated by the outgoing hot product. The product will then enter the main heating section with the heating media on one side, which could be steam, but would more likely be a steam/water mixture (to try and
34
Thermal technologies in food processing
reduce the level of fouling that will occur on the surface of the plates) to take the product temperature up to process temperature. The final section that the product will pass through is the regeneration zone, this time as the outgoing hot product giving up its energy to the incoming product and so reducing the amount of cooling capability required by the system. Preventative maintenance should take into account plate check for pinholes to ensure that there is no possibility for cross-contamination. Although the regeneration system ensures heat regeneration of over 90%, having a sterile product on one side of a plate and unsterile product on the other can cause problems of cross-contamination or re-infection. To try and reduce the possibility of contamination the regeneration system should be run so that the pressure on the sterilised product side is higher than that on the unsterilised side. It is for this reason that it is very important for such systems to have a sensitive, accurate pressure controller/recorder. 3.2.2 Tubular heat exchangers Tubular heat exchangers will process a variety of products from low viscosity product such as those processed on the plate packs, but can also handle products of higher viscosity that may contain particulates such as soups and sauces. In tubular heat exchangers product is pumped through a tube or multiple tubes, which are fixed inside a larger tube. In the space between the two tubes, heating or cooling media is pumped in counterflow to the product, maximising heat exchange efficiency. The mechanical strength of these tubes allows them to operate at high temperatures and pressures. Turbulence is achieved in the tubes by the velocity of the product and also by the presence of a corrugated surface to improve heat transfer efficiency. The amount of corrugation can be varied for compatibility between product and plant. A more angled corrugation can introduce turbulence without high velocity in low viscosity fluids such as water, juices and dairy products. Smoother corrugations which have more gradual angles and can impart a twist or turning motion, offer gentler handling of particulate and higher viscosity products. The more angled corrugations fill with such products and can reduce the efficiency of the exchanger. The simplest design for tubular heat exchangers is the monotube, basically a tube held within a tube. Figure 3.3 shows the basic design for a monotube. The product flows through the central tube and is surrounded by the outer tube, which contains the heating or cooling media. This design is the most frequently used system for processing particulate products as there are few problems with the particulate matter blocking the tubes and so causing processing problems and pressure build-up. As the heating media surround the product, this type of system allows for very gentle heating of particulate products. A more complex design, the concentric tubular heat exchanger is generally a single pass shell and tube exchanger with the product flowing through the gap between two heating (or cooling) media channels. The tubes tend to be smooth reducing the pressure drop that can occur when processing viscous products.
Continuous heat processing
Fig. 3.3
35
Typical monotube design (courtesy of Tetra Pak).
Concentric tubes have a single channel design where product flows through a tube which is surrounded by a second jacketed tube containing the heating (or cooling) media. Through the centre of the product tube is a further tube which also has the heating or cooling media flowing through it. In this way the product is surrounded by the media thus giving two heat transfer surfaces allowing for a more efficient heat transfer. As there is generally only one tube for the product to flow down this makes the plant easier to clean and to sterilise. Having a minimal effect on the flow patterns of the product, a uniform product quality for viscous products such as fruit pure´es, concentrates and sauces, such as mustard and mayonnaise can be achieved. The gap for the product flow can be designed depending on the application, giving wider gaps for products containing particulates. A third design for tubular heat exchangers is the multitube system (Fig. 3.4). This can be anything from two to several parallel tubes, through which the product flows, surrounded by a casing which contains the heating media allowing the heating media to be between and around each tube. The tubes can be corrugated or smooth depending on the level of turbulence and heat transfer efficiency required. The models tend to be modular in that the heat exchangers can be put in series depending on the level of heat energy required to achieve the processing temperature. This design of heat exchanger can work at high temperatures (approximately 160ºC) and pressures (approximately 6 mPa). A variation on the multitube is the multichannel, consisting of several tubes in tubes allowing the heating media to flow either side of the product channel. This type of set-up allows for a very large heat transfer surface and therefore high thermal efficiency. This design also allows for heat recovery in the form of
36
Thermal technologies in food processing
Fig. 3.4
Typical multitube design (courtesy of Tetra Pak).
product to product heating or cooling, as is achieved in plate heat exchangers. This design is based on narrow channels and is ideal for low viscosity products such as fruit juices. There are several advantages for the processor using tubular heat exchangers. Designs are available to produce a wide product range. They are able to produce particulate product up to 12 mm and maintain the particle integrity and quality throughout the process. One of the main advantages though, is in the very simple designs, which cut down on maintenance costs and downtime. One of the disadvantages with tubular heat exchangers is the tendency to form thermal cracks, due to the changes in temperature that occur in this type of process having hot product on one side of the tube and cold product or media on the other. To overcome this a floating end design may be used; this allows the internal tube bundle to move slightly within the outer shell, as they are not welded together (as in other designs). The floating end configuration also allows for changes in the tube configuration, allowing monotubes to be replaced by multitubes if a multipurpose system is required (e.g. in pilot scale or research units). Tubular heat exchangers can also suffer very high pressure drops in the system (due to the long pipe lengths used in the systems) and this can lead to practical processing problems and issues with recontamination. For example, if the processor is processing a fruit pure´e at the pasteurisation temperature 95ºC, the product will flow easily down the pipe, the viscosity of the product being much lower than at ambient temperatures. When the product is cooled the viscosity will rise again and cause a large pressure build-up in the system. It is
Continuous heat processing
37
for this reason that the pump used in the system must be well designed and able to stand large pressure changes to ensure that blowback does not occur. Tubular heat exchangers can also suffer from fouling and burn on. As the tubes tend to be long, the processor does not have the ability to open and inspect the plant after processing or cleaning so any fouling problems that can occur must be understood and strictly monitored. Finally, though regeneration is possible (currently for low viscosity products only), the maximum that can usually be achieved is 70–75%. 3.2.3 Scraped surface heat exchangers A more complex design than the plate or tubular heat exchanger, the scraped surface heat exchanger offers a way of processing highly viscous product containing particles that traditionally have been processed by the slower, batch operations and enables a higher quality, repeatable product to be produced. The basic design consists of a large tube in a tube (similar to the simple monotubes) with the heating or cooling media on the outer shell. The central processing tube contains a shaft which is connected to a motor and is supported by bearings at either end. The shaft has blades attached, which are designed to scrape the heating surface of the tube as the motor activates the rotation of the shaft. This design is ideal for viscous products as the rotation causes turbulence within the heating chamber, so increasing the heat transfer into the product and second, the blades scraping on the heating surface reduce the build-up of fouling that can occur with such products. Figure 3.5 outlines the main structure of scraped surface heat exchangers. The shell of the heating tube can be chrome plated nickel (due to the high thermal conductivity that it offers), stainless steel, bimetallic or chromed stainless steel, depending on the application for which it is to be used. The shell is usually of a standard diameter and manufacturers offer a range of central shafts (or rotors) for a specific set of conditions to optimise processing. A smaller diameter rotor will give a larger clearance within the heating chamber, therefore allowing for the processing of products with larger particles and also allow for a longer residence time in the heating unit. A larger diameter rotor will minimise the heating channel therefore minimising the residence time but allowing a more efficient heat transfer to occur. This design would be more appropriate for lower viscosity products with only small particles. Each rotor contains a set of blades and there are again several choices for the processor depending on application. The material of the blade should be compatible with the material of the shell of the heating tube. The scraper blades should not cause wearing of the outer shell (for example if stainless steel blades were used with a stainless steel shell) as this can cause damage to the heat exchanger, allow for foreign bodies within the product and make the plant difficult to clean. Manufacturers offer a wide range of materials such as durable plastics (that withstand operating temperatures and conditions) which are a good choice as they minimise the damage that can be done to the heat exchanger.
38
Thermal technologies in food processing
Fig. 3.5
Structure of a scraped surface heat exchanger (Contherm) (courtesy of Tetra Pak).
Stainless steel blades are also available for some specialist operations. The blades are attached to the rotor in different ways; the first is a ‘floating’ configuration, which enables blades to be easily attached to the rotor. As the rotor rotates through the product the blades scrape at the heating surface throughout the chamber. For very viscous products, some manufacturers offer different designs. The oval tube unit combines an oval shell containing a round rotor. By using this format the blade angle changes as the shaft rotates with product forced out from under the blade as the angle closes and moves under the blade as the angle opens. This method stops any mass movement that may occur in very viscous product as the shaft rotates. The second design is a spring-loaded
Continuous heat processing
39
blade which enables blade contact with the heating surface at slower rotation speeds (50 rpm). A different available design is a Multiscrape unit. Using multiple tubes within one shell, the scraper blades are mounted along a reciprocating piston. This reciprocating action enables efficient heat transfer into the product with minimal product damage. The rotor is sealed at each end by a rotary mechanical seal which is steam flushed to provide an aseptic seal (where necessary). The heating or cooling media for scraped surface heat exchangers can be brine, water, steam, freon and in some cases, oil, which can then achieve temperatures of up to 315ºC. Scraped surface heat exchangers can be installed in either the vertical plane or the horizontal plane. The vertical format can save floor space depending on plant design and so may be more advantageous in plants where there is little available space. Vertical designs now incorporate hydraulic units for the removal of the central shafts for servicing operations which makes for rapid inspection and maintenance. The vertical design also has a further advantage in that as the product enters the heating chamber from the base and travels upwards to the exit, this allows for effective purging of any air that may be present in the chamber so ensuring efficient heat transfer through the system. The air in the heating chamber will be transferred to the holding tube. Therefore this should be angled to allow purging of the air ensuring the correct volume in the holding tube is maintained. Horizontal scraped surface heat exchangers benefit from an equal loading on both bearings within the system (instead of all of the loading on the lower bearing as in the vertical designs). As the associated pipework within the system can be connected to either the front or the rear connectors, benefits can be made from having reduced pipework designed to optimise the system within the production plant. The advantages of the scraped surface heat exchanger system are that it can process very viscous product with particles up to 15 mm3. The process allows for reduced burn on (which would occur with this type of product) allowing for longer production runs. A lower pressure drop is seen in this type of system therefore reducing the problems with leakage and re-infection during processing. The main disadvantage of scraped surface systems is the cost, being expensive to set up and expensive to maintain in comparison with other types of exchanger. New blades and seals are required regularly, and bearings have to be replaced periodically. New tubes may also have to be replaced occasionally. More floor space is required for scraped surface plants than for other types of exchanger. Finally due to the high shear mixing that occurs during the process there can be damage to some fragile particles.
3.3
Direct heating
There are two main methods for heating product by direct methods: by injecting pressurised steam into product or by injecting product into steam (steam
40
Thermal technologies in food processing
Fig. 3.6
Typical heating curves for in-container, indirect and direct heating mechanisms.
infusion). Both systems work on the principle that as the steam comes into contact with the product it will condense and give up some latent heat so causing the product to heat up very quickly. Both methods give practically instantaneous heating, as opposed to the indirect methods or by the very slow batch or incontainer methods. Figure 3.6 compares typical heating curves for products heated by direct, indirect and batch processes. The basic principle for both systems is to pass the product from a balance tank to a preheating system, usually by a plate heat exchanger to 70–80ºC (for a UHT process). After this the product will pass through the main product pump through to the steam injection or infusion system. After holding the product for the required amount of time, the product passes through a reducing valve into the cooler. The cooling mechanism for such systems must also be quick and must remove additional water which will have been added by the heating method. To achieve this the product is passed into a vacuum chamber which will be held at a specific pressure corresponding to the product temperature before heating so causing the product to instantaneously boil and drive off the excess fluid whilst reducing the temperature to the preheat temperature before injection. The products processed by this method are then usually homogenised before cooling by indirect methods to ambient or chilled temperatures by indirect heat exchange (tube or plate). 3.3.1 Steam injection Steam injection is most suitable for low viscosity, homogenous products such as milk and juices. There are many different types of steam injector available, with probably the most varied designs of all the heat exchangers. There are both static and dynamic injectors and the methods of introducing steam into the product are
Continuous heat processing
41
aimed to heat the product as quickly as possible, to minimise any possible fouling by indirect heating methods (that is, from the surface of the injector itself) and to maximise the length of production run that can be carried out by the processor. The first requirement is thought to be maximised by the rapid condensation of the steam so rapidly heating the product. The product must be maintained at high pressure to prevent boiling at the injector; this is achieved by having a sufficient backpressure valve in the line before the injector system. The heating is also maximised by ensuring that the steam has good access to the product if the steam is introduced as small bubbles, or as a thin sheet. Although fouling is less of a problem in direct systems than indirect systems, there are still problems seen when the injector is poorly designed. As the steam will be injected through a nozzle or system of orifices, if this is not kept separate from the product the injector system will start to heat the product indirectly from the surface of the nozzle. This will cause fouling at the surface of the injector and therefore reduce the steam flow, so reducing the efficiency of the heating. To overcome this, manufacturers have designed systems that have the product entering the heating zone at right angles to the injector; this keeps the product away from the injector head and reduces fouling. After heating, the product passes through a reducing valve before entering the flash vacuum cooler which is held at a pressure suitable to cause the product to boil and drive off the added water from the steam injection. Control of the pressure in this system is needed for two reasons. First, this is to ensure that the correct amount of water is driven off, and second, that any undesirable volatile odours that may have developed during the injection process are removed. The flash vacuum chamber does bring its own problems however, as operating this at pressures lower than atmospheric pressure may allow contaminants to be sucked into the chamber. In order to reduce this, aseptic systems apply steam barriers to connecting joints moving seals and valves so that any material that may be sucked in will be sterilised before re-infecting both plant and product. Steam injection can be quite damaging to products due to the aggressive method of heating taking place. The rapid condensation of the steam causes changes in pressure in the liquid. In milk products this is thought to break down homogenisation of the product and can cause the formation of casein aggregates giving a chalky mouthfeel. In order to reduce this, it is necessary after direct steam injection to homogenise (aseptically) such products. 3.3.2 Steam infusion Steam infusers were designed to give a gentler process than that achieved by steam injection. Suitable for the same types of product, infusion systems can also be modified to handle small particles (such as cells in juices). The basis for this method is to introduce preheated product into a chamber containing steam at the temperature required for sterilisation. The steam chamber is usually a pressure vessel with a conical-shaped base through which the heated product falls into the holding section. Different types of infusion systems are available
42
Thermal technologies in food processing
and can be designed to allow for the direct processing of products containing small particulates. Other systems have varying numbers of diffusion holes through which the product is passed before entering the infusion chamber; the holes can be closed off depending on the heating capacity required by the plant. One of the main areas that must be strictly controlled within the infusion system is the holding period between heating and cooling. As the product, once heated, falls to the base of the pressurised vessel, the amount of time taken to travel to the cooler is dependent on level controllers which are connected to the product feed pump to ensure that a sufficient quantity of product is entering into the chamber but is also leaving so ensuring a sufficient holding period and reduced overheating of the product. A further problem with infusion systems is that on entering the heating chamber any dissolved gases in the product will leave the product (as the solubility of gases is lower at sterilisation than at atmospheric temperatures), which cause the mixture in the heating chamber to change from saturated steam to a steam/air mixture. This can have an effect on the efficiency of the heating and to overcome this very often the steam pressure in the system must be increased gradually to ensure the process temperature can be achieved. For both steam injection and infusion systems one critical part of the process is the steam quality. As the product and steam come into direct contact the steam must be of culinary standard. In the UK any producers should ensure that the steam adheres to the Food Safety (General Food Hygiene) Regulations 1995, specifically the requirement that any steam that comes into direct contact with food should be produced from potable water. Potable water means water which at the time of supply is or was not likely in a given case to affect adversely the wholesomeness of a particular foodstuff in its finished form. The steam should be dry saturated, oil free and should not contain volatile substances that can be carried from the steam into the product. The steam lines associated with direct heating systems should be stainless steel to prevent rust build-up and should also be filtered to remove any particulate matter.
3.4
Holding section
The holding section in a continuous system is the part of the plant where the product receives the cook and sufficient destruction of microorganisms to ensure that a commercially sterile and hence safe product is achieved. Usually a series of pipes and bends of known volume, product flows through the holding section at a known flow rate which allows the holding time from the entrance to the exit of the section to be determined. Temperatures in UHT continuous flow systems are much higher than those used for traditional in-container processing. Temperatures in the range 125ºC– 130ºC are typical and temperature in some exchangers can be greater than 145ºC. For this reason the holding times that are considered in continuous flow processing tend to be much shorter. As these times are much shorter, strict control is needed to ensure the correct time temperature regime is given to
Continuous heat processing
43
products as any deviation at high temperature can have a significant effect on the commercial sterility of the product. The product is held at constant temperature (as set by the controller) and held for a time (dictated by the length of the holding section and flow characteristics of the product). To ensure that worst case conditions are taken into account when calculating the time for which the product is held, it is necessary to know something of the product flow characteristics. When assessing the flow behaviour of a fluid, the physical properties of that fluid must be known, including density, specific gravity, temperature and apparent viscosity. From this data the processor can calculate the Reynolds number (Re), a dimensionless number, which characterises the flow. This is calculated by the following equation: Re
fluid density
kg m 3 pipe diameter
mfluid velocity
ms 1 fluid viscosity
Pa s
In general if the Reynolds number calculates to be less than 2100, the flow is considered to be laminar, if between 2100 and 10 000, the flow is transitional and if greater than 10 000, the flow is turbulent. When flow is laminar, the calculations are based on the fastest part of the flow (the centre) which is travelling twice as fast as the product, which is against the pipe wall. If the holding time is calculated on the mean velocity this would mean that the product at the centre only received half of the process required. For example, if the mean flow rate is 10 litres per minute and the volume of the holding section is 10 litres, the mean holding time would be 1 minute. However, the holding time would need to be doubled to 2 minutes to take into account the fastest moving fluid at the centre of the pipe to ensure that this part of the fluid received a 1 minute hold. The true relationship between the fastest moving particle and the pipe mean velocity depends on many factors. For example, the frictional resistance to the flow of the inside pipe walls, the changing temperature across and along the holding tube and the flow (rheological) properties are all important factors. Worst case scenarios for the holding section should also take into account particle sizes and quantities in the product. The size of particle that can be processed by continuous methods is limited by the time for conduction of heat to the centre of the particle. Usually for particulate product a laminar flow is assumed, where the particles may be flowing twice as fast as the mean flow. For modelling purposes, it should be assumed that the carrier fluid around the particle is static, because this will give the worst case heat transfer to the particle surface and therefore will underestimate the process delivered. When monitoring the temperature within the holding section, temperature losses along the pipeline should be taken into account. For this reason, it is necessary to securely lag the holding section to ensure any heat losses are kept to a minimum. In particular, the temperature sensors should be well lagged. Typically the controller temperature is located at the entrance to the holding section; however several degrees can be lost as the product passes along it, and
44
Thermal technologies in food processing
this should be monitored and taken into account when calculating the lethality of the process. The holding section should also be designed to avoid entrapment of air within the product by elevating the tube by a small angle. It may also be necessary to consider the use of in-line mixers for high viscosity products. The mixers should be placed into the line prior to temperature measurement at the holding tube exit, to ensure adequate mixing. In-line mixing should also be used to avoid channelling of product within the holding tube and several may be used in series if the tube is long or product viscosity is very high.
3.5
Future trends
Continuous heat processing is an economical way to produce large quantities of product ensuring safe, high quality reproducible products that were not feasible for all products when produced by traditional batch methods. Whether UHT processed and aseptically packed, or heated and hot-filled or chilled and kept under refrigerated conditions, this method of processing offers the food processor a flexible system. Plate heat exchangers offer a very well-established method for processing low viscosity homogenous products. Few developments are occurring in this area. Manufacturers of such equipment are offering easier to clean, quicker to service modules which enable a quick turnaround in terms of product downtime. Plate frames are designed to allow inspection to be carried out in place. The main development in this area is in the ability to produce product containing particulates with a reduced likelihood of blocking channels and again, reducing production time. Developments for tubular heat exchangers are currently being actively researched. Although tubular heat exchangers are designed for some level of regeneration (particularly in concentric tube designs) this is limited to low viscosity products such as milks and juices. With the introduction in the UK of the Climate Change Levy in April 2001, food manufacturers are being asked to provide information on how they are conserving energy within their companies. Work, looking at the use of higher viscosity products such as sauces and pure´es for regeneration is being carried out currently with the aim of enabling equipment manufacturers to develop tubular heat exchangers that have the capability of using these higher viscosity products to preheat and cool product. The aim is to try and enable regeneration of up to 75% with this type of product. Scraped surface heat exchangers are another well-established heating system in which few changes have occurred. Different materials are being used to aid heating and different plastics to try and extend the life of the blades between servicing. Companies such as Waukesha Cherry Burrell have updated their traditional horizontal processing equipment and have also introduced a vertical design, to compete directly with the other manufacturers. One development that has yet to be established commercially is the co-rotating disc scraped surface
Continuous heat processing
45
heat exchanger (CDHE). Having a rotating heating surface and a stationary scraping device this system claims a high heat transfer capacity having heat transfer coefficients as high as 600 W/m2K when processing viscous fluids. The CDHE develops large zones of reverse flow in the processing chamber so improving mixing in the chamber and improving heat transfer. The work carried out by the University of Denmark outlines that when placing several processing chambers in series the CDHE has great potential specifically for UHT processing of foods showing pseudoplastic behaviour and also particulate foods. Direct heating systems are also fairly static in terms of development being well-established systems. One of the areas where work is being carried out is in the use of direct systems for processing particulate matter. As already stated, infusion chambers can handle products containing small particles (such as juice cells) and further work is being carried out. Much needs to be understood about the method of heat transfer to the particles to ensure that on entry to the heating chamber the centre of the particle also achieves the required heat process so ensuring a safe product. Other trends in this area do not refer to the traditional heating methods already discussed. There are other methods of continuous processing that are being used for both development and commercial purposes. Ohmic heating will be discussed in a later chapter and so is not discussed here.
3.6
Sources of further information and advice
3.6.1 Equipment manufacturers Further detailed information on specific designs of equipment is available from the manufacturers of the equipment. Major manufacturers that can offer help and advice on types of exchangers suitable for specific processes are: Tetra Pak Ltd., 1 Longwalk Road, Stockley Park, Uxbridge, Middlesex UB11 1DL, UK Tel: +44 (0) 870 442 6000 Fax: +44 (0) 870 442 6001 Website: www.tetrapak.com HRS Heat Exchangers Ltd., 10–12 Caxton Way, Watford Business Park, Watford, Hertfordshire WD1 8UA, UK Tel: +44 (0) 1923 232335 Fax: +44 (0) 1923 230266 Website: www.hrs.co.uk Waukesha Cherry Burrell, 2300 One First Union Center, 301 South College Street, Charlotte, NC 28202, USA Tel: +1 800 252 5200 Fax: +1 800 252 5012 Website: www.waukesha-cb.com
46
Thermal technologies in food processing
FMC Food Tech. Food Processing Systems, Sint-Niklaas, Belgium Tel: +32 3 780 1211 Fax: +32 3 777 7955 Website: www.fmcfoodtech.com APV Systems, 23 Gatwick Road, Crawley, West Sussex RH10 2JB, UK Tel: +44 (0) 1293 527777 Fax: +44 (0) 1293 552640 Website: www.apv.com 3.6.2 Trade associations Trade associations offer advice on equipment and processors in this area. Further information can be found from: The Processing and Packaging Manufacturers Association (PPMA), New Progress House, 34 Stafford Road, Wallingford, Surrey SM6 9AA, UK Tel: +44 (0) 20 8773 8111 Fax: +44 (0) 20 8773 0022 Website: www.ppma.co.uk The Dairy Trade Federation, 19 Cornwall Terrace, London NW1 4QP, UK Tel: +44 (0) 20 7486 7244 Fax: +44 (0) 20 7487 4734 3.6.3 Professional bodies Professional bodies offer information and publications in this area. Further information can be found from: Institute of Food Science and Technology (IFST), 5 Cambridge Court, 210 Shepherds Bush Road, London W6 7NJ, UK Tel: +44 (0) 20 7603 6316 Fax: +44 (0) 20 7602 9936 Website: www.ifst.org Institute of Food Technology (IFT), 221 N. LaSalle Street, Ste. 300, Chicago IL 60601-1291, USA Tel: +1 312 782 8424 Fax: +1 312 782 8348 Website: www.ift.org 3.6.4 Research associations and universities Several universities and research associations can offer independent help and advice on processes and problems with processes. Further information can be found from:
Continuous heat processing
47
Campden and Chorleywood Food Research Association, Chipping Campden, Gloucestershire GL55 6LD, UK Tel: +44 (0) 1386 842000 Fax: +44 (0) 1386 842100 Website: www.campden.co.uk Reaseheath College, Nantwich, Cheshire CW5 6DF, UK Tel: +44 (0) 1270 625131 Fax: +44 (0) 1270 625665 Website: www.reaseheath.ac.uk University of Reading, Faculty of Agriculture and Food, Whiteknights, Reading RG6 2AP, UK Tel: +44 (0) 1189 318700 Fax: +44 (0) 1189 310080 Website: www.rdg.ac.uk Milk Marque, Product Development Centre, Reaseheath, Nantwich, Cheshire CW5 6TA, UK Tel: +44 (0) 1270 611051 Fax: +44 (0) 1270 611013 Website: www.milkmarque.co.uk School of Chemical Engineering, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK Tel: +44 (0) 121 414 5330 Fax: +44 (0) 121 414 5324 Website: www.bham.ac.uk 3.6.5 Further reading For further background reading the following texts offer specific advice and information in specialised areas: BURTON H,
Ultra High Temperature Processing of Milk and Milk Products. London, Elsevier Applied Science, 1988. HOLDSWORTH S D, Aseptic Processing and Packaging of Food Products. London, Elsevier Applied Science, 1992. ROSE D, Guidelines for the processing and aseptic packaging of low acid foods. Part 1. Principles of design, installation and commissioning. CCFRA Technical Manual 11, 1986. ROSE D, Guidelines for the processing and aseptic packaging of low acid foods. Part 2. CCFRA Technical Manual 11, 1987.
48
3.7
Thermal technologies in food processing
References
BUCHNER N,
Aseptic processing and packaging of food particulates. In: Aseptic Processing and Packaging of Particulate Foods. Ed. Willhoft, E M A, Glasgow, Blackie Academic and Professional, 1993. BURTON H, Ultra High Temperature Processing of Milk and Milk Products. London, Elsevier Applied Science, 1988. DAVID J R D, GRAVES R H, CARLSON, V R, Aseptic Processing and Packaging of Food. A food industry perspective. Boca Raton, USA, CRC Press Inc, 1996. FRIIS A, ADLER-NISSEN J, ‘The co-rotating scraped surface heat exchanger for food processing’. In: Advances in Aseptic Processing and Packaging Technologies. Ed. Ohlsson, T, Go¨teborg, Sweden, Kompendiet, 1995. NELSON P E, CHAMBERS J V, RODRIGUEZ J H, Principles of Aseptic Processing and Packaging. Washington DC, The Food Processors Institute, 1987. ROSE D, Guidelines for the processing and packaging of low acid foods. Parts 1 and 2. CCFRA Technical Manual 11, 1986, 1987.
Part II Measurement and control
This page intentionally left blank
4 Pressure and temperature measurement in food process control P.G. Berrie, Endress+Hauser Process Solutions AG, Reinach
4.1
Introduction
During thermal processing food safety is primarily assured by the control of temperature for a given time. The application of heat processes, for instance in the case of retort technology, requires not only the ability to heat and cool a product, but often the effective control of pressure too. Other measurements may also be required for full control. Thus trace moisture sensors are essential for monitoring the water content of blanket gases or dry air, for example. In-line pH measurements are often encountered in the manufacture of products such as milk. Occasionally, Redox and turbidity measurements may also be employed. As no further details will be given in this chapter of the methods used to measure these variables, the interested reader is referred to other literature on the subject.1 This chapter addresses temperature and pressure measurement, two process variables that account for over 80% of the measuring points in the food processing industry. Many instruments have been developed to measure them, but in this chapter only those that might be met in a food production facility will be considered. The chapter covers a number of devices for use in measuring pressure and temperature at a variety of points in production, both before and after, as well as during a specific thermal process. Effective temperature control remains important at all stages of production. Hygiene is an ever-present factor in food production, and a number of guidelines have been developed to assure the quality of the end product. Those especially applying to process instrumentation are discussed in Section 4.4. Here, instrument design is considered from the aspects of process connections, materials and cleaning-in-place.
52
4.2
Thermal technologies in food processing
Pressure measurement
In food processing, pressure measurements are required in pipe systems, across filters, as well as in closed tanks and process vessels. For piping in particular, the instruments must be designed to withstand overload pressures far beyond their normal operating range. As liquids or solids are pumped along the pipe, the pumping action causes regular surges in pressure. When valves are opened and closed, the pressure may increase to the normal operating pressure of the pump. If a valve closes abruptly, as is often the case, high pressure peaks can be created. Finally, pressure instruments must be immune to vibration and be able to withstand the temperatures and stresses induced by factory cleaning procedures. Pressure measurement devices are categorised according to the ‘type’ of pressure they display: • Absolute pressure devices measure the total pressure acting on the sensor. For atmospheric pressure, therefore, they display approximately 1000 millibar or 1 bar. • Gauge pressure devices display the pressure relative to atmospheric pressure. In this case, atmospheric pressure would produce a reading of 0 millibar – lower pressures are registered as negative values. • Differential pressure devices measure and display the pressure difference between two tapping points, e.g. in a pipeline or a tank. They are usually used to measure level and flow.
Measurement methods As far as the measuring principle is concerned, practically all pressure gauges fall under the category of force-type pressure devices. The way in which the force resulting from the acting pressure is sensed leads to a further subdivision into: • manometer-type measurement devices • mechanical-type measurement devices • electrical pressure transducers.
Of these methods, electrical measurements are most suited to automation. They are often seen alongside mechanical devices that are primarily installed on pressure vessels for safety reasons. Mechanical and electrical pressure transducers are most common in the heat treatment of foods. 4.2.1 Manometer-type instruments A manometer comprises a U-tube filled with a so-called manometric fluid (water, oil or mercury), one limb being connected to acting pressure and the other being open, closed or otherwise connected to an indicating, registering or recording mechanism (see Fig. 4.1). The pressure to be measured acts upon the liquid in one limb of the manometer, pushing it into the other limb. The liquid continues to be displaced until the forces acting in both limbs are equal. For a
Pressure and temperature measurement in food process control
Fig. 4.1
53
Manometer-type pressure measurement devices.
gas, the difference in the liquid levels in the two limbs of the manometer gives a reading of acting pressure. So-called well or J-tubes indicate pressure directly. Here the diameter of the pressure limb is much greater than the measuring limb, so that a minimal change in level causes a large change on the measuring side. Where small pressure differences are to be measured, the tube may be inclined to give more accurate readings. Other designs use a float and mechanical assembly which operates a pointer when the manometric fluid level changes. In addition to classical manometers, there are numerous more sophisticated variations, e.g. balance hollow ring, bell chamber or bellows. In these devices, the manometer fluid is used to separate the pressure limb from the measuring limb, on which acts a reference pressure. The difference in pressure is transmitted by a mechanical assembly to a dial gauge. In view of their simplicity, manometers are still a common sight in many plants, where they provide an optical control of the pressure. They are, however, losing ground to the other measurement methods. Classical manometers are also sensitive to temperature changes and must therefore be used within a fairly strict temperature range. As the manometric fluid is in direct contact with the process, manometers are not suitable for sterile applications. 4.2.2 Mechanical-type measurement devices Mechanical-type measuring devices translate the pressure acting upon them into a mechanical displacement. There are various designs of measuring element, of which diaphragms and Bourdon tubes are the most likely to be met in the food industry (see Fig. 4.2). A diaphragm element is a thin, flexible metal plate that separates the chamber connected to the pressure being measured from the atmosphere. It may be flat, corrugated or in the form of an introverted capsule. When both sides of the diaphragm are exposed to the same pressure, the diaphragm is in its equilibrium position. If a negative pressure is applied (less than atmospheric) the diaphragm is pulled in, if a positive pressure is applied, it is pushed out. The maximum displacement occurs at its centre.
54
Thermal technologies in food processing
Fig. 4.2 Mechanical-type pressure measurement devices.
A Bourdon tube element comprises a thin flattened metal tube, sealed at one end, which is connected to the pressure being measured. The tube is bent into a circular or helical form, or may simply be twisted. When positive pressure is applied, the tube will tend to straighten out, producing a linear or radial deflection that is proportional to the acting pressure. By the use of a suitable mechanical assembly this displacement or deflection can be used to give an indication of pressure, e.g. on a dial. If the dial is fitted with electrical contacts, pressure limits can be monitored. The pointer and dial display is easy to read and gives a quick indication of pressure. In automated plant, however, most serve only as optical back-ups to electrical pressure transmitters. 4.2.3 Electrical pressure transducer The majority of electrical pressure transducers use a flexible diaphragm as the pressure transmitting element. Unlike mechanical-type devices, the diaphragm is not a part of a pressure chamber connected to the process, but forms the front isolating element of a sensing chamber. Thus, by using the appropriate process connection, it is possible to produce flush mounted pressure devices that have no cavities and that are easy to clean. Figure 4.3 shows six typical measuring cells designs for absolute/gauge and differential pressure measurement. Type A is a resistive gauge/absolute pressure cell. The pressure acting on the isolating metal diaphragm is transmitted by a push rod onto the flexible beam, which deflects. Strain gauges located on the beam generate an electrical signal proportional to the acting pressure. The absolute cell is evacuated, the gauge cell is filled with fill fluid. Type B is a resistive gauge/absolute pressure cell. The pressure acting on the isolating metal diaphragm is transmitted by a fill fluid to a second, flat, sensing diaphragm. The sensing diaphragm is also deflected, whereby the degree of
Pressure and temperature measurement in food process control
55
Fig. 4.3 Measurement cells for electrical pressure transmitters.
deflection is greatest at its centre. Strain gauges or piezoelements on its surface produce an electrical signal proportional to the acting pressure. If the diaphragm is overloaded it sets back on the sensor body, avoiding any mechanical damage. The absolute cell is evacuated, the gauge cell is filled with fill fluid.
56
Thermal technologies in food processing
Type C is a capacitive gauge/absolute pressure cell. Two annular electrodes are to be found on the body, a third on the inner face of the diaphragm. When the diaphragm deflects, the capacitance between the inner body and diaphragm electrode changes in proportion to the acting pressure. Since there is virtually no movement towards the outer body electrode, this outputs a constant reference signal. If the diaphragm is overloaded it sets back on the sensor body, avoiding any mechanical damage. The cell is completely dry. The absolute pressure sensor is sealed and evacuated. Type D is a twin-chamber capacitive differential pressure cell. It comprises two corrugated isolating diaphragms on the low and high pressure sides. To prevent damage on overload, these lay back on a similarly formed substrate. The space between the substrate and isolating diaphragm as well as the entire sensing capsule is filled with fluid. Capillaries lead from the front to the centre of the cell, where a sensing diaphragm separates the two sides. Metal electrodes are deposited on the diaphragm and capsule walls, forming the capacitor which measures the pressure when the sensing diaphragm deflects. Inductive cells have a similar design, whereby the metal sensing diaphragm has two ferrite discs fixed to each side. These together with two fixed magnetic coils on each side of the sensing capsule body form a differential transformer. When the sensing diaphragm deflects, there is a change in inductivity that is proportional to the acting pressure. Type E is a twin-chamber resistive cell for differential pressure. The construction of the cell is essentially the same as the twin-chamber capacitance cell, with the exception that additional capillaries lead to an external sensing capsule. Any deflection at the isolating diaphragms is transmitted by the fill fluid to the external sensing diaphragm. Strain gauges or piezoelements then produce a pressure proportional electrical signal. Type F is a single-chamber capacitive cell for differential pressure. The two sides of the cell are connected together by a capillary and a fill fluid transfers the pressure from one side to another. The cell body and diaphragms are ceramic, each side of the cell having its own set of capacitance electrodes, which deliver an electrical signal which is a function of the pressure. Since the two sides are connected, it is possible to build a self-monitoring cell. As the temperature increases, the volume of the fill fluid increases and the capacitance of the sensor changes. Since the expansion coefficient of the fill fluid is known, this gives an indirect measurement of temperature. By comparing this temperature with an independent measurement of the sensor temperature, it is possible to detect whether the sensor is faulty. There are many other transducers on the market, e.g. resonance-wire or vibrating-beam, but these are often associated with a particular vendor and are beyond the scope of this book. There is also a further subdivision within the resistive type cells which relates to the sensing element. So-called metallic sensors use strain gauges of various types. Foil strain gauges are made from resistive alloys such as Konstantan that have been rolled into thin foils a few m thick, coated with an isolating organic substance and
Pressure and temperature measurement in food process control
57
stuck to a metal diaphragm. Thin-film strain gauges are produced by vapour deposition, spluttering or chemical vapour deposition techniques. Piezoresistive sensors are based on semiconductors that have pressuredependent electrical properties. There are two basic types of element. Polysilicon sensors are manufactured in the same way as metallic thin-film strain gauges and are doped to produce the desired electrical properties. Monosilicon sensors are made of pure silicon and are anisotropic, i.e. their resistance is direction-dependent, allowing them to produce highly sensitive sensors. Thick-film strain gauges are produced by silk-screen printing techniques. The individual layers required to produce the desired sensor geometry are built up on a ceramic substrate using several different masks. 4.2.4 Summary Table 4.1 summarises the typical operating conditions of the various measurement methods. The accuracy of measurement is not quoted since it depends on the individual transmitter: in general it comprises the measured error, the hysteresis and reproducibility measured at a reference temperature, usually 20ºC. Often all three are quoted as an ‘overall’ accuracy. Also to be considered is the temperature effect, which indicates by how much the accuracy changes for a given temperature rise. This can be quite considerable for certain fill fluids or if the sensor is operating with a diaphragm seal (an extension which allows flush mounting) or a remote seal and capillaries. The overload pressure indicates the ability to withstand pressure peaks without mechanical damage to the sensor: a re-calibration may be necessary, however. More information on the testing of pressure and other process transmitters can be obtained from IEC Standard 61298, Parts 1 to 4 and DIN 16086.2, 3
4.3
Temperature measurement
Temperature is by far the most frequently measured variable in process engineering. In food production, the monitoring and control of temperature is an important factor in assuring the safety and quality of the end product. Although there are many applications for temperature measurement in food processing, the process conditions to be encountered are not particularly hostile. Thus, the temperatures to be met range from about 50ºC in cold storage to 150ºC in sterilisation-in-place applications. Only in steam generation will higher temperatures be found. Accuracy is very important, and of course, hygiene is essential – see Section 4.4. Temperature measurement methods Temperature measuring devices that are used in food processing can be categorised as:
Table 4.1
Typical operating conditions of various process pressure devices
Type Manometer
Diaphragms
Temperature range 10ºC to 60ºC
Up to 100ºC
Min./max. pressure range
Differential pressure range
Remarks
0 mbar to 1 mbar 0 bar to 20 bar
—
Range depends on manometric fluid. Inaccurate if not operated at within the rated temperature range.
0 mbar to 10 mbar 0 bar to 25 bar
—
Significant temperature effect Up to 5 overload (max 40 bar)
Capsule
25ºC to 60ºC
0 mbar to 2.5 mbar 0 mbar to 600 mbar
—
Significant temperature effect
Bourdon gauges
20ºC to 60ºC
0 mbar to 600 mbar 0 bar to 4000 bar
—
Significant temperature effect
Electrical resistive
30ºC to 200ºC
0 mbar to 1 bar 0 bar to 400 bar
Up to 3 bar for static pressures up to 140 bar
Actual pressure and temperature range depend on cell design. Usually resistant to vibration and overload.
Electrical capacitive
30ºC to 200ºC
0 mbar to 400 mbar 0 bar to 40 bar (ceramic) 0 bar to 600 bar (metal)
Up to 3 bar for static pressures up to 420 bar (ceramic) Up to 500 bar for static pressures up to 500 bar (metal)
Actual pressure and temperature range depend upon cell designs. Resistant to vibration and overload. Ceramic cells are more corrosion resistant and self-monitoring.
Electrical inductive
20ºC to 120ºC
0 mbar to 100 mbar 0 bar to 160 bar
Up to 160 bar for static pressures up to 210 bar
Actual pressure and temperature range depend upon cell designs. Resistant to vibration and overload.
Notes: The values are guidelines only, for specific values see the manufacturer’s data sheets. For electrical transmitters, the measuring range can often be ‘turned down’ by 40:1 or more.
Pressure and temperature measurement in food process control
59
• force temperature devices • electrical temperature devices.
4.3.1 Force temperature devices Force temperature devices (FTDs) can be further divided into bimetallic FTDs and filled thermal system FTDs (see Fig. 4.4). They make use of the fact that the length or volume of a given mass of matter changes as its temperature increases. The change in length or volume with unit temperature is dependent upon the material and is characterised by a so-called coefficient of expansion. Bimetallic strip thermometers A bimetallic strip comprises two strips of metal with different coefficients of expansion that are bonded together, e.g. by riveting. When the temperature rises, each metal expands by a different amount causing the whole strip to bend. The amount of deflection is an indication of temperature. By a suitable mechanical assembly, this deflection can be used to give a continuous indication of temperature at a graduated display or to make contact with a switching element. Bimetallic strip gauges are inexpensive and good designs provide reasonable accuracy (2%–3% of full scale (FS)). They are built for specific temperature ranges. Continuously reading bimetallic thermometers can be used as optical back-ups to resistance thermometers and in applications where temperature is not used as a control variable. Bimetallic switches are often found in two-point temperature controls. Filled thermal systems Filled thermal systems make use of the thermal expansion of liquids to provide a direct indication of temperature. In the most common of all, the glass bulb thermometer, the bulb acts as a reservoir for the liquid and is connected to a
Fig. 4.4
Temperature measurement device designs.
60
Thermal technologies in food processing
graduated capillary tube. As the temperature rises, the liquid within the bulb expands into the capillary: the height of the liquid column indicates the current temperature. The liquids used are, for example, coloured alcohol or mercury. Thermometers are designed to measure over a specific and appropriate range. If the flow of liquid back to the bulb is restricted, for example by a kink in the capillary, they can also be designed to give more accurate measurements over smaller ranges. Should a glass bulb thermometer break, glass splinters and alcohol or mercury would enter the process and contaminate the product. For this reason, thermometers are mounted in protective metal sheaths. Here it is essential that there is good thermal contact between the bulb and the sheath, otherwise the temperature indication is inaccurate and the response slow. Alternatively metal bulb thermometers might be used. 4.3.2 Electrical temperature devices Due to the ease with which they can be integrated into control systems, electrical devices are the preferred method of temperature measurement in process control. They use the dependency of electrical properties of particular materials on temperature to provide a measurement (see Fig. 4.4). There are four different types: • • • •
resistance temperature detectors (RTDs) thermocouples silicon resistors semiconductors.
Only RTDs and thermocouples are of real interest in the production of food. Resistance temperature detectors Resistance temperature detectors (RTDs) or resistance thermometers are the most common type of temperature sensor to be found in process engineering. They comprise a thin-film or wire resistor with a standard resistance of 100 , 500 or 1000 . The resistor material may be platinum or nickel, the standardised designations being Pt100, Pt500, Pt1000 or Ni100. The sensors are very stable, have a wide operating range from 200ºC to 850ºC depending on type and exhibit a well-defined relationship between resistance and temperature. As a consequence they are interchangeable over a wide temperature range. RTDs may have a relatively slow time response and offer only a small change in resistance per unit increase in temperature. To limit or avoid errors in the determination of such small changes in resistance for measuring points, where the sensing element and the evaluating electronics are separated by some distance, three- or even four-wire measurement technology must be used. RTDs are also sensitive to vibration and shock, so care must be exercised in their design.
Pressure and temperature measurement in food process control
61
Thermocouples If two wires of different metals are joined together at their ends and the two junctions are at different temperatures, a potential difference is created that causes a current to flow around the loop. The difference in potential arises because the magnitude of the contact potential at each junction is dependent on the temperature. If one of the junctions is kept at a constant temperature, the magnitude of the current that flows is a measure of the temperature at the other. Thermocouples are simple, rugged, inexpensive and require no external power supply. They can also respond faster to temperature change than RTDs. On the other hand, the signal is non-linear, they exhibit low sensitivity and have relatively low stability. A reference junction is required and they must be compensated. Moreover, metallurgical changes and ageing sometimes causes a loss of performance. A more serious problem is that the low voltage output is susceptible to electromagnetic interference. 4.3.3 Design and assembly of RTDs Since RTDs are the temperature sensors most likely to be found in the food industry, their design, assembly and use will be examined in more detail. The design has features in common with other electrical temperature devices, since most comprise the sensing element in its protective sheath, a thermowell for introducing the sensor to the process and a head or remote transmitter for evaluating and processing the incoming signal as well as providing power to the sensor circuit. The use of a thermowell is also common practice for force temperature devices. Sensing element There are various designs of Pt100 sensing elements. The traditional sensing element is the ceramic Pt100 element which comprises a spiral of extremely fine platinum embedded in ceramic powder (frit) within capillaries of extremely pure alumina. Two platinum wires provide the contacts to the connecting wires. The glass Pt100 element comprises an extremely thin platinum tape wound around a glass mandrel. The element is protected by a glass sleeve which is fused to it. Platinum wires provide the contacts to the connecting wires. Due to the fusion of the sheath and element, this design has the advantage of a high vibration resistance. For the thin film Pt100 element, an extremely thin layer of platinum is evaporated onto a 0.5 mm thick ceramic substrate. By using a laser or etching technique, platinum is removed selectively from the surface in order to create a meander with a nominal resistance of 100 at 0ºC. The platinum coated connecting wires and thin film are protected by a fused glass coating. Table 4.2 lists the nominal operating temperature ranges for the three types of Pt100 sensing element. The resistance measured by the RTD sensing element is converted to a temperature by using the tables in European Standard EN 60 751 ( IEC 60 751).4 These values have been determined on the basis of the International
62
Thermal technologies in food processing
Table 4.2
Nominal operating temperatures of Pt100 sensing elements
Type
Dimensions
Design
Isolation
Pt100 Pt100 Pt100
2.5 mm dia 35 mm 2.5 mm dia 35 mm 10 mm 2 mm 0.5 mm
Wound Wound Thin film
Ceramic Glass Glass/ceramic
Temperature range 200ºC to 850ºC 200ºC to 600ºC 50ºC to 400ºC
Temperature Scale ITS90,5 and can be calculated using a polynomial equation, the coefficients of which are known exactly for each sensor element type, i.e. for Pt100, Pt500, Pt1000 and Ni100. The standard further defines two accuracy (tolerance) classes for Pt100 sensing elements: • Class A: Accuracy in ºC 0.15 0.002 (t) • Class B: Accuracy in ºC 0.30 0.005 (t)
where (t) is the unsigned numerical value of the temperature in ºC. This classification is often not adequate for industrial users and it is common to meet fractions or multiples of the classes. For example, Class 1/3 B would correspond to an accuracy in ºC of 0.10 0.005 (t) and 2 B to 0.60 0.005 (t). For fractional B classes, the temperature dependent part of the tolerance (0.005 (t)) sometimes has a lower value, thus assuring greater accuracy. Type of connection RTD sensing elements are manufactured by a few specialist firms around the world. The top five deliver products of comparable high quality to most of the major instrument manufacturers. It is only in the design beyond the element that differences are to be found. The two connecting wires of the sensing element carry the signal to the evaluating circuit which in its simplest form might comprise a constant current source and a means of measuring the resistance. For process instrumentation this circuitry is usually found in a head or remote transmitter. Since the evaluating circuit and sensing element are always removed from one another, the gap has to be bridged by connecting wires. If the whole sensing assembly is to be removable, e.g. for replacement in the event of a failure of one component, then it is usual that these wires end in terminals. Connecting leads are then required to carry the signal to the evaluating circuit. The connecting wires, leads and terminals may all be sources of error in the final measurement. For this reason, two-, three- and four-wire circuits are to be found (see Fig. 4.5). For a two-wire circuit, the resistance of the connecting wires, leads and terminals must be added to the resistance of the sensing element. The longer the leads, the higher the resistance and the larger the resulting error. For this reason two-wire circuits should only be used when the transmitter is located close to the sensing element. This is particularly critical for Pt100 and Ni100 RTDs.
Pressure and temperature measurement in food process control
63
Fig. 4.5 Types of connection for RTDs.
In a three-wire circuit, an additional wire is connected to one side of the sensing element. This is connected to the transmitter and acts as a reference for the lead resistance. Provided the leads are of identical material and length, this reference value can be subtracted from the detected resistance to provide a measurement that is influenced only by the terminal resistance. The remaining error is roughly 0.1% of the measured value, which provides sufficient accuracy for most industrial applications. The four-wire circuit is used for very accurate measurements. Two connecting wires on each side of the sensing element are connected to the transmitter. A constant current is supplied to the element via one pair of wires, whilst the other pair are used to measure the resulting voltage. By using this type of voltage compensation, unbalanced wires and leads, and their associated resistances, have no effect on the measurement. Intrinsic warming (self-heating) In order to measure temperature, an RTD must be supplied with a constant current. This current also causes the device to heat up. The degree of heating is dependent upon the design, and is particularly critical for thin-film elements. It is extremely difficult to determine the extent of this so-called intrinsic warming or self-heating; however, for a Pt100 sensor operating at 10 mA it is typically 0.02 K to 0.4 K in water and 0.9 K in air. For this reason it is usual to operate the RTD at a constant current of 1 mA, making the intrinsic warming effect negligible in comparison to the standard accuracy. Sensor insert Before the sensing element can be used in an industrial application, it must be protected against shorting and vibration. To this end, it is built into an exchangeable sensor insert (or inset). The insert comprises a metal sheath, usually AISI 316 stainless steel, containing the sensing element and connection wires (see Fig 4.6). To prevent them from shorting, the connection wires are threaded through thin ceramic tubes. The whole assembly is then packed in alumina powder to protect it against vibration.
64
Thermal technologies in food processing
1 Leads with ceramic isolation
Fig. 4.6
2 Metal or ceramic sheath
3 Alumina powder
Construction of sensor insert for RTDs and thermocouples.
A more advanced insert manufacturing technique uses magnesium oxide cables, i.e. extruded metal tubes that contain the connecting wires isolated by compressed MgO powder. This method has the advantage of providing excellent electrical isolation and provides protection against environmental humidity. Furthermore the insert is flexible, vibration-proof and can be operated at high temperatures. The top of the insert may be of two different types. Either the connecting wires are led out via ceramic sockets to a set of terminals or they are connected directly to a head transmitter which contains the evaluating circuits. The number of terminals depends upon the type of connecting circuit used. Thermowells For most applications, the sensor insert is not suitable for direct contact with the process. This is the case for food manufacture, where the external surface might offer conditions for bacterial contamination. The insert is therefore operated within a thermowell, which is often an integral part of the sensor assembly (see Fig. 4.7). For food processing, this comprises a highly polished, closed, stainless steel tube with a process connection or welding neck. In order to ensure the best possible transfer of heat, the sensor insert is usually in contact with the well bottom or set into a thermoconductive paste and often tapers towards the end to provide better heat transfer. Head or remote transmitter The transmitter is the ‘brains’ of the sensor and converts the signal from the sensing element into a standardised output. It also powers the sensor where necessary. Modern transmitters are equipped with a microcontroller which
Pressure and temperature measurement in food process control
65
Fig. 4.7 Construction of a temperature sensor, here shown with sensor insert with terminal connections.
provides a direct conversion of resistance (or voltage) to temperature for output at a local display or to a fieldbus system, as well as compensation functions, e.g. for non-linearity of the sensor signal. Alternatively, a simple temperatureproportional 4. . .20 mA or 0. . .10 V signal with freely assignable temperature range-end values may be provided. It is usual that either RTDs, thermocouples or voltage signals can be connected to the same transmitter. If the transmitter is to be found in the sensor assembly, it is known as a head transmitter. If on the other hand, the transmitter is not an integral part of the sensor, i.e. it is located in a nearby control cabinet or in the control room, it is known as a remote transmitter. Since the sensing element of all electrical temperature devices outputs an electrical signal, it is also possible to make a direct connection to an output device such as a chart or data recorder which has the facilities to handle temperature signals. Housing The housing of a temperature device must protect the sensor insert, the terminals and, where appropriate, the head transmitter from the ingress of water and dust. It must also provide good access to the terminals and sufficient space for the wiring. The size therefore varies according to type of transmitter. The connecting leads to the sensor are led through a cable gland into the inside of the housing. All housings have either an IP or NEMA rating which indicates the degree of protection they allow when the lid is firmly clamped or screwed into position and the cable gland screwed tight (see Section 4.4). For food applications, housings are often made of stainless steel, although polypropylene (FDA approved) is often encountered in North America. Standardised housings are to be found in DIN 43 729, whereby Type B is used most often.6
66
Thermal technologies in food processing
4.3.4 Installation conditions Even the most accurate temperature sensor will deliver incorrect measurements if it is not properly installed. As mentioned previously, for food applications electrical temperature devices will usually be installed in thermowells. These in turn will be built into the pipes and tanks where the medium to be measured is flowing or stored. Since the conditions to be met at the measuring points are different, e.g. gases, liquids or solids; flowing or stationary, various designs of thermowell are available for different measurement tasks. For measurements in flowing medium, the thermowell will have been designed for a particular range of flowrates. It is beyond the scope of this book to discuss thermowell design further, but more information is available from vendors’ catalogues. Heat transfer Electrical temperature devices must always be in good thermal contact with the medium to be measured so that any change in temperature is recorded as quickly as possible. Since the heat transfer coefficients of solids, liquids and gases differ considerably, each application requires the thermowell to be installed in a different way. Adequate heat transfer is assured by ensuring that a certain length of the thermowell, the so-called immersion depth, is exposed to the medium. If the immersion depth is too shallow, heat may be dissipated to the pipe or tank, resulting in an incorrect measurement. This phenomenon is enhanced when the sensing element is placed close to the process connection: the metallic mass radiates heat to the external environment. To avoid measurement errors, therefore, the minimum immersion depth should be at least 15 times the diameter of the thermowell tip, whereby, if possible 20 times the diameter should be used. In all cases, the technical data sheets and standards governing thermowells should be consulted. Response time A second important factor in temperature measurement is the speed with which a change in temperature is registered by the sensing element. Since this also depends on the rate of heat transfer and the thickness of the thermowell tip, tests are made under standard conditions. The characterising variable in EN 60751 is the response time,4 also called the settling time in the instrument test standards series IEC 60770 and IEC 61298.7, 2 The response time is the time the sensor output requires to reach a certain percentage of its final steady-state value after a step rise in the input temperature. The times are designated 0.5 for 50% steadystate value and 0.9 for 90%. The response time is determined under two sets of conditions. • In air flowing at a velocity of 3 m/s, immersion depth 15 times the diameter of the thermowell tip plus the sensor element length, start temperature between 10ºC and 30ºC, temperature jump between 10ºC and 20ºC.
Pressure and temperature measurement in food process control
67
Table 4.3 Response times under standard conditions for various thermowell thicknesses Diameter of tip
In water at 0.4 m/s 0:5 0:9
9 mm 10 mm 11 mm 12 mm 13 mm
30 s 34 s 38 s 44 s 50 s
80 s 103 s 125 s 143 s 160 s
0:5
In air at 3 m/s
133 s 144 s 154 s 180 s 205 s
0:9 390 s 398 s 405 s 430 s 455 s
• In water flowing at a velocity of 0.4 m/s, immersion depth 5 times the diameter of the thermowell tip plus the sensor element length, start temperature between 5ºC and 30ºC, temperature jump not greater than 10ºC.
Table 4.3 lists reference values for typical thermowell thicknesses. The response time decreases with increasing flowrate and vice versa. Mounting position Table 4.3 shows clearly the relationship between response time, thermowell diameter and immersion depth: the thinner the thermowell, the smaller the immersion depth and the quicker the response. On the other hand, the thermowell acts as a resistance to air or liquid flow and is subject to mechanical stresses and vibration. In some cases, it might also be subject to abrasion or chemical attack. These factors have to be weighed against each other when selecting a suitable thermowell. The choice of mounting position may also influence this decision by producing more favourable conditions. When temperature sensors are mounted in pipes, then a position or orientation should be chosen that reduces the surface area facing the direction of flow. This reduces the mechanical load on the thermowell. Since the sensor produces turbulence in the downstream flow, care should be taken that it is not positioned too close to any other instrumentation in the pipe. • The ideal position is in a pipe elbow. This offers the smallest cross-section and least mechanical load. • Where the pipe is wide enough, the sensor can be mounted vertically. • The sensor can be angled towards the oncoming flow if the pipe diameter does not allow vertical mounting. • For very small pipes, the sensor can be mounted on a nozzle which is wide enough to allow fluid to flow around the thermowell.
For tanks, the situation is somewhat better, in that there is usually space for vertical or horizontal mounting. The temperature sensor must be positioned such that it is covered with medium at all times when the signal is required. Care should be taken not to position it too close to moving parts such as agitators, pumps or valves, however, since the forces generated by the movement of fluid
68
Thermal technologies in food processing
may exceed the mechanical strength of the thermowell or lead to increased abrasion or corrosion.
4.4
General instrument design
The production of food is one of the most carefully controlled areas of industrial activity. In addition to being strictly controlled by government agencies, the industry itself has set up its own watchdogs to monitor advances in production engineering and make recommendations regarding safe, modern practice. The aim, of course, is to protect the consumer from exposure to unsafe products, resulting from the presence of, for example, foreign bodies, chemical contaminants or bacteria. The use of instrumentation manufactured according to these guidelines is a step towards safe food production. 4.4.1 Design factors When choosing instrumentation for food production, it is always wise to bear in mind that it is not the normal operation of a device that gives problems, but rather the unexpected event. Thus, the risk of chemical contamination can be eliminated by using a suitable material for the wetted parts of the device. The risk of bacterial contamination can be reduced by regular cleaning and the use of suitably designed process connections. The introduction of foreign bodies, however, is only partially covered by the adoption of a high degree of protection. The case where equipment in direct contact with the product fails, producing debris or releasing contaminants, must also be considered. Here it is essential that the user is warned and/or that the released products are not dangerous to health. 4.4.2 Wetted parts The wetted parts of a device are those parts which are in contact with the medium being measured. For temperature measurement this might be the thermowell, for pressure measurement the isolating diaphragm and for a contacting level measurement the sensing element itself. Even so-called noncontact devices must be considered to have wetted parts when they intrude into the pipeline or tank. Here it is not so much the contact with the medium, but crevices and their exposure to high temperatures and vapours which has to be considered. The positioning of the measurement device must also be examined. Flowing gases, liquids or solids may cause abrasion or generate high mechanical forces, which combined with high temperature or vibration enhance electrochemical attack or mechanical fatigue. Moreover, the wetted parts must be able to withstand the forces and temperatures generated during cleaning or sterilisationin-place procedures.
Pressure and temperature measurement in food process control
69
Materials In addition to the normal mechanical design factors, the toxicological and bacteriological compatibility of the materials used for wetted parts must also be taken into consideration. As far as the toxicological properties are concerned, a material approved by the USA Federal Drug Administration (FDA approval) or equivalent regulating body must be used. The bacteriological factor is a different matter. Although regular cleaning and gap-free design reduce the broad risk of infection, the proper design and finishing of the wetted parts is just as important. This basically means flowing contours and clean welding, no nooks and crannies, and no obstructions that might cause the product to gather and rot. Usually all components in the tank are of highly polished stainless steel to prevent the product from sticking. Not to be forgotten is the corrosion resistance of the wetted parts. This is not simply a matter of their resistance to the products and cleaning agents. Under high temperatures, strong vibration and mechanical stress, electrochemical corrosion or intergranular corrosion may be enhanced. The one results in surface pitting, providing an ideal breeding place for bacteria, the other in the depletion of the nickel and chromium at the grain interfaces, which means the component will rust. Table 4.4 lists some stainless steels suitable for the food industry. 4.4.3 Process connections The majority of process instruments are installed in pipes or tanks by means of threaded connections or flanges. In certain high-risk food manufacturing operations this is unacceptable as these connections and flanges offer crevices and gaps where the product can accumulate. In addition, the mounting and dismounting takes considerable effort, so cleaning becomes difficult. Ideally, a process connection should offer no gaps where the product can become trapped. One solution is to weld the instrument in place and then grind and polish the inside of the connection. Unfortunately this means that the instrument cannot be exchanged should it fail. For thermowells, where the
Table 4.4
Stainless steels suitable for the food industry
Material
AISI*
Properties
1.4301
304
1.4404
316 L
1.4435 1.4571
316 L 316 Ti
Good resistance against organic acids at moderate temperatures Good resistance against salt and alkalis at moderate temperatures Increased resistance against non-oxidising acids such as acetic acid, tartaric acid, phosphoric acid Increased resistance against pitting and intercrystalline corrosion Better corrosion resistance than Type 1.4404 Increased corrosion resistance against particular acids and salt water Resistance against pitting corrosion
* The AISI steels are equivalents but do not have identical compositions.
