Modern Drying Technology, Process Intensification [5, Volume 5 ed.] 3527315608, 9783527315604

Citation preview

Edited by Evangelos Tsotsas and Arun S. Mujumdar Modern Drying Technology

Modern Drying Technology Edited by E. Tsotsas and A. Mujumdar

Other Volumes Volume 1: Computational Tools at Different Scales ISBN: 978-3-527-31556-7

Volume 2: Experimental Techniques ISBN: 978-3-527-31557-4

Volume 3: Product Quality and Formulation ISBN: 978-3-527-31558-1

Volume 4: Energy Savings ISBN: 978-3-527-31559-8

Modern Drying Technology Set (Volumes 1 – 5) ISBN: 978-3-527-31554-3

Edited by Evangelos Tsotsas and Arun S. Mujumdar

Modern Drying Technology Volume 5: Process Intensification

The Editors: Prof. Evangelos Tsotsas Otto von Guericke University Thermal Process Engineering Universitätsplatz 2 39106 Magdeburg Germany Prof. Arun S. Mujumdar National University of Singapore Mechanical Engineering/Block EA 07-0 9 Engineering Drive 1 Singapore 117576 Singapore

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at . # 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Boschstr. 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: ePDF ISBN: ePub ISBN: Mobi ISBN: oBook ISBN: Cover Design Typesetting

978-3-527-31560-4 978-3-527-63171-1 978-3-527-65140-5 978-3-527-65139-9 978-3-527-63170-4 Adam Design, Weinheim Thomson Digital, Noida, India

Printing and Binding

Strauss GmbH, Mörlenbach

Printed on acid-free paper Printed in the Federal Republic of Germany

jV

Contents Series Preface XI Preface of Volume 5 XV List of Contributors XIX Recommended Notation XXIII EFCE Working Party on Drying: Address List XXIX 1 1.1 1.2 1.3 1.3.1 1.3.2 1.3.3 1.3.4 1.4 1.5 1.6

2 2.1 2.2 2.2.1 2.2.2 2.2.3 2.3 2.3.1 2.3.2 2.4 2.5

Impinging Jet Drying 1 Eckehard Specht Application 1 Single Nozzle 4 Nozzle Fields 7 Arrays of Single Nozzles 7 Hole Channels 12 Perforated Plates 13 Nozzles for Cylindrical Bodies 14 Summary of the Nusselt Functions 16 Design of Nozzle Field 17 Conclusion 23 References 24 Pulse Combustion Drying 27 Ireneusz Zbicinski, Tadeusz Kudra, and Xiangdong Liu Principle of Pulse Combustion 27 Pulse Combustors: Design and Operation 32 Pulse Combustors with Mechanical Valves 32 Pulse Combustors with Aerodynamic Valves 34 Frequency-Tunable Pulsed Combustors 35 Aerodynamics, Heat and Mass Transfer 36 Atomization 37 Heat and Mass Transfer 38 Modeling of Pulse Combustion Drying 42 Pulse Combustion in Drying 48 References 53

VI

j Contents 3 3.1 3.2 3.3 3.4 3.5 3.6

4

4.1 4.2 4.2.1 4.2.2 4.2.3 4.2.4 4.3 4.3.1 4.3.2 4.3.3 4.4 4.5

5

5.1 5.2 5.3 5.3.1 5.3.2 5.4 5.5 5.5.1 5.5.2 5.6 5.7

Superheated Steam Drying of Foods and Biomaterials 57 Sakamon Devahastin and Arun S. Mujumdar Introduction 57 Principle of Superheated Steam Drying (SSD) 58 Atmospheric-Pressure Superheated Steam Drying 61 Low-Pressure Superheated Steam Drying (LPSSD) 69 Application of LPSSD to Improve the Quality of Foods and Biomaterials 76 Concluding Remarks 82 References 83 Intensification of Fluidized-Bed Processes for Drying and Formulation 85 Evangelos Tsotsas, Stefan Heinrich, Michael Jacob, Mirko Peglow, and Lothar M€orl Introduction 85 Intensification by Apparatus and Flow Design 86 Different Types of Spouted Bed 86 Operating Characteristics of Spouted Beds 93 Mass and Heat Transfer in ProCell Units 100 Discrete Particle Modeling 107 Intensification by Contact Heating 112 General Principle 112 Main Effects and Influences 114 Further Remarks on Modeling 121 Further Methods of Intensification 126 Conclusion 127 References 128 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries 131 Roberto Pisano, Davide Fissore, and Antonello A. Barresi Introduction 131 Exergetic Analysis (and Optimization) of the Freeze-Drying Process 133 Process Intensification in Vacuum Freeze-Drying of Liquids 139 Regulation of Nucleation Temperature During Freezing 140 Use of Organic Solvents and Cosolvents 144 Atmospheric Freeze-Drying 146 Use of Combined Technologies for Drying Heat-Sensitive Products 150 Microwave-Assisted Drying 150 Ultrasound-Assisted Drying 152 Continuous Freeze-Drying 154 Conclusions 155 References 157

Contents

6 6.1 6.2 6.3 6.3.1 6.3.2 6.4 6.5

7

7.1 7.2 7.2.1 7.2.2 7.2.2.1 7.2.2.2 7.2.3 7.2.3.1 7.2.3.2 7.3 7.3.1 7.3.2 7.3.3 7.3.4 7.4 7.4.1 7.4.2 7.4.3 7.4.4 7.5

8

8.1 8.2

Drying of Foamed Materials 163 Ireneusz Zbicinski, Julia Rabaeva, and Artur Lewandowski Introduction 163 Foam Properties 164 Foam Spray Drying 167 Processing Principles 167 Final Product Properties 172 Foam-Mat Drying 181 Summary 187 References 188 Process-Induced Minimization of Mass Transfer Barriers for Improved Drying 191 Henry J€ager, Katharina Sch€ossler, and Dietrich Knorr Introduction 191 Structural Characterization of Plant Raw Materials and Impact of PEF and Ultrasound 192 Methods for Analysis of Tissue Structure and Quantification of Cell Damage 192 PEF: Principles and Impact on Plant Tissue Structure 195 Introduction to PEF Technology 195 PEF: Impact on Plant Tissue Structure 196 Ultrasound: Principles and Impact on Plant Tissue Structure 199 Introduction to Ultrasound Technology 199 Ultrasound: Impact on Plant Tissue Structure 200 Pulsed Electric Field (PEF) Application as a Pretreatment 204 Osmotic Dehydration 205 Air Drying 206 Impact of PEF on Freezing and Freeze-Drying Behavior of Raw Materials 208 Quality Characteristics Affected by PEF Pretreatment 211 Contact Ultrasound for Combined Drying Processes 216 Ultrasound in Osmotic Dehydration 217 Contact Ultrasound in Air Drying 218 Contact Ultrasound in Freeze-Drying 221 Quality Characteristics Affected by Ultrasound-Combined Drying Processes 224 Conclusion 226 References 230 Drying Assisted by Power Ultrasound 237 Juan Andres Carcel, Jose Vicente García-Perez, Enrique Riera, Carmen Rossello, and Antonio Mulet Introduction 237 Ultrasound 239

jVII

VIII

j Contents 8.2.1 8.2.1.1 8.2.1.2 8.2.1.3 8.2.1.4 8.2.2 8.3 8.3.1 8.3.2 8.3.3 8.3.3.1 8.3.3.2 8.4 8.4.1 8.4.1.1 8.4.1.2 8.4.1.3 8.4.2 8.4.3 8.5

9

9.1 9.2 9.2.1 9.2.2 9.2.3 9.2.4 9.3 9.3.1 9.3.2 9.3.3 9.3.4 9.4 9.4.1 9.4.2 9.4.2.1 9.4.2.2

Ultrasound Waves 239 Power 239 Frequency 240 Attenuation 240 Acoustic Impedance 240 Effects of Ultrasound on Mass Transfer 241 Ultrasonic Equipment 242 Source of Energy 243 Transducers 243 Application Systems 245 Treatments in Liquid Media 245 Treatments in Gas Media 247 Influence of the Main Process Variables on Drying Intensification by Ultrasound 250 Ultrasonic Power Applied 250 Ultrasonic Field Measurements 251 Ultrasonic Intensity and Effects 252 Influence of the Characteristics of the Medium on Ultrasonic Intensity 258 Drying Air Temperature 263 Ultrasound–Sample Interaction 266 Conclusions 272 References 273 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality 279 Yingqiang Wang, Min Zhang, and Arun S. Mujumdar Introduction 279 Microwave-Assisted Drying of Foods 281 Basic Principles of Microwave-Assisted Drying 281 Energy Absorption by Products During Dielectric Heating 283 Dielectric Properties 283 Penetration Depth 285 Microwave-Assisted Drying Equipment 285 Microwave-Assisted Convective Drying Equipment 286 Microwave-Assisted Vacuum Drying Equipment 287 Microwave-Assisted Freeze-Drying Equipment 290 Microwave-Assisted Spouted Bed Drying Equipment 291 Microwave-Assisted Drying Process 292 Moisture Loss 293 Temperature Distributions 295 Temperature Variations at Fixed Levels of Microwave Power 296 Temperature Variations at Variable Microwave Power without Controlling Temperature 298

Contents

9.4.2.3 9.4.3 9.4.4 9.4.5 9.4.6 9.5 9.5.1 9.5.2 9.6

10 10.1 10.2 10.2.1 10.2.2 10.2.3 10.3 10.3.1 10.3.2 10.3.3 10.3.4 10.4 10.4.1 10.4.2 10.4.3 10.5 10.5.1 10.5.2 10.6 10.7 10.7.1 10.7.2 10.7.3 10.7.4 10.8

Temperature Change with Time-Adjusted Power in Feedback Temperature Control 299 Energy Consumption 299 Dielectric Breakdown 302 Changes in Dielectric Properties 304 Quality Changes in Food during Microwave-Assisted Drying 305 Microwave-Assisted Drying Process Control and Optimal Operation 308 Factors Controlling Microwave-Assisted Drying Processes 308 Optimal Operation Strategy 308 Concluding Remarks 310 References 312 Infrared Drying 317 German Efremov Introduction 317 Radiation Heat Transfer 318 General Principles 318 Reflection, Absorption, and Transmission 319 Infrared Spectrum 321 Classification, Research, and Applications of Radiation Drying 323 Classification 323 Solar Drying 325 Infrared Drying 326 Catalytic Infrared Drying 329 Types of Radiators 332 General Considerations 332 Electric Radiators 333 Gas-Heated IR Radiators 335 Interaction between Matter and Infrared Radiation 337 General Relationships 337 Radiation Properties of Materials 339 Kinetics of Infrared Drying 342 Infrared Drying Combined with other Types of Drying 345 IR and Convective Drying 346 IR and Microwave Drying 347 IR and Freeze-Drying 348 IR with other Types of Drying 348 Conclusions 351 References 352 Index 357

jIX

jXI

Series Preface The present series is dedicated to drying, that is, to the process of removing moisture from solids. Drying has been conducted empirically since the dawn of the human race. In traditional scientific terms it is a unit operation in chemical engineering. The reason for the continuing interest in drying and, hence, the motivation for the series, concerns the challenges and opportunities. A permanent challenge is connected to the sheer amount and value of products that must be dried – either to attain their functionalities, or because moisture would damage the material during subsequent processing and storage, or simply because customers are not willing to pay for water. This comprises almost every material used in solid form, from foods to pharmaceuticals, from minerals to detergents, from polymers to paper. Raw materials and commodities with a low price per kilogram, but with extremely high production rates, and also highly formulated, rather rare but very expensive specialties have to be dried. This permanent demand is accompanied by the challenge of sustainable development providing welfare, or at least a decent living standard, to a still-growing humanity. On the other hand, opportunities emerge for drying, as well as for any other aspect of science or living, from either the incremental or disruptive development of available tools. This duality is reflected in the structure of the book series, which is planned for five volumes in total, namely: Volume Volume Volume Volume Volume

1: 2: 3: 4: 5:

Computational tools at different scales Experimental techniques Product quality and formulation Energy savings Process intensification.

As the titles indicate, we start with the opportunities in terms of modern computational and experimental tools in Volumes 1 and 2, respectively. How these opportunities can be used in fulfilling the challenges, in creating better and new products, in reducing the consumption of energy, in significantly improving existing or introducing new processes will be discussed in Volumes 3, 4 and 5. In this sense, the first two volumes of the series will be driven by science; the last

XII

j Series Preface three will try to show how engineering science and technology can be translated into progress. In total, the series is designed to have both common aspects with and essential differences from an extended textbook or a handbook. Textbooks and handbooks usually refer to well-established knowledge, prepared and organized either for learning or for application in practice, respectively. On the contrary, the ambition of the present series is to move at the frontier of “modern drying technology”, describing things that have recently emerged, mapping things that are about to emerge, and also anticipating some things that may or should emerge in the near future. Consequently, the series is much closer to research than textbooks or handbooks can be. On the other hand, it was never intended as an anthology of research papers or keynotes – this segment being well covered by periodicals and conference proceedings. Therefore, our continuing effort will be to stay as close as possible to a textbook in terms of understandable presentation and as close as possible to a handbook in terms of applicability. Another feature in common with an extended textbook or a handbook is the rather complete coverage of the topic by the entire series. Certainly, not every volume or chapter will be equally interesting for every reader, but we do hope that several chapters and volumes will be of value for graduate students, for researchers who are young in age or thinking, and for practitioners from industries that are manufacturing or using drying equipment. We also hope that the readers and owners of the entire series will have a comprehensive access not to all, but to many significant recent advances in drying science and technology. Such readers will quickly realize that modern drying technology is quite interdisciplinary, profiting greatly from other branches of engineering and science. In the opposite direction, not only chemical engineers, but also people from food, mechanical, environmental or medical engineering, material science, applied chemistry or physics, computing and mathematics may find one or the other interesting and useful results or ideas in the series. The mentioned interdisciplinary approach implies that drying experts are keen to abandon the traditional chemical engineering concept of unit operations for the sake of a less rigid and more creative canon. However, they have difficulties of identification with just one of the two new major trends in chemical engineering, namely process-systems engineering or product engineering. Efficient drying can be completely valueless in a process system that is not efficiently tuned as a whole, while efficient processing is certainly valueless if it does not fulfill the demands of the market (the customer) regarding the properties of the product. There are few topics more appropriate in order to demonstrate the necessity of simultaneous treatment of product and process quality than drying. The series will try to work out chances that emerge from this crossroads position. One further objective is to motivate readers in putting together modules (chapters from different volumes) relevant to their interests, creating in this manner individual, task-oriented threads trough the series. An example of one such thematic thread set by the editors refers to simultaneous particle formation and drying, with a focus

Series Preface

on spray fluidized beds. From the point of view of process-systems engineering, this is process integration – several “unit operations” take place in the same equipment. On the other hand, it is product engineering, creating structures – in many cases nanostructures – that correlate with the desired application properties. Such properties are distributed over the ensemble (population) of particles, so that it is necessary to discuss mathematical methods (population balances) and numerical tools able to resolve the respective distributions in one chapter of Volume 1. Measuring techniques providing access to properties and states of the particle system will be treated in one chapter of Volume 2. In Volume 3, we will attempt to combine the previously introduced theoretical and experimental tools with the goal of product design. Finally, important issues of energy consumption and process intensification will appear in chapters of Volumes 4 and 5. Our hope is that some thematic combinations we have not even thought about in our choice of contents will arise in a similar way. As the present series is a series of edited books, it can not be as uniform in either writing style or notation as good textbooks are. In the case of notation, a list of symbols has been developed and will be printed in the beginning of every volume. This list is not rigid but foresees options, at least partially accounting for the habits in different parts of the world. It has been recently adopted as a recommendation by the Working Party on Drying of the European Federation of Chemical Engineering (EFCE). However, the opportunity of placing short lists of additional or deviant symbols at the end of every chapter has been given to all authors. The symbols used are also explained in the text of every chapter, so that we do not expect any serious difficulties in reading and understanding. The above indicates that the clear priority in the edited series was not in uniformity of style, but in the quality of contents that are very close to current international research from academia and, where possible, also from industry. Not every potentially interesting topic is included in the series, and not every excellent researcher working on drying contributes to it. However, we are very confident about the excellence of all research groups that we were able to gather together, and we are very grateful for the good cooperation with all chapter authors. The quality of the series as a whole is set mainly by them; the success of the series will primarily be theirs. We would also like to express our acknowledgments to the team of Wiley-VCH who have done a great job in supporting the series from the first idea to realization. Furthermore, our thanks go to Mrs Nicolle Degen for her additional work, and to our families for their tolerance and continuing support. Last but not least, we are grateful to the members of the Working Party on Drying of the EFCE for various reasons. First, the idea about the series came up during the annual technical and business meeting of the working party 2005 in Paris. Secondly, many chapter authors could be recruited among its members. Finally, the Working Party continues to serve as a panel for discussion, checking and readjustment of our conceptions about the series. The list of the members of the working party with their affiliations is included in every volume of the series in the sense of acknowledgment,

jXIII

XIV

j Series Preface but also in order to promote networking and to provide access to national working parties, groups and individuals. The present edited books are complementary to the regular activities of the EFCE Working Party on Drying, as they are also complementary to various other regular activities of the international drying community, including well-known periodicals, handbooks, and the International Drying Symposia. December 2006

Evangelos Tsotsas Arun S. Mujumdar

jXV

Preface of Volume 5 Volume 5 of “Modern Drying Technology” is dedicated to “Process intensification”. This is a natural conclusion for the series. “Computational tools at different scales” and “Experimental techniques”, presented in Vol. 1 and Vol. 2, respectively, were used to discuss “Product quality and formulation” in Vol. 3, and “Energy savings” in Vol. 4. Now the goal is not as specific as in Vol. 4, but more general and quite ambitious; namely, to use as intensive drying processes as possible in order to reduce the drying time and hence the equipment size. Insights from all previous volumes of the series must be implemented and applied to this purpose, leading to the following ten chapters of Vol. 5: Chapter 1: Chapter 2: Chapter 3: Chapter 4: Chapter 5:

Impinging jet drying Pulse combustion drying Superheated steam drying of foods and biomaterials Intensification of fluidized-bed processes for drying and formulation Intensification of freeze-drying for the pharmaceutical and food industries Chapter 6: Drying of foamed materials Chapter 7: Process-induced minimization of mass transfer barriers for improved drying Chapter 8: Drying assisted by power ultrasound Chapter 9: Microwave-assisted drying of foods – Equipment, process and product quality Chapter 10: Infrared drying Frequent mention of foods, biomaterials and pharmaceuticals in the list of contents shows that process intensification is not a stand-alone perspective, but a challenge which usually must be addressed under serious constraints set by the quality requirements of valuable products that might suffer damage during processing. On the other hand, the list of contents of this final volume is longer than the lists of previous volumes in this series, indicating that a large variety of approaches and methods can lead to process intensification in practice, and that the international drying community is continually and persistently working on their further development and implementation.

XVI

j Preface of Volume 5 External, gas-side heat and mass transfer resistances can seriously limit the rate of drying processes, but they can also be radically reduced by high velocity flow impingement on the surface of the material to be dried. This is easy to implement for flat products such as paper, textiles, tissues, tiles or wood veneer, but connected with questions about how many nozzles shall be used, and how these nozzles shall be placed and operated. Answers to those questions are provided in Chapter 1, in a concise way that refrains from the consideration of less significant details and aims at immediate engineering applicability. The same purpose of lowering external heat and mass transfer resistances can be achieved by imposing, instead of a steady turbulent flow, an oscillatory flow of drying gas around the material to be dried. This can be realized by drying in flue gas coming from a special, pulse burner via a tailpipe to the drying chamber. Intensification of the external heat and mass transfer may not be as spectacular as in case of impinging jets, but the method is applicable to virtually any kind of convective dryer, i.e. it is not restricted to flat products. Construction and operation of the respective combustors, enhancement of heat and mass transfer, and modeling are discussed in Chapter 2. In Chapter 3, the focus is shifted to the use of superheated steam, instead of hot air of flue gas, as the drying agent, which has an influence on both, the external and the internal heat and mass transfer. A major advantage of this process is that energy can much easier be recovered from exhaust steam, than from the wet exhaust air of conventional drying processes. The necessity of operating above the boiling point of the liquid to be removed, usually water, may be turned into an advantage by combining the drying process with, e.g., sterilization or cooking of foods and biomaterials. Damage that such materials might suffer at the boiling temperature of water at ambient pressure can be prevented by reducing the operating pressure, i.e. by low-pressure superheated steam drying. Alternatively, drying rates can be boosted in fluidized beds by combining heat transfer from the fluidization air with indirect heat transfer, usually from immersed steam tubes. Fundamentals and applications of respective processes are discussed in Chapter 4. It is pointed out that the resulting process intensification can be used for increasing the capacity of the dryer, or for reducing the temperature level in order to protect thermally sensitive products. Moreover, the process can be significantly intensified by applying spouted beds, instead of conventional fluidized beds. The background of this behavior is that regions of extremely high gas velocity can be realized in specially designed spouted beds with adjustable air inlet. However, there are foods and pharmaceuticals which are so sensitive, that they must be dried from the frozen state. Purposeful use of the notoriously slow process of freeze drying increases the necessity and urgency of process intensification measures. Various such measures are available, as discussed in Chapter 5, including automatic control for better drying cycles, favorable templating of the solid matrix to be dried by controlled freezing, the use of organic solvents, freeze drying under atmospheric conditions, or the transition from batch to the continuous operation mode. Moreover, hybrid processes can be applied, such as microwave or ultrasound assisted freeze drying.

Preface of Volume 5

Sometimes, product quality requirements meet with the goal of more intense drying processes for hard-to-dry products, such as fruit pulps, juices, or dairy. Foam drying techniques provide attractive solutions for such cases. Foamed products can be produced in spray dryers by injecting inert gas to the feed of the dryer, before or during atomization. Alternatively, solutions or dispersions can be whipped to foam that is subsequently dried in any appropriate type of equipment, which is denoted by foam-mat drying. Respective process configurations, enhancement of drying by increased surface area and more open structures, and resulting product properties are discussed in Chapter 6. In some other cases, biological materials to be dried contain natural barriers to mass transfer, such as cell membranes. Then, drying can be enhanced by applying pulsed electric fields to create pores in the membranes or disintegrate the cells, followed by osmotic dehydration, hot-air drying, or freeze drying. Similar effects can be attained by application of ultrasound to support and assist the mentioned drying processes. Principles and results of these novel technologies are presented in Chapter 7, along with methods for the structural and textural characterization of the materials, and quality characteristics of the resulting products. A more detailed treatment of the application of ultrasound is provided in Chapter 8, along with a discussion of the principles of generation and transmission of ultrasound energy to the material to be treated. It is pointed out that power ultrasound can be used to assist both, liquid-solid processes, such as brine treatment, and drying. The acoustic field is shown to enhance, by a number of mechanisms, both, the external and the internal mass transfer when combined with hot air or atmospheric freeze drying of vegetables and fruits. The more porous the material, and the lower the permissible temperature and gas velocity, the higher is the intensification that can be reached by application of power ultrasound. Another method of hybrid or assisted processing is to support hot-air drying, vacuum drying, freeze drying, or spouted bed drying by microwaves. Microwaves have the unique property of targeting heat supply to the consumer, i.e. to the wet interior of drying materials. Respective processes, equipment, and the enhancement of drying rate that can be achieved by means of the microwaves are thoroughly presented in Chapter 9. Moreover, issues of energy consumption, automatic control, and product quality are addressed. Agricultural products and food materials are, again, in the focus of the discussion. Infrared radiation usually does not penetrate deep into materials, but it can significantly intensify drying processes by supplying significant and well controllable amounts of energy to the surface. In a comprehensive treatment of infrared drying in Chapter 10 different types of radiators, including gas-fired ones, are presented, the necessity of matching the infrared spectrum used with the properties of the material to be dried is stressed, and opportunities to improve product quality by intermittent radiation supply are pointed out. Combinations of infrared heat supply with hot-air drying, microwave drying, freeze drying, and heat-pump drying are discussed. Volume 5 brings several thematic threads set in previous volumes, for instance on fluidized bed drying or food processing, to their contemporary completion. It

jXVII

XVIII

j Preface of Volume 5 reflects the interdisciplinary and multi-scale character of modern drying technology in a similar way as the previous volumes of the series. Therefore, we hope that this final volume and the entire series have at least partially attained the goal of providing all people working on drying in industry and research with a map that shows where drying science and technology are, where they are presently growing to cope with increasing challenges and application demands, and where they may be in the future. In other words, we hope that the series can contribute to the solution of specific, well defined practical tasks (“improve quality, save energy, cut costs”, as promised in the flyer of the publisher), but that it can also motivate further exploratory work and inspire to innovation in an important and rewarding field of engineering science. In this farewell preface, we would like to renew our profound acknowledgement of all persons who have made this series possible our families and co-workers, the excellent editorial team of the publisher, all our outstanding and esteemed colleagues and friends who have served as chapter authors, the countless engineers and scientists whose contributions are quoted in the book series, but also those who have contributed to drying science and practice without finding individual citation. Summer 2013

Evangelos Tsotsas Arun S. Mujumdar

jXIX

List of Contributors Editors Prof. Evangelos Tsotsas Otto von Guericke University Magdeburg Thermal Process Engineering PSF 4120 39106 Magdeburg Germany Email: [email protected] Prof. Arun S. Mujumdar McGill University Department of Bioresource Engineering 2111 Lakeshore Road Sainte-Anne-de-Bellevue Quebec H9X 3V9 Canada Email: [email protected] Authors Prof. Antonello A. Barresi Politecnico di Torino Dipartimento di Scienza Applicata e Tecnologia Corso Duca degli Abruzzi 24 10129 Torino Italy Email: [email protected]


rcel Prof. Juan Andr
es Ca Universidad Polit
ecnica de Valencia Departamento de Tecnología de Alimentos Cami de Vera s/n 46022 Valencia Spain Email: [email protected] Prof. Sakamon Devahastin King Mongkut’s University of Technology Thonburi Department of Food Engineering 126 Pracha u-tid Road Bangkok 10140 Thailand Email: [email protected] Prof. German Efremov Moscow State Open University Street Krasnogo Mayaka 13a cor. 2, app. 71 117570 Moscow Russia Email: [email protected] Dr. Davide Fissore Politecnico di Torino Dipartimento di Scienza Applicata e Tecnologia Corso Duca degli Abruzzi 24 10129 Torino Italy Email: davide.fi[email protected]

j List of Contributors

XX

Prof. Jos
e Vicente García-P
erez Universidad Polit
ecnica de Valencia Departamento de Tecnología de Alimentos Cami de Vera s/n 46022 Valencia Spain Email: [email protected] Prof. Stefan Heinrich Hamburg University of Technology Institute of Solids Process Engineering and Particle Technology Denickestrasse 15 21073 Hamburg Germany Email: [email protected]

Dr. Tadeusz Kudra CanmetENERGY 957 de Salieres St. Jean-sur-Richelieu Quebec J2W 1A3 Canada Email: [email protected] Artur Lewandowski Lodz University of Technology Faculty of Process and Environmental Engineering ul. Wolczanska 213 93-924 Lodz Poland Email: [email protected]

Dr. Michael Jacob Glatt Ingenieurtechnik GmbH Nordstrasse12 99427 Weimar Germany Email: [email protected]

Prof. Xiangdong Liu China Agricultural University College of Engineering 17 Qinghua East Rd. Beijing 100083 P. R. China Email: [email protected]

€ger Dr. Henry Ja Technische Universit€at Berlin Department of Food Biotechnology and Food Process Engineering Koenigin-Luise-Strasse 22 14195 Berlin Germany Email: [email protected]

Prof. Lothar M€orl Otto von Guericke University Magdeburg Chemical Equipment Design PSF 4120 39106 Magdeburg Germany Email: [email protected]

Prof. Dietrich Knorr Technische Universit€at Berlin Department of Food Biotechnology and Food Process Engineering Koenigin-Luise-Strasse 22 14195 Berlin Germany Email: [email protected]

Prof. Arun S. Mujumdar McGill University Department of Bioresource Engineering 2111 Lakeshore Road Sainte-Anne-de-Bellevue Quebec H9X 3V9 Canada Email: [email protected]

List of Contributors

Prof. Mirko Peglow IPT-PERGANDE GmbH Wilfried-Pergande-Platz 1 06369 Weißandt-G€olzau Germany Email: [email protected] Dr. Roberto Pisano Politecnico di Torino Dipartimento di Scienza Applicata e Tecnologia Corso Duca degli Abruzzi 24 10129 Torino Italy Email: [email protected] Prof. Antonio Mulet Pons Universidad Polit
ecnica de Valencia Departamento de Tecnología de Alimentos Cami de Vera s/n 46022 Valencia Spain Email: [email protected] Dr. Julia Rabaeva Lodz University of Technology Faculty of Process and Environmental Engineering ul. Wolczanska 213 93-924 Lodz Poland Email: [email protected] Dr. Enrique Riera Instituto de Seguridad de la Informaci
on (ISI) Grupo de Sistemas Ultras
onicos CSIC Serrano 144 28006 Madrid Spain Email: [email protected]

Prof. Carmen Rossell
o English University of Illes Balears Spanish Departamento de Química Ctra. Valldemossa km 7.5 07122 Palma Mallorca Spain Email: [email protected] Dr. Katharina Sch€ossler Technische Universit€at Berlin Department of Food Biotechnology and Food Process Engineering Koenigin-Luise-Strasse 22 14195 Berlin Germany Email: [email protected] Prof. Eckehard Specht Otto von Guericke University Magdeburg Thermodynamics and Combustion PSF 4120 39106 Magdeburg Germany Email: [email protected] Prof. Evangelos Tsotsas Otto von Guericke University Magdeburg Thermal Process Engineering PSF 4120 39106 Magdeburg Germany Email: [email protected]

jXXI

XXII

j List of Contributors Dr. Yingqiang Wang Jiangnan University School of Food Science and Technology 214122 Wuxi Jiangsu Province P. R. China

Prof. Ireneusz Zbicinski Lodz University of Technology Faculty of Process and Environmental Engineering ul. Wolczanska 213 93-924 Lodz Poland Email: [email protected]

and Longdong University College of Agriculture and Forestry Lanzhou Road 45 745000 Qingyang Gansu Province P. R. China Email: [email protected]

Prof. Min Zhang Jiangnan University School of Food Science and Technology 214122 Wuxi Jiangsu Province P. R. China Email: [email protected]

jXXIII

Recommended Notation  Alternative symbols are given in brackets  Vectors are denoted by bold symbols, a single bar, an arrow or an index (e.g., index: i)  Tensors are denoted by bold symbols, a double bar or a double index (e.g., index: i, j)  Multiple subscripts should be separated by colon (e.g., rp;dry : density of dry particle) A aw B b C (K) c D D (d) d E F _ FðVÞ f f G G g H H H h h (a) h~ ðhN Þ Dhv I

surface area water activity nucleation rate breakage function constant or coefficient specific heat capacity equipment diameter diffusion coefficient diameter or size of solids energy mass flux function volumetric flow rate relative (normalized) drying rate multidimensional number density shear function or modulus growth rate acceleration due to gravity height enthalpy Heaviside step function specific enthalpy (dry basis) heat-transfer coefficient molar enthalpy specific enthalpy of evaporation total number of intervals

m 2 – kg 1 m 1 s m 3 various J kg 1K 1 m m2 s 1 m J – m 3s 1 – – Pa kg s 1 ms 2 m J – J kg 1 Wm 2K 1 J mol 1 J kg 1 –

1

XXIV

j Recommended Notation J J _ JÞ j ðm; K k (b) L M (m) ~ (M, MN) M _ ðWÞ M _ ðJ; jÞ m _ m N N N_ ðW N Þ n n n n_ ðJ N Þ P P p Q_ ðQÞ q_ ðqÞ R R ~ ðRN Þ R r r S S s T t u u V V_ ðFÞ v v W _ W ðMÞ w X x

numerical flux function Jacobian matrix mass flux, drying rate dilatation function or bulk modulus mass transfer coefficient length mass molecular mass mass flow rate mass flux, drying rate volumetric rate of evaporation number molar amount molar flow rate molar density, molar concentration number density outward normal unit vector molar flux power total pressure partial pressure/vapor pressure of component heat flow rate heat flux equipment radius individual gas constant universal gas constant radial coordinate pore (throat) radius saturation selection function boundary-layer thickness temperature time velocity, usually in z-direction displacement volume, averaging volume volumetric flow rate specific volume general velocity, velocity in x-direction weight force mass flow rate velocity, usually in y-direction solids moisture content (dry basis) mass fraction in liquid phase

– various kg m 2 s 1 Pa ms 1 m kg kg kmol 1 kg s 1 kg m 2 s 1 kg m 3 s 1 – mol mol s 1 mol m 3 m 3 mol m 2 s 1 W kg m 1 s 2 kg m 1 s 2 W Wm 2 m J kg 1 K 1 J kmol 1 K m m – s 1 m K, C s ms 1 m m3 m3 s 1 m3 kg 1 ms 1 N kg s 1 ms 1 – –

1

Recommended Notation

x x x0 x~ ðx N Þ Y y y (v) ~y ðyN Þ z Operators ! ! D Greek letters a (h) b (k) b d d (D) e e e e h u k l m m n p r S s s s s t F w w v v (y)

particle volume in population balances general Eulerian coordinate, coordinate (usually lateral) general Lagrangian coordinate molar fraction in liquid phase gas moisture content (dry basis) spatial coordinate (usually lateral) mass fraction in gas phase molar fraction in gas phase spatial coordinate (usually axial)

m3 m m – – m – – m

gradient operator divergence operator difference operator heat-transfer coefficient mass-transfer coefficient aggregation kernel Dirac-delta distribution diffusion coefficient voidage emissivity small-scale parameter for periodic media strain efficiency angle, angular coordinate thermal diffusivity thermal conductivity dynamic viscosity moment of the particle-size distribution kinematic viscosity circular constant density, mass concentration summation operator surface tension Stefan–Boltzmann constant for radiative heat transfer standard deviation (of pore-size distribution) stress dimensionless time characteristic moisture content relative humidity phase potential angular velocity mass fraction in gas phase

Wm 2K ms 1 s 1 m2 s 1 – – – – – rad m2 s 1 W m 1K kg m 1 s various m2 s 1 – kg m 3 Nm Wm m Pa – – – Pa rad –

1

1 1

1 2

K

4

jXXV

XXVI

j Recommended Notation Subscripts a as b bed c c cr D dry dp eff eq f g H I i, 1, 2, . . . i, j, k in l m max mf min N o out P p pbe ph r rel s S surf V v w w wb wet 1

at ambient conditions at adiabatic saturation conditions bound water bed cross section capillary at critical moisture content drag dry at dewpoint effective equilibrium (moisture content) friction gas (dry) wet (humid) gas inner component index, particle index coordinate index, i,j,k = 1 to 3 inlet value liquid (alternative: as a superscript) mean value maximum at minimum fluidization minimum molar quantity outer outlet value at constant pressure particle population balance equation at the interface radiation relative velocity solid (compact solid phase), alternative: as a superscript at saturation conditions surface based on volume vapor, evaporation water wall at wet-bulb conditions wet at large distance from interface

Recommended Notation

Superscripts, Special symbols v   or h i a or h ia 

volumetric strain rheological strain at saturation conditions average, phase average intrinsic phase average spatial deviation variable

jXXVII

jXXIX

EFCE Working Party on Drying: Address List Prof. Odilio Alves-Filho Norwegian University of Science and Technology Department of Energy and Process Engineering Kolbjørn Hejes vei 1B 7491 Trondheim Norway odilo.fi[email protected] Prof. Julien Andrieu (delegate) UCB Lyon I/ESCPE LAGEP UMR CNRS 5007 batiment 308 G 43 boulevard du 11 novembre 1918 69622 Villeurbanne cedex France [email protected] Dr. Paul Avontuur (guest industry) Glaxo Smith Kline New Frontiers Science Park H89 Harlow CM19 5AW United Kingdom [email protected] Prof. Christopher G. J. Baker Drying Associates Harwell International Business Centre 404/13 Harwell Didcot Oxfordshire OX11 ORA United Kingdom [email protected]

Prof. Antonello Barresi (delegate) Dip. Scienza dei Materiali e Ingegneria Chimica Politecnico di Torino Corso Duca degli Abruzzi 24 10129 Torino Italy [email protected] Dr. Rainer Bellinghausen (delegate) Bayer Technology Services GmbH BTS-PT-PT-PDSP Building E 41 51368 Leverkusen Germany [email protected] Dr. Carl-Gustav Berg Abo Akademi Process Design Laboratory Biskopsgatan 8 20500 Abo Finland cberg@abo.fi Dr. Catherine Bonazzi (delegate) AgroParisTech - INRA JRU for Food Process Engineering 1 Avenue des Olympiades 91744 Massy cedex France [email protected]

j EFCE Working Party on Drying: Address List

XXX

Paul Deckers M.Sc. (delegate) Bodec Process Optimization and Development Industrial Area ‘t Zand Bedrijfsweg 1 5683 CM Best The Netherlands [email protected] Henk van Deventer M.Sc. (delegate) TNO Utrechtseweg 48 3704 HE Zeist The Netherlands [email protected]

Prof. Dr. Istvan Farkas (chairman) Szent Istvan University Dep. of Physics and Process Control Pater K. u. 1 2103 Godollo Hungary [email protected] Dr. Dietrich Gehrmann Wilhelm-Hastrich-Str. 12 51381 Leverkusen Germany [email protected]

Dr. German I. Efremov Pavla Korchagina 22 129278 Moscow Russia [email protected]

Prof. Adrian-Gabriel Ghiaus (delegate) Thermal Engineering Department Technical University of Civil Engineering Bd. P. Protopopescu 66 021414 Bucharest Romania [email protected]

Prof. Trygve Eikevik Norwegian University of Science and Technology Dep. of Energy and Process Engineering Kolbjørn Hejes vei 1B 7491 Trondheim Norway [email protected]

Prof. Gheorghita Jinescu University “Politehnica” din Bucuresti Department of Chemical Engineering 1 Polizu street, Room F210 78126 Bucharest Romania [email protected]

Dr.-Ing. Ioannis Evripidis Dow Deutschland GmbH & Co. OHG P.O. Box 1120 21677 Stade Germany [email protected]

Prof. Dr. Gligor Kanevce St. Kliment Ohridski University Faculty of Technical Sciences ul. Ivo Ribar Lola b.b. Bitola FYR of Macedonia [email protected]

EFCE Working Party on Drying: Address List

Ir. Ian C. Kemp (delegate) Glaxo SmithKline, R&D Gunnels Wood Road Stevenage SG1 2NY United Kingdom [email protected] Prof. Matthias Kind Karlsruhe Institute of Technology € r Thermische Institut fu Verfahrenstechnik Kaiserstr. 12 76128 Karlsruhe Germany [email protected] Prof. Stefan J. Kowalski Poznan University of Technology Institute of Technology and Chemical Engineering ul. Marii Sklodowskiej Cuvie 2 60965 Poznan Poland [email protected] Prof. Magdalini Krokida (delegate) National Technical University of Athens Department of Chemical Engineering Politechnioupoli, Zografou Campus 15780 Athens Greece [email protected] Dr. Ir. Angelique Leonard (delegate) Universite de Liege Departement de Chimie Appliquee Laboratoire de Genie Chimique B^atiment B6c – Sart-Tilman 4000 Liege Belgium [email protected]

Prof. Avi Levy (delegate) Ben-Gurion University of the Negev Department of Mechnical Engineering Beer-Sheva 84105 Israel [email protected] Jean-Claude Masson Rhodia Operations, CRTL-GI-DIPH 85 Avenue des Freres Perret 69196 Saint-Font Cedex France [email protected] Prof. Natalia Menshutina Mendeleyev University of Chemical Technology of Russia (MUCTR) High Technology Department Muisskaya sq.9 125047 Moscow Russia [email protected] Dr. Thomas Metzger BASF SE GCP/TT-L540 67056 Ludwigshafen Germany [email protected] Prof. Antonio Mulet Pons (delegate) Universitat Politecnica de Valencia Departament de Tecnologia d’Aliments Cami de Vera s/n 46071 Valencia Spain [email protected]

jXXXI

XXXII

j EFCE Working Party on Drying: Address List Prof. Zdzislaw Pakowski (delegate) Lodz University of Technology Faculty of Process and Environmental Engineering ul. Wolczanska 213 90924 Lodz Poland [email protected] Prof. Patrick Perre (delegate) Ecole Centrale Paris Laboratoire de Genie des Procedes et Materiaux Grande Voie des Vignes 92295 Ch^atenay-Malabry France [email protected] Dr. Romain Remond AgroParisTech – ENGREF 14, Rue Girardet 54042 Nancy France [email protected] Dr. Roger Rentr€om Karlstad University Department of Environmental and Energy Systems Universtietsgatan 2 65188 Karlstad Sweden [email protected] Prof. Michel Roques Universite de Pau at des Pays de l’Adour 5 Rue Jules-Ferry ENSGTI 64000 Pau France [email protected]

Prof. Carmen Rossell o (delegate) University of Illes Baleares Dep. Quimica Ctra. Valldemossa km 7.5 07122 Palma Mallorca Spain [email protected] Dr. Panayiotis Scarlatos SusTchem Engineering LTD 144 3rd September Street 11251 Athens Greece [email protected] Dr. Michael Sch€ onherr BASF SE GCT/T – L 540 67056 Ludwigshafen Germany [email protected] Dr. Andreas Schreiner Novartis Pharma AG WSJ-145.1.54 Lichtstr. 35 4056 Basel Switzerland [email protected] Dr. Milan Stakic Vin9ca Institute for Nuclear Sciences Center NTI P.O. Box 522 11001 Belgrade Serbia [email protected] Dr. Andrew Stapley (delegate) Loughborough University Department of Chemical Engineering Loughborough, Leicestershire LE11 3TU United Kingdom [email protected]

EFCE Working Party on Drying: Address List

Prof. Stig Stenstrom (delegate) Lund University Institute of Technology Department of Chemical Engineering P.O. Box 124 22100 Lund Sweden [email protected] Prof. Ingvald Strommen (delegate) Norwegian University of Science and Technology Department of Energy and Process Engineering Kolbjørn Hejes vei 1b 7491 Trondheim Norway [email protected] Prof. Czeslaw Strumillo (delegate) Lodz University of Technology Faculty of Process and Environmental Engineering ul. Wolczanska 213 90924 Lodz Poland [email protected] Prof. Radivoje Topic (delegate) University of Belgrade Faculty of Mechanical Engineering 27 Marta 80 11000 Beograd Serbia [email protected]

Prof. Evangelos Tsotsas (delegate) Otto von Guericke University Thermal Process Engineering P.O. Box 4120 39016 Magdeburg Germany [email protected] Thorvald Ullum M.Sc. GEA, Niro Gladsaxevej 305 2860 Soeborg Denmark [email protected] Dr. Bertrand Woinet (delegate) SANOFI-PASTEUR, CDP B^atiment 8600 31-33 Quai Armand Barbes 69683 Neuville sur Sa^ one cedex France Bertrand.woinet@sanofi-aventis. com Prof. Ireneusz Zbicinski Lodz University of Technology Faculty of Process and Environmental Engineering ul. Wolczanska 213 90924 Lodz Poland [email protected]

jXXXIII

1

1 Impinging Jet Drying Eckehard Specht 1.1 Application

Flat products such as tiles, tissue, paper, textiles and wood veneer are often dried using nozzle arrays (Mujumdar, 2007). Figure 1.1 illustrates the basic principle of the drying process using a nozzle array. Ambient air of temperature Ta is heated in a combustion chamber or in a heat exchanger to temperature T0, requiring the _ 0. The heated air is blown through the nozzle array in order to dry the energy H product. These conditions result in the product temperature TS. The evaporating _ v. The nozzle array is characterized by the inner mass flow has the enthalpy H diameter d of the individual nozzle, by the pitch t between the nozzles, and the velocity w of the released gas. There are basically three possible designs of nozzle arrays which differ with regard to the spent flow of the air (Fig. 1.2). In a field of individual nozzles the air can flow unimpeded between almost all nozzles; however, in a hole channel the air can flow only between those above. In a perforated plate the air can only continue to flow laterally and then escape. Hole channels and perforated plates are easier to produce than single nozzles, as they only require holes to be perforated. However, the heat transfer is the highest for nozzle fields and the lowest for perforated plates, as will be subsequently shown. For the design of the nozzle array the energy consumption needed for drying is essential. This is the energy for heating the air: _ 0 ¼ r0 V_ 0 cp ðT0 H

Ta Þ;

ð1:1Þ

where T0 and Ta represent the temperatures of the heated air and the environment, respectively, and cP is the average specific heat capacity between these two temperatures. The density and the volume flow rate refer to the heated air. Volume flow rate depends on the discharge velocity w and the number n of the nozzles with a diameter of d: p V_ 0 ¼ n d2 w: 4

ð1:2Þ

Modern Drying Technology Volume 5: Process Intensification, First Edition. Edited by Evangelos Tsotsas and Arun S. Mujumdar. Ó 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

2

1 Impinging Jet Drying

Fig. 1.1 Drying using a nozzle array.