70
Thermal technologies in food processing
Table 4.5
Sanitary couplings for the food industry
Type
Use
Description
Dairy coupling (to DIN 11851)
Pipes and tanks
Reasonably priced coupling that is frequently found in the food industry. Its weakness lies in the hygienic adaptation to the process which does not allow flush mounting. The coupling is made by a threaded boss and slotted sleeve. A conical seating and tapered nozzle with sealing ring comprise the process seal.
Aseptic coupling (to DIN 11864)
Pipes Tanks in preparation
Introduced in 1998 as a replacement for the dairy coupling. Designed to EHEDG recommendations (see Table 4.9), offering better hygiene thanks to a flush sealing construction. The mechanical coupling is via bolts or a threaded sleeve, the seal being flush with the pipe wall.
VariventÕ coupling
Pipes
In-line housing that allows the flush mounting of the sensor, which is attached to the housing by means of a screw clamp. Three housing types cover a wide range of pipe diameters. For the majority of process sensors Type 3, for pipes of DN 40 upwards, is required. This facilitates the exchange of instruments.
APV coupling
Pipes
In-line housing of similar construction to the Varivent coupling. The sensor, however, is bolted in position.
SMS coupling
Pipes and tanks
Reasonably priced, Scandinavian standardised screw coupling which is also used in France. Its weakness lies in the hygienic adaptation to the process which does not allow flush mounting.
IDF coupling
Pipes and tanks
International Dairy Federation screw coupling
Tri-clampÕ coupling
Pipes and tanks
Sanitary coupling with bevel seating produced by the Tri-Clover Company in America. Instruments are quickly mounted and fixed with snap-on clamps. The couplings find widespread use in America.
sensor insert is easily replaced, and for flowmeters, however, it is quite feasible and is often encountered. Process instruments are generally installed by means of so-called sanitary couplings. These combine the need for a gap-free mounting with that of easy mounting and dismounting, allowing them to be removed quickly for cleaning. Over the years a number of different designs have come on to the market, a selection of which are summarised in Table 4.5. 4.4.4 Ingress and explosion protection Just as important as the wetted parts and process connection is the design of the housing of a process instrument. Depending upon the instrument type, this may
Pressure and temperature measurement in food process control
71
contain only the connecting terminals or the entire evaluating electronics. In both cases it must provide protection: • from the ingress of dust or moisture from the outside • when the sensor is used in an explosion hazardous area, from the egress of a spark or flame from the inside to the outside.
The former can be ensured a suitable degree of protection, the latter by a suitable type of protection. Ingress protection As far as the ingress of dust and moisture is concerned, the world is divided into two camps. One half uses the IP standard (IEC 60 529)8 and the other the American NEMA Standard No. 250. Nowadays, however, many manufacturers quote both in their technical specifications. The IP standard is a description of the measures designed for the protection of the housing and the equipment within the housing. The degree of protection is indicated by a two-part code, e.g. IP 65. The first number is concerned with the protection from the ingress of solid matter, the second with water. As can be seen from Table 4.6, in order to withstand the frequent cleaning in a food production facility, housings with ratings of IP 65 or better are required. The NEMA standard comprises 14 type codes which deal with practical requirements on housings suitable for indoor and outdoor use. It also makes a statement about the protection from external influences and conditions such as mechanical impact, corrosion, humidity, mould, pests, dust, etc. As can be seen from Table 4.7, which lists only a selection of codes, a NEMA 4X enclosure is best suited to the requirements of the food industry. Explosion protection In comparison to the chemical industry, there is less need for explosion protection in the production of food. If flammable liquids or easily ignitable gases are present, however, then the instrumentation must be approved for use in explosion hazardous areas. Powders can also be a problem, since clouds of dust Table 4.6
Ingress protection categories to IEC 60 529
Code
Ingress protection against solids
Code
0 1 2 3 4 5 6
Not protected 50 mm diameter, e.g. hand 12.5 mm diameter, e.g. finger 2.5 mm diameter, e.g. tool 1 mm diameter, e.g. wire Protected from dust Dust-proof
0 1 2 3 4 5 6 7 8
Ingress protection against water Not protected Vertical dripping Dripping (15º inclination) Water spray Splash water Jet of water Strong jet of water Temporary submersion Total submersion
72
Thermal technologies in food processing
Table 4.7
Degree of protection of enclosures as per NEMA Standard 250 (selection)
Type
Indoor
1 2 3
yes yes
4 4X
yes yes
5
yes
6
yes
Table 4.8
Outdoor Degree of protection
yes yes yes
yes
Protection against contact with equipment within the housing Protection against a specified quantity of water droplets and dirt Protection against blown dust, rain, sleet, rain and snow, and external ice formation Protection against blown dust, jets splashes and jets of water Protection against corrosion, blown dust, jets splashes and jets of water Protection against dust falling, dirt and lubricating noncorrosive fluids Protection against water penetration during occasional temporary submerging in limited depth
Types of protection to EN 50 014
Type of protection
Significance
Intrinsic safety ‘i’
Favoured by European manufacturers. The protection is achieved by limiting the current flowing in the device circuits. This means that there are restrictions regarding the installation and cabling. The electronics are designed such that high temperatures and sparking cannot occur during normal operation. This type of protection is met, for example, in four-wire devices and power supplies. The entire electronics of the device are potted, so that no spark can enter the surrounding atmosphere The enclosure (housing) is designed such that any spark or explosion is retained within it. Frequently met in America where the entire cabling is routed in conduits. Also used in Europe for Dust-Ex protection.
Enhanced safety ‘e’
Encapsulated ‘m’ Explosion-proof ‘d’
are easily combustible under certain conditions. For milling, storage, conveyance and bagging operations, therefore, the Dust-Ex equipment should be used. The types of explosion protection are standardised in EN 50 014.9 Table 4.8 lists and explains the ones of interest to food processing. 4.4.5 Standardisation authorities As mentioned previously, the food industry is regulated by both government and industrial bodies. When choosing equipment, therefore, it pays to check whether a corresponding approval or authorisation has been granted. Table 4.9 lists several bodies which are of particular importance to safe food manufacture.
Pressure and temperature measurement in food process control
73
Table 4.9 Some regulatory bodies and associations important to process instrumentation Body
Field of activity
3-A sanitary standards Tests and authorises process instrumentation regarding its fitness for use in food production, in particular the dairy industry. The use of 3-A standards is entirely voluntary, but the quality is such that 3-A authorisation is accepted worldwide. European Hygienic Equipment Design Group (EHEDG)
Issues guidelines on hygiene, tests and certifies equipment and publicises state-of-the-art technologies. An EHEDG approval means that the equipment at hand has been tested successfully with respect to its suitability for food applications.
US Food and Drug Administration
Issues licences for products, whereby the processes, constituents, materials and constructional details are subject to examination. The use of an FDA-approved material, for example a stainless steel or elastomer, is a guarantee that the component concerned will not be attacked by the food product.
National governmental Lay down the permissible levels of trace organic and inorganic agencies substances in food products. Since this is an area of permanent research, it is possible that regulations vary from country to country. International Dairy Federation
4.5 1. 2.
3. 4. 5. 6. 7.
Responsible for standards and recommendations in the dairy industry which are often published as ISO standards.
References (ed) Food and Beverages, Measurement and Automation. Endress+Hauser, CH-4153 Reinach, Switzerland. IEC STANDARD 61 298, Parts 1–4, Process measurement and control devices, General methods and procedures for evaluating performance. Part 1 (1995) General considerations Part 2 (1995) Tests under reference conditions Part 3 (1998) Tests for the effects of influence quantities Part 4 (1995) Evaluation report content. DIN 16 086, Electrical pressure measuring instruments; pressure sensors, pressure transmitters, pressure measuring instruments; concepts, specifications on data sheets. Beuth Verlag, Berlin, May 1992. IEC STANDARD 60 751, Industrial platinum resistance thermometer sensors (currently under review), 1983. The International Temperature Scale of 1990 (ITS-90), see e.g. www.its90.com DIN 43 729, Electrical temperature sensor; connection heads for thermocouple-thermometers and resistance thermometers. IEC STANDARD 60 770, Transmitters for use in industrial-process control systems. BAUR M
74
8. 9.
Thermal technologies in food processing Part 1 (1999) Methods for performance evaluation Part 2 (1989) Guidance for inspection and routine testing (under review). IEC STANDARD 60 529, Degrees of protection provided by enclosures (IP code) Issued 1989, Amendment 1999. EN 50 014, Electrical apparatus for potentially explosive atmospheres – General requirements.
5 Validation of heat processes G.S. Tucker, Campden & Chorleywood Food Research Association, Chipping Campden
5.1 Introduction: the need for better measurement and control There is an extensive range of food products preserved using heat, encompassing many different process types from mild pasteurisation treatments to full sterilisation cooks, and using numerous heating systems to effect the heat treatments. Section 5.2 presents the challenge this poses to food companies as they strive to prove the microbiological safety of their products. The remaining sections present details of the more common methods available to companies for measuring pasteurisation and sterilisation values, and describes how the data can be used to demonstrate the thermal efficacy of a process. Sections 5.3 and 5.4 deal with conventional temperature probe systems, which should be, and are likely to remain as, the first option for the validation method chosen by most companies. Section 5.3 describes temperature distribution testing and Section 5.4 heat penetration testing. The methodology for temperature distribution and heat penetration testing evolved in the early part of the 1900s for the canned foods sector where full sterilisation processes in metal cans were the norm.1 In recent years, the temperature measurement hardware and analysis software have developed to allow measurements to be made for the plethora of pasteurisation treatments2 that have emerged in response to the consumer demands for less processed foods. To determine the microbiological adequacy of a thermal process, first order reaction kinetics for the destruction by heat of the target organisms are commonly used. Measured times and temperatures at the food product thermal centre are converted to equivalent processes or to log reductions in the organism numbers. However, estimating the log reductions in microorganisms attributed
76
Thermal technologies in food processing
to a thermal process is becoming increasingly difficult because of the growing number of process/product/package combinations that stretch the applicability of temperature probe systems. For example, continuous in-line processes for foods containing particulates or continuous oven and fryer processes (intended for products stored under refrigeration) require methods without trailing wires and dataloggers. Alternatives to probes must be used, such as immobilised spores in alginate beads, encapsulated biochemical time-temperature integrators or microbiological challenge tests. Sections 5.5 and 5.6 consider microbiological and biochemical methods respectively to illustrate where they can be best applied.
5.2
Validation methods: objectives and principles
One of the core activities involved with establishing a thermal process is the measurement of product temperatures that represents the ’worst case conditions’ likely to be experienced during normal production. This is the current thought process used by most food companies. Having considered the conditions that combine to give rise to this scenario3 this process typically involves two stages of temperature measurements. These are required irrespective of whether the process is established using the traditional General method for process value calculation, or with a model for data analysis such as Ball, CTemp or NumeriCAL. Chapter 6 in this book concentrates on modelling thermal processes, therefore in this section these modelling methods will be mentioned but no further details provided. For an in-container process, there are two main stages in a process validation exercise; temperature distribution or TD tests to identify the location of the zone of slowest heating in the retort, and heat penetration or HP tests to measure the temperature response at the product cold point. For a continuous flow process, a specialised form of the HP test is required, with various assumptions made regarding the positions of temperature measurement that make the TD test redundant. Some of the methods used for continuous flow processes are common with those for in-container processes and are described in this chapter, but further details on continuous processing can be found in Chapter 3. 5.2.1 Temperature distribution: locating the retort cold point Any thermal processing system (retort, autoclave or steriliser) will contain regions in which the temperature of the heating medium is lower than that measured by the master temperature indicator (MTI). For example, with water processes this can be caused by poor circulation of hot water throughout a close packed basket of glass jars. The location of these cold spots should be determined by performing ‘temperature (or heat) distribution’ (TD) tests throughout the system. The concepts for TD testing are simple, however the practicalities of making the relevant measurements are fraught with difficulty.
Validation of heat processes 77 For example, regions of low temperature may exist within a retort crate because of the flow restrictions imposed by the close packed containers, but to get at a container near the crate centre it is necessary to manually re-pack the crate and trail thermocouple wires between containers which may open up the flow channels. These are just two examples of hazards that must be dealt with when conducting a TD test. Further guidance on how TD tests should be carried out can be found in Section 5.3. 5.2.2 Heat penetration This is usually sub-divided into two further stages when conducting the tests, first to locate the product cold point in the container, and second to establish the process conditions that will lead to the scheduled process. Locating the product cold point Within each food container there will be a point or region that heats up more slowly than the rest. This is referred to as the ‘slowest heating point’ or ‘thermal centre’ and should be located using thermocouples or some other sensing method positioned at different places in a food container. For foods that heat mainly by conduction, the slowest heating point will be at the container geometric centre. However, for foods that permit movement and can thus convect heat, this point is between the geometric centre and approximately onetenth up from the base (in a static process). During a thermal process the food viscosity will decrease in response to increasing temperature, and as a result the slowest heating point will move downwards from the container geometric centre. The critical point is when the lethal effect on the target microbiological species is at its most significant, which will be towards the end of the constant temperature hold phase. If the process utilises rotation or agitation, the slowest heating point will be at the container geometric centre. Establishing the scheduled process time and temperature The thermal process is finally established by measuring the temperature at the container slowest heating point for a number of replicates that are placed in the cold spot(s) of the thermal processing system. The data obtained are usually referred to as ‘heat penetration’ data. A point open to discussion is the number of replicates required for confidence in the data. In general, this depends on the variability between data sets, with CCFRA4 recommending three sensors from each of three replicate runs, and NFPA5 suggesting at least ten working sensors from a run with replicate runs required where variability is found. Further details on heat penetration testing are given in Section 5.4. 5.2.3 Process establishment methods The need for caution in process establishment can be reinforced by reminding ourselves that the calculation of an integrated process value from heat
78
Thermal technologies in food processing
penetration data requires the use of three mathematical models. Each model introduces an error in the calculated process value, thus the potential clearly exists for these errors to be additive. The models are for: • Decimal reduction time (or DT value) for the target organism, which is calculated by using first order death kinetics to model the spore survivor curves. The DT value is the time (usually in minutes) which is required at a constant heating temperature to reduce the numbers of surviving spores by a factor of ten. • Kinetic factor (or z value) for the target organism, which is calculated by using a first order relationship to model the relationship between the DT value and heating temperature. The z value is a measure of the relative ‘killing power’ of the heating temperature, and is the temperature difference required to effect a ten-fold change in the DT value. • Integrated lethal rate, usually referred to as the F0 value (or P-value, Pu), which is calculated by integrating the area beneath the curve obtained when the lethal rates are plotted against time (see Fig. 5.1). It is usual to employ the trapezoidal rule to effect this integration, which is one further model with an associated error.
The calculation procedure outlined above is referred to as the General method and derives from Bigelow et al.6 Process values calculated using the General method should not be considered exact values due to the need for using the three models outlined above, but as estimates and therefore quoted to one decimal place (d.p.). 5.2.4 Setting the target process value Commercial sterility for a thermally processed product is the target condition, which depends on the types and numbers of organisms present before and after the process, and on the intended storage conditions. When the process value (F)
Fig. 5.1
Evolution of lethal rate with time for a sterilised petfood can.
Validation of heat processes 79 is divided by the decimal reduction time (DT) this gives the number of log reductions of surviving spores (see equation 5.1). Ninitial F DT log 5:1 Nfinal where, Nfinal is the final number of organisms after a specific time-temperature history; Ninitial is the initial number of organisms; and DT is the decimal reduction time at a fixed temperature (T) to reduce the number of organisms by a factor of ten (in minutes). For example, a ‘12D cook’ is often quoted as the commercially acceptable minimum for Clostridium botulinum spores in ambient stable products, which assumes a starting position of 1 viable spore per container (N0) and a final container with 10 12 viable spores. In fact, this actually equates to a probability of 1 processed container in 1012 containing a viable spore (N). The 12D minimum botulinum cook is quoted as F03, which is calculated using a D121.1 of 0.21 minutes and rounding up the F-value to the nearest integer. In commercial thermal processing practice, however, it would be common to operate at substantially increased safety margins, which would equate to log reductions upwards of 25. This will result in a fully-sterilised food product. Pasteurisation processes on the other hand are usually operated to only 6 log reductions of the target organism, presumably because of the less lethal nature of the organisms when compared with Clostridium botulinum. Further details on pasteurisation treatments can be found in CCFRA.2 The method for calculating the process values is similar to that for the sterilised products, and utilises equation 5.1. Having calculated the target process value from data on the heat resistance of the target pathogenic or spoilage organisms, the thermal process achieved in the food containers must be validated using temperature or other appropriate inprocess measurements. These methods are discussed in the subsequent sections.
5.3
Temperature distribution testing
Temperature distribution (TD) tests in the heating medium are the first stage in a heat penetration study. However, a uniform TD throughout the retort does not necessarily imply that the lethalities delivered to the product are also uniform, since uniform temperature does not guarantee uniform heat transfer. Therefore, the uniformity in temperature is the minimum that has to be studied and an additional heat distribution study is advisable if there are concerns about air entrapment or heat transfer coefficient reductions throughout a container load. Rotary retorts using excessive air overpressure are one example where the potential exists for air to collect at the crate centres. For a steam retort, if the TD is unsatisfactory this can normally be resolved simply by increasing the length of the period of air removal at the start of the process (venting). This contrasts with non-steam retorts where a large
80
Thermal technologies in food processing
temperature range may be attributed to the design/loading of the retort and simple corrective action is not possible, so testing can only provide the relationship between the retort instrumentation and the retort interior but does not necessarily improve it. 5.3.1 When is TD testing required? The TD within a retort should be tested on its installation, with intermittent retesting being required as factors change that could affect the retort performance. Retorts require, as a minimum, re-testing in the event of any engineering work likely to affect the TD of the retort, such as: • relocation of the retort or installation of another retort that uses the same services • modification to the steam, water or air supply • failure of the key components (for example pumps and valves) • repair or modification to water or steam circulation systems within the retort • if there are any doubts about the performance of the circulation system.
In addition, if the load to be processed in a retort changes, re-testing is required. Such circumstances include the use of: • new container sizes and shapes • new container loading patterns • new crate or layer pad design, or mode of use.
It is also necessary to ensure that a retort’s performance does not deteriorate over a period of time, as corrosion or fouling in the steam, water or air supply pipes builds up. Retort instrumentation and process records should be inspected regularly to identify when a TD problem has arisen. Regular re-testing of a retort’s TD is good practice to ensure that these faults are not overlooked. 5.3.2 What are the objectives of TD tests? Although TD tests basically sample the performance of retort systems, in industry they are frequently used as an opportunity to ‘audit’ the installation to ensure long term compliance. There are no UK regulations of general application to retort processing, although there are specific regulations, derived from EC Directives, which apply to the heat processing of certain products. These contain some implications for thermal processing of these products (e.g. milk, milk products, egg products, meat and fish products). Food manufacturers producing products not covered by these specific regulations must comply with the Food Safety Act (General Food Hygiene Regulations) 1995, implementing EC Directive 931431EC. The Good Manufacturing Practice guidelines for TDs in batch retorts as defined in DH,7 section 10.3.4, are as follows:
Validation of heat processes 81 In steady state operation, the temperature spread across the sterilising vessel should ideally be 1ºC or less. However, when this degree of control is not achievable due to design or characteristics of the equipment, any deviation from the limit should be allowed for in the scheduled process. If the retort uses condensing steam as the media, it is necessary to establish the time of vent in order that the distribution of temperatures across the retort is reduced to an acceptable limit. For venting trials, the following guidance is derived from CCFRA:8 (a) Note precisely the time at which the retort reaches 100ºC. (b) Do not close the main vent until all thermocouples reach the same temperature within 0.5ºC. (c) Close the vent and record when the master temperature indicator and chart recorder reach process temperature. (d) All thermocouples should indicate the same temperature within one minute of the first thermocouple indicating that (process) temperature. (e) Record the venting time as the number of minutes for which the main vent was left open after 100ºC was reached on the thermocouple in the thermometer pocket. The vent test is specific to condensing steam retorts and is required in addition to TD testing. For retort systems utilising water or mixtures of steam and air, the TD tests are unlikely to result in a 1ºC distribution in temperatures across the crates at the start of the hold phase. This is because of less favourable heat transfer coefficients with these heating media when compared with condensing steam and also the reduced quantity of heat available in, for example, a raining water system. It is common practice to quote a time into the hold phase by which the temperature distribution has stabilised to within 1ºC, and to take this into account when establishing the hold time at constant temperature.
5.4
Heat penetration testing
Various methods can be employed to collect accurate heat penetration (HP) data. The aim of an HP study is to determine the heating and cooling behaviour of a specific product in order to establish a safe thermal process regime and to provide the data to analyse future process deviations. Design of the study must ensure that all of the critical factors are considered to deliver the thermal process to the product slowest heating point. The numbers of instrumented sample containers and replicate retort runs have been subject to much discussion (CCFRA4; NFPA5; IFTPS9), with the final decision linked to the measured variability between samples and between runs. Modern datalogging systems can provide the facility for taking multiple temperature measurements, therefore large quantities of data can be taken more
82
Thermal technologies in food processing
easily than was the case when the CCFRA guidelines were written in 1977. These recommendations were for three samples in three replicate runs, providing a total of nine measurements. The more common situation now is to take up to ten samples in two replicate runs, providing that the variability between runs is within acceptable limits. However, there can be limitations on the number of probes that can be inserted through a packaging gland or through the central shaft of a rotating system, and in these situations at least two replicate runs should be completed. 5.4.1 When is HP testing required? The HP study should be carried out before commencing production of a new product, process or package. Changes to any of the criteria that may change the time-temperature response at the product slowest heating point will require a new HP study to be conducted. The conditions determined in the study are referred to as the scheduled heat process and must be followed for every production batch, with appropriate records taken to confirm that this was followed. No further temperature measurement within containers is required in production, although some companies do measure temperatures in single containers at defined frequencies. However, the conditions used in single container testing will not represent the worst case, and it would be expected that the instrumented container would show a process value in excess of that measured from the HP study. Such data are intended to show due diligence and are more of a comfort factor. 5.4.2 What are the objectives of the HP tests? It has been stated above that the HP study should be conducted using worst case conditions. Therefore, by inference it should not be possible for a normal production batch to heat more slowly than the combination of factors evaluated as worst case. To determine the worst case conditions it is necessary to consider the product, process and package separately. The following lists suggest the factors that should be addressed in an HP study, although the lists are not intended to be exhaustive. Product factors • Formulation; weight variation in ingredients, e.g. high starch levels that could lead to increased viscosity. • Fill weight; percent overfill of the key components, e.g. solids content. • Consistency or viscosity of the liquid components; before and after processing. • Solid components; size, shape and weight before and after processing, potential for matting and clumping. • Preparation methods; e.g. blanching. • Rehydration of dried components. • Heating mode; convection, conduction, mixed or broken heating.
Validation of heat processes 83 Container factors • Type; metal cans, glass jars, pouches, semi-rigid containers. • Nesting of low profile containers; geometry of the divider plates. • Vacuum and headspace; residual gases with flexible containers. • Orientation. • Fill method; initial temperature and the effects of delays in getting instrumented containers into the retort • Symmetry of rotation. Retort factors • Type; steam, steam/air, water immersion, raining water. Venting schedule if steam. Overpressure profile if water or steam/air. • Retort come-up time; this should be as short as possible to minimise the quantity of heat absorbed by the product during this phase. • Racking and dividing systems. • Rotation; slowest heating position usually along the retort axis. 5.4.3 Instrumentation for TD and HP testing Modern dataloggers are typically multi-channel systems with digital outputs allowing data to be recorded directly to a laptop PC for display and to maintain permanent records. Thermocouples based on type T (copper/constantan) with PTFE insulation are most common because they are inexpensive, accurate over the desired temperature range, and respond rapidly to changing temperature. Other types, based on a change in electrical resistance with temperature, such as thermistors and platinum resistance thermometers (pt100), are used in dataloggers where the logging unit is remote (e.g. Ball Datatrace and Ellab Tracksense). These are referred to as resistance temperature detectors, or RTDs. Calibration of a temperature sensor against a traceable instrument is essential each time it is used in a set of TD or HP trials. This can be achieved using the master temperature indicator on the retort (MTI) which must be calibrated at no less than six-monthly intervals.7
5.5
Microbiological spore methods
Microbiological spore methods are often referred to as direct methods, but they in fact rely on measuring the achieved log reductions for a process using a nonpathogenic microorganism and converting this to a process value for the target pathogen using equation 5.1. For example, in a sterilisation process where the target was Clostridium botulinum spores, the marker organism could be spores of Bacillus stearothermophilus which are reported to have similar death kinetics, or specifically the kinetic factor or z-value. F-values can also be calculated by integrating the killing power of a thermal process over the time-temperature history experienced by the product, measured
84
Thermal technologies in food processing
using a temperature sensor. An F-value calculated using equation 5.1 will be the same as that calculated from the time-temperature integration, provided that first order kinetics have been followed for the microorganism’s destruction by heat (see equation 5.2).
t Ninitial P 10
T Tref =z dt DT log 5:2 Nfinal 0 where, T is the product temperature, ºC; Tref is the reference temperature for the DT value, ºC; t is the process time, min; and z, the kinetic factor, is the temperature change required to effect a ten-fold change if the DT value (ºC). It is critical that the z-value is close to that for the target microbial species, otherwise significant errors in the calculated process values can be introduced.10,11 If the z-value is not the same as the target value, the processing temperature should be close to the reference temperature, otherwise significant errors can arise between values estimated with organisms and probes. Also, the decimal reduction time should allow sufficient log reductions to be measured in order that the process can be correctly calculated from equation 5.2. A test that results in no surviving spores does not allow the process F-value to be calculated, and raises doubt as to where or when the total kill occurred. A microbiological method can be conducted using organisms distributed evenly throughout a food product or concentrated in small beads. 5.5.1 Inoculated containers This method is also known as the count reduction method and involves inoculating the entire food with organisms of known heat resistance. It is essential that some organisms survive the heat process in order that the containers can be incubated and the surviving organisms counted. The average thermal process received by a container can be calculated using equation 5.1. If the product is liquid it is relatively easy to introduce the organisms but for solid products it is necessary to first mix the organisms in one of the ingredients to ensure that they are dispersed evenly throughout the container. Few studies on sterilisation processes use Clostridium botulinum spores because of the hazards associated with their handling, but also due to the very low number of surviving spores that would result from a commercial process. The theory of a sterilisation process is that three minutes equivalent at 121.1ºC will result in at least 12-log reductions in spores. Almost all commercial processes operate to safety margins many times larger than three minutes, so it would be impractical to incubate the vast numbers of containers required to find the surviving spores. Incubation and testing of full production runs would be required, with little chance of finding the surviving spore. Hence, a nonpathogenic organism with a high D121.1 value is used, such as spores of Clostridium sporogenes or Bacillus stearothermophilus. Typical levels of the inoculum are between 103 and 105 spores per container. An alternative is to use
Validation of heat processes 85 a gas-producing organism and estimate the severity of the process by the number of blown cans. 5.5.2 Encapsulated spores or organisms This method allows the organisms to be placed at precise locations within a container or within the food particulates, by encapsulating known numbers of spores in an alginate bead.12 The alginate bead can be made up with a high percentage of the food material so that the heating rate of the bead is similar to the food. This method has been used for continuous processes where the food contains particulates that require evaluation at their centres, and conventional temperature sensing methods cannot be used. Large numbers of alginate beads are used to determine the distribution of F-values that can occur in continuous processes as a result of the distribution of particle residence times.13 Estimating the exact number to use in a test is not straightforward because it depends on the F-value distribution, which is not known until after the test is conducted and the results analysed. The number of organisms used will greater than for an inoculated container test and can be of the order of 106 per bead, but it is also important that not all are destroyed by the heat process otherwise it is not possible to estimate a process value using equation 5.1. If there are no surviving organisms then it is only possible to conclude that the process achieved greater than 6-log reductions for the example of a 106 initial loading. In this situation, there will be uncertainty as to whether the organisms died as a result of the process, during transportation to or from the factory, or if the spores germinated during the come-up time making them more susceptible to destruction at milder temperatures than for the heat resistant spores. Hence, controlling how these tests are performed is critical and the expertise to conduct a test using encapsulated spores or organisms tends to be restricted to a limited number of microbiology laboratories.
5.6
Biochemical time and temperature integrators
The use of time and temperature integrators (TTIs) as an alternative means of process evaluation, to either temperature or microbial systems, has received considerable attention recently.10–11, 14–15 A TTI can be an enzyme, such as amylase or peroxidase, that denatures (an unwinding of the structure) as it is heated. If the reaction kinetics of the temperature-induced denaturation match those of the microbial death kinetics for the target species, it is possible to use such TTIs as non-biological markers of a process. Many pasteurised products heated in steam-jacketed vessels or in heat exchangers require a microbiological or TTI method to measure the process Fvalues. Table 5.1 presents the key attributes of one such system, an amylase from Bacillus amyloliquefaciens16 that has suitable kinetics for estimating
86
Thermal technologies in food processing
Table 5.1
Key attributes of the Bacillus amyloliquefaciens -Amylase TTI
Category
Description
Operating principle
Reduction in amylase activity in response to time and temperature Amylase assay to measure absorbance rate, using a spectrophotometer 60–100ºC 9.7 0.3ºC D80.7 18.7 min ‘pasteurisation-value’ 0.02 mL
Measurement method Active temperature range Kinetic factor, or z-value Decimal reduction time Process value Sample size
pasteurisation, or P-values, in processes. This TTI has been used recently for studies on fruit processing and poultry cooking where the target processes were within the measurement range. The heat denaturation rate of this amylase is minimal at ambient temperatures, simplifying transportation of the amylase TTI from the laboratory to the factory. P-values for TTIs are calculated from the initial and final activities using equation 5.3, which assumes that the thermal death time model applied. This is the same equation for calculating integrated process values when using microbiological spore methods except that the number of organisms is replaced by the amylase activity. Ainitial P DT log 5:3 Afinal where, Afinal is the final activity after a specific time-temperature history, and Ainitial is the initial activity. Figure 5.2 shows the form that a typical data set can take, in which a large number of amylase cubes were introduced to a well-mixed batch and collected after cooling through a Spiraflo (Tetra Pak) tubular heat exchanger. Each amylase cube contained 20 L of amylase solution at the centre of a silicone cube. A distribution of values was expected and it is important to understand the shape of this distribution in order to estimate the likely lowest value expected in any batch. 5.6.1 Applications of TTIs to retort processes To validate an in-container process for a product containing particulates, a large food chunk or piece is usually attached to the end of a probe, so that temperatures can be measured at the chunk centre during a process. While this is an acceptable method for static processes where thermocouple conduction is the only uncertainty, questions can be raised about the conditions experienced by the chunk during rotary processing that could affect the measured process value. For example, the greater relative velocity of the sauce over its surfaces may lead to increased surface heat transfer coefficients, and the probe hole may physically enlarge to allow hot
Validation of heat processes 87
Fig. 5.2 P-value distribution for a 430 kg batch of 10 mm pineapple fruit preparation, calculated using an amylase TTI with a D85 6.95 minutes. Sample size was 44.
liquid to penetrate. By using a TTI encapsulated in an artificial chunk of an impervious material such as silicone or as a small tube of the TTI inserted directly into the chunk, a more representative process value can be measured.
5.7
Future trends
The use of time and temperature measurements to validate the degree of thermal process achieved is likely to remain the most widely used method. However, the advances in microchip technology will provide more computing power to analyse these results and increase the accuracy in defining the calculated process value. Traditional canning processes have been evaluated using lethal rates at time steps of one minute, because of the capabilities of available data recorders, but modern process values can be estimated from temperature measurements taken at much reduced frequencies, for example at every second. The increased data storage capabilities also allow for more temperature probes to be used in each TD or HP test, with multiple loggers linked together. With such systems it is easily possible to exceed the number of suggested working probes to define an HP test; however, the limitation on the number of probes that can go through the packing gland of a retort remains the same. To overcome this limitation, the diameter of the thermocouple cables must be reduced without compromising the strength or measuring accuracy of the system. The use of more advanced mathematical models to evaluate and predict process times and temperatures will also increase as the computing power
88
Thermal technologies in food processing
available on desk-top PCs increases. Examples of predictive models currently used are the CTemp and NumeriCAL software, both of which utilise a numerical approach to solving complex equations using finite differences. At present, most companies rely on the General method for their scheduled times and temperatures, but this does not allow process deviations to be assessed. A predictive modelling approach would not only help with deciding the fate for batches of product that had undergone a process deviation, but the task of process establishment would be made more straightforward by applying the models to evaluate low initial product temperatures, short come-up times or low retort temperatures. In the introduction to this chapter it was stated that the scope of the market for thermally processed foods is increasing, as the pasteurised foods sector becomes more prominent. This is a trend that will continue because of consumer demand for products of high quality that can only be produced by applying minimal thermal processes. There will always be demand for fully sterilised canned food, but it is unlikely that this industry sector will increase in popularity. To service the needs of the pasteurised foods sector, the methods available to the food processor for validating the process must evolve. Many of these process types cannot be validated using conventional wire-based systems, and so alternative methods are required. For food products manufactured in a cook-chill or cook-freeze system, dataloggers are available that can pass with the product as it moves between ovens and chillers or freezers. The types of products manufactured in this way have traditionally been processed to an end of oven product temperature, but are now being evaluated in terms of a pasteurisation process and the data converted to pasteurisation values. Selfcontained logging units are also a recent introduction, which can be inserted into food containers that are processed in continuous retorts such as hydrostatic or reel and spiral sterilisers. While these units have enabled continuous processes to be evaluated without recourse to process simulators, improvements can be made to their robustness and longevity. Research is active in the application of TTIs to various pasteurisation and sterilisation processes, particularly to those where the food is heated and cooled in large vessels or heat exchangers. These processes cannot be evaluated using wire-based systems. In conclusion, the scope of the thermally processed foods sector will increase to include products, packages and processes of increased complexity, and in response to this, the methods used to validate the process severity will also need to develop. The requirements of the traditional canned foods industry are disappearing, with an exciting new thermal processing sector developing with all of its demands on process validation techniques.
5.8
Sources of further information and advice
Many of the useful texts for advice on thermal process evaluation are given in the list of references, so will not be repeated here. However, the guidelines
Validation of heat processes 89 produced by the Campden & Chorleywood Food Research Association2, 4, 8, 17–19 and the Institute for Thermal Processing Specialists9, 20 are of particular relevance for their use as industry standard documents. The CCFRA guidelines for overpressure retorts17–19 contained information that updated some aspects of Technical Manual No. 3,4 and the IFTPS protocols for carrying out temperature distribution20 and heat penetration20 studies provided detailed listing of factors to consider for TD and HP testing respectively. Both CCFRA and IFTPS documents were produced in consultation with representatives from the food industry. One further reference text produced from an industrial working party is the Department of Health guidelines for establishing safe thermal processes.7 This booklet covers all aspects of thermal processing, from handling of raw materials through the processing stages and dispatch of processed containers.
5.9 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
11. 12.
References BALL C O and OLSEN F C W, Sterilization in Food Technology. Theory, Practice and Calculation. New York, McGraw-Hill Book Co., 1957. CCFRA, Pasteurisation Heat Treatments, CCFRA Technical Manual No. 38, CCFRA, Chipping Campden, Glos., GL55 6LD, 1992. MAY N S, Thermal Underware, Food Manufacture, 1992. CCFRA, Guidelines to the establishment of scheduled heat processes for low-acid foods, CCFRA Technical Manual No. 3, CCFRA, Chipping Campden, Glos., GL55 6LD, 1977. NFPA, Guidelines for Thermal Process Development for Foods Packaged in Flexible Containers, NFPA, Washington, DC, USA, 1985. BIGELOW W D, BOHART G S, RICHARDSON A L and BALL, C O, Heat Penetration in Processing Canned Foods, National Canners Association Bulletin, 16L: 128, Washington, USA. DEPARTMENT OF HEALTH, Guidelines for the safe production of heat preserved foods. HMSO, London, 1994. CCFRA, Canning retorts and their operation, Technical Manual No. 2, CCFRA, Chipping Campden, Glos., GL55 6LD, 1975. IFTPS, Protocol for carrying out heat penetration studies. IFTPS, PO Box 2764, Fairfax, VA, USA, 1995. HENDRICKX M, MAESMANS G, DE CORDT S, NORONHA J, VAN LOEY A and TOBBACK P, Evaluation of the integrated time-temperature effect in thermal processing of foods. Critical Reviews in Food Science and Nutrition, 1995 35 (3) 231–62. VAN LOEY A M, HENDRICKX M E, DE CORDT S, HAENTJENS T H and TOBBACK P P, Quantitative evaluation of thermal processes using time-temperature integrators. Trends in Food Science and Technology, 1996 7 16–26. BROWN K L, AYRES C A, GAZE J E and NEWMAN M E, Thermal destruction of bacterial spores immobilised in food/alginate particles. Food Microbiology, 1984 1 187–98.
90 13. 14. 15.
16. 17. 18. 19. 20.
Thermal technologies in food processing and WITHERS P M, Determination of residence time distribution of non-settling food particles in viscous food carrier fluids using Hall effect sensors. Journal of Food Process Engineering, 1992 17 401–22. MAESMANS G, HENDRICKX M, DE CORDT S, VAN LOEY A., NORONHA J and TOBBACK P, Evaluation of process value distribution with time temperature integrators. Food Research International, 1994 27 413–23. TUCKER G S, Comment calculer les valeurs de pasteurisation dans le produits avec morceaux avec l’integrateur temps-temperature amylase. Symposium Technique International de l’appertise UPPIA/CTCPA `Securite et appertisation: de nouveaux outils pour la maitrise des treatments thermiques’, Paris, 1998. ADAMS J B, Determination of D80ºC for -amylase inactivation. CCFRA Internal Project Report, Ref: 12598/1, CCFRA, Chipping Campden, Glos., GL55 6LD, 1996. CCFRA, Guidelines for batch retort systems – full water immersion – raining/spray water – steam/air. Guideline No. 13, CCFRA, Chipping Campden, Glos., GL55 6LD, 1997. CCFRA, Guidelines for performing heat penetration trials for establishing thermal processes in batch retort systems. Guideline No. 16, CCFRA, Chipping Campden, Glos., GL55 6LD, 1997. CCFRA, Guidelines for establishing heat distribution in batch overpressure retort systems. Guideline No. 17, CCFRA, Chipping Campden, Glos., GL55 6LD, 1997. IFTPS, Temperature distribution protocol for processing in steam still retorts, excluding crateless retorts. IFTPS, PO Box 2764, Fairfax, VA, USA, 1992. TUCKER G S
6 Modelling and simulation of thermal processes B.M. Nicolaı¨, P. Verboven and N. Scheerlinck, Katholieke Universiteit, Leuven
6.1
Introduction
Thermal food processes are one of the few production processes in industry which rely on a mathematical model to ensure the safety of the process.1 For a proper process design, the food centre temperature must be known during the subsequent stages of the preparation process, so that the effect of the thermal treatment on the microbiological and sensory quality can be evaluated using wellestablished methods. Mathematical process models and simulation software represent a powerful alternative to the traditional, time-consuming temperature measurements and quantitative microbiological and food quality analyses. The objective of this chapter is to give an overview of the existing models for conduction and convection heat transfer during food manufacture. As these models can only be solved for very simple problems, numerical solution is usually mandatory. A wide variety of numerical techniques and corresponding software is now available to solve these models. The outline of the chapter is as follows. In Section 6.2 the Fourier equation for conduction heat transfer will be introduced, along with its corresponding boundary and initial conditions. Some analytical solutions will be given. The Navier-Stokes equations which describe convective transport phenomena will be described in Section 6.3. It will be shown how the transport equations can be modified to take into account turbulence effects. Several types of boundary conditions which are relevant to food processes will be described as well. In Section 6.4 several numerical methods to solve heat and mass transfer problems will be introduced, including the finite difference, finite element and finite volume methods. An overview will be given of commercially available software packages. In section 6.5 some illustrative examples will be given.
92
Thermal technologies in food processing
6.2 Modelling of conduction heat transfer: the Fourier equation 6.2.1 Derivation Heat transport in solids is governed by lattice waves induced by atomic motion and is called conduction. Transient heat conduction in an isotropic object with boundary is governed by the Fourier equation2 @T rkrT Q on
6:1 @t where is density (kg m 3), c is heat capacity (J kg 1 ºC 1 ), k is thermal conductivity (W m 1 ºC 1 ), Q is volumetric heat generation (W m 3 ), T is temperature (ºC), t is time (s), and xi is coordinate (m). The thermophysical parameters k, , and c may be temperature dependent so that the problem becomes nonlinear. Thermophysical properties of various agricultural and food products are compiled in various reference books (e.g., the compilation made by the ASHRAE).3 Further, equations have been published which relate the thermophysical properties of agricultural products and food materials to their chemical composition. In general, both the heat capacity and the density can be calculated with sufficient accuracy, but the models for the thermal conductivity require some assumptions about the orientation of the different main chemical constituents with respect to the direction of heat flow which is not always obvious. In conventional thermal food processes the heat generation Q is zero. However, in the case of volumetric heating techniques such as microwave and ohmic heating, Q is the driving force of the heat transfer. The modelling of these techniques is a very active research area.4–7 The initial condition for the Fourier equation can be described as a spatial dependent function at time t 0: c
T
x; y; z; t T0
x; y; z at t 0
6:2
At the boundary of the heated or cooled object, fixed temperature (Dirichlet), convection or radiation conditions may apply: T
x; y; z; t f
x; y; z; t @ 4 T h
T1 T "
T1 T 4 on 6:3 @n? with f (x, y, z, t) a known function (e.g., it was measured, or it is known from control procedures), n? the outward normal to the surface, h the convection coefficient (W/m2 ºC), T1 the (known) ambient temperature, " the emission coefficient, and the Stefan-Boltzmann constant. The surface heat transfer coefficient h must be considered as an empirical parameter. k
6.2.2 Analytical solutions Equation [6.1] can be solved analytically under a limited set of initial and boundary conditions for simple geometries only. Several solution techniques
Modelling and simulation of thermal processes 93 such as separation of variables, Green functions and variational methods are discussed in the many books on partial differential equations.10, 11 A large number of analytical solutions of the Fourier equation were compiled by Carslaw and Jaeger.12 Usually the Fourier equation is rewritten in dimensionless coordinates by introducing a dimensionless temperature # and a dimensionless time Fo which is called the Fourier number T1 T1
6:4
Fo kt=cL2
6:5
#
T T0
with L a characteristic length, e.g., the half-thickness of a slab. For different geometries such as slab, cylinder and sphere, it can be shown that there exists a linear relationship between the logarithm of # and Fo. For example, for a slab of half-thickness L subjected to convection boundary conditions, # is given by
1 X
4 sin
n exp
&n2 Focos
&x 2& sin
2& n n n1
6:6
and the discrete values of n are positive roots of the transcedental equation n tan
n Bi
6:7
where the Biot number Bi is defined as hL 6:8 k For Fo 0:2, it can be shown that the infinite series in equation [6.6] can be approximated by the first term of the series. The graphical representation of the resulting relationship is commonly known as a Heissler chart and can be found in any standard textbook on heat transfer.13 Bi
6.3
The Navier–Stokes equations
6.3.1 Conservation equations In fluids, transport of heat and mass is more complicated than in solid foods, as besides diffusion also convective transport of liquid particles may take place. The driving force behind convective transport is a pressure gradient in the case of forced convection, e.g. due to a fan in an oven, or density differences because of, e.g. temperature gradients. Navier and Stokes independently derived the equations for convective transport which now bear their names. For simplicity we will restrict the discussion to a single Newtonian fluid system. This means that we will only consider fluids for which there is a linear relationship between shear stress and velocity gradient, such as water or air. More complicated fluids such as ketchup, starch solutions, etc., are so-called non-Newtonian fluids, and the reader is referred to standard books on rheology for more details.14
94
Thermal technologies in food processing
When we apply the conservation principle to a fixed infinitesimal control volume dx1dx2dx3 we obtain the continuity, momentum and energy equations, written in index notation for Cartesian coordinates xi (i 1, 2, 3 for the x-, y- and z-direction, respectively), and whenever an index appears twice in any term, summation over the range of that index is implied (for example, @uj =@xj becomes
@u1 =@x1
@u2 =@x2
@u3 =@x3 ): @ @uj 0 @t @xj @ui @uj ui @ @ui @uj @ 2 @uj p fi @xj @xj @xi @xi 3 @xj @t @xj @H @uj H @ @T @p Q k @xj @t @xj @xj @t
6:9 6:10 6:11
where ui (i 1, 2, 3) Cartesian components of the velocity vector U (m s 1), T is temperature (ºC), H is static enthalpy (J kg 1), p is pressure (Pa), is density (kg m3), k is thermal conductivity (W m 1 ºC 1), is dynamic viscosity (kg m 1 s 1), fi is external body forces (N m 3), and Q is heat source or sink (W m 3). For a full derivation of these equations we refer to any textbook on fluid mechanics.15, 16 The system of five equations (Eq. [6.9]–[6.11], three equations for the velocity components plus the continuity and the energy equation) contains seven variables (u1, u2, u3, p, h, T, ). We therefore need additional equations to close the system. The thermodynamic equation of state gives the relation between the density and the pressure p and temperature T. The constitutive equation relates the enthalpy h to the pressure and the temperature. For an ideal gas we can use the following equations: pM RT @H c @T p
6:12 6:13
with M the molecular weight of the fluid (kg mol 1) and R the universal gas constant (J mol 1 K 1). When the heat capacity is assumed constant, the constitutive equation reduces to a linear relation between H and the difference between the actual temperature T and a reference temperature. Since only relatively low velocities are encountered in the food processes under consideration, the flow is often assumed incompressible and these equations can be applied. For isothermal fluids we can assume that the density is constant so that the continuity equation vanishes. In the case of non-isothermal flows the Boussinesq approximation is often applied, in which it is assumed that density is the only parameter which depends on the temperature.16
Modelling and simulation of thermal processes 95 6.3.2 Turbulence Many heat transfer processes in food operations often involve turbulent flow of air or water. Turbulence can be induced by the presence of flow obstructions such as baffles, shelves and the foods themselves. Turbulence is a state of the flow which is characterised by fluctuations of the flow variables (eddies) over a large range of scales, both in time and space. This complex pattern of motion enhances heat transfer rates considerably but also causes additional pressure drops which must be taken into account in the design of the equipment. Turbulence must therefore be incorporated in the governing models unless a laminar flow regime can be guaranteed. Although the Navier–Stokes equations are general conservation equations which are equally well applicable to turbulent flow, the large variation of spatial scales introduces severe numerical problems, and only for simplified cases and low Reynolds numbers is it currently possible to perform such direct numerical simulations on supercomputers.17 Simulation shortcuts are possible at different levels of complexity and approximations. The least approximations are needed in large eddy simulations, in which case the largest eddies are resolved but the effects of smaller eddies are estimated by additional models.16 This approach is now being used more widely, since it is almost within reach of current computer power. The most popular approach is based on the Reynolds Averaged Navier– Stokes (RANS) equations, which are obtained from averaging out the governing equations (Eq. 6.10) and including the effect of the turbulent fluctuations by additional models for the new terms appearing in the RANS equations. In the Boussinesq approach, the turbulence is accounted for by a turbulent ‘viscosity’ which is incorporated in the viscous and thermal diffusion transport terms. In K– " models, originally proposed by Jones and Launder, the turbulent viscosity is obtained as a function of the turbulent variables K, which represents the turbulent kinetic energy associated with the fluctuating components of the flow velocities, and e, the turbulent energy dissipation rate:18 K2 6:14 " The constant C may be assumed constant for equilibrium conditions, where the turbulence production nearly equals the turbulence dissipation. Additional transport equations have been derived for these turbulent flow variables. Several undefined constants appear in the model equations, which together with several assumptions and the specific near-wall treatment render this model empirical. There are three popular K–" models, namely the standard k–" model, a RNG (Renormalisation Group) k–" model and a LRN (Low Reynolds Number) k–" model.18–20 Verboven et al. compared these three turbulence models for a typical forced convection heating process of complexly shaped foods, and concluded that the boundary layers are badly represented by the wall function approach and the departure from local equilibrium is not accounted for.21 A correction function can be added to correct for the latter t C
96
Thermal technologies in food processing
behaviour in conjunction with the Low Reynolds Number model (see the work of Yap).22 Nevertheless, it was found that experimental input for these corrections is needed in order to determine important constants. More complex closures for the RANS models are based on dynamic equations for the Reynolds stresses and fluxes themselves in the RANS equations. In addition to the equations for the mean flow, this approach results in seven more partial differential equations. These models are believed to be more accurate but require a better insight into the process of turbulence and care must be taken with their numerical solution. Finally, it must be noted that new turbulence models are constantly proposed and tested. 6.3.3 Initial and boundary conditions Unlike the diffusion equations, there are no conclusive general rules for the implementation of boundary conditions for the Navier–Stokes equations in order to have a well-posed problem because of their complex mathematical nature. For a full account, we refer to Hirsch.23 For incompressible and weakly compressible flows, it is possible to define Dirichlet boundary conditions (fixed values of the variables, mostly upstream), Neumann boundary conditions (fixed gradients, mostly downstream) and wall boundary conditions (a wall function reflecting the behaviour of the flow near the wall). Initial values must be provided for all variables. Difficulties arise when the exact conditions are unknown. This is especially true in turbulent flows, where the exact values of the turbulence energy and energy dissipation rate are often unknown at the inlet, and need to be guessed using information about the velocity and the flow geometry. The direction of the flow at boundaries may be difficult to specify, but may have considerable influence when the flow contains swirls. The effect of the pressure resistance (e.g. in a cool room) on the fan flow rate may be considerable and cannot always be taken into account appropriately. In any case a sensitivity analysis can be useful to obtain an error estimate associated with approximate or guessed boundary conditions. 6.3.4 Additional equations In the case of an air flow through a bulk of products (e.g. cooling of horticultural products), or a flow of multiple fluids (e.g. the dispersion of disinfectants in a cool room or the injection of water for air humidification), the separate phases need to be considered. Depending on the flow conditions, a multi-phase modelling or a mixed-fluid modelling approach can be applied. The reader is referred to the literature for more details.24 When the problem contains chemical kinetics (e.g. microbial inactivation), an additional transport equation must be introduced. Further, the chemical reaction must be solved. Therefore the reactions rates, property changes and heat releases must be calculated as part of the solution. The heat of reaction can be calculated
Modelling and simulation of thermal processes 97 from the heats of formation of the species and depends on temperature. The reaction leads to sources/sinks in the species conservation and energy equations. For more details, see reference 25.
6.4
Numerical methods
6.4.1 Numerical discretisation For realistic – and thus more complicated – heat and mass transfer problems usually no analytic solution is available, and a numerical solution becomes mandatory. For this purpose the problem is reduced significantly by requiring a solution for a discrete number of points (the so-called grid) rather than for each point of the space-time continuum in which the heat and mass transfer proceed. The original governing partial differential equations are accordingly transformed into a system of difference equations and solved by simple mathematical manipulations such as addition, subtraction, multiplication and division, which can easily be automated using a computer. However, as a consequence of the discretisation the obtained solution is no longer exact, but only an approximation of the exact solution. Fortunately, the approximation error can be decreased substantially by increasing the number of discretisation points at the expense of additional computer time. Various discretisation methods have been used in the past for the numerical solution of heat conduction problems arising in food technology. Among the most commonly used are the finite difference method, the finite element method, and the finite volume method. It must be emphasised that – particularly in the case of nonlinear heat transfer problems – the numerical solution must always be validated. It is very well possible that a plausible, convergent but incorrect solution is obtained. At least a grid dependency study must be carried out to verify whether the solution basically remains the same when the computational grid is refined. 6.4.2 The finite difference method Principle The finite difference method is the oldest discretisation method for the numerical solution of differential equations and had been described as long ago as 1768 by Euler. The method is based on the approximation of the derivatives in the governing equations by the ratio of two differences. For example, the first time derivative of some function T(t) at time ti can be approximated by dT T
ti1 T
ti 6:15 dt ti t with t ti1 ti . This expression converges to the exact value of the derivative when t decreases. The power of t with which the so-called truncation
98
Thermal technologies in food processing
error decreases is called the order of the finite difference approximation, and can be obtained from a Taylor series approximation of T at time ti. Equation [6.15] is called a forward difference as it uses the future value of the function and it is of order 1. A backward difference of order 1 is given by dT T
ti T
ti 1 6:16 dt ti t Equations [6.15] and [6.16] are also called forward and backward Euler schemes. Likewise, finite difference formulas can be established for second order derivatives. The so-called central difference formula is of order 2 and is defined by d 2 T T
ti1 2T
ti T
ti 1 6:17 dt2 ti t2 The finite difference method will be illustrated for a two-dimensional heat conduction problem. For this purpose the computational domain is subdivided in a regularly spaced grid of lines which intersect at common nodal points (Fig. 6.1). Subsequently, the space and time derivatives are replaced by finite differences. For example, if central differences are used it is easy to see that the following expression is obtained for the Fourier equation: @Ti; j kt Ti1; j 2Ti; j Ti 1; j Ti; j1 2Ti; j Ti; j 1 6:18 c @t
x2
y2 Similar equations can be established for all interior nodes of the grid, and special procedures are available to discretise the boundary conditions in the nodes which are on the boundary of the grid. The large number of equations
Fig. 6.1 Finite difference grid of a two-dimensional rectangular region. The nodes which are involved in the computation of the temperature at position (i,j) are indicated by dots.
Modelling and simulation of thermal processes 99 (equal to the number of nodal points) can conveniently be ordered into a differential system of the general form d u Ku f 6:19 dt with u [u1 u2 uN]T the nodal temperature vector. This vector differential equation can be discretised in time, and typically leads to a system of algebraic equations which must be solved by appropriate means. The system matrices contain many zeros, and this feature can be exploited advantageously to reduce the required number of computations. C
6.4.3 The finite element method Principle In the finite element method,26 a given computational domain is subdivided as a collection of a number of finite elements, subdomains of variable size and shape, which are interconnected in a discrete number of nodes. The solution of the partial differential equation is approximated in each element by a low-order polynomial in such a way that it is defined uniquely in terms of the (approximate) solution at the nodes. The global approximate solution can then be written as a series of low-order piecewise polynomials with the coefficients of the series equal to the approximate solution at the nodes. Substitution of the approximate solution in the differential equation produces in general a non-zero residual. In the Galerkin method, the unknown coefficients of the low-order piecewise polynomials are then found by orthogonalisation of this residual with respect to these polynomials. This results in a system of algebraic or ordinary differential equations which can be solved using the well-known techniques. The Galerkin finite element method A first step in the construction of a finite element solution of a partial differential equation is the subdivision of the computational domain in a grid of finite elements, which are interconnected at a discrete number of common nodal points. The elements may be of arbitrary size and shape. A large number of element shapes have been suggested in the literature and are provided in most commercial finite element codes. Typical 2D and 3D element shapes are shown in Figs 6.2 and 6.3. The unknown solution is expressed in each element as a piecewise continuous polynomial in the space coordinates with the restrictions that (i) continuity between elements must be preserved and (ii) any arbitrary linear function could be represented.26 In general, the unknown temperature field T (x, y, z, t) can then be approximated by T
x; y; z; t NT
x; y; zu
t
6:20
with N a vector of so-called shape functions and u a vector containing the temperatures at the nodes of the finite element grid. In general the approximate
100
Thermal technologies in food processing
Fig. 6.2
Typical 2D and 3D finite element shapes.
temperature field T is not identical to T, and when T is substituted in the heat conduction equation, a non-zero residual r is obtained: r c
@T @t
rkrT
Q
6:21
This residual is subsequently orthogonalised with respect to the shape functions N:
@T rkrT Q d 0 N c 6:22 @t
It can be shown that after the application of Green’s theorem and some matrix algebra a system of the form [6.19] is obtained.26–28 C and K are now called the capacitance matrix and the stiffness matrix, respectively; f is the thermal load vector. The matrices C, K and f are constructed element-wise. As in the case of the finite difference method, the system [6.19] is solved using traditional finite
Modelling and simulation of thermal processes
Fig. 6.3
101
3D finite element grid for a food container. Because of symmetry reasons only a quarter of the food container needs to be modelled.
difference methods. Note that K and C are positive definite, symmetric and banded. These very important features can be exploited advantageously to significantly reduce the computational effort and memory requirements. Special attention has been paid recently to stochastic finite element methods which were developed to take into account random variability of product and process parameters.29, 30, 31 6.4.4 The finite volume method Principle The finite volume method of discretisation is most widely used in commercial CFD (computational fluid dynamics) codes at the moment. It owes its popularity to the fact that it obeys the clear physical principle of conservation on the discrete scale. The concepts of the method are easy to understand and have physical meaning. The system of general conservation equations can be written in coordinatefree notation and integrated over a finite control volume V with surface A. Applying Gauss’s theorem to obtain the surface integral terms, the equations have the following form, with the transported quantity:
@ dV
UndA
rndA S dV 6:23 V @t A A V This equation states the conservation principle on a finite scale for all relevant quantities in the system when the surface integrals are the same for volumes
102
Thermal technologies in food processing
sharing a boundary. Moreover, the finite volume form of the model becomes independent on the coordinate system. When the physical domain is subdivided into control volumes, a grid only defines the boundaries of the volumes. This is advantageous for modelling complex geometries. The volume integrals are approximated in terms of the volume-centered value of . The values at the volume faces are required for solving the surface integrals in equation [6.23]. This requires interpolation in terms of volumecentered values. Some interpolation schemes may be highly accurate, but produce unbounded solutions when grids are too coarse. Others are unconditionally stable, but have a low accuracy and produce erroneous results called numerical or false diffusion. The reader is referred to the literature for a more elaborate discussion about the limits and benefits of different approximating formulas.15, 16 The time discretisation in the control volume method is carried out using finite differences in the time domain, explained above. Solution of the discretised equations Discretisation results in the following set of equations, in matrix-vector notation: Au Q
6:24
where A is a square sparse matrix containing the coefficients resulting from the discretisation, u is a vector containing the unknowns at the control volume centres and Q is a vector containing the variable-independent source terms. Equation [6.24] is still non-linear: the flow variables appear in the coefficients. An iterative method is therefore required in which the non-linear terms have to be linearised. The least expensive and most common approach is the Picard iteration. In this method coefficients are updated using the most recent solution of the system. This approach requires more iterations than Newton-like methods, which use a Taylor series expansion, but do not involve the computation of complex matrices and are found to be much more stable. The solution of the linearised equations can be performed by direct methods, which are computationally very costly and generally do not benefit from the mathematical properties of the linear system. It is therefore advantageous to use an iterative method. The iterative method should have certain properties in order to guarantee a valid solution. The main requirement for convergence of the solution is that the matrix A be diagonally dominant, which has been shown by Scarborough:32 P jAnp j 1 at all P 6:25 jAP j < 1 at all P at least where np are the neighbouring nodes of the node P. Several iterative solvers are available. A detailed discussion is given by Ferziger and Peric.16 To verify the validity of the mathematical solution, the solution change during the iterative procedure should be monitored. One can then stop the
Modelling and simulation of thermal processes
103
iteration process, based on a predefined convergence criterion and be assured of a convergent solution of the discretised equations. The convergence error e nc can be defined as:16 e nc u
un
6:26 n
where u is the converged solution of equation [6.24] and u is the approximate solution after n iterations. It is not possible to obtain e nc directly and it is even hard to calculate a suitable estimation of the value. In practice, the residual rn can be used to test for convergence: Au n Q
rn
6:27
When the residual goes to zero, the convergence error will be forced to decrease as well, because: Ae nc rn
6:28
The reduction of the norm of the residual is a convergence criterion to stop the iterations. The residual should be reduced by three to five orders of magnitude. It may happen that the residual decreases much faster than the actual convergence error, in which case care should be taken and the iteration procedure continued. 6.4.5 Commercial software Some general characteristics of finite difference, finite element and finite volume methods are compared in Table 6.1. Because of their generality, most CFD codes are currently based on finite element and finite volume methods. Most commercial CFD codes for fluid flow analysis are available on UNIX as well as NT platforms. Parallel versions are often available as well. Some of the Table 6.1 method
Characteristics of the finite difference, finite element and finite volume Finite Finite element difference
Finite volume
Geometry of problem Boundary conditions Nonlinearities Complexity Main application area
simple difficult difficult low general
moderately complex easy easy high computational fluid dynamics
Physical background
none
Availability of commercial few codes Price low
complex easy easy high mechanical engineering depends on discretisation many
physical balances satisfied many
high
very high
104
Thermal technologies in food processing
commercial codes dedicated to CFD analysis are described below. Some general-purpose numerical codes, like ANSYS (Ansys Inc, Swansee, USA) also include CFD features, but are mainly intended for structural and conduction heat transfer analysis. The main packages are CFX/TASCflow (AEA, Harwell, U.K.), Fluent/FIDAP (Lebanon, NH, USA), PHOENICS (Cham Ltd., London, U.K.), STAR-CD (Computational Dynamics Ltd., London, U.K.). The basic functionality of these packages is very similar.