The number of nozzles will depend on the nozzle pitch t and the surface area A of the material: n¼

A : t2

ð1:3Þ

The air temperature is calculated from the condition that the transferred heat has to cover the enthalpy of vaporization and the enthalpy to heat the dry material flow from ambient temperature to the saturation temperature TS: a AðT0

_ s cs ðTS _ v A D hv þ M TS Þ ¼ m

Ta Þ:

ð1:4Þ

_ v is the evaporation flux, D hv is the evaporation enthalpy, and cs is In Eq. 1.4, m the specific heat capacity of the material. The evaporating mass flux is obtained from the relationship for the mass transfer _v¼b m

P P ln Rv TS P

Pa : PS

ð1:5Þ

Here, the influence of one-side diffusion is taken into consideration. The gas constant of the vapor is represented by Rv, the total pressure by p, the partial pressure of the vapor in the ambient air by pa, and the saturated vapor pressure by pS. Additionally, the analogy of the Nusselt and Sherwood function is applied to the ratio of the heat and mass transfer coefficients: a ¼ brcP Le0:6 ;

ð1:6Þ

wherein for the exponent of the Prandtl number in the Nusselt function the value 0.4 was used. The saturation pressure is approximated from the equilibrium relationship 
 Dhv 1 1 PS ¼ P0 exp ð1:7Þ Rv TS T0

1.1 Application

Fig. 1.2 Types of nozzle arrays. (a) Single-nozzle array; (b) Hole channel; (c) Perforated plate.

with the reference condition P0, T0, for example, P0 ¼ 1 bar, T0 ¼ 373 K. The minimum required energy is the enthalpy of vaporization of the water _v¼m _ v D hv : H

ð1:8Þ

3

4

1 Impinging Jet Drying

_ 0 related to the The specific drying energy is the energy for heating the air H enthalpy of evaporation. From the above equations results
 _ v D hv m 2
 þ T pd r c w T P S a 0 _0 pd2 r0 cP w 1 TS Ta H a : ð1:9Þ ¼ ¼ þ _v _ v D hv _ v D hv 4t2 m 4t2 a m H The enthalpy to heat up the dry material was omitted for clarity purposes. Specific drying energy consumption according to Eq. 1.9 is dependent on the heat transfer coefficient. Particularly at high rates of evaporation, the specific energy consumption is lower with a high heat transfer coefficient. This strongly influences the drying rate and the size of the apparatus and, as a result, an increasing heat transfer coefficient increases the rate of drying which in turn allows a reduction in the size of the apparatus. The setting and the regulation of the heat transfer coefficient is therefore of great importance. The heat transfer of bodies in a crossflow is relatively low. In generating a high heat transfer, nozzle arrays are implemented wherein the jet emerging from the nozzles is perpendicular to the body. Such flows are called stagnation point flows. Nozzles may be either round or slot-shaped. The fields of nozzles can be made from single nozzles, or hole channels, or from perforated plates with aligned or staggered arrangements, permitting a variety of geometric parameters. The heat transfer coefficient of nozzle arrays is therefore considered in more detail in the following.

1.2 Single Nozzle

First, an air jet emerging from a single nozzle is considered. In Fig. 1.3, the generated flow field of a nozzle is shown schematically. From the nozzle with the diameter d, the flow exits with the approximately constant speed w. The jet impinges the surface virtually unchanged with a constant velocity as long as

Fig. 1.3 Boundary layer of a stagnation point flow.

1.2 Single Nozzle

Fig. 1.4 Local heat transfer coefficients for a single nozzle.

the relative distance between the nozzle and the surface h/d is less than 6. At greater distances the core velocity decreases reducing the heat transfer. In the impinged zone, also called the stagnation point region, the flow is redirected in a radial direction, and in this region the flow can always be considered as laminar. At the outer edge of the emerging jet, circular vortices form. In the deflection region at r/d  1, a large vortex forms but this disintegrates at approximately r/d  2 into many small vortices. Eventually, the flow becomes increasingly turbulent (refer to Angioletti et al., 2003, for photographs of the visualized flow). Increasing the nozzle distance from the stagnation point decreases the velocity, which in turn results in a decrease of the heat transfer coefficient. A completely laminar flow only occurs at Reynolds numbers smaller than 50; however, such small Reynolds numbers possess no meaning for technical heat-transfer processes. The local heat transfer coefficient depends on three geometric dimensions: the nozzle diameter d; the nozzle distance h; and the radial distance r. Figure 1.4 shows the local heat transfer coefficients as a function of relative radius r/d with the relative distance h/d as a parameter exemplarily for a nozzle with a diameter of 10 mm and an outlet velocity of 81 m s 1. It is evident for small distances of h/d ¼ 1 to 4 that the heat transfer coefficient at the stagnation point (r/d  0.5) is approximately constant, but then drops slightly to a relative minimum at r/d  1. In the region of the vortex degradation at r/d  2 the coefficient passes through a maximum and then decreases continuously. At radial distances r/d > 3.5, all profiles coincide for h/d  6. Only higher distances result in a lower heat transfer. For technical processes, the average heat transfer coefficient for a circular area, ðr 1 a ¼ 2 aðrÞ2prdr; pr 0

ð1:10Þ

5

6

1 Impinging Jet Drying

Fig. 1.5 Average Nusselt number of a single nozzle.

is of interest rather than the local heat transfer coefficient. The nozzle diameter exerts the greatest influence of all three geometric parameters. Therefore, the Nusselt number (Nu) and Reynolds number (Re) are defined as Nu ¼

ad wd ; Re ¼ : l n

ð1:11Þ

In Fig. 1.5, the average Nusselt numbers are shown depending on the specific radius r/d, with the relative distance h/d as a parameter for a given Reynolds number. In the figure it is evident that from a distance of approximately r=d > 3, all profiles coincide for h=d  6; greater distances will then result in a lower heat transfer. As a consequence of Figs 1.4 and 1.5, it can be concluded that:  The relative nozzle distance should be h/d < 6 because for higher distances the heat transfer decreases. At a distance less than h/d ¼ 4, the local variation in the stagnation region increases.  Up to relative stagnation point distances of approximately r/d ¼ 3, the average heat transfer is largely independent of the stagnation point distance. Only for greater distances will the heat transfer begin to decrease. The Nusselt functions specified in the literature deviate from each other particularly with regard to the exponent of the Reynolds number; this is because the exponent depends on the radial distance. At the stagnation point the flow is always laminar, and the exponent is therefore 0.5. Turbulence fully develops only at a considerable distance from the stagnation point, and as a result the exponent must be 0.8. Figure 1.6 shows the increase of the exponent related to the radius (Adler, 2004). At a distance of r/d ¼ 3, the exponent has reached a value of only about 0.62. Different exponents are obtained depending on the size of the area considered in averaging.

1.3 Nozzle Fields

Fig. 1.6 Dependence of the exponent of the Reynolds number on distance.

The Nusselt functions reported in the literature for the stagnation point (r/d   0.5) can be approximated by NuSt ¼ 0:72 Re0:5 Pr0:4 :

ð1:12Þ

For distances around r/d ¼ 3, the average heat transfer can be accurately approximated with a power of 0.67 Nu ¼ 0:12 Re0:67 Pr0:4 :

ð1:13Þ

In the literature, values ranging from 0.67 to 0.7 are mostly reported and, in order to improve comparability, in all of the following discussions the Reynolds exponent is approximated by 0.67. Nusselt functions are typically determined for air, which is the typical medium used during application.

1.3 Nozzle Fields

In the case of larger areas to be cooled or heated, several nozzles in a so-called nozzle field must be used. However, the relative nozzle pitch t/d must now be taken into consideration as an additional geometric parameter. Initially, arrays of single nozzles are discussed, and it is assumed that the pitch is equal in both directions. 1.3.1 Arrays of Single Nozzles

The profiles of the local Nusselt number Nulo for a row of three nozzles is shown in Fig. 1.7 (Attalla, 2005). Placement of the nozzles is indicated in the figure, where it can be seen how the typical profiles of the single nozzles overlap.

7

8

1 Impinging Jet Drying

Fig. 1.7 Profiles of the local Nusselt number in an array of individual nozzles.

In Fig. 1.2a, the view of a nozzle array is shown with the impacted area. The nozzles may be arranged either in-line or staggered (in this figure the nozzles are in-line). In any case, each nozzle is influenced by the square area t2 of the nozzle pitch, and by using this area the average of the heat transfer can be determined. In Fig. 1.8, the average Nusselt number is shown in exemplary fashion as a function of the pitch t/d in an in-line assembly, with the relative nozzle distance h/ d as a parameter for two Reynolds numbers. It is evident that a pitch of t/d ¼ 6 will always result in a pronounced maximum. In Fig. 1.9, the average Nusselt number is shown as a function of the relative nozzle distance with the pitch as a parameter. The profiles for both the in-line and staggered arrangements are shown. The Nusselt numbers are approximately constant up to a relative nozzle distance of about five, and then decrease continuously. When considering the profiles of the average Nusselt number, no significant difference can be seen between the in-line and staggered arrangements.

Fig. 1.8 Average heat transfer dependent on the pitch of the individual nozzles.

1.3 Nozzle Fields

Fig. 1.9 Average heat transfer dependent on the relative nozzle distance.

However, the profiles of the local Nusselt number differ considerably (this point will be discussed further). For t/d ¼ 6 (maximum heat transfer), the Nusselt function in the stagnation area can be expressed as r/d   0.5 NuSt ¼ 0:82 Re0:5 Pr0:4 ;

ð1:14Þ

The average Nusselt number over the entire surface area for h/d < 5 is Nu ¼ 0:16 Re0:67 Pr0:4

ð1:15Þ

(Attalla and Specht, 2009). In comparison with respect to the Nusselt functions for single nozzle (Eqs 1.12 and 1.13), it is evident that the heat transfer in the nozzle array is higher than that of the single nozzle under the condition of similar Reynolds numbers. Regarding the average heat transfer, this increase is approximately 30%. In the following, the influence of the nozzle arrangement is considered in more detail. In Fig. 1.10 infrared images of the temperature field of a metal plate for both nozzle arrangements are shown. The average heat transfer is equal in both arrangements, as long as the relative nozzle distance is equal, as shown previously. However, there are differences in the profiles of the local heat transfer coefficient, and thus in the uniformity of heat transfer. In both images two lines are shown where the heat transfer is either maximal or minimal. The maximum heat transfer always occurs along the line passing through the stagnation points of the nozzles; accordingly, the minimum heat transfer always occurs on the line passing the middle between the stagnation points. As a consequence, the heat transfer is different in the longitudinal and transverse directions on a belt that passes beneath the nozzle array. In Fig. 1.11 for a specific example, the profiles of the Nusselt number along the lines of the maximum and minimum heat transfer for the in-line arrangement and the staggered arrangement are shown, respectively. For the in-line arrangement,

9

10

1 Impinging Jet Drying

Fig. 1.10 Temperature field of a plate with (a) an in-line and (b) a staggered arrangement of the nozzles.

heat transfer along line A is always higher than along line B, so that the heat transfer is different in the cross-section. The high difference along line A is especially noteworthy. In this example the Nusselt number fluctuates between the values 19 and 3, which is approximately a factor of 6. Also on line B the Nusselt number fluctuates by about a factor of 4, and the heat transfer is therefore very uneven. In the staggered arrangement, both lines intersect so that the heat transfer takes alternatingly high and low values. It is also worth noting, that the differences

Fig. 1.11 Local heat transfer along lines A and B in (a) an in-line and (b) a staggered nozzle arrangement.

1.3 Nozzle Fields

Fig. 1.12 Average heat transfer along lines A and B (cf. Fig. 1.10) in (a) an in-line and (b) a staggered arrangement.

between the maximum and minimum values of a line are significantly lower than for the in-line arrangement. On line A, the factor between the highest Nusselt number (95) and the lowest Nusselt number (32) amounted to only approximately three, compared to six for the in-line arrangement. Also, the difference between the maximum and minimum Nusselt numbers on line B for the staggered arrangement was less by a factor of 2 than for the in-line arrangement. In Fig. 1.12, along lines A and B, the average Nusselt numbers for both arrangements are shown as a function of the relative distance for two Reynolds numbers at the optimum nozzle pitch. Up to a relative distance of 5, the Nusselt number is again approximately constant and then decreases. It is also significant that the average Nusselt number of line A (maximum value) for the in-line arrangement is always about 20% larger than that of line B (minimum value). In the staggered arrangement, however, no remarkable differences between the Nusselt numbers can be seen for the two lines. Therefore, although the difference between the two arrangements indicates that heat transfer for the staggered arrangement is much more uniform than for the in-line arrangement, the average heat transfer is approximately the same in both arrangements. In Fig. 1.13, the average heat transfer coefficient for a nozzle array is shown as a function of the discharge velocity for selected nozzle diameters. It can be seen that

11

12

1 Impinging Jet Drying

Fig. 1.13 Average heat transfer coefficient in a nozzle array.

it is possible for a nozzle array to achieve very high heat transfer coefficients in comparison to other forms of flow. 1.3.2 Hole Channels

Nozzle arrays are technically easier to manufacture in the form of hole channels than in the form of individual nozzles; such a hole channel is shown, in principle, in Fig. 1.2b. In the rectangular channel, the holes are punched in the smaller lateral surface and serve as nozzle openings. In this case, the resistance of the outflow must be significantly higher than for the axial flow in the channel to ensure that the same amount of air emerges through all holes. This is usually the situation if the crosssectional area of the channel is about 2.5-fold greater than the sum of the areas of the perforations. Uneven punching, for example by nonuniform ridges, will result in the jet streams not being perpendicular to the surface, and this in turn will lead to an uneven distribution of heat transfer. In Fig. 1.14, the average heat transfer is shown as a function of the pitch of the perforations, for example for a Reynolds number with relative distance as a parameter. Regarding other Reynolds numbers in principle gives similar curves. Up to a pitch of about 6 the heat transfer will remain substantially constant, but after this point it will decrease. There is, therefore, no pronounced maximum at a

1.3 Nozzle Fields

Fig. 1.14 Average heat transfer in a hole channel.

pitch of 6, as occurs with a field of individual nozzles, although at a smaller pitch the heat transfer will remain identical. Nonetheless it is recommended that, even with hole channels, a pitch is selected that is no less than 6 in order to minimize the flow rate, and thus the fan performance. The heat transfer will again be highest for relative nozzle distances in the range 2 < h/d < 5, but will then decrease continuously with increasing distance. The Nusselt functions for the stagnation point and the average value of the field can be approximated by (Attalla, 2005): NuSt ¼ 0:47 Re0:5 Pr0:4 ; 0:67

Nu ¼ 0:10 Re

0:4

Pr :

ð1:16Þ ð1:17Þ

Based on comparisons with the corresponding functions for the single-nozzle array, it is evident that the heat transfer in hole channels is less, by about 35%. 1.3.3 Perforated Plates

In perforated plates, spent holes or slots must be installed so that the injected air can flow out again. The size and number of these outlets depends on the manufacturer of the perforated plates, and consequently various cross-flows can occur between the target area and the plate. How this affects the heat transfer is not yet known. Nonetheless, Martin (1977) developed a Nusselt function with which the heat transfer for a majority of perforated plates can be approximated, by providing the following correlation for aligned and hexagonally arranged perforations: "
 # 0:05 h F 6 1 2:2 F 0:67 0:42 Nu ¼ Re Pr 1þ F : ð1:18Þ d 0:6 1 þ 0:2ðh=d 6ÞF Here, h/d is again the relative nozzle distance, and F is the square-root of the area ratio of the nozzle openings to the total area:

13

14

1 Impinging Jet Drying

Fig. 1.15 Influence of the pitch and the nozzle distance on the heat transfer for perforated plates.

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Anozzles : F¼ Atotal

For plates with equally distributed pitch t, it follows that rffiffiffi pd F¼ : 4t

ð1:19Þ

ð1:20Þ

The influence of the relative distance and the nozzle pitch is shown in Fig. 1.15, where heat transfer is seen to decrease continuously with the relative distance. The maximum heat transfer occurs in the case of pitches in the range of 3 to 5, depending on the distance. The heat transfer is lower than that of an array of individual nozzles. As the maximum occurs at a smaller pitch than 6 (as for singlenozzle arrays), a higher flow rate must also be applied. Geers et al. (2008) have largely confirmed the correlation of Martin. Heikkil€a and Milosavljevic (2002) concluded that the correlation provides slightly toohigh values only when air temperature is above 400  C. Huber and Viskanta (1994) provided a somewhat more simple correlation for pitches t/d  4 right of the maximum (see Tab. 1.2), whereby the Nusselt number was seen to decrease continuously with the pitch. 1.3.4 Nozzles for Cylindrical Bodies

For nonplanar bodies the curvature arises as an additional geometric parameter. In the following discussions, a cylindrical body is considered as the fundamental case. In Fig. 1.16, a cylinder of diameter D is shown, which is impinged from a slot of

1.3 Nozzle Fields

Fig. 1.16 Flow to a cylinder from a slot nozzle.

width s. With regards to the influence of the geometric parameters and the Reynolds number, divergent results have occasionally been reported in the literature. For example, Gori and Bossi (2003) and Chan et al. (2002) noted that a maximum in the heat and mass transfer occurred at a relative nozzle distance of h/ s ¼ 8, whereas Nada (2006) gave this maximum as between h/d ¼ 4 and 6. McDaniel and Webb (2000) measured a maximum at h/d ¼ 5, but only from nozzles with rounded edges. In contrast, for sharp-edged nozzles the heat transfer was continuously decreased with distance. Olsson et al. (2004) also indicated a decreasing heat transfer with distance, but this was very moderate with an exponent of only 0.077. For all of these authors the influence of the distance was weak. As noted previously for single nozzles and flat surfaces, a distance up to values of h/d ¼ 5 had no effect, whereas heat transfer was slightly decreased at higher distances. The Nusselt functions given below are therefore based on a distance in the range 2  h/s  8. The influence of the ratio D/s has been reported by the various authors with exponents ranging from 0.03 to 0.22; consequently, in order to provide a better comparison their results were approximated with an average exponent of 0.1. The exponents for the Reynolds number specified by the above-mentioned authors have values from 0.4 to 0.82, depending partly on the geometric size. In Fig. 1.17, the measured Nusselt numbers of various authors are compared using D/s ¼ 2 and h/s ¼ 5 as examples. The length of the dotted lines represents the range of the investigated Reynolds numbers. The bold line with the gradient 0.67 represents an average value; this line can be approximated in

15

16

1 Impinging Jet Drying

Fig. 1.17 Nusselt number dependent on the Reynolds number for cylinders with slot nozzles for D/s ¼ 2 and h/s ¼ 5.

the range 2  h/s  8 with 0:1 0:4 NuD ¼ 0:20 Re0:67 D Pr ðD=sÞ :

ð1:21Þ

The dimensionless numbers are formed with the cylinder diameter D: NuD ¼

aD ; l

ReD ¼

wD : n

ð1:22Þ

The values of the Nusselt numbers are similar to those resulting from the Nusselt function corresponding to cross-flow of cylinders if the cross-flow length ðp DÞ=2 is used as a characteristic dimension in the Nusselt and Reynolds numbers. The Nusselt function can alternatively be formed with the slot width s as the characteristic dimension resulting in Pr0:4 ðD=sÞ Nus ¼ 0:15 Re0:67 s

0:23

;

ð1:23Þ

with Nus ¼

as ; l

Res ¼

ws : n

ð1:24Þ

1.4 Summary of the Nusselt Functions

The Nusselt functions for the single nozzle are summarized in Tab. 1.1, and for the nozzle fields in Tab. 1.2. The Nusselt and Reynolds numbers are defined with the inner nozzle diameter d regarding round nozzles, and with the width s regarding slot nozzles. The exponent of the Reynolds number in correlations for the average

1.5 Design of Nozzle Field Tab. 1.1

Nusselt functions for single nozzles.

Nozzle shape

Nusselt function

Scope

Reference

Circular

Stagnation point region 0:4 Nud ¼ 0:72 Re0:5 d Pr 0:5 0:4 Nus ¼ 0:70 Res Pr Averaged values Pr0:4 Nud ¼ 0:12 Re0:67 d

r/d   0.5; 2  h=d  5 2  h=s  5 0 < r/d < 3

Attalla and Specht (2009) Vader et al. (1991) Adler (2004)

Slot Circular

Tab. 1.2

Nusselt functions for nozzle fields impinging flat surfaces.

Arrays of individual nozzles (in-line or staggered): 0:4 Stagnation point r=d  1: Nud ¼ 0:82 Re0:5 d Pr 0:4 Average: Nud ¼ 0:16 Re0:67 Pr d Pitch t/d ¼ 6, Distance 2  h=d  5 Hole channels: 0:4 Stagnation point r=d  1: Nud ¼ 0:47 Re0:5 d Pr 0:4 Average: Nud ¼ 0:10 Re0:67 Pr d Pitch t/d  6, Distance 2  h=d  4; Width sK =d  2 Perforated plates: "


 # 0:05 h F 6 1 2:2 F Pr 1þ F Nud ¼ d 0:6 1 þ 0:2ðh=d 6ÞF rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffi Anozzles pd , Fðround nozzleÞ ¼ F¼ Atotal 4t Re0:67 d

0:42

Attalla and Specht (2009)

Attalla and Specht (2009)

Martin (1977)

Pitch 1.4  t/d  14, Distance 2  h/d  12

Nud ¼ 0:43 Re0:67 Pr0:4 ðh=dÞ 0:123 ðt=dÞ 0:725 d Pitch 4  t/d  8, Distance 0.25  h/d  6

Huber and Viskanta (1994)

heat transfer in nozzle fields has various reported values ranging from 0.66 to 0.72. It is mostly a matter of choice which exponent is used for the approximation of data within the spread of the measurement results. In order to enhance comparability of the Nusselt functions, the measurement results of the various references were respectively approximated by the exponent 0.67.

1.5 Design of Nozzle Field

Based on the given Nusselt functions for the required heat transfer coefficient in Eq. 1.9, nozzle arrays can now be designed. It is again assumed that the heat is predominantly transferred for evaporation, while the enthalpy to heat up the dry material is again neglected. First, an array of individual nozzles will be considered for which a distinct maximum in the heat transfer results at a pitch of t ¼ 6d. Thus,

17

18

1 Impinging Jet Drying

Fig. 1.18 Specific energy consumption, air, and product temperature as a function of the drying rate.

for the heat transfer coefficient from Eq. 1.15 with Pr ¼ 0.7, is obtained: a ¼ 0:14

l 0:67 w d n0:67

0:33

:

ð1:25Þ

The material properties are to be formed with the average temperature ðT0 þ TS Þ=2. The specific energy consumption thus depends only on the three _ v , d, and w – the influence of which must be calculated numerically. parameters – m In Fig. 1.18, the specific energy consumption, as well as the air and product temperatures, are shown as a function of the drying rate. As an example, a velocity of 50 m s 1, a nozzle diameter of 5 mm, and an ambient temperature of 20  C was used. The higher the drying rate, the more the air must be heated to achieve the enthalpy of evaporation because the velocity, and therefore the flow rate, remain constant. The air does not need to be heated for drying rates less than about 2 g m 2 s 1, as an ambient air temperature of 20  C is sufficient. Consequently, the specific energy consumption will be zero. At higher drying rates, the specific energy consumption is initially increased sharply and then passes through a maximum; however, the specific energy consumption decreases steadily thereafter. In Fig. 1.19, specific energy consumption is shown separately as a function of the evaporation rate, with the nozzle diameter as a parameter. The nozzle exit velocity is assumed to be constant at 50 m s 1. It is evident that the specific energy consumption is less with smaller nozzles, with the maximum being displaced towards higher evaporation rates. However, the lowest possible value of nozzle diameter is limited, first by the increase in production costs, and second through the minimum distance of the nozzle to the product. This distance should not be greater than approximately five-times the nozzle diameter; otherwise the heat transfer will be decreased. Rather, the smaller the nozzle the closer the field must be shifted towards the product. In Fig. 1.20, specific energy consumption is shown as a function of the drying rate, with the nozzle discharge velocity as a parameter; as an example, a nozzle

1.5 Design of Nozzle Field

Fig. 1.19 Specific energy consumption as a function of the drying rate for different nozzle diameters.

Fig. 1.20 Specific energy consumption as a function of the drying rate for different air velocities.

diameter of 5 mm is used. Here, the lower the discharge velocity, the lower is the specific energy consumption, and the maximum is shifted towards lower drying rates. Lower velocities have the further advantage that only a small pressure drop, and therefore less fan power, will be required. However, the gas (i.e., the air) must be heated to higher temperatures in order to transfer the heat required. In Fig. 1.21, the heating temperature of the nozzle flow is shown as a function of the drying rate, with the nozzle exit velocity as a parameter; the nozzle diameter is again 5 mm. The required temperature increases approximately linearly with the drying rate, and the lower the velocity the steeper the gradient. The lowest value of the velocity is thus limited by the maximum temperature of the gas flow. This maximum temperature is dependent on the type of apparatus used for heating, the material strength, and the thermal sensitivity of the product. Due to the strength of the steel material the gas temperatures are generally limited to 500  C.

19

20

1 Impinging Jet Drying

Fig. 1.21 Influence of the drying rate and the velocity on the required air temperature.

Fig. 1.22 Influence of the nozzle diameter.

Figures 1.22 and 1.23 show the specific energy consumption, the product temperature and the gas temperature as a function of the nozzle diameter and of the discharge velocity, respectively, as an example for the drying rate 10 g m 2 s 1. All three parameters decrease with decreasing nozzle diameter, while the gradients become larger with smaller nozzle diameters. With decreasing velocity the specific energy consumption also decreases and both temperatures are increased. The lower the velocity, the larger is the gradient of variation of the considered parameters. For hole channels, analogous results are valid, and a pitch of t ¼ 6d is again recommended. A larger pitch will result in a decrease in the heat transfer, which will remain constant with lower pitches; however, the number of nozzles and thus the flow rate, will be increased. The specific energy consumption and the required gas temperatures are slightly higher for hole channels than for single-nozzle arrays, because the heat transfer is somewhat lower.

1.5 Design of Nozzle Field

Fig. 1.23 Influence of the air velocity.

Fig. 1.24 Comparison of single-nozzle array and perforated plate for different drying rates.

In the case of perforated plates the heat transfer is somewhat less but they are the easiest to manufacture and consequently are commonly used. In Fig. 1.24, the specific energy consumption, the air temperature and the temperature of the material are again shown as a function of the drying rate, with a pitch of t ¼ 6d as an example. The values for the perforated plate are compared here with values for the single-nozzle array. It is clear that, in order to achieve the same drying rate with perforated plates, higher air temperatures – and thus higher specific energy consumptions – are required to compensate for the reduced heat transfer. Figure 1.25 shows the influence of the diameter on specific energy consumption and both temperatures, using examples of drying rates of 10 g m 2 s 1 and of 50 g m 2 s 1; the pitch is again t ¼ 6d. When the drying rate is 10 g m 2 s 1, the specific energy consumption of the perforated plate is almost 50% higher than that of the single-nozzle array. However, at a higher drying rate of 50 g m 2 s 1 for diameters larger than 3 mm, the material can no

21

22

1 Impinging Jet Drying

Fig. 1.25 Influence of the nozzle diameter for two drying rates.

longer be dried; otherwise, the air temperatures must be above 500  C. With regards to perforated plates, the heat transfer according to Fig. 1.15 also depends on the pitch and nozzle distance. Thus, the maximum heat transfer will occur with a pitch in the range of four, which is a smaller value than for single-nozzle arrays. Heat transfer will also depend on the nozzle distance, and increases with decreasing distance. The influence of the pitch is shown for perforated plates in Fig. 1.26, where the pitch is with four smaller than before and with eight greater than before, while the diameter and velocity are held constant. With regards to the pitch of four, lower air temperatures will be needed as the heat transfer will be stronger, although the flow rate will be increased due to the larger number of perforations and, as a result, the specific energy consumption will also increase. In terms of low energy consumption, a large pitch is desirable; however, the maximum possible drying rates are limited due to increasing temperatures. Finally, the influence of the nozzle distance is shown in Fig. 1.27 where, the lower the distance, the lower is the energy consumption. For a given air temperature with lower distances, higher drying rates can be achieved.

1.6 Conclusion

Fig. 1.26 Influence of the pitch for perforated plates at different drying rates.

Fig. 1.27 Influence of the nozzle–surface distance for perforated plates.

1.6 Conclusion

The highest convective heat transfer coefficients in drying processes can be achieved using nozzle fields, and drying is strongly intensified in such a field. As a consequence, the heat transfer area, and therefore the size of the apparatus, can be reduced, saving investment costs. The energy required to heat the drying air decreases with smaller nozzle diameters and slower outflow velocities, and consequently the operation costs are reduced. The lowest possible value of the nozzle diameter is, however, limited to ensure that the distance between the nozzle and material is not below four diameters. The lower the air velocity, the higher the hot air temperature must be. The temperature of the hot air is limited to 500  C due to the strength of steel; however, the hot air temperature must normally be kept much lower in order to protect sensitive products against thermal damage. The pitch of the nozzles has an optimum value of six diameters for

23

24

1 Impinging Jet Drying

the heat transfer with regards to single-nozzle arrays and hole channels. With regards to perforated plates, the optimum value for heat transfer occurs at pitches in the range of four, while the minimum for the specific energy consumption is achieved with a nozzle pitch of between eight and ten.

Additional Notation Used in Chapter 1

D d F _ H h m s t

diameter of cylindrical drying body inner diameter of nozzle square-root of nozzle opening area to total area enthalpy flow rate nozzle distance from product exponent of Reynolds number slot width, thickness pitch between nozzles

m m Js m

1

m m

Subscripts

lo s st 0

local value saturation (assumed product condition) stagnation point/region outlet of heater (entrance of dryer)

References Adler, W., 2004. Experimentelle Bestimmung channels. Diss., Otto von Guericke des W€arme€ ubergangs bei der Prallstr€omung University Magdeburg, Germany. € ber einen hohen Reynoldszahlenbereich u Attalla, M., Specht, E., 2009. Heat transfer mittels Infrarot-Thermografie. Diss., Otto characteristics from in-line arrays of free von Guericke University Magdeburg, impinging jets. Heat Mass Transfer 45(5): Germany. 537–543. Angioletti, M., Di Tommaso, R. M., Nino, Chan, T. L., Leung, C. W., Jambunathan, K., E., Ruocco, G., 2003. Simultaneous Ashforth-Frost, S., Zhou, Y., Liu, J. H., 2002. visualization of flow field and evaluation Heat transfer characteristics of a slot jet of local heat transfer by transitional impinging on a semi-circular convex surface. impinging jets. Int. J. Heat Mass Transfer Int. J. Heat Mass Transfer 45: 46: 1703–1713. 993–1006. Attalla, M., 2005. Experimental investigation of Geers, L. F. G., Tummers, M. J., Bueninck, T. heat transfer characteristics from arrays of J., Hanjalic, K., 2008. Heat transfer free impinging circular jets and hole correlation for hexagonal and in-line arrays

References of impinging jets. Int. J. Heat Mass Transfer 51: 5389–5399. Gori, F., Bossi, L., 2003. Optimal slot height in the jet cooling of a circular cylinder. Appl. Therm. Eng. 23: 859–870. Heikkil€a, P., Milosavljevic, N., 2002. Investigation of impingement heat transfer coefficient at high temperatures. Drying Technol. 20(1): 211–222. Huber, A. M., Viskanta, R., 1994. Effect of jet-jet spacing on convective heat transfer to confined, impinging arrays of axisymmetric jets. Int. J. Heat Mass Transfer 37: 2859–2869. Martin, H., 1977. Heat and mass transfer between impinging gas jets and solid surface, in Advances in heat transfer, (eds J. R. Hartnett, T. F. Irvine), Academic Press, New York, USA, pp. 1–59.

McDaniel, C. S., Webb, B. W., 2000. Slot jet impingement heat transfer from circular cylinders. Int. J. Heat Mass Transfer 42: 1975– 1985. Mujumdar, A. S. (ed.), 2007. Handbook of industrial drying, CRC Press, Boca Raton, USA. Nada, S. A., 2006. Slot/slots air jet impinging cooling of a cylinder for different jets cylinder configurations. Heat Mass Transfer 43(2): 135–148. Olsson, E. E. M., Ahrne, L. M., Tr€agardh, A. C., 2004. Heat transfer from a slot air jet impinging on a circular cylinder. J. Food Eng. 63: 393–401. Vader, D. T., Incropera, F. P., Viskanta, R., 1991. Local convective heat transfer from a heated surface to an impinging, planar jet of water. Int. J. Heat Mass Transfer 34: 611–623.

25

27

2 Pulse Combustion Drying Ireneusz Zbicinski, Tadeusz Kudra, and Xiangdong Liu 2.1 Principle of Pulse Combustion

Pulse combustion is an advanced technology of significant potential in thermal processing, since it is claimed to increase productivity, maximize utilization of the input fuel, and reduce pollutant emissions in comparison with conventional (continuous) combustion (Kudra et al., 1994; Zbicinski et al., 2002; Mujumdar and Wu, 2004; Kudra and Mujumdar, 2007). Because of enhanced momentum, heat and mass transfer, pulse combustion drying has been identified by the European Process Intensification Center (EUROPIC) as one of 60 promising process intensification technologies. To make it available to drying professionals, the technology report on pulse combustion drying of EUROPIC (code 4.1.4) has been reproduced in the Drying Technology Journal (Kudra, 2008). The term “pulse combustion” originates from the intermittent (pulse) combustion of solid, liquid, or gaseous fuel in burners of special design, consisting basically of the combustion chamber and the diffuser, also termed the tailpipe. Such periodic combustion generates intensive pressure, velocity and, to a certain extent, temperature waves propagated from the combustion chamber through a tailpipe to the process volume (an applicator) such as a dryer, calciner, or incinerator. Due to the oscillatory nature of the momentum transfer, pulse combustion intensifies the rates of heat and mass transfer thus affects greatly such processes as thermal drying. The mechanism behind the operation of a pulse combustor is a complex interaction between an intermittent combustion process and pressure/velocity waves that are propagated from the combustor (Fig. 2.1). The process of pulse combustion is initiated when air required for combustion and fuel in the form of a gas jet or a liquid spray are admitted to the combustion chamber to make-up an explosive mixture, which is ignited by a spark plug and burns instantly in an explosion-like manner. At this moment, the air and fuel inlet ports are closed, either by mechanically operated valves or due to the hydrodynamic action of the rapidly rising pressure. The combustion-generated pressure forces the combustion products to flow out through the tailpipe to the process volume. As the hot flue Modern Drying Technology Volume 5: Process Intensification, First Edition. Edited by Evangelos Tsotsas and Arun S. Mujumdar. Ó 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

28

2 Pulse Combustion Drying

Fig. 2.1 Operation of a pulse combustor with flapper valves.

gases discharge, the resulting outward momentum causes the pressure in the combustion chamber to decrease (Fig. 2.2). When the pressure reaches its minimum, the inlet ports open and admit fresh fuel and air into the combustion chamber. This new charge ignites itself due to contact with remnants of hot flue gases left in the tailpipe from the preceding cycle that reenter the combustion chamber during the minimum pressure period. These combustion cycles repeat themselves at a definite frequency, which depends on the design of the pulse combustor and characteristics of the tailpipe. In addition, the startup fan is normally shut off once self-aspiration is established. Typically, pulse combustors operate at frequencies from 20 to 250 Hz. Pressure oscillation in the combustion chamber of 10 kPa produces a velocity oscillation in the tailpipe of about 100 m s 1, so that the instantaneous velocity of a gas jet at the tailpipe exit varies from 0 to 100 m s 1 (Keller et al., 1992). The amplitude of the pressure rise may vary from 10% (domestic heating applications) to 100% as for

2.1 Principle of Pulse Combustion

Fig. 2.2 (a) Theoretical and (b) experimental pressure trace in a pulse combustor. A: air and fuel enter the combustion chamber; B: fresh charge ignited, pressure rises as combustion gases heat up, air and fuel inflow are stopped;

C: combustion complete, pressure decreases as flue gases are vented; D: momentum of exhausting gases creates negative pressure in the combustion chamber.

heavy-duty pulse combustors for industrial use (Kentfield, 1993). The output power for commercially available pulse combustors ranges from 70 to 1000 kW. Numerous studies on a variety of pulse combustor designs have demonstrated that pulse combustion offers numerous benefits over continuous combustion, of which the most important are (Stewart et al., 1991; Keller et al., 1992; Ozer, 1993; Kudra, 2008):    

increased heat and mass transfer rates (by a factor of 2 to 5); increased combustion intensity (by a factor of up to 10); higher combustion efficiency with low excess air values; reduced pollutant emissions (especially NOx, CO, and soot) by a factor of up to 3;

29

30

2 Pulse Combustion Drying

Fig. 2.3 Typical emissions from pulse combustor (courtesy of Novadyne Ltd, Hastings, ON, Canada).

 improved thermal efficiency by up to 40%; and  reduced space requirements for the combustion equipment. An added benefit of pulse combustion is its contribution to environment protection. The rapid combustion, which allows extremely short times for the formation of nitrogen oxides, and lower peak temperatures as compared to continuous combustion, result in very low NOx emissions from the pulse combustors (Fig. 2.3). This is especially advantageous when drying foods and biomaterials, provided that the pulse combustor is not fed with sulfur-containing fuels such as propane or natural gas (Kudra, 1998; Kudra, 1999). Recent studies on pulse combustion have indicated that the design of the combustion chamber and the geometry of a combustor–dryer system may affect the level of noxious emissions. Noticeably lower concentrations of CO, NO and NOx were obtained for those combinations of the volume of the combustor chamber and length of the tailpipe that provide smoother and sinusoidal pressure fluctuations (Zbicinski et al., 1999). The influence of the tailpipe length and fuel flow rate on NOx emission is presented in Fig. 2.4. A higher residence time in longer tailpipes results in increased NOx emission, and the same effect is observed for the fuel flow rate. During measurements performed at optimal geometry and operating conditions, the NOx level ranged between 5 and 98 ppm and the CO emission between 3 and 40 ppm. When the air-to-fuel ratio was lower than the stoichiometric ratio, leading to incomplete combustion, the CO emission was over 20 000 ppm (Smucerowicz, 2000). The results of experiments showed that the pulse combustor could function over a wide range of air-to-fuel ratios, but that the optimal range from the low-emissions viewpoint was narrow. With regards to other gaseous emissions in drying applications, the high temperatures of combustion gases when in contact with the wet material fed into the tailpipe may result in the destruction of odor, toxic contaminants, and other

2.1 Principle of Pulse Combustion

Fig. 2.4 Effect of tailpipe length and fuel flow rate on NOx emission.

volatile organic compounds (VOCs), either by thermal dissociation or by freeradical oxidation reactions (Rafson, 1998; Guy et al., 1997). From another viewpoint, concerns of odors are lessened in pulse combustion dryers where the wet material is injected directly to the tailpipe; this occurs as the result of an extremely short residence time (fractions of a second), as well as the effect of evaporative cooling which keeps the product temperatures sufficiently low so as to prevent the release of volatile substances. The major disadvantage of pulse combustion systems is the extreme noise generated under normal operation, which may be a deterrent factor in industrial applications. The sound pressure level (SPL) depends on the operating pressure and frequency; the higher the operating pressure, the higher the SPL. From a technical viewpoint, large pressure amplitudes are desirable in processes such as drying because of increased rates of heat and mass transfer, and thus a reduced size of the burner with the same heat output. From a practical viewpoint, however, noise emission may be a much more important design criterion than minimizing the burner size. Although the SPL at the tailpipe outlet attains 110 dB (Zbicinski, 2002), this noise is greatly attenuated in a drying chamber, with the major source of emitted noise occurring at the inlet to the combustor. As indicated by Duncan (2012), building the combustor from thick-walled Inconel and encasing the entire combustor in an outer shell will reduce the noise level to less than 85 dB. Mounting a sound suppressor built as a tube-in-tube segment, where the in-between tubes volume is filled with a porous material, will reduce the sound level from 160 dB to 140 dB and 90 dB, respectively, when the suppressor is attached at the tailpipe outlet and the combustor inlet.

31

32

2 Pulse Combustion Drying

Fig. 2.5 Design principle of pulse combustor with rotary valve (after Lockwood, 1987).

Another solution to the sound problem is a rotary-valve combustor enveloped by a steel shell, as shown in Fig. 2.5. Because, for spray-drying applications, such a combustor is running out of its resonant frequency, the sound level will be less than 85 dB at a distance of 1 m from the dryer (Rehkopf, 2012).

2.2 Pulse Combustors: Design and Operation

Pulse combustors may be categorized into three distinct classes according to the specific acoustic system on which their operation depends: (i) the Quarter-wave (or Schmidt) combustor; (ii) the Helmholtz combustor; and (iii) the Rijke-type combustor. In contrast to the Rijke combustor, which operates with solid fuels, both the Schmidt and Helmholtz combustors accept liquid and gaseous fuels. The Helmholtz combustor is preferred for drying applications because the larger volume of the combustion chamber and the smaller (but longer) tailpipe allows for multivalve assembly. Detailed information on these types of combustor is available in articles by Zinn (1985) and Kudra and Mujumdar (2009). Some combustors also exploit the resonance phenomenon; these are referred to as frequency-tunable pulse combustors. 2.2.1 Pulse Combustors with Mechanical Valves

Based on the manner in which fuel and air charge the combustion chamber, pulse combustors can be divided into two general categories: (i) those with mechanical valves; and (ii) those with aerodynamic valves (also called “valveless” combustors).

2.2 Pulse Combustors: Design and Operation

Fig. 2.6 Mechanical valves. (a) Flapper-type; (b) Reed-type with petal-shaped lamella; (c) Reedtype with rectangular lamella.

Mechanical valves can be further divided into three subtypes of flapper valves, reed valves, and rotary valves. Since mechanical valves provide a physical barrier to the backflow of combustion products through the combustor inlet during the positive-pressure phase of the pulse combustion cycle, the unidirectional flow is the fundamental feature of valved pulse combustors. There are, however, certain problems associated with the design of mechanical valves, such as minimizing valve inertia, protection from corrosion, and resistance to material fatigue due to thermal stress. These specific problems are of major importance in heavy-duty pulse combustors operated at large pressure amplitudes (Kentfield, 1993). The basic consideration in the design of the flapper (membrane) valves, as shown schematically in Fig. 2.6a, is the ability to rapidly open and close the openings in the base plate according to the combustor operating frequency. Moreover, the valves must withstand the maximum pressure in the combustion chamber, yet open with only a small pressure difference. The reed valves normally used in heavy-duty pulse combustors are made from thin-sheet spring-steel (Fig. 2.6b and c), and the spring action of reed valves is such that, when normally shut, the valves are sprung lightly. In order to ensure a vigorous mixing of the fuel and air, the fully open flow area of the inlet reed valves must be considerably smaller than the cross-sectional area of the combustion zone. The major problem often encountered with reed-type mechanical valves is fatiguebased failure. A schematic diagram of a pulse combustor with flapper valves is shown in Figure 2.7. In contrast to flapper and reed valves, the operation of rotary valves is independent of the combustor design, so they can accommodate a broad range of combustor frequencies and firing rates. Other important features of rotary valves include durability, flexibility of use, and resistance to oil and dirt accumulation. Generally, the rotary valve is built on two plates: a motor-driven “butterflyshaped” rotating plate; and a stationary plate with two (or more) slots displaced by

33

34

2 Pulse Combustion Drying

Fig. 2.7 Pulse combustor with flapper valves (after Kitchen, 1987).

180 (or less, depending on the number of slots in the stationary plate). The air for combustion passes through the valve, perpendicular to the rotating plate, and enters the combustion chamber. The height and width of the slots on the stationary plate determine the open area for airflow through the rotary valve; alternatively, the plates may have circular or other shaped openings. The second design of rotary valve is based on three coaxially disposed sleeves with a plurality of slots in their lateral surface (c.f., Fig. 2.5). When rotated by a motor with controllable speed, the slots periodically admit the intake air at the rate required for optimum combustion, whereas the fuel is typically injected via the nozzle. In such a design, as the rotation of the valve determines the running frequency of the pulse combustor, an adequate “feedback system” is required to synchronize the rotation of the valve with the resonant frequency of the burner. The speed of the valve, which influences the resonant frequency, is determined through the sensing and processing of pressure pulses in the combustion chamber. 2.2.2 Pulse Combustors with Aerodynamic Valves

Aerodynamic valves employ the properties of the fluid entering a specially designed inlet to the combustor to create an artificial barrier to the backflow of combustion products out of the combustor through its inlet section. The main advantage of aerodynamic valves is a lack of mechanical parts that are prone to failure. The design concept of aerodynamic valves exploits the principle of a fluid diode which are, of course, inferior in performance when compared to mechanical valves as the backflow of combustion products cannot be fully eliminated. One mechanism that is known to limit the amount of backflow is to include a nonuniform crosssectional area. A tapered inlet that diverges gradually towards the combustion chamber initially accelerates the inlet stream of air and then diffuses it before

2.2 Pulse Combustors: Design and Operation Tab. 2.1

Operational characteristics of pulse combustor fed with various fuels (Wu, 2007).