6.5
Applications
6.5.1 Design of thermal processes The design of the thermal sterilisation process is well-established and is based on the analysis of the heat penetration in the sterilised food and the kinetics of thermal inactivation of microorganisms. Teixeira et al.33 suggested solving numerically the Fourier equation by means of the finite difference method, and to use the computed centre temperature as an input for the calculation of the process lethality by numerical integration. As a further improvement, the use of time-varying retort temperature profiles was considered34 in order to maximise the retention of thiamin while safeguarding the required process value. This eventually led to the STERILMATE software package for computer-aided design of sterilisation processes.35 More elaborate computer-aided optimisation procedures have been described in the literature.36, 37 Applications of the finite element method include the simulation of conduction heat transfer in foods with complicated geometrical shapes such as chicken legs,38 a baby food jar,39 broccoli stalks,40 tomatoes,41 and lasagna.42 Integrated software packages have appeared which incorporate heat transfer simulation models with microbial kinetics and food process engineering knowledge. The CookSim package43 was essentially a knowledge-based system to guide the user towards a safe thermal process design by automatically solving the mathematical models underlying the heat transfer process and the associated microbial kinetics. A related approach was followed in the development of the ChefCad package for computer-aided design of complicated recipes consisting of consecutive heating/cooling steps.44–46 The data and knowledge base contains the declarative and procedural knowledge of the system. The declarative knowledge encompasses all the data in the system, including the current recipe, a list of food ingredients (the complete food table is in the system), species of microorganisms and the parameters of their growth/ inactivation models, equipment types such as ovens and refrigerators, etc. The procedural knowledge base contains finite element routines for the numerical solution of 2D heat conduction problems, an automatic finite element grid generator, routines to calculate the thermophysical properties from the chemical composition of the food, routines to calculate the surface heat transfer coefficient of the heating/cooling fluid, differential equation solvers for the microbial growth/inactivation and texture changes. The inference engine is the
Modelling and simulation of thermal processes
Fig. 6.4
105
Main window of the ChefCad package.
core of the system. It is a part of the programming environment and contains procedural knowledge for making logical inferences. It is not immediately accessible to the programmer. The inference engine processes the user requests which arrive through the user interface. The necessary declarative data are fetched from the data and knowledge base, and passed to the calculation routines which are then fired. The calculation results are then transferred back to the user interface for visualisation. Also, a microbial safety diagnosis of the recipe is made by inferencing appropriate rules. The main window of the package is shown in Fig. 6.4. 6.5.2 Design of forced convection ovens Non-homogeneous heating of foods may cause considerable microbial risks and a large non-uniformity of the food quality. Industrial appliances for the convection heating of foods have been shown to produce a strongly inhomogeneous distribution of the main processing parameters, such as the processing temperature and the surface heat transfer coefficient (see, for example, the work of Sheard and Rodger47). The heat transfer implications of an airflow that varies in direction as well as in magnitude may be large, especially at low velocities: it has been shown that when the surface heat transfer coefficient is small, small deviations in its value may result in large deviations in the food product temperature.48 It is expected that this non-uniformity can be attributed to a bad distribution of the heating medium, which results from an
106
Thermal technologies in food processing
Fig. 6.5 Calculated time-temperature profiles at the food centre, during condensation heating with different values for the air humidity and an ambient temperature of 60ºC.50
improper oven design. For example, even at a mild value of the air velocity of 1.5 m/s and a turbulence intensity of 38%, a velocity deviation of 5% will lead to a typical product centre temperature deviation of 1ºC, if a food is heated from 5ºC to 90ºC, using a first order sensitivity analysis of the Fourier equation for heat conduction in foods with fluctuating boundary conditions.49 Even more pronounced is the effect of an improper distribution of moisture in the air. When, due to an inhomogeneous flow distribution, local humidity levels differ, very large temperature deviations can be expected, as given in Fig. 6.5. This effect is due to different rates of moisture condensation onto the product surface at different humidity levels. This was a study performed by means of a model for combined airflow, heat and mass transfer during heating of a flat surface of a food (thickness 5 cm, length 30 cm, heating from 5ºC to 60ºC, air velocity 4 m/s). The model was implemented and solved using the commercial CFD code CFX (AEAT, Harwell, UK).50 As a consequence of these findings, it is required to have a comprehensive tool to optimise the airflows and consequently the design of ovens. A CFD model of an oven was developed and solved using the commercial code CFX (AEAT, Harwell, UK).51, 52 The model consists of the Reynolds-averaged conservation equations of mass, momentum and energy, a k–" turbulence model and additional equations to describe the fan operation, including its rotation, and heat transfer from the heating coils. Model calculations were validated against hot film velocimetry (TSI, St Paul, MN) and temperature measurements by means of thermocouples. The model was solved by means of the commercial CFD code CFX (AEAT, Harwell, UK). The code employs a finite volume formulation for the numerical solution of the conservation equations. Hybrid
Modelling and simulation of thermal processes
107
Fig. 6.6 Calculated airflow and temperature distribution in an oven cavity, arrows: air velocity vectors (white: > 6 m/s, black: 0 m/s), contours: temperature distribution (grey: < 100ºC, white: > 105ºC).
differencing is used for the convection terms, central differencing for the diffusion terms. Time stepping is performed using the implicit backward scheme. A fine body-fitted structured grid with 55,944 rectangular volumes (15 20 70) was used. The mesh has control volumes of dimensions approximately equal to 0.025 m by 0.025 m by 0.01 m in each direction. The flow pattern and the temperature distribution are shown in Fig. 6.6. Recognise the hot spots (light grey contours) and the regions of lower velocity (small arrows). Figure 6.7 shows the temperature response of polymer bricks subject to forced convection heating in the oven, compared to experiments. Obviously, in quantitative terms, the CFD model lacks some accuracy, due to complex features such as turbulence and fan rotation. However, the cold and hot spots are correctly predicted by means of this CFD model.
108
Thermal technologies in food processing
Fig. 6.7 Measured (black symbols) and calculated (clear symbols) centre temperature responses of polymer bricks at different positions in the oven: top (R), bottom (P).
6.6
Conclusions
Because of the increasing power of computers, food engineers now have tools to solve complicated heat transfer problems involving a variety of heat transfer mechanisms. The reliability of the numerical solution largely depends on the availability of suitable thermophysical properties and the complexity of the governing models. The numerical solution of convective transport problems described by the Navier–Stokes equations remains a difficult task, particularly when turbulence is involved. The empirical constants involved in the most popular turbulence models necessitate a careful validation of the results obtained. The currently available finite difference, finite element or finite volume codes can only be used appropriately by highly trained engineers who have a formal knowledge of both heat transfer theory and numerical discretisation methods. There is, hence, a need for more user friendly software packages which can also be used by operators who have less knowledge about the physical mathematical details so that they can concentrate on the actual application. These software packages should combine heat transfer simulations with prediction of microbial growth/inactivation and quality changes, and HACCP. That this is at least possible for relatively simple heat conduction problems is shown by the ChefCAD package.
Modelling and simulation of thermal processes
6.7
109
Acknowledgements
The authors wish to thank the European Union (projects FAIR-CT96-1192 and INCO IC15 CT98 0912) and the Flemish Government (COF 99-003) for financial support.
6.8 1. 2. 3. 4.
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15.
References HOLDSWORTH S D,
‘Optimisation of thermal processing – a review’, Journal of Food Engineering 1985 4 89–116. INCROPERA F P, DE WITT D, Fundamentals of Heat and Mass Transfer, 3rd ed. New York, Chichester, Brisbane, John Wiley & Sons, 1990. ANON., ASHRAE Handbook – Fundamentals. Atlanta, American Society of Heating, Refrigeration and Air-Conditioning Engineers, 1993. DATTA A K, PROSETYA H M, HU W, ‘Mathematical modeling of batch heating of liquids in a microwave cavity’, in Yano, T., Matsuno, R., Nakamura, K., eds, Developments in Food Engineering. London, Glasgow, Weinheim, Blackie Academic & Professional, 1994, 325. OHLSSON T, ‘In-flow microwave heating of pumpable foods’, in Yano, T., Matsuno, R., Nakamura, K., eds, Developments in Food Engineering. London, Glasgow, Weinheim, Blackie Academic & Professional, 1994. ZHANG L, LIU S, PAIN J-P, FRYER P J, ‘Heat transfer and flow in solid–liquid food mixtures’, Food Engineering in a Computer Climate. New York, Philadelphia, Hemisphere Publishing Company, London, 1992, 79. DE ALWIS A A, FRYER P J, ‘A finite element analysis of heat generation and transfer during ohmic heating of food’, Chemical Engineering Science, 1990 45 1547–60. BIRD R B, STEWART W E, LIGHTFOOT E N, Transport Phenomena. New York, John Wiley & Sons, 1960. NI H, DATTA A K, ‘Moisture, oil and energy transport during deep-fat frying of food materials’, Trans IChemE, 1999 77(C) 194–204. ZAUDERER E, Partial Differential Equations of Applied Mathematics. Singapore, New York, Chichester, John Wiley & Sons, 1989. KEVORKIAN J, Partial Differential Equations – Analytical Solution Techniques. Pacific Grove, California, U.S.A, Wadsworth, Brooks and Cole, 1990. CARSLAW H S, JAEGER J C, Conduction Heat Transfer in Solids, 2nd ed. London, Oxford University Press, 1959. MILLS A F, Heat Transfer. Homewood, Irwin, 1992. CROCHET M J, DAVIES A R, WALTERS K, Numerical Simulation of NonNewtonian Flow, Rheology series 1. Amsterdam, Elsevier, 1984. VERSTEEG H K, MALALASEKERA W, An Introduction to Computational Fluid Dynamics: The Finite Volume Method. Harlow, Longman Scientific & Technical, 1995.
110
Thermal technologies in food processing
16.
FERZIGER J H, PERIC M,
17. 18. 19. 20. 21.
22. 23. 24. 25. 26. 27. 28. 29. 30.
31.
Computational Methods for Fluid Dynamics. Heidelberg, Springer-Verlag, 1996. MOIN P, KIM J, ‘Tackling turbulence with supercomputers’, Scientific American, 1997 276(1) 46–52. JONES W P, LAUNDER B E, ‘The prediction of laminarisation with twoequation model of turbulence’, Int J of Heat Mass Transf, 1972 15 301–13. YAKHOT V, ORSZAG S A, THANGAM S, GATSKI T B , SPEZIALE C G, ‘Development of turbulence models for shear flows by a double expansion technique’, Phys Fluids A, 1992 4(7) 1510–20. LAUNDER B E, SHARMA B I, ‘Application of the energy dissipation model of turbulence to the calculation of flow near a spinning disc’, Lett Heat and Mass Transf, 1974 1 131–8. VERBOVEN P, SCHEERLINCK N, DE BAERDEMAEKER J, NICOLAI¨ B M, ‘Performance of different turbulence models for the calculation of the surface heat transfer coefficient’, Conference/Workshop on Thermal Properties and Behaviour of Foods during Production, Storage and Distribution, Prague, Food Research Institute, 1997. YAP C, Turbulent heat and momentum transfer in recirculating and impinging flows. Ph.D. Thesis, Manchester, Faculty of Science and Technology, 1987. HIRSCH C, Numerical Computation of Internal and External Flows. Vol. 1: Fundamentals of Numerical Discretization. Chichester, John Wiley & Sons, 1991. XU Y, BURFOOT D, ‘Predicting condensation in bulks of foodstuffs’, Journal of Food Engineering 1999 40 121–7. TOLEDO R T, Fundamentals of Food Process Engineering. New York, Van Nostrand Reinhold, 1991. ZIENKIEWICZ O C, The Finite Element Method. Maidenhead, New York, St. Louis, San Francisco, McGraw-Hill Book Company, 1977. CONNOR J J, BREBBIA C A, Finite Element Techniques for Fluid Flow. London, Boston, Newnes-Butterworths, 1976. SEGERLIND L, Applied Finite Element Analysis. New York, Chichester, John Wiley & Sons, 1984. NICOLAI¨ B M, DE BAERDEMAEKER J, ‘A variance propagation algorithm for the computation of heat conduction under stochastic conditions’, International Journal for Heat and Mass Transfer, 1999 42 1513—20. NICOLAI¨ B M, VERBOVEN P, SCHEERLINCK N, DE BAERDEMAEKER J, ‘Numerical analysis of the propagation of random parameter fluctuations in time and space during thermal food processes’, Journal of Food Engineering, 1999 38 259–78. NICOLAI¨ B M, VERLINDEN B, BEUSELINCK A, JANCSOK P, QUENON V, SCHEERLINCK N, VERBOVEN P, DE BAERDEMAEKER J, ‘Propagation of stochastic temperature fluctuations in refrigerated fruits’, International Journal of Refrigeration, 1999 22(2) 81–90.
Modelling and simulation of thermal processes 32. 33. 34. 35.
36. 37.
38. 39. 40. 41. 42.
43. 44. 45.
46.
SCARBOROUGH J B,
111
Numerical Mathematical Analysis, 4th ed. Baltimore, University Press, 1958. TEIXEIRA A A, DIXON J R, ZAHRADNIK J W, ZINSMEISTER G E, ‘Computer optimization of nutrient retention in thermal processing of conductionheated foods’, Food Technology, 1969 23 137. MANSON J E, ZAHRADNIK J W, STUMBO C R, ‘Evaluation of lethality and nutrient retentions of conduction-heating foods in rectangular containers’, Food Technology, 1970 24 109–13. KIM K H, TEIXEIRA A A, BICHIER J, TAVARES M, ‘STERILMATE: software for designing and evaluating thermal sterilization processes’, ASAE paper no. 93–4051, St. Joseph, Michigan, USA, American Society of Agricultural Engineers, 1993. BANGA J R, PEREZ-MARTIN J M, GALLARDO J M, CASARES J J, ‘Optimization of the thermal processing of conduction-heated canned foods: study of several objective functions’, Journal of Food Engineering, 1991 14 25–51. SILVA C M, HENDRICKX M, OLIVEIRA F, TOBBACK P, ‘Critical evaluation of commonly used objective functions to optimize overall quality and nutrient retention of heat-preserved foods’, Journal of Food Engineering, 1992 17 241–58. DE BAERDEMAEKER J, SINGH R P, SEGERLIND L J, ‘Uses of the finite element method in food processing’, J Food Process Eng, 1977 45 1547–59. NAVEH D, KOPELMANN I J, PFLUG I J, ‘The finite element method in thermal processing of foods’, J Food Sci, 1983 48 1086–93. JIANG H, THOMPSON D R, MOREY R V, ‘Finite element model of temperature distribution in broccoli stalks during forced-air precooling’, Transactions of the ASAE, 1987 30, 1473–7. PAN J C, BHOWMICK S R, ‘The finite element analysis of transient heat transfer in fresh tomatoes during cooling’, Trans ASAE, 1991 34 972–6. NICOLAI¨ B M, VAN DER BROECK P, SCHELLEKENS M, DE ROECK G, MARTENS T, DE BAERDEMAEKER J, ‘Finite element analysis of heat conduction in lasagna during thermal processing’. International Journal of Food Science and Technology, 1995 30(3) 347–64. RACE P R, POVEY M J, ‘CookSim: a knowledge based system for the thermal processing of food’, Expert Systems and their Applications. Avignon, France, 1990 115–27. SCHELLEKENS M, MARTENS T, ROBERTS T, MACKEY B M, NICOLAI¨ B M, VAN IMPE J F, DE BAERDEMAEKER J, ‘Computer aided microbial safety design of food processes’, International Journal of Food Microbiology, 1994 24 1. NICOLAI¨ B M, VAN IMPE J F, SCHELLEKENS M, ‘Application of expert systems technology to the preparation of minimally processed foods: a case study’, Journal A Benelux Quarterly Journal of Automatic Control, 1994 35 50–5. NICOLAI¨ B M, OBBELS W, SCHELLEKENS M, VERLINDEN M, MARTENS T, DE BAERDEMAEKER J, ‘Computational aspects of a computer aided design package for the preparation of cook-chill foods’, Proceedings of the Food
112
47.
48. 49.
50.
51.
52.
Thermal technologies in food processing Processing and Automation Conference III, Orlando, USA, ASAE, St. Joseph, 1994, 190–9. SHEARD M A, RODGER C, ‘Optimum heat treatments for ‘‘sous vide’’ cookchill products, FLAIR, Food-linked Agro-industrial Research 1989–1993. First European ``Sous vide’’ Cooking Symposium Proceedings, Leuven, Belgium, 1993, ALMA Sous Vide Competence Center, 117–26. NICOLAI¨ B M, DE BAERDEMAEKER J, ‘Sensitivity analysis with respect to the surface heat transfer coefficient as applied to thermal process calculations’, Journal of Food Engineering, 1996 28 21–33. VERBOVEN P, SCHEERLINCK N, DE BAERDEMAEKER J, NICOLAI¨ B M, ‘Sensitivity of the food temperature with respect to the air velocity and the turbulence kinetic energy’, Journal of Food Engineering, in press, 2000. VERBOVEN P, Investigation of the uniformity of the heat transfer to food products in ovens by means of Computational Fluid Dynamics, PhD Thesis, Faculty of Agricultural and Applied Biological Sciences, K.U. Leuven, Leuven, Belgium, 1999. VERBOVEN P, SCHEERLINCK N, DE BAERDEMAEKER J, NICOLAI¨ B M, ‘Computational fluid dynamics modelling and validation of the isothermal airflow in a forced convection oven’, Journal of Food Engineering, 2000 43 41–53. VERBOVEN P, SCHEERLINCK N, DE BAERDEMAEKER J, NICOLAI¨ B M, ‘Computational fluid dynamics modelling and validation of the temperature distribution in a forced convection oven’. Journal of Food Engineering, 2000 43 61–73.
7 Modelling particular thermal technologies S. Bakalis, P. W. Cox and P. J. Fryer, University of Birmingham
7.1
Introduction
7.1.1 Types of thermal process Thermal processing is a ubiquitous operation in food processing. Many of the commonest food processing operations, such as canning, baking and pasteurisation, rely on heating: • In a number of cases the effect of heat is intended for preservation alone, i.e. to kill bacteria and inactivate enzymes, such as in the pasteurisation of milk and the sterilisation of canned food. In this case the aim is to deliver the required microbial kill with as little damage to the structure of the food as possible. • In the processing of foods, such as meats and vegetables, heat acts also to develop taste and flavour, so that in addition to sterilisation heat is required to carry out physical changes to the food. • However, in many other situations food is heated to develop the structure of the material, such as in baking of bread or biscuits, where heating acts both to change the starch structure and function and also to develop the bubble structure within the material (for examples, see Campbell et al., 1999). • There are also situations, such as drying and frying, in which heat transfer is accompanied by mass transfer, and the two effects must be considered as coupled; evaporation of moisture in drying requires heat transfer to provide the necessary heat.
Each of the above points has been the subject of whole books! – to attempt a full summary is impossible. The modelling of each type of process has followed the same general trend:
114
Thermal technologies in food processing
• purely empirical correlations or graphical solutions • models based around very simple approximations, such as simple geometries (spheres or cylinders) and uniform physical properties • finite difference models, in which the basic equations are discretised on a simple grid system • finite element (FE) and finite volume models, in which more complex geometries can be used.
As computer power has increased, FE models have become easier to solve, the codes have become more robust and more accurate, and the computers on which the codes run have become cheaper. For simplicity, this review will concentrate on the case of heating for microbial cook alone, and will not consider taste or texture development. Hopefully it will show how modelling is being used and how modelling techniques can be applied in this area. The problem with using heat for sterilisation is in demonstrating that the material is safe; models are critical in minimising the amount of experiments that have to be done and in convincing regulatory bodies. 7.1.2 Basic equations Microbiological and quality kinetics The rate of thermal processing is commonly quantified through the integrated lethality of a thermal process calculated using equation (7.1) (from the work of Ball, 1923):
t T
t Tref F 10 dt
7:1 z 0 where z is the increase in temperature that gives an increase in rate of a factor of 10, and Tref is a reference temperature. F has the units of time; it is the length of time that the food would have to be held at the reference temperature to obtain the same effect as the actual process with T. The thermal history of the processed product may equally be applied to product sterility or nutritional quality (where z will become a rate of cooking and not microbial lethality). Equation (7.1) is difficult to justify other than as an experimental fit; it is a local approximation to the Arrhenius expression, only accurate over a narrow temperature range, and it is not clear whether the death of microorganisms follows the Arrhenius kinetics. Thermal transport equations Heat is transferred by three mechanisms: conduction through solids or stationary liquids or gases, convection through flowing fluids, and radiation. For conduction and convection the rate of heat transfer is proportional to the temperature difference, whilst for radiation it is the difference between the fourth power of the temperatures. There are a multitude of good books on heat transfer, amongst them ¨ zisik (1993) and Carslaw and Jaeger (1980). Incorpera (1981), O For products where conduction is the sole mechanism of heat transfer the temperature profile may be estimated from the partial differential equation:
Modelling particular thermal technologies
115
@T r
rT
7:2 @t the solution of which requires knowledge of the spatial variation of the thermal conductivity, , the density, and the specific heat, cp, of the product. At the edges of the solid, different boundary conditions may apply. The simplest is constant temperature; however, a heat transfer boundary condition is often necessary, in which the flux to the surface is given by, for example, a convective heat transfer coefficient, or by radiation. The overall rate of heating of a solid will depend on consecutive processes; heat must move to the product and then within it. The relationship between external and internal thermal transport can be estimated using the Biot number: cp
hd
7:3 where h is the interfacial heat transfer coefficient and d some characteristic dimension of the body being heated. The higher the Biot number, the greater is the effect of heat transfer coefficient; in practice, a Bi > 10 implies that the slowest heat transfer process will be conduction within the solid particle. For a low Biot number (< 1) the process is controlled externally, with the solid essentially isothermal. The heating of fluids is more complex because of fluid motion, so that both thermal and fluid transport equations need to be solved. Solution of the NavierStokes equation is needed for the flow field. Simplified equation sets are often used; for example, in a tubular geometry the partial differential equations describing the heat and momentum transport are (Bird et al., 1964): Bi
Equation of continuity: 1@ @u
rv 0 r@ @z Equations of motion: @u @u v u @r @z @v @v v u @r @z
@P 1 @ @v @u @ @u r 2 @z r @ @z @r @z @z @P 2 @ @v @ @v @u r @r r @ @r @z @r @z
Equation of energy: @T @T 1 @ @T @2T u Cp v 2 @r @z r @r @r @z The assumptions used in deriving these equations are: • • • • •
that the flow is axisymmetric there is negligible thermal generation by viscous dissipation the effects of natural convection are also negligible the liquid is homogeneous constant density, specific heat and thermal conductivity.
7:4
7:5
7:6
116
Thermal technologies in food processing
The types of boundary conditions usually applied are: for both velocity and temperature, a known profile at the inlet of the heater, with a known temperature profile at the wall of the heater and cooler, with a no-slip boundary for the velocity. In the case where a holding tube is used an adiabatic boundary condition is applied at the wall. The next sections review the modelling of heat transfer to packed and flowing foods using both conduction/convection and heat generation models.
7.2
Processing of packed and solid foods
7.2.1 Introduction The classical method of thermal preservation is canning, still the basis of a very large industry, despite the reduced quality of many canned products. Canned food is not sterile when it is packed; filled cans are exposed to a temperature-time profile sufficient to give a safe product. However, any process within which heat is applied externally will cause the centre of the product to lag the surface and bulk product temperatures. Therefore, estimates of temperatures at the slowest heating point are required if equation (7.1) is to be used effectively and, more importantly, safely. The heating rate at the slowest heating point may also be used to ensure that a product is not over processed and therefore experience an impractical, in terms of quality and process costs, thermal history. Estimates of F values using equation (7.1) to describe integrated lethality were first proposed by Ball (1923). Within this work with canned low acid foods (pH < 4.5), the proposed reference temperature (Tref) was 121.1ºC and a Z value of 10 was used. The Z value was determined from the slope of a decimal reduction time against temperature graph for the thermally tolerant spores of the spoilage pathogen Clostridium botulinum. A thermal process was then considered to be safe if the slowest heating point of a can reached an F value of three minutes. Mathematical modelling allows estimation of temperature at the slowest heating point. In addition, using a model it is possible to explore processing variation upon the calculation of a desired F value and therefore process operation. This is particularly true if a new product or processing method is to be established, especially where a priori knowledge of the product and process variability is not available. Although tables exist to predict the measured temperature responses and thermal diffusivities of various foods and packages (e.g., in Tucker and Holdsworth, 1991), many practical and theoretical investigations have involved the examination of simplified regular geometries, such as cylinders or spheres (e.g. Kim and Teixeira, 1997). It was proposed that for conducting solids, and for nonflowing conducting liquids, a geometric simplification was valid as the thermal diffusivity () need only be estimated across the diameter of a cylinder describing the shortest chord of the original solid. Paramount to successful modelling is availability of accurate physical data with which models can be constructed and used for process simulations. In addition to static isothermal estimates of
Modelling particular thermal technologies
117
parameters, such as density, viscosity, thermal conductivity and specific heat capacity, knowledge of how these parameters change with heating and process time must also be used. For example, it is difficult to estimate the thermal diffusivity () and the interfacial heat transfer coefficient h in the same experiment; should be inferred from heating curves where an infinite surface heat transfer coefficient can be assumed, i.e. with condensing steam (Kim and Teixeira, 1997). 7.2.2 Conduction in simple solids: solutions of the conduction equation Before personal computers became commonplace, calculating the integrated lethality was complex and time consuming even for geometrically simplified conduction cooked products. In place of the repetitive manual calculations ‘simpler’ methods were used, for example, F value estimates could be made by first measuring temperature (using a thermocouple), calculating the lethality and then examining the area under a lethality rate versus process time graph, which could be directly correlated with F (for examples of this see Lopez, 1987). However, for products where conduction is the sole mechanism of heat transfer the temperature profile can be estimated using equation (7.2). Graphical solutions for this partial differential equation can be produced for various situations of simple-shaped solids and constant thermal diffusivity, e.g. the Heisler and Gurney-Lurie charts, examples of which can be seen in Toledo (1991). These charts plot dimensionless temperature () against Fourier number (t/R 2) for the inverse of various Biot numbers (1/Bi). Solutions for the appropriate geometry allows the temperature at the geometric centre to be estimated. However, these methods employ simplifying assumptions: • internal heat transfer is solely by conduction: external heat transfer is by uniform external heat transfer coefficient or wall temperature • only simple shapes (spheres, cubes, cylinders) can be solved • the product is homogeneous and isotropic • the initial temperature is uniform.
Akterian (1999) summarised the application of equation (7.2) to conductive heat transfer in terms of a partial differential equation, incorporating a shape factor G to account for a product with symmetrical geometry (equation 7.7a). Equation (7.7b) describes the boundary conditions for surface convection and equation (7.7c) is the boundary condition for the line of symmetry. 2 @T @ T G @T
7:7a @t @x2 x @x @T @x
h
T Tm @T 0 @x
7:7b
7:7c
118
Thermal technologies in food processing
Such a simplified modelling approach has been employed to examine the thermal inactivation of bacteria in model foods, for example Bellara et al. (1999) modelled the growth of pathogenic bacterial species after thermal treatment within agar cylinders. Investigations modelled the inactivation of the bacteria and explored the notion that slow heating rates ( 2ºC/min) may increase thermotolerance in potentially pathogenic bacteria (Mackey and Derrick, 1987; Quintavalla and Campanini, 1991). The model could predict successfully the rate and position of bacterial destruction across the radius of the model sausage, using equation (7.8) and predict the reduction in bacterial numbers, N. dN 2:303
T 10 dt Dref
Tref =z
N
7:8
Hendrickx et al. (1993) describe how thermal processing of solids might be optimised, using computer-based finite difference calculations. Early attempts using digital computers include Timbers and Hayakawa (1967), Hayakawa (1969) and Teixeira et al. (1969). Hendrickx et al. (1993) then extended the discussion and development of a model for variable sterilisation temperatures. Numerical simulations could be used to examine and explore varying factors such as the target F value, Z value, product quality (Zq), initial product temperature and retort come up time. However, the model of Hendrickx et al. (1993) still used simple product geometries, i.e. infinite cylinder, slab and spheres, and the assumed simplifications, detailed above, for the exploration of various process parameters. 7.2.3 More complex models: non-uniformity and convective flows Real products are rarely of a regular geometry, have thermal properties which vary with temperature and have different heat resistances along the boundary. For example, in retorts, where condensing steam is used as the heating method, condensation may adversely affect the uniformity of heat transfer to the product surface; heat transfer to a dry surface will be very high, but the presence of a film of liquid will reduce the heat transfer rate (Verboven et al., 1997). Non-isotropic aspects of conductive cooking have been addressed by Pan et al. (2000) in the modelling of the cooking of frozen hamburgers. Their approach, which involved unequal cooking to both the major external surfaces of the patty, considered the enthalpy changes associated with the melting of ice and fat as well as resulting mass transfer effects. The base equation used in the model (equation 7.9) was considered to be valid throughout the whole structure of the hamburger patty regardless of whether it was frozen or not, or whether it was a surface crust or the bulk of the patty. @H @ @T
H
H
7:9 @t @x @x
Modelling particular thermal technologies
119
where the specific enthalpy function is dependent on and T. It was claimed that consideration of both enthalpy changes and of mass transfer effects made the proposed final model more accurate than obtainable with a heat transfer model alone. FE methods of examining heat transfer now allow partial differential equations to be explored in complex geometries. For example, Tewkesbury et al. (2000) used a computational model to predict the cooling of chocolate within a polycarbonate mould. The use of commercially available software, in this case FIDAP (Fluid Dynamics International Inc., Evanston, Illinois), allowed the authors to model the conduction cooling of chocolate through the mould, and took into account the different thermophysical properties of each component, particularly the change in the effective specific heat capacity of chocolate as a function of temperature and cooking rate. The model developed by Tewkesbury et al. (2000) gave a solution for a non-isotropic system, although the properties of each of the components were constant, and was able to predict accurately the heat conduction between the chocolate, the mould and to the environment. FE techniques no longer require very complicated packages; Fig. 7.1 shows the solution of the conduction equation for heat transfer for a complex product geometry using the PC package Matlab (The MathWorks Inc., Natick, MA). The two-dimensional solution (i.e. assuming a transverse section of an infinite slab) assumes that the product has the thermal properties of water, an initial temperature of 20ºC and a boundary temperature of 100ºC. The simulation is
Fig. 7.1 Partial differential equation solution for heat transfer into a complex product. The key indicates the final temperature after 100 seconds.
120
Thermal technologies in food processing
over 100 seconds and the temperature profile takes only five seconds to calculate. Importantly, for a large proportion of commercially processed products, heating is rarely by pure conduction. A prime example of this is canning, where liquids or soft solids of varying viscosities are routinely processed in batch retorts; heat transfer occurs by a combination of conduction and natural convection within the can. Computational fluid dynamics (CFD) can be used to solve for the flow field. A large body of literature exists to describe this complex situation. Reviewed here are several recent papers that demonstrate progression to the current limit of understanding of such systems. All of the papers cited contain useful broad introductions to the area. Kumar et al. (1990) and Kumar and Bhattacharya (1991b) describe the heating of canned viscous liquids that have temperature dependent viscosity. Within their CFD model an initial conductive heat transfer phase reduces the viscosity of the liquid at the periphery of the can. This less viscous and more buoyant hot liquid rises and then re-circulates to the centre of the can. The equations that describe this action (continuity – equation (7.4), energy conservation – equation (7.6) and momentum in the radial direction – equation (7.5)) and the effects of these phenomena are discussed more fully below in the context of fluid movement within pipes. In naturally convecting cans the momentum equation for the liquid movement incorporates a natural convection term (i.e. the density driving force g; Abdul Ghani et al., 1999a): @u @u @u @p 1@ @u @2u v u
7:10 r 2 g @t @r @z @z r @r @r @z Through their model Abdul Ghani et al. (1999a) provided an estimate of the axial velocity of fluid elements. In carboxy-methyl cellulose it was estimated to be 10 4 10 5 ms-1 and for water: 10 2 10 1 ms 1. The differences in magnitude were explained by the ratio of the buoyancy force to the viscous force in each case (expressed as the Grashof number). However, the rates quoted for water have not been observed by these authors when using a positron emission tomography method to study can heating, albeit at lower temperatures than those reported (unpublished results: Bakalis, Cox and Fryer; also see section 7.5). Additionally, Abdul Ghani et al. (1999b) introduced a rate for bacterial inactivation within convecting liquids, combining convection and reaction terms within a single equation: @C @C @C 1@ @C @2C v u De
7:11 kt C r 2 @t @r @z r @r @r @z All cans contain some headspace, i.e. air above the surface of the fluid. As Kumar et al. (1990) pointed out headspace may perform one of two roles. When water is heated within a can containing a headspace the free space may quickly become saturated and give a high heat transfer rate as if it were condensing
Modelling particular thermal technologies
121
steam. In contrast, the space above a viscous liquid may not become wholly saturated and therefore the heat transfer rate would be reduced (< 70 Wm 2 K 1). Abdul Ghani et al. (1999a) suggest that the headspace acts as an insulator, and that natural convection within a can may induce a temperature difference of approximately 10 12ºC depending on the rate of liquid movement and headspace insulation. A secondary point was made that the bottom surface of a can – resting on a retort crate, for example – may not receive as much heat as the side wall increasing the temperature difference. As a consequence of convection and fluid displacement the slowest heating point is no longer the geometric centre; it was found to be within the bottom tenth of the can height and not on the axis of symmetry. Differences in external heat transfer have prompted the simulation of the temperature distributions throughout batch retorts. As pointed out above, the boundary conditions experienced by a product may have a profound effect upon the estimation of the integrated lethality, and reports of temperature deviation of up to 10ºC within a process can be found (Akterian et al., 1998). Additionally, model loads have been investigated, albeit as conduction-heated products, within the virtual processes. Akterian (1999) and Akterian et al. (1998) suggested the use of previously calculated sensitivity functions, these factors correct for arbitrary fluctuations in the heating medium during different parts of the heating holding and cooling cycles. Incorporating these functions with an estimate of the bulk temperature deviation, estimated from the expected lethality rate, allowed for the process to be controlled using a simple microprocessor controller. Ryckaert et al. (1999) linked a process model to a PID controller that operated an industrial oven. The modelling of the heat balance with the heated chamber was used with temperature estimates from 32 thermocouples throughout the oven volume. The modelled temperature distribution was then used to tune the PID terms and so improve the oven control through the different heat input rates that correspond to different parts of the process cycle. Varga et al. (2000) used an FE method to model horizontal cascading water retorts, and verified their model with industrial scale measurements. It was stated that a reliable mathematical solution required a model that should be combined with an appropriate statistical method to estimate real-world variability, as well as an understanding of the quantitative implications of the variance upon the system. Similarly, Verboven et al. (2000) used CFD to examine temperature deviations within an industrial scale forced convection oven. The model could predict oven temperatures to within 5ºC of measured values (albeit in a trial with a reduced product load), although it was pointed out that because of the assumptions made about wall effects, in order to make the computation time nonprohibitive, at present the model was of qualitative use only. What remains as a clear goal is the accurate modelling of the inside of an oven or retort with a full load of convection heated products with temperature dependent physical properties. This aspiration, although not yet realised in the literature, may soon be reached as computation becomes quicker and models more accurate and descriptive.
122
Thermal technologies in food processing
7.3
Continuous heating and cooling processes
Continuous thermal processing was developed to solve the problems associated with traditional batch processing, such as low heat transfer rates, the long processing times needed to achieve the required lethality and the resulting high nutritional losses. The advantage of continuous thermal processing lies in the fact that the activation energy of microbial death is higher than that of the reactions occurring during thermal destruction of nutrients. As a result, high temperature short time (HTST) processes offer the potential to give the same level of sterility for a reduced quality loss. It is also possible to get higher heat transfer rates in fluids flowing through heat exchangers. Aseptic processes first appeared in 1927 while the first patent was granted to Ball in 1936 (Ramaswamy et al., 1995). Commercial application of the technology was used extensively only after the introduction of aseptic packaging and flexible packages, in the 1960s – however, packing methods for high-solids fraction foods are still not well developed. Depending on the nature of the processed food the physics of the process are quite different. Therefore, simulations will be examined separately for liquid and particulate foods. In both cases a variety of approaches have been used. Historically, early investigators used over-simplifying assumptions in order to derive analytical solutions more recently researchers have used discretised techniques to solve the partial differential equations describing the problem. 7.3.1 Sterilisation of liquid foods Continuous sterilisation of foods is commonly used for products like milk, juices, sauces and soups. The application of this technology has gained a lot of attention with the introduction of aseptic packaging, which resulted in self-stable products in convenient flexible packages. In a typical sterilisation process the product passes through a heater, in order to inactivate the bacteria and is then cooled. In many cases sterility is achieved in a holding tube placed between the heater and the cooler. Although a number of heat exchangers are used in the industry, most of the published mathematical models use a simplified pipe geometry. The process variables in this case are the product flow rate (typically 100 l/min), the pipe diameter (typically 0.03 m), the temperature of the heater (typically 140ºC) and the physical properties of the fluid (rheological properties, specific heat capacity and thermal conductivity). In most cases relevant to industrial applications the flow is laminar. The concept of F value, originating from traditional canning, is also used in aseptic processing to describe microbial death. For the non-isothermal case of continuous flow the equation for the total lethality F at a radial position r over an axial length L, can be modified as follows:
t
L 1 T
r; t T0
r; tZ T
r;t T0
r; tZ F
r; L 10 dt 10 dz dz=dt 0 0
Modelling particular thermal technologies
L 10 0
T
r; z T0
r; zZ
123
1 dz u
r; z
7:12
To derive this equation a variable transformation was performed. This equation assumes that bacterial spores remain on the streamlines throughout the process. The lethality given by equation (7.12) is used for the slowest heating zone, i.e. the centre line. A bulk (volume average) lethality is often estimated as follows:
R
R F
r; zu
r; z2dr F
r; zu
r; z2dr ^ F
z 0
R
7:13 0 Q_ u
r; z2dr 0
For first order kinetics of nutrient destruction the same equations can be used to calculate the nutrient retention. The lethality and the average lethality can be estimated by using the appropriate Z value. Typically for nutrient retention bulk lethality is estimated. If the velocity and temperature fields are known, one can use the above equation to estimate the microbial death and the nutrient retention. Viscosity is a strong function of temperature: the flow pattern and thus the residence time at different radii depends on the temperature distribution. There has been no reported analytical solution for the above system of equations for temperature dependent viscosity for a pipe flow. Simpson and Williams (1974) developed an analytical method for the design of aseptic processes by neglecting the effect of temperature on viscosity. The authors suggested a total dimensionless length ( z/(R 2Uavg) of 1.2, with 0.8 for the heating section and 0.4 for the cooling (a holding section was not considered). Based on their calculations this value is accurate within 2% for all power law fluids with 0.3 n 1 and 4 4, where characterises the strength of dependence on temperature. The dimensionless length was estimated to give appropriate inactivation of Clostridium botulinum spores at the centreline. It is important to note that the structure of the dimensionless length , i.e. the length of the pipe used during aseptic processing, should increase proportionally to the square of the pipe radius. This relationship does not consider any effects of the wall temperature. Kumar and Bhattacharya (1991a) used a commercial finite element program (FIDAP) to simulate the process for a shear thinning fluid with temperature dependent viscosity. Lethality was calculated along the centreline. The length of the heater was selected through a trial so a sterility of at least 6 min was reached at the exit. The heater was followed by a 10 m cooling section. The mesh used consisted of 25 nodes in the radial direction not equally distributed. In the axial direction 100 nodal points were used for the first two metres of the heating section, for the rest of the pipe the spacing between two nodes was equal to 0.025 m. Nine node isoparametric elements were used. The variables considered in this study were the wall temperature in the heating section, the tube diameter and the flow rate. As expected the length of the heating section increased with an increase in flow rate, for a given temperature and diameter but decreased with an increase in temperature.
124
Thermal technologies in food processing
Jung and Fryer (1999), again using FIDAP, demonstrated how to optimise the quality of food in a sterilisation process for a Newtonian and a shear thinning fluid. Both lethality and quality were estimated as bulk averages. The agreement of numerical predictions with various analytical solutions was tested. A constant length of 12 m was considered for the heater while the length of the holding section was estimated so that a final sterility of 3 min was reached at the exit. Fifty elements were used in the radial direction. The authors concluded that the common practice of the food industry to estimate lethality, assuming a Newtonian isothermal flow could lead to significant overprocessing. In addition, the conventional High Temperature Short Time treatment could fail under some circumstances. A heater temperature of about 170ºC was found to be optimal for the conditions studied; this lies above the usual operating temperature for food processes. Recently Liao et al. (2000), again using FIDAP, investigated the sterilisation of a starch suspension. The effect of gelatinisation on viscosity was included. Microbial death was calculated at the shortest heating zone, while a volume average was used for the nutrient retention. The authors did not include any heat needed for the starch gelatinisation in the energy equation. In addition, even at temperatures higher than the gelatinisation temperature, where the viscosity is relatively low, velocity distributions appear to be fairly uniform, indicating high viscosity values. The problem of sterilisation of liquid foods in tubular heat exchangers is well understood. Future areas of interest include studies of foods with complex rheological properties, such as slip on the wall and yield stress. In addition, application of the existing models to industrial situation for optimisation is expected to minimise cost, and improve the nutrient content and quality of the processed foods. 7.3.2 Sterilisation of foods containing particulates Especially since approval has been granted by the FDA, aseptic processing of foods containing particulates has become of great interest to the food industry. There are a number of aseptically processed foods, such as soups, that contain solid fractions up to 60% with sizes typically between 3 and 20 mm (Lareo et al., 1997a, b). Aseptic processing is expected to lead to a new range of self-stable products in flexible packages that are preferable to the institutional and food service sector when compared to the commonly used #10 cans. Although the flow of a single particle has been studied extensively, particulate flows are not very well understood (Lareo et al., 1997c). Complicated particle-particle interactions and the often-complex rheological properties of the liquid, result in a non-uniform and often unpredictable flow. In order to study thermal treatment of foods, data such as the heat transfer coefficient between the fluid and the particle are needed. Although a number of correlations for the heat transfer coefficient have been published in the literature (Barigou et al., 1998) the accuracy of these models for food systems is questionable. Furthermore, particles, as with the fluid, experience a wide range
Modelling particular thermal technologies
125
of residence times. In order to ensure a safe product the time temperature history of the fastest moving particle should be considered. Locating this particle in real-life cases is not trivial. As a result it is often a common practice in the food industry to assume a laminar particle flow, which results in overprocessing. In order to model particulate flows a set of equations describing the motion of the fluid and the particles has to be solved simultaneously. The motion of fluid is described from the continuity and momentum equations (equations (7.4) and (7.5)), while the motion of particles is given from a force balance on each particle. A description of the forces acting on the particles is given by Sastry et al. (1989). The energy equations for the particles and the fluid have to be solved to estimate temperature at various locations. One has to keep in mind that the process is transient. From the above it is clear that solving the full problem for realistic process conditions poses a tremendous computational challenge. A number of investigators have tried to simulate the process, often using over-simplifying assumptions. In an early attempt, Manson (1974) assumed infinite fluid-particle heat transfer coefficient and demonstrated the importance of the residence time distribution. Larkin (1989) suggested a modification of Ball’s method, in order to predict the sterility in the middle of a particle. Sastry (1986) developed energy balances for a continuous steriliser. Thermal balances were used to estimate the temperature profile along the tube, assuming that solid and liquid velocities were the same, and the effect of these profiles on different particles (size and residence time) was studied. The author concluded that most of the lethality is taking place in the holding tube and that increasing particle size requires a longer tube length. Mankad et al. (1995) presented a model to study the importance of slip velocity, i.e. the difference between the velocity of the liquid and the solid. The model was based on energy and mass balances for the liquid and the solid phases. The resulting equations are as follows: Energy balance for the liquid: Heat change by thermal @Tf @ 2 Tf ax 2 @t @x Accumulation of heat
Heat exchange with the wall
vf
@Tf 1 2hw
Tw Tf hp a
Tf @x
Cf
1 act R t
Heat change by convective fluid flow
Energy balance for the particle (heat conduction): @Ts 1 @ @Ts s 2 r2 R p @r @t @r
Ts
7:14
Heat exchange with a particle
7:15
126
Thermal technologies in food processing
The equations were solved as an initial boundary value problem, for steady state conditions using implicit and explicit finite difference methods. The effect of the slip velocity and fraction of the delivered solids on the length of the heating tube required to heat the solid and the liquid to 100ºC and the residence time were examined. The authors showed that increase on slip velocity increased the fluidparticle heat transfer and reduced the required heating tube length while increasing the delivered solids concentration had the opposite effect. Limitations of the model arise from the assumption of uniform radial profile for the velocity and the temperature. Furthermore, the authors acknowledged the difficulty in predicting accurately the heat transfer coefficient between the particle and the fluid. This model was extended by Mankad and Fryer (1997) to account for more realistic flows occurring in particulate foods. The flow was divided in two regimes, a sedimented bed with a low velocity and a low fraction region above it. The temperature and the velocity of the fluid were uniform in each zone. The model reveals the importance of slip velocity upon the temperatures of the two phases. Flows where the velocity differences were minimised appear to be best in terms of process time. Recently Sandeep et al. (2000) simulated an isothermal two-phase flow in straight and helical holding tubes. The motion of fluid was described by the continuity and the three momentum equations. A source term added in the momentum equations accounted for the effect of the particle on the fluid. The translation of particles was predicted from three linear dynamics equations and the rotation from three angular dynamics equations. The equations were solved using a finite difference scheme (fourth-order four-stage explicit Runge–Kutta). Although the authors used a shear-thinning fluid (CMC) the form of Navier– Stokes used was for constant properties. No boundary conditions were used for the fluid particle interface. In general, it is not clear if the authors differentiated between the nodes used for the fluid and the particles, and as a result it could be possible for the fluid and the particles to share the same physical space. This limits the application to low particle concentrations. The authors demonstrated that the existence of particles enhanced the secondary flow that improves the mixing characteristics and product uniformity. An increase of the particle diameter or flow rate appeared to narrow the residence time distribution (RTD). The geometrical characteristics of the tubes appeared to have minimal effect on the RTD (less than one second for both average residence time and difference between fastest and slowest moving particle). Aseptic processing has great industrial potential. The limited understanding of the physics makes optimisation very challenging, in practice, over processed products, with inferior quality characteristics often being produced. The problem of sterilisation of liquid foods in tubular heat exchangers is well understood. Future areas of interest include studies of foods with complex rheological properties, such as slip on the wall and yield stress. Experimental investigation of particulate flows will give greater insight to the phenomena occurring and provide means to validate existing numerical simulations. One area of research that shows promise is the use of magnetic resonance imaging (MRI) to provide
Modelling particular thermal technologies
127
temperature maps of foodstuffs in the sterilisation of foods containing particulates (Sun et al., 1993, 1994; Hulbert et al., 1997; Kantt et al., 1998). Increasing computational power will provide the means to handle multiple particles flowing under non-isothermal conditions. In addition, application of existing models to industrial applications for optimisation will minimise cost, and improve the nutrient content and quality of the processed foods.
7.4
Heat generation methods: ohmic and microwave heating
Conventional thermal processes are limited to the standard three ways of heating, by convective, conductive and radiative transport. In practice, conduction is a slow process, which limits the practical heating rates of solid foods and of foods in packages, such as cans. Alternative methods have been sought in which heat is supplied using other techniques; here, heat generation methods are briefly considered. In these, heat is generated by the material in situ as the result of interaction with an external field. This field can be applied by shear (for example, the viscous heating of a solid during shear in an extruder barrel) or by some external electric field. The two best-studied examples of heat generation processes are those of microwave and ohmic heating, in which external electric fields are used. In microwave heating a high frequency field is passed through the food, stimulating the vibrational frequencies of chemical bonds to heat the material: details of the process are found in the excellent review of Metaxas (1996). In ohmic heating, an electric current is passed through a food material which then heats as a result of its inherent electrical resistance; a review is given by Fryer and Davies (2001). Modelling these processes is useful for process and product designers: the need is to show that uniform heating can be provided to commercial products, to ensure safety and optimise product quality. In both cases, the conventional thermal conduction equation within a solid: @T r2 T
7:16 @t must be modified by inclusion of a source term Q, the amount of heat generation per unit volume: cp
@T r2 T Q
7:17 @t Within a liquid, a convective mixing term must be included; where there is fluid motion or relative motion between particles and liquids the full Navier–Stokes equation must be solved in the appropriate geometry. This is complex in itself, even if the heating term were constant. However, the heating term results from the presence of the field; variations in local field strength can thus result in different local heating rates. It is necessary to solve simultaneously for both the cp
128
Thermal technologies in food processing
electric field and for the thermal field that results; this is difficult both theoretically and computationally. The source term in microwave heating is due to the interaction of the external field and the material; this is a function of the field frequency and strength and the ability of the food to absorb the microwave energy. The two Maxwell’s equations used in deriving the field equations for microwaves are Ampere’s law: r H E
@D @t
7:18
relating the variation of the magnetic field H to the electric flux density D and the electric field E, and Faraday’s law rE
@B @t
7:19
which relates E to the flux density B. For a dielectric material this becomes the wave equation: r2 E "
@2E @t2
7:20
and the power dissipated per unit volume is 1 Q e jEj2
7:21 2 The source term in electrical resistance (ohmic) heating is of the same form but is here due to the resistance of the food: Q E2
rV 2
7:22
where E is the voltage gradient and the electrical conductivity. To calculate this requires solution of Laplace’s equation for the voltage field within a system in which the electrical conductivity varies with position: r
rV 0
7:23
throughout the material. In both cases the equations are coupled through the temperature dependence of the physical properties of the system: electrical conductivity and permittivity. Temperature variation results from variation in the field: in microwave heating there is a penetration depth as a result of absorption of the external field, which is generally written as p Dp
4:8=f "0 ="00
7:24 where the penetration depth is in centimetres when the frequency f is in GHz, and "0 and "00 are the dielectric constant and the dielectric loss factor respectively. In electrical heating the heating variation occurs as a result of inhomogeneous distribution of the electric field, which is distorted by the presence of different
Modelling particular thermal technologies
129
electrical conductivities, such as conductors and insulators. Current flows around an insulator, and through a conductor; this distorts the uniformity of the field. A number of attempts have been made to model these processes and to demonstrate the complexity of the heating patterns that can result; these are summarised below. 7.4.1 Microwave heating Excellent discussions of the problems inherent in modelling microwave processes, and in validating them for commercial production, are given by Bows (2000) and Bows et al. (1999). The interaction between the applied field and the food material is key. Multimode resonant applicators, used in all domestic ovens and almost all industrial applications produce one resultant heating pattern from the complex interaction of the field, material and applicator. Strategies to minimise temperature differences (rotating turntables, mode stirrers or moving the product) are not always sufficient to overcome undesirable effects. These can include runaway heating when thawing frozen foods (Buffler and Stanford, 1995), centre focusing in spheres and cylinders (Ohlsson and Risman, 1978) and edge or corner overheating in trays (Bows and Richardson, 1990). For frozen food heated in a domestic oven: thermal runaway heating is observed when thawed areas preferentially absorb a greater proportion of the microwave field than the remaining frozen areas. Thermal runaway is chiefly a property effect; as a material’s dielectric properties change (with temperature for some food materials, but particularly on thawing), the electric field pattern also changes. Bows et al. (1999) describe a novel method of microwave heating, referred to as phase control. The method uses constructive interference techniques: when the microwave fields interacting within a product are coherent (vector addition heating), many heating patterns can be generated at an instant in time, and more controlled heating of food can be carried out. A 3D finite element time domain code, using edge elements, was used to simulate phase controlled heating using the method developed by Dibben and Metaxas (1994). The model solves Maxwell’s equations in 3D and includes a waveguide input feed with a surface excitation plane. The time domain was used because it was found that the conductivities of foods caused ill-conditioned matrices in the frequency domain. Good agreement between the images and experiments was found; however, the code was complex, with a mesh of 73 500 tetrahedral elements in the foodstuff and 135 000 tetrahedral elements in the whole solution domain. A Silicon Graphics Indigo 2 XZ with an R4000 processor and 128 MB RAM took 85 hours of CPU time to obtain a complete solution. This is a valuable tool but, unlike conventional heating, requires advanced computers to solve the model. Nott et al. (1999) used MRI to map quantitatively in three dimensions the complex temperature distributions induced by microwave heating of food materials: results are compared with infrared thermal images. MRI is limited to
130
Thermal technologies in food processing
non-metallic systems such as paper, plastic, ceramic and glass, but this is an approach which allows non-invasive sampling. 7.4.2 Ohmic heating Although ohmic heating is thought of as a novel process, its use in food processing goes back a century (de Alwis and Fryer, 1990). Over the last twenty years, interest in the process has followed the development of a commercial unit by APV Baker (Parrott, 1992). In this process, food is pumped past a series of electrodes connected to three-phase supply at 50–60 Hz. Heating rates of the order of 1 C/s are possible, in field strengths on the order of 10 V/cm. The process has found use in sterilisation of high-solids fraction (30–40% solids) mixtures as well as the production of high-value materials such as pasteurised fruit pieces for yoghurts. Numerical models of electrical heating have concentrated on a set of problems on different length and timescales: • Heating rates within a solid-liquid mixture (such as de Alwis et al., 1989; Palaniappan and Sastry, 1991). • The types of temperature pattern found inside solid-liquid mixtures (such as Fryer et al., 1993; Kemp et al., 1999). • Predicting the temperatures of mixtures undergoing ohmic heating (such as Zhang and Fryer, 1993, 1994; Benabderrahmane and Pain, 2000).
Laplace’s equation for the electric field can only be solved analytically in very simplistic cases, such as for an isolated sphere or infinite cylinder of constant physical properties in a uniform field. de Alwis et al. (1989) show that for such cases the ratio of the heating rate in a solid and undisturbed liquid is given by: R Q
sphere
9s L
s 2L 2
; R Q
cylinder
4s L
s 2L 2
7:25
The heating rate in the cylinder can thus never exceed that of the surrounding fluid, whilst that in the sphere can exceed that of the liquid for 1 < S =L < 4. That ohmic heating can result in solids that overheat the liquid was shown by de Alwis et al. (1989). In general, for non-uniform shapes and non-uniform physical properties, computer models are needed to solve for the heating patterns. The first interaction between the particle size and shape and the electric field was shown by de Alwis and Fryer (1990), using a code written especially for the problem. It is now possible to use commercial codes, such as ANSYS (Zhang and Fryer, 1994) or FIDAP, with which Kemp et al. (1999) validate a computational model for the heating pattern around an insulating particle. Here, both simulation and experiment show that it is possible to have over- and underheating in the same particle; underheating results from the region of low electric field behind the insulating particle, and overheating is due to the high conductivity of the second particle. Local heating and cooling effects have been shown by thermocouple and in elegant MRI experiments by Ruan et al. (1999).
Modelling particular thermal technologies
131
Various models have been proposed to identify the coldest spot in an ohmic formulation: finding this is key to deciding what a process should be. In conventional processing, of course, the coldest spot will occur in the solid; with the correct choice of solid and liquid electrical conductivity, the particles will overheat the liquid, and the coldest spot will be found in the fluid. The amount of fluid mixing is then critical; high viscosity fluids have a higher range of temperature differences between them than less viscous ones (see Fryer et al., 1993). Models for the whole process have been developed which involve a series of simplifications. Zhang and Fryer (1993, 1994) used a finite-element model to predict the behaviour of a sphere in a well-mixed fluid, and then used that model as the basis for a Fortran model of a flowing mixture. Heat generation techniques have found some applications in industry – obviously microwave ovens have achieved significant penetration into the kitchen as a result of the rapidity of heating possible. Developments in modelling will be used (i) to make industrial application easier, and (ii) to make products which are of higher quality when heated in the domestic microwave oven.
7.5
Developments in the field
The modelling of thermal processing of foods is an active research subject around the world. The subject is developing in a number of areas: 7.5.1 Ease of solving models Advances in computing continue to make it more straightforward to run the types of programs that are needed to solve these problems. As shown above, FE software can now run efficiently on a PC, whereas only a few years ago it required workstation or mainframe capabilities. It is likely that problems that currently are at the limits of computing power, such as the efficient modelling of microwave heating, will be simple to solve in a few years time. This means that models could be used as the basis for real-time control systems: at the moment most run too slowly. 7.5.2 Realistic physical properties Many of the papers described above have treated systems with simplistic physical properties, such as constant thermal diffusivity. In practice, many physical properties vary with temperature; the strong variation of viscosity and electrical properties has been shown in some of the papers described above to lead to very strongly coupled problems, where the thermal and other fields have to be solved together. To do this accurately requires accurate data: the more accurate the data, the better the fit of the model to reality. In some cases, it is the
132
Thermal technologies in food processing
lack of data rather than the lack of appropriate theory that is limiting the accuracy of the models. 7.5.3 New types of model The models described above have used finite element, volume or difference approaches to model heat transfer. Other types of approach are being used in some areas, such as discrete element and lattice Boltzmann modelling of singleand two-phase flows in complex geometries, and cellular automata models for the interaction of microbes and foods. These different types of model have some advantages, particularly where FE meshes are complicated and change with time, as in particle flows. At the moment, they are difficult to use, and commercial codes are not well developed; as these types of model become simpler to use they will be applied to food problems. 7.5.4 Kinetic models for food processes Given accurate physical property data and computer codes, thermal models are capable of predicting the temperatures throughout food solids and liquids. However, these predictions cannot be used fully if information on the rates of the processes affected by temperature are not available. The microbiological models used in many thermal models are probably simplistic; if equation (7.1) does not describe the effect of temperature on a microbial population then the results of the model will not apply. More research is needed to determine models for microbial growth and inactivation, and on the rates of development of texture and flavour in foods. This type of data, combined with effective thermal models, will lead to the production of models which can predict food quality and safety, and act as the basis for optimising production. Some of our recent work (Bakalis, Cox and Fryer, unpublished results) has studied ways of validating flow and heating models. The School of Physics at the University of Birmingham, UK, has developed a unique way of following flows in opaque fluids using tracers, the so-called Positron Emitting Particle Tracking (PEPT). In this technique a radioactive tracer particle is monitored during its passage through a system. The tracer emits positrons, which then generate a pair of back-to-back gamma rays on collision with an electron (an extremely rapid and geometrically close annihilation event). By detecting the trajectories of the gamma rays (with a 180º of separation) the position and velocity of the tracer can be followed by triangulating their position at many times per second. The technique can also be used in pilot-scale equipment as the gamma rays can penetrate reasonable thicknesses of metal. As an example of how it can be used in foods, PEPT was used to examine particle paths in a canning process. A typical metal can was filled to 90% of its volume with vegetable soup obtained from a local market, and a 600 micron wide tracer particle was inserted in a piece of potato from the soup and was followed as the can was rotated axially at a speed of 25 rpm. Figure 7.2
Modelling particular thermal technologies
Fig. 7.2
133
PEPT traces of a metal can filled with soup.
represents the particle path over a period of three minutes. It can be seen that the particle follows a D shape trajectory with the headspace affecting the flow. This pattern is very different when compared to a fully filled can, where the particles are moving in circles. Overall PEPT appears to be a promising technique that can give invaluable insight on thermal processing and provide means to validate numerical simulations.
7.6
References
ABDUL GHANI A G, FARID M M, CHEN X D, RICHARDS P
(1999a). Numerical simulation of natural convection heating of canned food by computational dynamics. J. Food. Eng., 41, 55–64. ABDUL GHANI A G, FARID M M, CHEN X D, RICHARDS P (1999b). An investigation of deactivation of bacteria in a canned liquid food during sterilization using computational fluid dynamics. J. Food. Eng., 42, 207–14. AKTERIAN S G (1999). Online strategy for compensating for arbitrary deviations in heating medium temperature during batch thermal sterilization processes. J. Food Eng., 39, 1–7. AKTERIAN S G, SMOUT C, HENDRICKX M E, TOBBACK P P (1998). Application of sensitivity functions for analysing the impact of temperature nonuniformity in batch sterilisers. J. Food Eng., 37, 1–10. BALL C O (1923). Thermal process time for canned foods. Bulletin of the National Resources Council, 7, 1. BARIGOU M, MANKAD S, FRYER P J (1998). Heat transfer in two-phase solid-liquid food flows: A review. Trans. IChemE., 76, C, 3–29. BELLARA S R, FRYER P J, MCFARLANE C M, THOMAS C R, HOCKING P M, MACKEY B M (1999). Visualization and modelling of the thermal inactivation of bacteria in a model food. Appl. Env. Microbiol., 65, 3095–9.