Parameter

Propane

Ethanol

Fuel oil

Frequency, Hz Pressure amplitude, Pa Average temperature in chamber, K Average gas velocity, m s 1 Fuel flow rate, g s 1 Combustion heat, MJ kg 1 Power output, kW Average trust, N Specific impulse, Ns kg 1 Trust/power output, N kW 1

116 5082/þ6035 2094 16.55 0.1929 46.36 8.95 0.4244 2200 0.0474

120 6351/þ6800 2106 15.69 0.3077 28.0 8.64 0.4018 1310 0.0465

112 5283/þ6404 2070 17.67 0.2162 40.53 8.76 0.4595 2130 0.0524

Fig. 2.8 Configuration of University of Calgary pressure-gain valveless pulse combustor (courtesy of J. A. C. Kentfield).

entering the combustion chamber. Such a tapered inlet acts effectively as a nozzle restricting the backflow of combustion products. Figure 2.8 presents, as an example, the schematics of a reverse-flow, pressure-gain, valveless pulse combustor designed at the University of Calgary. Details for the designs of valves and various pulse combustors can be found elsewhere (Kentfield, 1993; Speirs, 1989; Olorunmaie and Kentfield, 1989). The performance characteristics of a flapper valve pulse combustor fed with various fuels are listed in Tab. 2.1 (Wu, 2007). 2.2.3 Frequency-Tunable Pulsed Combustors

Pulsed combustors usually operate at a frequency which is much lower than the inherent frequency, often controlled by an ignition, fuel injection, or a valve open/

35

36

2 Pulse Combustion Drying

close sequence. Thus, valveless or flapper valve combustors fall into a category of pulse combustors while mechanically driven valves (e.g., rotary valve) used to control either air or fuel inflow, flue gas discharge (or both) are categorized as frequency-tunable pulse combustors. One possible design of a frequency-tunable pulse combustor takes advantage of the natural non-longitudinal acoustic modes of the process chamber, such as a dryer or incinerator (Zinn and Daniel, 1988). In order to obtain maximum benefit from the pressure/velocity oscillations in the process chamber, the pulse combustor can be tuned to one (or more) of these acoustic modes by modulating the flow of fuel to the combustion chamber. Such modulation may be accomplished by exciting the acoustic resonance within the fuel line, or via periodic interruptions of the fuel flow, using a rotary valve. The frequency-tunable pulse combustor combines a combustion zone and exhaust zone. The reaction between fuel and air, which takes place in the combustion zone, induces an acoustic wave in the combustor that is then used to control the flow of fuel and air into the combustion chamber (Lockwood, 1987). The acoustic characteristics of the combustor, which may deliberately be altered to provide a selectively variable (or frequency-tunable) pulsating combustion, are briefly described by Kudra and Mujumdar (2009).

2.3 Aerodynamics, Heat and Mass Transfer

Since transport phenomena in the pulse combustion chamber are beyond the scope of this book, the following sections will be restricted to momentum, and to heat and mass transfer between the combustion gases discharged through the tailpipe and the particles of a drying material or droplets of atomized liquid. Even though the microscopic effects in pulse combustion drying have been well identified and exploited in industrial dryers owing to the professional experience of dryer manufacturers and field testing (e.g., Rehkopf, 2012; Duncan, 2012), an insufficient understanding of the process mechanism and a scarce mathematical description of the transfer phenomena (Wu and Mujumdar, 2006a; Xiao et al., 2008; Liewkongsataporn et al., 2008) has hampered the science-based dryer design and scale-up. Aside from exploratory research by Kudra et al. (1994) to compare drying rate of solid wood in pulsed and continuous combustion, as well as recent studies on the pulse combustion drying of a paper sheet (Liewkongsataporn et al., 2006; Liewkongsataporn et al., 2008), the majority of publications have described the pulse combustion drying of dispersed materials (Kudra and Mujumdar, 2009). Thus, the following sections will provide an overview of the investigations conducted on transfer phenomena in droplets and granules exposed to the oscillating flow of pulse combustion gases.

2.3 Aerodynamics, Heat and Mass Transfer

2.3.1 Atomization

Following on from the aforementioned characteristics, an advantage of pulse combustion is that the strongly oscillating hot-gas jet from the combustion chamber, and also from the tailpipe, can promote the dispersion of liquids, pastes and clumps of particulate materials (Wu et al., 2012). Hence, in the case of liquids, pulse combustion spray-dryers can refrain from using conventional disk or nozzle atomizers, which translates into a simple design and lower capital and operating costs, as well as an enhanced product quality due to the creation of finer droplets with a narrow size distribution. The process of liquid atomization by the oscillating gas jet emerging from a Helmholtz-type pulse combustor was studied experimentally for water and aqueous maltose solutions with different viscosities (Xiao et al., 2008). The frequency of the oscillating flow of combustion gases ranged from 61 to 100 Hz, while the pressure spanned from 93.7 to 114.8 kPa with an average amplitude of about 21 kPa. Because temporal variations of temperature and velocity were difficult to measure for such high pulse frequency fluctuations, the cycle-averaged values of air flow temperature and velocity were used in the data interpretation. The air temperature in the tailpipe was maintained at 423  10 K using a water-cooled system. The oscillating air flow transient velocity was between 100 m s 1, and the mean velocity about 26.8 m s 1. The droplet size was presented in terms of Sauter mean diameter (SMD), which is the diameter of a droplet having the same volume-tosurface area ratio as the ratio of the total volume of all droplets to the total surface area of these droplets. Figure 2.9 presents the effect of liquid feed rate, liquid viscosity and oscillating frequency on the SMD of the dispersed liquid. Effects of all three parameters were evident with, most notably, larger droplets being obtained for a higher liquid feed rate. When the liquid feed rate was below 28 l h 1, however, the atomized droplets were so small that they vaporized immediately and the optical analyzer had insufficient time to capture their image, and consequently all subsequent tests were performed at a liquid feed rate of 35 l h 1. Smaller droplets were obtained at a higher frequency of gas flow oscillation, with the best atomizing effect being achieved at a liquid feed rate of 35 l h 1 and the gas flow oscillating at a frequency of 100 Hz. The experimental results presented in Fig. 2.9 show that the liquid viscosity will have a major impact on the characteristics of atomization. An inversion point was observed at a viscosity of 0.007 Pa
s whilst, at a relatively low viscosity ( 0.05). HAD: Hot air drying; HPD: Heat pump drying; IMC: Intermediate moisture content.

If there is a desire to further improve the quality of a dried product, then a multistage drying process may be employed. Such a process can be realized by conducting SSD during the first stage, followed by a second-stage drying process at a lower temperature to avoid exposing the drying material to the high-temperature drying medium (viz. superheated steam) for an extended period of time. Namsanguan et al. (2004) validated the aforementioned concept by drying shrimp in a multistage fashion. The process started with SSD at 140  C, followed by heat pump drying (HPD) at 50  C. The results were compared with those obtained through first-stage SSD at 140  C and then hot-air drying (HAD) at 50  C; single-stage SSD at 140  C was also conducted, and the results were compared. These investigators also studied the effect of varying the intermediate moisture content (IMC), which is the moisture content of shrimp at the end of the first drying stage (in this case SSD), to determine how the drying time of the different drying processes would affect the overall drying kinetics and the quality of the dried product. Indeed, a higher IMC implied that the time spent by the material during the first drying stage is shorter. All of the tested conditions are listed in Tab. 3.2. Figure 3.8a, which illustrates the drying kinetics of shrimp undergoing different drying process combinations, clearly shows that the moisture content of shrimp decreased most rapidly if only SSD was applied (Case 1). This is because the shrimp was exposed to the high-temperature superheated steam throughout the whole period of drying. In the cases of multistage drying, however, a higher IMC resulted in a longer overall drying time as the period with the higher drying rate (SSD) was shorter, as mentioned earlier. Nevertheless, it is interesting to see from Fig. 3.8b that SSD/HPD required as much as a 1.4 to 1.5-fold shorter drying time than SSD/HAD at the same IMC. This point is ascribed to the fact that HPD involves drying at a much lower air relative humidity as compared to HAD. In terms of the dried product quality, it was noted that the shrimp which had undergone multistage drying (especially SSD/HPD) suffered significantly less shrinkage than the shrimp dried only by SSD (see Tab. 3.2). This was due to the fact that superheated steam-dried shrimp was exposed to the high-temperature

3.3 Atmospheric-Pressure Superheated Steam Drying

Fig. 3.8 (a) Drying curves and (b) drying times of shrimp at different conditions. Modified from Namsanguan et al. (2004).

drying medium for an extended period of time, leading to a collapse of the porous structure that had been formed as a result of water vaporization during an early period of SSD; in the case of SSD, the shrimp temperature eventually reached a value which was much higher than the glass transition temperature,

67

68

3 Superheated Steam Drying of Foods and Biomaterials

Fig. 3.9 Microstructural images of shrimp dried at different conditions. (a) Only SSD at 140  C; (b)SSDat140  CfollowedbyHPD(IMC40%w.b.);(c)SSDat140  CfollowedbyHPD(IMC30%w.b.). Reproduced with permission from Namsanguan et al. (2004).

and this resulted in a rather severe phase transition and collapse. On the other hand, if the partially superheated steam-dried shrimp, which still possessed an intact high-porosity structure, was subsequently transferred into a lowertemperature drying environment, the structure of shrimp would be much better preserved, with less shrinkage. This hypothesis was indeed verified by the microstructural images of shrimp shown in Fig. 3.9, which clearly show that the shrimp which had undergone SSD/HPD at a lower IMC exhibited a higherporosity structure than the shrimp that had been dried only by SSD. However, it can be seen from the data in Tab. 3.2 that the shrimp which undergone SSD/ HAD had a similar (or even higher) level of shrinkage than those dried only by SSD. This may be because SSD/HAD was a relatively slow process compared to SSD and SSD/HPD, especially when the IMC was higher. A longer drying process is indeed known to lead to a greater degree of structural collapse and shrinkage of the drying material. In terms of color, shrimp subjected to SSD/HPD was seen to possess a redder color than shrimp dried only by SSD. The shrimp dried by SSD/HPD was also softer; the reason for these observations being similar to those applied earlier to explain the collapse and shrinkage phenomena. One major advantage of using superheated steam to dry shrimp is that the boiling stage, which normally needs to be conducted prior to drying, can be

3.4 Low-Pressure Superheated Steam Drying (LPSSD)

eliminated, and the shrimp can simply be immersed in saline solution and introduced to SSD. Boiling and drying would then occur simultaneously. This represents an alternative approach to reducing the time and energy consumption of the entire dried shrimp production process. Moreover, dried shrimp obtained via such a process also taste better as the flavoring ingredients are not lost with the boiling solution.

3.4 Low-Pressure Superheated Steam Drying (LPSSD)

Since atmospheric-pressure (or near atmospheric-pressure) SSD involves the use of a high-temperature drying medium (i.e., superheated steam at a temperature higher than 100  C), this type of process may not be appropriate for heat-sensitive materials such as fruits, vegetables, herbs and spices, and many other biomaterials. For this reason, an alternative drying technique has been developed to combine the advantages of SSD with an ability to conduct drying at a lower temperature. The proposed alternative is based on the fact that water vaporizes at a lower temperature if the pressure of the system is lower; superheated steam can then be produced at a lower temperature under such a lower pressure. The so-called low-pressure superheated steam drying (LPSSD) process has been tested with a wide array of foods and biomaterials, with the first comprehensive test covering both the drying kinetics and quality of the dried product (carrots) having been conducted by Devahastin et al. (2004); a schematic diagram of the employed LPSSD set-up is shown in Fig. 3.10. This type of setup can be attached to a boiler, which is available in a typical food-processing plant. However, as a typical boiler generates saturated steam at a pressure of around 6–8 bar, the steam cannot be directly introduced into the drying chamber. A reservoir, along with a pressure regulator, is installed to maintain the saturated steam at around 1.5–2 bar prior to feeding it to the drying chamber. Once the saturated steam flows into the low-pressure drying chamber, the steam would spontaneously become superheated steam as the temperature of the entering steam is already higher than the saturation temperature of water at the pressure of the drying chamber. Nevertheless, to allow a more precise control of the temperature within the drying chamber, an electric heating coil is installed to help regulate the superheated steam temperature. At an industrialscale setting, this electric heating coil may be replaced by a coil through which hot oil or high-pressure steam can run. When carrots were subject to drying at 60–80  C over a pressure range of 7–13 kPa, they were found to experience a slight increase in moisture content during the early period of drying, which is a typical characteristic of any superheated steam drying process. The initial condensation decreased on reducing the pressure of the drying chamber, as the reduced boiling point at a lower pressure (at a constant superheated steam temperature) led to a

69

70

3 Superheated Steam Drying of Foods and Biomaterials

Fig. 3.10 Low-pressure superheated steam drying (LPSSD) set-up. (1) Boiler; (2) Steam valve; (3) Steam reservoir; (4) Pressure gage; (5) Steam trap; (6) Steam regulator; (7) LPSSD chamber; (8) Steam inlet; (9) Electric fan; (10) Sample holder; (11) Electric heater;

(12) Temperature sensor and recorder; (13) Vacuum break-up valve; (14) Insulator; (15) Load cell and mass recorder; (16) Vacuum pump; (17) PC and data acquisition system. Reproduced with permission from Devahastin et al. (2004).

higher degree of superheating (i.e., the difference between the superheated steam temperature and saturation temperature of water at a specific pressure). In addition, the drying temperature was noted to have a more significant effect on the drying kinetics of the product than did the pressure, notably because the temperature has a more dominant effect on the steam thermophysical properties than the pressure. Nevertheless, the drying rate of carrots increased as the pressure of the drying chamber decreased (at a constant superheated steam temperature), as the boiling point of water decreased at a lower pressure and this resulted in a faster removal of water from the carrots. If a comparison is made between LPSSD and vacuum drying (VD), the time needed to dry a material using VD would be less than that required for LPSSD at the same temperature and pressure. This difference can be ascribed to the fact that the material receives more energy from the heater in the case of VD because the heater is the sole energy source for drying; thus, the material absorbs more radiated energy from the heater and exhibits a higher temperature and, hence, a higher drying rate. Nevertheless, the difference in the required drying time would be smaller at a higher drying temperature, which is indeed a characteristic of the inversion phenomenon. In order to preserve the quality of the dried product, however, it is not always recommended to conduct LPSSD at the inversion temperature or higher; it should be noted that the inversion temperature in the case of LPSSD may be as high as 90  C (Suvarnakuta et al., 2005b). Figure 3.11 illustrates the drying curves and temperature evolution patterns of a material (carrots) undergoing LPSSD. These data show clearly that the drying temperature and pressure both have significant effects on the moisture and

3.4 Low-Pressure Superheated Steam Drying (LPSSD)

Fig. 3.11 Evolutions of moisture content and temperature of carrots during LPSSD. (a) 60  C, 7 kPa; (b) 70  C, 7 kPa; (c) 80  C, 7 kPa; (d) 80  C, 13 kPa. ~: moisture content; : steam temperature; : carrot temperature. Reproduced with permission from Devahastin et al. (2004).





temperature evolution patterns. In general, the material temperature would increase rapidly from the initial temperature to the saturation temperature of water corresponding to the pressure of the drying chamber, and would then remain at the saturation temperature until the end of a constant drying rate period. The material temperature would subsequently increase and approach the superheated steam temperature towards the end of drying. One important observation that can be made upon examining Fig. 3.11 is that when the superheated steam temperature increases (at a constant pressure), the period where the material temperature would remain constant (i.e., the constant drying rate period) would become shorter. In some cases where the superheated steam temperature is higher and the pressure of the drying chamber is lower (e.g., at a temperature of 80  C and a pressure of 7 kPa; see Fig. 3.11c), the material temperature would continuously increase and approach the superheated steam temperature without having a constanttemperature period. In contrast, if the superheated steam temperature is lower and the pressure of the drying chamber is higher, then the period of constant material temperature would be longer (see Fig. 3.11d). These findings confirm

71

72

3 Superheated Steam Drying of Foods and Biomaterials

that a constant drying rate period – or, in other words, the critical moisture content – is dependent on not only the type of material but also the drying conditions. Figure 3.12 illustrates sample drying curves and temperature evolution patterns of a material (again carrots) undergoing VD at the same drying conditions as those shown in Fig. 3.11. It can be seen that the moisture and temperature evolution patterns (or heat and mass transfer behavior) of the material are quite different from those of the sample undergoing LPSSD. In the case of VD, the material temperature continuously increases toward the drying temperature without having a constant-temperature period. However, the vacuum-drying rates are seen to be higher than the LPSSD rates at the same operating conditions, as mentioned earlier. In terms of dried product quality, the volume and density of the carrots changed in a similar manner upon both VD and LPSSD. However, although both vacuumdried and low-pressure superheated steam-dried carrots exhibited similar volumetric shrinkage (calculated from the change in the volume of a sample compared

Fig. 3.12 Evolutions of moisture content and temperature of carrots during vacuum drying. (a) 60  C, 7 kPa; (b) 70  C, 7 kPa; (c) 80  C, 7 kPa; (d) 80  C, 13 kPa. ~, moisture content; , drying temperature; , carrot temperature. Reproduced with permission from Devahastin et al. (2004).





3.4 Low-Pressure Superheated Steam Drying (LPSSD)

Fig. 3.13 Photographic images of dried and reconstituted (rehydrated) carrots.

to the initial volume of the sample prior to drying), the deformation patterns of the two sets of samples were quite different. Carrots which had undergone VD suffered a much more nonuniform shrinkage than those that had undergone LPSSD. This difference may be because VD led to a higher drying rate and, in turn, causing casehardening of the carrot surface before the inner part of the carrot could be dried. In other words, the inner carrot dried much more slowly than the surface, and this naturally led to a nonuniform shrinkage (see Fig. 3.13). It can be seen from the microstructural images shown in Fig. 3.14 that carrots dried by LPSSD exhibited a more uniform microstructure than those dried by VD. Notably, because the superheated steam-dried carrots deformed more uniformly, they were more easily reconstituted when immersed in water – a very desirable characteristic sought by the instant food industry. In addition to studies on the changes of various physical characteristics of materials undergoing LPSSD, several investigations have been made into the use of this drying technique to preserve various heat-sensitive bioactive compounds. For

Fig. 3.14 Microstructural images of carrots after (a) LPSSD and (b) vacuum drying. Reproduced with permission from Devahastin et al. (2004).

73

74

3 Superheated Steam Drying of Foods and Biomaterials

example, Methakhup et al. (2005) prepared dried Indian gooseberry (Phyllanthus emblica L.), which can be used as an ingredient of a herbal tea, by using both VD and LPSSD. Although LPSSD was noted to require a longer drying time, the technique resulted in a higher retention of vitamin C (ascorbic acid) in the dried gooseberry than did VD (see Tab. 3.3). In addition, the gooseberry dried by LPSSD exhibited much less of a color change than did gooseberries treated with VD. It can also be seen from the data in Tab. 3.3 that the change in the VD temperature significantly affected the ascorbic acid content and the color of the dried gooseberry, whereas the superheated steam drying temperature (and pressure) had no significant effects on the quality parameters of the dried gooseberry. These findings may help confirm the negligible oxygen environment of LPSSD, which led to a negligible oxidation of ascorbic acid and other phenolic compounds in the gooseberry, and thus resulted in smaller losses of ascorbic acid and less color change. Rather, the loss of ascorbic acid and color changes after LPSSD were considered due to thermal degradation and other reactions that do not involve oxygen, including nonenzymatic browning reactions (e.g., the Maillard reaction, which occurs between amino acids and reducing sugars). In addition to the retention of vitamin C in carrots, both VD and LPSSD can also significantly reduce losses of b-carotene when compared to hot air drying, as demonstrated by Suvarnakuta et al. (2005a) (see Tab. 3.4). This effect is, again, due to the lack of oxygen during VD and, in particular, during LPSSD. The retention of b-carotene in carrots undergoing VD and LPSSD did not differ greatly, unlike that of ascorbic acid, because the oxygen content plays a more significant role in the loss of the latter than of the former. Thus, it can be concluded that LPSSD has clear advantages over other competing drying techniques when applied to heat-sensitive and, in particular, oxygen-sensitive materials. From the consideration of food safety, which involves the inactivation or growthsuppression of microorganisms, LPSSD has been noted to be more effective than other commonly used drying techniques at the same drying temperature. For example, Phungamngoen et al. (2011) tested two different drying techniques that have been commonly applied to drying foods and biomaterials – namely hot air drying and VD – and compared these with LPSSD over a temperature range of 50 to 70  C. The three drying techniques were evaluated for their effectiveness in decontaminating a pathogenic bacterium (Salmonella Anatum), which was inoculated onto the surfaces of cabbage leaves. The results, shown in terms of the survival curves of S. Anatum (see Fig. 3.15), showed that LPSSD reduced the bacterial numbers faster than did hot air drying and VD, especially at 60 and 70  C, despite requiring 30% and 14% longer times than VD to reduce the cabbage moisture content to the desired final value at 60 and 70  C, respectively. This is because the temperature of the cabbage rose rapidly in the case of LPSSD. The initial condensation, which occurred only in the case of LPSSD, also helped increase (or at least maintain) the activity of water on the surfaces of the cabbage. As noted earlier, the inactivation of microorganisms is more effective under a high water activity condition (Chiewchan, 2011), and this renders LPSSD more effective for microbial inactivation than the alternative drying techniques that were tested.

76

3 Superheated Steam Drying of Foods and Biomaterials b-carotene contents of fresh carrot and carrots dried at different conditions. Modified from Suvarnakuta et al. (2005a).

Tab. 3.4

Sample

Fresh carrot

b-carotene content (mg per 100 g 1)

Retention (%)

Dry basis

Wet basis

4.86  0.65

51.11  6.89

—

54.92 50.32 38.81

43.94 41.24 31.55

83  2a 76  4b 74  2b

24.28 24.05 27.85

74  3c 71  3c,d 68  3d

21.46 23.30 22.05

62  2e 62  3e 58  3e

Carrots dried by LPSSD 60  C 70  C 80  C

Carrots dried by vacuum drying 60  C 70  C 80  C

26.95 25.81 31.61

Carrots dried by hot air drying 60  C 70  C 80  C

22.56 35.66 36.26

Same letters in the same column indicate that the mean values are not significantly different when considered at a confidence level of 95% (p > 0.05).

Figure 3.16a shows that S. Anatum contaminated on the surface of raw cabbages exhibited an intact rod shape, and the bacterial cells were distributed quite uniformly on the leaves. Upon drying, however, the cells suffered significant morphological changes, especially in the cases of VD and LPSSD (see Figs. 3.16c and d).

3.5 Application of LPSSD to Improve the Quality of Foods and Biomaterials

It is generally recognized that foods and biomaterials suffer quality degradation, in one way or another, upon drying. However, if an appropriate drying technique, possibly combined with appropriate drying conditions, can be employed then it might be possible to improve the quality or selected properties of foods and biomaterials, and perhaps even make the dried products superior to the fresh materials. Some examples of the use of LPSSD, in combination with other assisting drying processes or heating methods, are briefly reviewed in the present section. Emphasis is placed on the production of fat-free snacks that currently attract attention from health-conscious consumers as an alternative to deep-fried snacks. Highly crisp, fat-free snacks may be produced through the combined use of LPSSD and far-infrared radiation (FIR). For example, Nimmol et al. (2007a, 2007b) applied FIR-assisted LPSSD (LPSSD-FIR) to produce dried banana chips, while FIR-assisted vacuum drying (VACUUM-FIR) and LPSSD were also evaluated over

3.5 Application of LPSSD to Improve the Quality of Foods and Biomaterials

Fig. 3.15 Survival curves of Salmonella Anatum contaminated on cabbage surfaces undergoing different drying processes and conditions (log S is the log of the number of bacteria at any time t to

the number of bacteria at t ¼ 0). (a) Hot air drying; (b) Vacuum drying; (C) LPSSD. &: 50  C; ^: 60  C; : 70  C. Reproduced with permission from Phungamngoen et al. (2011).