134
Thermal technologies in food processing
BENABDERRAHMANE Y, PAIN J-P
(2000). Thermal phase behaviour of a solid/ liquid mixture in an ohmic heating sterilizer – slip phase model – sensitivity to process parameters. Chem. Eng. Sci., 55, 1371–84. BIRD R B, STEWART W E, LIGHTFOOT, E W (1964). Transport Phenomena. John Wiley & Sons, Inc. BOWS J R (2000) A classification system for microwave heating of food. Int. J. Food Sci Technol., 35, 417–30. BOWS J R, RICHARDSON P S (1990). Effect of component configuration and packaging materials on microwave reheating a frozen 3-component meal. Int. J. Food Sci Technol. 25, 538–50. BOWS J R, PATRICK M L, JANES R, DIBBEN D C (1999). Microwave phase control heating. Int. J. Food Sci. Technol., 34, 295–304. BUFFLER C R, STANFORD M A (1995). Effects of dielectric and thermal properties on the microwave heating of foods. Microwave World, 16, 5–10. CAMPBELL G M, WEBB C, PANDIELLA S S, NIRANJAN K (1999). Bubbles in Food. Eagan Press. CARSLAW H S, JAEGER J C Conduction of Heat in Solids. Oxford University Press. DE ALWIS A A P, FRYER P J (1990). A finite element analysis of heat generation and transfer during ohmic heating of food. Chem. Eng. Sci.. 45, 1547–60. DE ALWIS A A P, HALDEN K, FRYER P J (1989). Shape and conductivity effects in the ohmic heating of foods. Chem. Eng. Res. Des., 67, 159–68. DIBBEN D C, METAXAS A C (1994). Finite element time domain analysis of multimode applicators using edge elements. J. Microwave Power and Electromagnetic Energy, 29, 242–51. FRYER P J, DAVIES L J (2001). Modelling electrical resistance (‘ohmic’) heating in foods. To be published in Irudayraji, J. (ed.), Food Process Operations Modelling: design and analysis. Marcel Dekker. FRYER P J, DE ALWIS A A P, KOURY E, STAPLEY A G F, ZHANG L (1993). Ohmic heating of solid-liquid foods; heat generation and convection effects, J. Food. Eng., 18, 101–25. HAYAKAWA K (1969). New parameters for calculating mass average sterilizing value to estimate nutrients in thermally active foods. Can. Inst Food Technol. J., 2, 165–72. HENDRICKX M, SILVA C, OLIVERA F, TOBBACK (1993). Generalised (semi-) empirical formulae for optimal sterilisation temperatures of conductionheated foods with infinite surface heat transfer coefficients. J. Food Eng., 19, 141–58. HULBERT G J, LITCHFIELD J B, SCHMIDT S J (1997). Determination of convective heat transfer coefficients using 2D MRI temperature mapping and finite element modelling. J. Food. Eng., 34, 193–201. INCORPERA F P (1981). Fundamentals of Heat Transfer. Wiley. JUNG A, FRYER P J (1999). Optimising the quality of safe food: Computational modelling of a continuous sterilisation process. Chem. Eng. Sci., 54, 717– 30. KANTT C A, SCHMIDT S J, SIZER C E, PALANIAPPAN S, LITCHFIELD J B (1998).
Modelling particular thermal technologies
135
Temperature mapping of particles during aseptic processing with magnetic resonance imaging. J. Food. Sci., 63, 305–11. KEMP M, DAVIES L, FRYER P J (1999). The geometry of shadows: effects of inhomogeneities in electrical field processing. J. Food Eng., 40, 245–58. KIM K H, TEIXEIRA A A (1997). Predicting internal temperature responses to conduction-heating of odd shaped solids. J. Food Process Eng., 20, 51–63. KUMAR A, BHATTACHARYA M (1991a) Numerical analysis of aseptic processing of a non-Newtonian liquid food in a tubular heat exchanger. Chem. Eng. Comm., 103, 27–51. KUMAR A, BHATTACHARYA M (1991b). Transient temperature and velocity profiles in a canned non-Newtonian liquid food during sterilisation in a still cook retort. Int. J. Heat Mass Transfer, 34, 1083–96. KUMAR A, BHATTACHARYA M, BLAYLOCK J (1990) Numerical simulation of natural convection heating of canned thick viscous food products. J. Food Sci., 55, 1403–11, 1420. LAREO C, BRANCH C A, FRYER P J (1997a). Particle velocity profiles for solidliquid food flows in vertical pipes. 1. Single particles. Powder Tech, 93, 23–34. LAREO C, NEDDERMAN R M, FRYER P J (1997b). Particle velocity profiles for solidliquid food flows in vertical pipes. 2. Multiple particles. Powder Tech, 93, 35–45. LAREO C, FRYER P J, BARIGOU M (1997c). The fluid mechanics of two-phase solid-liquid food flows: A review. Trans. IChemE., 75, C, 73–105. LARKIN J W (1989). Use of a modified Ball formula method to evaluate aseptic processing of foods containing particulates. Food Technol., 43, 124–31. LIAO H J, RAO M A, DATTA A K (2000). Role of thermo-rheological behaviour in simulation of continuous sterilization of a starch dispersion. Trans. IChemE., 78, C, 48–56. LOPEZ A (1987). Heat process determinations. In: A Complete Course in Canning and Related Processes, Vol 2, 21–30. The Canning Trade Inc., Baltimore. MACKEY B M, DERRICK C M (1987). Changes in the heat resistance of Salmonella typhimurium during heating at rising temperatures. Lett. Appl. Microbiol., 4, 13–16. MANKAD S, FRYER P J (1997). A heterogeneous flow model for the effect of slip and flow velocities on food steriliser design. Chem. Eng. Sci., 52, 1835–43. MANKAD S, BRANCH C A, FRYER P J (1995). The effect of particle slip on the sterilization of solid-liquid food mixtures. Chem. Eng. Sci., 50, 1323–36. MANSON J E (1974). Thermal process simulation for aseptic processing of foods containing discrete particulate matter. J. Food Sci., 39, 1084–9. METAXAS A C (1996). Foundations of Electroheat: A Unified Approach. John Wiley & Sons, Inc. NOTT K P, BOWS J R, PATRICK M L (1999). Three-dimensional MRI mapping of microwave induced heating patterns. Int. J. Food Sci. Technol., 34, 305–17. OHLSSON T, RISMAN P O (1978). Temperature distribution of microwave heating, spheres and cylinders. J. Microwave Power, 13, 303–10.
136
Thermal technologies in food processing
(1993). Heat Conduction, 2nd edition. John Wiley & Sons, Inc. (1991). Modelling of electrical conductivity of liquid-particle mixtures. Trans. IChemE., 69, C, 167–74. PAN Z, SINGH R P, RUMSEY T R (2000). Predictive modelling of contact-heating process for cooking a hamburger patty. J. Food Eng., 46, 9–19. PARROTT D L (1992). Use of ohmic heating for aseptic processing of particulates. Food Technol., 46, 68–72. QUINTAVALLA S, CAMPININI M, (1991). Effect of rising temperature on the heat resistance of Listeria monocytogenes in meat emulsion. Lett. Appl. Microbiol., 12, 184–7. RAMASWAMY H S, ABDELRAHIM K A, SIMPSON B K, SMITH J P (1995). Residence time distribution (RTD) in aseptic processing of particulate foods: a review. Food Res. Int., 28, 291–310. RUAN R, CHEN P, CHANG K, KIM H-J, TAUB I A (1999). Rapid food particle measurement temperature mapping during ohmic heating using FI-ASH MRI. J. Food Sci., 64, 1024–6. RYCKAERT V G, CLAES J E, VAN IMPE J F (1999). Model-based temperature control in ovens. J. Food. Eng., 39, 47–58. SANDEEP K P, ZURITZ C A, PURI V M (2000). Modelling non-Newtonian two phase flow in conventional and helical tubes. Int. J. Food Sci. and Tech., 35, 511–22. SASTRY S (1986). Mathematical evaluation of process schedules for aseptic processing of low-acid foods containing discrete particulates. J. Food Sci., 51, 1323–8. SASTRY S, HESKITT B, BLAISDELL J (1989). Experimental and modelling studies on convective heat transfer at the particle-liquid interface in aseptic processing systems. Food Tech., 43, 132–6. SIMPSON S G, WILLIAMS M C (1974). An analysis of high temperature/short time sterilization during laminar flow. J. Food Sci., 39, 1047–5. SUN X Z, LITCHFIELD J B, SCHMIDT S J (1993). Temperature mapping in a model food gel using magnetic resonance imaging. J. Food Sci., 68, 168–72, 181. SUN X Z, SCHMIDT S J, LITCHFIELD J B (1994). Temperature mapping in a potato using half Fourier transform MRI of diffusion. J. Food Proc. Eng., 17, 423–37. TEIXEIRA A A, ZINSMEISTER J W, ZINSMEISTER G E (1969). Computer simulation of variable retort control and container geometry as a possible means of improving thiamine retention in thermally processed foods. J. Food Sci., 40, 656–9. TEWKESBURY H, STAPLEY A G F, FRYER P J (2000). Modelling temperature distributions in cooling chocolate moulds. Chem. Eng. Sci., 55, 3123–32. TIMBERS G E, HAYAKAWA K (1967). Mass average sterilizing value for thermal process. Part 1 Comparison of existing procedures. Food Technol., 21, 1069. TOLEDO R T (1991). Thermal process calculations. In: Fundamentals of Food Process Engineering, 2nd edition, 315–97. Van Nostrand Reinhold. ¨ ZISIK M N O
PALANIAPPAN S, SASTRY S K
Modelling particular thermal technologies TUCKER G S, HOLDSWORTH S D
137
(1991). Mathematical modelling of sterilisation and cooking processes for heat preserved foods. Trans. IChemE., 69, C, 5– 12. VARGA S, OLIVERIA J C, SMOUT C, CHANTAL C, HENDRICKX M E (2000). Modelling temperature variability in batch retorts and its impact on lethality distribution. J. Food. Eng., 44, 163–74. VERBOVEN P, NICOLAI¨ B M, SCHEERLINCK N, DE BAERDEMAEKER J (1997). The local surface heat transfer coefficient in thermal food process calculations: A CFD approach. J. Food Eng., 33, 15–35. VERBOVEN P, SCHEERLINCK N, DE BAERDEMAEKER J, NICOLAI¨ B M (2000). Computational fluid dynamics modelling and validation of the temperature distribution in a forced convection oven. J. Food Eng., 43, 61–73. ZHANG LI, FRYER P J (1993). Models for the electrical heating of solid-liquid food mixtures. Chem. Eng. Sci., 48, 633–43. ZHANG LI, FRYER P J (1994). Food processing by electrical heating; the sensitivity of product sterility and quality to process parameters. AIChEJ, 40, 888–98.
8 Thermal processing and food quality: analysis and control A. Arnoldi, University of Milan
8.1
Introduction: the importance of thermal processing
Most foods are submitted to thermal processes such as cooking, baking, roasting, extrusion cooking, pasteurisation or sterilisation. The first processes serving mostly to obtain particular sensory or texture features, the last two to assure microbiological safety and to eliminate some enzymatic activities that reduce food preservation. The reactions that occur are of great importance for the production of aroma, taste and colour. From this point of view foods can be divided in two classes. In some cases, these modifications must be reduced to a minimum because a natural and fresh appearance is required, as, for example, in drinkable milk or fruit juices. In other cases they are desirable, because they produce the specific sensory and texture features of foodstuffs, such as bread, cereals, chocolate, coffee, nuts, malt and cooked meat. However, when these reactions occur during food preservation, they always have a negative impact on quality. One of the most important processes involved is the reaction of amino acids, peptides and proteins with reducing sugars and vitamin C, the process generally known as Maillard Reaction (MR) or non-enzymatic browning. This is a cascade of complex competitive reactions and even if most of them are fundamental organic reactions, such as eliminations, aldol condensations, retroaldol fragmentations, oxidations and reductions, the fact that they occur simultaneously and give rise to many reactive intermediates makes difficult their interpretation and control (Yaylayan 1997). For many years these transformations have been studied for their technological impact, but more recently it has been shown that they may be accompanied by a reduction in the nutritive value of foods and by the formation of toxic compounds. The loss in nutritional quality
Thermal processing and food quality: analysis and control
139
and potentially in safety is attributed to the destruction of essential amino acids, interaction with metal ions, decrease of digestibility, inhibition of proteolytic and glycolytic enzymes, and formation of anti-nutritional and toxic compounds.
8.2
The importance of the Maillard reaction
The first investigations on the MR were performed around 90 years ago by Maillard (1912). Subsequently, Amadori reported on the formation of a rearranged stable product from glucose and amino acids which was named after him and Heyns reported a similar compound from fructose. An overall picture of the reactions involved was proposed by Hodge (1953). One of the most detailed descriptions of the pathways that lead to the main Maillard Reaction Products (MRPs) can be found in a very complete review by Ledl and Schleicher (1990). Only some relevant points will be discussed here. Reducing sugars are necessary for this reaction; generally they are a monosaccharide, glucose or fructose, or a disaccharide, maltose or lactose, and in some cases a pentose. Non-reducing disaccharides, such as sucrose, or bound sugars as in glycoproteins, glycolipids, and flavonoids, react only after hydrolysis, a process often facilitated by fermentation. The counterparts are free amino acids or proteins. Fermentation is useful also to increase the concentration of free amino acids, whereas in some cases (e.g. cheese) biogenic amines can react as amino compounds. Small amounts of ammonia can be produced from amino acids during the Maillard reaction and large amounts are added for the preparation of a particular kind of caramel colouring. The mechanism of this reaction has been studied very seldom in real foods, because it is too complex and the separation of non-volatile products is very difficult. On the contrary, most authors have used simple model systems in order to improve control of all the parameters and very often these results have been extrapolated to foods quite efficiently. Consider now an aldose. The initial step is the condensation of the aldehydic group of the sugar with an amino group to give a relatively unstable glycosylamine 1 which undergoes a reversible rearrangement to give an aminoketose 2 (Amadori compounds) (Fig. 8.1). These intermediates have been fully characterised in model systems and detected in many foods. The equivalent rearrangement of the fructose + amino acid adduct produces an amino aldose which is called Heins product. The reader will find a detailed description of the properties and reactivity of the Amadori Heins compounds in Yaylayan and Huyggues-Despointes (1994). At low water content and pH 3–6, they can be considered the main precursors of reactive intermediates in model systems. Deoxydiketoses and deoxyaldoketoses are degradation products of aminoketoses in the pH range 4–7. Ring opening followed by 1, 2 or 1, 3-enolisation are crucial steps in this transformation and are followed by dehydrations and fragmentations with the formation of many very reactive dicarbonyl fragments. One of the most important end products of the reaction of 3-deoxyhexulose is 5-hydroxymethyl-furancarboxaldehyde 3 (HMF).
140
Thermal technologies in food processing
Fig. 8.1
Amadori rearrangement.
In the presence of large amounts of primary amines, this compound is suppressed and replaced by pyrrole aldehydes of structure 4 as well as betaines 5 which have been demonstrated to derive from a common intermediate and not from HMF. Compounds 4 and 5 can be obtained also from maltose and lactose. A typical degradation product of 1-deoxydiketoses is furanone 6 which easily splits off formaldehyde to give furanone 7 which is the base for the formation of some relevant low molecular weight coloured compounds. Furanone 7 is produced also from the equivalent reaction of pentoses (Fig. 8.2). Small C2, C3 and C4 fragments are produced from sugars by retroaldol cleavage (Fig. 8.3). Many of these compounds are very reactive and condense
Thermal processing and food quality: analysis and control
Fig. 8.2
141
Relevant end products from 3-deoxydiketoses and 1-deoxydiketoses.
readily with other retroaldol derivatives or sugar derivatives in which the carbon skeleton is still intact. The reaction with the amino group of amino acids results in the incorporation of nitrogen, a critical step for the formation of melanoidins and important flavour compounds such as pyrazines. One of the first observations by Maillard was related to the production of CO2 in the reaction mixture. This process is named Strecker degradation (Fig. 8.4). The mechanism involves the reaction of amino acids with -dicarbonyl compounds to produce an azovinylogous beta-ketoacid that is decarboxylated. Thus the amino acid is converted in an aldehyde containing one less carbon than
Fig. 8.3 Some of the compounds produced by retroaldol cleavage.
142
Thermal technologies in food processing
Fig. 8.4 Strecker degradation.
the original amino acid. Most of these aldehydes are very reactive intermediates or have very peculiar sensory properties. In addition to the normal compounds, with cysteine the reaction yields also hydrogen sulfide, with methionine 2methylthiopropanal and methanethiol. The Strecker degradation is therefore responsible for the incorporation of nitrogen and sulfur in many volatile and non-volatile end products. In recent years, very detailed experiments on the reaction between 13C labelled monosaccharides or disaccharides (Tressl et al. 1995; Tressl and Rewicki 1999) and amino acids have clarified the different Maillard pathways of these two important classes of food components, and a revised scheme for the Maillard reaction was proposed. In proteins the most relevant effect of the Maillard reaction is the non-enzymatic glycosylation which involves mostly lysine. The first glycation products are then converted to the Amadori product fructosyllysine that can form cross-links with adjacent proteins or with other amino groups. The resulting polymeric aggregates are called advanced glycation end products (AGEs).
8.3
Thermal processing and food safety
The negative side of thermal processing is the possible formation of potentially carcinogenic MRPs, among which heterocyclic amines (HAs) have attracted a growing interest during the past two decades. HAs include about twenty different derivatives that are primarily found at ppb level in cooked muscle foods (Sugimura et al. 1977). Many HAs have been shown to be carcinogenic in mice, rats, and non-human primates studies. However, epidemiologic studies showed conflicting data: some have shown an association between cooked meat and fish intake and cancer development and others no significant relationship
Thermal processing and food quality: analysis and control
143
(Sugimura et al. 1993; Steineck et al. 1993). The International Agency for Research on Cancer (IARC) regards some HAs as possibly or probably carcinogenic to humans, and recommends minimising our exposure to them (IARC 1993). Minimising their formation requires knowledge of precursors, cooking conditions, reaction mechanisms and kinetics, so that the food industry can choose optimal conditions for designing food processes and food process equipments. HAs are found mostly in the crust of grilled and fried meat and fish and in the pan residue, but to a much lesser extent in the interior of meat (Skog et al. 1995). Cooking time and temperature are critical parameters in determining their amount (Knize et al. 1994). The low concentration of HAs and the complex sample matrix of cooked foods make the analysis of these compounds very difficult. Most data reported in literature refer to polar HAs in cooked muscle foods that are found at low ng/g level (for reviews see Skog 1993; Layton et al. 1995). An improved method for the determination of non-polar HAs has been published recently (Skog et al. 1998). A kinetics model for the formation of polar HAs in a meat model system has been proposed (Arvidson et al. 1998). A detailed overview of the risk assessment can be found in a review by Friedman (1996). Data suggest that HAs are the only known animal colon carcinogens that humans (except vegetarians) consume every day and, although very difficult, it would be desirable to control their level in food (Fig. 8.5).
8.4
Thermal processing and nutritional quality
The glycation of protein by sugars affects negatively the nutritional value of proteins (Friedman 1996). In fact, beside the destruction of particularly reactive amino acids, such as lysine, all essential ones must be liberated from food proteins by digestion and factors that decrease digestibility as cross-linking reduces their bioavailability. Loss of nutritional quality has been demonstrated in heat treated casein, casein-glucose, and casein-starch and was explained with decreased nitrogen digestibility, possibly also through impaired intestinal absorption (Gumbmann et al. 1993). On the basis of labelled experiments, Mori and Nakatsuji (1977) affirmed that the reduction of the nutritive value depends on the reduction in intestinal absorption of the Maillard-induced lysine derivatives. Besides the change in physiological properties of non-enzymatic browning products, there are other factors that could affect the nutritional quality, such as the formation of toxic compounds during the heat treatment of food: heterocyclic amines (HAs), melanoidins and lysinoalanine (LAL). LAL is not exactly an MRP, but a cross-linking product which occurs in proteins treated under alkaline conditions owing to the fact that the "-amino group of lysine in the protein chain reacts with dehydroalanine, derived from -elimination of good leaving groups from cystine and O-substituted serine (Friedman 1999). The cross-linking reaction occurs mostly intramolecularly and results in the
144
Thermal technologies in food processing
Relevant heterocyclic amines. Fig. 8.5
Thermal processing and food quality: analysis and control
145
reduction of digestibility of the proteins involved. The loss of nutritional quality is particularly relevant in infant nutrition where milk is the only source of proteins and amino acids. Erbersdobler (1989) observed other adverse effects in experiments with animals fed with thermally treated proteins, in particular, damage of the renal function, but these negative influences do not seem to be significant in practical conditions. Another important nutritional consequence of the MR in food is the formation of antioxidative materials. Chiu et al. (1991) carried out studies on the formation of antioxidants in the reaction between tryptophan and glucose or fructose. These MRPs have a positive effect in preventing lipid oxidation in foods and the scavenging of active oxygen species has been related to these antioxidative effects. Chelation of metals is another important contribution. There are also some scattered data on the formation of antibacterial materials (Einarsson et al. 1983, 1988). Products deriving from arginine/glucose and arginine/xylose seemed particularly efficient. However, the reported data refer only to model systems and it is not clear if these compounds really contribute to the preservation of foods.
8.5
Thermal processing, food flavour and colour
The MR is responsible for most colour formation during thermal processing of food, e.g. bread baking, roasting of coffee and nuts, roasting of meat, kiln-drying of malt. Although nobody can deny its importance in determining the quality of foods, the progress of knowledge in this field has been slowed by many methodological difficulties, because most of the colour is due to high molecular weight polymers, which are named melanoidins. Some attempts have been made to isolate melanoidins from foods, e.g. soy sauce (Hashiba 1973; Lee et al. 1987), dark beer (Kuntcheva and Obretenov 1996), malt or roasted barley (Milic et al. 1975; Obretenov et al. 1991), coffee (Maier and Buttle 1973; Steinhart and Packert 1993), but their very complex, probably non-repetitive, structure has limited their structural characterisation. Until now the only coloured compounds which have been fully characterised are some low molecular weight ones (Ames and Nursten 1989; Arnoldi et al. 1997; Ravagli et al. 1999; Hofmann 1998b, 1998d; Tressl et al. 1998a, 1998b; Wondrack et al. 1997). Figure 8.6 shows the structure of some coloured compounds isolated in recent years. Generally foods contain proteins and more rarely free amino acids. In a recent work (Hofmann 1998c), it was demonstrated that different compounds are produced reacting glucose with glycine and alanine or -casein. With free amino acids the majority of coloured compounds were shown to have molecular weight below 1000 amu, whereas the reaction between glucose and casein leads to a drastic increase of the molecular weight. Hofmann (1998a) has proposed colour dilution analysis, a method that permits selection of the compounds of a mixture that contribute mostly to colour. Another very important consequence of the MR is the formation of flavour, a complex sensory feature related both to taste and
146
Thermal technologies in food processing
Fig. 8.6 Examples of low molecular weight coloured compounds recently reported in literature: 19 (Ravagli et al. 1999); 20 (Hofmann 1998d): 21 (Hofmann 1998a); 22 (Hofmann 1998b); 23 (Tressl et al. 1998a); 24 (Tressl et al. 1998b).
smell sensations. From the viewpoint of taste, there is a generalised shift from sweet to bitter sensations in part from the destruction of sugars and in part the formation of bitter compounds. The latter aspect has not been studied in detail yet, partly for methodological difficulties. Much more can be said about aroma. Odorous compounds must be relatively small in size and lack polarity, therefore they can be studied by GC-MS much more easily than any other MRP. Hundreds of compounds have been identified. Nursten (1980–81) has suggested dividing them into three groups:
Thermal processing and food quality: analysis and control
Fig. 8.7
1. 2. 3.
147
Relevant volatiles from sugar dehydration/fragmentation (Nursten 1980–81).
simple sugar dehydration/fragmentation products, such as furans, pyrones, cyclopentenes, carbonyls, acids (Fig. 8.7); simple amino acid degradation products, Strecker aldehydes and sulfur compounds; volatiles produced by other interactions: pyrroles, pyridines, imidazoles, pyrazines, oxazoles, thiazoles, aldol condensation products (Fig. 8.8).
Pyrazines give a major contribution to roasted and cooked aroma (Maga 1982; Arnoldi 1992); their organoleptic properties depend strictly on structure and quantitative structure-activity relationships (QSARs) have been proposed. Flavours developed by using thermal treatment of selected precursors have become standard commercial flavours sold all over the world, and they are generally known in Europe as ‘process flavours’ (Manley 1994). They are generally sold as the final flavour, however microwave flavours may be sold as precursor mixtures prior to the development of flavour. Of course not only the absolute concentration, but also the typical note and the odour thresholds of flavour are important in determining the sensory properties of a food. The reader will find many sensory data of MRPs in a very complete compilation by Fors (1983). Threshold values can sometimes be relatively high (ppm range) or extraordinarily low (ppb level or less), for example 3-sec-butyl-2-methoxypyrazine can be perceived in water at 0.001 ppb, and 1,3,5-trithiane in water at 0.04 ppb. Although MRPs are relevant impact compounds in foods, the presence of a odour-active volatile is not a direct measure of its importance to aroma. Specific techniques have been developed to solve this problem: the most common are the gas-chromatography-olfactometry (GCO) technique (Cunningham et al. 1986) and the aroma extraction dilution analysis (AEDA) (Schieberle et al. 1990). The
148
Thermal technologies in food processing
Fig. 8.8 Relevant volatiles from other interactions (Nursten 1980–81).
GCO bioassay characterises potent odour-active volatiles by sniffing the gas chromatograph effluent. Using a sniffing port, a trained operator notes the presence of an odour at a particular retention index and records the sensory characteristics. In subsequent runs the sample is diluted by factors of three and the analysis is repeated until the odour is no longer perceivable. The combination of these runs produces a corrected chromatogram with odour potency defined as the area of the chromatographic peaks. Using this technique, for example, it was possible to determine the major odour potent compounds in glucose-proline model systems (Roberts and Acree 1994) which are burnt caramel, cotton candy and popcorn. They are diacetyl, 2-acetyl-1-pyrroline, 2acetyl-1,4,5,6-tetrahydropyridine, 2-acetyl-3,4,5,6-tetrahydropyridine and furaneol, and all of them are very small and apparently irrelevant peaks in the FID chromatogram.
8.6
Maillard reaction and lipid oxidation
Another important reaction which occurs in processed and cooked foods and influences the aroma is lipid degradation. The intermediates of these reactions are involved in the formation of food flavour. Experiments carried out with
Thermal processing and food quality: analysis and control
149
model systems added with fat have demonstrated that several volatile compounds are product by the interaction between MRPs and lipid oxidation products (Whitfield 1992). Most of them are heterocyclic compounds containing nitrogen or sulfur such as alkylthiazoles, long-chain alkylpyrazines (Chiu et al. 1990; Huang et al. 1987), pyridines and thiophenes. Phospholipids were added to model systems to mimic the formation of meat aroma (Farmer and Mottram 1990; Farmer et al. 1989), and recently it was demonstrated that some edible oils can affect the pyrazine profile (Negroni et al. 2000).
8.7
Controlling factors in the Maillard reaction
Understanding the factors that influence the MR is critical for achieving its control. The formation of colour and flavour and the loss in nutritional value are the aspects that have been studied in detail. The most important parameters that affect the MR are: temperature, pH, water activity, and the structure of amino acids and sugars involved. However, most of the information available has been collected in model systems and not in real foods. The effect of temperature is particularly strong and involves every aspect of the MR. For example, several experiments have shown that an increase in temperature and/or time of heating leads to an increase of colour development and aroma profile. Not only the amount of the MRPs is increased, but also their nature is modified. For example it has been demonstrated that the carbon-tonitrogen ratio, the degree of unsaturation and the chemical aromaticity of the melanoidins formed in model systems increase with temperature and time. The formation of undesired compounds as HAs (see Section 8.3) is particularly sensitive to high temperature and a careful choice of the cooking conditions allows them to be reduced to a minimum. Considering flavour, at intermediate temperatures caramel or cooked flavours are produced, while at higher temperatures the aroma profile is toasted or roasted. Nevertheless, the MR proceeds slowly also at room temperature and browning and off-flavour formation are responsible for the deterioration of food during storage. The MR proceeds faster at low moisture level (Nursten 1980–81), and it is generally accepted that moisture values corresponding to a water activity around 0.65–0.75 are the most favourable. The differences in colour and flavour of the outer part (where fast dehydration occurs) and wet inner part of baked or roasted foods are a very clear proof of these effects. However, not all the MRPs are sensitive in the same way and studies on flavour have demonstrated that different classes of volatile components are more or less sensitive depending on whether or not water is required for their formation. The complex network of reactions that produce browning is far from being disclosed, but certainly many condensations and dehydrations are involved and water activity is a critical parameter of the reaction kinetics.
150
Thermal technologies in food processing
pH is another important parameter. Browning is much greater at pH above 7 and the rate of colour formation can be reduced by decreasing pH. Flavour is also influenced, for example pyrazines reductones and fission products are preferred at high pH values, while furans, especially furfurals, prevail at lower pH. Nevertheless, mixtures of compounds are formed at each pH. Careful control of the pH is necessary in model systems because of acid formation, whereas food pH is more stable because they have a buffered environment. Another important parameter is the structure of the reactants: pentoses react more quickly than hexoses, that react faster than disaccharides, and aldoses react faster than ketoses. Non-reducing sugars such as sucrose, dextrines and bound sugars are involved only after hydrolysis. Aspartic and glutamic acids are relatively unreactive, whereas lysine, the most reactive free amino acid, is reactive also in proteins owing to the "-amino group. The structure of the amino acids determines the formation of particular MRPs such as sulfur compounds from cysteine and methionine, that, however, react only after hydrolysis or fermentation.
8.8
Methods of measurement
Browning is certainly a macroscopic effect of the MR and sometimes it is directly measured to determine the progress of the reaction. However, recently much more sophisticated parameters have been suggested and used. Non-enzymatic glycosylation of proteins is one of the most important nutritional consequences and many methods have been developed for its estimation. The first glycation product or Schiff base rearranges to a more stable ketoamine or Amadori product. The Amadori product can form cross-links between adjacent proteins or with other amino groups. The resulting polymeric aggregates are called advanced glycation end products (AGEs). During acid hydrolysis of proteins, the glucose derived Amadori product fructosyllysine reacts to give furosine in about 30% yields and a small amount of pyridosine, whereas 50% reverts to lysine. Furosine was first detected by Erbersdobler and Zucker (1966) in foods and it can be easily analysed by HPLC with 280 nm detection. The corresponding reaction in milk and milk products produces lactulosyllysine that can be estimated by the furosine method, too. The furosine test has been used by several authors to study the progress of the Maillard reaction in different foods (Chiang 1983; Hartkopf and Erbersdobler 1993, 1994; Henle et al. 1995; Resmini et al. 1990). Carboxymethyllysine is another useful parameter. It was detected for the first time in milk by Bu¨ser and Erbersdobler (1986) and an oxidative mechanism was proposed for its formation (Ahmed et al. 1986). Hewedy et al. (1991) have compared several damage indicators for the classification of UHT milk, but have shown that it is suitable only for monitoring very severe damage because it is formed only in small amounts. Another useful parameter is 5-hydroxymethylfurancarboxaldehyde (HMF). Its formation in foods has been explained in two ways: via the Amadori products
Thermal processing and food quality: analysis and control
151
through enolisation (in the presence of amino groups) and via lactose isomerisation and degradation, known as the Lobry de Bruyn-Alberda van Ekenstein transformation (Ames 1992). Owing to this, recently it has been proposed to measure separately the HMF formed only by the acidic degradation of Amadori products, called bound HMF and directly related to the MR, and total HMF related also to the degradation of other precursors (Morales et al. 1997). This method is more reliable than the previous spectrophotometric one (Keeney and Bassette 1959). "-Pyrrolelysine is another parameter that has been proposed to measure the MR in foods. It was observed for the first time in the reaction between glucose and lysine (Nakayama et al. 1980). It is particularly useful in dry foods because it is very stable. Resmini and Pellegrino (1994) have proposed a method to measure protein-bound "-pyrrolelysine in dried pasta. The formation of this MRP parallels very well the formation of furosine. LAL, though not deriving from the MR, can be a useful parameter in cheese where the very low residual concentration of lactose impairs the MR and makes other decomposition routes not involving sugars more probable. Pellegrino et al. (1996) have proposed a very sensitive method to detect the addition of caseinates to mozzarella cheese based on LAL determination by HPLC with fluorimetric detection..
8.9
Application to the processing of particular foods
As indicated above, the MR is of critical importance for the production of aroma, taste and colour of foods, that can be roughly divided into two classes. The first class contains foods where a natural and fresh appearance is required such as, for example, drinkable milk or fruit juices. In this case, the thermal treatment is applied to sterilise and the modifications induced by the MR must be reduced to a minimum. Typical cooked, roasted or baked foods such as bread, cereals, chocolate, coffee, nuts, malt, and cooked meat belong to the second class. In this case, the MR is the tool to attain the specific sensory and texture features. In this chapter three foods are discussed as examples of a very mild process (milk), a medium temperature process (bread), and a very drastic one (coffee). 8.9.1 Milk Milk is heated to improve its shelf-life and to kill disease-causing microorganisms and viruses. Three processes are used: 1. 2. 3.
pasteurisation at 85ºC for 2–3 s or at 72–75ºC for 15–30 s in a plate heater; UHT treatment in coils or plate at 136–138ºC for 5–8 s or by steam injection at 140–145ºC for 2–4 s; autoclave sterilisation in closed bottles.
152
Thermal technologies in food processing
In each of them the time and temperature conditions must be selected carefully to reduce the MR to a minimum. Drinkable milk is probably one of the foods that has been more studied from this point of view and great importance is attributed to the preservation of its nutritional value. Many parameters can be used to identify correctly the different milk classes and some of them are based on the MR. Although the pH is neutral and the sugar concentration is around 3.5%, the recent industrial technologies of the production of pasteurised and UHT milk are rather mild and milk contains only lactose, a disaccharide relatively slow to react. In these conditions only the first steps of the MR occur and lysine glycation is the main consequence, while aroma and colour remain practically unchanged. The main product is therefore lactulosyllysine, that can be quantified indirectly after acid hydrolysis as furosine. Many methods have been proposed for its quantification, on one of the best being proposed by Resmini et al. (1990). In raw and pasteurised milk, the level of furosine is around 3–5 mg/100 g protein, in UHT milk is 5–220 mg/ 100 mg, in sterilised milk is > 300 mg/100 g. Abnormal values indicate fraud, for example the addition of reconstituted milk powder. Great interest has been dedicated to the nutritional value of dry milk powder, whose most important use is for the preparation of infant formulas. The main processes used include preconcentration using film evaporating systems to 40– 50% solids and spray drying. An alternative method is drum drying where the liquid is applied in a thin layer to a heated cylinder and after some time removed by a knife. With the latter method, however, the thermal exposure is much higher. 8.9.2 Baked products Wheat flour contains only very small amounts of free sugars and the fermentation process is very important to generate glucose and maltose that are indispensable MR precursors. This can be improved by the addition of flour from malted grains very rich in -amylase. Bread baking time and temperatures vary very much as a function of the bread dimensions and constituents, in general small products (about 45 g) may be heated for 18–20 min at 240–250ºC, while big products (1000 g) may be heated at 240–220ºC for 55–65 min. Naturally, especially in the case of large products, the heat transfer occurs slowly and there is a temperature gradient from the outer to the inner part of the dough. Water evaporates efficiently only in the crust, while the inner part remains softer and colder. The extent of the MR and its consequences are not homogeneous: the differences in colour, flavour and texture of the crumb and the crust are probably part of the hedonistic pleasure of bread consuming. In crumb the most important flavours are autoxidation products of linoleic acid, methional and diacetyl, while 2-methyl-3-ethylpyrazine, 2-acetylpyrazine, 2acetyl-1-pyrroline and 5-methyl-(5H)cyclopenta(b)pyrazine and furaneol are responsible for the caramel, malty and roasty notes of crust. Free proline is the precursor of many volatile compounds that have been characterised in detail
Thermal processing and food quality: analysis and control
153
(Tressl et al. 1985). The pleasant colour of crust is due in part to the MR and in part to caramelisation of sugars. 8.9.3 Coffee During roasting, coffee beans undergo many pyrolytic reactions, which lead to the substances responsible for its particular colour and flavour. The knowledge of green coffee composition is still incomplete and during roasting other reactions take place besides the MR, so that the composition of roasted coffee, especially the non-volatile part, is still far from being clear. The process can be divided into two phases: drying and roasting. An important parameter is heat transfer which is correlated with time and temperature of roasting. The intensity of colour is correlated with the final roasting temperature. The taste changes from slightly sweet (green coffee) to bitter. Most of the constituents are transformed, with some exceptions such as caffeine. Free amino acids and sucrose, which is around 8.0% in arabica green coffee and 4.0% in robusta green coffee, disappear completely, but also polysaccharides are transformed, together with other compounds not directly involved in the MR like chlorogenic acid. Any effect that changes the concentration of the precursors in green beans before roasting influences the quality of the roasted material. At the end the condensation and caramelisation products are about 25%. More than 700 volatile compounds have been identified in coffee aroma: important classes deriving from the Maillard reaction are pyrans, pyrazines, pyridines, pyrones, pyrroles, thiazoles and thiophenes (Illy and Viani 1995).
8.10
Future trends
Extrusion cooking is a high-temperature/short-time technique was introduced in 1930s in the cereal industries. Nowadays this process is used extensively in the food industry for the production of expanded snacks, breakfast cereals, textured soy protein, and other foods. Due to the nature of this technique, involving high temperatures, pressures and shear forces, the food matrix is subjected to chemical changes (gelatinisation of starch molecules, cross-linking of proteins) and the production of flavours can be induced by the Maillard reaction (Riha et al. 1996). Another recent technique for food processing is microwave heating. As experienced also during domestic cooking, the aroma profile of foods cooked in this way is a very critical point. The lack of volatiles is due to the speed of heating and surface moisture and temperature; actually many aroma compounds are steam volatile and during microwave heating they evaporate rapidly; moreover the surface aw, that generally is around 1.0, is very unfavourable for the Maillard reaction. In order to enhance the flavour profile of microwaved foods several approaches have been attempted, e.g. addition of commercial
154
Thermal technologies in food processing
flavourings or Maillard precursors, use of microwave absorbing susceptor sheets, combination of heating treatment (van Eijk 1992).
8.11
Sources of further information and advice
The reader will find a useful discussion of the consequences of thermal treatment of the most relevant foods in Belitz and Grosch (1999). Every 3–4 years the researchers meet for international symposia, whose proceedings are a very useful source of information (Waller and Feather 1983; Fujimaki et al. 1986; Finot et al. 1990; Parliment et al. 1994; O’Brien et al. 1998).
8.12
References
AHMED M U, THORPE S R, BAYNES J W
(1986), ‘Identification of N-"-carboxymethyllysine as a degradation product of fructoselysine in glycated protein’. J. Biol. Chem., 61, 4889–94. AMES J M (1992), ‘The Maillard reaction’. In: Hudson, B. J. F., Biochemistry in Food Proteins. Elsevier, London, 1992, 99–153. AMES J M, NURSTEN H E (1989), ‘Recent advances in the chemistry of coloured compounds formed during the Maillard reaction’. In: Lien, W. S., Foo, C. W. Trends in Food Science. Institute of Food Science and Technology, Singapore, 1989, 8–14. ¨ GERSTAD M (1998), ‘Kinetics of ARDVISSON P, VAN BOEKEL M A J S, SKOG K, JA formation of polar heterocyclic amines in a meat model system’. J. Chromat. A., 227–33. ARNOLDI A (1992), ‘Pyrazines. Part 1’. Riv. It. EPPOS 11, 25–31. ARNOLDI A (1993), ‘Pyrazines. Part 2’. Food Occurence Riv. It. EPPOS. 7, 33–9. ARNOLDI A, CORAIN E A, SCAGLIONI L, AMES J M (1997), ‘New colored compounds from the Maillard reaction between xylose and lysine’. J. Agric. Food Chem., 45, 650–5. BELITZ H-D, GROSCH W (1999), Food Chemistry, Springer, Berlin. ¨ SER W, ERBERSDOBLER H F (1986), ‘Carboxymethyllysine, a new compound of BU heat damage in milk products’. Milchwissenschaft, 41, 780–5. CHIANG G H (1983), ‘A simple and rapid high-performance liquid-chromatography procedure for determination of furosine lysine-reducing sugar derivative’. J. Agric. Food Chem., 31, 1373–4. CHIU E-M, KUO M-C, BRUECHERT L J, HO C-T (1990), ‘Substitution of pyrazines by aldehydes in model systems’. J. Agric. Food Chem., 38, 58–61. CHIU W K, TANAKA M, NAGASHIMA Y, TAGUCHI T (1991), ‘Prevention of sardine lipid oxidation by antioxidative Maillard reaction products prepared from fructose-tryptophan’. Nippon Suisan Gakaishi 57, 1773–81. CUNNINGHAM D G, ACREE T E, BARNARD J, BUTTS R M, BRAELL P A (1986), ‘Charm analysis of apple volatiles’. Food Chem. 19, 137–47.
Thermal processing and food quality: analysis and control EINARSSON H, GORAN S, SNYGG B G, ERIKSSON C
155
(1983), ‘Inhibition of bacterial growth by Maillard reaction products’. J. Agric. Food Chem. 31, 1043– 7. EINARSSON H, EKLUND T, NES I F (1988), ‘Inhibitory mechanisms of Maillard reaction products’. Microbios, 53, 27–36. ERBERSDOBLER H F (1989), ‘Protein reactions during food processing and storage – Their relevance to human nutrition’. Bibl. Nutr. Dieta, 43, 140–55. ERBERSDOBLER H F, ZUCKER H (1966), ‘Untersuchungen zum gehalt an lysin und verfu¨gbarem lysin in trockenmagermilch’. Milchwissenschaft, 21, 564–8. FARMER L J, MOTTRAM D S (1990), ‘Interaction of lipid in the Maillard reactions between cysteine and ribose: the effect of a triglyceride and three phospholipids on the volatile compounds’. J. Sci. Food Agric. 53, 505–25. FARMER L J, MOTTRAM D S, WHITFIELD F B (1989), ‘Volatile compounds produced in Maillard reactions involving cysteine, ribose and phospholipid’. J. Sci. Food Agric., 49, 347–68. FINOT P A, AESCHBACKER H U, HURRELL R F, LIARDON R (1990), The Maillard Reaction In Food Processing, Human Nutrition and Physiology. Adv. Life Science. Birkha¨user, Basel. FORS S (1983), ‘Sensory properties of volatile Maillard reaction products and related compounds. A literature review’. In: Waller G. R., Feather M.S. The Maillard Reaction in Food and Nutrition. ACS Symp. Ser. 215, American Chemical Society, Washington DC, 1983, 185–286. FRIEDMAN M (1996), ‘Browning and its prevention: an overview’. J. Agric. Food Chem. 44, 631–53. FRIEDMAN M (1999), ‘Chemistry, biochemistry, nutrition, and microbiology of lysinoalanine, lanthionine, histidinoalanine in food and other proteins’. J. Agric. Food Chem. 47, 1295–319. FUJIMAKI M, NAMIKI M, KATO H (1986), Amino-carbonyl reactions in food and biological systems. Development in Food Sciences 13., Elsevier, Amsterdam. GUMBMANN M R, FRIEDMAN M, SMITH G A (1993), ‘The nutritional values and digestibilities of heat damaged casein and casein-carbohydrates mixtures’. Nutr. Rep. Int. 355–61. HARTKOPF J, ERBERSDOBLER H F (1993), ‘Stability of furosine during ionexchange chromatography in comparison with reverse-phase HPLC. J. Chromatogr., 635, 151–4. HARTKOPF J, ERBERSDOBLER H F (1994), ‘Model studies of conditions for the formation of N-"-carboxymethyllysine in food’. Z. Lebensm. Unters. Forsch., 198, 15–19. HASHIBA H (1973), ‘Non enzymic browning of soy sauce. Use of ion exchange resin to identify types of compounds involved in oxidative browning’. Agric. Biol. Chem., 37, 871–7. HENLE T, ZEHTNER G, KLOSTERMEYER H (1995), ‘Fast and sensitive determination of furosine in food’. Z. Lebensm. Unters. Forsch., 200, 235–7. HEWEDY M, KIESNER C, MEISSNER K, HARTKOPF J, ERBERSDOBLER H F (1991),
156
Thermal technologies in food processing
‘Effect of UHT heating milk in an experimental plant on several indicators of heat treatment’. J. Dairy Res., 61, 304–9. HODGE J E (1953), ‘Chemistry of browning reactions in model systems’. J. Agric. Food Chem., 1, 928–43. HOFMANN T (1998a), ‘Characterization of the most intensely coloured compounds from Maillard reactions of pentoses by application of colour dilution analysis’. Carbohydr. Res., 313, 203–13. HOFMANN T (1998b), ‘Characterization of precursors and elucidation of the reaction pathway leading to a novel coloured 2H, 7H, 8aH-pyrano[2, 3b]pyran-3-one from pentoses by quantitative studies and application of 13 C-labelling experiments’. Carbohydr. Res., 313, 215–24. HOFMANN T (1998c), ‘Studies on the relationship between molecular weight and color potency of fractions obtained by thermal treatment of glucose/amino acid and glucose/protein solutions by using ultracentrifugation and color dilution techniques’. J. Agric. Food Chem., 46, 3891–5. HOFMANN T (1998d), ‘Identification of novel colored compounds containing pyrrole and pyrrolinone structures formed by Maillard Reactions of pentoses and primary amino acids’. J. Agric. Food Chem., 46, 3902–11. HUANG T C, BRUECHERT L J, HARTMAN T-C, ROSEN R T, HO C-T (1987), ‘Effect of lipids and carbohydrates on thermal generation of volatiles from commercial zein’. J. Agric. Food Chem., 35, 985–90. IARC (1993), Monographs on the Evaluation of Carcinogenic Risk to Humans: Vol. 58. Some Naturally Occurring Aromatic Amines and Mycotoxins, International Agency for Research on Cancer, Lyon, 163–242. ILLY A, VIANI R (1995), Espresso Coffee – The Chemistry of Quality. Academic Press, London. KEENEY M, BASSETTE R (1959), ‘Detection of intermediate compounds in the early stages of browning reaction in milk products’. J. Dairy Sci., 42, 945–60. KNIZE M G, CUNNINGHAM P L, AVILA J R, GRIFFIN E A JR, FELTON J S (1994) ‘Formation of mutagenic activity from amino acids heated at cooking temperatures’. Food Chem. Toxicol., 32, 55–60. KUNTCHEVA M J, OBRETENOV T D (1996), ‘Isolation and characterization of melanoidins from beer’. Z. Lebensm. Untersch., 202, 238–43. LAYTON D W, BOGEN K T, KNIZE M G, HATCH F T, JOHNSON V M, FELTON J S (1995), ‘Cancer risk of heterocyclic amines in cooked foods; an analysis and implications for research’. Carcinogenesis, 16, 39–52. LEDL F, SCHLEICHER E (1990), ‘New aspects of the Maillard reaction in foods and in the human body’. Angew. Chem. Int. Ed., 29, 565–708. LEE Y S, HOMMA S, AIDA K (1987), ‘Characterization of melanoidins in soy sauce and fish sauce by electrofocusing and high performance gel permeation chromatography’. J. Jap. Soc. Food Sci. Technol. 34, 313–19. MAGA J A (1982), ‘Pyrazines in Flavour’. In: Morton, I. D., MacLeod, A. J., Food Flavours. Part A. Introduction., Elsevier, Amsterdam, 283–323. MAIER H G, BUTTLE H (1973), ‘Isolation and characterization of brown compounds of coffee’. Z. Lebensm. Untersch., 150, 331–4.
Thermal processing and food quality: analysis and control
157
(1912), ‘Action des acides amines sur les sucres. Formation des melanoidines par voie methodologique’. C. R. Acad. Sci., 154, 66–8. MANLEY C H (1994), ‘Process flavour and precursors systems’. In: Parliment T. H.; Morello M. J.; McGorrin, R.J. Thermally Generated Flavors. Maillard, Microwave and Extrusion Processes. ACS Symp. Ser. 543, American Chemical Society, Washington DC, 16–25. MILIC B L J, GRUJIC INJAC B, PILETIC M V, LAJSIC S, KOLAROV L J A (1975), ‘Melanoidins and carbohydrates in roasted barley’. J. Agric. Food Chem., 23, 960–3. MORALES F, ROMERO C, JIMENEZ-PEREZ S (1997), ‘Chromatographic determination of bound hydroxymethylfurfural as an index of milk protein glycosylation’ J. Agric. Food Chem., 45, 1570–3. 14 MORI B, NAKATSUJI H (1977), ‘Utilization in rats of C-L-lysine-labeled casein browned by the amino-carbonyl reactions’. Agric. Biol. Chem., 41, 345–50. NAKAYAMA T, HAYASE F, KATO H (1980), ‘Formation of "-(2-formyl-4hydroxymethyl-pyrrol-1-yl)-L-norleucine in the Maillard reaction between D-glucose and L-lysine’. Agric. Biol. Chem., 44, 1201–2. NEGRONI M, D’AGOSTINA, ARNOLDI A (2000), ‘Autoxidation in xylose/model systems’. J. Agric. Food. Chem., 48, 479–83. NURSTEN H E (1980–81), ‘Recent developments in studies of the Maillard reaction’. Food Chem., 6, 263–77. O’BRIEN J, NURSTEN H E, CRABBE M J C, AMES J M (1998), The Maillard Reaction in Foods and Medicine. Royal Society of Chemistry, Cambridge, 1998. OBRETENOV T D, KUNTCHEVA M J, MANTCHEV S C, VALKOVA G D (1991), ‘Isolation and characterization of melanoidins from malt and malt roots’. J. Food Biochem., 15, 279–94. PARLIMENT T H, MORELLO M J, MCGORRIN R J (1994), Thermally Generated Flavors. Maillard, Microwave, and Extrusion Processes. ACS Symposium Ser. 543, Washington DC. PELLEGRINO L, RESMINI P, DE NONI I, MASOTTI F (1996), ‘Sensitive determination of lysinoalanine for distinguishing natural from imitation mozzarella cheese’. J. Dairy Sci., 79, 725–34. RAVAGLI A, BOSCHIN G, SCAGLIONI L, ARNOLDI A (1999), ‘Reinvestigation of the reaction between 2-furancarboxyaldehyde and 4-hydroxy-5-methyl-3(2H)furanone’. J. Agric. Food Chem., 47, 4962–9. RESMINI P, PELLEGRINO L (1994), ‘Occurence of protein-bound lysylpyrrolealdehyde in dried pasta’. Cereal Chem., 71, 254–62. RESMINI P, PELLEGRINO L, BATELLI G (1990), ‘Accurate quantification of furosine in milk and dairy products by a direct HPLC method’. Ital. J. Food Sci., 3, 173–83. RIHA W E III, HO C-T (1996), ‘Formation of flavors during extrusion cooking’. Food Rev. Int., 12, 351–73. ROBERTS D D, ACREE T (1994), ‘Gas chromatography – olfactometry of glucoseproline Maillard reaction products’. In: Parliment T. H.; Morello M. J., McGorrin, R. J. Thermally Generated Flavors. Maillard, Microwave and MAILLARD A C
158
Thermal technologies in food processing
Extrusion Processes. ACS Symposium Ser. 543, American Chemical Society, Washington DC, 71–9. SCHIEBERLE P, OFNER S, GROSCH W (1990), ‘Evaluation of potent odorants in cucumbers (Cucumis sativus) and muskmelons (Cucumis melo) by aroma extract dilution analysis’. J. Food Sci., 55, 193–5. SKOG K (1993), ‘Cooking procedures and food mutagens: a literature review’. Food Chem. Toxicol., 31, 655–75. ¨ GERSTAD M (1998), ‘Analysis of non SKOG K, SOLYAKOV A, ARDVISSON P, JA polar heterocyclic amines in cooked foods and meat extracts using gas chromatography-mass spectrometry’. J. Chromat. A., 803, 227–33. ¨ GERSTAD M (1995), ‘Effect of cooking SKOG K, STEINECK G, AUGUSTSSON K, JA temperature on the formation of heterocyclic amines in fried meat products and pan residues’ Carcinogenesis, 16, 861–7. ¨ VERVIK E (1993), ‘The STEINECK G , GERHARDSSON DE VERDIER M, O epidemiological evidence concerning intake of mutagenic activity from fried surface and the risk of cancer cannot justify preventive measures’. Eur. J. Cancer Prev., 2, 293–300. STEINHART H, PACKERT A (1993), ‘Melanoidins in coffee. Separation and characterization by different chromatographic procedures’. Colloq. Sci. Int. Cafe [C. R.], 15th (vol. 2), 593–600. SUGIMARA T, NAGAO M, KAWACHI T, HONDA M, YAHAGI T, SEINO Y, SATO S, MATSUKARA N, SHIRAI A, SAWAMURA M, MATSUMOTO H (1977), Mutagens – carcinogens in food, with special reference to highly mutagenic pyrolytic products in boiled foods’. In: Hiatt H. H., Watson J. D., Winsten J. A., Origins of Human Cancer. Cold Spring Harbor Laboratory, 1561– 77. SUGIMARA T, WAKABAYASHI K, NAGAO M, ESUMI H (1993), ‘A new class of carcinogens: heterocyclic amines in cooked food’. In: Parke D. V., Ioannides C., Walker R., Food, Nutrition and Chemical Toxicity. SmithGordon and Company, London, 259–76. TRESSL R, REWICKI D, HELAK B, KAMPERSCHROER H (1985), ‘Formulation of pyrrolidines and piperidines on heating L-proline with reducing sugars’. J. Agric. Food Chem., 33, 924–8. TRESSL R, REWICKI D (1999), ‘Heat generated flavors and precursors’. In Flavor Chemistry: 30 Years of Progress. ACS Symposiums Series, American Chemical Society, Washington DC, 305–25. TRESSL R, NITTKA CH, KERSTEN E, REWICKI D (1995), ‘Formation of isoleucine specific Maillard products from [1-13C]-D-glucose and [1-13C]-Dfructose’. J. Agric. Food Chem., 43, 1163–9. ¨ GER R P, REWICKI D (1998a), ‘New melanoidin-like TRESSL R, WONDRAK G T, KRU Maillard polymers from 2-deoxypentoses’. J. Agric. Food Chem., 46, 104– 10. ¨ GER R P, REWICKI D, GARBE L-A (1998b), ‘Pentoses TRESSL R, WONDRAK G T, KRU and hexoses as source of new melanoidin-like Maillard-polymers’. J. Agric. Food Chem., 46, 1756–76.
Thermal processing and food quality: analysis and control
159
(1992), ‘Flavor and Flavorings in microwave foods: an Overview’. In: Parliment, T. H., Morello M. J., and McGorrin, R. J., Thermally Generated Flavors, ACS Symposiums Series 543, American Chemical Society, Washington DC, 395–404. WALLER G R, FEATHER M S (1983), The Maillard Reaction in Foods and Nutrition. ACS Symposium Ser. 215, Washington DC. WHITFIELD F B (1992), ‘Volatiles from the interactions of the Maillard reaction and lipids’. Crit. Rev. Food Sci. Nutr., 31, 1–58. WONDRAK G T, TRESSL R, REWICKI D (1997), ‘Maillard reaction of free and nucleic acid-bound 2-deoxy-D-ribose and D-ribose with !-amino acids’. J. Agric. Food Chem., 45, 321–7. YAYLAYAN V A (1997), ‘Classification of the Maillard reaction: a conceptual approach’. Trends Food Sci. Technol. 8, 13–18. YAYLAYAN V A, HUYGGUES-DESPOINTES A (1994), ‘Chemistry of Amadori rearrangement products: analysis, synethesis, kinetics, reactions, and spectroscopic properties’. Crit. Rev. Food Sci. Nutr., 34, 321–8. VAN EIJK T
This page intentionally left blank
Part III New thermal technologies
This page intentionally left blank
9 Radio frequency heating A. T. Rowley, EA Technology Ltd, Chester
9.1
Introduction
Radio frequency (RF) or high frequency (HF) heating, is more correctly described as RF dielectric heating, and is a well established thermal processing technology which has found applications in many areas. The most notable of these are found in the plastics (welding1), textiles (drying2), paper and board (drying3), wood (gluing4 and drying4) and food industries. RF heating has much in common with microwave and ohmic heating – all three are electroheat technologies in which heat is generated volumetrically throughout a product rather than having to rely on the slow conduction of heat through its surface. Following an overview of the fundamentals of RF heating, this chapter will outline why the use of this technology in the food processing industry can lead to clear advantages over other thermal techniques. In particular, the technology will be compared to microwave and ohmic heating. The limitations of the technology will also be discussed. A major section of this chapter describes the technology associated with RF heating and how it is generally used for food processing. Two case studies, RF-assisted baking and meat defrosting, will be presented which emphasise the importance of RF heating to the food processing industry. Finally, the chapter concludes by considering the future direction of the technology, and the implications for the thermal processing of food.
9.2
Basic principles of RF heating
In very simple terms, radio frequency heating of foods arises from the direct conversion of electrical energy to heat within the volume of the food itself. This
164
Thermal technologies in food processing
electrical energy is provided by a high frequency electric field applied between the plates of a capacitor (the RF applicator). In electrical terms, foods are dielectrics (materials which increase the charge stored in a capacitor) which possess a significant dielectric loss factor (an indication of the amount of energy which will be dissipated). The term dielectric heating can equally be applied to radio frequency and microwave systems – in both cases the heating is due to the fact that energy is absorbed by a lossy dielectric when it is placed in a high frequency electric field. In foods, at radio frequencies, this loss principally arises from the electrical conductivity of the food, and the heating mechanism is simply resistance heating (i.e. similar to ohmic heating). Although microwave heating also relies on a dielectric loss to provide the heat, the principal loss mechanism in food products at microwave frequencies is different (resonant dipolar rotation5). The radio and microwave frequency bands occupy adjacent sections of the electromagnetic spectrum, with microwaves having higher frequencies than radio waves. The actual dividing point between the two frequency bands is imprecisely defined, with, for example, some applications at around 900 MHz being referred to as RF (cellular telephones), and some as microwaves (dielectric heating). However, the technology used to generate and transmit the high frequency electric fields can be used to distinguish them. RF systems are generally based on high power electrical valves (to produce the RF power), transmission lines (to carry the RF energy), and applicators in the form of capacitors; whereas microwave systems use magnetrons (to generate the microwaves), waveguides (to transport the microwaves) and cavities (in which the microwaves are applied). There is a relatively small number of internationally agreed and recognised frequency bands which can be used for RF and microwave heating. These are known as the Industrial, Scientific and Medical (ISM) bands, and are defined in Table 9.1. Electromagnetic compatibility (EMC) regulations set very low limits for any emissions outside of these bands, and, in most countries, compliance is a legal obligation. Consequently, virtually all RF process heating equipment will operate at one of the three allowed ISM frequencies. It is worthwhile noting that the wavelength at radio frequencies is substantially greater than at microwave frequencies – 11 m at 27.12 MHz compared with only 12 cm at 2450 MHz. It is this difference which leads to a number of significant advantages of RF over microwaves, particularly for industrial food processing applications. Table 9.1
The ISM bands available for dielectric heating
Heating technology
Frequency
Radio frequency
13.56 MHz 0.05% ( 0.00678 MHz) 27.12 MHz 0.6% ( 0.16272 MHz) 40.68 MHz 0.05% ( 0.02034 MHz) 900 MHz (depending on country) 2450 MHz 50 MHz
Microwave
Radio frequency heating
165
Fig. 9.1 The effect of a dielectric on a capacitor.
A calculation of the actual amount of energy (or power) absorbed by a dielectric body is essential to a full understanding of radio frequency (or microwave) heating. An expression for the power dissipated in a dielectric can be derived following directly from the premise that, in essence, all applicators used for RF dielectric heating are some form of capacitor. When a dielectric material such as food (with a dielectric constant, "r0 , and a dielectric loss factor, "r00 ) is placed in this capacitor, it will be affected in two ways (see Fig. 9.1). First, the new effective capacitance, C0 , will be greater than the original capacitance (Co) by a factor er’ (by definition, "r0 is always greater than one), and secondly, a finite resistance, R (proportional to 1/Co"r00 ), will appear across the capacitor. The increase in capacitance arises from a change in the distribution of electric charge within the RF applicator, and the presence of the resistance gives the possibility of heat generation within the dielectric. Assuming that the power, P, dissipated in this resistance to be equal to V2/R, then it can be shown that the power dissipation per unit volume or power density, Pv, is given by:6 Pv 2f "0 "00r E2
9:1
where f is the frequency of the applied field (RF), "0 is a constant (the permittivity of free space) and E is the electric field strength in the dielectric. Although derived in another way, the same equation is used to describe microwave dielectric heating.5 Inspection of equation [9.1] reveals that the power density is proportional to the frequency of the applied field and the dielectric loss factor, and is proportional to the square of the local electric field. This equation is crucial in determining how a dielectric will absorb energy when it is placed in a high frequency electric field. For a given system, the frequency is fixed and f and "0 are both constants. The dielectric loss factor "00r can, in principle, be measured. The only unknown left in equation [9.1] is the electric field, E. To evaluate this, the effect of the dielectric material itself on the applied electric field (due to the RF voltage across the RF applicator) must be considered. For materials, such as food products, where the dielectric loss arises principally from the electrical conductivity, then the loss factor, "00r , is given by "00r =2f "0 , and equation [9.1] can be further reduced to Pv E2
9:2
166
Thermal technologies in food processing
which is the same equation as that used to describe ohmic heating. However, unlike ohmic heating, in RF heating this electric field can be achieved without the need for the electrodes to be in contact with the product. This is a consequence of the much higher frequency used.
9.3
Application to food processing
Radio frequency heating has been used in the food processing industry for many decades. In particular, RF post-baking of biscuits and cereals, and RF drying of foods are well established applications. More recently, RF thawing equipment has demonstrated substantial benefits over conventional techniques and over comparable microwave tempering systems. Furthermore, as the public concern over food safety issues continues to grow, and as the demand for convenience foods increases, RF pasteurisation and sterilisation processes are becoming more important. 9.3.1 Baking The post-baking of biscuits7 is one of the most accepted and widely used applications of RF heating in the food processing industry. The addition of a relatively small RF unit to the end of a conventional baking line results in a substantial increase in product throughput, together with improvements in product quality. Similarly, the same process has also been applied to cereal, pastry and bread products. More recently, the RF system has been incorporated directly into the hot air oven, allowing RF-assisted baking of a wide range of products to be carried out in a very compact unit. 9.3.2 Drying The principal role of RF heating in baking is the removal of moisture, particularly at the end of the process when conventional heating is inefficient. RF drying is intrinsically self-levelling,6 with more energy being dissipated in wetter regions than in drier ones. This RF levelling leads to improvements in product quality and more consistent final products. As well as baking applications, RF drying applications in the food industry include the drying of: food ingredients (e.g. herbs, spices, vegetables); potato products (e.g. French fries), and a number of pasta products. 9.3.3 Defrosting A more recent, and rapidly growing, application of RF in the food processing industry is its use for the bulk defrosting of meats and fish. Conventionally, large blocks of meat are thawed slowly, often over a period of days. The volumetric nature of RF heating allows the thawing process to be accelerated, whilst still
Radio frequency heating
167
maintaining control of the temperature distribution within the food product. Typically, RF defrosting times of 1–2 hours are possible. 9.3.4 Pasteurisation/sterilisation With the increasing demand for convenience foods and ‘ready meals’, re-heating of food products is becoming more common, often carried out in a microwave oven, and it can no longer be assumed that all of the food is heated to a high enough temperature to kill bacteria. Consequently, there is a demand for more in-package pasteurisation. As a non-contact volumetric heating technique, RF is an ideal process for this application. Increased public concern and awareness of food safety issues are leading to more RF pasteurisation and sterilisation applications being investigated, and it is likely that a number of these will become industrial processes in the near future.
9.4
Advantages and disadvantages of RF heating
9.4.1 Advantages In common with microwave and ohmic heating, the volumetric nature of radio frequency heating gives rise to a number of significant advantages over more traditional, surface heating techniques. The most important of these to the food industry are: • Improved food quality. The main reason for using RF heating in food processing (rather than any other thermal technology) is improved food quality. First, the volumetric process leads to more uniform heating, removing the risk of overheating food surfaces while trying to heat the centre of products. Secondly, the selective nature of RF heating, with energy being dissipated according to the local loss factor, can produce very uniform products, even when there are relatively large variations in the unprocessed food. • Increased throughput. Conventional surface heating often has to heat foods relatively slowly to avoid the risk of overheating the surface. Moreover, once the surface of foods have dried out, they often form a good thermal barrier layer, making it even more difficult to heat the centres. By contrast, volumetric RF heating avoids these effects, allowing production lines to operate much faster. • Shorter process lines. As an alternative to increased throughput, food processing lines which include RF systems can be significantly shorter for a given throughput. • Improved energy efficiency. Since the RF energy is dissipated directly within the product being heated, processing lines using this technology can be very efficient, particularly when the increased throughput is also taken into account.
168
Thermal technologies in food processing
• Improved control. Since the power dissipated within the food is due to the presence of an electric field (equation [9.1]), if this field is changed, or switched off, the heating of the food products responds almost instantaneously. In this way, RF heating can be controlled very precisely, again leading to improvements in food quality
In food processing, RF heating also has a number of advantages over the alternative volumetric technologies, namely microwave and ohmic heating. The main ones are: • Contactless heating. Although the heating mechanism is essentially the same as with ohmic heating, RF does not require the electrodes to be in contact with the food. This removes the constraint that the food product has to be pumpable, and allows RF heating to be applied to solid as well as liquid heating. • Increased power penetration. The longer wavelength at radio frequencies compared with microwave frequencies, and the dielectric properties of foods, mean that RF power will penetrate further into most products than microwave power. For example, the penetration depth (the distance for the power to fall to 1/e of its initial value) in unfrozen meat products is typically only a few millimetres at microwave frequencies, but tens of centimetres at RF frequencies. • Simpler construction. Large RF applicator systems are generally simpler to construct than microwave ones. In particular, the longer wavelength at radio frequencies allows relatively large entry and exit ports to be designed – 2 m wide ports are not untypical. Moreover, the geometries of RF applicator systems (see Section 9.5.3) naturally lend themselves to industrial food processing applications. • Improved moisture levelling. In food products, the variation of the dielectric loss factor with moisture content is generally greater at radio than at microwave frequencies. Consequently, the use of RF heating for baking and drying applications leads to improved moisture levelling and correspondingly higher quality final products.
9.4.2 Disadvantages When compared with conventional heating techniques, and, to some extent with ohmic heating, the main disadvantages of RF heating relate to equipment and operating costs. In comparison with microwave heating, the main limitation of RF heating arises from a lower power density. • Equipment and operating cost. For an equivalent power output, RF heating equipment is more expensive than conventional convection, radiation or steam heating systems. It is also more expensive than an equivalent ohmic heating system. However, in some applications, improvements in product quality and throughput often more than justify the initial capital investment.
Radio frequency heating
169
As an electroheat technology, the unit energy costs of a RF system will be higher than an equivalent conventional heating system. Nevertheless, when factors such as increased energy efficiency and increased throughput are taken into account, the total energy cost may be comparable to (or even less than) a conventional system. • Reduced power density. Equation [9.1] shows that the power density is directly proportional to the frequency. Given that the electric field is limited to avoid the occurrence of an electrical breakdown, then the power density will be much higher at microwave than radio frequencies. The main consequences of this are that RF systems are usually significantly larger than microwave heating systems of the same power rating, and that faster heating rates can often be achieved with a microwave system.
9.5
RF heating technologies
The available systems for producing and transferring RF power to dielectric heating applicators can be divided into two distinct groups; the more widespread conventional RF heating equipment,8 and the more recent 50 RF heating equipment.9 Although conventional RF equipment has been used successfully for many years, the ever tightening EMC regulations, and the need for improved process control, are leading to an increased use of RF heating systems based on 50 technology. 9.5.1 Conventional RF equipment In a conventional system,8 the RF applicator (i.e. the system which applies the high frequency field to the product) forms part of the secondary circuit of a transformer which has the output circuit of the RF generator as its primary circuit. Consequently, the RF applicator can be considered to be part of the RF generator circuit, and is often used to control the amount of RF power supplied by the generator. In many systems, a component in the applicator circuit (usually the RF applicator electrodes themselves) is adjusted to keep the power within set limits. Alternatively, the heating system is set up to deliver a certain amount of power into a standard load of known conditions, and then allowed to drift automatically up or down as the condition of the product changes. In virtually all conventional systems, the amount of RF power being delivered is only indicated by the DC current flowing through the high power valve (usually a triode) within the generator. A typical conventional RF heating system is shown schematically in Fig. 9.2. 9.5.2 50 RF equipment RF heating systems based on 50 equipment are significantly different, and are immediately recognisable by the fact that the RF generator is physically separated from the RF applicator by a high power coaxial cable (Fig. 9.3).
170
Thermal technologies in food processing
Fig. 9.2 Conventional RF dielectric heating system.
The operation frequency of a 50 RF generator is controlled by a crystal oscillator and is essentially fixed at exactly 13.56 MHz or 27.12 MHz (40.68 MHz is seldom used). Once the frequency has been fixed, it is relatively straightforward to set the output impedance of the RF generator to a convenient value – 50 is chosen so that standard equipment such as high power cable and RF power meters can be used. For this generator to transfer power efficiently, it must be connected to a load which also has an impedance of 50 . Consequently, an impedance matching network has to be included in the system which transforms the impedance of the RF applicator to 50 . In effect, this matching network is a sophisticated tuning system, and the RF applicator plates themselves can be fixed at an optimum position.
Fig. 9.3
50 RF dielectric heating system.
Radio frequency heating
171
The main advantages of this technology over the conventional systems are: • Fixed operation frequency makes it easier to meet onerous international EMC regulations. • The use of 50 cable allows the RF generator to be placed at a convenient location away from the RF applicator. This is of particular importance to food processing applications, where the RF applicator will need to be cleaned regularly. • The RF applicator can be designed for optimum performance, and is not itself part of any tuning system. • The use of a matching network gives the possibility of an advanced process control system. The positions of components in the matching network give on-line information on the condition of the dielectric load (such as its average moisture content). This information can be used to control the RF power, the speed of conveyor or the temperature of air in applicator as appropriate.
9.5.3 RF applicators Whether conventional or 50 dielectric heating systems are used, the RF applicator has to be designed for the particular product being heated or dried. Although the size and shape of the applicator can vary enormously, they mostly fall into one of three main types – throughfield, fringefield or staggered throughfield. Whatever the type of applicator, RF food processing systems often benefit from the combination of RF with hot air convection heating.10 This hot air can either be introduced conventionally into the applicator enclosure, or directed onto the surface of the product through the electrodes themselves. This combination of volumetric and conventional surface heating optimises the cooking, baking and drying processes in such a way that relatively small amounts of RF energy can lead to large improvements in throughput and food quality, whilst minimising the size of the combination heating unit. Throughfield applicator Conceptually, a throughfield RF applicator is the simplest, and the most common, design, with the electric field originating from a high frequency voltage applied across the two electrodes which form a parallel plate capacitor (Fig. 9.4a). This type of system can be used for both batch and continuous processing applications, and is mainly used with relatively thick products, or blocks of material. For example, this electrode arrangement is found in RF meat defrosting systems. Fringefield applicator An alternative RF applicator arrangement, often used in drying applications, is known as the fringefield system. In this case, the product passes over a series of bars, rods or narrow plates which are alternately connected to either side of the
172
Thermal technologies in food processing
Fig. 9.4
Alternative RF electrode configurations.
Radio frequency heating
173
RF voltage supply (Fig. 9.4c). The major advantage of this configuration is that the product runs close to the electrode bars and there is only a small air gap between the RF applicator and the product. This ensures that there will be a virtually constant electric field in the material between the bars (an important requirement to maximise moisture levelling performance). It also reduces the electric field that has to be applied between the electrodes to generate a given power density within the product. The major disadvantage of this arrangement is that only relatively thin layers of product can be used, otherwise there will be an electric field variation throughout the product thickness. This electrode arrangement is found in some pasta drying and cereal baking applications. Staggered throughfield applicator For intermediate thickness products, a modified form of the throughfield applicator is often used. This is known as a staggered throughfield applicator (Fig. 9.4b). This arrangement reduces the overall capacitance of the applicator which, in turn, makes the overall system tuning easier. It also reduces slightly the voltage that has to be applied across the electrodes to produce a given RF power density within the product. This electrode configuration is commonly used in RF post-baking applications.
9.6
Case studies
9.6.1 RF-assisted biscuit baking Conventional process In a conventional baking oven, convective and/or radiative energy is applied simultaneously to the top and bottom surface of the biscuits. The biscuit drying proceeds from these surfaces towards the biscuit centre. Once the surface has dried the conventional heat starts to bake the surface. A typical industrial biscuit baking oven is about 60 m long. The main disadvantages arise from the difficulty associated with the use of conventional surface heating to remove the small amount of moisture in the centre of the biscuit at the end of the process. These are: • the biscuit baking line has to be very long, due to the disproportionately high amount of energy needed towards the end of the biscuit baking process • for the same reason, conventional biscuit baking is very energy intensive • the conventional process leads to only average or even poor quality biscuits.
RF process The traditional RF solution to these problems is to add a relatively small (3–4 m long) RF drying unit at the end of the conventional processing line. The intrinsic moisture levelling associated with RF heating allows the final thin layers of moisture at the centre of the biscuit to be rapidly and efficiently removed. Such a RF post-baking unit can increase throughput by up to 30%. More recently, the
174
Thermal technologies in food processing
RF applicator systems have been incorporated directly into the conventional high temperature ovens allowing hot air and RF energy to be applied simultaneously. This RF-assisted biscuit baking system is comprised of a number of RF zones, with typically 40–50 kW of RF power per zone. The main advantages of the RF-assisted unit, over both conventional and RF post-baking systems, are: • the length of the biscuit oven can be substantially shorter • it is a very energy efficient process • biscuit quality is substantially improved: the biscuits are typically larger, have a less dense and more uniform structure, are more difficult to break and have a less ‘pasty’ taste.
9.6.2 Meat defrosting Conventional process Once removed from cold storage, large (typically 10–20 kg) blocks of meat have to be carefully stacked in a temperature controlled room where they slowly thaw out over a period of several days. Even at this slow thawing rate, there is often a large variation in temperature within individual blocks and between different blocks. The main disadvantages of the conventional process are: • the slow processing speed means that the supply of defrosted meat cannot respond to a rapidly changing demand • during conventional thawing, there is a large drip loss from the meat which can account for up to 12% of the product volume, and which reduces significantly the value of the product (meat is sold by weight), and also gives effluent handling problems • the conventional process is very labour intensive • the slow thawing process can lead to significant biological growth which reduces the product shelf-life, and may present a potential health risk • a large amount of floor space has to be allocated to the thawing process.
RF process The RF meat defroster is a continuous, conveyorised unit made up of three independent RF zones, each with a relatively simple throughfield electrode arrangement. A generator with a maximum output power of typically 30–40 kW is connected to each zone. This RF unit can continuously defrost around one tonne of meat per hour, with a defrost time of less than two hours. A typical unit would be around 20 m long with a conveyor belt width of about 2 m. The main advantages of the RF meat defroster are: • with a thawing time of less than 2 hours, the RF unit allows the supply of thawed meat to respond rapidly to any changes in demand • the drip loss is reduced to less than 1% – increasing the value of the final product and reducing any effluent handling problems
Radio frequency heating
175
• any biological growth is minimised, increasing the shelf-life of the meat products, and reducing any health hazard • the RF defrosting process is not labour intensive – up to 75% less than the conventional process • compared with the conventional thawing process, the RF unit is very compact.
9.7
Future trends in RF heating
The future direction of the use of RF heating in the food processing industry will be influenced by many factors. Developments in RF technology could significantly benefit both existing and emerging food applications. Similarly, changes in consumer food preferences could lead to new applications, but could also lead to some existing applications becoming redundant. Moreover, changes in the food (or radio) regulations could have a major impact on the future use of RF heating in the food industry. 9.7.1 Technology Although predicting the future direction of any technology is, at best, difficult, there are a number of general trends in RF heating technology which are likely to influence food processing applications. 50 RF equipment specifically developed for the dielectric heating market is now commercially available, and the technology has been proven for a range of food applications. The clear benefits of this technology over conventional equipment (see Section 9.5.2) will lead to its more widespread future use, both for new processing lines, and as replacements for conventional RF equipment in existing lines. At present, virtually all RF generators used for the thermal processing of foods use thermionic valves. However, future developments in transistor technology (particularly in MOSFETs), will lead to the upper power limit of solid state generator systems being increased from its present level of 5–10 kW. Ultimately, cost competitive solid state generators will be available in the range 20–50 kW typically used in most industrial food processing applications. Such RF generators will be compact, light and very controllable. The same developments in transistor technology could also lead to low power (i.e. 0.5–2 kW) RF generators becoming much cheaper. At present, microwave heating completely dominates dielectric heating in the commercial (and domestic) food processing sectors. Even though RF heating has a number of clear advantages over microwave heating (see Section 9.4.1), the cost of low power RF systems is prohibitively expensive compared to equivalent microwave systems. The availability of low cost RF power sources could lead to a major growth in the use of RF heating in the commercial food sectors.
176
Thermal technologies in food processing
9.7.2 Applications The current and increasing demand for high quality food products will mean that RF post-baking and RF-assisted baking will continue to be important stages in the processing of biscuit, cereal and pastry products. Similarly, public concern over food hygiene issues will continue to require rapid and safe food thawing techniques, such as RF meat and fish defrosting systems. Increasing public awareness of general food safety issues, and the rising demand for convenience, pre-packaged foods will lead to a growth in the demand for RF (and microwave) pasteurisation and sterilisation techniques.
9.8 9.8.1
Sources of further information and advice Further reading
BLEANEY B I, BLEANEY B,
Electricity and Magnetism, Oxford, Oxford University Press, 1978. METAXAS A C, MEREDITH R J, Industrial Microwave Heating, IEE Power Engineering Series, London, Peter Peregrinus Ltd, 1983. METAXAS A C, Foundations of Electroheat – A Unified Approach, Chichester, John Wiley & Sons Ltd., 1996. HULLS P (Secretary, dielectric heating working group), Dielectric Heating for Industrial Processes, UIE, 1992. Journal of Microwave Power and Electromagnetic Energy, The International Microwave Power Institute (IMPI). 9.8.2 Organisations and other contacts EA Technology Capenhurst Chester CH1 6ES, UK Website: http:///www.eatechnology.com EA Technology provide advice and supply RF heating equipment to the food industry. They also have expertise in microwave and ohmic heating systems. British National Committee for Electroheat (BNCE) 30 Millbank London SW1P 4RD, UK Website: http://www.electricity.org.uk/services/bnce/ The International Union for Electroheat (UIE) Tour Atlantique Cedex 06 92080 Paris La Defense, France The International Microwave Power Institute (IMPI) 10210 Leatherleaf Court
Radio frequency heating
177
Manassas, VA 20111, USA Website: http:/www.impi.org/ The Electric Power Research Institute (EPRI) Hillview Avenue, Palo Alto California 94304, USA Website: http:/www.epri.com/
9.9
Acknowledgements
The author would like to thank the Directors of EA Technology for permission to publish this work. In compiling the case studies, the author would also like to gratefully acknowledge the assistance of Petrie Technologies Ltd., Ackhurst Road, Chorley, Lancs PR7 1NH, UK.
9.10 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.
References FEDERATION OF HF WELDERS,
High Frequency Welding Handbook, Ashburton, Federation of High Frequency Welders, 1995. JONES P L, Radio Frequency Drying in the Textile Industry, International Dyer, June 1980. JONES P L, Radio Frequency Processing in Europe, J Microwave Power and Electromagnetic Energy 1987 22(3) 143–53. POUND J, Radio Frequency Heating in the Timber Industry, London, E & FN Spon Ltd., 1973. METAXAS A C, MEREDITH R J, Industrial Microwave Heating, IEE Power Engineering Series, London, Peter Peregrinus Ltd, 1983. JONES P L, ROWLEY A T, Dielectric Dryers. In Baker C J, ed. Industrial Drying of Foods, London, Chapman and Hall, 1997. HOLLAND J M, Dielectric Post-baking in Biscuit Making, Baking Industries Journal, 1974 8(6). HULLS P, Dielectric Heating for Industrial Processes, UIE, 1992. MARCHAND C, Recent Developments in Industrial Radio Frequency Technology, UIE, Proc. High Frequency and Microwave Processing and Heating, Arnhem, 26–29 Sept. 1989, 4.4.1–4.4.10. JONES P L, ROWLEY A T Dielectric Drying, Special issue of Drying Technology, 1996, 14(5) 1063–98.
10 Microwave processing M. Regier and H. Schubert, University of Karlsruhe
10.1
Introduction
In this chapter an overview of microwave heating as one method of thermal food processing is presented. Due to the limited space, this overview cannot be complete; instead some important theoretical information and also examples of practical uses at home and in industry are shown. This chapter provides a starting point, and the interested reader is directed to the references, where more information about the special themes discussed in this chapter can be found. Additional to the references in the text the interested reader is also referred to two bibliographies that cover more or less all the published work on microwaves (Goldblith and Decareau 1973; Dehne 1999). 10.1.1 History of microwave heating The total sales number of microwave ovens in the United States stays at a constant level of approximately 10 million per year (Anon. 1998). The corresponding number in Europe is in the same range. These enormous sales point to the importance of microwave heating today. Nevertheless, it took some time for the development of this technique starting from the microwave source invention in 1921 by Hill (Knutson et al. 1987). The first continuous magnetron (see Section 10.2.4) was built by Randall and Boot who tried to produce a microwave source to power radar sets for the British military during World War II (Reynolds 1989). It was brought to the United States in order to use America’s production potential. Raytheon Co. was the company that received a contract to make copies of the magnetron, where the electrical engineer Spencer improved its manufacturability for large productions. He filed a patent in 1942 concerning
Microwave processing
179
the ameliorated magnetron that was issued nine years later. In 1945 Spencer occasionally observed the heating of various substances (a legend tells about his own body, popcorn and an exploding egg) by the microwave energy of the antenna horn (Reynolds 1989). In the same year, he applied for a patent (issued in 1950) called ‘method of treating foodstuffs’ describing for the first time a closed microwave oven. This technique was applied in Raytheon’s Radarange oven in 1946. With further development and falling prices (also due to the expiring of the basic patent) the domestic microwave oven market grew very fast, starting in the late 1960s, reaching a peak of 12 million ovens sold in the United States in 1988, later becoming constant near 10 million per year. The development of industrial dielectric heating applications started in the radio frequency range in the 1930s (Pu¨schner 1966). Due to the proportionality of the electromagnetic power loss to the used frequency (see equation [10.24]) the energy rate could be enhanced by increasing the frequency. The first patent describing an industrial conveyor belt microwave heating system was issued in 1952 (Spencer 1952). However, the first conveyor belt microwave application only started in 1962 due to the slow development of high power microwave generators. Its first major applications were the finish drying of potato chips, pre-cooking of poultry and bacon, tempering of frozen food and drying of pasta (Decareau 1985). 10.1.2 Today’s uses, advantages and disadvantages of microwave heating applications Today’s uses range from these well known applications over pasteurisation and sterilisation to combined processes like microwave vacuum drying (see also Section 10.3). The rather slow spread of food industrial microwave applications has a number of reasons: there is the conservatism of the food industry (Decareau 1985) and its relatively low research budget. Linked to this, there are difficulties in moderating the problems of microwave heating applications. One of the main problems is that, in order to get good results, they need a high input of engineering intelligence. Different from conventional heating systems, where satisfactory results can be achieved easily by intuition, good microwave application results often do need a lot of knowledge or experience to understand and moderate effects like uneven heating (e.g. edge heating or focused heating) (see Section 10.2.2), or the thermal runaway (see Section 10.2.3)). Another disadvantage of microwave heating as opposed to conventional heating is the need for electrical energy, which is its most expensive form. Nevertheless, microwave heating has a number of quantitative and qualitative advantages over conventional heating techniques that make its adoption a serious proposition. One main advantage is the place where the heat is generated, namely the product itself. Because of this, the effect of small heat conductivities or heat transfer coefficients do not play such an important role. Therefore, larger pieces can be heated in a shorter time and with a more even temperature distribution. These advantages often yield an increased production
180
Thermal technologies in food processing
rate and/or an improved product quality. Another advantage is the almost entire energy conversion from electromagnetic radiation into heat, where it is needed. Depending on the various applications, there could be further advantages like space savings or low noise levels.
10.2
Physical principles
10.2.1 MW-frequency range Microwaves are a kind of electromagnetic wave within a frequency band of 300 MHz to 300 GHz. By equation [10.1] the frequency f is linked by the velocity of light c to a corresponding wavelength . cf
10:1
The term microwaves is a little bit misleading since the vacuum wavelength of them is in the range between 1 m and 1 mm. Their name rather points to their wavelength within the matter, where their wavelength can be in the micrometer range. In practice, for microwave heating applications not all the microwave spectrum is used, in effect there are some discrete frequency bands, which have been set aside from telecommunication applications for industrial, scientific and medical (so called ISM) applications. The most important and most used ISM microwave frequency bands are 915 25 MHz and 2450 50 MHz, where a certain limited radiation level has to be tolerated by other applications (like communication devices). 10.2.2 Maxwell’s equations, wave equations and exemplary solutions Maxwell’s equations As already mentioned above, microwaves belong to electromagnetic waves, which can be basically described with Maxwell’s equations [10.2–10.5]: !
rD !
rE !
10:2 !
@B @t
rB 0 ! ! ! @D rH j @t
10:3 10:4 10:5
In order to include the interactions of matter with electromagnetic fields, the material equations, also called constitutive relations [10.6–10.8], have to be added, where the permittivity or dielectric constant " (interaction of nonconducting matter with an electric field), the conductivity and the permeability (interaction with a magnetic field) appear to model their behaviour (see also Section 10.2.3). The zero indexed values describe the behaviour of a vacuum, so that " and are relative values.
Microwave processing !
!
D "0 " E !
!
B o H !
!
j E
181 10:6 10:7 10:8
In the most general form all these material parameters, describing the properties of matter, can be complex tensors (with directional dependent behaviour). For practical use with food substances some simplifications are possible: the relative permeability can be set to 1, since food behaves non-magnetically, and the permittivity tensor can be reduced to a complex constant with real ("0 ) and imaginary part ("00 ), which may include the conductivity (see Section 10.2.3). Wave equations and exemplary solutions Starting from Maxwell’s equations, with the simplifications of no charge ( 0) ! and no current density ( j 0), there is an easy way to infer the wave equations for electric and magnetic fields (here only shown for the electric field case). Applying the curl-operator (r) on [10.3] yields [10.9]: !
! @B @
r B r
r E r 10:9 @t @t Using the material equation for the magnetic field [10.7], supposing to be constant and introducing [10.5] into [10.9], this can be transformed to [10.10]: ! ! @ @D r
r E 0 10:10 @t @t !
The last step is to utilise the material equation for the electric field [10.6], ! ! Maxwell’s equation [10.2] and the vector identity r
r X r
r X ! X to get the following well known wave equation [10.11]: !
@2E 10:11 E 0 "0 " 2 0 @t Similarly the corresponding wave equation for the magnetic component can be derived, yielding [10.12]: !
!
@2B 0 10:12 @t2 By comparing the wave equations [10.11] and [10.12] with the standard one, one can infer that in this case the wave velocity is defined by [10.13]: !
B
0 "0 "
1 c0 c p p " 0 "0 "
10:13
In order to illustrate the nature of solutions of [10.11] or [10.12], we consider the case where the electric field has only a component in the z-direction Ez (socalled linearly polarised) and depends only on the x coordinate (so-called plane wave). Additionally the material parameters should be frequency independent. Equation [10.11] then reduces to
182
Thermal technologies in food processing
@ 2 Ez 1 @ 2 E z 0 10:14 c2 @t2 @x2 Often used as solutions also for the more complex case [10.11], [10.12] are time harmonic functions: !
!
!!
!
! E0
!!
E E0 cos
k x
!t;
sin
k x !t; E ! ! !! E < E0 exp i
k x !t
10:15
!
Here k is the wave vector pointing to the direction of propagation with its absolute value defined by !2 10:16 c2 It should be added that the magnetic and the electric field are not independent from each other, since the wave equations cannot completely replace Maxwell’s equations. These lead to further conditions listed in Table 10.1. Both the dispersion (the dependence of c on !) and the absorption are included in this theory, the latter by a complex permittivity and with this a complex wave vector. Another case of absorption occurs when a current ! ! j E is allowed due to finite conductivity in [10.10]. Then [10.11] becomes: !2
k
!
!
@2E @2E 0 "0 " 2 0 10.11a E 0 @t @t Using time-harmonic functions for the electric field as above, [10.11a] reduces to: ! ! 10.11b E !2 0 "0 " i E 0 "0 ! !
This equation shows that a finite conductivity is equivalent to an imaginary term in the permittivity ". Coming back to an exemplary solution, the time harmonic plane wave in the case of an absorbing material, where the permittivity " has an imaginary part " "0 i"00 total, Table 10.1 Further conditions on electric and magnetic fields of an electromagnetic wave Transversality !
!
!
!
k E0 0 k B0 0
Correlation of electric and magnetic fields !
!
!
!
!
k E0 !B0 k B0
!
!0 "0 "E0
Microwave processing
183
Table 10.2 Boundary conditions for the electric and the magnetic fields, which have to be satisfied by the wave solutions at the boundaries Prerequisite
Boundary condition
No surface charge – No surface current – Ideally conducting wall (metallic) Ideally conducting wall (metallic)
Continuity Continuity Continuity Continuity Ejj 0 B? 0
"00total "00
of of of of
D? B? Hjj Ejj
"0 !
10:17
has to be a solution of [10.11c]. @ 2 Ez !2 0 "0
"0 i"00total Ez 0 10.11c @x2 A similar equation can be derived for the magnetic component of the plane wave, leading to a general solution with g, h, m and n constants to satisfy the boundary conditions (see Table 10.2): Ez g expf
ik xg h expf
ik xg Hy m expf
ik xg n expf
ik xg
10:18
One boundary condition, namely the continuity of E|| has to be emphasised, since it can explain the often observed effect of edge or corner overheating. Later it will be shown that the power dissipation in a sample volume is proportional to the squared electric field [10.24]. At edges (corners) the microwaves cannot only intrude from two (three) directions, but also at these volumes electric fields of two (three) polarisations find a parallel surface to intrude continuously, which means without amplitude decrease. Therefore the heat generation there will be very large. The solution [10.18] is an exponentially damped wave, with wave number k and damping constant , where the dependency on " can be derived by solving !2 0 "0
"0 leading to
i"00
ik2
r o "0 "0 k! 2
sr ! "002 1 02 1 "
r o "0 "0 ! 2
sr ! "002 1 02 1 "
and
10:19
10:20
10:21
184
Thermal technologies in food processing
The corresponding electric field penetration depth, the distance in which the electric field is reduced to 1/e is defined by e1 1=. An important consequence of the frequency dependency of is that microwaves of 915 MHz do have an approximately 2.5 times larger penetration depth than waves of 2450 MHz, when similar permittivities at both frequencies are assumed. With the assumption of the excitation and the propagation of a plane wave, that satisfies the boundary conditions, first estimations of the field configurations are possible. This yields, for example, the laws of the geometric optics, which are also valid for microwaves, when a typical object size is much larger than the wavelength. With this approach the particular centre heating of objects of cm-dimensions with convex surfaces (like eggs) can be easily understood, since at the convex surface the microwave ‘rays’ are refracted and focused to the centre. In order to calculate the temperature change within an object by microwave heating, it is important to determine the power density, starting from the electromagnetic field configuration. Since normal food substances are not significantly magnetically different from vacuum ( 1), in most cases the knowledge of the electric field is enough to calculate the heat production, by power dissipation. This power dissipation (per unit volume) pV is determined by ohmic losses which are calculable by 1 ! ! pV <
E j 10:22 2 The value of current density is determined by the conductivity and the electric ! ! field j total E . As in equation [10.17] the equivalence of an imaginary part of the permittivity and a conductivity is shown, here the conductivity also consists of the pure d.c. conductivity and the imaginary part of the permittivity ": total !"0 "00
10:23
The resulting power dissipation can be written in terms of the total conductivity or the total imaginary part of the permittivity, the so-called loss factor: ! ! 1 1 10:24 pv total j E j2 !"0 "00total j E j2 2 2 With this result (the dependence on the squared electric field magnitude), it is clear that the power dissipation penetration depth power is only half the value of the electric field penetration depth el.
10.2.3 Dielectric properties As mentioned above the permittivity, also called the dielectric constant, describes the interaction of an electric field with non- or low-conducting matter. Starting with the simplest case of a static electric field acting on a material, its
Microwave processing
185
!
reaction will be a polarisation P . This is due to a displacement of the charge centres or an orientation of dipoles on account of Coulomb forces. In simple and often used models the polarisation is proportional to the electric field, which ! leads to the constitutive relation for the displacement D , using the permittivity. !
!
!
!
!
!
D "0
E P "0
E E "0 " E
10.6b
When a time dependent electric field is applied, another effect can be observed. Due to the mass and the corresponding inertia of the dipoles to rotate or the charge centres to displace, there will be a difference in phase between the applied field and the resulting polarisation. For time harmonic fields this influence can be taken into account by an imaginary part of the permittivity. Origin of dielectric losses in food substances Important for microwave applications of food substances are two types of origins of the polarisation and the corresponding losses into heat: the ion conductivity and the dipole-orientation. These two types are addressed in more detail here; different mechanisms that are more prominent at other frequencies can be found, for example, in Hasted (1973). As already mentioned in the previous section, the ion conductivity can be included into the dielectric loss factor by: "00total "00 10:17 "0 ! where the second term is determined by the ion conductivity divided by the circular frequency. The ion conductivity itself is dependent on the ion concentrations ni, the ion valences zi and their mobilities i (which are unfortunately only in the first approximation independent of the concentrations): X ni zi i 10:25 i
The frequency dependence of the effect of dipoles (for example, water) can be described by a relaxation behaviour (Debye-relaxation). Starting from the suggestion that the change of polarisation per unit time is equal to the difference of the instantaneous and the steady state value divided by a typical relaxation time , one gets the frequency dependence of the permittivity: "
0 1 1 1 !2 2 ! "00
! "
0 10:26 1 !2 2 For real materials the dielectric behaviour of atomic polarisation at light frequencies has to be taken into account, additionally. The zero-frequency value "(0) is predominantly determined by the dipole moment of the molecule and the molecular concentration. "0
!
186
Thermal technologies in food processing
Fig. 10.1
Frequency dependencies of the dielectric loss factor due to dipolar relaxation, ion conductivity and their superposition.
The typical frequency dependence of both important effects on the loss factor can be seen schematically in Fig. 10.1 together with their temperature dependency. This figure shows that the relaxation peak as well as the contribution of the ion conductivity in "00 are shifted to higher frequency due to the smaller viscosity of the solution and the corresponding higher mobility of ions and dipoles (Feher 1997; Mudgett 1985). For microwave applications the values of the permittivity at 915 MHz and 2.45 GHz are the most interesting ones. Nevertheless, their temperature dependence cannot be predicted without knowing the influences of the temperature dependence on the relaxation frequency, for example. Let us take a look at four examples: 1.
2.
Pure water above the melting point. Since the relaxation frequency of pure water at 20ºC is in the range of 20 GHz, the microwave frequency of 2.45 GHz is smaller than the peak point of "00 . Heating the water will shift the peak to higher frequencies, which will result in a decreasing loss factor value. Salt solution. As above, the heating of a salt solution will shift the relaxation peak to higher frequencies yielding a smaller "00 value; on the other hand,
Microwave processing
3.
4.
187
the higher mobility of the salt ions results in an increased "00 . For a certain temperature and concentration range this could lead to a stable total value of the loss factor. Ethanol. The corresponding relaxation frequency peak for ethanol at 10ºC is below 1 GHz. It shifts towards 2.45 GHz at approximately 50ºC yielding an increased "00 value with temperature in this range. The consequence is an enhanced heating rate with increasing temperature, which can lead to fast overheating of localised spots. This unstable behaviour of a higher "00 with increasing temperature is called ‘thermal runaway’ and also occurs for example in cellulose, and should normally be avoided. Pure water in the melting point range. A similar effect is observed during the melting of water. When the water is frozen, the dipoles do not have the mobility to rotate with the applied electric field. The corresponding "00 value is therefore very small (approximately 0.003). As soon as melting occurs, the high molecular mobility yields an "00 value, which is approximately 3000 times higher than in the frozen state (see Table 10.3). As a consequence an interesting effect to observe is the boiling of water in the presence of water in frozen state, an effect far from equilibrium.
Food material mostly consists of a complex mixture of various ingredients. As already mentioned, water and salts are the most important ingredients, showing the strongest interactions with microwaves. Therefore the dielectric properties are strongly dependent on their concentrations. The interaction of microwaves with components other than ions and dipoles are rather negligible, yielding small values of the permittivity. Nevertheless, only in simple cases of non-interacting mixtures, can model calculations completely describe their dielectric behaviour (Datta et al. 1995; Erle et al. 2000; Persch 1997). In Table 10.3 Dielectric properties of various materials at 2.45 GHz. Apart from the case of ice, the values are valid for room temperature. Corresponding to the variation in natural materials the dielectric property values are given in different accuracy here Material
"0
"00
Water Ice 1 molar NaCl-solution 2 molar NaCl-solution Ethanol 10% Ethanol-solution 10% Sucrose-solution Vegetable oil Beef tissue Cooked ham Mashed potatoes Carrot tissue Apple tissue
78.1 3 74.8 65.6 7.5 71.5 74.5 3.1 50 45 65 71 64
10.4 0.003 21.4 69.1 7.1 13.8 13.1 0.4 15 25 21 18 13
Remarks #
2ºC
(weight %) (weight %) (uncooked) (water content 89.7%) (water content 84.0%)
188
Thermal technologies in food processing
particular, for mixtures of highly interactive molecules, for example different types of dipoles, simple theories predicting permittivities of the mixtures do not succeed. The reason is the existence of molecular clusters of different dipole moments and inertia and therefore a changed relaxation frequency, that lead to completely different dielectric constants. The consequence is, that often only measured values can act as starting values for accurate calculations. Some measured values of different food substances are shown in Table 10.3; further dielectric properties can be found in the literature (e.g. Mudgett 1985; Bengtsson 1971; Datta et al. 1995) or in a www database presenting a collection of physical properties of foods (www.nel.uk/fooddb/). Measurement of dielectric properties of food in the microwave frequency range Among many possible methods for the measurement of dielectric properties in the microwave frequency range (Rost 1978), we choose two well-suited and widely spread methods for food material to be presented here: (a) the openended coaxial line dielectric probe, and (b) resonator methods. The open-ended coaxial line dielectric probe The set-up consists of a network analyser, which is the source and the detector of electromagnetic waves of defined frequency. It is coupled by a coaxial line to an open-ended probe, which has to be in direct contact with the sample to be characterised (a schematic view is shown in Fig. 10.2(a)). The end of the coaxial line, which is defined by the probe itself and the sample, represents a capacity composed of the internal probe capacity and the fringing field capacity. Within this capacity the electromagnetic wave is reflected, whereby the amplitude and the phase of the reflected wave are influenced by the fringing field capacity, which is dependent on the permittivity of the sample material. By changing the frequency of the incident wave, frequency dependent dielectric properties can be measured, and can be used to estimate their temperature dependencies. For homogeneous fluids or samples with flat surfaces, in particular, it is quite simple to use. Disadvantages of the system are that a flat surface of the sample has to be guaranteed to make a direct contact between probe and sample possible. Also the sample should be homogeneous in the sensible depth of the probe, which is in the mm-range, and the values of both the real and the imaginary part of the permittivity must not be very small for good accuracy. Resonator methods Resonator methods (Fig. 10.2(b)) consist of a microwave resonator (a metallic cavity), where at certain frequencies (called resonant frequencies) standing waves can exist. In these frequency ranges resonant curves of the microwave cavity change due to a dielectric filling, which can be detected by a network analyser. Generally by inserting a dielectric material the resonant frequency decreases and the width of the resonant curve increases with increasing "0 or "00 , respectively. For special geometries and symmetries (namely the concentric cylindrical geometry) the governing Maxwell’s equation can be analytically
Microwave processing
189
Fig. 10.2 Schematic view of measuring systems for dielectric properties: (a) Openended coaxial line dielectric probe. (b) Partially filled resonator system.
solved yielding equations for "0 , "00 (Rost 1978; Regier and Schubert 2000). In other cases calibration procedures that use materials with known dielectric properties and correlate them with measured resonant curve changes, may be used (e.g. Bengtsson 1971). The most prominent advantage of the resonator methods is their suitability for more heterogeneous material, since they sample a larger volume and average over inhomogeneties. Also both high and low values of permittivities can be measured with high accuracy but with different kinds of resonators (completely or partially filled cavities). The major disadvantage of the resonator methods is that each resonator can only determine the permittivity at its resonant frequencies, which is additionally dependent on the investigated material. 10.2.4 Microwave sources, waveguides and applicators Magnetrons By far the microwave source most used for industrial and domestic applications (Metaxas (1996) mentions a figure of 98%) is the magnetron tube. Therefore, we limit our discussion here to the description of a magnetron and only from a
190
Thermal technologies in food processing
phenomenological point of view. More detailed descriptions can be found in, for example, Metaxas and Meredith (1983) and Pu¨schner (1966). A magnetron is a vacuum tube with a central electron emitting cathode of highly negative potential (see Fig. 10.3), which is surrounded by a structured anode. The anode structure forms cavities, which are coupled by their fringing fields and have the intended microwave resonant frequency. Due to the high electric d.c. field, the emitted electrons are accelerated radially but are deflected by an orthogonal magnetic d.c. field, yielding a spiral motion. If the electric and the magnetic field strength are chosen appropriately, the resonant cavities take energy from the electrons which can be coupled out by a circular loop antenna in one cavity into a waveguide or a coaxial line. The power output of a magnetron can be controlled by both the tube current and the magnetic field strength. The maximum power is generally limited by the temperature of the anode; practical limits at 2.45 GHz are approximately 1.5 kW and 25 kW for air- or water-cooled anodes, respectively (Roussy and Pearce 1995). Due to their larger size, 915 MHz magnetrons can achieve higher powers per unit. The efficiencies of modern 2.45 GHz magnetrons range at approximately 70% most limited by the magnetic flux of the economic ferrite magnets used (Yokoyama and Yamada 1996), whereas the total efficiency of microwave heating applications often are lower due to unmatched loads. Waveguides For guiding an electromagnetic wave, transmission lines (e.g. coaxial lines) and waveguides can be used. At higher frequencies like microwaves, waveguides have lower losses and are therefore used for power applications. Principally, waveguides are hollow conductors of normally constant cross-section, whereby
Fig. 10.3
Schematic view of the set-up and the function of a magnetron.
Microwave processing
191
rectangular and circular forms are of most practical use. Its size defines a minimum frequency fc (the so-called cut-off frequency) by the solution of the wave equations (eq. [10.11] and [10.12]) and appropriate boundary conditions (Table 10.2), below which waves do not propagate. Within the waveguide the wave may spread out in so-called modes, which define the electromagnetic field distribution within the waveguide. These modes can be split into transversal electric (TE) and transversal magnetic (TM) ones, describing the direction of the electric or the magnetic field, respectively, towards the propagation direction. The most commonly used waveguide is of rectangular cross-section with a width equal to double the height in TE10 mode, which is depicted in Fig. 10.4. Microwave applicators Already the waveguide can be used as an applicator for microwave heating, when the material to be heated is introduced by wall slots and the waveguide is terminated by a matched load. This configuration is then called a travelling wave device, since the location of the field maxima change with time. A radiation through the slots only occurs if wall current lines are cut and the slots exceed a certain dimension, which can be avoided (Roussy and Pearce 1995). More common in the food industrial and domestic field are standing wave devices described in the next section, where the microwaves irradiate by slot arrays (that cut wall currents) or horn antennas (specially formed open ends) of waveguides. One should distinguish between three types of applicators by the type of field configurations: (a) near field applicators, (b) single-mode applicators, and (c) multi-mode applicators. Near field applicators In the case of near field applicators, the microwaves originating from a horn antenna or slot arrays ‘hit’ directly the product to be heated, and are almost completely absorbed by it. The transmitted microwaves have to be transformed into heat in dielectric loads (usually water) behind the transmitted product. This case is very similar to the travelling wave device, since a standing wave cannot develop. Consequently no standing wave pattern can be formed, which can yield a relatively homogeneous electrical field distribution (depending on the mode irradiated from the waveguide) within a plane orthogonal to the direction of propagation of the wave. Single-mode applicators The near field applicators as well as the travelling wave devices work best with materials with high losses. For substances with low dielectric losses, applicators with resonant modes, which enhance the electric field at certain positions, are better suited. The material to be heated should be located at these positions. Single-mode applicators consist generally of a feeding waveguide and a relatively small microwave resonator with dimensions in the range of the wavelength. As in the case of the resonator measurements (Section 10.2.3) a
192
Thermal technologies in food processing
Z X Y
a.
Z X Y
b.
Fig. 10.4 Electric (a) and magnetic (b) field distribution in a rectangular TE10 waveguide. The arrows point to the actual field direction, their size shows the magnitude of the field amplitudes.
standing wave (resonance) exists within the cavity at a certain frequency. The standing wave yields a defined electric field pattern, which can then be used to heat the product. An example of such a system is shown in Fig. 10.5, where a cylindrical TM010 field configuration, with high electric field strength at the centre, is used to heat a cylindrical product that could be transported through tubes (e.g. liquids).
Microwave processing
Fig. 10.5
193
Schematic view of a TM010 resonator, as an example of a single-mode heating device.
The small dimensions of the applicator are needed to avoid different modes from the used one, since the number of modes per frequency range grows very fast with cavity dimensions. It has to be noted that this type of applicator has to be well matched to the load, since the insertion of the dielectric material naturally alters the resonant modes. Multi-mode applicators By increasing the dimensions of the cavity a fast transition from the single mode to the multi-mode applicator occurs, due to the fast growth of mode density with applicator size and the fact that microwave generators like magnetrons do not emit a single frequency but rather a frequency band. In industrial and domestic applications the multi-mode applicators play by far the most important role, since most of the conveyor-belt-tunnel applicators and the microwave ovens at home are of the multi-mode type due to their typical dimensions. Despite the high number of stimulated modes, a non-homogeneous field distribution (constant in time) will develop depending on the cavity and the product geometry and the dielectric properties of the material to be processed. In opposition to the case of the single mode application, normally this inhomogeneous field distribution, which would result in an inhomogeneous heating pattern is not desired. Possible remedies are either moving the product (conveyor belt, turn table) or changing the field configuration by varying cavity geometries (e.g. mode stirrer), which are explained in more detail in the corresponding application sections.
194
10.3
Thermal technologies in food processing
Microwave applications
A detailed and more extensive description of many microwave uses in the food sector can be found in the monograph by Decareau (1985). Here only the main purposes are outlined together with some examples, that illustrate the large spectrum of different applications with the various demands. 10.3.1 Household ovens and product engineering The principal set-up of a household microwave oven consists of a magnetron tube, which is coupled by a waveguide and an aperture to a commonly rectangular cavity. Due to reflections a standing wave develops, which leads to an inhomogeneous heating pattern even in a homogeneous sample. Three possible remedies can effectively reduce this undesirable heating behaviour. The simplest method is to use low microwave power mostly achieved by pulsing the microwave irradiation. The consequence is, that the relatively slow heat conduction mechanism levels the temperature gradients within the product. Another often used system is the turntable of microwave ovens. The movement of the dielectric material enhances the power uniformity in two ways: by averaging different electric field strength areas and by changing the electric field pattern due to the varied geometrical set-up, which yields different field configurations. Concurrent to this turntable, mode stirrers can be used, which are placed just before the aperture of the waveguide to the cavity. This mode stirrer also changes the geometrical set-up of the complete cavity and therefore yields time dependent field configurations, leading to more even product heating. In most cases these methods have to be combined to get sufficient results. Another way to get a more uniform or a desired temperature distribution is to change product properties (e.g. the ingredients, the geometrical set-up of the product, the packaging, etc.), instead of the processing (microwave heating) method. This way of product enhancement is called product engineering or product formulation. In the literature and in various patents possible approaches have been developed; a detailed overview is presented in Decareau (1992). Here only some illustrative examples are described. In order to get similar heating rates and temperatures in products with different dielectric or thermodynamic properties, containers covered by a metal foil with appropriate apertures have been proposed. A different approach is changing the location of different foods on the plate. Alternatively, the incorporation of susceptors into the packaging is possible. These susceptors consist of material with high losses; consequently they can reach high temperatures which they can transfer to the desired location by irradiation or conduction. With these materials even surface browning is possible in microwave ovens, which is normally prevented by too low surface temperatures due to evaporation. The browning and crisping effects are also the objective of combination ovens, which mostly use an additional conventional (resistive) heating
Microwave processing
195
equipment within the microwave oven. Recently, a method called jet impingement, where very hot (up to 500ºC) and very fast air jets flow contrary to the primary microwave ‘fronts’, is introduced in the microwave market. The resulting problem of complicated cooking power programs should be overcome with Internet access, where these recipes can be downloaded directly to the oven (Franke and Pool 2000). 10.3.2 Industrial ovens Industrial applications mostly need continuous processing due to the desired high throughputs. Therefore continuous microwave applicators had to be developed, starting in 1952 with the first conveyor belt oven patent (Spencer 1952). Nevertheless, due to the lack of high power microwave generators the start of its industrial use was nearly ten years later. Today’s industrial ovens (a more complete overview can be found in Decareau (1985)) may be differentiated into two groups by the number and power of microwave sources: high power single magnetron and low power multi-magnetron devices. Whereas for a single mode unit only a single source is possible, in all other systems (multi-mode, near field or travelling wave system) the microwave energy can be irradiated by one high power magnetron or several low power magnetrons. An important hurdle for all continuous ovens is the avoiding of leakage radiation through the product in- and outlet. The leakage radiation is limited by law to 5 mW/cm2 at any accessible place. For fluids or granular products with small dimensions (cm-range), this value can be guaranteed by in- and outlet sizes together with the absorption in the entering product, sometimes with additional dielectric loads just in front of the openings. Especially in the case of larger product pieces, inlet and outlet gates, which completely close the microwave application device, have to be used. A conveyor belt oven with its alternative power sources and openings is shown schematically in Fig. 10.6. 10.3.3 Industrial processes In industry a variety of microwave applications have been and are still used. For a more complete survey, illustrating many real commercial processes, see Decareau (1985), Metaxas and Meredith (1983), Metaxas (1996), Roussy and Pearce (1995), Buffler (1993) and various authors (1986). The use of microwave energy in food processing can be classified into six unit operations: (re)heating, baking and (pre)cooking, tempering, blanching, pasteurisation and sterilisation, and dehydration. Although their objectives differ, these aims are established by similar means: an increase in temperature. Nevertheless, for each special use (different from pure microwave heating), different advantages and disadvantages have to be taken into account. These are presented in the next sections together with some examples of real industrial applications.
196
Thermal technologies in food processing
Fig. 10.6 Schematic view of a continuous conveyor belt microwave tunnel, with alternative microwave energy inlets (above) and various product in- and outlets (at bottom).
Baking and cooking In the process of baking bread, cakes, pastry, etc., microwaves have been used and studied by several authors; references can be found in Rosenberg and Bo¨gl (1987a). The major task of the microwaves is to accelerate the baking, leading to an enhanced throughput with negligible additional space required for microwave power generators. Often combined with conventional or infrared surface baking, microwave use avoids the remedy of lack of crust formation and surface browning. With the fast combined process also different flour can be used with high -amylase and low protein content (for example from European soft wheat). In contrast to conventional baking the microwave heating inactivates this enzyme fast enough (due to a fast and uniform temperature rise in the whole product) to prevent the starch from extensive breakdown, and develops sufficient CO2 and steam to produce a high porous good (Decareau 1986). One difficulty in the microwave baking process was to find a microwavable baking pan, that is sufficiently heat resistant and not too expensive for
Microwave processing
197
commercial use. But already by 1981 and 1982 patents were issued overcoming this problem using metal baking pans in microwave ovens (Schiffmann et al. 1981, Schiffmann 1982). Today, the main use of microwaves in the baking industry is microwave finishing. While the conventional oven technique is used at the beginning with high moisture dough, microwaves improve the end baking, where the low heat conductivity would lead to considerably higher baking times. Microwaves can also be applied as part of a parallel process. One example is the frying of donuts with microwave assistance (Schiffmann 1986), resulting in a shorter frying time and a lower fat uptake. Also the (pre)cooking process can be accelerated with the help of microwaves, as has been established for (pre)cooking of poultry (Decareau 1986), meat patty and bacon. A convective air flow removes the surface water using microwaves as the main energy source, thus rendering the fat and coagulating the proteins by an increased temperature. Also this process yields a valuable by-product namely rendered fat of high quality, which is used as food flavorant (Schiffmann 1986). Tempering Another widely used industrial microwave application is the tempering of foods (Metaxas (1996) mentions a figure of 250 units all over the world). Tempering is defined as the thermal treatment of frozen foods to raise the temperature from below 18ºC to temperatures just below the melting point of ice (approximately 2ºC). At these temperatures the mechanical product properties are better suited for further machining operations (e.g. cutting or milling). The time for conventional tempering strongly depends on the low thermal conductivity of the frozen product and can be in the order of days for larger food pieces such as blocks of butter, fish, fruits or meat. Due to the long time, the conventional process needs large storage rooms, there is a not-negligible drip loss and the danger of microbial growth. By using microwaves (mostly with 915 MHz due to their larger penetration depth) the tempering time can be reduced to minutes or hours (Edgar 1986) and the required space is diminished to one sixth of the conventional system (Metaxas 1996). Another advantage is the possibility to use the microwaves at low air temperatures, thus reducing or even stopping microbial growth. Of very high importance is the heating uniformity and the control of the end temperature to avoid localised melting, which would be coupled to a thermal runaway effect. The best homogeneity in this application is reached in a multisource multi-mode cavity, equipped with mode stirrers (Metaxas 1996). Drying The main cause for the application of microwaves in drying is the acceleration of the processes, which are (without using microwaves) limited by low thermal conductivities, especially in products of low moisture content. Correspondingly sensory and nutritional damage caused by long drying times or high surface temperatures can be prevented. Another advantage is the possible avoidance of
198
Thermal technologies in food processing
case hardening, due to a more homogeneous drying, without large moisture gradients. Generally, microwave drying can be subdivided into two cases, the drying at atmospheric pressure and that with applied vacuum conditions. Until now, more common in the food industry are combined microwave-air-dryers, that could again be classified into a serial or a parallel combination of both methods. In the serial process, mostly the microwaves are used to finish partly dried food or food of low moisture content, where an intrinsic levelling effect is advantageous: the loss factor is often dominated by the water concentration, therefore the places of high water content transform more of the microwave energy into heat and are selectively dried. Well studied and still applied examples for a serial hot air and microwave dehydration are pasta drying (Decareau 1985) and the production of dried onions (Metaxas and Meredith 1983). Intermittently successful in the 1960s and 1970s was the finish drying of potato chips, but the process did not survive due to microwave-intrinsic and extrinsic reasons (O’Meara 1977). The combination of microwave and vacuum drying also has a certain potential. Though the microwave assisted freeze drying is well studied, as can be read in detail in Sunderland (1980), up to now practically no commercial industrial application can be found, due to high costs and a small market for freeze dried food products (Knutson et al. 1987). It seems that microwave vacuum drying with pressures above the triple point of water has more commercial potential. Clearly, the benefit of using microwave energy is overcoming the disadvantage of very high heat transfer and conduction resistances, leading to higher drying rates. These high drying rates correspond also to the retention of water insoluble aromas (Erle 2000) and to less shrinkage. The use of vacuum pressures is very favourable for high quality food substances, since the reduced pressure limits the product temperatures to lower values, as long as a certain amount of free water is present. This enables the retention of temperature sensitive substances like vitamins, colours, etc. Commercial applications of microwave vacuum dehydration are the concentration or even powder production of fruit juices and drying of grains in short times without germination (Decareau 1985). A relatively new and successful combination of pre-air-intermittent microwave vacuum (called puffing) and post-air-drying is predominantly used to produce dried fruits and vegetables, with improved rehydration properties (Ra¨uber 1998). By the conventional pre-drying due to case hardening the form can be stabilised, the microwave vacuum process opens the cell structures (puffing) due to the fast vaporisation and an open pore structure is generated. The consecutive post-drying reduces the water content to the required moisture. Pasteurisation and sterilisation Since microwave energy can heat many foods (containing water or salts) effectively and fast, its use for pasteurisation and sterilisation has also been intensively studied. Many references can be found in the review by Rosenberg and Bo¨gl (1987b).
Microwave processing
199
In early work, beside the thermal heating effect on microorganisms also nonthermal effects seemed to be found. Physically a non-thermal effect on molecules is very improbable, as becomes clear when the quantum energy of photons of microwaves, of a thermal radiator and the energy of molecular bonds are compared. The quantum energy of a photon of f 2:45 GHz is defined by E hf 1 10 5 eV, the typical energy of a photon radiated from a body of 25ºC 298K equals E kT 0:26 eV and the energy of molecule bonds are in the eV range. Since the collection of energy with time for bound electrons is forbidden by quantum mechanics, only multi-photon processes, which are very unlikely, could yield chemical changes. More likely is the induction of voltages and currents within living cell material, where eventual consequences are still in discussion (Sienkiewicz 1998). From the practical point of view, the nonthermal effects, found in early work could either not be reproduced or could be explained later by the influence of inhomogeneous temperature distributions. No longer in doubt are the thermal effects of microwaves, which can be used for pasteurisation and sterilisation. Academic and industrial approaches to microwave pasteurisation or sterilisation cover the application for prepacked food like yogurt or pouch-packed meals as well as the continuous pasteurisation of fluids like milk (Decareau 1985; Rosenberg and Bo¨gl 1987b). For the packed food systems, conveyor belt systems were intended (e.g. Harlfinger 1992). For continuous fluid pasteurisation or sterilisation tubes intersecting waveguides or small resonators were developed (Sale 1976). Whereas the pasteurisation process can take place at atmospheric pressures, in the case of sterilisation only temperatures of more than 100ºC may be used, in order to achieve satisfactory short sterilisation times and to maintain high product quality. For products which contain free water, like many food products, the reachable temperature at atmospheric pressure is limited to the boiling point at around 100ºC. Therefore, the pressure during the sterilisation process has to be increased to overpressure values. The consequence is the need for special compression and decompressing systems, e.g. sliding gates, that have to be connected to the microwave heater. Advantages of using microwaves in microorganism deactivation are the possibly high and homogeneous heating rates, also in solid foods (heat generation within the food) and the corresponding short process times, which can yield a very high quality. For both processes it is extraordinarily important to know or even to control the lowest temperatures within the product, where the microoganism destruction has the slowest rate. Since both calculation and measurement of temperature distributions are still very difficult (see Section 10.4), this is one reason that up to now microwave pasteurisation and sterilisation can be found very seldom in industrial use, and then only for batch sterilisation operations.