Î

the drying temperature range of 70–90  C at pressures of 7 and 10 kPa. The banana temperature (see Figs 3.17 and 3.18) was seen to slightly decrease from the initial temperature during an early period of drying. This was due to a reduction in the pressure of the drying chamber, leading to a flash evaporation of the surface moisture and hence a cooling of the banana. Subsequently, the banana temperature in the case of LPSSD-FIR rapidly increased and approached (but did not equal) the saturation temperature of water at the dryer operating pressure. The temperature evolution pattern in this case was quite different from that of LPSSD, where the material temperature increased to the saturation temperature of water at

77

78

3 Superheated Steam Drying of Foods and Biomaterials

Fig. 3.16 Scanning electron microscopic images of Salmonella on cabbage surfaces. (a) Prior to drying (fresh cabbage); (b) After hot air drying for 6 h; (c) After VD for 1 h; (d) After

LPSSD for 1 h. The drying temperature was 60  C in all cases. Reproduced with permission from Phungamngoen et al. (2011).

the dryer operating pressure. This difference was due to the use of FIR as an extra source of energy for drying; drying is no longer purely convective when the FIR is present. When a constant drying rate period had elapsed, the banana temperature was noted to increase to a value that was higher than the superheated steam temperature; this was again due to the existence of FIR. Nevertheless, the banana temperature did not rise further as the fruit possessed much less moisture at this point and hence was unable to absorb any more radiated energy. It was noted further that, at the same drying temperature, the banana temperature towards the end of the drying process in the case of LPSSD-FIR (Fig. 3.17) was higher than that in the case of VACUUM-FIR (Fig. 3.18). This was due to the fact that during LPSSD-FIR, the radiation intensity at the position of a temperature sensor used to measure the medium temperature (which was in turn used to control the infrared radiator) was lower than that in the case of VACUUM-

3.5 Application of LPSSD to Improve the Quality of Foods and Biomaterials

Fig. 3.17 Evolutions of moisture content and temperature of bananas during LPSSD-FIR at different conditions. (a) 80  C, 7 kPa; (b) 80  C, 10 kPa; (c) 90  C, 7 kPa; (d) 90  C, 10 kPa. :

moisture content; : drying temperature; : banana temperature. Reproduced with permission from Nimmol et al. (2007b).

Î

FIR. This situation occurred because part of the radiated energy had been absorbed by the superheated steam, which possesses a higher ability to absorb FIR than does the air. For this reason, in the case of LPSSD-FIR, the radiator must be switched on more frequently in order to maintain the required radiation level, leading to a higher surface temperature of the banana than in the case of VACUUM-FIR. In terms of dried product quality, banana chips produced by LPSSD-FIR exhibited a darker color than those produced by VACUUM-FIR in all cases. This was because the banana temperature during LPSSD-FIR was higher than that during VACUUM-FIR, as mentioned earlier. The higher temperature expectedly led to accelerated rates of browning reactions, and hence to darker colors. It should be noted that the dominant browning reactions were nonenzymatic in nature, as the amount of oxygen in the drying systems was very low (almost negligible in the case of LPSSD-FIR). In addition, when comparing the processes with and without FIR, it was noted that the color of banana chips which had undergone the processes without FIR suffered less changes. This

79

80

3 Superheated Steam Drying of Foods and Biomaterials

Fig. 3.18 Evolutions of moisture content and temperature of bananas during VACUUM-FIR at different conditions. (a) 70  C, 7 kPa; (b) 80  C, 10 kPa; (c) 90  C, 7 kPa; (d) 90  C, 10 kPa. :

moisture content; : drying temperature; : banana temperature. Reproduced with permission from Nimmol et al. (2007b).

Î

was again related to the changes in the material temperature during drying (Nimmol et al., 2007a; Thomkapanish et al., 2007). Interestingly, the radial shrinkage of bananas subjected to lower-temperature drying (e.g., at 70 and 80  C) was noted to be less than that of bananas undergoing drying at a higher temperature. This was ascribed to the fact that casehardened skins had been formed during lower-temperature drying, and these had helped to retard the radial shrinkage. Both LPSSD-FIR and VACUUM-FIR were also noted to lead to a more extensive formation of the casehardened skins than LPSSD alone, again due most likely to the different temperature evolution patterns of the processes with and without FIR (as mentioned earlier). In terms of rehydration behavior, which is an important characteristic of any instant food, the banana chips produced by drying at a higher temperature were better able to absorb water than those produced at a lower temperature. This was because high-temperature drying had led to chips with a higher porosity as a result of a more intensive vaporization of the internal water (as illustrated comparatively in Fig. 3.19). LPSSD-FIR was noted to produce chips with a better rehydration ability than VACUUMFIR for reasons related to the changes in the material temperature and the internal

3.5 Application of LPSSD to Improve the Quality of Foods and Biomaterials

Fig.3.19 Microstructuralimagesofbananachipsdriedatdifferentconditions.(a)LPSSD-FIRat80  C and 7 kPa; (b) VACUUM-FIR at 80  C and 7 kPa; (c) LPSSD-FIR at 90  C and 7 kPa; (d) VACUUM-FIR at 90  C and 7 kPa. Reproduced with permission from Nimmol et al. (2007a).

vaporization of water (as described in Section 3.4). However, when drying was performed at a higher temperature of 90  C, the chips prepared by both processes exhibited similar rehydration behaviors. This most likely occurred because the levels of internal water vaporization of both groups of samples were no longer different, and this led to insignificantly different porosities of the chips (Nimmol et al., 2007a). The results of textural evaluations of banana chips prepared under different conditions are listed in Tab. 3.5. These results are expressed in terms of the maximum force in a force-deformation curve, which is a representative of the hardness of the samples. The results are also expressed in terms of the number of peaks in the force-deformation curve, which is a representative of the crispness of the samples. It can be seen that chips prepared by VACUUM-FIR were harder than those prepared by LPSSD-FIR. The difference was due to variations in the microstructure of the chips, as illustrated in Fig. 3.19. The drying temperature and pressure nevertheless had no significant effect on the hardness of the chips. In terms of crispness, LPSSD-FIR resulted in chips with higher numbers of peaks (and hence were crispier) than did VACUUM-FIR, especially when drying was conducted at a lower temperature, such as 80  C. This was again due to the more extensive porous structure of the chips produced by LPSSD-FIR. When combining LPSSD with FIR, it is clear that the hardness of the chips significantly decreased, whereas the crispness was significantly increased

81

82

3 Superheated Steam Drying of Foods and Biomaterials Texture of banana chips dried at different conditions. Modified from Nimmol et al. (2007a); LPSSD data from Thomkapanish et al. (2007).

Tab. 3.5

Drying method

Temperature ( C)

Pressure (kPa)

Maximum force (N)

Number of peaks

LPSSD-FIR

70

7 10 7 10 7 10 7 10 7 10 7 10 7 7 7

N/A N/A 17.09  3.15a 13.30  3.60a 16.39  3.57a 16.89  4.58a 18.44  3.80a 19.12  4.07a 19.95  3.55a 18.16  4.51a 16.72  3.19a 17.81  3.63a N/A 21.52  2.23 24.09  1.26

N/A N/A 37  3d 36  4d 38  4d 38  5d 22  4a,b 21  5a 25  5b,c 26  5c 36  3d 36  4d N/A 27  3 28  6

80 90 VACUUMFIR

70 80 90

LPSSD

70 80 90

N/A implies that the desired final moisture content could not be achieved. Same letters in the same column indicate that the mean values are not significantly different when considered at a confidence level of 95% (p > 0.05).

compared to chips produced only by LPSSD. This again confirms the potential benefits of the FIR-assisted drying processes in the production of highly crisp, fat-free snacks.

3.6 Concluding Remarks

The basic fundamentals of both conventional atmospheric-pressure Superheated steam drying and more novel low-pressure Superheated steam drying are presented in this chapter. The application of these processes to the drying of a wide variety of foods and biomaterials is illustrated and discussed, with emphasis placed on the use of LPSSD to preserve or even improve the properties of the drying materials.

Additional Notation Used in Chapter 3

a b DE L S

redness/greenness yellowness/blueness overall color change lightness survival ratio

Arbitrary unit Arbitrary unit Arbitrary unit Arbitrary unit –

References

Subscript

0

initial

Abbreviations

CRP d.b. FIR HAD HPD IMC LPSSD SSD VD w.b.

constant rate period dry basis far infrared hot air drying heat pump drying intermediate moisture content low-pressure superheated steam drying superheated steam drying vacuum drying wet basis

References Caixeta, A. T., Moreira, R., Castell-Perez, M. E., 2002. Impingement drying of potato chips. J. Food Proc. Eng. 25: 63–90. Chiewchan, N., Pakdee, W., Devahastin, S., 2007. Effects of water activity and hot airdrying on thermal resistivity of Salmonella Krefeld on rawhide surface. Int. J. Food Microbiol. 114: 43–49. Chiewchan, N., 2011. Effect of processing on microbial growth and inactivation in foods, in Physicochemical aspects of food engineering and processing, (ed. S. Devahastin), CRC Press, Boca Raton, USA, pp. 85–104. Devahastin, S., Suvarnakuta, P., 2004. Superheated-steam drying of food products, in Dehydration of products of biological origin, (ed. A. S. Mujumdar), Science Publishers, Enfield, USA, pp. 493–512. Devahastin, S., Suvarnakuta, P., 2008. Lowpressure superheated steam drying of food products, in Drying technologies in food processing, (eds X. D. Chen, A. S. Mujumdar), Wiley-Blackwell, West Sussex, UK, pp. 160–189. Devahastin, S., Suvarnakuta, P., Soponronnarit, S., Mujumdar, A. S., 2004. A comparative study of low-pressure superheated steam

and vacuum drying of a heat-sensitive material. Drying Technol. 22: 1845–1867. Hiranvarachat, B., Suvarnakuta, P., Devahastin, S., 2008. Isomerisation kinetics and antioxidant activities of b-carotene in carrots undergoing different drying techniques and conditions. Food Chem. 107: 1538–1546. Iyota, H., Nishimura, N., Nomura, T., 2001. Drying of sliced raw potatoes in superheated steam and hot air. Drying Technol. 19: 1411–1424. Iyota, H., Konishi, Y., Inoue, T., Yoshida, K., Nishimura, N., Nomura, T., 2005. Popping of amaranth seeds in hot air and superheated steam. Drying Technol. 23: 1273–1287. Kittiworrawatt, S., Devahastin, S., 2009. Improvement of a mathematical model for low-pressure superheated steam drying of a biomaterial. Chem. Eng. Sci. 64: 2644–2650. Kongsoontornkijkul, P., Ekwongsupasarn, P., Chiewchan, N., Devahastin, S., 2006. Effects of drying methods and tea preparation temperature on the amount of vitamin C in Indian gooseberry tea. Drying Technol. 24: 1509–1513.

83

84

3 Superheated Steam Drying of Foods and Biomaterials Leeratanarak, N., Devahastin, S., Chiewchan, N., 2006. Drying kinetics and quality of potato chips undergoing different drying techniques. J. Food Eng. 77: 635–643. Li, Y. B., Seyed-Yagoobi, J., Moreira, R. G., Yamsaengsung, R., 1998. Superheated steam impingement drying of tortilla chips. Proceedings of the 11th International Drying Symposium, Halkidiki, pp. 1221–1228. Li, Y. B., Seyed-Yagoobi, J., Moreira, R. G., Yamsaengsung, R., 1999. Superheated steam impingement drying of tortilla chips. Drying Technol. 17: 191–213. Methakhup, S., Chiewchan, N., Devahastin, S., 2005. Effects of drying methods and conditions on drying kinetics and quality of Indian gooseberry flake. LWT – Food Sci. Technol. 38: 579–587. Mujumdar, A. S., 1990. CEA report 817 U 671. Canadian Electrical Association, Montreal, Canada. Mujumdar, A. S., 2007. Superheated steam drying, in Handbook of industrial drying, (ed. A. S. Mujumdar), 3rd edn, CRC Press, Boca Raton, USA, pp. 439–452. Namsanguan, Y., Tia, W., Devahastin, S., Soponronnarit, S., 2004. Drying kinetics and quality of shrimp undergoing different twostage drying processes. Drying Technol. 22: 759–778. Nimmol, C., Devahastin, S., Swasdisevi, T., Soponronnarit, S., 2007a. Drying of banana slices using combined low-pressure superheated steam and far-infrared radiation. J. Food Eng. 81: 624–633. Nimmol, C., Devahastin, S., Swasdisevi, T., Soponronnarit, S., 2007b. Drying and heat transfer behavior of a food product undergoing combined low-pressure superheated steam and far-infrared radiation drying. Appl. Therm. Eng. 27: 2483–2494.

Phungamngoen, C., Chiewchan, N., Devahastin, S., 2011. Thermal resistance of Salmonella enterica serovar Anatum on cabbage surfaces during drying: Effects of drying methods and conditions. Int. J. Food Microbiol. 147: 127–133. Prachayawarakorn, S., Soponronnarit, S., Wetchacama, S., Jaisut, D., 2002. Desorption isotherms and drying characteristics of shrimp in superheated steam and hot air. Drying Technol. 20: 669–684. Schwartze, J. P., Brocker, S., 2002. A theoretical explanation for the inversion temperature. Chem. Eng. J. 86: 61–67. Sheikholeslami, R., Watkinson, A. P., 1992. Rate of evaporation of water into superheated steam and humidified air. Int. J. Heat Mass Transfer 35: 1743–1751. Soponronnarit, S., Nathakaranakule, A., Jirajindalert, A., Taechapairoj, C., 2006. Parboiling brown rice using superheated steam fluidization technique. J. Food Eng. 75: 423–432. Suvarnakuta, S., Devahastin, S., Mujumdar, A. S., 2005a. Drying kinetics and b-carotene degradation in carrot undergoing different drying processes. J. Food Sci. 70: S520–S526. Suvarnakuta, S., Devahastin, S., Soponronnarit, S., Mujumdar, A. S., 2005b. Drying kinetics and inversion temperature in a lowpressure superheated steam drying system. Ind. Eng. Chem. Res. 44: 1934–1941. Tang, Z., Cenkowski, S., Muir, W. E., 2000. Dehydration of sugar-beet pulp in superheated steam and hot air. Trans. ASAE 43: 685–689. Thomkapanish, O., Suvarnakuta, S., Devahastin, S., 2007. Study of intermittent lowpressure superheated steam and vacuum drying of a heat-sensitive material. Drying Technol. 25: 205–223.

85

4 Intensification of Fluidized-Bed Processes for Drying and Formulation Evangelos Tsotsas, Stefan Heinrich, Michael Jacob, Mirko Peglow, and Lothar M€orl 4.1 Introduction

Fluidized beds are widely used in industry for both the drying and wet formulation of particulate products. Wet formulation is achieved by spraying a solution or suspension (slurry) onto fluidized particles and evaporating the solvent or carrier liquid, which leads to a solid-phase formation in the form of layers (coating, granulation) or bridges (agglomeration). Consequently, formulation of the product is inherently and intimately coupled to drying in terms of both, process capacity and product structure, similar to the case of spray-drying. Due to their importance in practice, drying fluidized-bed processes have been intensively discussed in Modern Drying Technology, especially in Chapter 6 of Volume 1, where population balance equations have been presented as an instrument for the macroscopic description of wet formulation, and in Chapter 7 of Volume 3. In the latter case, the influence of material properties, operating conditions and apparatus design on the type, structure and properties of products obtained in spray fluidized-bed dryers has been discussed in detail. In the present chapter, previous discussions are extended through the aspect of process intensification. Two approaches that may significantly enhance fluidizedbed processes – namely, intensification by apparatus and flow design, and intensification by contact heating – will be treated in detail. Intensification by apparatus and flow design (Section 4.2) is based on the idea of deliberately changing the geometry and the construction of the fluidization chamber and gas entrance, which leads to so-called spouted beds. Consequently, different types of such equipment are discussed in Section 4.2.1, and details of their operating characteristics (pressure drop, operating regimes) are presented in Section 4.2.2. The intensification of gas-side mass and heat transfer that can be attained in a specific type of spouted bed equipment is quantified in Section 4.2.3. Finally, emerging possibilities for understanding the advantages or disadvantages of different types of fluidized-bed equipment by means of discrete particle modeling (DPM) are outlined in Section 4.2.4. Intensification by contact heating, which can

Modern Drying Technology Volume 5: Process Intensification, First Edition. Edited by Evangelos Tsotsas and Arun S. Mujumdar. Ó 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

86

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

be achieved by either immersed heating elements or by heating of the fluidization chamber wall, is discussed in Section 4.3. Following an explanation of the general principle in Section 4.3.1, the main effects and influences – as can be elucidated from experimental results – are presented in Section 4.3.2, where some modeling issues are also discussed. Since the numerous routes to process intensification presented in other chapters of this volume are also applicable to fluidized-bed drying, these discussions will not be repeated here; rather, some cross-references and short additional hints will be provided in Section 4.4, and the chapter will be closed with a brief summary in Section 4.5.

4.2 Intensification by Apparatus and Flow Design 4.2.1 Different Types of Spouted Bed

Spouted beds have a special apparatus design to create spatially correlated flow patterns of particles and gas. Whereas, gas bubbles rising through conventional fluidized beds invoke a random motion and mixing of particles, in spouted beds the apparatus design gives rise to an upwards-oriented gas jet. Velocities higher than the particle entrainment velocity force the solids to follow this jet; however, the cross-sectional area of the apparatus increases with its height in spouted beds, so that the gas velocity decreases gradually and becomes, at a certain position, too small for entrainment. Consequently, the particles fall onto relatively well-defined trajectories back to the bed, where they are recaptured by the jet and the same type of motion is then repeated. The result is that the gas velocity, which must lie between minimal fluidization and the particle entrainment velocity in conventional fluidized beds, can be much higher than the entrainment velocity in certain regions of spouted beds, but without any significant elutriation occurring from the process chamber. Due to the regularity of the imposed circulatory motion of the solids, particles will appear periodically at the same position in the equipment, at which point either a nozzle can be placed for product formulation by coating, or a second gas inlet for rapid catalytic reaction. Moreover, the very high local values of gas velocity in spouted beds enable the stable fluidization of bulk materials that cannot be fluidized via conventional processes because of their large size and/or highly irregular particle shape. The basic design concepts of conventional fluidized beds and spouted beds are shown comparatively in Fig. 4.1. In the conventional, bubbling fluidized bed (Fig. 4.1a) the gas enters the fluidization chamber via a distributor plate at a superficial velocity that is between the minimal fluidization and particle entrainment velocity. Small bubbles are formed at the distributor, and these grow by coalescence while ascending through the bed. The bubble size and their rising velocity will depend on the specific two-phase system (gas and particle properties) in a manner that can be estimated according to Geldart (1973). The bubbles erupt at

4.2 Intensification by Apparatus and Flow Design

(a)

(b)

(c)

(d)

ΔPbed

ΔPbed

1

2

3

4

1 2 3

u0 Fig. 4.1 (a) Design principle and (c) bed pressure drop of a conventional fluidized bed in comparisonwith (b) design principle and (d)bed pressure drop of a conical spouted bed. Operating regions of conventional fluidized beds: 1, packed bed; 2, transitionregion; 3, stable

4

5

6

u0 bubbling fluidization; 4, pneumatic conveying. Operating regions of spouted beds: 1, packed bed; 2, spout formation; 3, transition region; 4, stable operatingregion;5, pulsation;6, plug flow. Bed pressure drop is plotted over gas velocity.

the upper boundary of the bed, expelling particles that can be elutriated even at less than the particle entrainment velocity in the freeboard. In order to reduce elutriation, an expansion section (not shown in the figure) with an increasing cross-sectional area is usually placed on top of a conventional fluidized bed; this allows the gas velocity to be decreased such that the expelled particles can fall out from the gas flow. The ascending bubbles convey particles in their wake, thus providing an intensive mixing of solids in conventional fluidized beds. The influence of the bubbles on interface heat and mass transfer is rather negative, due to a back-mixing of the gas that reduces the driving transfer potentials and some bypass-flow action.

87

88

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

The totally different flow pattern in a spouted bed is shown in Fig. 4.1b where, as noted above, the cross-section of the spouted bed expands with its increasing height. As a result of this geometry the gas velocity will exceed the particle entrainment velocity at the bottom, which creates the spout. However, at the top the gas velocity will be less than the entrainment velocity, so that particles can fall out and enter a regular recirculation. The typical behavior of bed pressure drop with increasing gas velocity is depicted in Fig. 4.1c and d for conventionally fluidized beds and for spouted beds, respectively. The graph for a conventional fluidized bed (Fig. 4.1c) can be subdivided into four regions, corresponding to: 1, the packed bed; 2, transition to bubbling fluidization; 3, stable bubbling fluidization; and 4, pneumatic conveying. In the case of a spouted bed, the behavior becomes more complex and depends, to a certain extent, on the type of spouted bed being considered (see discussion below). However, this dependence is usually moderate so that, according to Piskova (2002), six regions can be identified for all types of spouted bed on the graph of Fig. 4.1d. In region 1, the pressure drop increases with increasing gas flow rate, corresponding to the pressure drop of a packed bed, and reaches a maximum. The bed pressure drop is decreased again in region 2 as a spout begins to take form, creating a path of easy flow for the gas through the particles. In region 3 the spout is still very irregular, so that gas breakthrough occurs by means of both the developing spout and large gas bubbles, accompanied by strong pulsations of pressure. Beyond a certain gas velocity a stable operating regime of the spouted bed can be observed, as denoted by region 4 in Fig. 4.1d. The spout is well established and regular in this region, and the amplitude of pressure pulsations is relatively low. The average value of the bed pressure drop no longer increases with the increasing gas flow rate in region 4; rather, it remains constant at a level which is much lower than in a stably operated conventional fluidized bed of the same mass (corresponding to region 3 of Fig. 4.1c). This means that the weight of the bed, which defines the bed pressure drop in conventional fluidization, does not necessarily need to be outbalanced by the bed pressure drop of a spouted bed. Depending on the precise apparatus configuration, an excessive increase in gas velocity leads first to a new regime of pulsation (region 5) and then to a plug-like, unidirectional flow in a large part of the apparatus (region 6). The ultimate operational limit is reached when the largest particles can be entrained by the gas even in the largest cross-section of the equipment. The major goal of spouted bed apparatus design is a stable operation in region 4. In both, conventionally fluidized and spouted beds, bed porosity starts with the packed bed value in region 1 and then increases continually with increasing gas velocity in regions 2 and 3 or 2 to 6, respectively. Concerning the spatial distribution, both configurations have almost particle-free regions – the bubbles in conventionally fluidized or the spout in spouted beds – surrounded by dense regions with nearly packed bed porosity – the so-called “suspension phase” in conventionally fluidized and the region of sliding particle recirculation in spouted beds (cf. Fig. 4.1a and b).