10.4
Modelling and verification
In the beginning of microwave processing the product and process development was essentially a trial-and-error procedure. Due to the lack of powerful
200
Thermal technologies in food processing
computers it was nearly impossible to calculate realistic temperature or even electromagnetic field distributions within microwave applicators when products were present. The reason for this is the number of coupled partial differential equations, describing the physical problems of electromagnetism, heat and mass transfer, which have to be solved in a parallel manner. Meanwhile for the separated problems, numerical software packages are available and there is also progress in the solution of the coupled problem, which best describes real processes. The governing equations of electromagnetism, Maxwell’s equations, have already been described in Section 10.2.2, together with the wave solutions for the simple one-dimensional plane wave example. The resultant exponentially damped wave within a material with dielectric losses have also often been used for the three-dimensional case, in order to simply estimate the power distribution within products. More sophisticated approaches calculating more realistic solutions are presented later in this chapter, but first the governing equations for heat and mass transfer and the coupling to electromagnetism are introduced. Starting from the continuity equation, the thermal energy equation and Fick’s law, a general equation for heat transfer can be described by: X @T cp r
k rT r qR hi Ii Qem 10.26a @t i While the left side of this equation is well known from the traditional heat conduction equation, the terms on the right side have to be added for heat transfer by radiation and by a mass sink or source (e.g. phase change of water or due to diffusion or convection) and for the heat source by the dielectric losses, respectively (Metaxas 1996). If the product consists of a solid but moist material, the radiative term has only to be taken into account at surfaces to gaseous materials yielding additional boundary conditions and the mass sink and source can be replaced by the moisture content changing: @T @M r
k rT "v hevap Qem 10:26b @t @t Especially for the case of drying, this equation is coupled to the mass transfer equation, here written for the moisture content M: cp
@M r
M rM T rT p rp 10:27 @t This equation becomes even more difficult when a product porosity has to be taken into account with its capillary forces and the possible shrinkage. The equation describing the behaviour of the total pressure can also be found in Metaxas (1996): @p p p @t
"v @Ml : ca @t
10:28
Microwave processing
201
The above equations and the electromagnetic equations are coupled in two ways: explicitly by the values of the temperature, the moisture content, the pressure and the electromagnetic heat generation (eq. [10.24]), but also implicitly by the temperature and moisture dependency of the material properties. In general, these material properties also have a directional dependency, which has to be expressed by tensor properties. This is especially valid for inhomogeneous natural material; as a well known example the variation of mass transfer coefficient along and orthogonal to the predominant fibre orientation can be mentioned. Additionally these equations are already simplified by the assumption that there would be no volume change of the solid material structure. Indeed volume shrinkage can often be observed during heating and especially drying. At the product surface, boundary conditions come into operation to take into account the external heat, mass and pressure transfer, respectively. Although there is a fast development of numerical calculation power, a complete calculation could not be done without some simplifications, up to now. These simplifications have to be established, neglecting some of the abovementioned dependencies. For the case of pure electromagnetics, commercial numerical software packages are available (a comparison of their potential for microwave heating is addressed by Yakovlev (2000), but is not yet finished), but also some home-built software codes are described in the literature. Most of them originate from the telecommunications area and are not perfectly suited for microwave heating applications, with their special demands. General to all numerical techniques is the discretisation of the partial differential equations or their corresponding integral equations together with the suitable boundary conditions on a calculation grid. In practical use most common are the method of finite differences in time domain (FDTD), the finite integration method (FIM), the finite element method (FEM), the method of moments (MOM) and the transmission line matrix method (TLM), and also methods using optical raytracing codes. Again, we have to refer to special publications (Metaxas 1996; Lorenson 1990), for a more detailed overview. Some approaches are mentioned here, together with the articles, where the interested reader can find more information. For short times and high microwave power densities, the heat transfer, which is in this case much slower than the microwave heat generation itself, can be neglected. While the one-dimensional example has already been addressed analytically in Section 10.2.2, which has educational value, of course, for more realistic problems two or three dimensions are needed. The temperature rise in a defined volume is then directly proportional to the microwave heat generation rate, which can be inferred from the effective electric field value and the dielectric loss factor (eq. [10.24]). Relatively new results using this approximation can be found for example in Fu and Metaxas (1994), Liu et al. (1994), Sundberg et al. (1998), Dibben and Metaxas (1994) and Zhao and Turner (1997). In only a few papers is electromagnetism already coupled to a thermal model; examples including heat conduction can be found in Torres and Jecko (1997) and Ma et al. (1995). In Haala and Wiesbeck (2000) additional to heat
202
Thermal technologies in food processing
conduction, heat transport by radiation is addressed by a raytracing algorithm. Since for most food applications the temperatures are more moderate in comparison to ceramics sintering, where the latter software code originates, this radiation seems to be more negligible than heat transport by convection or evaporation. While in a recent publication (Zhang and Datta 2000) the heat transfer from the product surface by free convection in a microwave oven is addressed by the corresponding boundary condition, both ways of heat and mass transport within the product are taken into account only in more phenomenological studies: either the microwave heating phenomenon is simplified by using Lambert’s or Mie’s equations for special geometries (Lian et al. 1997; Jun et al. 1999), or the heat and mass transport is modelled by the use of non-local balances (Roques and Zagrouba 1997; Erle 2000; Zhou et al. 1995). Generally, it has to be concluded that the published model calculations are limited mostly to special cases or to very similar ones, where they have been applied successfully. Therefore, after model calculations their verification is also a very important task, and in the case of microwave applications this is not at all simple. The electromagnetic fields are not easily measurable without changing them by the measurement procedure itself. The same has to be said for the measurement of temperature distributions. A relatively old bibliography of different temperature indication methods in microwave ovens can be found in Ringle and Donaldson (1975). Without enormous efforts normal thermocouples could not be used in microwave devices, since their metallic wires at least change the field distribution. Moreover only a poor local resolution would be achievable. While the first problem can be overcome by fibre optic thermometers, the reachable local resolution is often not sufficient, either. Besides, the fibre optic sensors are rather delicate and expensive in comparison to conventional thermocouples. A different approach to determine temperature distributions in microwave devices is the use of model substances that change their properties when a certain temperature is reached. One published method uses the colour change of the model substance (Risman et al. 1993), another one a coagulation effect (Wilhelm and Satterlee 1971). Care has to be taken to match the other important properties of the model to the real product (e.g. the dielectric and thermal properties). In order to determine surface temperatures, infrared photographs and also liquid crystal foils (Gru¨newald and Rudolf 1981) or thermofax paper (Feher 1997) are established methods. By the development of solid phantom foods that can be taken apart practically instantly along pre-cut planes, this disadvantage could be weakened. All the above-mentioned methods can give useful hints of the temperature distribution, but cannot really prove the calculated temperature values within the product. For this task nowadays the sophisticated method of nuclear magnetic resonance imaging can be used, that can measure three-dimensional temperature distributions without destruction of the product (Nott et al. 1999). The disadvantages of this technique are its huge costs and that inline measurements in an industrial scale process are up to now practically impossible.
Microwave processing
10.5
203
Summary and outlook
Microwave ovens are commonplace in households and are established there as devices of everyday use. Their primary function is still the reheating of previously cooked or prepared meals. The relatively new combination of microwaves with other (e.g. conventional, infrared or air jet) heating systems should enhance their potential for a complete cooking device, that could replace conventional ovens. Unfortunately, in industry the distribution of microwave processes is still far away from such high numbers. Only a relatively low number of microwave applications can be found in actual industrial production, compared with their indisputable high potential. These successful microwave applications range over a great spectrum of all thermal food processes. The most prominent advantages of microwave heating are the reachable acceleration and time savings and the possible volume instead of surface heating. Reasons mentioned for the failure of industrial microwave applications range from high energy costs, which have to be counterbalanced by higher product qualities, over the conservatism of the food industry and relatively low research budgets, to the lack of microwave engineering knowledge and of complete microwave heating models and their calculation facilities. The latter disadvantage has been partly overcome by the exponentially growing calculating power which makes it possible to compute more and more realistic models by numerical methods. Very important for the task of realistic calculations is the determination of dielectric properties of food substances by experiments and theoretical approaches. Nevertheless in order to estimate results of microwave heating applications and to check roughly the numerical results, knowledge of simple solutions of the one-dimensional wave propagation like the exponentially damped wave is of practical (and also educational) relevance. But still the best test for numerical calculations are experiments, which yield the real temperature distributions within the product, which is really important especially in pasteurisation and sterilisation applications. While more conventional temperature probe systems, like fibre optic probes, liquid crystal foils or infrared photographs only give a kind of incomplete information about the temperature distribution within the whole sample, probably magnetic resonance imaging has the potential to give very useful information about the heating patterns. Hopefully, this together with the enormous calculation and modelling power will give the microwave technique an additional boost to become more widespread in industrial food production. The breakthrough of microwave technology in the food industry due to its high potential has been predicted many times before, but it has been delayed every time up to now. That is why we are cautious in predicting the future of microwaves in industrial use. However, we think that the potential of microwave technology in the food industry is far from being exhausted.
204
Thermal technologies in food processing
10.6
References
(1998), ‘1998 Microwave Oven Sales Comparison’, Microwave World, 19 (2), 8. BENGTSSON, N E (1971), ‘Dielectric Properties of Food at 3 GHz as Determined by a Cavity Perturbation Technique’, Journal of Microwave Power, 6 (2), 101–23. BUFFLER CH R (1993), Microwave Cooking and Processing: Engineering Fundamentals for the Food Scientist, New York, AVI Book. DATTA A K, SUN E, SOLIS A (1995), Food Dielectrical Property Data and their Composition Based Prediction, Engineering Properties of Foods. New York, Marcel Dekker Inc. DECAREAU R V (1985), Microwaves in the Food Processing Industry. Orlando, Academic Press Inc. DECAREAU R V (1986), ‘Microwave Food Processing Equipment Throughout the World’, Food Technology, June 1986, 99–105. DECAREAU R V (1992), Microwave Foods: New Product Development. Trumball, Food & Nutrition Press Inc. DEHNE L I (1999), ‘Bibliography on Microwave Heating of Food’, Bundesinstitut fu¨r gesundheitlichen Verbraucherschutz und Veterina¨rmedizin, 04/1999. DIBBEN D C, METAXAS A C (1994), ‘Finite-Element Time Domain Analysis of Multimode Application Using Edge Elements’, Journal of Microwave Power & Electromagnetic Energy, 29 (4), 242–51. EDGAR R (1986), ‘The Economics of Microwave Processing in the Food Industry’, Food Technology, June 1986, 106–12. ERLE U (2000), ‘Untersuchungen zur Mikrowellen-Vakuumtrocknung von Lebensmitteln’, PhD-Thesis, Universita¨t Karlsruhe. ERLE U, REGIER M, PERSCH CH, SCHUBERT H (2000), ‘Dielectric Properties of Emulsions and Suspensions: Mixture Equation and Measurement Comparisons’, Journal of Microwave Power & Electromagnetic Energy, 35 (1), 25–33. FEHER L E (1997), Simulationsrechnungen zur verfahrenstechnischen Anwendung von Millimeterwellen fu¨r die industrielle Materialprozesstechnik, Wissenschaftliche Berichte FZ KA 5885, Karlsruhe, Forschungszentrum Karlsruhe. FRANKE S, POOL J (2000), ‘Review of Advances in Microwave & Combination Cooking in 1999’, 35th Microwave Power Symposium, Montreal, July 2000. FU W, METAXAS A C (1994), ‘Numerical Prediction of Three-Dimensional Power Density Distribution in a Multi Mode Cavity’, Journal of Microwave Power & Electromagnetic Energy, 29 (2), 67–75. GOLDBLITH S A, DECAREAU R V (1973), An Annotated Bibliography on Microwaves, their Properties, Production and Applications to Food Processing. Cambridge, The Massachusetts Institute of Technology Press. ¨ NEWALD TH, RUDOLF M (1981), ‘Messung der Temperatur und der GRU ANON.
Microwave processing
205
Temperaturverteilung im Mikrowellenfeld’, Zeitschrift fu¨r die Lebensmittelwirtschaft, 32 (3), 85–8. HAALA J, WIESBECK W (2000), ‘Simulation of Microwave, Conventional and Hybrid Ovens Using a New Thermal Modelling Technique’, Journal of Microwave Power & Electromagnetic Energy, 35 (1), 34–43. HASTED J B (1973), Aqueous Dielectrics. London, Chapman and Hall. HARLFINGER L (1992), ‘Microwave Sterilization’, Food Technology, Dec, 57–61. JUN W, JING-PING Z, JIAN-PING W, NAI-ZHANG X (1999), ‘Modelling Simultaneous Heat and Mass Transfer for Microwave Drying of Apple’, Drying Technology, 17 (9), 1927–34. KNUTSON K M, MARTH E H, WAGNER M K (1987), ‘Microwave Heating of Food’, Lebensmittel-Wissenschaft und -Technologie, 20, 101–10. LIAN G, HARRIS C H, EVANS R, WARBOYS M (1997), ‘Coupled Heat and Moisture Transfer During Microwave Vacuum Drying’, Journal of Microwave Power & Electromagnetic Energy, 32 (1), 34–44. LIU F, TURNER I, BIALKOWSKI M (1994), ‘A Finite Difference Time Domain Simulation of Power Density Distribution in a Dielectric Loaded Microwave Cavity’, Journal of Microwave Power & Electromagnetic Energy, 29 (3), 138–48. LORENSON C (1990), ‘The Why’s and How’s of Mathematical Modelling for Microwave Heating’, Microwave World, 11 (1), 14–23. MA L, PAUL D L, POTHECARY N, RAILTON C, BOWS J, BARRAT L, MULLIN J, SIMONS D (1995), ‘Experimental Validation of a Combined Electromagnetic and Thermal FDTD Model of a Microwave Heating Process’, IEEE Transactions on Microwave Theory and Techniques, 43 (11), 2565–71. METAXAS A C (1996), Foundations of Electroheat. Chichester, John Wiley & Sons. METAXAS A C, MEREDITH R J (1983), Industrial Microwave Heating. London, Polu Pelegrinus Ltd. MUDGETT R E (1985), Dielectric Properties of Foods, Microwaves in the Food Processing Industry. Orlando, Academic Press Inc., 15–56. NOTT K P, HALL L D, BOWS J R, HALE M, PATRICK M L (1999), ‘Three-Dimensional MRI Mapping of Microwave Induced Heating Patterns’, International Journal of Food Science and Technology, 34, 305–15. O’MEARA J P (1973), ‘Why Did they Fail? A Backward Look at Microwave Applications in the Food Industry’, Journal of Microwave Power & Electromagnetic Energy, 8 (2), 167–72. PERSCH CH (1997), ‘Messung von Dielektrizita¨ tskonstanten im Bereich von 0,2 bis 6 GHz und deren Bedeutung fu¨r die Mikrowellenerwa¨rmung von Lebensmitteln’, PhD-Thesis, Universita¨t Karlsruhe. ¨ SCHNER H A (1966), Heating with Microwaves. Berlin, Philips Technical PU Library. ¨ UBER H (1998), ‘Instant-Gemu RA ¨ se aus dem o¨stlichen Dreila¨ndereck’, Gemu¨se, 10’98. REGIER M, SCHUBERT H (2000), ‘Dielectric Properties at Microwave Frequencies
206
Thermal technologies in food processing
Studied in Partially Filled Cylindrical TE011 Cavities’, Journal of Microwave Power & Electromagnetic Energy, 354 (1), 25–33. REYNOLDS L (1989), ‘The History of the Microwave Oven’, Microwave World, 10 (5), 11–15. RINGLE E CH, DONALDSON D B (1975), ‘Measuring Electric Field Distribution in a Microwave Oven’, Food Technology, 29 (12), 46–54. RISMAN P O, OHLSSON T, LINGNERT H (1993), ‘Model Substances and their Use in Microwave Heating Studies’, SIK Report, 588, 1–10. ROQUES M A, ZAGROUBA F (1997), ‘Analysis of Heat and Mass Fluxes During Microwave Drying’, Drying Technology, 15 (9), 2113–27. ¨ GL W (1987a), ‘Microwave Thawing, Drying and Baking in the ROSENBERG U, BO Food Industry’, Food Technology, June, 85–91. ¨ GL W (1987b), ‘Microwave Pasteurization, Sterilization, ROSENBERG U, BO Blanching and Pest Control in the Food Industry’, Food Technology, June, 92–121. ROST A (1978), Messung dielektrischer Stoffeigenschaften. Braunschweig, Vieweg Verlag. ROUSSY G, PEARCE J A (1995), Foundations and Industrial Applications of Microwaves and Radio Frequency Fields. Chichester, John Wiley & Sons. SALE A J H (1976), ‘A Review of Microwave for Food Processing’, Journal of Food Technology, 11, 319–29. SCHIFFMANN R F (1982), Method of Baking Firm Bread, U.S. Patent 4,318,931. SCHIFFMANN R F (1986), ‘Food Product Development for Microwave Processing’, Food Technology, June. SCHIFFMANN R F, MIRMAN A H, GRILLO R J (1981), Microwave Proofing and Baking Bread Utilizing Metal Pans, U.S. Patent 4,271,203. SIENKIEWICZ Z (1998), ‘Biological Effects of Electromagnetic Fields’, Power Engineering Journal, 12 (3), 131–9. SPENCER P (1952), Means for Treating Foodstuffs, U.S. Patent 2,605,383, 605,383. SUNDBERG M, KILDAL P S, OHLSSON T (1998), ‘Moment Method Analysis of a Microwave Tunnel Oven’, Journal of Microwave Power & Electromagnetic Energy, 33 (1), 36–48. SUNDERLAND J E (1980), ‘Microwave Freeze Drying’, Journal of Food Process Engineering, 4, 195–212. TORRES F, JECKO B (1997), ‘Complete FDTD Analysis of Microwave Heating Processes in Frequency-Dependent and Temperature-Dependent Media’, IEEE Transactions on Microwave Theory and Techniques, 45 (1), 108–17. VARIOUS AUTHORS (1986), ‘Microwave Food Processing: an Industrial Overview’, Food Technology, June. WILHELM M S, SATTERLEE L D (1971), ‘A 3-Dimensional Method for Mapping Microwave Ovens’, Microwave Energy Application Newsletter, 4 (5), 3. YAKOVLEV V (2000), ‘Comparing Commercial Electromagnetic Modelling Software for the Microwave Power Industry’, 35th Microwave Power Symposium, Montreal, July 2000.
Microwave processing YOKOYAMA R, YAMADA A
207
(1996), ‘Development Status of Magnetrons for Microwave Ovens’, Proceedings of 31st Microwave Power Symposium, 132–5. ZHANG H, DATTA A K (2000), ‘Coupled Electromagnetic and Thermal Modeling of Microwave Oven Heating of Foods’, Journal of Microwave Power & Electromagnetic Energy, 35 (2), 71–85. ZHAO H, TURNER I (1997), ‘A Generalised Finite-Volume Time-Domain Algorithm for Microwave Heating Problems on Arbitrary Irregular Grids’, 13th Annual Review of Progress in Applied Computational Electromagnetics, Monterey, 352–9. ZHOU L, PURI V M, ANANTHESWARAN R C, YEH G (1995), ‘Finite Element Modeling of Heat and Mass Transfer in Food Materials During Microwave Heating – Model Development and Validation’, Journal of Food Engineering, 25, 509–29.
11 Infrared heating C. Skjo¨ldebrand, ABB Automation Systems (formerly Swedish Institute of Food Research (SIK)), Tumba
11.1
Introduction; principle and uses
Sir William Herschel discovered in the 1800s infrared – or heat-radiation when he was attempting to determine the part of the visible spectrum with the minimum associated heat in connection with the astronomical observations he was making. In 1847 A.H.L. Fizeau and J.B.L. Foucault showed that infrared radiation has the same properties as visible light. It was being reflected, refracted and capable of forming an interference pattern (Encyclopedia Internet 2000). There are many applications of infrared radiation. A number of these are analogous to similar uses of visible light. Thus, the spectrum of a substance in the infrared range can be used in chemical analysis much as the visible spectrum is used. Radiation at discrete wavelengths in the infrared range is characteristic of many molecules. The temperature of a distant object can also be determined by analysis of the infrared radiation from the object. Medical uses of infrared radiation range from the simple heat lamp to the technique of thermal imaging, or thermography. It has also been used for drying of dye and lacquer for cars, drying of glue for wallpaper, drying of paper in a paper machine, drying of dye to plastic details, shrinkage of plastics, activation of glue in the plastic industry, etc. The electromagnetic spectra within infrared wavelengths can be divided into three parts: long waves (4 m–1 mm), medium waves (2–4 m) and short waves (0.7–2 m). The short waves appear when temperatures are above 1000ºC; the long waves appear below 400ºC and medium waves between these temperatures. The electromagnetic spectrum is shown in Fig. 11.1. For food the technique has been used in many applications, as the long waves are one of the main heat transfer mechanisms in ordinary ovens or other heating
Infrared heating
Fig. 11.1
209
The electromagnetic spectrum (from Anon. 1974).
equipment. However, using short waves is new to the food industry. In the 1950s in the Soviet Union Lykow and others reported in general terms the results of their theoretical and experimental studies of infrared drying (Ginzburg 1969). In the 1960s in East Germany, Jubitz carried out substantial work on infrared heating and in France De´ribe´re´ and Leconte did some work on different applications on infrared irradiation in various industries. During this time Pavlov in the Soviet Union carried out a lot of work on infrared heating and food. On an industrial scale long wave radiation was already being used in the United States during the 1950s in many industrial food processes. During the early 1970s there were many discussions about finding new methods for industrial frying/cooking of meat products (Skjo¨ldebrand 1986). Deep fat frying, the process most often used in industrial frying, was criticised because of the fat and flavour exchange and surface appearance. Also environmental and nutritional aspects had to be considered. The consumer also wanted products more like the ones cooked at home. One of the new techniques that was discussed was near infrared heating (NIR) or short-wave infrared heating. This technique is used in the car industry for drying of coatings. Also the paper and textile industry use it for drying. Thus, like many other processes in the food industry, infrared heating was transferred from other industries. Why, then, has short wave infrared radiation not been used before? The answer is that there was a lack of knowledge about many of the factors concerning this process. On the one hand the radiators, the reflectors and the different systems for cassettes were developed during the 1960s; on the other hand there was not very much knowledge about the optical properties of the foodstuffs and how these develop during processing. Other problems then were braking of the radiators and cleaning of the equipment.
210
Thermal technologies in food processing
During the 1970s and 1980s most of the research work on food in Sweden was carried out at SIK the Swedish Institute for Food and Biotechnology ¨ sterstro¨m 1979; Skjo¨ldebrand 1986; Skjo¨ldebrand and (Dagerskog and O Andersson 1989). New knowledge was gained after a lot of studies. In this chapter the following topics of infrared heating for food processing will be covered: • • • • • •
theories and infrared properties technologies infrared heating equipment applications and case studies future trends sources of information and advice.
11.2
Theories and infrared properties
The basic concepts of infrared radiation are high heat transfer capacity, heat penetration directly into the product, fast process control and no heating of surrounding air. These qualities indicate that infrared radiation should be the ideal source of energy for heating purposes. The penetration characteristics are such that a suitable balance between the surface and body heating can be achieved, which is necessary for an optimal heating result. The following factors are critical to control to get an optimal heating result (Ginzburg 1969; Hallstro¨m et al. 1988): • • • •
radiator temperature radiator efficiency infrared reflection/absorption properties infrared penetration properties.
11.2.1 Radiators The main component of IR equipment for heating is the radiator, of which there are various types and shapes. They may be divided into the following main groups • gas heated radiators (long waves) • electrically heated radiators – tubular/flat metallic heaters (long waves) – ceramic heaters (long waves) – quartz tube heaters (medium, short waves) – halogen tube heaters (ultra short waves).
There is more information about radiators in Section 11.4.
Infrared heating
211
11.2.2 Heat transfer As indicated earlier electromagnetic radiation having a wavelength in the range from 75 10 6 cm to 100,000 10 6 cm (0.000075–0.1 cm) covers the infrared spectra (Fellows 1988). Infrared rays thus occupy that part of the electromagnetic spectrum with frequency less than that of visible light and greater than that of most radio waves although there is some overlap. The name infrared means ‘below the red’, i.e. beyond red, or lower-frequency (longer wavelength) end of the visible spectrum. Infrared radiation is thermal for objects whose temperature is above 10 K. The radiation gives up its energy to heat materials when it is absorbed. The rate of heat transfer depends on • the surface temperature of the heating and receiving materials • the surface properties of the two materials • the shape of the emitting and receiving bodies.
The amount of heat emitted from a perfect radiator (termed black body) is calculated using the Stefan-Boltzmann equation Q AT 4
11:1
1
8
where Q (Js ) is the rate of heat emission, ( 5.7 10 J s m K 4) the Stefan-Boltzmann constant, A (m2) the surface area and T (K ºC 273) the absolute temperature. This equation is also used for a perfect absorber of radiation, again known as a black body. Figure 11.2 shows the spectral characteristics of black body radiation from objects at different temperatures. These curves give the maximum possible radiation that can be emitted by any body for the temperatures shown. A black body produces maximum intensity according to Planck’s law Q C1 5
ec2 =T
1
1
1
2
11:2
Here Q is the radiant flux emitted per unit area per unit increment of wavelength w/m2/microns. C1 3:740 10 12 W cm2 and C2 is 1.438 cm, K. It should be remembered that Wien’s law states that the product of the wavelength corresponding to maximum emittance and the emission temperature of the black body is constant if is expressed in microns max T 2900 m K
11:3
Table 11.1 shows some data for different IR radiators. The wave length distribution is important both for penetration, since it is not a distinct wavelength, and for the energy transfer. However, radiant heaters are not perfect radiators and foods are not perfect absorbers, although they do emit and absorb a constant fraction of the theoretical maximum. To take account of this, the concept of grey bodies is used, and the Stefan-Boltzmann equation is modified to Q "AT 4
11:4
212
Thermal technologies in food processing
Fig. 11.2 The spectral characteristics of black body radiation from objects at different temperatures.
Here " is the emissivity of the grey body expressed as a number from 0 to 1. Emissivity varies with the temperature of the grey body and the wavelength of the radiation emitted. Table 11.1
Some data for IR radiators
IR radiator
max (m)
Temp (K)
Max energy flux (kW/m2)
Radiation <1.25 m (%)
Ultra short wave Short wave 1 Short wave 2 Medium wave Long wave
1.0 1.12 1.24 1.8 3.0
2627 2316 2066 1338 694
4010 2547 1697 1697 50
41.1 32.9 25.8 7.0 0.2
Infrared heating
213
11.2.3 Optical properties/IR reflection When infrared waves hit a material they are reflected r, transmitted t or absorbed (see Fig. 11.3). The amount of radiation absorbed by a grey body is termed the absorptivity and is numerically equal to the emissivity. Radiation, which is not absorbed, is reflected and this is expressed as the reflectivity 1 . The amount of absorbed energy, and hence the degree of heating, varies from zero to complete absorption. This is determined by the components of food, which absorb radiation to different extents, and the temperature of the source determines the wavelength of infrared radiation. Higher temperatures produce shorter wavelengths and greater depth of penetration. The net rate of heat transfer to a food therefore equals the rate of absorption minus the rate of emission: Q "A
T14
T24
11:5
where T1 (K) is the temperature of emitter and T2 (K) the temperature of absorber. The absorbed waves are transformed to heat and the temperature of the material increases. When the waves penetrate the material the vibrations and rotation of the molecules are changed. The two fundamental vibrations that occur are stretching and bending. Stretching means a decrease or an increase of the distance between atoms and bending means a movement of the atoms. When the infrared light hits a molecule, energy is absorbed and the vibration changes. When the state of the molecule returns to the absorbed, energy will be transformed to heat (Fig. 11.4). The infrared absorption properties of foodstuffs are not as easy to describe. Regular reflection takes place on the surface of a material and is approximately 4% for most organic materials. In the case of body reflection the light enters the material, diffuses because of the scattering, and undergoes some absorption (Dagerskog 1978). Regular reflection produces only the gloss or shine of the
Fig. 11.3
The total energy (E) is either reflected (r), absorbed () or transmitted (t).
214
Thermal technologies in food processing
Fig. 11.4
Molecular vibrations.
polished surfaces while body reflection produces the colour and patterns that constitute most of the information we obtain visually. For materials with rough surfaces, both regular and body reflection will become diffuse. The reflection at short wavelength ( < 1:25m) is generally high, approximately 50% and at longer wavelengths is long: less than 10%. It is important to remember this fact in order to choose the right type of infrared radiators. The actual reflection has to be measured for each foodstuff at different stages of the process and translated into what it means at the actual lamp distribution when calculating heat transfer. The penetration properties are important for optimising the system. The penetration depth is defined as 37% of the unabsorbed radiation energy. For short waves, the penetration ability is ten times higher than for long waves. The Table 11.2 The calculated penetration depths for crust and crumb for radiators used in baking ovens (Skjo¨ldebrand et al. 1988) Power level (%)
Maximum energy wavelength
Spectral range (nm)
Penetration depth Crumb
Crust
100
1300
75
1320
50
1410
800–1250 1250–2500 800–2500 800–1250 1250–2500 800–2500 800–1250 1250–2500 800–2500
3.8 1.4 1.9 3.8 1.4 1.9 3.8 1.4 1.8
2.5 0.6 1.2 2.5 0.6 1.1 2.5 0.6 1.1
Infrared heating Table 11.3 1979) Product
Potato Potato Pork Bread
215
¨ sterstro¨m Measured penetration depths for some foods (Dagerskog and O Radiation source max
m
Penetration depth
1.12 1.24 1.12 1.12
4.76 4.17 2.38 6.25
Wavelength range (m) < 1:25 1:25 < < 1:51 0.48 0.47 0.28 1.52
> 1:51 0.33 0.31
direct penetration ability of infrared radiation makes it possible to increase the energy flux without burning the surface and thus reduce the necessary heating time that conventional heating methods require. This is especially true for thin products. In a special study a method was developed to determine optical properties of bread at different degree of baking (Skjo¨ldebrand et al. 1988). The results were that the transmission by the crust is less than the crumb. Even the thinnest dough sample did not transmit any radiation. Reflection curves for crust and dough are very similar while the reflection for the crumb is about 10–15% less. Table 11.2 shows calculated penetration depths for crust and crumb for radiators used in baking ovens. Measurements have been carried out for other foods and Table 11.3 shows some examples (Dagerskog and ¨ sterstro¨m 1979). O
11.3
Technologies
In IR heating, heat is transferred by radiation, the wavelength of which is determined by the temperature of the body – the higher the temperature, the shorter the wavelength. Present interest in industrial heating applications centres on short wave IR (wavelengths around 1 m) and intermediate IR (around 10 m), since these wavelengths make it possible to start up and reach working temperatures in seconds, while also offering rapid transfer of high amounts of energy and excellent process control. In some food materials, moreover, short wave IR demonstrates a penetration depth of up to 5 mm (see Section 11.1). The best known industrial applications (for non-food uses) are in the rapid drying of automobile paint and drying in the paper and pulp industry (see Section 11.1) For paper drying IR has superseded microwaves because it offers superior process control and economy. IR technology has long been under-estimated in the food field, despite its great potential. The main commercial applications of IR heating are drying low moisture foods (examples are drying of breadcrumbs, cocoa, flours, grains, malt, pasta products and tea). The technique is often used as one part of the whole process very often at the start to speed up the first increase of the surface
216
Thermal technologies in food processing
temperature. Such processes are frying, baking and drying. Radiant heating is used in baking or roasting ovens and is also used to shrink packaging film. In a recent search in the literature it was found that there is some use of the IR technique on drying of fish products (Wei-Renn-Lein and Wen-Rong-Fu 1997). A lot of research in recent years has been done in Taiwan, China and Japan (Afzal and Abe 1998; Wen-Rong-Fu and Wei-Renn-Lien 1998). They have studied drying of fish, drying of rice, rice parboiling and potato. The basic characteristics of infrared radiation are the high heat transfer capacity, heat penetration directly into the product, fast regulation response and good possibilities for process control. These qualities indicate that infrared radiation should be an ideal source of energy for heating purposes. As distinguished from microwave heating the penetration properties are such that suitable balance for surface and body heating can be reached which is necessary for optimal heating (see Section 11.1). Some empirical work in this field can be found in the literature by, for example, Ginzburg (1969). Suggestions have been made that radiant heating elements should be operated at temperatures between 1200 and 1800ºC as only wavelengths longer than 2 m are effective in developing colour. Successful results have been reported for several frying applications (Dagerskog 1978). Asselberg et al. (1960) used quartz tube heaters (1000 1300ºC) at 2.2 W/m2 for braising of beef stew. For the parboiling procedure, for a similar degree of heat treatment as compared with conventional technology, the infrared treatment required a shorter time (83%) with lower weight losses (50%). The flavour, colour and texture of the infrared braised meat were claimed to be superior. An industrial process for pre-cooking of bacon in a continuous infrared oven at Swift & Company has been investigated by Hlavacek (1968). Electric resistance heaters below the seamless stainless steel belt supplemented the 288 kW of infrared radiant heating from overhead quartz lamps. The frying time was 2–3 minutes and pre-cooked bacon was found to taste as good or better than freshly fried bacon. Several studies have been reported by Soviet investigators concerning the frying of meat with infrared radiation, but the only work published in English is that of Bolshakov et al. (1976) on the production of baked pork meat products. By analysing transmittance spectrographs of lean pork they showed that the maximum transmission of infrared radiation is for the wavelength region of 1.2 m. For wavelengths longer than 2.5 m the transmission was negligible. Consequently, it is necessary to use sources with the maximum radiation falling in the region of maximum transmission to achieve deep heating of pork. For heat treatment of the pork surface radiators in the region of maximum transmittance and reflectance (max > 2:3 m must be used. The authors therefore designed a two-stage frying process. In the first stage surface heat transfer was bought about by a radiant flux with max at 3.5–3.8 m. In the second stage the product was subjected to an infrared radiation flux with max at 1.04 m providing deep heating of the product. The result showed that the final moisture content and sensory quality of the product heated by the two-stage process were higher than those heated by conventional methods.
Infrared heating
217
The effect of radiation intensity (0.125; 0.250; 0.375 and 0.500 W/cm2 and slab thickness (2.5, 6.5 and 10.5 mm) on moisture diffusion coefficient of potato during far IR drying have been investigated by Afzal and Abe in 1998 in Japan. They found that the diffusivity increased with increasing radiation intensity and with slab thickness. In contrast activation energy for moisture desorption decreased with increasing slab thickness and resulted in higher drying rates for slabs of greater thickness. In Taiwan also far IR has been used for dehydration of fish. Over 90% of the far IR dried products had higher quality than currently marked sun dried product (Wei-Renn-Lein and Wen-Rong-Fu 1997).
11.4
Equipment
IR ovens or equipment of various sizes and constructions have been developed and tested in many countries. The main component – the radiator – may be of various types and shapes. Early tests with tube heaters revealed that successful radiant cooking required both the quality and quantity of energy used should be suitably balanced. As described in Section 11.2.1 the radiators may be divided into the following main groups: 1. 2.
Gas-heated radiators (long waves) Electrically heated radiators – tubular/flat metallic heaters (long waves) – ceramic heaters (long waves) – quartz tube heaters (medium- and short wave) – halogen heaters (ultra short waves)
Various reflector systems are also used (see Fig. 11.5) (Hallstro¨m et al. 1988): • individual metallic/gold reflectors • individual gilt twin quartz tube • flat metallic/ceramic cassette reflector.
Some of the high intensity radiators need water or compressed air cooling to avoid overheating. Table 11.4 shows the infrared emitter characteristics. IR equipment may be either of batch or continuous type. The radiator cassettes are positioned above the transport belt, which usually are wire mesh as indicated in Fig. 11.6. Some equipment also uses IR heating from below if the product allows this from a contamination point of view. The equipment shown in Fig. 11.7 has been used in reheating and frying in catering, utilises individual pans for the product to avoid fat dripping. In other equipment the IR system is combined with air convection to control the surrounding air temperature and humidity. On most equipment the degree of heating is controlled by thyristor systems. The simplest way is to use pure on/off systems where the number of radiators/tubes operating at any one time is controlled by switches. Advantages
218
Thermal technologies in food processing
Fig. 11.5
(a) Individual reflector (b) Individual gilt twin quartz tube (c) flat metallic ceramic cassette reflector.
Fig. 11.6
Fig. 11.7
Continuous process infrared oven.
Infrared oven in which heating is from above and below the product: this oven utilises individual pans for the product.
Table 11.4
Infrared emitter characteristics (Fellows 1988)
Type of emitter
Radiant heat
Convection heat
Heatingcooling time
(kW/m2)
Maximum process temperature (ºC)
(%)
(%)
(s)
2200 2300 2200
10 2 80
300 1600 600
75 98 80
25 2 20
1 1 1
5000 h – 5000 h
Medium wavelength Quartz tube
950
60
500
55
45
30
Years
Long wavelength Element Ceramic
800 700
40 40
500 400
50 50
50 50
<120 <120
Years Years
Short wavelength Heat lamp IR gun Quartz tube
Maximum running temperature (ºC)
Maximum intensity
Expected life
220
Thermal technologies in food processing
Fig. 11.8
Infrared Teknik oven.
of microcomputer control systems over convection ovens include: • • • •
very short heating time fast heating due to penetration of radiation into the product easy to pre-program and regulate the heating cycle for different products high energy efficiency.
At SIK, the Swedish Institute for Food and Biotechnology, two types of equipment were tested. The first one was built by Infrared Teknik in Sweden. This equipment was constructed as a baking oven. It was 4.3 1 m (see Fig. 11.8), continuous, i.e. it had a conveyor that transported the bread through the
Infrared heating
221
oven, and the speed of the conveyor could be varied, and the oven is divided into three zones with different effects of IR radiators. All the lamps installed are of short waves. The lamps were placed in cassettes that in turn were placed both above and under the conveyor. The total input was 125 kW and this was distributed in the three zones as follows:
Zone 1 Zone 2 Zone 3
Cassette above kW 30 16 12
Cassette under kW 32 16 8
The radiators at the top are placed at an angle to the conveyor belt so the radiation falls on the side of the bread. Each radiator can be switched on or off separately. The IR effect and the speed of the conveyor can be varied and controlled from the control panel. The top cassette can be regulated to have different distances between the product and the radiators. The distance is varied between 24 and 41 cm. The cassettes below the product are at a distance of 18 cm and the baking time can be varied between 2 to 30 minutes. There is a steam zone before the radiating zones. The oven has been used for baking in a bakery as well as in the laboratory. The second baking oven was built by a Swedish company (TRIAB = TRI Innovation AB) in collaboration with SIK (see Fig. 11.9) (Skjo¨ldebrand and Andersson 1987). The oven is semi-continuous, which means that the bread is transported on a conveyor into the oven chamber. The conveyor stops during the baking procedure. The oven has two cassettes, which have 24 radiators each. One of the cassettes is placed at the bottom and one at the top of the oven. 12 of the 24 radiators emit electromagnetic waves mainly within the medium range of the infrared spectrum and 12 within the short wave range. The size of the oven cabinet is 1800 mm500 mm. A fan is placed in one of the walls, which makes it possible to combine the infrared heating with convection. The air temperature can be varied between 90ºC and 300ºC and the air velocity between 0 and 2 m/s.
Fig. 11.9
The research oven at SIK.
222
Thermal technologies in food processing
Fig. 11.10 Wave length distribution of the lamps in the research oven.
The maximum temperature of the short wave radiators is 2228 K which corresponds to a max of 1.3 m. The corresponding temperature for medium infrared radiation is 1113 K which corresponds to a max of 2.7 m. Also the power level per unit surface area can be varied continuously between 0 and 100%. The wavelength distribution of the short wave radiators and the corresponding power are shown in Fig. 11.10. The cassettes can be moved to create different distances between product and radiator. The parameters that can be varied are wavelength, level of power, air velocity and temperature. A computer-based system is connected to the oven, which will aid in giving reproducible experiments. Similar equipment from the same manufacturer is sold to the food industry. These often form only part of the process, for example the first step in a biscuit-baking process or in a meat frying process.
11.5
Applications: case studies and modelling
In this section some applications with the most important results will be summarised. Also some examples of modelling of the heating procedure will be described. Most applications of IR within the area of food came during the 1950s to 1970s from the United States, the Soviet Union and the eastern European countries. During the 1970s and 1980s SIK did a lot of basic work applying this technique within the area of food. More recently work has also been carried out in Japan, Taiwan and other countries. Most of this work is still of an experimental nature. Applications are mainly from the following areas
Infrared heating • • • • • • • •
223
drying of vegetables, fish drying of pasta, rice heating of flour frying of meat roasting of cereals roasting of coffee roasting of cocoa baking of pizza, biscuits and bread.
The technique has also been used for thawing, surface pasteurisation of bread and pasteurisation of packaging materials. 11.5.1 Baking When baking with irradiation it seems that short-wave radiators should be used. The short-wave infrared radiation may be combined with convection for drying out the surface with good results. Ginzburg divided the baking process using infrared radiation into three periods: • The first phase is characterised by an increase of the surface temperature (1– 2 mm) to 100ºC. Very little weight loss occurs during this period. • The second period is characterised by the start of mass transfer. An evaporation zone forms, which moves towards the central parts. Energy is used to evaporate water and to heat the dough. • In the third and final period the central parts have reached 90ºC and the temperature increases further about 8º at the end of baking. The duration of this period amounts to about 25% of the total time of baking.
When comparing time-temperature relations between infrared radiation and conventional baking it is clear that IR radiation is more efficient both at the surface parts and the central sections. Using the baking oven at SIK mentioned in Section 11.4, the following results were achieved using short wave infrared heating (Skjo¨ldebrand et al. 1994): • the baking time was 25–50% shorter compared to an ordinary baking oven, depending on the thickness of the product • energy consumption was comparable to ordinary baking • weight losses were 10–15% lower • quality was comparable.
These results show that infrared heating for bakery products is very promising compared to other heating techniques. Further studies at SIK have shown that baking of bread using the short-wave infrared heating technique is a very interesting alternative to traditional baking (Skjo¨ldebrand and Andersson 1987). The baking time can be reduced by 25%
224
Thermal technologies in food processing
but in some cases 50%. The reduction depends on the thickness of the product. This is due to penetration of the waves into the product. The penetration properties of the bread change during baking: at the start they are almost zero and the crust has poorer penetration depth than the porous water-rich breadcrumb. The baking time reduction is also due to the more effective heat transfer to the surface than convection or conduction heating. Using short wave infrared radiation may reduce the weight losses. In some of the experiments it was found that the water content in the centre had increased during baking causing better and longer storage. The sensory quality does not seem to be poorer for bread baked in an infrared oven. The advantages of using short wave infrared radiation were found to be • • • • •
high and effective heat transfer heat penetration/reduction of baking time no heating of air in the oven quick regulation and control compact and flexible ovens.
11.5.2 Mathematical modelling Several attempts have been made to model infrared heating of foods. To calculate temperature distribution during combined infrared and convection heating, the heat conduction equation was solved using the finite difference method where the material is divided into a number of layers. For calculation of heat generated by infrared radiation an exponential penetration was assumed. The following discussion shows the principal equations of the finite difference solution to the heat transport by combined conduction and infrared radiation. The first attempt to find a mathematical model for infrared heating was done by Dagerskog (1978). Skjo¨ldebrand and Falk (1990) made some further developments to the program. The computer program enables us to simulate the time-temperature relation at different points in the food process. Radiation can be combined with convection in the program. The program was developed using the American developed scientific software called ‘Asyst’ originally developed for use in datalogging. The program calculates the temperature distribution in heated foods by solving the heat conduction equations using a numerical method based on the finite difference technique. Theories and basic equations are developed for surface temperature, centre temperature and temperatures in the product between surface and centre. Since no pure conduction occurs, so-called effective or apparent thermal properties have been used. Energy balances are used for the calculations. Three different energy balances are made up for the following volume increments: • at the surface of the product • in the product • in the centre.
Infrared heating
225
The program consists of three main parts: 1. 2. 3.
initiation of the calculations calculations results of the calculations.
The program was tested on three cases: 1. 2. 3.
baking in conventional ovens baking using short wave infrared radiation grilling of meat.
The results for baking in a short wave infrared oven were that it was possible to simulate the temperature and these showed the same tendency as the ones measured, but the values are somewhat lower. The explanation for the deviation is complex since there could be errors in the thermal properties as such, but also other assumptions such as the volume changes in time or temperature of the thermal properties and the chain in the rate of heat transfer during the baking. For grilling it was easier to simulate the temperature curves as no calculation is needed for the surface thermal properties. The air temperature rose to quite a high level, no crust was formed on the meat and the surface temperature was below 100ºC. Skjo¨ldebrand et al. (1994) looked at the ratio between the radiate and the convective heat transfer of the total heat flux under different process conditions in a baking oven. Bread and a ‘dummy’ food were studied. An energy balance was set and the time-temperature relation was studied. It was found that • it is appropriate to assume that the radiate energy at low temperatures (100– 150ºC) is transferred to the product at an infinitesimal surface layer • the radiate heat transfer coefficient was 16–43 W/m2K at the radiant source temperature of 250 and 370ºC. • the proportion of radiate to convective energy is correlated to time more than correct surface temperature • the ratio of radiation and convective heat transfer (free or forced) ranges from 0.65 to 8. The convective coefficient ranges from 5–10 (free) to 30–40 W/mK • for dry bread the proportion of radiation to total amount of energy is 53% to 83% • attempts to simulate the temperature inside the bread with the developed model did not give any satisfactory results.
11.6
Future trends
I will describe future developments on infrared heating in the following vision of the future. Let us visit a small town in Europe in the year 2003. Both households and small companies are located in the town. One of the companies is a small enterprise called ‘Local Food’. It produces fresh bread and convenience foods based on bread. About 400 different articles are produced. The products are
226
Thermal technologies in food processing
distributed to households close to ‘Local Foods’, i.e. the country where the production is carried out. The products are filled French rolls, crepe, and pies pizza. New products are developed now and then. These new products are often developed based on the technique used, available raw material properties and wanted product properties. The company has just started producing ‘functional foods’ based on bread and fillings. The customers can buy and order their food via the Internet. The ordered product will be sent out at the time the customer has indicated. If he/she wants something special this is possible. The order goes directly to the production computer that controls the processes. The production unit is fully automated and the process line is based on flexible automation and is tailor-made to the product to be produced. New heating techniques like short wave infrared radiation and microwaves are used frequently and where it is appropriate. The aim is minimal processing based on knowledge about wanted product properties and its interaction with the processes. A computer is installed to provide decision support for the operator. It is possible to trace components or foreign substances throughout the whole production line. Sensors on-line are installed to measure product properties and feed forward process control is used. When the product is ready, the processes stop and the product is sent to the customer. The operators have good knowledge about their processes. This scenario shows the different developments that are reality in present time and will influence the future application of, for example, infrared heating techniques. To be able to fulfil this scenario the following are important points for future developments. • More knowledge about the interaction between processes and products needs to be gained. The relationship between raw material properties and how these are affected by the process to obtain the desired properties in the end product should be studied. These are necessary for the success of using new techniques like NIR or short wave infrared heating equipment. • IR heating should be particularly useful for continuous baking, drying and grilling as well as for surface pasteurisation. • As different heating techniques have their own limitations, areas of application and possibilities, good background knowledge may combine these in an optimal way. More knowledge is important for the success of the heating technique. • The use of IR technology in the food industry today is quite limited, and the available equipment is not optimised for the various heating operations along the processing lines for baking, drying, etc. Its application is certain to grow as food equipment manufacturers begin to realise its full potential. • Along with the development of process control and information technology the IR technique will show its full potential with fast regulation of radiators and rapid heat transfer. • IR heating certainly will fulfil its role in the requirement of flexible production units.
Infrared heating
227
• With the development of new products the heating technique used will be important. New flavours can be created via both the recipe and the heating technique.
11.7
References
AFZAL T M, ABE T
1998, Diffusion in potatoes during far infrared radiation drying. Journal of Food Engineering 37 4 353–365. ANON. 1974 The Infrared Handbook. Svenska Philips, Stockholm. ASSELBERG E A, MOHR W P, KEMP J G 1960, Studies on the application of infrared food in processing. Food Technology 14 449. BOLSHAKOV A S, BOUSKOV V G, KASULIN G N, ROGOV F A, SKRYABIN U P, ZHUKOV N N 1976, Effects of infrared radiation rates and conditions of preliminary processing of quality index on baked products. 22nd European Meeting of Meat Research Workers, Malmo¨, Sweden. DAGERSKOG M 1978, Stekning av livsmedel, Ph.D thesis, Chalmers University of Technology, Gothenburg, Sweden. ¨ STERSTRO ¨ M L 1979, Infrared radiation for food processing. I. A DAGERSKOG M, O study of the fundamental properties of infrared radiation. Lebensmittel Wissenschaft u. Technologie 12 237–42. FELLOWS P 1988, Food Processing Technology Principles and Practice. Ellis Horwood, Chichester, England and VCH, Weinheim, Germany. GINZBURG A S 1969, Application of Infrared Radiation in Food Processing. Leonard Hill, London. ˚ RDH C, SKJO ¨ M B, TRA ¨ GA ¨ LDEBRAND C 1988, Heat Transfer and Food HALLSTRO Products. Elsevier Sciences, London. HLAVACEK R G 1968, Bacon pre-fried in continuous infrared oven has excellent taste. Food Proc. 29 7 50–2. ¨ LDEBRAND C 1986, Cooking by infrared radiation. In: Proceedings from SKJO Progress in Food Preparation Processes. Swedish Institute for Food and Biotechnology, Gothenburg, 157–73. ¨ LDEBRAND C, ANDERSSON C G 1987, Baking using short wave infrared SKJO radiation. In: I. D. Morton, ed., Proceedings from Cereals in a European Context. First European Conference on Food Science and Technology. Ellis Horwood Ltd, Chichester, 364–76. ¨ LDEBRAND C, ELLBJA ¨ R C, ANDERSSON C G, ERIKSSON T 1988, Optical SKJO properties of bread in the near infrared range. Journal of Food Engineering 8 129–39. ¨ LDEBRAND C, ANDERSSON C G 1989, A comparison of infrared bread baking SKJO and conventional baking. Journal of Microwave Power and Electromagnetic Energy 24 2 91–101. ¨ LDEBRAND C, FALK C 1990, Development of a simulation programme for SKJO IR-heating. In: W. E. L. Spiess and H. Schubert, eds, Engineering and Food, Vol. 1: Physical properties and process control. Proceedings of the
228
Thermal technologies in food processing
Fifth International Congress on Engineering and Food, 28 May–3 June 1989, Cologne, Federal Republic of Germany. Elsevier Applied Science, London, 869–86. ¨ LDEBRAND C, VAN DEN HARK S, JANESTAD H, ANDERSSON C G 1994, SKJO Radiative and convective heat transfer when baking dough products. Proceedings from the 4th Bath Food Engineering Conference, University of Bath, UK. Campden & Chorleywood Food RA, Chipping Campden, 19– 21 September. WEI-RENN-LEIN, WEN-RONG-FU 1997, Small fish dehydration by far infrared heating. Food Science Taiwan 24 (3) 348–56 (in Chinese).
12 Instant and high-heat infusion J. Andersen, APV Systems, Silkeborg
12.1 Instant infusion: an introduction Instant infusion is a new technology enabling heat treatment of products previous considered too difficult or heat sensitive to heat treat, including products of medium to high total solid contents. These products include whey protein concentrate and other concentrates such as infant formula, processed cheese, egg-based products and desserts with high viscosity. Instant infusion provides either mild pasteurisation, or sterilisation, depending on product requirements. To ensure effective pasteurisation or sterilisation whilst preserving product quality, the following demands must be met: • fast and short heating time • short and accurate holding time at ‘sterilising’ temperature • fast cooling time.
Instant infusion meets these demands. Some of the characteristics for instant infusion of medium to high total solid content products are: • • • •
a
Capacity ranges: Max. viscositya at sterilisation temperature: Viscosity range prior to sterilisation: Max. dry matter, depending on viscosity:
Viscosity for Newtonian liquids.
50–20,000 l/h 500 cp 1 to 1000 cp normally 52%TS
230
12.2
Thermal technologies in food processing
Instant infusion in detail
An instant infusion plant typically consists of: • a preheating section using a plate/tubular heat exchanger • an instant infusion section including flash chamber • a final cooling section using plate tubular or scraped surface heat exchangers.
A flow diagram of an instant infusion plant is shown in Fig. 12.1. The following description takes the example of a milk dessert with, in brackets, the temperature range for other product types. 12.2.1 Preheating The product is preheated indirectly from 20ºC to 75ºC (40–110ºC) in the plate/ tubular heat exchanger before it enters the top of the infusion chamber. The typically indirect heating rate is 0.5 to 2ºC/s. In the infusion chamber the product obtains the ‘UHT’ temperature 143ºC (65ºC–160º) in less than 0.2 second giving a heating rate of approximately 400ºC/s. 12.2.2 The infusion chamber The infusion chamber needs both to heat the product gently and quickly, and avoid fouling. This is achieved using a number of features. The top of the infusion chamber is equipped with a special lamination nozzle. Through the nozzle, the product is dispersed in a circular pattern of individual ‘strings’ falling to the bottom of the chamber. Each string is in turbulent flow, creating a large surface for heat transfer. A spray ball is mounted in the middle of the lamination nozzle which is deployed during CIP. The steam enters the chamber from the side walls of the chamber. From here it moves towards the centre at laminar flow, condensing into the product without cavitation, which ensures no uncontrolled homogenisation effect on the product. During the fall impact, steam and product are effectively mixed and, at the same time, direct contact between product and the wall of the infusion chamber is avoided. This ensures maximum heat transfer whilst foam building is kept at a minimum. The lower part of the infusion chamber where product comes into contact with the wall is equipped with a cooling jacket. This creates a thin condensate film on the inside of the chamber cone which eliminates product fouling. The lamination nozzle, the heating during the free fall, and the cooling jacket are all factors contributing to the minimisation of foam building and burning on of the product, thus contributing to the attainment of a maximum product quality. Temperature control in the infusion system is carried out via pressure control. The sterilising temperature is controlled by maintaining a set steam pressure in the chamber, which is more accurate than using a temperature sensor. The temperature sensor mounted in the product discharge pump housing is used to monitor the actual sterilisation temperature and will adjust the steam pressure
75ºC STEAM
PRODUCT
FILLING
2
8
COOLING WATER
7
COOLING WATER
6
3
o\APV
143ºC VACUUM
4 73ºC
5ºC
25ºC
<25ºC
STEAM
1
5
5
COOLING WATER
1. 1. Plate Platepreheaters preheaters 2. chamber 2. Steam Steaminfusion infusion chamber
5. Plate coolers
3. 4. 8.
3. Flashvessel vessel Flash 4. Aseptic homogeniser Aseptic homogeniser 5. Plate coolers Condenser
Fig. 12.1
Instant infusion plant.
6. tank Aseptic tank 6. Aseptic 7. loop loop 7. SterilisingSterilising 8. Condensor
232
Thermal technologies in food processing
controller, if necessary. This combination allows the temperature in the chamber to be down to 2–3ºC of the final product sterilising temperature for some products, compared to injection and SSHE technology where a minimum T of 12–15ºC is required. At the bottom of the infusion chamber, the product falls directly into a positive rotary pump. The rotary pump ensures a uniform and defined holding time throughout the full running time. The effective holding time is considered as the holding time from the outlet of the infusion chamber to the outlet of the positive product discharge pump, where the product flashes instantaneously. This results in a consistent and effective holding time of < 0.5 seconds without use of any holding cell or back pressure valve. At the outlet of the pump the product is flashed down to 73ºC instantaneously. 12.2.3 The vacuum chamber The vacuum chamber is also called the flash chamber or flash vessel. In the vacuum chamber all the steam added during the heating is removed and, at the same time, the product is flash cooled down to 73ºC, 2–3ºC below preheating temperature. The 2–3ºC represents the amount of water added due to condensation of steam in the infusion chamber caused by cooling on the cone and the nozzle plate. The vacuum chamber is connected to a condenser which condenses the steam. The product is pumped from the vacuum chamber to a plate tubular/scraped surface heat exchanger section for final cooling to an outlet temperature of typically 25ºC.
12.3
Advantages and disadvantages of instant infusion
The main advantages of instant infusion are: • • • • • •
gentle heat treatment of the product, without partial overheating accurate holding time uniform bacteriological kill rate for both low and high viscosity products reduced chemical heat damage to both low and high viscosity products less fouling due to the elimination of hot surface contact during heating less homogenising effect on the product as the system operates without a back pressure valve prior to the flash chamber • uniform heating-holding-cooling time profile for low and high viscosity (1– 500 cp) products through the same installation • long operating times. Some of these advantages are discussed in more detail below. 12.3.1 Gentle heat treatment Instant infusion can be used on products that traditionally have been hard to heat treat. Instant infusion minimises chemical changes due to heat treatment.1 The
Instant and high-heat infusion
233
sterilisation temperature is reached within 1/5 of a second. Minimum fouling in the heating section, combined with the instantaneous flash cooling on the discharge pump outlet, ensures minimum or no changes in colour, flavour and fat stability. 12.3.2 Accurate heating and holding time Temperature control in infusion systems is carried out via pressure control. Pressure monitors react much faster than any temperature sensor. The temperature sensor mounted in the product discharge pump housing is normally used to monitor the sterilisation temperature and will alter the steam pressure controller, if necessary. This combination allows the temperature in the chamber to be within 2–3ºC of the final product sterilising temperature, resulting in no overheating of any product constituents. In the infusion chamber the product is heated to the sterilisation temperature in less than 0.2 seconds. The effective holding time is considered as the holding time from the outlet of the infusion chamber to the outlet of the positive product discharge pump, where the product flashes immediately. This gives an effective holding time of < 0.5 seconds, with the rotary pump ensuring a uniform and defined holding time. 12.3.3 Reduced fouling The short heating time in instant infusion results2 in a high concentration of unfolded -lactoglobulin which will foul when it gets in contact with metal. However, the product is heated during its free fall in the infusion chamber without any contact with hot surfaces. The product falls directly into a positive rotary pump, minimising fouling and ensuring no product deposit in the bottom of the infusion chamber. At the bottom of the chamber a ‘cooling’ jacket ensures a very thin condensate film, further reducing fouling at this point. 12.3.4 Long operating times Instant infusion technology provides long operating times for products with medium to high total solid contents. APV has successfully installed infusion plants operating continuously for up to 28 h with concentrates of 52%TS, depending on product composition. 12.3.5 The main disadvantages The main disadvantages of instant infusion are: • relatively high capital costs compared to indirect systems • relatively high operating costs due to lower heat regeneration and higher maintenance costs than PHE and THE, but lower than the maintenance costs for SSHE systems • the requirement for culinary steam
234
Thermal technologies in food processing
• whilst minimal chemical changes to products is usually an advantage, for those with high enzyme content, residual enzymatic activity can cause problems such as age gelling flavour and colour defects.
12.4
High-heat infusion: an introduction
Since the first reported case in 1985, the problem of heat resistant spores (HRS) has increased significantly. This development is a major reason for the development of high-heat infusion plants. Other factors in the development of high-heat technology include improved energy recovery compared to normal direct technology. High-heat infusion plants are used mainly for dairy products such as UHT milk, lactose-reduced milk, flavoured milks, various sauces and dressings, and, more recently, functional foods. The unique feature of high heat technology is that the product is ‘flash cooled’ prior to heat treatment. The main process phases can be summarised as follows:3 • the product is preheated to 95ºC for protein stabilisation (preheating is done in a tubular heat exchanger) • it is then vacuum cooled to 70ºC, removing air and water, the latter in an amount equivalent to the steam added later in the steam infusion chamber • the product is then further preheated indirectly to 120–130ºC • final heating to 150ºC takes place in the steam infusion chamber, where the product is cooled indirectly to 75ºC before homogenisation • the product may then be homogenised and further cooled to filling temperature (using tubular heat exchangers).
12.5
The problem of heat resistant spores (HRS)
In 19854 the first incident of highly heat-resistant spores was detected in southern Europe. This problem has later been detected in other European countries as well as Mexico and the United States.5 Bacillus sporothermodurans (HRS) consists of highly heat-resistant endospores. Growth in UHT milk is slow and the total number only reaches 105 bacteria/ml. The principal problem represented by HRS is that of spoilage. The heat resistance of the spores of Bacillus sporothermodurans is much higher than that of the other strains of Bacillus which usually give problems in UHT and ESL products. Huemer et al.6 showed that one strain of B. stearothermophilus had a z value of 9.1ºC and a D140 of 0.9 seconds where B. sporothermodurans had a z value of 13.1–14.2 and D140 values ranging from 3.4 to 7.9 seconds. These results show how difficult it is to kill Bacillus sporothermodurans. HRS are found to be homogeneous in molecular, physical and phenotypic characteristics. Phenotypically, HRS bacteria are negative for many standard tests used to identify Bacillus. HRS bacteria are identified as B.
Instant and high-heat infusion
235
aneurinolyticus and B. badius using API systems. However, they can be distinguished by additional tests.7 Until the first cases of heat resistant spores (HRS) were observed in 1985, a heat treatment giving a B* value > 1 was generally accepted as giving a commercial sterile product. To kill these HRS (classified as bacillus sporothermodurans strains) a heat treatment of at least B* 8 (or F0 value 40) is needed. Some references report even higher values than this.8 However, EU/International Dairy Federation Guidelines,9 as well as Good Manufacturing Practice (GMP) guidelines in many countries and companies, are based on maximum heat degradation in milk of 600 mg/kg lactulose. These guidelines prevent processors in Europe in particular from simply increasing sterilisation time or temperature in indirect UHT plants. The time/temperature combination needed to kill HRS is 150ºC in 2–6 seconds, depending on contamination load. To both kill HRS and minimise damage to products during heat treatment, it is generally agreed some sort of direct heating has to be applied. High-heat infusion has been developed in part to meet this objective.
12.6
High-heat infusion in detail
The temperature-time profile for high-heat infusion is shown in Table 12.1. The time-temperature profile for high-heat infusion is compared to that for indirect UHT treatment in Fig. 12.2. The following sections look in more detail at key steps in the process. 12.6.1 The infusion chamber An infusion chamber is shown in Fig. 12.3. The product is pumped into the nozzles at the top of the chamber, dividing the product into long ‘strings’ which fall through the chamber. These ‘strings’ create a large product surface, ensuring a short heating time and a low T. Steam, pressure and temperature are closely related, with steam pressure used to control the temperature. To Table 12.1
Temperature-time profile for high-heat infusion
Step Preheating – 1 Holding time Flash Preheating – 2 Infusion Cooling Homogeniser Final cooling
Temperature up to 95ºC at 95ºC to 70ºC to 125ºC to 150ºC to 70ºC
Time min. 30 s instant instant (0.2 s)
236
Thermal technologies in food processing
Fig. 12.2
UHT of products with HRS (comparative temperature profiles with F0 40).
prevent fouling, the nozzles at the top of the chamber can be cooled whilst, at the bottom cone of the chamber, the wall is cooled by water at 120ºC. This infusion chamber is for instant infusion with a rotary pump below the chamber. For normal and high-heat infusion, the chamber is equipped with a centrifugal pump. 12.6.2 The flash chamber The functions of the flash chamber are to cool the product down rapidly to the desired temperature. At the same time an efficient de-aeration will take place (see Section 12.7.1). The desired temperature for normal direct plants is usually 70–80ºC whilst, for high-heat infusion, it is 60–70ºC. At the same time the flash chamber removes the same amount of water as will be added as steam during the direct heating. The main components of a flash chamber are: • a back pressure valve needed to ensure pressure prior to the flash chamber and vacuum in the chamber itself • the extract pump which pumps the product from the flash chamber and further into the system • the vacuum pump and condensers, which create the vacuum in the flash chamber • the entrance to the chamber which is placed at a tangent to separate the product and the steam.
Instant and high-heat infusion
Fig. 12.3
237
238
12.7
Thermal technologies in food processing
Advantages and disadvantages of high-heat infusion
The main advantages of high-heat infusion include: • its capacity to kill HRS spores • longer operating times, due to reduced fouling • increased safety, because the flash chamber (and in some cases the homogeniser) is placed prior to UHT heat treatment • increased cost efficiency, especially in energy costs and the cost of maintenance, inspection, etc. • the possibility of aseptic flavour addition, giving the technology an advantage in the production of flavoured milk or ice cream mixes, for example • the capacity to manufacture a wider range of products.
Some of these advantages are discussed in more detail below. 12.7.1 De-aeration prior to final heat treatment In the high-heat infusion process, the product is flash-cooled from 95ºC to 70ºC in the preheating step, prior to final heating. As a result an amount of water equal to the amount later added as steam is removed. This process optimises deaeration. Effective de-aeration means that fouling is reduced and that unwanted volatile aroma components, such as flavour components derived from animal feed, will be removed together with the water and air. For flavoured milk, etc., the recipe may need to be modified in order to compensate for loss of flavours in the flash chamber. 12.7.2 Reduced risk of contamination In a normal direct heat treatment process (either infusion or injection) the milk is preheated followed by direct heating and consequent flash-cooling in a flash chamber. This sequence means that flash-cooling equipment has to be able to prevent contamination, for example through the use of steam barriers. In the highheat infusion technology, the flash chamber is placed prior to final heating, meaning that it need not be aseptic. In a high-heat infusion system, the product is under pressure from the moment it has been sterilised until it reaches the sterile tank. This sequence reduces the risk of contamination compared to a normal direct system. In a normal direct UHT plant an aseptic homogeniser is needed downstream. This is because the mixing or presence of steam bubbles during UHT heating and flash-cooling, combined with conditions at the back pressure valve, will damage the fat globules so much that a homogenisation is needed to re-establish the emulsions. 12.7.3 Flexibility in adding flavour components In high-heat technology, volatile flavour components are removed in the flash chamber at the end of the process. In high-heat infusion, flash-cooling occurs
Instant and high-heat infusion
239
prior to final heat treatment. This sequence opens up the option of a non-aseptic in-line addition of flavour/aroma components prior to sterilising. This flexibility is important in the manufacture of flavoured milks, for example. 12.7.4 Cost The capital cost of a high-heat infusion system is broadly comparable to that of other direct heating systems. It is, however, possible to upgrade existing UHT plant to incorporate high-heat infusion technology, reducing cost and installation time. The main cost advantage is in energy recovery. A high heat infusion plant has an energy recovery of 75–80% compared to the recovery of a normal direct plant of 40–50%, representing a considerable saving on running costs. Maintenance costs in high-heat infusion plants are also lower since only one process requires aseptic conditions, reducing the need for inspection. 12.7.5 Longer operating times High-heat infusion plant can be operated for longer than a normal infusion plant. The reason for this is that the milk is ‘pre-heat treated’ prior to the infusion treatment, reducing the amount of fouling in the infusion chamber and the holding cell.2 12.7.6 Increased product range High-heat infusion plant can be supplied as part of an integrated plant combining normal and high-heat infusion technology. Such integrated facilities make it possible to produce, for example, ESL-milk (milk with a shelf-life of two months but of comparable quality to normal pasteurised milk), normal UHT milk, and high-heat infusion milk products, creams, ice cream mixes, custard and other dairy desserts. 12.7.7 The main disadvantages The main disadvantage of high-heat infusion technology is that, to kill HRS, products undergo more severe chemical damage compared to the normal infusion process. In milk production the process also requires downstream homogenisation for the longest possible shelf-life to be obtained.
12.8 1. 2.
References DE JONG P, WAALEWIJN R, VAN DER LINDEN H J L J,
Performance of a steaminfusion plant for heating milk. Netherlands Milk & Dairy Journal, 1994, 48 181–99. DE JONG P, BOUMAN S, VAN DER LINDEN H J L J, Fouling of heat treatment
240
3. 4. 5.
6. 7. 8.
9.
Thermal technologies in food processing equipment in relation to the denaturation of -lactoglobulin. Journal of the Society of Dairy Technology, 1992 45 (1), February. APV, Long life dairy, food and beverage products, Technology Update. DUQUET J P, TROUVAT A, MOUNIQUA, ODET G, CERF O, Heat resistant spores in milk used in the manufacture of UHT milk. Laid, 67 393. KLIJN N, HERMAN L, LANGEVELD L, VAEREWIJCK M, WAGENDORP A A, HUEMER I, WEERKAMP A H, Genotypical and phenotypical characterization of Bacillus sporothermodurans strains surviving UHT sterilisation. Int. Dairy Journal, 1997 7 421–8. HUEMER I, KLIJN N, VOGELSANG H W J, LANGEVELD L P M, Thermal death kinetics of spores of Bacillus sporothermodurans isolated from UHT milk. Int. Dairy Journal, 1998 8 851–5. PETTERSSON B et al. Bacillus sporothermodurans, a new species producing highly heat-resistant endospores. International Journal of Systematic Bacteriology, 1996 July 759–64. HAMMER P, LEMBKE F, SUHREN G, HEESCHEN W, Characterization of a heat resistant Meosphilis Bacillus species affecting quality of UHT milk – a preliminary report. Kieler Milchwirtschaftliche Forschungsberichte, 1995 47 297–305. EU Council Directive 92/46/EEC, June 1992.
13 Ohmic heating R. Ruan, X. Ye, P. Chen and C.J. Doona, University of Minnesota and I. Taub, US Army Natick Soldier Center
13.1
Introduction
Over the past two decades, there has been an increasing shift from batch thermal operation towards continuous High Temperature Short Time (HTST) processing of foods. In HTST processes, food is processed continuously through plate or scraped surface heat exchangers at temperatures as high as 140ºC. At this temperature, only a few seconds are needed for sterilization, during which the products suffer only a slight quality deterioration. HTST processes rely on rapid convection heat transfer and are thus well suited to liquid foods. They are, however, limited in application to particulates since, for particles more than a couple of millimeters thick, the processing time is insufficient for heat to transfer to the center to give sterility. In ohmic heating processes, foods are made part of an electric circuit through which alternating current flows, causing heat to be generated within the foods due to the electrical resistance of the foods. Therefore, in a liquid-particulate food mixture, if the electrical conductivity of the two phases are comparable, heat could be generated at the same or comparable rate in both phases in ohmic heating. In other conditions, heat can also be generated faster in the particulate than in the liquid. Ohmic methods thus offer a way of processing particulate food at the rate of HTST processes, but without the limitation of conventional HTST on heat transfer to particulates. 13.1.1 History The concept of ohmic heating of foods is not new. In the nineteenth century, several processes were patented that used electrical current for heating flowable materials. In the early twentieth century, ‘electric’ pasteurization of milk was achieved by passing1 milk between parallel plates with a voltage difference
242
Thermal technologies in food processing
between them, and six states in the United States had commercial electrical pasteurizers in operation.2 In the design of McConnel and Olsson,3 frankfurter sandwiches were cooked by passing through electric current for a predetermined time. Schade described a blanching method of preventing the enzymatic discoloration of potato using ohmic heating.1 It was thought at that time that lethal effects could be attributed to electricity. The technology virtually disappeared in succeeding years apparently due to the lack of suitable inert electrode materials and controls. Since that time, the technology has received limited interest, except for electroconductive thawing.4 Within the past two decades, new and improved materials and designs for ohmic heating have become available. The Electricity Council of Great Britain has patented a continuous-flow ohmic heater and licensed the technology to APV Baker.5 The particular interest in this technology stems from ongoing food industry interest in aseptic processing of liquid-particulate foods. Conventional aseptic processing systems for particulates rely on heating of the liquid phase which then transfers heat to the solid phase. Ohmic heating apparently offers an attractive alternative because it heats materials through internal heat generation. 13.1.2 Principles The principles of ohmic heating are very simple as illustrated in Fig. 13.1. Ohmic heating is based on the passage of alternating electrical current (AC) through a body such as a liquid-particulate food system which serves as an electrical resistance in which heat is generated. AC voltage is applied to the electrodes at both ends of the product body. The rate of heating is directly proportional to the square of the electric field strength, E, and the electrical conductivity. The electric field strength can be varied by adjusting the electrode gap or the applied voltage. However, the most important factor is the electrical conductivity of the product and its temperature dependence. If the product has more than one phase such as in the case of a mixture of liquid and particulates, the electrical conductivity of all the phases has to be considered. The electrical conductivity increases with rising temperature, suggesting that ohmic heating becomes more effective as temperature increases, which could theoretically result in runaway heating. A difference in the electrical resistance and its temperature dependence between the two phases can make the heating characteristics of the system very complicated. Since electrical conductivity is influenced by ionic content, it is possible to adjust the electrical conductivity of the product (both phases) with ion (e.g. salts) levels to achieve effective ohmic heating. In ohmic heating, microbes are thought to be thermally inactivated. Other contributions to the kill mechanism have also been suggested. A mild electroporation mechanism may occur during ohmic heating operating at low frequency (50–60 Hz) which allows electrical charges to build up and form pores across cell walls.
Ohmic heating
243
Fig. 13.1 Schematic diagram showing the principle of ohmic heating.
13.1.3 Current status of commercial uses Currently, at least eighteen ohmic heating operations have been supplied to customers in Europe, Japan, and the United States. The most successful of these systems have been for the processing of whole strawberries and other fruits for yogurt in Japan, and low acid ready-to-eat meals in the United States.6 Currently there are two commercial manufacturers of ohmic heating equipment: APV Baker, Ltd., Crawley, UK, and Raztek Corp, Sunnyvale, CA, US. In the United States, a consortium of 25 partners from industry (food processors, equipment manufacturers, and ingredient suppliers), academia (food science, engineering, microbiology and economics) and government was formed in 1992 to develop products and evaluate the capabilities of the ohmic heating system. A 5 kW pilot-scale continuous-flow ohmic system manufactured by APV Baker, Ltd., Crawley, UK, was evaluated by the consortium at Land-O’Lakes, Arden Hills, Minnesota, from 1992 to 1994. A wide variety of shelf-stable lowand high-acid products, as well as refrigerated extended-shelf-life products were developed. They were found to have texture, color, flavor, and nutrient retention that matched or exceeded those of traditional processing methods such as freezing, retorting, and aseptic processing. The consortium concluded that the technology was viable. In addition to the technical evaluation, an economic study was also initiated. Ohmic operational costs were found to be comparable to those for freezing and retorting of low-acid products.7 While the economics and technology appear favorable, there still remains the business risk associated with start-up costs and unknown market potential. For this reason, the consortium did not pursue commercialization of ohmic heating processes. Although the technological approaches associated with aseptic food processing (i.e., pumps, fillers, heater electrode, etc.) have developed significantly, the identification, control, and validation of all the critical control points required to demonstrate that an ohmically processed multiphase food product has been uniformly rendered commercially sterile have yet to be
244
Thermal technologies in food processing
generally accepted. Consequently, the Food and Drug Administration (FDA) has no current filing of continuous ohmically processed multiphase food products. The ohmic heating process has the promise to provide food processors with the opportunity to produce new, high-value-added, shelf-stable products with a quality previously unrealized with current sterilization techniques. Applications that have been developed include aseptic processing of high-value-added readyprepared meals for storage and distribution at ambient temperature; preheating of food products prior to in-can sterilization; and the hygienic production of high-value-added ready-prepared foods for storage and distribution at chilled temperatures. Ohmic heating can also be used for heating of high-acid food products such as tomato-based sauces prior to hot-filling, with considerable benefits in product quality. Other potential applications include rapid heating of liquid food products, which are difficult to heat by conventional technologies.8 Other potential future applications for ohmic heating include blanching, evaporation, dehydration, fermentation, and extraction. The disadvantages of ohmic heating are associated with its unique electrical heating mechanisms. For example, the heat generation rate may be easily affected by the electrical heterogeneity of the particle, heat channeling, complex coupling between temperature and electrical field distributions, and particle shape and orientation. All these make the process complex and contribute to non-uniformity in temperature, which may be difficult to monitor and control. 13.1.4 Advantages The advantages of ohmic heating technology claimed in the previous research are summarized as follows.5, 9 • Heating food materials by internal heat generation without the limitation of conventional heat transfer and some of the non-uniformity commonly associated with microwave heating due to limited dielectric penetration. Heating takes place volumetrically and the product does not experience a large temperature gradient within itself as it heats. • Higher temperature in particulates than liquid can be achieved, which is impossible for conventional heating. • Reducing risks of fouling on heat transfer surface and burning of the food product, resulting in minimal mechanical damage and better nutrients and vitamin retention. • High energy efficiency because 90% of the electrical energy is converted into heat. • Optimization of capital investment and product safety as a result of high solids loading capacity. • Ease of process control with instant switch-on and shut-down. • Reducing maintenance cost (no moving parts). • Ambient-temperature storage and distribution when combined with an aseptic filling system. • A quiet environmentally friendly system.
Ohmic heating
13.2
245
Ohmic heating process and equipment
13.2.1 Flow chart and key equipment Figure 13.2 is a schematic of a continuous flow ohmic heating process. A viscous food product containing particulates enters the continuous-flow ohmic heating system via a feed pump hopper. The product then flows past a series of electrodes in the ohmic column, where it is heated to process temperature. Then the product enters the holding tubes for a fixed time to achieve commercial sterility. Next, the product flows through tubular coolers and into storage tanks, where it is stored until filling and packaging. Most ohmic heating system configurations consist of three modules: heater assembly, power supply and control panel. 13.2.2 Equipment design Equipment design is a critical factor that should be considered. The reason for the early failure of ohmic heating to be applied widely on a commercial scale was the absence of inert electrode materials and control equipment accurate enough to keep the temperature within the necessary range and sufficiently robust to withstand the conditions of commercial production. Currently, commercially available designs include electrodes that are located at various positions along the length of the product flow path (in-line field), or those located perpendicular to the flow (cross-field), differing principally in distribution of electric field strength.
Fig. 13.2
Schematic of a continuous-flow ohmic heating process.
246
Thermal technologies in food processing
13.2.3 Electrode Previous designs attempted to use a range of electrode materials from graphite to aluminum or stainless steel. In food processing, high standards of hygienic design are required; electrodes must be carefully designed. In the early designs the electrolytic effect that causes the dissolution of the metallic electrodes was completely neglected, and material technology had not progressed to the stage that a suitable electrode material was available. For recent technologies, such as the APV ohmic process, the use of a food-compatible electrode electrical material and the correct current density has eliminated contamination problems. Other ways to overcome this problem include utilizing high power frequency, since at alternating frequencies above 100 kHz, there is no apparent metal dissolution. 13.2.4 Control Getchell10 was the first author to emphasize the importance of controlling the ohmic heating process. Since that time, significant developments in semiconductor technology have taken place, increasing the sophistication of possible control equipment and strategies. In continuous processing, problems can result if a single electrode pair is used to heat food material through a large change in temperature; substantial changes in liquid conductivity and thus in heating rate, may result along the length of the electrode. Multiple sets of electrodes are easier to control. However, Biss et al.11 report that pure feedback control was not suitable for the APV Baker ohmic process due to large time constants, and describe the development of a feed-forward control scheme. For feed-forward processes, some idea of the characteristics of the system are needed; control depends as much on process knowledge as on the design of the control loops. 13.2.5 Commercial equipment A typical commercial ohmic heating system for liquid-particulate mixture is the APV Baker ‘ohmic heating’ process. The process was originally developed by the UK Electricity Council Research Centre, and was then licensed to APV Baker who have developed it into a commercial system. A diagram of the process is given in Fig. 13.3. Food is pumped through a vertical pipe containing a series of cylindrical electrodes connected to a 50 Hz three-phase supply. Electric current thus flows through the food in the pipes connecting the electrodes. The food material is rapidly heated to sterilization temperature, then passes to a holding section and finally to an aseptic packaging plant. Unlike most previous techniques, the advantage of ohmic heating in giving rapid sterility is not lost due to the packaging process. The use of multiple electrodes gives a much greater degree of control than is possible in other techniques, together with a uniform electric field in the pipe sections. Specially-developed electrode material is used to eliminate polarization and contamination. The process allows food products containing particulates up to 25 mm to be heated to sterilization temperatures up
Ohmic heating
Fig. 13.3
247
The 5 kW pilot scale ohmic heating system by APV Baker, Ltd. (Courtesy of APV Baker, Ltd.)
to 140ºC in less than 90 seconds, after which they are cooled back down to ambient temperatures within 15 minutes. These processing times are significantly shorter than the typical two-hour process cycles for in-can sterilization techniques. The whole process is automatically controlled using a programmable logic controller. Because all early attempts at commercialization failed due to problems associated with electrode degradation and uneven heating of the product, APV conducted research on several key areas of the technology: electrode arrangement design and establishment of safety standards for such equipment; validation procedures to ensure proper sterilization of food products; and research to ensure the absence of any toxicological effect on food products. These studies resulted in a successful submission in 1991 to the UK Advisory Committee for Novel Foods and Processes for approval of the APV ohmic heating system for the production of ambient stable low-acid, ready-to-eat meals in the United Kingdom. The development also earned the Institute of Food Technologists’ 1996 Industrial Achievement Award.
13.3
Monitoring and modeling of ohmic heating
13.3.1 Mathematical modeling The future utilization of ohmic heating by the industry will depend on development of adequate safety and quality assurance protocols. One crucial
248
Thermal technologies in food processing
component in understanding the process lies in the development of mathematical models, which can then be used to simulate various effects of critical factors. Two different modeling approaches currently have been published. De Alwis-Fryer model De Alwis and Fryer12 use the solution to Laplace’s equation to calculate heat generation rate together with transient energy balance equations to model a single particulate immersed in a fluid medium without convection. This model has been extended by Zhang and Fryer13 to include multiple spheres uniformly distributed on a lattice within a non-convective fluid. A typical situation is illustrated in Fig. 13.4, wherein a cylindrical particle is stationed in the middle of a tube filled with stationary, non-convective fluid. An electric field is applied along the length of the tube. The electric field or voltage distribution can be developed from Maxwell’s equations, or by combining Ohm’s law and the continuity equation for electrical current: r
i rV 0
13:1
where V voltage, r gradient, i electrical conductivity of phase I which can take on different values for the particles and liquid. Ignoring convection effects, the heat transfer problem is one of pure conduction with internal energy generation: r
ki rT u_ i i Cpi
@T @t
13:2
where i again represents the phase, k thermal conductivity, u_ specific internal energy generation rate, density, Cp specific heat capacity, T temperature, and t time. The external boundary condition is one of convection to the surroundings: !
kiS rT n U
TiS
T1
13:3 !
where kiS thermal conductivity of phase i at surface, n unit normal vector, U overall heat transfer coefficient, TiS surface temperature of phase i, and T1 surrounding temperature.
Fig. 13.4
Situation simulated in de Alwis–Fryer model. (Redrawn from Sastry and Salengke.14)
Ohmic heating
249
The internal energy generation term in Equation [13.3] is given by: u_ i jrV j2 0i
1 mi T
13:4
where rV voltage gradient, 0i initial electrical conductivity, mi temperature compensation constant, T temperature. The system of Equations [13.1–13.4] can be solved by the Galerkin–Crank– Nicolson algorithm, a hybrid-spatially finite element, temporally finite difference scheme. Sastry-Palaniappan model Sastry and Palaniappan15 used circuit analogy to approximate electrical conductivity and thus the heat generation for a static heater with a particle immersed in a well-mixed fluid (assuming infinite convective heat transfer within the fluid). A typical situation is as illustrated in Fig. 13.5. Sastry16 extended this approach to a continuous flow ohmic heater for high solid concentration. The effective electrical resistance can be determined using the circuit analogy. The effective resistance of the cell is calculated as: R R fs1 R fs2
R fp R sp R fp R sp
13:5
where the resistance R’s can be calculated from the electrical conductivity and the geometry. The average voltage gradients (rV ) in the liquid and particulate can simply be calculated as: rV
V L
or
rV
IR V and I L R
13:6
where V is the applied voltage and L the distance between the two electrodes.
Fig. 13.5
Situation simulated in Sastry–Palaniappan model. Also shown is the circuit analogy. (Redrawn from Sastry and Salengke.14)
250
Thermal technologies in food processing
The energy balance on the well-mixed fluid phase is: Mf Cpf
dTf u_ f vf np hfp Ap
TsSm dt
Tf
UAw
Tf
T1
13:7
where the subscript f represents liquid phase and p particulate phase, Mf mass of fluid, Cp specific heat capacity, u_ f specific internal heat generation rate, v volume, np number of particulates, hfp liquid-particulate convective heat transfer coefficient, Ap surface area of one particle, TsSm mean particulate surface temperature, U overall heat transfer coefficient to surroundings, Aw area of heater wall. The specific internal heat generation rate for the liquid phase u_ f is given by: u_ f jrV j2 0 f
1 mf T
13:8
The particulate heat according to the conduction heat transfer equation with internal energy generation: r
ks rTs u_ s s Cps
dTs dt
13:9
where u_ s jrV j2 0 s
1 ms Ts
13:10
Temperatures of the two phases are linked by the convective boundary condition: !
krT n js hfp
TsS
Tf
13:11
This system of equations can be solved by forward differences for the liquid phase, and the Galerkin–Crank–Nicolson method for the solid phase. For any aseptic process, safety deserves the highest priority. The limitations of these two models and the lack of understanding of the temperature distribution mandate that the conservative approaches be used. Any lowtemperature regions that result will extend the process time needed to ensure satisfactory commercial sterility and will make it more likely that other regions of the food will be overcooked. Ohmic heating thus will lose its most attractive advantage. The effects of convection were over simplified in the two models. In reality, the convection heat transfer rate is neither zero nor infinity. A more general model is thus required and convection effects must be included to model the temperature profile. 13.3.2 Experimental and instrumental monitoring Invasive Locating the coldest spot within an ohmically heated food system is of significant concern for food engineers and processors. For a high temperature short time thermal process such as heating, the determination of the spatial and temporal
Ohmic heating
251
distribution of temperature within the particulate is necessary and important. Not only will it provide information for calculation of lethality and cook value, but it will also provide a guide to mathematical modeling and process control. However, mapping the intra-particle temperature distribution in food material undergoing dynamic ohmic heating is a difficult and complex task. Thermal couples were used for modeling verification and process monitoring. Temperature could only be obtained from selected points and the integrity of the process could be disturbed. Moreover, it is difficult to monitor the temperature of a flowing particle. Non-invasive: magnetic resonance imaging Magnetic resonance imaging (MRI) has contributed greatly to a diversity of scientific disciplines in recent years. Applications in the food area have increased significantly as researchers have discovered the power and flexibility of this technique.17 Food-related MRI research has gone beyond static imaging experiments to experiments involving dynamic processes, diffusion, water mobility, flow, water and oil distribution, and temperature distribution.17 The major advantages of MRI are that • it is nondestructive and non-invasive to the material being imaged and the complexity of foods can be probed without disturbing the sensitive balance • it can provide high spatial resolution • it can provide diverse information such as proton density (which is related to moisture or fat concentration), internal structure, chemical shift, diffusion, temperature and flow.
Among all the food-related MRI techniques, MRI temperature thermometry is emerging as an attractive and promising temperature mapping method.18–22 As pointed out by Hills23 the true potential of MRI lies not in static structure determination but in non-invasive, real time and dynamic changes as foods are processed, stored, packaged and distributed. When using MRI to map temperature distribution in a dynamic process like ohmic heating, the data acquisition should be as fast as possible so that a real-time measurement can be accomplished. If this temporal resolution cannot be achieved, the measurement will be inaccurate or perhaps even meaningless. Significant progress has been made to apply MRI techniques to mapping temperature distribution of ohmically heated food systems. Principles and methodology MRI is an extension of nuclear magnetic resonance (NMR) spectroscopy. Briefly, NMR spectroscopy is based on magnetic behavior of certain nuclei (such as protons in water and fat) in a sample placed in an external bulk magnetic field, and subjected to a radio frequency (RF) pulse. The nucleus of an atom possesses a positive charge, and is considered as spinning about an axis. The spinning of the nucleus generates a magnetic field, which is similar to that
252
Thermal technologies in food processing
generated by a simple bar magnet, and is termed the nuclear magnetic dipole. When placed in a magnetic field B0 generated by a permanent magnet, for instance, the nucleus will interact with the applied field via its magnetic dipole, and tend to precess about the direction of the applied field at a specific frequency known as Larmor frequency !, which is proportional to the strength of the applied field B0, and governed by the following equation: ! B0 ;
13:12
where is the magnetogyric ratio, a fundamental property of the nucleus. If a second magnetic field generated by using a radio frequency (RF) coil, for instance, whose frequency exactly matches the Larmor frequency !, is applied as a RF pulse to the nucleus, then resonant absorption of energy occurs. This resonance effect is hence termed nuclear magnetic resonance. In the time following the excitation, the excited spins give up energy and return to their equilibrium state. The energy is released in the form of an RF wave, characterized by its Larmor frequency, and discharged into the environment through two mechanisms: spin-lattice and spin-spin relaxation processes, each of which is characterized by a time constant. The time constant for spin-lattice relaxation is called T1 while for spin-spin relaxation it is called T2. In general, T1 and T2 can be functions of structure, molecular mobility, temperature, solute and water concentrations, and maybe other physical and chemical properties of the samples. The RF coil also functions as an antenna to receive the RF signals emitted from the nucleus during the relaxation processes, by which the signal decay or relaxation behavior of the spins with characteristic T1 and T2 can be recorded. MRI is based on a function of spatial position of magnetic fields instead of using a uniform static field. If a linear field gradient Gx is superimposed on the main magnetic field along the x direction, the resonance frequency at which the spin precesses is a function of spatial position along x: !
x
B0 Gx x
13:13
Transformation of the data then yields not only the magnitude but also spatial information. The same principle can be extended to achieve spatial information in two or three dimensions, which can be used to construct two- or threedimensional magnetic resonance images. NMR parameters that cause image contrast include nucleus intensity (signal strength), T1, T2, magnetization transfer rates, self-diffusion coefficients, proton resonance frequency shift, and velocity profiles, and are often termed ‘contrast agents’. Using different contrast agents, MR images or maps of NMR parameters and structure of a cross-section of a sample can be obtained. The fact that many of these contrast agents are a function of temperature makes constructing temperature maps with MRI instruments possible. Among these contrast agents, T1, proton resonance frequency shift (PRF), and signal intensity are often the choices for temperature mapping.
Ohmic heating
253
T1 methods Simply put, the T1 methods involve the following procedures: (1) (2) (3) (4) (5)
acquisition of T1 relaxation data, fitting of T1 values, constructing T1 images, experimental calibration between T1 and temperature, and translating T1 image to temperature image.
Inversion recovery and spin echo imaging sequences are common T1 imaging methods. Both methods are considered time consuming and unsuitable for monitoring fast changing ohmic heating process. Inversion Recovery-Flip Low Angle Snap Shot (IR-FLASH) is one of the fast pulse sequences for T1 image acquisition. In one IR-FLASH sequence, magnetization gradually recovers after an inversion pulse; a FLASH image is acquired at a designated time (recovery time) during the recovery. A series of IR-FLASH sequences is repeated, but only with varying recovery time. The images acquired at different recovery times are used to construct the T1 image by fitting the magnetization values of the images into a single-exponential recovery equation. Total acquisition time for a T1 image, which usually requires ten or more inversion recovery FLASH images, is about 30 seconds to 1 minute or more. Although this time is much shorter than the two conventional methods mentioned above, it is still long for dynamic food processes like ohmic heating. To shorten the data acquisition time, the pulse sequence can be altered so that we could take a snapshot of the sample at different recovery times during a single recovery without having to repeatedly revert the signal to its original position (negative maximum) for each image. To map T1, a series of FLASH images are acquired during magnetization relaxation following a single inversion pulse, almost without delay between two adjacent images. After T1 values are computed and a T1 image is constructed, the next step is to correlate T1 to temperature and construct a temperature image based on the T1 image. Calibration between T1 and temperature is done by measuring the signal intensity of a spot in the MR image where the temperature is measured using fiber optic temperature sensors. This procedure is repeated for several spots and at different temperatures. The relationship between T1 and temperature () is determined using a linear function as follows T1 a b
13:14
where a and b are constants related to the nature of the material. This calibration should be carried out for each material. To translate the T1 data into temperature for a particle, the relationship described by the above equation is applied within the boundary of the sample. This translation procedure can be automated using a computer program that discerns the true signals from background noise and then applies the temperature fit to the true signals only.
254
Thermal technologies in food processing
PRF method The temperature sensitivity of the proton resonance frequency (PRF) was first observed by Hindman in 1966.24 The fractional change of water proton resonance frequency (!=!) with temperature is defined as . It is also referred to as proton chemical shift or PRF shift. Mathematically, ! T !
13:15
where T is the temperature change. Experimentally, has been estimated as 0.01 ppm/ºC in water22 and is about the same for ex vivo tissues.25 It is generally assumed that the water PRF shift to lower frequency with higher temperature is caused by rupture, stretching, or a small amount of bending of the hydrogen bonds.24, 26 This means a reduction in the average degree of association of water molecules, and hence that these shifts are evidence of an increased average shielding constant of the protons. The proton resonance frequency shift imaged at a static field strength B0 after having undergone a temperature change of T is: ! B0 T
13:16
where is the gyromagnetic ratio. This frequency change manifests as a phase change when imaged with a gradient-echo sequence having an echo time TE. This phase change can be expressed as: B0 T TE
13:17
To use the PRF shift technique to map temperature, a reference phase image is first acquired at a known temperature, and then subtracted from subsequent phase images taken at different temperatures. Temperature maps can therefore be obtained based on the reference temperature and the echo time TE of the image sequence according to equation (13.17) which also indicates that B0 TE T
13:18
The temperature sensitivity of phase images depends on two factors, i.e., the main magnetic field and the echo time of the image sequence. For the same echo time, the temperature sensitivity for a 4.7 Tesla scanner would be three times more than that for a 1.5 Tesla scanner. The actual PRF shift of a material may depart slightly from the apparent value, 0.01 ppm/ºC. Therefore, the shift is usually regressed to the known temperature in order to find the actual . Ohmic heating of potato, carrot, and beef Examples Potato, carrot, and beef were cut into about 2.5 cm3 cubes. The carrier medium was made up of 50 g/kg starch and 1 g/kg NaCl. An experimental ohmic heating device was constructed of glass. It consisted of a 38 mm dia and 318 mm long
Ohmic heating
255
Fig. 13.6 Schematic diagram of the experimental set-up for MR imaging of ohmically heated food particulates.
glass tube with a rubber stopper at each end (Fig. 13.6). A 30 mm diameter stainless still electrode was fixed to each of the stoppers and connected to a transducer corresponding to a 50 Hz AC power supply. A hole was made in one of the stoppers to allow release of pressure built up within the tube during heating. Two fiber optic temperature sensors could be inserted into the tube through this hole, one near the end of the tube and the other in the center. A constant 140 volts was applied. The MRI images (Fig. 13.7) constructed from these images show that the temperature in the particulates and the corresponding rate of heating the particulates increased with heating time. The increased heating rate was consistent with the rise in electrical conductivity of the particulates with temperature. Wang and Sastry27 reported a higher electrical conductivity for ohmically heated materials compared to unheated raw materials. An increase in heating rate with time was also observed in the liquid phase, reflecting the increase in electrical conductivity with temperature. These images indicate differences in temperature among different particulates in the same carrier fluids. Beef appeared to heat faster than the others. The differences in heating rate among different materials could be due to differences in electrical conductivity9 and/or non-uniformity in the electric field. When two temperature probes were placed in different locations within the heating tube, the observed temperature variation was considerable. The MR images also show that the temperatures in the center regions of a particulate were higher than in the outer regions, indicating that the particulate heated intensively and transferred heat to the colder carrier liquid. Variation in temperature was observed within a particle, probably due to spatial variation in
256
Thermal technologies in food processing
Fig. 13.7 Two-dimensional MR images showing temperature distribution in particulate potato, beef, and carrot undergoing ohmic heating. The numbers indicate the heating time (heating was terminated after 13.75 minutes).
electrical conductivity. Two additional images (Fig. 13.7), taken 28.75 and 38.75 min after heating was terminated show that the heat redistributed within the particulates. These results indicate that, in modeling the ohmic heating process, the generation of heat and associated distribution of temperatures during the heating stage and the redistribution of temperatures during the holding stage all need to be taken into consideration. Ohmic heating of whey gels Examples Whey gels composed of 20% Alacen whey protein powder (New Zealand Milk Products) and 80% distilled deionized water along with NaCl solutions were used as models of a liquid-particulate mixture. Two similar samples of the model system were prepared. It consisted of a 305 mm long hollow cylinder of whey gel and 0.01% NaCl solution. 1.5% NaCl based on the mass of water was also added to the whey gels to adjust their electrical conductivity. A PVC thermal/electrical barrier was inserted into the hollow whey gel to form an insulated passage in the center of the gel cylinder as shown in Fig. 13.8. The model system was configured to resemble electrical circuits and ohmically heated by the application of an AC power supply with a constant voltage of 143 V and frequency of 50 Hz.
Ohmic heating
Fig. 13.8
257
Schematic diagram of the experimental set-up.
An experimental ohmic heating device was constructed (Fig. 13.8). It consisted of a Plexiglas vessel that was 43 mm in inner diameter and a nylon stopper at each end. A 35 mm diameter stainless steel electrode was fixed to each stopper and connected to the power supply. The distance between the two electrodes was 305 mm. A small hole was drilled in one of the stoppers and the Plexiglas vessel to allow the release of pressure build-up during heating. Two fluorescent fiber-optic temperature sensors were inserted through the holes into the whey gel and the solution at the same cross-sectional location that would be scanned to monitor the temperature for calibration. The absolute accuracy of the fiber-optic measurements was 0.2ºC. MR susceptibility artifacts were eliminated by using the non-metal temperature sensors. Figure 13.9 shows the phase change, , plotted against the temperature change in the whey gel of the sample and the regression line obtained by using linear regression through origin. The slope of the regression line was 0.0567 radians/ºC and the standard error for the phase was 0.0574 radians. These values correspond to a PRF shift of 0.0098 ppm/ºC and a temperature uncertainty of 1.01ºC. Figure 13.10 is the corresponding graph for the NaCl solution. The PRF shift and the temperature uncertainty in this case were 0.0096 ppm/ºC and 2.07ºC. Using the phase reference image, phase difference images and the derived PRF shift values, temperature maps of the sample at time 2, 4, and 8 minutes during heating were constructed and are shown in Fig. 13.11. The spatial resolution and temporal resolution were 0.94 mm and 0.64 s respectively. PRF shift was linearly and reversibly proportional to the temperature change. The temperature uncertainties determined were about 1ºC for the whey gel and about 2ºC for the NaCl solution. The temperature maps show that there existed a slight gradient along the radial direction. The existence of this gradient is due to the internal heat generation of ohmic heating process and the radiation heat transfer from the particle surface through the vessel wall to the ambient. Therefore, the cold spots of the particle should be the surface and corners.
258
Thermal technologies in food processing
Fig. 13.9
Phase change vs. temperature change and the regression line in the whey gel.
Fig. 13.10 Phase change vs. temperature change and the regression line in the solution.
Intrinsic chemical marker Kim et al.28, 29 developed an intrinsic chemical marker approach that was used in conjunction with direct microbiological measurements to map the temperature distribution and calculate lethality for aseptic processing of particulate food. Selecting conditions in which particulates heated faster than the liquid, temperature gradients within the particulates (containing precursor and bacteria) were demonstrated in terms of the concentration of thermally produced compounds (chemical markers) and of the surviving bacterial population.
Ohmic heating
259
Fig. 13.11 Temperature maps of whey gel during ohmic heating (2, 4, and 8 min).
Chemical markers can be viewed as time-temperature integrators over an HTST domain relevant to thermal processing of foods. At a given temperature, the concentration of a suitable chemical marker should be directly proportional to the log reduction in bacterial population. Measuring the yields of intrinsic chemical markers produced in hollow whey cylinders subjected to ohmic heating provided another method for mapping the temperature distribution generated in the sample that produced results remarkably consistent with those found using MRI. The experimental conditions were similar in each case. For the marker method, a 25.2 cm hollow cylinder of whey filled the
260
Thermal technologies in food processing
entire length of the ohmic cell. In addition to containing 1.5% NaCl, the 20% whey preparation also contained 0.4 M ribose, the precursor for chemical marker M-2. The initial applied voltage was 140 V. To monitor the temperature during heating, individual fiber optic probes were inserted through the top of the ohmic cell to a depth of approximately 11 cm in each of the particulate and solution phases. After the ohmic heating, the whey cylinder was cooled and sectioned, then M-2 was harvested from distinct locations of the sample. The yield of M-2 was analyzed at its absorption maximum of 285 nm using HPLC/UV-Vis. Three primary differences distinguish the MRI and chemical marker experiments. First, MRI measurements are made during heating, and marker yields are determined post-processing. Second, the ohmic cell is oriented horizontally for MRI analysis, but vertically for marker experiments. Third, the marker experiments were controlled to maintain constant high sample temperatures by gradually reducing the applied voltage during the heating process. In conditions in which the temperature of the whey was maintained constant at a temperature of 94ºC (Fig. 13.12a) generated uniform M-2 yields along the long axis of the whey cylinder sample (Fig. 13.12b). The extent of conversion of ribose to M-2 ranged from 20–25% of the maximum attainable yield. The slight trend of increasing marker yield from the bottom to top of the whey sample represents stratification in which heat rises from the lower regions to heat the upper regions faster. Although MRI methods have excellent resolution along the radial axis, measuring marker yields, which are not constrained to a single slice, conveniently complement the MRI results by determining the temperature along the long axis of the sample. In a similar experiment, the temperature-time profile for each phase (Fig. 13.12c) demonstrates that the high electrical conductivity whey cylinder heats faster than the low conductivity central solution. This type of heating pattern has been successfully mathematically modeled using the Unit Cell Model.30 Temperature-sensitive liquid crystal Sastry and Li31 reported a method that used temperature-sensitive liquid crystal sheets to monitor temperature of particulate flowing in a continuous ohmic heater. Transparent solid particles were suspended in fluid and a liquid crystalcoated sheet changed color from black to red to green to blue to black over a specified range of temperature. The sheet was placed parallel to the electric field so the interference with the electric field was minimal and the entire temperature profile inside a solid object could be visualized. This method provided useful information for the temperature distribution and for the model verification.
13.4 Major challenges and needs for future research and development Three major challenges hindering the commercialization of ohmic heating processing are:
Ohmic heating
261
Fig. 13.12 Ohmic heating of whey gel and NaCl solution in the chemical marker experiment: (a) temperature of whey gel as a function of ohmic heating time, (b) chemical marker yield in whey under constant temperature condition as shown in a, (c) effect of electrical conductivity on heating of whey and solution.
262 1.
2. 3.
Thermal technologies in food processing Lack of a complete model that takes into account differences in electrical conductivity between the liquid and solid phases and the responses of the two phases to temperature changes, which affect relative heating rates and distribution. Lack of data concerning critical factors affecting heating, including residence time, orientations, loading levels, etc. Lack of applicable temperature validating techniques for locating cold/hot spots.
These issues are in part being addressed, as the preceding sections indicate. 13.4.1 Develop reliable, predictive models of ohmic heating patterns Modeling ohmic heating is difficult owing to the unique character of this mode of heating, which requires much understanding of the critical factors. The ohmic heating rate is critically dependent on the electrical conductivity of the foods being processed for which only limited information is available. If the electrical conductivities of the liquid and particulate phases are the same, the mixture can be heated rapidly and uniformly to a high temperature irrespective of particle size. If the conductivity of the particulate is higher than the liquid, then it heats faster and transfers heat to the liquid, which has advantages for ensuring process adequacy.9 Nevertheless, possible heat channeling, which could cause coupling between temperature and electrical field distributions as well as sensitivity to process parameters, e.g. particle shape and orientation, could contribute to the complexity of the process. To ensure sterilization, the heating behavior of the food must be understood, so process reliability and safety could be demonstrated. Kim et al9 showed that monitoring the temperature at the entrance and exit of the holding tube could provide such assurances. Mathematical modeling allows insight into the heating behavior of the process. Spatial and temporal temperature distribution obtained from a reliable mathematical model that incorporates the critical factors can provide information for the calculation of lethality and cook value. It will also save time and money for validation experiments, process and product design. Modeling of a continuous ohmic heating process is difficult, because a number of different physical phenomena occur during the heating process. The verification of the predictions by any model will be limited to selected regions within the system. These limitations require that direct or indirect measurements of the temperature within the product and its constituents be made when establishing a process. Using appropriate conservative assumptions can compensate for some of these limitations (at the expense of the product quality). A general and reliable model is needed. 13.4.2 Develop product specifications and process parameters for specific products Particulates are the centerpiece around which an ohmic heating formulation is built. Contrary to conventional heating, in which we expect little difference in heat transfer due to changes in particle orientation, the heating pattern of an
Ohmic heating
263
ohmically-heated food system would be influenced by particle orientation. De Alwis et al.12 showed the heating of identically-shaped potato particles differed depending on whether the particles were aligned parallel or perpendicular to the electric field. De Alwis et al. explained that this difference is due to changes in the electric field, but it also reflects a change in the equivalent resistance of the overall circuit. When both the heating and cooling stages are considered, a practical limitation on particle size and shape might be expected. Cooling of particulates will always be thermal conduction controlled. The center of large particles may cool too slowly and thus become overprocessed during prolonged cooling. Consequently, particulate size is limited to 1 in3. Various combinations of particulates can be successfully processed when accompanied by suitable product and process control. Optimizing the combination will results in excellent particulate texture through uniform heating. The liquid phase must also be optimized. Viscosity should be determined at various temperatures to assure adequate suspension of particulates and to facilitate liquid/particle interface heat transfer. Overall product specification is important in determining how much lethal treatment is delivered during the process. Critical factors include particle size, shape, and orientation, viscosity, pH, specific heat, thermal conductivity, solidliquid ration, and electrical conductivity. Process design, a complete description of the critical processing conditions, is also important in ensuring lethality and optimizing quality. It should include batch formulation procedures, initial temperature, flow rate or particle residence time, exit temperature, and solid loading levels. These parameters will be specific for individual systems and formulations, and their impact on heating behavior needs to be understood. 13.4.3 Develop real-time temperature monitoring techniques for locating slowest heating regions Pioneers of ohmic heating research have documented that a particle does not heat uniformly during an ohmic heating process because of the non-uniform nature of the electric field and the differences in the physical properties of the food materials.32 As with any other thermal process, it is important to have information on the temperature-time history of the coldest point within the liquid-particulate system undergoing ohmic heating. It is assumed that the agitation of a continuous processing system minimizes any disparity among the temperature profiles of the particulates. However, there is no sufficient published evidence to indicate what the temperature is within a particle and how the temperature profile changes during a continuous process. It is clear that, for a particle with a homogeneous electrical conductivity greater than that of the liquid, the particle heats faster than the liquid phase, its coldest spot occurs at its surface1 and the liquid is colder than the particle.9 The location of the slowest heating part of the system is especially important, because its thermal lethality must be ensured. This is the key factor to determine the processing time.
264
Thermal technologies in food processing
Conventional tools such as thermocouples and fiber optic probes are invasive when used to measure ohmic-heated food systems. A non-destructive and noninvasive technique that can be used to monitor the spatial distribution of temperature is important for understanding and controlling ohmic heating. The MRI temperature mapping technique described here is essential for model development and the validation of the novel ohmic heating process. There is a need to further improve the MRI technique, using it to collect data under various product specifications and processing conditions, and to use it to validate mathematical models.
13.5 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11.
12. 13. 14.
References SCHADE A L,
Prevention of enzymatic discoloration of potatoes, 1951, US patent 2,569,075. SASTRY S K, PALANIAPPAN S, Ohmic heating of liquid-particle mixtures, Fd Technol., 1992, 46, 64–7. MCCONNEL S V, OLSSON R P, Wiener vending machine, 1938, US Patent 2,139,690. DE ALWIS A A P, FRYER P J, The use of direct resistance heating in the food industry, J Fd Eng, 1990, 11, 3–27. SKUDDER P J, Ohmic heating: new alternative for aseptic processing of viscous foods, Food Engineering, 1988, 60, 99–101. ANON., Ohmic heating garners 1996 industrial achievement award, Fd Technol., 1996, 20, 114–15. ZOLTAI P, SWEARINGEN P, Product development considerations for ohmic processing, Fd Technol., 1996, 50, 263–6. PARROTT D L, Use of ohmic heating for aseptic processing of food particulates, Fd Technol., 1992, 46, 68–72. KIM H J, CHOI Y M, YANG T C S, TAUB I A, TEMPEST P, SKUDDER P J, TUCKER G, PARROTT D L, Validation of ohmic heating for quality enhancement of food products, Fd Technol., 1996, 253–61. GETCHELL B E, Electric pasteurization of milk, Agric. Engng., 1935, 16, 408. BISS C H, COOMBES S A, SKUDDER P J, The development and application of ohmic heating for the continuous heating of particulate foodstuffs, Engineering innovation in the food industry, Bath, UK, Institution of Chemical Engineers, 1987, 11–20. DE ALWIS A A P, FRYER P J, A finite element analysis of heat generation and transfer during ohmic heating of foods, Chem. Eng. Sci., 1990, 45, 1547–59. ZHANG L, FRYER P J, Models for the electrical heating of solid-liquid food mixtures, Chem. Eng. Sci., 1993, 48, 633–43. SASTRY S K, SALENGKE S, Ohmic heating of solid-liquid mixtures: a comparison of mathematical models under worst-case heating conditions, Journal of Food Process Engineering, 1998, 21, 441–58.
Ohmic heating 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.
SASTRY S K, PALANIAPPAN S,
265
Mathematical modeling and experimental studies on ohmic heating of liquid-particle mixtures in a static heater, Journal of Food Process Engineering, 1992, 15, 241–61. SASTRY S K, Application of ohmic heating to continuous sterilization of food, Indian Food Industry, 1992, 11, 28–30. RUAN R R, CHEN P L, Water in foods and biological materials: A nuclear magnetic resonance approach, Lanaster: Technomic Publishing Co., 1998. CHANG K, RUAN R R, CHEN P L, FULCHER R G, BASTIAN E, Moisture, fat and temperature mapping of cheese block during cooling using MRI, Paper No. 956535, 1995, presented at ASAE Summer meeting. HULBERT G J, LITCHFIELD J B, SCHMIDT S J, Temperature mapping in carrot using T1 weighted magnetic resonance imaging, J. Fd Sci., 1995, 60, 780–5. RUAN R R, LONG Z, CHANG K, CHEN P L, TAUB I A, Glass transition temperature mapping using magnetic resonance imaging, Transactions of the ASAE, 1999, 42, 1055–9. SUN X, LITCHFIELD J B, SCHMIDT S J, Temperature mapping in a model food gel using magnetic resonance imaging, J. Fd Sci., 1993, 68, 168–72, 181. SUN X, SCHMIDT S J, LITCHFIELD J B, Temperature mapping in a potato using half Fourier transform MRI of diffusion, J. Fd Proc. Engr., 1994, 17, 423–37. HILLS B, Food processing: an MRI perspective, Trends in Food Science & Technology, 6, 111–17. HINDMAN J C, Proton resonance shift of water in the gas and liquid states, J. Chem. Phys., 1966, 44, 4582–92. PETERS R D, HINKS R S, HENKELMAN R M, Heat-source orientation and geometry dependence in proton-resonance frequency shift magnetic resonance thermometry, Magn. Reson. Med., 1999, 41, 909–18. MULLER N, REITER R C, Temperature dependence of chemical shifts of protons in hydrogen bonds, J. Chem. Phys., 1965, 42, 3265–9. WANG W C, SASTRY S K, Changes in electrical conductivity of selected vegetable stuffs during multi-thermal treatments, IFT Annual Meeting 1995, Anaheim, CA, 1995. KIM H-J, TAUB I A, Intrinsic chemical markers for aseptic processing of particulate foods, Food Technol., 1993, 91–9. KIM H J, CHOI Y M, HESKITT B, SASTRY S K, LI Q, Chemical and microbiological investigation of ohmic heating of particulate foods using a static ohmic heater. IFT Annual Meeting, Anaheim, CA, 1995. DOONA C J, POULIN N, SEGARS R, TAUB I A, KANDLIKAR S, RUAN R, YE X, Modeling ohmically-heated whey protein gel circuits, IFT Annual Meeting, Chicago, IL, 2000. SASTRY S K, LI Q, Modeling the ohmic heating of foods, Food Technol., 1996, 50, 246–8. FRYER P J, DE ALWIS A A P, KOURY E, STAPLEY A G F, ZHANG L, Ohmic processing of solid-liquid mixtures: heat generation and convection effects, Journal of Food Engineering, 1993, 18, 101–25.
14 Combined high pressure thermal treatment of foods L. Ludikhuyze, A. Van Loey, Indrawati and M. Hendrickx, Katholieke Universiteit, Leuven
14.1
Introduction
The prospects of high pressure as a food preservation method were first reported by Hite (1899) who observed spoilage microorganisms in milk and meat to be reduced by high pressure treatment and subsequently expanded his work to fruits and vegetables. In 1914, Bridgman stated that egg white coagulated under specific conditions of pressure and temperature, establishing that in addition to killing microorganisms high pressure could modify protein structure. Nevertheless, no attempts were made at that time to introduce high pressure in food preservation and processing (Knorr, 1995a), probably because of technical difficulties associated with pressure processing units and packaging materials. During the last decade, high pressure gained renewed interest in the area of food preservation and processing, mainly for two reasons: (i) the growing consumer demand for high-quality, fresh-like foods that are safe and additivefree stimulated research efforts in the area of new ‘nonthermal’ technologies and (ii) the current status of high pressure technology is such that operating, process, control and safety requirements imposed by the food industry can readily be met (Mertens, 1995). As a result of these research efforts, it has been shown that high pressure allows inactivation of vegetative microorganisms and spoilage enzymes while only minimally affecting quality attributes such as colour, flavour and nutritional value. Bacterial spores, on the other hand, cannot be reduced by high pressure alone. In this case, combination with other preservation techniques, in particular mild temperature elevation, is required. Next to enzyme and microbial inactivation, some key effects of high pressure include
Combined high pressure thermal treatment of foods 267 protein modification (denaturation, gelation, texturisation) as well as changes in product functionality (phase transitions, density, textural properties). This chapter will give an overview of the effects of high pressure on different aspects relevant for food processing and preservation and will try to review the current status of high pressure technology with respect to its application in the food industry.
14.2
Effect of high pressure on microorganisms
14.2.1 Vegetative microorganisms At moderate pressure, growth and reproduction rate of vegetative bacteria is retarded while at higher pressures, inactivation occurs. Although pressure stability is largely dependent on the type of microorganism, the species and the medium conditions, it is generally admitted that pressures between 200 and 600 MPa at room temperature are sufficient to cause a substantial reduction of viable vegetative cells. Vegetative forms of prokaryotes such as yeasts and moulds are most pressure sensitive and inactivated by pressures between 200 and 300 MPa. Gram bacteria can be inactivated by pressures of about 300 MPa and are, in their turn, less pressure stable than Gram bacteria, for which pressures higher than 400 MPa are required for inactivation. However, numerous exceptions on these general statements can be found. One of the most pressure sensitive groups of bacteria is the Gram Vibrio species (Berlin et al., 1999). V. paraheamolyticus (106 CFU/ml) in buffer and clam juice can be completely inactivated by exposure to pressure of 170 MPa for 30 and 10 min respectively (Styles et al., 1991). On the other hand, some very pressure resistant strains of E. coli O157:H7 were found by Benito et al. (1999). Although for E. coli these pressure resistant strains also seemed to be rather heat resistant, the relation between heat and pressure resistance could not be generalized. For L. monocytogenes only a very weak correlation was found whereas for Salmonella no correlation (Smelt, 1998). Since medium conditions largely affect pressure resistance of microorganisms, results of studies in buffer or laboratory media cannot directly be extrapolated to real food situations. Microorganisms are often observed to be more stable in real food products. Listeria monocytogenes was found to be completely inactivated at 340 MPa in buffer but exposed higher resistance in UHT and raw milk (Styles et al., 1991). For Salmonella seftenberg and thyphimurium, inactivation occurred more rapidly in phosphate buffer as compared to a chicken-base product (Metrick et al., 1989). Shigehisa et al. (1991) reported pressures as high as 600 MPa to be required for inactivation of Micrococcus luteus, Staphylococcus aureus or Streptococcus faecalis in pork slurries. In general, the protective effect of real food products has been attributed to the presence of proteins and sugars. On the other hand, synergistic effects between pressure and acidification or addition of anti-microbial substances can be exploited to lower pressure resistance of microorganisms (Hauben et al., 1997; Garcia-Graells et al., 1998).
268
Thermal technologies in food processing
14.2.2 Bacterial spores The formation of so called ‘spores’ is a unique strategy of survival of some bacterial genera in extreme stress conditions, among which the most important for food preservation are Bacillus and Clostridium. The extreme resistance to physical and chemical treatments gained upon sporulation is believed to be related with factors such as core dehydration, impermeability of coat layers or membranes and presence of small acid soluble proteins. When the stress factor is removed and conditions become suitable for outgrowth and multiplication, germination can take place, i.e. the conversion of the dormant to the vegetative state. It is generally admitted that, although spore counts can be lowered by exposure to high pressure, combination with other preservation methods, such as mild temperature elevation, is required for substantial reduction of viable spore counts (Hoover, 1993). Larson et al. (1918) observed that pressure treatments up to 1800 MPa at room temperature were not sufficient to obtain commercial sterility of food products. On the other hand, some authors convincingly showed that comparably low pressures (< 200 MPa) can trigger spore germination, the pressure and temperature level to induce a maximal germinative effect being strongly dependent on the strain under investigation (Clouston and Wills, 1970; Gould et al., 1970; Murell and Wills, 1977). Hence, Sojka and Ludwig (1994, 1997) suggested the use of a two step process to overcome the problems associated with the extreme pressure resistance of bacterial spores. Such process includes an initial mild pressure treatment to induce spore germination followed by a treatment at higher pressure and temperature to kill the germinated spores. However, biological diversity in germinability within one spore population and the lack of information on the kinetics of germination seems to limit this approach (Heinz, 1997; Wuytack, 1999). Therefore, the combined effect of high pressure and temperature on spore inactivation has been investigated. Many authors reported combination of pressure with temperatures of 60ºC and higher to be required for extensive inactivation of spores: the lower the pressure applied, the higher the required temperature to induce a preset extent of inactivation (Sale et al., 1970; Heinz, 1997; Wuytack, 1999). At temperatures below 60ºC in combination with pressure of about 400 MPa, maximal three log-cycle reductions were obtained for Clostridium sporogenes and Bacillus coagulans spores (Roberts and Hoover, 1996; Mills et al., 1998). As a conclusion, the major benefit of high pressure treatment for food preservation is the reduction of the thermal resistance of the spores. However, this synergistic effect seems to be somewhat impaired at higher temperature.
14.3
Effect of high pressure on food quality related enzymes
In view of the specificity of enzymatic reactions, enzymes may be affected by pressure in several ways (Cheftel, 1992): (i) pressurisation at room temperature
Combined high pressure thermal treatment of foods 269 may bring about reversible or irreversible, partial or complete enzyme inactivation resulting from conformational changes in the protein structure, (ii) enzymatic reactions may be enhanced or retarded by pressure, depending on the positive or negative reaction volume, (iii) a macromolecular substrate (protein, starch) may become more sensitive to enzymatic depolymerization or modification once it has been pressurized and (iv) intracellular enzymes may be released in extracellular fluids or cell cytoplasm due to alteration of the membranes by pressure, thereby facilitating enzyme-substrate reactions. With respect to enzymes related to quality of fruit and vegetable products, for which high pressure treatment is believed to offer great potential in the area of preservation and processing, research efforts have mainly been focused on enzyme inactivation, while the release from the membrane and the reactions they catalyse to a much lesser degree. Some key enzymes in fruit and vegetable processing include polyphenoloxidase (PPO), which is responsible for enzymatic browning and the concomitant quality deterioration, lipoxygenase (LOX) which induces changes in flavour, colour and nutritional value, pectinmethylesterase (PME) which is responsible for cloud destabilization and consistency changes and peroxidase (POD) which gives rise to unfavourable flavours. Because of their importance in food industry, thermal inactivation kinetics of these food quality related enzymes have been studied extensively in the past and are generally well documented. Throughout the last decade studies on pressure inactivation have consistently been increasing. While initially the potentials of high pressure for enzyme inactivation were investigated on a qualitative basis, more systematic, quantitative results are becoming available today. PPO, whatever its origin, is not extremely heat resistant (Lourenc¸o et al., 1990; Yemenicioglu et al., 1997; Weemaes et al., 1998a) as treatment at temperatures exceeding 70ºC are in most cases sufficient for partial or total destruction of its catalytic function (Vamos-Vigyazo, 1981). High pressure research carried out so far revealed that upon pressurization, PPO may display, depending on its source, either activation, i.e. enhancement of catalytic activity, and/or inactivation. Comparison of literature data allows us to conclude that pressures needed to induce substantial inactivation of PPO vary between 200 and 1000 MPa, depending on the enzyme origin and micro-environmental conditions such as medium composition, pH, presence of salts, sugars, . . . (Weemaes, 1998). Detailed kinetic studies on pressure inactivation of different fungal and plant PPO at room temperature were carried out by Weemaes (1998). At room temperature the following pressure stability ranking was observed: apple, grape, avocado, mushroom, pear, plum. In case of plum PPO, no inactivation was achieved by treatment at 900 MPa for 3 hours. Subsequently, inactivation of avocado PPO (pH 7) was investigated in a broad pressure (0.1 900 MPa) and temperature (25 80ºC) domain. In this case a strong antagonistic effect between low pressure and high temperature was observed, i.e. low pressure application protects the enzyme from thermal inactivation (Weemaes et al., 1998b). For many sources (e.g. apple, pear, potato, strawberry),
270
Thermal technologies in food processing
activation of PPO was observed at pressures below those needed for inactivation (Jolibert et al., 1994; Anese et al., 1995), which could be ascribed to conformational changes or to conversion of a latent enzyme form to an active form by release from the membrane (Asaka et al., 1994; Gomes and Ledward, 1996). In tissues, apparent activation of PPO may take place as a result of membrane alterations and decompartmentation of the enzyme and its substrate (Butz et al., 1994; Jolibert et al., 1994). PME from different fruits has been reported to be quite thermoresistant: temperatures between 80 and 95ºC are required to induce significant inactivation and in some cases a considerable percentage of remaining activity (up to 50%) after treatment was observed (Van den Broeck, 2000), which was ascribed to the presence of heat labile and heat stable PME isozymes (Versteeg et al., 1980; Wicker and Temelli, 1988; Van den Broeck et al., 2000b). Tomato PME on the other hand, displayed lower thermal stability and no thermostable fraction was observed (Van den Broeck et al., 2000a). Pressure stability has mainly been investigated for orange PME and to a lesser degree for grapefruit, guava and tomato PME. Threshold pressures for inactivation at room temperature of PME from different sources have been reported to vary largely from about 150 to 1200 MPa, depending on the origin and the medium in which the inactivation is carried out (Van den Broeck, 2000), i.e. inactivation occurs faster in acid medium and is protected by an increased amount of soluble solids (Ogawa et al., 1990). Most studies report only partial inactivation of PME, which is ascribed to the presence of isozymes with different pressure resistance, in accordance with the existence of a thermoresistant PME. Complete kinetic characterization of inactivation of PME from oranges in a broad pressure (0.1 800 MPa) and temperature (15 65ºC) domain revealed a slight antagonistic effect of low pressure and high temperature (Van den Broeck et al., 2000b). In contrast to thermal resistance, tomato PME was found to be much more pressure resistant than orange PME and an extreme antagonistic effect of high temperature and pressure was noted in this case. At 60ºC, a temperature where inactivation at atmospheric pressure occurs, pressure up to 700 MPa completely inhibited inactivation. At higher pressure, inactivation again occurred although the inactivation rate was still slower at 900 MPa as compared to atmospheric pressure (Crelier et al., 1995; Van den Broeck et al., 2000a). Because of the extreme pressure stability of tomato PME, also its catalytic activity under pressure has been investigated. At atmospheric pressure, optimal activity was found at 55ºC. Application of low pressure increased the activity of PME, which became maximal at a pressure of 100 200 MPa in combination with a temperature of 60 65ºC (Van den Broeck et al., 2000a). For LOX, thermal stability at atmospheric pressure largely varies with the enzyme source and medium, i.e. temperatures for inactivation range from 40 to 130ºC (Indrawati, 2000). As to pressure inactivation, detailed studies have been performed for tomato, soybean, green bean and pea LOX. In literature, threshold pressures for inactivation in a narrow range between 400 and 600 MPa have
Combined high pressure thermal treatment of foods 271 been reported (Heinisch et al., 1995; Ludikhuyze et al., 1998a; Tangwonchai et al., 1999; Indrawati et al., 1999, Indrawati, 2000). For soybean, green bean and pea LOX, complete kinetic characterization of the inactivation kinetics has been accomplished in a pressure-temperature domain from 0.1 to 650 MPa and from 10 to 80ºC. For green bean and peas it was noted that pressure stability of LOX decreased with increasing system complexity, i.e. inactivation occurred faster in situ (in the intact vegetable) as compared to a crude extract (Indrawati, 2000). For soybean LOX on the other hand, higher pressure stability was observed in milk as compared to buffer solution (Seyderhelm et al., 1996). Similarly for avocado PPO and orange PME, an antagonistic effect between low pressure and high temperature was noted for pea LOX. In the case of soybean and green bean LOX, an antagonistic effect between temperature lower than 30ºC and pressure higher than 500 MPa has been observed (Ludikhuyze et al., 1998b; Indrawati et al., 1999). POD, which is generally considered to be the most heat stable vegetable enzyme, is at least in some cases also extremely pressure resistant. In green beans, pressure treatment of 900 MPa merely induced slight inactivation at room temperature while a combination with elevated temperature enhanced the inactivating effect at 600 MPa (Quaglia et al., 1996). Contradictory results were found by Cano et al. (1997) who reported POD in strawberry pure´e and orange juice to be increasingly inactivated at room temperature with pressure up to 300 and 400 MPa respectively, whereas at higher pressure activity decreased again. At higher temperature (45ºC), a decrease in activity was found for all pressures (50 400 MPa).