4.2 Intensification by Apparatus and Flow Design

In order to fully use the major advantage of spouted beds – namely, their ability to treat very different types of particulate material – various design alternatives have been proposed and developed. An evident distinction results from the shape of the process chamber, which may be conical, conical-cylindrical or prismatic, corresponding to either circular or rectangular cross-sections. Solids are present only in the conical part of the equipment depicted in Fig. 4.1b which is, therefore, a conical spouted bed (Olazar et al., 1992; Olazar et al., 1993). In contrast, the solids also fill a significant part of the cylinder, set on top of the cone in the set-up shown in Fig. 4.2, so that this then belongs to the conical-cylindrical type. It should be noted that the angle and height of the cone, as well as the height of the cylindrical part, can be varied over a wide range (Mathur and Epstein, 1974). Nevertheless, the operating regions of conical and conical-cylindrical spouted beds are similar, as illustrated in Fig. 4.1d. Because of high turbulence in the spout and the regularity of particle mixing, different types of spouted bed are highly appropriate to the placement of spray nozzles for the purpose of particle formulation, such as granulation or coating. Two examples in which liquid spraying occurs directly at the entrance of the gas jet, and which results in a concurrent, upward flow of the droplets and gas with frequent droplet–particle collisions, are shown in Fig. 4.3. The equipment shown in Fig. 4.3a is of the conical-cylindrical type, whereas the cylindrical device of Fig. 4.3b is named Wurster equipment, after its inventor (Wurster, 1959). The main feature of

Fig. 4.2 Conical-cylindrical spouted bed, in contrast to the conical equipment of Fig. 4.1b.

89

90

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

(a)

(b)

Fig. 4.3 Spouted beds with circular cross-section and liquid spraying. (a) Conical-cylindrical; (b) Wurster equipment.

the Wurster equipment is a centrally placed riser tube, wherein droplets are sprayed and deposited onto the particles. Segmentation of the distributor plate with a higher permeability below the riser causes the gas flow to be significantly stronger in this tube than at the sides, thus creating a toroidal circulation of the particles. Either one-fluid or two-fluid nozzles can be used in the Wurster equipment. Unfortunately, spouted beds with a cylindrical cross-section cannot be deliberately enlarged, and their scale-up is difficult. However, this problem proved to be an important motivation for the development of prismatic spouted beds with a rectangular cross-section. A simple configuration of such equipment, as proposed by Mitev (1967), is depicted in Fig. 4.4a. This and other similar equipment can be simply scaled-up by prolongation, because an increase in the length of the apparatus (the direction perpendicular to the plane of Fig. 4.4) does not alter the operational regime as long as the pressure drop in the empty equipment does not (a)

(b)

Fig. 4.4 Spouted bed with a rectangular cross-section. (a) With one gas feed (via the plenum (Mitev, 1967); (b) With two gas feeds (via the plenum and the right-hand side bottom plate (Piskova, 2002).

4.2 Intensification by Apparatus and Flow Design

fall below a critical value. It is not generally feasible, however, to spray in the gas jet of the equipment of Fig. 4.4a, because this jet is created on the side wall which would, as a consequence, be wetted by droplets and covered by solid deposits after drying. Rather, it is possible to construct one of the bottom plates of the fluidization chamber as an empty jacket, and to use this jacket for a second gas supply, in addition to the primary gas feed via the plenum, as proposed by Jordanova et al. (2000) and shown in Fig. 4.4b. By using this approach, rapid reactions involving the two gas streams and the particles can be conducted. In both design variants of Fig. 4.4 a porous distributor is indicated by a dashed vertical line in the gas entrance gap (enlarged post of the figure). This type of distributor is not required when the gas velocity in the gap is selected to be sufficiently high to prevent particles from falling out into the plenum. However, the ratio between the gas entrance cross-sectional area (cross-sectional area of the gap) and the surface area of the stagnant bed is an important geometric feature of both devices shown in Fig. 4.4, as noted by Mitev (1967, 1979) (cf. Eq. 4.11). Replication of the geometry of Fig. 4.4a by projection to the one or to the other side gives rise to the design configurations of Fig. 4.5. In the case of Fig. 4.5a, a particle fountain is located in the middle. Particles move to the sides at the top of the fountain, slide down the inclined bottom walls of the process chamber, and are then pushed by the gas jets back into the fountain. In this way, two symmetrically arranged rotation cells with high porosity in their horizontal axes occur over the entire length of the equipment. In the case of a gas–particle reaction (e.g., combustion) in the equipment of Fig. 4.5a, the reaction zone would be in the middle, far away from the external side walls. In contrast, in the equipment shown in Fig. 4.5b the direction of rotation of the longitudinal cells is exactly opposite to that in Fig. 4.5a, with particles rising with the two gas jets at the outer bottom plates of the fluidization chamber, moving to the middle, and then descending onto the inner bottom plates before re-entering the circulation. There is no central particle fountain in this case. The two configurations of Fig. 4.5 can be expanded for liquid spraying, as illustrated in Fig. 4.6. Spraying takes place from the bottom at the edge of the central partition in Fig. 4.6a, which means that liquid droplets are sprayed into the (a)

(b)

Fig. 4.5 Prismatic, dual-gas-entrance spouted beds which are symmetric aside their middle plane. (a) Equipmentwith centralfountain ofparticles accordingto Mitev(1979);(b) Equipmentwithgas jets and ascending particles at the sides according to Mitev (1979) and Piskova (2002).

91

92

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

(a)

(b)

Fig. 4.6 Expansion of the prismatic spouted beds of Fig. 4.5 for (a) bottom spray and (b) top spray of liquid.

ascending central fountain of particles. The high momentum of particles and high gas velocities in this fountain, which result in intensive drying, are rather in favor of coating or layering granulation than of agglomeration in this type of equipment. The configuration of Fig. 4.6b differs in two ways: (i) in the placement of the nozzle (it is still central, but here is employed as top-spray); and (ii) in the direction of particle cell rotation. By combining these two effects, however, liquid droplets can be sprayed onto a rather dense packing of quite slowly descending particles, and this may lead to an increased agglomerate formation. Stable rotation cells, resulting from correct equipment design, will lead to a regular, periodic passage of particles through the spray zone in both cases of Fig. 4.6. This feature may be used either to improve the uniformity of coating layers or to reduce the width of the particle size distribution of agglomerates. Many spraying (usually twofluid) nozzles can be placed periodically along the equipment, while scale-up can be achieved by the above-mentioned prolongation and by further lateral replication, to create virtually any size. A common limitation of all types of prismatic spouted bed equipment with lengthy gas entrance gaps described so far results from the constant width of these gaps. The consequence of a constant gap width is that the pressure drop increases with the square of the gas volume flow rate, which is unfavorable in terms of automatic control. Moreover, the entrance may be blocked by solids at some places, decreasing the gas flow rate for given blower characteristics, or increasing the pressure drop and energy consumption. From these points of view, constructive solutions that would enable adjustment of the width of the gas entrance and the respective pressure drop during operation would appear desirable. Such a constructive solution has been developed by M€ orl et al. (2000) and is illustrated in Fig. 4.7. This solution is based on the use of roller valves, made from cylindrical drums which are cut flat on one side, so that rotation of the valve opens the gas entrance gap in one direction and closes the gap in the other direction, thus creating the desired adjustability of cross-sectional area and pressure drop (Gryczka, 2009). This concept has been introduced successfully to the market by the Glatt Co., under the trade name ProCell, and will be further discussed in Sections 4.2.3 and 4.2.4. It can be applied to any spouted bed with a rectangular cross-section, as illustrated by the two examples in Fig. 4.7.

4.2 Intensification by Apparatus and Flow Design

(a)

(b)

Fig. 4.7 Prismatic spouted beds with adjustable gas inlet. (a) Equipment with central particle fountain without liquid spraying; (b) Equipment with gas jets at the sides and top spray.

4.2.2 Operating Characteristics of Spouted Beds

Several efforts to chart fluidization and the behavior of fluidized particles have been undertaken (Mathur and Ghisler, 1955; Mitev, 1967; Mitev, 1979; Li and Kwaik, 1980; Kunii and Levenspiel, 1991; Olazar et al., 1993). One such diagram, shown exemplarily in Fig. 4.8, is the modified version of a plot that was originally proposed by Grace (1986) but later expanded by information provided by additional authors and recommended for use by Piskova (2002).

Pneumatic transport

Fast luidized beds

10

ut

u*

1 Spouted beds 10-1 umf

Bubbling luidized beds

A

B

10-2

D

10

-3

1

10

d*

102

Fig. 4.8 Characteristic regions of fluidization.

93

94

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

The modified gas velocity at the ordinate of the plot of Fig. 4.8 is defined as: 2

r2g

313

Re0  5 ¼ 1=3 ; u ¼ u0 4
Ar mg rp  rg g


ð4:1Þ

where u0 is the superficial gas velocity and Re0 is the respective Reynolds number. The abscissa plots a modified particle diameter  313 2
r r  r g g p g
5 ¼ Ar13 d ¼ d4 2 mg

which is equal to the cubic route of the Archimedes number (Ar):
 gd3 rp  rg Ar ¼ : n2g rg

ð4:2Þ

ð4:3Þ

Different groups of the classification of solids according to Geldart (Geldart, 1973; Molerus, 1982) can be identified on the abscissa of the plot of Fig. 4.8. However, this plot provides only a coarse estimate of the material properties and operating conditions leading to spouting fluidization, which is greatly influenced by the precise construction of the equipment and, especially, of the spout zone. In order to reliably design spouted bed equipment for practical application, the following three quantities must be known, depending on the specific configuration, for example, cylindrical or prismatic:  The minimal gas volume flow rate required for fluidization, which corresponds to the upper end of region 1 in Fig. 4.1d.  The maximal pressure drop in the bed which is observed at starting fluidization, that is, simultaneously with the minimally necessary gas flow rate (highest value of pressure drop in Fig. 4.1d).  The range of stable operation, corresponding to region 4 in Fig. 4.1d, especially its lower limit. A very simple, but also coarse, method of determining the minimally required gas volume flow rate is to define it as the flow rate creating conditions of minimal fluidization at the upper surface of a stagnant bed of solids placed in the equipment under consideration. This derivation will be illustrated for the ProCell unit depicted schematically in Fig. 4.9. The superficial gas flow velocity at the free surface of the respective stagnant bed is the ratio of the gas volume flow rate V_ g to the respective cross-sectional area, which is twofold the product of the width W0 and the length L of the apparatus, perpendicular to the plane of Fig. 4.9. By setting the superficial gas flow velocity equal to the minimal fluidization velocity umf and the gas volume

4.2 Intensification by Apparatus and Flow Design

Vg 2

Vg 2

W0

Vg 2

Vg 2

Fig. 4.9 Schematic of a symmetrical ProCell unit, with two gas inlet gaps (M€ orl et al., 2000).

flow rate equal to its minimally required value V_ g;min , the relationship umf ¼

V_ g;min 2W0 L

ð4:4Þ

is obtained. With known umf, Eq. 4.4 can be used to derive V_ g;min . To this purpose, the minimal fluidization velocity can be calculated from the respective Reynolds number, which, according to Todes (Goroshko et al., 1958), is Remf ¼

Ar pffiffiffiffiffi : 1400 þ 5:22 Ar

ð4:5Þ

It should be noted that the same authors derive the Reynolds number at elutriation (i.e., at the ultimate operation limit) to Reelu ¼

Ar pffiffiffiffiffi : 18 þ 0:61 Ar

ð4:6Þ

In this way, the minimally required gas volume flow rate is obtained to 2W0 Lng Ar pffiffiffiffiffi : V_ min ¼  d 1400 þ 5:22 Ar

ð4:7Þ

The previously explained method for the determination of the minimally required gas volume flow rate is of general applicability, but approximate. Alternatively, correlations that refer to a specific type of spouted bed may be used. A number of correlations of this type for conical and conical-cylindrical spouted beds are summarized in Tab. 4.1, based on an evaluation by Piskova (2002). The geometric quantities involved are illustrated in Fig. 4.10. All correlations of Tab. 4.1 lead to the minimal value of a Reynolds number: Rein ¼

ug;in d ng

ð4:8Þ

built with the superficial gas velocity ug;in ¼ V_ g =Ain

ð4:9Þ

95

4 Intensification of Fluidized-Bed Processes for Drying and Formulation Tab. 4.1 Correlations for the determination of the minimally required gas volume flow rate in conical and conical-cylindrical spouted beds.

Source

Correlation

Olazar et al. (1992)

Rein;min ¼ 0:126 Ar0:5

Mathur and Epstein (1974)

Rein;min

Romankow and Raschkovskaja (1968)

Rein;min

 D0 1:68 ðtan g Þ0:57 Din  0:48  1:27 H0 D0 ¼ 2:8  102 Ar0:57 Din Din  0:82 D ðtan g Þ0:1 ; 1 ¼ 0:364 Reelu Din 

Reelu from Eq. 4.6

 0:1  0:25 Din H0 D0 D0  0:85 D 0 ¼ 0:174 Ar0:5 ðtan g Þ1:25 Din

Nikolaev and Golubev (1964)

Rein;min ¼ 0:05 Ar0:59

Gorshtein and Mukhlenov (1964)

Rein;min

Mitev (1979)

Rein;min ¼ 12:2 Remf G0:8 ; Remf from Eq. 4.5, G ¼ 1:09  103 Ar0:507 þ 4:1

D

d D0

γ H0

96

Din

Fig. 4.10 Characteristic dimensions of conical-cylindrical spouted beds.

4.2 Intensification by Apparatus and Flow Design

in the narrowest cross-section of the equipment – that is, in the gas entrance tube with the diameter Din and a cross-sectional area of Ain in Fig. 4.10. With Rein,min known from Tab. 4.1, the respective minimal velocity ug,in,min can be obtained from Eq. 4.8 and multiplied with Ain to evaluate V_ g;min . The gas load that is at least necessary for fluidization in prismatic spouted beds can be estimated in a similar way, based on correlations from Tab. 4.2. As noted above, the second quantity necessary for the design of spouted bed equipment is the maximal pressure drop corresponding to the minimal necessary gas flow rate at the onset of fluidization, as the blower must be able to provide this pressure drop during the start-up of the process. Bed pressure drops measured by Gryczka (2009) in a ProCell apparatus are depicted in Fig. 4.11, in dependence of the gas volume flow rate. The plot shows very clearly that the maximal pressure drop in the bed is much higher – typically, two- to fourfold – than the bed pressure drop during stable operation of the spouted bed. By comparing with the more general Fig. 4.1d, it is recognized that the bed pressure drop falls very quickly after the maximal value, which means that region 2 of Fig. 4.1d is small for the ProCell unit investigated by Gryczka (2009). In contrast, the stable operating region 4 is quite larger, without any significant fluctuations of pressure drop on changing the gas flow rate. The results of investigations of the maximal bed pressure drop with different types of spouted bed are summarized in Tab. 4.3. Therein DPf ¼ gð1  emf Þ ðrp  rg ÞH0

ð4:10Þ

denotes the pressure drop that would outbalance the weight of the solids, corrected for buoyancy by the gas. Apart from the maximal pressure drop, the bed pressure drop during the stable operation of spouted beds is of interest. As demonstrated in Fig. 4.11, the latter will usually be significantly smaller. Respective literature results are presented in Tab. 4.4. The third parameter to be determined is the range of stable operation, and in particular the lower limit of this range that defines the gas volume flow rate which is at least necessary in order to reliably operate the spouted bed. A stable operation means the stability of particle circulation trajectories, which may be different depending on equipment construction (as noted above), but should not fluctuate too much in time. Investigations of the stable operating range have been conducted

Tab. 4.2 Correlations for the determination of the minimally required gas volume flow rate in prismatic spouted beds.

Source

Correlation

Piskova (2002)

Rein;min ¼ 3:093 Reelu

Mitev (1979)

Rein;min ¼ 12:2 Reelu G0:8 , Remf from Eq. 4.5, G ¼ 3:9  103 Ar0:443 þ 1:1



H0 Din

0:6638

ðtang Þ0:92 , Reelu from Eq. 4.6

97

4 Intensification of Fluidized-Bed Processes for Drying and Formulation Tab. 4.3 Correlations for the determination of the maximal bed pressure drop in conical, conicalcylindrical and prismatic spouted beds.

Source

Equipment

Asenjo et al. (1977)

Conical

Kmiec (1980)

Conical-cylindrical

Mitev (1979)

Prismatic

Correlation  

2H0 DPmax ¼ DPf 1 þ 2:8 exp 0:156 D0  

2H0 DPmax ¼ DPf 1 þ 0:206 exp 0:62 D0

DPmax ¼ K H0 ð1  emf Þrp g; K ¼ 1:2 for H0 < 180 mm K ¼ 0:72  101:26H0 for H0 > 180 mm

6000

ΔPbed, Pa

98

Mbed=1.98 kg Mbed=1.45 kg Mbed=0.95 kg

4000

2000

0 0

50

.

Vg, m3 h-1

100

150

Fig. 4.11 Bed pressure drop as a function of gas volume flow rate for a ProCell unit according to measurements by Gryczka (2009).

Tab. 4.4 Correlations for the determination of bed pressure drop during stable operation in conical, conical-cylindrical, and prismatic spouted beds.

Source

Equipment

Correlation

Raschkowskaja (1969)

Conical

DPst ¼ 0:45 rbed gH0 1 þ

Mathur and Epstein (1974) Mitev (1979) Flisjuk et al. (1984)

Conicalcylindrical Prismatic Prismatic

"

0:13 K ¼ 1  0:053 Ar"

DPst ¼ Krbed gH0 1 þ



D Din

1

þ 3

0:81ðtan g Þ2 f



D Din

2 # ;

!  #1 D0 d Din H0 D2in

DPst ¼ Kð1  emf Þrbed gH0 ; K ¼ 0:8ð0:92 H0 þ 0:5Þ DPst ¼ Krbed gH0 , K ¼ 1  0:053 Ar0:13

4.2 Intensification by Apparatus and Flow Design

1000000 Onset of stable luidization

E

Rein

100000

10000

End of stable luidization

D

Ar = 2.6 107

C B

Ar = 1.7 105 Ar = 3.2 104

A

1000

Ar = 3.7 103 Ar = 1.8 103 100 0.1

1.0

G [%]

10

Fig. 4.12 Determination of the stable operating range of symmetrical ProCell equipment with two parallel gas entrances with different particles (A to E) according to Gryczka (2009).

by several authors for different types of spouted bed (Mitev, 1979; Olazar et al., 1992; Piskova, 2002; Gryczka, 2009). Here, a method developed by Mitev (1979) und Piskova (2002) will be described, based on a Re-G-Ar-plot. The Reynolds number of this plot refers to the gas entrance, and is defined according to Eq. 4.8. The parameter G combines geometric features with the hold-up, defined as G¼

Ain ; Abed;0

ð4:11Þ

where Ain is the cross-sectional area of the gas entrance and Abed;0 is the area of the free surface of the stagnant bed in the considered equipment. Finally the Archimedes number (Ar) expresses the material properties of the particle system (Eq. 4.3). The procedure is to vary the parameter G by variation of the hold-up and, thus, of Abed,0 for a given piece of equipment and a certain particulate material (i.e., at Ar ¼ constant) until the limits of the stable operation range have been identified and the respective values of G and Rein have been plotted on the Re-G-Ar-diagram. The same determination is then repeated for another particulate material, that is, with another value of Ar. This procedure is illustrated in Fig. 4.12 on the basis of experimental results by Gryczka (2009) for a ProCell unit. The plot of Fig. 4.12 may be extended to include results for other types of spouted beds. In this way, Fig. 4.13 was obtained by Gryczka (2009). Reynolds numbers – as defined by Eq. 4.8 – that correspond to the lower limit of the stable operating regimes of Fig. 4.13 may be denoted by Rein,st. Correlations developed for Rein,st are summarized in Tab. 4.5. However, it should be noted that Fig. 4.13 and Tab. 4.5 provide only coarse estimates for the stable operation ranges of different types of spouted bed. The reason for this is that even small variations in the apparatus geometry – especially variations in the construction of the gas entrance – can have a significant influence on the fluid dynamics of the spouted

99

100

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

Fig. 4.13 Working diagram for the determination of the stable operation regions of different types of spouted beds according to Gryczka (2009). Between A and B: Conventional spouted beds (Mitev, 1979); C to D: Prismatic spouted bed with two parallel gas entrances

(Piskova, 2002); E to F: Conical-cylindrical spouted beds (Olazar et al., 1993); G to H: Prismatic spouted bed with one gas entrance (Mitev, 1979); I to J: Prismatic ProCell unit with two gas entrances (Gryczka, 2009), same data as in Fig. 4.12.

bed. Moreover, the presented results are mainly based on experiments conducted with near-spherical, non-sticky particles with narrow size distributions. 4.2.3 Mass and Heat Transfer in ProCell Units

In the previous sections it was shown that the flow field in spouted beds is spatially correlated, resulting in a characteristic, circulatory motion of particles; this is in contrast to the complex but, on average, essentially uniform flow patterns of conventional fluidized beds. Apart from the chamber geometry, special apparatus Tab. 4.5 Correlations for the Reynolds number at the lower limit of the stable operating region for various types of spouted bed equipment.

Source Olazar et al. (1993) Mitev (1979)

Equipment

Conicalcylindrical Prismatic, one gas inlet Piskova (2002) Prismatic, two gas inlets Gryczka ProCell, two (2009) gas inlets

Correlation

Validity

Rein;st ¼ 12:2Reelu G0:8

3  103 Ar0:507 < G < 1:09  103 Ar0:507 þ 4:1 0:7  103 Ar0:443 þ 1:1 < G < 3:9  103 Ar0:443 6  108 Ar þ 1:8094 < G < 2  107 Ar þ 6:085 0:81 Ar0:0089 < G < 3:7Ar1:0486

Rein;st ¼ 12:2Reelu G0:8 Rein;st ¼ 42:3Reelu G1:979 Rein;st ¼ 190:8Reelu G1:0486

4.2 Intensification by Apparatus and Flow Design

design features can be used to adapt and control the flow field in spouted beds, such as air inlet openings of adjustable width in the ProCell equipment. This enables the processing of materials that can hardly be treated in conventional fluidized beds, namely large, heavy, non-spherical, polydispersed or sticky particles. However, the question arises as to whether such units can be expected to have an advantage also in terms of overall heat and mass transfer between the gas and the particles, or not. Hoffmann et al. (2011) examined this process intensification issue for a ProCell unit, and the following discussions are based mainly on these investigations. The underlying experiment involved drying batches of water-loaded a-Al2O3 particles in air within the first drying period. The a-version of Al2O3 is, in contrast to the c-version of the same material, almost utterly non-hygroscopic and shows a distinct first drying period of constant drying rate. The particles used were nearly monodispersed, with a mean diameter of d ¼ 1.4 mm and a dry particle density of rp,dry ¼ 1950 kg m3. All experiments were conducted using laboratory-scale equipment with a cylindrical top section (diameter 150 mm; height 235 mm) and a ProCell bottom section (Fig. 4.14). The triangularly prismatic ProCell piece had a depth of 100 mm from its transparent front plane, and a height of 110 mm. The bottom angle measured 60 , with the ProCell partition in the middle, and one roller valve at each side of it. As previously explained, regulation of the air inlet is performed by the

Fig. 4.14 Frontal view of the apparatus used in the measurements. From Hoffmann et al. (2011).