14.4
Effect of high pressure on food structure and texture
In general pressures up to 350 MPa can be applied to plant systems without any major effect on overall texture and structure (Knorr, 1995b). Several studies revealed that pressure treatment of fruits and vegetables can cause both firming and softening (Basak and Ramaswamy, 1998), the effects being dependent on pressure level and pressurization time. In general, the softening curves revealed that texture changes due to pressure occurred in two phases, i.e. a sudden loss as a result of the pulse action of pressure followed by further loss or gradual recovery during pressure holding phase. At low pressure (100 MPa), instantaneous pressure softening was caused by compression of cellular structures without disruption, while at higher pressure (> 200 MPa) severe texture loss occurs due to rupture of cellular membranes and consequent loss of turgor pressure. During pressure holding time, the instantaneous texture loss can be gradually recovered and some products become even more firm than their fresh counterparts. In many cases, pressure treated vegetables do not soften during subsequent cooking, which is attributed to the action of PME that is only partially inactivated by pressure. Simultaneous disruption of cell structures allows interaction of the enzyme with the pectic substrate. Hence, the de-
272
Thermal technologies in food processing
esterified cell wall pectin can cross-link with divalent ions, leading to increased compactness of cellular structure. When various muscles are subjected to pressure, very firm and contracted raw meat is obtained. However, after cooking, pressurized meat is more tender and has higher moisture content and lower shear values than non-pressurized meat, mainly as a result of lower contraction and less drip loss. Sensory analysis showed pressurized meat to be less juicy, but more tender than control meat. The fact that tenderisation instead of toughening occurs, indicates a more severe damage of the sarcomere structure. In general it can be concluded that brief exposure of pre-rigor meat to pressures in the range of 100 200 MPa alters meat texture and is effective for tenderisation (Elgasim and Kennick, 1980; Ohmori et al., 1991). For post-rigor application of pressure, which is far more important from a commercial point of view, beneficial effects in counteracting toughening by cold-shortening are merely noted when combining pressure up to 150 MPa with heat (55 60ºC). This tenderisation effect was solely attributed to modification of myofibrillar structure but not of connective tissue (Cheftel and Culioli, 1997). Next to studies on overall texture and structure, some individual compounds which may be important for food structure engineering have been studied in more detail: starch, proteins and polysaccharides. In general it was found that high-pressure-induced physico-chemical changes in starch systems, such as loss of crystallinity, loss of anisotropic order, hydration and increase in viscosity, are very similar to those induced by heat while the rheological properties differ greatly. For high-amylose waxy maize starch no swelling of starch granules and no loss of birefringence was observed under severe conditions such as 900 MPa for 50 minutes. For barley starch (10%), the pressure-induced (550 MPa) gel was composed of closely packed, slightly swollen starch granules. In the heat-induced gel on the other hand, starch granules were more swollen, amylose and amylopectin were phase separated and amylose leaching occurred. For potato starch (10%), in contrast to heat treatment, treatment at pressure higher than 650 MPa resulted in a very rigid and quite elastic gel, as a result of an enormous swelling of the starch granules (Autio et al., 1999). From the limited studies on the effect of pressure on polysaccharides (hydrocolloids), it seemed that these compounds are not affected by pressure. Pectin, a heteropolysaccharide, was chemically not affected by pressure and its solubility did not change. At low temperature, pressure treatment at 400 MPa of a high-methoxyl pectin led to about tenfold increase in viscosity, whereas at higher temperature and pressure the effect was much lower (Michel et al., 1998). In contrast to the effect of pressure on pectin, protein modification by pressure in the context of texture engineering has been extensively studied, i.e. protein denaturation, aggregation, depolymerisation, gel formation. These effects result from the rupture of protein non-covalent interactions and the subsequent reformation of intra- and intermolecular bonds within or between protein molecules. Various studies demonstrated that differences in protein denaturation
Combined high pressure thermal treatment of foods 273 and aggregation induced by heat and pressure occur in food proteins (Funtenberger et al., 1997). All whey proteins, except BSA, showed more or less the same thermal denaturation behaviour, i.e. 10% denaturation after 75ºC for 5 minutes while for BSA 50% denaturation was noted under the same conditions. Pressure denaturation of whey proteins started at 200 MPa and complete denaturation was found at 800 MPa, depending on pH and temperature (Felipe et al., 1997). In this case, stability of BSA was higher than for the other proteins. As to gel formation, pressure-induced gels are weaker, less elastic and more exudative than heat induced gels (Cheftel and Dumay, 1996). Recently, some studies on the effect of pressure on proteins in presence of polysaccharides have been carried out (Tolstoguzov, 1998). Studies on the effects of pressure (0.1 800 MPa) and temperature (25 40ºC) on a binary system (12% whey protein/1.5% pectin) showed that a combination of pressure, temperature and pressure holding time can be used to induce phase separation in miscible pectin/WPI mixtures, the extent being dependent on the degree of denaturation. However, significant changes in texture could only be achieved when the protein is gelled, which implies that all WPI species must have a denaturation degree above 60%.
14.5 Effect of high pressure on sensorial and nutritional properties of foods Important characteristics of high quality foods are texture, colour, flavour and nutritional value. Although it is generally admitted that high pressure treatment only minimally affects overall food quality (Galazka and Ledward, 1995; Thakur and Nelson, 1998), an advantage ascribed to the fact that high pressure keeps covalent bonds intact, the effects on these individual quality characteristics have not been extensively studied hitherto. In the context of high pressure processing it is however important to know the effects of pressure on the chemical or biochemical reactions that can bring about undesirable changes in or deterioration of these quality attributes. For many fruit and vegetable products such as fruit jam, strawberries, tomato juice, guava pure´e, avocado pure´e and banana pure´e, high pressure treatment was noted to largely preserve fresh colour (Watanabe et al., 1991; Poretta et al., 1995; Donsi et al., 1996; Yen and Lin, 1996; Lopez-Malo et al., 1998). Brightness (L-colour value) and redness/greenness (a-colour value) of pressure treated products were found to be superior as compared to their thermally treated counterparts. However, during storage of guava and banana pure´e, green colour gradually decreased because of browning as a result of residual PPO activity (Lopez-Malo et al., 1998; Palou et al., 1999). Longest acceptability storage time was achieved by using high pressure, low pH and refrigerated storage. A detailed kinetic study regarding the combined effect of pressure and temperature on colour of broccoli juice revealed chlorophyll content and green colour (avalue) to be stable up to 4 hours treatment at 800 MPa and 40ºC. Only when high
274
Thermal technologies in food processing
pressure is combined with temperature higher than 50ºC, some colour changes were noted. Degradation of chlorophyll content was described by a first order model, with chlorophyll a being less pressure stable than chlorophyll b. On the other hand, loss of green colour was described by a consecutive step model because both conversion of chlorophyll to pheophytin and further conversion to pyropheophytin occurred (Van Loey et al., 1998; Weemaes et al., 1999). In contrast, many authors reported meat discoloration due to pressure processing as a result of (i) a whitening effect in the pressure range 200 350 MPa due to globin denaturation and heme displacement and (ii) a loss of red colour due to oxidation of ferrous myoglobin into ferric myoglobin above 400 MPa (Cheftel and Culioli, 1997). Hence, pressure processing of fresh red meat is not to be envisaged unless subsequent cooking is carried out, while for cured or white meats no serious colour problems are to be expected. For most fruit juices, the potentials of high pressure mainly arise from the fact that fresh flavour can be maintained during pressure treatment. Many authors reported trained sensory panel unable to differentiate between fresh and pressurized juice made from the same raw material (Ogawa et al., 1990; Watanabe et al., 1991; Bignon, 1996). For tomato and onions on the other hand some flavour defects due to pressure treatment were perceived: tomato showed a rancid taste while onions smelled less intensely and more like fried onions (Butz et al., 1994; Poretta et al., 1995). In the former case, the rancid flavour was attributed to a marked increase in n-hexanal, which is largely responsible for fresh tomato flavour in a concentration of 1 2 mg/kg. Higher concentrations impart the rancid flavour. For onions, pressure treatment was reported to diminish dipropylsulfide, a compound responsible for pungency and characteristic odour of fresh onions and to increase transpropenyldisulfide and 3,4dimethylthiophene concentrations leading to a flavour of braised or fried onions. The taste of pressurized meat products has, on some occasions, been reported to be sweeter than that of control meat (Cheftel and Culioli, 1997). Bignon (1996) observed that vitamin A, C, B1, B2, and E content of fruit and vegetable products is not significantly affected by pressure treatment in contrast to thermal treatment. Besides, in case of strawberries and guava pure´e, the decrease in vitamin C content during storage after pressure treatment (400 600 MPa/15 30 min) was found to be much lower as compared to the fresh products (Sancho et al., 1999). A more detailed kinetic study on pressuretemperature stability of ascorbic acid in buffer, orange juice and tomato juice was performed by Van den Broeck et al. (1998). They found only significant degradation of ascorbic acid when pressure of about 850 MPa was combined with temperatures between 60 and 80ºC, and more in tomato and orange juice than in buffer. Next to vitamins some minor studies on other health characteristics such as antimutagenicity, allergenicity and toxicity have been performed in recent years. Fruits and vegetables such as carrots, cauliflower, kohlrabi, leek and spinach are characterized by strong antimutagenic potencies, which were found to be sensitive to heat but not to pressure. For beet and tomatoes antimutagenic activity was affected, but only at very extreme
Combined high pressure thermal treatment of foods 275 conditions, i.e. 600 MPa/50ºC or 800 MPa/35ºC (Butz et al., 1997a). Selective elimination of -lactoglobulin, a major food allergen in milk could be achieved by hydrolysis with thermolysin at elevated pressure. At this high pressure, lactalbumin is quite resistant to hydrolysis due to the presence of four disulfide bridges while -lactoglobulin can be hydrolysed faster and more completely (Hayashi et al., 1987). Next to these beneficial effects, some drawbacks of high pressure processing have been observed. Pressure treatment (600 MPa/60ºC/ 3 min) of aspartame in diet chocolate milk induced 50% loss of the active substance while the non-sweet diketopiperazine, a toxic compound is formed (Butz et al., 1997b).
14.6 The use of integrated kinetic information in process design and optimisation In recent years, systematic kinetic studies on inactivation of some microorganisms and food spoiling enzymes have resulted in the development of pressuretemperature kinetic diagrams, which are two-dimensional diagrams indicating combinations of pressure and temperature resulting in the same inactivation rate and of mathematical models capable of describing the combined pressuretemperature dependence of this inactivation rate (Sonoike et al., 1992; Hashizume et al., 1995; Ludikhuyze et al., 1998b; Weemaes, 1998; Indrawati et al., 1999; Indrawati, 2000; Reyns et al., 2000; Van den Broeck, 2000). In the context of process optimisation, kinetic information on both food safety and quality aspects has been combined. A theoretical case study on high pressure process optimisation was elaborated by combining pressure-temperature kinetic diagrams of some food quality related enzymes (PPO, LOX, PME, ALP) with pressure-temperature kinetic diagrams of microbial inactivation (Sonoike et al., 1992; Hashizume et al., 1995; Reyns et al., 2000) on the one hand and of chlorophyll degradation on the other hand (Van Loey et al., 1998). Therefore pressure-temperature combinations resulting in a six log-unit reduction of microbial load and in 90% loss of enzyme activity and total chlorophyll content after a process time of 15 minutes are combined in Fig. 14.1. In contrast to the enzymes, the shapes of the pressure-temperature kinetic diagrams for the different vegetative microorganisms shown in this figure are similar. It can be seen that enzymes are generally more resistant than vegetative microorganisms with respect to pressure-temperature treatments. Hence, food quality related enzymes, rather than vegetative microorganisms, can become the critical issue in defining optimal pressure processes. Moreover, it can be seen that at pressure-temperature combinations resulting in sufficient inactivation of food spoiling enzymes and microorganisms, total chlorophyll content is only slightly affected. This supports the general statement that nutritional and sensorial quality is only minimally affected by pressure.
276
Thermal technologies in food processing
Fig. 14.1 Simulated pressure-temperature combinations resulting in six log-unit reduction of microorganisms, 90% reduction of enzyme activity and 90% of chlorophyll loss after a treatment time of 15 minutes: PPO (V); ALP (Q); BSAA (}); pea LOX in situ (); pea LOX in juice (X); green bean LOX in situ (O); green bean LOX in juice (U); soybean LOX (Z); PME (); total chlorophyll content ( ) yeast (D); Z. bailii (o); L. casei (—); E. coli (— —)
14.7
High pressure processing technology and products
In recent years, a number of manufacturers have developed ultra high pressure equipment for the food industry, each of them starting from their own disciplines. The two types of equipment used in food industry are batch systems, derived from cold isostatic processing, and semi-continuous systems. Using batch systems, both liquid and solid products can be processed, but these have to be pre-packed. New developments in batch high pressure processing include an internal pressure intensifier, a pre-stressed composite vessel, and fast opening and closing systems (Van den Berg et al., 1998). In-line systems can be applied to pumpable products only (for instance orange juice). The product is pumped in the pressure vessel and pressurized using a floating system, which separates the product from the pressure medium. After treatment the product is transferred to a surge vessel after which filling takes place. By coupling a number of pressure vessels, the energy saved in the pressurized vessel can be used to pressurize a second one, thus saving energy and process time. Because of these new developments in hardware and methods of treatment, the level of costs of HP treatment is decreasing. Further optimisation of equipment design should allow further reduction of the cost and to increase the efficiency. The renewed interest in high pressure processing together with the availability of pressure equipment that can meet the requirements imposed by
Combined high pressure thermal treatment of foods 277
Fig. 14.2
High pressure processed products on the Japanese market in 1995 (Rovere, 2001).
the food industry (i.e. high capacity as well as efficient cleaning and sanitation) and some public or governmental financing led to the appearance of a first generation of products on the Japanese market (Fig. 14.2), although there were only slight differences in quality between the traditionally and the high pressure processed foods. The application of the technique was here at its first stage and the resulting products were a compromise between Japanese food regulation (i.e. thermal treatment requirements), achievable operating pressure level and nutritional habits. Anyway some of them still are good examples of an emerging new processing technology. The following application of the process appeared in France in 1996: for the first time a company (ULTI) started to produce a HPP orange juice. Today the product is commercialized under the name Pampryl and the production is continuing. The company refers to the technology in the product labelling as follows: ‘High pressure preservation and the absence of pasteurisation guarantee exceptional preservation of flavour and vitamins of freshly pressed juices’. The next application emerged in the USA where a New Mexican company, Avomex Inc., started to market various avocado products for bulk and retail use in 1996. Guacamole became in a few years the best example of a HPP product being superior to any other traditional competitor. The refrigerated shelf life of this product is extended to 45 days without major quality losses. Its success is based on the heat sensitivity of the avocado pulp and on the low pressure stability of the polyphenoloxidase. The latest application of HPP was started in Spain in 1999. A plant was installed to process cooked ham in Espuna SA company (Olot-Spain) in order to reduce the re-contamination during slicing and to prolong its commercial shelf life to approximately 8 weeks (Fig. 14.3).
278
Thermal technologies in food processing
Fig. 14.3
14.8
High pressure processing of cooked ham in Espuna SA (Rovere, 2001).
Conclusive remarks
Where high pressure research in the past used to be rather qualitative and fragmentary, detailed quantitative studies on the effects of pressure on different food related aspects have become available throughout the last decade. In recent years, the use of a systematic kinetic approach in the area of enzyme (in)activation, has clearly revealed that thermal stability ranking of enzymes cannot be extrapolated to higher pressures. As to microorganisms, further research on inactivation kinetics and stability ranking is required, especially in view of defining, by analogy with thermal treatment, a suitable target microorganism for pressure processing. Indeed, kinetic information on microbial and enzyme inactivation, together with quantitative data on the effect of pressure on sensorial and nutritional quality attributes, is indispensable in the context of regulatory approval (FDA, EU Novel Food Regulation) and would facilitate a larger-scale industrial breakthrough of this new technology. However the pressure-processed products today available on the market clearly show that high pressure technology has to be seen as a ‘novel’ technology creating products with new and unique functional/quality properties, rather than an alternative or replacing technology for existing thermal treatments.
14.9
Acknowledgement
The authors would like to thank NFSR (National Fund for Scientific Research) and the European Commission (project FAIR-CT96-1175) for financial support.
14.10
References
ANESE M, NICOLI M C, DALL’AGLIO G, LERICI C R.
1995. Effect of high pressure treatments on peroxidase and polyphenoloxidase activities. J. Food Biochem. 18: 285–93.
Combined high pressure thermal treatment of foods 279 ASAKA M, AOYAMA Y, NAKANISHI R, HAYASHI R.
1994. Purification of a latent form of polyphenoloxidase from La France pear fruit and its pressureactivation. Biosci. Biotechnol. Biochem. 58: 1486–9. AUTIO K, STOLT M, SNAUWAERT J, HEREMANS K. 1999. Pressure-induced microstructural changes in starch suspensions. Lebensm. Wiss. Technol., submitted. BASAK S, RAMASWAMY H S. 1998. Effect of high pressure processing on texture of selected fruit and vegetables. J. Text. Stud. 29: 587–601. BENITO A, VENTOURA G, CASADEI M, ROBINSON T, MACKEY B. 1999. Variation in resistance of natural isolates of Eschericia coli O157 to high hydrostatic pressure, mild heat and other stresses. Appl. Environ. Microbiol. 65: 1564– 9. BERLIN D L, HERSON D S, HICKS D T, HOOVER D G. 1999. Response of pathogenic Vibrio species to high hydrostatic pressure. Appl. Environ. Microbiol. 65: 2776–80. BIGNON, J. 1996. Cold pasteurizers Hyperbar for the stabilization of fresh fruit juices. Fruit processing 2: 46–8. BRIDGMAN P W. 1914. The coagulation of albumen by pressure. J. Biol. Chem. 19: 511–12. BUTZ P, KOLLER D, TAUSCHER B. 1994. Ultra-high pressure processing of onions: chemical and sensory changes. Lebensm. -Wiss. u. -Technol. 27: 463–7. BUTZ P, EDENHARDER R, FISTER H, TAUSCHER B. 1997a. The influence of high pressure processing on antimutagenic activities of fruit and vegetable juices. Food Res. Int. 30(3/4): 287–91. BUTZ P, FERNANDEZ A, FISTER H, TAUSCHER B. 1997b. Influence of high hydrostatic pressure on aspartame: instability at neutral pH. J. Agric. Food Chem. 45: 302–3. CANO M P, HERNANDEZ A, DE ANCOS. 1997. High pressure and temperature effects on enzyme inactivation in strawberry and orange products. J. Food Sci. 62: 85–8. CHEFTEL J-C. 1992. Effects of high hydrostatic pressure on food constituents: an overview. In: High Pressure and Biotechnology, vol. 24. Balny, C., Hayashi, R., Heremans, K., Masson, P. (eds.), John Libbey Eurotext, Montrouge, 195–202. CHEFTEL J-C, DUMAY E. 1996. Effects of high pressure on dairy proteins: a review. In: High Pressure Bioscience and Biotechnology. Hayashi, R., Balny, C. (eds.), Elsevier, Amsterdam, 299–308. CHEFTEL J-C, CULIOLI J. 1997. Effects of high pressure on meat: a review. Meat Sci. 46(3): 211–36. CLOUSTON J G, WILLS P A. 1970. Kinetics of initiation of germination of Bacillus pumilis spores by hydrostatic pressure. J. Bacteriol. 103: 140–3. ^ ´ M-C, RENKEN A, RAETZ E. 1995. High pressure for the CRELIER S, TACHE inactivation of enzymes in food products. Poster presentation at the 9th World Congress on Food Science and Technology, July 31–August 4, Budapest, Hungary.
280
Thermal technologies in food processing
DONSI G, FERRARI G, DI MATTEO M.
1996. High pressure stabilization of orange juice: evaluation of the effects of process conditions. Ital. J. Food Sci. 2: 99–106. ELGASIM E A, KENNICK W H. 1980. Effect of pressurization of pre-rigor beef muscles on protein quality. J. Food Sci. 45: 1122–4. FELIPE X, CAPELLAS M, LAW A J R. 1997. Comparison of the effects of highpressure treatments and heat pasteurization on the whey protein in goat’s milk. J. Agric. Food Chem. 45: 627–31. FUNTENBERGER S, DUMAY E, CHEFTEL J-C. 1997. High pressure promotes lactoglobulin aggregation through SH/S-S interchange reactions. J. Agric. Food Chem. 45: 912–21. GALAZKA V B, LEDWARD D A. 1995. Developments in high pressure food processing. Food Technol. Int. Europe: 123–5. GARCIA-GRAELLS C, HAUBEN K, MICHIELS C W. 1998. High pressure inactivation and sublethal injury of pressure resistant Eschericia coli mutants in fruit juices. Appl. Environ. Microbiol. 64: 1566–8. GOMES M R A, LEDWARD D A. 1996. Effect of high-pressure treatments on the activity of some polyphenoloxidases. Food Chem. 56: 1–5. GOULD G W, STUBBS J M, KING W L. 1970. Structure and composition of resistant layers in bacterial spores coats. J. Gen. Microbiol. 60: 347–55. HASHIZUME C, KIMURA K, HAYASHI R. 1995. Kinetic analysis of yeast inactivation by high pressure treatment at low temperatures. Biosci. Biotech. Biochem. 59: 1455–8. HAUBEN K J, WUYTACK E Y, SOONTJENS C C, MICHIELS C W. 1997. High pressure transient sensitization of Eschericia coli to lysozyme and nisin by disruption of outer membrane permeability. J. Food Technol. 59: 350–5. HAYASHI R, KAWAMURA Y, KUKUGI S. 1987. Introduction of high pressure to food processing: preferential proteolysis of -lactoglobulin in milk. J. Food Sci. 52: 1107–108. HEINISCH O, KOWALSKI E, GOOSSENS K, FRANK J, HEREMANS K, LUDWIG H, TAUSCHER B. 1995. Pressure effects on the stability of lipoxygenase: Fourier transform-infrared spectroscopy (FT-IR) and enzyme activity studies. Z. Lebensm. Unters. Forsch. 201: 562–5. HEINZ V. 1997. Wirkung hoher hydrostatischer Dru ¨ cke auf das Absterbe- und Kiemungsverhalten sporenbildender Bakterien am Beispiel von Bacillus subtilis ATCC 9372. Ph.D. Dissertation, Technische Universita¨t, Berlin. HITE B H. 1899. The effect of pressure in the preservation of milk. West Va. Univ. Agr. Expt. Sta. Bull. 58: 15–35. HOOVER D G. 1993. Pressure effects on biological systems. Food Technol. 47(6): 150–5. INDRAWATI, VAN LOEY A M, LUDIKHUYZE L R, HENDRICKX M E. 1999. Soybean lipoxygenase inactivation by pressure at subzero and elevated temperatures. J. Agric. Food Chem. 47: 2468–74. INDRAWATI 2000. Lipoxygenase inactivation by high pressure treatment at subzero and elevated temperatures: a kinetic study. Ph.D. Dissertation.
Combined high pressure thermal treatment of foods 281 Katholieke Universiteit Leuven, Belgium. JOLIBERT F, TONELLO C, SAGEGH P, RAYMOND J.
1994. Les effets des hautes pressions sur la polyphe´nol oxydase des fruits. Bios Boissons 251: 27–35. KNORR D. 1995a. Hydrostatic pressure treatment of food: microbiology. In: New methods of food preservation. Gould, G.W. (Ed.), Blackie Academic and Professional, Glasgow, pp. 159–75. KNORR D. 1995b. High pressure effects on plant derived foods. In: High Pressure Processing of Foods. Ledward, D.A., Johnston, D.E., Earnshaw, R.G., Hasting, A.P.M. (eds.), Nottingham University Press, Loughborough, pp. 123–35. LARSON W P, HARTZELL T B, DIEHL H S. 1918. The effect of high pressure on bacteria. J. Infectious Dis. 22: 271–9. LOPEZ-MALO A, PALOU E, BARBOSA-CANOVAS G V, WELTI-CHANES J, SWANSON B G 1998. Polyphenoloxidase activity and color changes during storage of high hydrostatic pressure treated avocado pure´e. Food Res. Int. 31(8): 549–56. LOURENC ¸ O E J, DE SOUZA LEAO J, NEVES V A. 1990. Heat inactivation and kinetics of polyphenoloxidase from palmito (Euterpe edulis). J. Sci. Food Agric. 52: 249–59. LUDIKHUYZE L, INDRAWATI, VAN DER BROECK I, WEEMAES C, HENDRICKX M. 1998a. Effect of combined pressure and temperature on soybean lipoxygenase: I. Influence of extrinsic and intrinsic factors on isobaricisothermal inactivation kinetics. J. Agric. Food Chem. 46: 4074–80. LUDIKHUYZE L, INDRAWATI, VAN DER BROECK I, WEEMAES C, HENDRICKX M. 1998b. Effect of combined pressure and temperature on soybean lipoxygenase: II. Modeling inactivation kinetics under static and dynamic conditions. J. Agric. Food Chem. 46: 4081–86. MERTENS B. 1995. Hydrostatic pressure treatment of food: equipment and processing. In: New methods of food preservation. Gould, G.W. (ed.), Blackie Academic and Professional, Glasgow, 135–58. METRICK C, HOOVER D G, FARKAS D F. 1989. Effects of high hydrostatic pressure on heat-resistant and heat-sensitive strains of Salmonella. J. Food Sci. 54(6): 1547–64. MICHEL M, LESER M, SYRBE A, M F, C, I B, VON SCHACK M, WATZKE H J. 1998. Structure formation in protein-polysaccharide mixtures upon pressure treatment. In: Fresh novel foods by high pressure. Autio, K. (ed.), Julkaisija Utgivare publisher, Espoo, 69–82. MILLS G, EARNSHAW R, PATTERSON M F. 1998. Effects of high hydrostatic pressure on Clostridium sporogenes spores. Lett. Appl. Microbiol. 26: 227–30. MURELL W G, WILLS P A. 1977. Initiation of Bacillus spore germination by hydrostatic pressure: Effect of temperature. J. Bacteriol. 129: 1272–80. OGAWA H, FUKUHISA K, KUBO Y, FUKUMOTO H. 1990. Pressure inactivation of yeasts, moulds and pectinesterase in Satsuma mandarin juice: effects of juice concentration, pH and organic acids and comparison with heat sanitation. Agric. Biol. Chem. 54: 1219–25.
282
Thermal technologies in food processing
OHMORI T, SHIGEHISA T, TAJI S, HAYASHI R.
1991. Effect of high pressure on the protease activities in meat. Biol. Chem. 55(2): 357–61. PALOU E, LOPEZ-MALO A, BARBOSA-CANOVAS G V, WELTI-CHANES J, SWANSON B G. 1999. Polyphenoloxidase activity and color of blanched and high hydrostatic pressure treated banana pure´e. J. Food Sci. 64: 42–5. PARISH M E. 1998. Orange juice quality after treatment by thermal pasteurization and isostatic high pressure. Lebensm. -Wiss. u. -Technol. 31: 439–42. PORETTA S, BIRZI A, GHIZZONI C, VICINI E. 1995. Effects of ultra-high hydrostatic pressure treatments on the quality of tomato juice. Food Chem. 52: 35–41. QUAGLIA G B, GRAVINA R, PAPERI R, PAOLETTI F. 1996. Effect of high pressure treatments on peroxidase activity, ascorbic acid content and texture in green peas. Lebensm. -Wiss. u. -Technol. 29: 552–5. REYNS K M, SOONTJES C C, CORNELIS K, WEEMAES C A, HENDRICKX M E, MICHIELS C W. 2000. Kinetic analysis and modeling of combined pressuretemperature inactivation of the yeast Zygosaccharomyces bailii. J. Food Microbiol. 56: 199–210. ROBERTS C M, HOOVER D G. 1996. Sensitivity of Bacillus coagulans spores to combinations of high hydrostatic pressure, heat, acidity and nisin. J. Appl. Bacteriol. 81: 363–8. ROVERE P P, 2001. High pressure processing of real foods. In: High hydrostatic pressure treatment of foods, Hendrickx, M., Knorr, D. (eds.), Aspen Publ., Gaithersburg, in press. SALE A J H, GOULD G W, HAMILTON W A. 1970. Inactivation of bacterial spores by hydrostatic pressure. J. Gen. Microbiol. 60: 323–34. SANCHO F, LAMBERT Y, DEMAZEAU G, LARGETEAU A, BOUVIER J-M, NARBONNE JF. 1999. Effects of ultra-high hydrostatic pressure on hydrosoluble enzymes. J. Food Eng. 39: 247–53. SEYDERHELM I, BOGULAWSKI S, MICHAELIS G, KNORR D. 1996. Pressure induced inactivation of selected food enzymes. J. Food Sci. 61: 308–10. SHIGEHISA T, OHMORI T, SAITO A, TAJI S, HAYASHI R. 1991. Effects of high hydrostatic pressure on characteristics of pork slurries and inactivation of microorganisms associated with meat and meat products. Int. J. Food Microbiol. 12: 207–16. SMELT J P P M. 1998. Recent advances in microbiology of high pressure processing. Trends Food Sci. Technol. 9: 152–8. SOJKA B, LUDWIG H. 1994. Pressure-induced germination and inactivation of Bacillus subtilis spores. Pharm. Ind. 56: 660–3. SOJKA B, LUDWIG H. 1997. Release of dipicolinic acid and amino acids during high pressure treatment of Bacillus subtilis spores. Pharm. Ind. 59: 355–9. SONOIKE K, SETOYAMA T, KUMA Y, KOBAYASHI S. 1992. Effect of pressure and temperature on the death rates of Lactobacillus casei and Eschericia coli. In: High pressure and biotechnology, Colloque Inserm, vol. 224. Balny, C., Hayashi, R., Heremans, K., and Masson, P. (eds.), John Libbey Eurotext, Montrouge, 297–301. STYLES M F, HOOVER D G, FARKAS D F. 1991. Response of Listeria monocytogenes
Combined high pressure thermal treatment of foods 283 and Vibrio parahaemolyticus to high hydrostatic pressure. J. Food Sci. 56(5): 1404–7. TANGWONCHAI R, LEDWARD D A, AMES J M. 1999. Effect of high pressure on lipoxygenase activity in cherry tomatoes. In: Advances in high pressure bioscience and biotechnology. Ludwig, H. (ed.), Springer-Verlag, Berlin and Heidelberg, pp. 435–8. THAKUR B R, NELSON P E. 1998. High pressure processing and preservation of food. Food Rev. Int. 14(4): 427–47. TOLSTOGUZOV V B. 1998. Functional properties of protein-polysaccharide mixtures. In: Functional properties of food molecules. Hill, S.E., Ledward, D.A., Mitchell, J.R. (eds.). Aspen Publishers Inc., Gaithersburg. VAMOS-VIGYAZO L. 1981. Polyphenol oxidase and peroxidase in fruits and vegetables. Crit. Rev. Food Sci. Nutr. 5: 49–127. VAN DEN BERG R W, BARTELS P V, VAN SCHEPDAEL L J J M. 1998. High pressure apparatus. Patent WO97/43914. VAN DEN BROECK I, LUDIKHUYZE L, WEEMAES C, VAN LOEY A, HENDRICKX M. 1998. Kinetics of isobaric-isothermal degradation of L-ascorbic acid. J. Agric. Food Chem. 46: 2001–6. VAN DEN BROECK I. 2000. Kinetics of temperature and pressure inactivation of pectinesterase from oranges and tomatoes. Ph.D. Dissertation. Katholieke Universiteit Leuven, Belgium. VAN DEN BROECK I, LUDIKHUYZE L, VAN LOEY A, HENDRICKX M. 2000a. Effect of temperature and/or pressure on tomato pectinesterase activity. J. Agric. Food Chem. 48: 551–8. VAN DEN BROECK I, LUDIKHUYZE L R, VAN LOEY A M, HENDRICKX M E. 2000b. Inactivation of orange pectinesterase by combined high pressure and temperature treatments: a kinetic study. J. Agric. Food Chem., in press. VAN LOEY A, OOMS V, WEEMAES C, VAN DEN BROECK I, LUDIKHUYZE L, INDRAWATI, DENYS S, HENDRICKX M. 1998. Thermal and pressuretemperature degradation of chlorophyll in broccoli (brassica oleracea L. italica) juice: a kinetic study. J. Agric. Food Chem. 46: 5289–94. VERSTEEG C, ROMBOUTS F M, SPAANSEN C H, PILNIK W. 1980. Thermostability and orange juice cloud destabilizing properties of multiple pectinesterases from orange. J. Food Sci. 45: 969–71. WATANABE M, ARAI E, KUMENO K, HOMMA, K. 1991. A new method for producing non-heated jam sample: The use of freeze concentration and high pressure sterilisation. Biol. Chem. 55(8): 2175–6. WEEMAES C. 1998. Temperature and/or pressure inactivation of polyphenoloxidases for prevention of enzymatic browning in foods: a kinetic study. Ph.D. Dissertation, Katholieke Universiteit Leuven, Belgium. WEEMAES C, LUDIKHUYZE L R, VAN DEN BROECK I, HENDRICKX M E, TOBBACK P P. 1998a. Activity, electrophoretic characteristics and heat inactivation of polyphenoloxidases from apples, avocados, grapes, pears and plums. Lebensm. -Wiss. u.-Technol. 31: 44–9. WEEMAES C, LUDIKHUYZE L R, VAN DEN BROECK I, HENDRICKX M E. 1998b. Effect
284
Thermal technologies in food processing
of pH on pressure and thermal inactivation of avocado polyphenoloxidase: a kinetic study. J. Agric. Food Chem. 46: 2785–92. WEEMAES C, OOMS V, INDRAWATI, LUDIKHUYZE L, VAN DEN BROECK I, VAN LOEY A, HENDRICKX M. 1999. Pressure-temperature degradation of green color in broccoli juice. J. Food Sci. 64: 504–8. WICKER L, TEMELLI F. 1988. Heat inactivation of pectinesterase in orange juice pulp. J. Food Sci. 53: 162–4. WUYTACK E Y. 1999. Pressure-induced germination and inactivation of Bacillus subtilis spores. Ph.D. Dissertation. Katholieke Universiteit Leuven. YEMENICIOGLU A, OZKAN M, CEMEROGLU B. 1997. Heat inactivation kinetics of apple polyphenoloxidase and activation of its latent form. J. Food Sci. 62: 508–10. YEN G.-C, LIN H-T. 1996. Comparison of high pressure treatment and thermal pasteurization effects on the quality and shelf life of guava pure´e. Int. J. Food Sci. Technol., 31, 205–13.
Index
absolute pressure devices 52, 54–6 absorptivity 213 ABV Baker continuous ohmic heating process 242, 243, 246–7 advanced glycation end products (AGEs) 142, 150 agitation, shaking-type 25–6 aldose 139–42 alginate beads 85 alternating electrical current (AC) 242 Amadori compounds 139, 140, 150 amino acids 139, 145, 150 aminoketoses 139–40 ammonia 139 amylase activity 85–6, 87 antimutagenicity 274–5 antioxidants 145 Appert, N. 7 aroma 146–8 aseptic processing 30, 122 modelling 124–7 automatic loading systems 26 automatic temperature control 9, 10 Bacillus 268 amyloliquefaciens 85–6 sporothermodurans (HRS) 234–5 stearothermophilus 83, 84 bacteria 114, 132 effect of high pressure 267–8
microbiological spore methods for validation 83–5 modelling inactivation of 118, 120 safety and instrument design 69 see also spores baking 113 IR heating 223–4 Maillard reaction 152–3 microwave processing 196–7 RF heating 166, 173–4 batch high pressure systems 276 batch retorts 13–21, 26 basic retort cycle 7–11 batch rotary retorts 21 beef 254–6 bending 213, 214 -lactoglobulin 275 betaines 140 bimetallic strip thermometers 59 biochemical time and temperature integrators (TTIs) 85–7 Biot number 93, 115 biscuits 173–4 black body radiation 211, 212 body reflection 213–14 boundary conditions 116 Navier-Stokes equations 96 wave equations 183–4 Bourdon gauges 54, 58 Boussinesq approximation 94, 95 bread 152–3, 223–4
286
Index
browning 150 see also Maillard reaction Campden and Chorleywood Food Research Association (CCFRA) 89 cans/canning modelling and 116, 120–1 and retort systems 11 ring pull cans 25 capacitance 165 capacitive pressure measuring cells 55, 56, 58 capsule-type pressure measurement 54, 58 carboxymethyllysine 150 carcinogenicity 142–3 carrot 254–6 cassettes 26 cell structures, plant 271–2 chart recorder 9 ChefCad package 104–5 chlorophyll 273–4 circuit analogy 249 cleaning 33 Climate Change Levy 44 Clostridium 268 botulinum 79, 83, 84 sporogenes 84 coffee 153 cold spots ohmic heating 131 slowest heating point 77, 263–4 TD tests for retorts 76–7 colour effect of high pressure 273–4 Maillard reaction 145–8 colour distribution analysis 145 combination ovens 194–5 combined high pressure thermal treatment see high pressure processing combined microwave-air-dryers 198 commercial software 103–4 computational flow dynamics (CFD) 120–1 CFD model of an oven 106–8 computing power 131 concentric tubular heat exchanger 34–5 conduction 127 modelling Fourier equation 92–3, 114–15 solutions 117–18 conductivity see electrical conductivity
connections general design of process connections 69–70 types for RTDs 62–3 conservation equations 93–4 contactless heating 168 containers see packaging/containers contamination, risk of 238 continuous flow processing 29–48 direct heating 39–42 steam infusion 41–2 steam injection 40–1 future trends 44–5 holding section 42–4 indirect heating 30–9 plate heat exchangers 31–4, 44 scraped surface heat exchangers 37–9, 44–5 tubular heat exchangers 34–7, 44 modelling 122–7 sterilisation of foods containing particulates 124–7 sterilisation of liquid foods 122–4 ohmic heating 242, 243, 245, 246–7 continuous microwave applicators 195, 196 continous process IR oven 217, 218 continuous retorts 22–5, 26 contrast agents 252–4 control intelligent 27 IR heating 217–20 need for better control 75–6 ohmic heating 246 RF heating 168 temperature control 9, 10, 230–2 convection 114, 250 modelling 118–21 Navier-Stokes equations 93–7, 115–16 convergence error 103 conveyor belt microwave ovens 195, 196 cook-chill foods 88 cook-freeze foods 88 cooking 196–7 CookSim package 104 cooling, retort 10 co-rotating disc scraped surface heat exchanger (CDHE) 44–5 corrosion resistance 69 corrugation 31–2, 34 cost high-heat infusion systems 239 RF heating 168–9
Index count reduction method 84–5 crateless retorts 21 crust 152–3 CTemp 88 De Alwis-Fryer model 248–9 de-aeration 238 Debye relaxation 185–8 decimal reduction time 78 defrosting 166–7 meat 174–5 density 94 deoxyaldoketoses 139 deoxydiketoses 139–40 Department of Health 89 design high pressure process 275–6 mathematical modelling and 104–8 design of forced convection ovens 105–8 design of thermal processes 104–5 ohmic heating equipment 245–6 RF heating and 168 RTDs 61–5 deviations, process 27 diaphragm pressure devices 53, 54, 58 dielectric heating 164 see also microwave heating; radio frequency heating dielectric losses 165 origin of 185–8 dielectric properties 165, 185–9 measurement of in microwave frequency range 188–9 differential pressure devices 52, 55, 56 dipolar relaxation 185–8 direct heating 39–42, 45 steam infusion 41–2 steam injection 40–1 discretisation, numerical 97 double-separation plate heat exchanger 32–3 dry milk powder 152 drying infrared heating 215–16, 217 microwave drying 197–8 RF heating 166 E. coli 267 easy open cans 25 electrical conductivity microwave processing 180, 182–3, 184, 185–8
287
ohmic heating 242, 262–3 electrical pasteurisation 241–2 electrical pressure transducers 54–7, 58 electrical temperature devices installation conditions 66–8 RTDs 60, 61–5 thermocouples 59, 61 electrodes 246 electromagnetic spectrum 208, 209 EN standard 50 014 72 encapsulated spores/organisms 85 energy efficiency 167 environment 26 enzymes 268–71 "-pyrrolelysine 151 equipment costs 168–9 ethanol 187 exemplary solutions 181–4 experimental monitoring 247–60 explosion protection 71–2 extrusion cooking 153 F values 78, 85–6, 87, 114, 116, 122–4 feed-forward control 246 fermentation 139 50 RF technology 169–71, 175 filled thermal systems 59–60 finite difference method 97–9, 103, 104 finite element method 99–101, 103, 104–5, 119–20 finite volume method 101–3 5-hydroxymethylfurancarboxaldehyde (HMF) 139–40, 141, 150–1 Fizeau, A.H.L. 208 flash chamber 232, 236–7 flash cooling 232, 234, 236–7, 238 flavour high-heat technology and adding flavour components 238–9 high pressure and 274 Maillard reaction and 145–8, 149 flexibility 26, 238–9 floating end tubular heat exchanger 36 flow behaviour 43 food quality 124, 138–59 future trends 153–4 high pressure treatment and enzymes related to quality 268–71 food structure and texture 271–3 sensorial and nutritional quality 273–5 quality kinetics 78, 114 RF heating 167 see also Maillard reaction
288
Index
food safety 142–3 Food Safety (General Food Hygiene) Regulations 1995 42 food structure see structure, food force temperature devices 59–60 forced convection ovens 105–8, 121 Foucault, J.R.L. 208 fouling 33, 233 Fourier equation 92–3, 114–15 free amino acids 139, 145, 150 freeze drying 198 frequency microwave heating 164, 180 radio frequency heating 164 fringefield RF applicator 171–3 fructosyllysine 142, 150 frying 216 full water immersion retorts 15–18 furanone 140, 141 furosine 150 Galerkin finite element method 99–101 gas-chromatography-olfactometry (GCO) technique 147–8 gauge pressure devices 52, 54–6 General method 77–8 gentle heat treatment 232–3 glass containers 11 Gram bacteria 267 Gram bacteria 267 guacamole 276 ham, cooked 277–8 head transmitters 64–5 headspace 25, 120–1 heat exchangers 30–9 plate 31–4, 44 scraped surface 37–9, 44–5 tubular 34–7, 44 heat generation methods modelling 127–31 see also microwave heating; ohmic heating; radio frequency heating heat penetration (HP) testing 76, 77, 81–3 instrumentation 83 objectives 82–3 when required 82 heat resistant spores (HRS) 234–5 heat transfer electrical temperature devices 66 modelling 114–16 conduction 92–3, 114–15, 117–18 convection 93–7, 115–16, 118–21
infrared heating 211–12 microwave heating 200–2 Heissler chart 93 Herschel, W. 208 heterocyclic amines (HAs) 142–3, 144 high acid products 30 high-heat infusion 234–9 advantages and disadvantages 238–9 flash chamber 236–7 infusion chamber 235–6 problem of heat resistant spores 234–5 high pressure processing 3, 266–84 food quality related enzymes 268–71 food structure and texture 271–3 microorganisms 267–8 processing technology and products 276–8 sensorial and nutritional properties 273–5 use of integrated kinetic information in process design and optimization 275–6 high temperature short time (HTST) processing 25–6, 122, 241 HMF (5-hydroxymethylfurancarboxaldehyde) 139–40, 141, 150–1 holding section 42–4 holding time 43, 233 horizontal scraped surface heat exchangers 39 hot fill systems 30 household microwave ovens 194–5 housings general instrument design 70–2 temperature devices 65 hydrolock retorts 23 hydrostatic retorts 22–3 impedance matching network 170, 171 in-line mixers 44 in-line pH measurements 51 indirect heating 30–9, 40, 44–5 plate heat exchangers 31–4, 44 scraped surface heat exchangers 37–9, 44–5 tubular heat exchangers 34–7, 44 indirect UHT treatment 235, 236 inductive pressure measuring cells 56, 58 Industrial, Scientific and Medical (ISM) bands 164, 180 infrared (IR) heating 3, 208–28 applications 222–5 equipment 217–22
Index future trends 225–7 mathematical modelling 224–5 principle and uses 208–10 technologies 215–17 theories and IR properties 210–15 heat transfer 211–12 optical transfer 213–15 radiators 210 Infrared Teknik oven 220–1 infusion 39–40, 41–2, 45, 229–40 high-heat 234–9 instant 229–34 infusion chamber 230–2, 235–6, 237 ingress protection 71, 72 injection, steam 39–41 inoculated containers 84–5 inserts, sensor 63–4 instant infusion 229–34 advantages and disadvantages 232–4 infusion chamber 230–2 preheating 230 vacuum chamber 232 Institute for Thermal Processing Specialists (IFTPS) 89 instrumental monitoring 250–60 instrumentation design 51, 68–73 design factors 68 ingress and explosion protection 70–2 process connections 69–70 standardisation authorities 72–3 wetted parts 68–9 pressure measurement 9, 51, 52–7 retort systems 9–10 for TD and HP testing 83 temperature measurement 9–10, 51, 57–68, 83 integrated kinetic information 274–5 integrated lethal rate (F value) 78, 85–6, 87, 114, 116, 122–4 intelligent control 27 intrinsic warming 63 invasive temperature monitoring 250–1 Inversion Recovery – Flip Low Angle Snap Shot (IR-FLASH) 253 IP standard (IEC 60 529) 71 iteration 102–3 jars 11 jet impingement 195
289
K-" models 95–6 kinetic factor (z value) 78, 83–4 kinetic models for food processes 132–3 kinetic pressure-temperature diagrams 275–6 lactulosyllysine 150, 152 laminar flow 43 lamination nozzle 230 Larmor frequency 252 leakage radiation 195 lethality, integrated 78, 85–6, 87, 114, 116, 122–4 lipid oxidation 148–9 lipoxygenase (LOX) 269, 270–1 liquid foods 122–4 Listeria monocytogenes 267 long operating times 239 low acid products 30 low power RF technology 175 lysinoalanine (LAL) 143, 151 magnetic resonance imaging (MRI) 126–7, 251–60 microwave heating 129–30, 202 ohmic heating 263–4 potato, carrot and beef 254–6 whey gels 256–60 PRF method 254 principles and methodology 251–2 T1 method 253 magnetrons 178, 189–90 Maillard reaction (MR) 138–59 applications 151–3 controlling factors 149–50 food flavour and colour 145–8 food safety 142–3 future trends 153–4 importance 139–42 and lipid oxidation 148–9 methods of measurement 150–1 nutritional quality 143–5 manometer-type instruments 52–3, 58 mass transfer 113 microwave processing 200–2 master temperature indicator (MTI) 9–10 materials for wetted parts 69 mathematical modelling see modelling Maxwell’s equations 180–1 measurement Maillard reaction 150–1 need for better 75–6 pressure 9, 51, 52–7
290
Index
measurement continued temperature 9–10, 51, 57–68, 83 see also validation meat 254–6 defrosting 174–5 high pressure processing 272, 274 mechanical pressure instruments 53–4, 58 melanoidins 145 melting point 187 metallic sensors 56–7 microbiological kinetic models 114, 132 microorganisms see bacteria; spores microwave heating 2, 153–4, 164, 165, 178–207 applications 194–9 household ovens and product engineering 194–5 industrial ovens 195 industrial processes 195–9, 203 history 178–9 microwave applicators 191–3 modelling 127–8, 129–30, 199–202 outlook 203 physical principles 180–93 dielectric properties 185–9 frequency range 164, 180 Maxwell’s equations 180–1 wave equations and exemplary solutions 181–4 uses, advantages and disadvantages 179–80 verification 199–202 milk 151–2 milk powder 152 mixing 25, 44 mode stirrers 194 model substances 202 modelling 27, 113–37 continuous heating and cooling 122–7 design of forced convection ovens 105–8 developments 131–3 ease of solving models 131 kinetic models for food processes 132–3 new types of model 132 realistic physical properties 131–2 infrared heating 224–5 microwave heating 127–8, 129–30, 199–202 ohmic heating 127–8, 128–9, 130–1, 247–50, 260 predictive 27, 87–8, 260
processing of packed and solid foods 116–21 thermal processes 91–112, 113–16 applications 104–8 design of thermal processes 104–5 Fourier equation 92–3, 114–15 Navier-Stokes equations 93–7, 115–16 numerical methods 97–104 moisture level 51, 149 moisture levelling 168 molecular vibrations 213, 214 monosilicon sensors 57 monotube heat exchangers 34, 35 mounting position 67–8 multichannel heat exchanger 35–6 multi-mode microwave applicators 193 Multiscrape unit 39 multitube heat exchanger 35, 36 Navier-Stokes equations 93–7, 115–16 additional equations 96–7 conservation equations 93–4 initial and boundary conditions 96 turbulence 95–6 near field microwave applicators 191 near infrared heating (NIR) 209 see also infrared heating NEMA standard 250 71, 72 Newtonian fluids 93 non-enzymatic browning see Maillard reaction non-enzymatic glycosylation 142, 150 non-uniformity 118–21 nuclear magnetic resonance (NMR) spectroscopy 251–2 see also magnetic resonance imaging NumeriCAL 88 numerical methods 97–104, 201 commercial software 103–4 finite difference method 97–9, 103, 104 finite element method 99–101, 103, 104–5, 119–20 finite volume method 101–3 numerical discretisation 97 nutrient retention 123 nutritional quality effect of high pressure 273–5 Maillard reaction and 143–5 ohmic heating 2, 241–64 advantages 244
Index commercial equipment 242, 243, 246–7 current status of commercial uses 243–4 experimental and instrumental monitoring 250–60, 263–4 future research and development 260–4 history 241–2 modelling 127–8, 128–9, 130–1, 247–50, 262 principles 242–3 process and equipment 245–7 open-ended coaxial line dielectric probe 188, 189 operating cost 168–9 operating times, long 239 optical properties 213–15 optimization, process 275–6 orientation, particle 262–3 overpressure batch retorts 13–14, 15–21 P value 78, 85–6, 87, 114, 116, 122–4 packaging/containers 1–2 and HP tests 82 retort technology and 11, 26 packed foods 116–21 Pampryl 277 paneling 10 particulate foods 241 continous processing 43, 45 modelling sterilisation 124–7 particle orientation 262–3 particle size 43, 263 see also ohmic heating pasteurisation 113, 151–2, 241–2 continuous flow 30 microwave processing 198–9 RF heating 167 validation 79, 85–6, 87, 88 peaking 10 pectin 272 pectinmethylesterase (PME) 269, 270 penetration depth 214–15 permittivity 180, 182–3, 184, 185 peroxidase (POD) 269, 271 pH 150 in-line pH measurements 51 phase control 129 photons, quantum energy of 199 physical changes 113 physical properties, realistic 131–2 Picard iteration 102–3 piezoresistive sensors 57
291
Planck’s law 211 plant cell structure 271–2 plastic containers 11, 26 plate heat exchangers 31–4, 44 polyphenoloxidase (PPO) 269–70 polysaccharides 272 polysilicon sensors 57 Positron Emitting Particle Tracking (PEPT) 132–3 potable water 42 potato 254–6 pouches 11, 26 power density 165, 169, 184 power penetration 168 pre-air-intermittent microwave vacuum drying 198 precool stage 15 predictive modelling 27, 87–8, 262 preheating 230 preservation 113 pressure changes in tubular heat exchangers 36–7 high pressure processing see high pressure processing requirements for a retort 13 pressure-temperature kinetic diagrams 275–6 pressure measurement 9, 51, 52–7 electrical pressure transducers 54–7 manometer-type instruments 52–3 mechanical-type instruments 53–4 operating conditions 57, 58 process connections 69–70 process design 104–5, 275–6 process deviations 27 process flavours 147 process lines 167 process optimization 275–6 process parameters 262–3 process timer 10 process validation see validation process value 77–9 General method 77–8 setting the target process value 78–9 product cold point 77 product engineering 194–5 product specifications 262–3 products high pressure processing 276–8 and HP testing 82 range 239 proline 152
292
Index
proteins 142, 145 effect of high pressure 271–2 proton resonance frequency (PRF) shift 252, 254 Pt100 sensing elements 61–2 puffing 198 pulsed electric fields 3 pyrazines 147 pyrrole aldehydes 140 quality, food see food quality quality kinetics 78, 114 quantum energy 199 radiation 114, 201–2, 215 radiators 210, 217–22 radio frequency (RF) heating 2, 163–77 advantages 167–8 applications 166–7, 176 basic principles 163–6 case studies 173–5 disadvantages 168–9 future trends 175–6 technologies 169–73, 175 conventional RF equipment 169, 170 50 equipment 169–71, 175 RF applicators 171–3 raining water retorts 19–21 Raytheon Co. 178–9 realistic physical properties 131–2 Redox measurements 51 reducing sugars 139–42 reel and spiral retorts 23–5 reflection, IR 213–15 reflectors 217, 218 regeneration system 33–4 regular reflection 213–14 relaxation, Debye 185–8 remote transmitters 64–5 residence time distribution (RTD) 126 resistance temperature detectors (RTDs) 59, 60, 61–5 head or remote transmitter 64–5 housing 65 intrinsic warming 63 sensing elements 61–2 sensor inserts 63–4 thermowells 64, 65, 66 type of connection 62–3 resistive pressure measuring cells 54–5, 56–7, 58 resonator methods 188–9 response times 66–7
retort technology 7–28 applications of TTIs 86–7 basic retort cycle 7–11 cold point 76–7 container selection 11 future trends 25–7 influence of heating medium on performance 13–25 batch retorts 13–21 continuous retorts 22–5 selection of a retort 11–13 and validation 82 retroaldol cleavage 140–1 Reynolds Averaged Navier-Stokes (RANS) equations 95–6 Reynolds number 43 ring pull cans 25 risk of contamination 238 rotary batch retorts 21 rotary pump 232 runaway heating 129 Salmonella 267 salt solution 187 sanitary couplings 70 Sastry-Palaniappan model 249–50 saturated steam retorts 7–11, 13, 14 scraped surface heat exchangers 37–9, 44–5 self-contained logging units 88 self-heating 63 semi-continuous systems 276 sensor inserts 63–4 sensing elements 61–2 sensory properties 272–4 see also aroma; colour; flavour; texture shaking retort system 25–6 simulation see modelling single-mode microwave applicators 191–3 slip velocity 125–6 short-wave IR heating 209 slowest heating point 77, 263–4 sniffing port 148 software, commercial 103–4 solid foods 116–21 specifications, product 261–2 Spencer, P. 178–9 spin echo imaging sequences 253 spin-lattice relaxation time constant (T1) 252, 253 spores heat resistant 234–5
Index high pressure processing 268 microbiological spore methods for validation 83–5 see also bacteria sprayed water retorts 19–21 staggered throughfield applicator 172, 173 stainless steels 69 standardisation authorities 72–3 starch 272 steam/air retorts 15, 16 steam infusion see infusion steam injection 39–41 steam quality 42 steam retorts, saturated 7–11, 13, 14 Stefan-Boltzmann equation 211 sterilisation 79, 113, 151–2 microbiological spore methods 84–5 microwave processing 198–9 modelling 104, 122–7 foods containing particulates 124–7 liquid foods 122–4 RF heating 167 STERILMATE software package 104 stochastic finite element methods 101 Strecker degradation 141–2 stretching 213, 214 structure, food 113 effect of high pressure 271–3 sugars 150 dehydration/fragmentation products 147 reducing 139–42 susceptors 194 T1 (spin-lattice relaxation time constant) 252, 253 target process value 78–9 temperature control 9, 10, 230–2 establishing for validation 77 holding section 42–3, 43–4 and Maillard reaction 149 pressure-temperature kinetic diagrams 275–6 requirements for a retort 12–13 temperature-time profiles 235, 236 time and temperature integrators (TTIs) 85–7 temperature distribution (TD) 76–7 modelling 121 microwave processing 199–202 monitoring ohmic heating 250–60, 263–4
293
testing 79–81, 87 instrumentation 83 objectives 80–1 when required 80 temperature measurement 9–10, 51, 57–68, 83 electrical devices 60–5 RTDs 60, 61–5 thermocouples 59, 61 force devices 59–60 installation conditions 66–8 tempering 197 tenderisation 272 texture 271–3 thermal centre 77 thermal runaway 129 thermal transport equations 92–7, 114–16 thermocouples 59, 61 thermometers 59–60 thermophysical parameters 92 thermowells 64, 65, 66 thick-film strain gauges 57 thin-film strain gauges 57 throughfield RF applicator 171, 172 throughput 167 time holding time 43, 233 long operating times 239 response times for temperature measurement 66–7 scheduled process time 77 time and temperature integrators (TTIs) 85–7 timer, process 10 toxicological safety 69 trace moisture sensors 51 transistor technology 175 transitional flow 43 transmitters, sensor 64–5 trays 11 TRIAB infrared oven 221–2 tubular heat exchangers 34–7, 44 turbidity measurements 51 turbulence 43, 95–6 ‘12D cook’ 79 UHT processing 29, 30, 42, 151–2 indirect 235, 236 see also continuous flow processing umbrella effects 19 vacuum chamber 232, 236–7 vacuum and microwave drying 198
294
Index
validation 75–90 biochemical time and temperature integrators 85–7 future trends 87–8 heat penetration testing 76, 77, 81–3 methods 76–9 process establishment methods 77–8 setting the target process value 78–9 microbiological spore methods 83–5 need for better measurement and control 75–6 ohmic heating 250–60 temperature distribution testing 76–7, 79–81, 83, 87 Vario hydrostatic retort system 26 vegetative microorganisms 267 vent test 81 venting 8–9 verification 199–202
vertical scraped surface heat exchangers 39 vibrations, molecular 213, 214 viscosity 120 vitamins 274 water dielectric properties and microwave processing 186, 187 potable 42 water activity 51, 149 wave equations 181–4 waveguides 190–1, 192 well manometers 53 wetted parts 68–9 whey gels 256–8 Wien’s law 211 z value 78, 83–4