101

102

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

ϕ r1

R1 h1

R h2

r x

x0

Fig. 4.15 Geometric data of the adjustable air inlet in a scale of 3 to 1: R ¼ 15 mm; R1 ¼ 16 mm; r ¼ 10 mm; r1 ¼ 20 mm; h1 ¼ 21 mm; h2 ¼ 13 mm; x0 ¼ 3 mm. From Hoffmann et al. (2011).

two roller valves, each of which consists of a cylindrical drum that is cut flat on one side so as to create a planar surface. The vertical distance between the partition contour and the line where the plane meets the cylindrical surface of the drum is denoted by x in Fig. 4.15, and defines the width of the air entrance gap. If the drums are turned against each other, the two gaps either open or close; closing of the gap increases the local gas velocity (air inlet velocity) at a constant gas mass flow rate. The central partition unit combines the gas flows coming from the right and from the left to a spout. Symmetry of the flow field in respect to the central plane and stability of the jet are important features that must be attained by the correct design of the partition. The parameters varied in the 19 experiments reported by Hoffmann et al. (2011) _ g (from were the dry solids hold-up Ms (from 68 to 291 g), the air mass flow rate M 1 20 to 40 kg h ), the angular position of the gas entrance valves w (from 0 to 15 ; see Fig. 4.15), and the gas inlet temperature Tg,in (from 40 to 60  C). The variation of hold-up corresponds to static bed heights H0 ranging from 29 to 58 mm. The width of the gas inlet opening gap x can be calculated from the angle w to values between x ¼ 3 mm (for w ¼ 0 ) and x ¼ 1.38 mm (for w ¼ 15 ). The variation of air mass flow rate and inlet temperature leads to superficial velocities u0 in the upper, cylindrical part of the apparatus ranging from 0.27 to 0.56 m s1. Gas velocities in the adjustable inlet gaps ug,in were very much higher, ranging from 12 to 30 m s1. The main measuring result of each experiment was the moisture content of the outlet air, Yout, which remains constant over time as long as drying occurs within the first drying period. This was determined by infrared spectroscopy (as described in Chapter 1, Volume 2 of this series). The moisture content of the inlet air, Yin, was measured in the same way. With a known temperature, Tg,in, and moisture content, Yin, of the inlet air, the adiabatic saturation temperature of the air, Tas, and the respective moisture content, Yas, can either be computed or read from the Mollier

4.2 Intensification by Apparatus and Flow Design

diagram. On the basis of the above-mentioned air moisture contents, the mass transfer efficiency of the equipment can be evaluated as: g¼

Yout  Yin Yas  Yin

ð4:12Þ

This efficiency can range from g ¼ 0, which means that the air leaves the particle system without any uptake of water vapor (Yout ¼ Yin), to a maximum of g ¼ 1 for outlet gas saturation (Yout ¼ Yas). It should be noted that a distinction between wet bulb conditions at low values of g and adiabatic saturation conditions at high mass transfer efficiencies is not necessary, due to approximately equal values of wet bulb and adiabatic saturation temperatures for the evaporation of water in air. The next step of evaluation is to translate the efficiency, g, to the number of gasside transfer units (NTU) of the considered particle system. In this context, the motion of the particles does not play a direct role, because conditions at the surface of all particles are identical everywhere in the dryer during the first drying period. However, the flow field of the gas, which is influenced by the motion of the particles, is of consequence, and it is necessary either to perform a detailed, fullscale CFD or to make assumptions concerning the gas flow. The second option has been implemented by Hoffmann et al. (2011), assuming the gas as either perfectly back-mixed or in ideal plug flow. The former corresponds to a so-called continuous stirred tank reactor (CSTR), and the latter to a plug flow reactor (PFR). In case of the CSTR assumption, the function g(NTU) is g¼

NTU : 1 þ NTU

ð4:13Þ

The same dependence is described by g ¼ 1  eNTU

ð4:14Þ

for the PFR limiting case. In one or in the other way, numbers of transfer units, defined as NTU ¼

b rg Abed ; _g M

ð4:15Þ

are obtained from the measured mass transfer efficiencies. With known Abed (total _ g (gas mass flow rate), values particle surface area in the bed), rg (gas density) and M of the overall particle-to-gas mass transfer coefficient, b, and, finally, Sherwood (Sh) numbers Sh ¼

bd d

ð4:16Þ

can be calculated (where d is the binary diffusion coefficient of water vapor in air). The results of the evaluation are presented in Fig. 4.16, where the black squares and the gray circles are Sherwood numbers calculated according to the CSTR and the PFR assumptions, respectively. The Reynolds number of the abscissa is defined with the superficial gas velocity in the upper cylindrical part of the apparatus, u0, to

103

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

10

2

packed bed packed bed single particle single particle Sh FB Sh FB Sh app Sh app 10

(Ho min) (Ho min)

1

Sh

app

Shapp

(Ho max) (Ho max)

•PFTR PFR

Sh [-]

104

10

CSTR CSTR

0

-1

10 0 10

10

1

10

2

Re [-] 0

Fig. 4.16 Sherwood numbers in the ProCell unit, evaluated with either the back-mixing (CSTR) or the plug flow (PFR) assumption and compared with regions for the respective Sherwood numbers of conventional fluidized beds. From Hoffmann et al. (2011).

Re0 ¼

u0 d : ng

ð4:17Þ

The variation of operating parameters other than Re0 is not reflected in Fig. 4.16. The Schmidt number (Sc ¼ ng/d) was almost constant for all experiments at approximately 0.66. Figure 4.16 allows for comparison of mass transfer in the ProCell unit with the mass transfer in conventional fluidized beds. To this purpose, Sherwood numbers calculated according to the well-known correlations of Gnielinski (2010) for a single particle (broken curve) and for a packed bed (solid curve) are plotted. The Sherwood numbers for conventional fluidized beds should lie in the stripe between these two lines, because the packed-bed condition corresponds to minimal fluidization, and the single particle condition to elutriation. In between, the Sherwood number should be almost constant, irrespective of the value of Re0, at about Sh ¼ 9 for the particles used in the experiments (the horizontal straight line in the plot). Extensive previous analyses of interface heat and mass transfer in conventional fluidized beds (Groenewold and Tsotsas, 1997, 1999) have shown that such Sherwood numbers should be combined with the assumption of perfect back-mixing of the gas in the bed. Consequently, the value of approximately Sh ¼ 9 for the conventional fluidized bed should be

4.2 Intensification by Apparatus and Flow Design

compared with Sherwood numbers obtained for the ProCell unit in a similar manner, that is, by means of the CSTR assumption of Eq. 4.13. Such values are represented by the black squares in Fig. 4.16, with a range from 11.2 to 36.5 and an average of Sh ¼ 20.3, which is more than twofold larger than Sh ¼ 9. This is a clear demonstration of a very significant enhancement of particle-tofluid mass transfer in the ProCell unit, compared to a similarly operated conventional fluidized bed. The same conclusion can be attained via the PFR evaluation. The respective data from the ProCell unit (gray cycles) are located in Fig. 4.16 significantly lower than Sherwood numbers evaluated with the CSTR assumption, because back-mixing reduces the efficiency of interface heat and mass transfer and, thus, leads to lower values of transfer coefficients, if taken into account in such coefficients instead of being considered in the balance equation of the model. For the reference case of the conventional fluidized bed, so-called “apparent” Sherwood numbers (Shapp) must be used in combination with the PFR model (Groenewold and Tsotsas, 1999). By setting the right-hand sides of Eqs 4.13 and 4.14 equal to each other and expressing NTU in terms of dimensionless quantities NTU ¼

Sh AV H Re0 Sc

the apparent Sherwood number can be calculated as:   Re0 Sc Sh AV H ln 1 þ ; Shapp ¼ AV H Re0 Sc

ð4:18Þ

ð4:19Þ

where Av denotes the particle surface area per bed volume. For spherical particles it is AV H ¼ 6ð1  eÞ

H : d

ð4:20Þ

Equation 4.20 can be applied with the expanded bed height and porosity, H and e, respectively, or with the packed bed porosity of approximately 0.40 and the stagnant bed height H0. By using the smallest and largest H0 from the experiments in Eq. 4.20 and Sh ¼ 9 in Eq. 4.19, two curves are obtained for Shapp of the conventionally fluidized particles in Fig. 4.3. It should be noted that Shapp can be smaller than the single-particle Sherwood number. For conventional fluidized beds, Shapp (the PFR Sherwood number) depends on H0 (it decreases with increasing H0), whereas Sh (the CSTR Sherwood number) is independent from H0 (approximately equal to 9 for the particles used by Hoffmann et al. (2011)). The difference between gray circles and the stripe of Shapp in Fig. 4.16 demonstrates the intensification of mass transfer in the ProCell unit. Based on their data, Hoffmann et al. (2011) proposed for the mass transfer in ProCell units empirical correlations of the type  p  q D m D : ð4:21Þ Sh ¼ a Re0 H0 x

105

106

4 Intensification of Fluidized-Bed Processes for Drying and Formulation Tab. 4.6 Pre-factor and exponents of the correlation for the Sherwood number (Eq. 4.21), in dependence on the assumption concerning the flow.

Assumption

a

m

p

q

CSTR PFR

0.0468 0.0142

0.57 0.80

0.48 1.36

0.79 0.23

Although not discernible in Fig. 4.16, the parameters H0 and x were varied in the experiments. In contrast, the diameter D of the upper, cylindrical part of the apparatus was constant, and consequently dimensionless representation is the only justification for the appearance of this quantity in the correlation. The Schmidt number was also constant, and has been incorporated in the pre-factor of Eq. 4.21. The values of the pre-factor, a, and of the exponents, m, p, and q, which are listed in Tab. 4.6, depend on whether the Sherwood numbers shall be used in combination with the CSTR or with the PFR assumption in order to calculate particle-to-gas mass transfer in the unit. Nusselt numbers for gas-to-particle heat transfer can be obtained by analogy. To this purpose, the pre-factor should be split to a ¼ b Scn

ð4:22Þ

with Sc ¼ 0.66 and an assumed value of n ¼ 1/3, then the Schmidt number should be replaced by the Prandtl number. The necessity of assuming the value of the exponent n, results from the fact that Sc was not varied in the experiments. The influence of the Reynolds number is higher in the case of plug flow (m ¼ 0.80 according to Tab. 4.6), but still existent under the CSTR assumption (m ¼ 0.57). The first finding is similar as for conventional fluidized beds, but the second constitutes a difference because Sherwood numbers of conventional fluidized beds evaluated on the basis of the CSTR assumption are nearly independent of Re0 (the exponent m is almost zero, as discussed above). This is an indication of the significant influence of the unidirectional flow in the region of the spout on interface transport phenomena. The static bed height H0 (i.e., the holdup) has a great influence on Sherwood numbers evaluated by means of the PFR assumption (p ¼ 1.36 according to Tab. 4.1), but it has only small influence when evaluating for a CSTR (p ¼ 0.48). This is similar to the above-discussed behavior of conventional fluidized beds. Finally, Eq. 4.21 and Tab. 4.6 show that the mass transfer is intensified by decreasing the opening width of the valve, x, which is equivalent to increasing the inlet gas velocity, ug,in. Hoffmann et al. (2011) attempted in their studies also a comparison between the ProCell unit and spouted beds with conventional geometry and air inlet. To this purpose, they used for conventional spouted beds on the one hand correlations by Uemaki and Kugo (1968) and El-Naas et al. (2000), and on the other hand more recent measurements and CFD simulation results (Kmiec et al., 2009; Szafran and Kmiec, 2004). The correlations indicated similar efficiencies of interface transfer in conventional spouted beds as in conventional fluidized beds, with the already discussed large advantage for ProCell units. This advantage becomes smaller when comparing with the data from Kmiec and coworkers, but it is still existent.

4.2 Intensification by Apparatus and Flow Design

Single-particle correlations appear to be applicable for the computation of local heat and mass transfer between the phases (Bialobrzewski et al., 2008). The resulting transfer coefficients are spatially rather uniform in conventional fluidized beds, but very unevenly distributed in spouted beds (Oliveira et al., 1998), with large maximal values in the spout. The relation between high and low heat and mass transfer regions seems to be more favorable in ProCell units than in conventional spouted beds, most likely due to extremely high air velocities in the gas entrance region of such equipment. Although the resulting process intensification must be seen in connection with pressure drop and the respective operation cost, it is an advantage by itself, as well as in terms of a smaller required apparatus volume, a smaller footprint, and less thermal exposition of heat-sensitive products. 4.2.4 Discrete Particle Modeling

Discrete particle modeling (DPM) is an advanced computational technique for particulate systems (in this case, fluidized beds) that has already been presented in Chapter 7 of Volume 3, Modern Drying Technology. DPM combines continuous (Eulerian) CFD for the gas phase with a discrete (Lagrangian) consideration of the particle phase by means of a discrete element method (DEM), and is therefore often also denoted as DEM-CFD. Its application enables the resolution of not only interactions between the gas and the particle phase, but also of particle–particle and particle–wall interactions, in the sense of a four-way coupling (compare also with Chapter 5 in Volume 1, Modern Drying Technology). The method has been recently applied by Fries (2012) for the inter-comparison of three different fluidized-bed configurations: a conventional fluidized bed with top spray (Fig. 4.17a, d); Wurster equipment (Fig. 4.17b, e); and a ProCell spouted bed unit (Fig. 4.17c, f). All three units correspond to apparatus produced by the Glatt Co. (types GF3, Wurster GF3, and ProCell 5, respectively). Some geometric data are provided in Fig. 4.17, while other data such as operating conditions, material properties and parameters of the simulations are summarized in Tab. 4.7. It should be noted that the particles were assumed to be spheres of equal size, and their number (see Tab. 4.7) corresponds to a holdup of 0.94 kg at a particle density of 1500 kg m3. The hold-up, fluidization gas flow rate and nozzle gas flow rate were the same for all types of investigated equipment. In order to keep the simulations tractable, droplets were not considered; however, a certain volume of each bed was defined as the spray zone. Based on results produced by Fries et al. (2011), the spray zone was assumed to be biconical, with a total length of 45 mm and a spraying angle w ¼ 40 , as depicted in Fig. 4.18. The distributor plate of the Wurster equipment was segmented, with larger orifices at the center. Typical snapshots of particle positions and velocities are shown in Fig. 4.19 for the three investigated apparatuses. The top-spray fluidized bed and the Wurster equipment were cut along the central vertical plane in order to show the particles at

107

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

(a)

(b)

(c)

Nozzle Nozzle Air distributor (sieve plate) z Air distributor (perforated plate)

y x

(e)

(d)

(f)

250 mm 250 mm

Nozzle

hw = 200 mm

dw = 72 mm Wurster tube

hcone = 330 mm hn = 90 mm

108

30°

Nozzle Annulus

180 mm

180 mm

Air inlet

Air inlet

Nozzle

Air inlet

Air inlet

Fig. 4.17 Three-dimensional and cross-sectional representations of the simulated fluidized bed equipment. (a, d) Conventional fluidized bed with top spray; (b, e) Wurster equipment; (c, f) ProCell spouted bed.

Tab. 4.7

Parameters in the simulations of Fries (2012).

Parameter

Value

Particle diameter Number of particles Fluidization air mass flow rate Nozzle air mass flow rate Gap distance below Wurster Normal coefficient of restitution Sliding friction coefficient of particles Rolling friction coefficient of particles Poisson’s ratio of particles Shear modulus of particles DEM simulation time step CFD simulation time step Number of CFD grid cells Total simulation time

2 mm 150 000 360 kg h1 5.7 kg h1 30 mm 0.8 0.1 0.05 0.25 108 Pa 106 s 106 s 9600–17 600 4s

30 mm

d = 50 mm

15 mm

4.2 Intensification by Apparatus and Flow Design

ϕ

length L

Vsprayzone = 29.5 cm3 Nozzle Fig. 4.18 Volume attributed in the simulations to the spray zone.

(a)

(b)

Wurster tube

Nozzle

Nozzle

z

z y

y

x

x

(c)

2.0

1.2 0.8

z

0.4

y x

Nozzle

0.0

Particle velocity [m/s]

1.6

Fig. 4.19 Instantaneous particle positions and particle velocity distributions at identical process conditions. (a) Conventional fluidized bed with top spray; (b) Wurster equipment; (c) ProCell spouted bed (Fries, 2012). Colors indicate the particle velocities.

109

110

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

the center of the fluidization chamber. The images closely resemble the particle motion patterns qualitatively discussed in Section 4.2.1, and illustrate the strong dependence of such motion patterns on apparatus design. An agitated dense bed with particle eruptions caused by bubble breakup at the top is found for the conventional top-spray equipment (Fig. 4.19a). Air from the nozzle, which was assumed to be a two-fluid nozzle, can quite strongly dilute the bed in a region located just below the atomizer, with a shape similar to that of Fig. 4.18. A dense bed with a low particle motion is seen in the lower annulus region of the Wurster equipment (Fig. 4.19b). The solids volume fraction of this region is higher than in the conventional fluidized bed. In the draft tube, particles are transported upwards and drawn towards the center of the tube by the gas flow. The gas and particle velocities are very high, due to segmentation of the distributor plate with larger orifices at the center and concurrent air injection from the atomizer. Even in the expansion chamber above the Wurster tube the gas velocity reaches 5 m s1, compared to a maximum of 0.5 m s1 in the other two devices at the same height. A relatively large number of particles are present in the dilute region above the bed in the Wurster equipment, but no particles rise higher than 300 mm above the bottom in the other two devices. The bed is very dense at the bottom of the ProCell spouted bed (Fig. 4.19c), especially towards the inlet slots along the inclined side walls of this equipment. The spout is much more dilute and dominates the flow pattern. There is a certain fluctuation of the spout asides the central plane of the equipment in time, so that the spout is more or less strongly tilted to one or to the other side in any snapshot picture. Selected results of the DPM study by Fries (2012) are summarized in Tab. 4.8. It can be seen that the relative velocity of collisions between particles, at 0.40 m s1, is much higher in the spray zone of the Wurster unit (zone in front of the nozzle tip, as defined in Fig. 4.18) than in the spray zones of the two other devices (0.23 m s1 in ProCell, 0.29 m s1 in the conventional unit). Since high collision velocities mean that the particles will rather rebound than stick together on wet spots, it may be concluded that the Wurster equipment is more adequate for coating processes than are the other two configurations. A similar, though weaker, trend is observed when considering the relative collision velocity in the whole bed (overall value), and not only in the spray Tab. 4.8

Selected results from the simulations of Fries (2012).

Parameter Relative collision velocity in spray zone Relative collision velocity, overall Particle-particle collision frequency, spray zone Particle-particle collision frequency, overall Collision velocity particle-wall Time-averaged maximal energy dissipation

Unit 1

ms m s1 s1 s1 m s1 mJ

Wurster

ProCell

Conventional

0.40 0.13 136 357 0.26 34.9

0.23 0.11 433 1149 0.51 51.5

0.29 0.08 73 1300 0.40 12.6

4.2 Intensification by Apparatus and Flow Design

zone (relative collision velocity of 0.13 m s1 in Wurster, 0.11 m s1 in ProCell, and 0.08 m s1 in the conventional fluidized bed). Certainly, coating can also be conducted by spraying from the top on a conventional fluidized bed or (and more easily) by spraying from the bottom in a spouted bed, if the operating conditions are appropriately selected for the material to be processed. However, assuming that agglomeration takes place in the spray zones of these two units, it may be concluded from the data listed in Tab. 4.8 that it would be faster in the ProCell, as the frequency of collisions between particles is much higher in the spray zone of this equipment (namely, 433 s1) than in the (diluted) spray zone of the conventional fluidized bed (only 73 s1). If applied to the entire bed, however, the same argument would be slightly in favor of the conventional unit, because of somewhat higher overall particle–particle collision frequency (1300 s1 in the conventional, instead of 1149 s1 in the ProCell unit). Consequently, the rate of the agglomeration process might depend on whether aggregation can take place only in the spray zones or in the entire equipment, which is a matter of drying, because the intensive drying of sessile droplets in the spray zone makes them unusable for agglomeration in the rest of the bed, whereas mild drying preserves the droplets, but would limit the equipment’s capacity. The highest particle–wall collision velocities in the ProCell (0.51 m s1 instead of 0.26 m s1 or 0.40 m s1), and also highest values of maximal energy dissipated during collisions (energy dissipated by the strongest collision at every point of time, averaged over the duration of the simulation: 51.5 mJ in ProCell instead of 34.9 mJ or 12.6 mJ in the two other pieces of equipment), indicate that the highest consolidation (lowest porosity) and the highest mechanical strength of agglomerates are produced in the spouted bed. However, this is combined with the highest risk of breakage of the as-yet unconsolidated agglomerates (which would slow the process), and with the highest risk of abrasion from coated or granulated particles. In total, the analysis by Fries (2012) indicated that spouted beds – in this case, the ProCell unit – might possess the highest versatility for the wet formulation of various products (agglomerates, coated particles, layered granules), coupled with the highest potential for intensification of the respective processes by apparatus and flow design. One important drawback of the DPM results from its high computational costs, which imposes limits on the size and number of particles that can be considered, and on the process time that can be simulated (cf. Tab. 4.7). Any expansion to consider, for example, the droplets, gas-side mass transfer (drying) or the change of particle size by the wet formulation process requires intelligent multiscale approaches. In order to obtain realistic results, parameters defining the interactions between different phases, or among the elements of one and the same phase, must be selected in an educated and elaborate manner, and the same holds for the scaletransition (coarsening) rules. However, such labor can be rewarded by insights into the influence of apparatus design on product properties and process efficiency, as noted above.

111

112

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

4.3 Intensification by Contact Heating 4.3.1 General Principle

The heat necessary in order to evaporate the moisture from wet particles is usually provided in fluidized beds by the gas, so that this gas – typically air – acts as both the fluidizing and the drying agent. However, additional thermal energy can be inserted to the process by heating elements. The function of a heating element can be fulfilled by the jacketed apparatus wall, by some other structural part of the equipment (e.g., the partition of a ProCell spouted bed), or by panels or tubes immersed in the fluidized bed, with steam being the usual heating medium. In all of these cases, heat is transferred to the fluidized particles mainly by contact with the hot surface. This is not meant to completely replace, but rather to augment the convective heat supply from the fluidization gas, which is still present. Therefore, drying processes carried out in this way may be described as “combined convective and contact drying.” A schematic comparison of convective with combined convective and contact drying is presented in the Mollier diagram of Fig. 4.20. The benchmark for this comparison is pure convective drying along the line AC. Here, A represents the thermodynamic condition of inlet air with a temperature of Tg,in, C is the respective condition of adiabatic saturation at a temperature of Tas, and AC is a line of constant adiabatic saturation temperature (prolongation of the fog isotherm at Tas from the oversaturated to the unsaturated region of the plot). The condition of inlet air A will change from A to C along every possible type of convective dryer, if operated adiabatically. In order to prevent condensation, and due to kinetic limitations, the condition of the outlet dryer air will usually be at a certain distance from C, which is denoted in Fig. 4.20 by point B. Projection of the vector AB on the

Fig. 4.20 Schematic comparison of convective drying (AC) with combined convective and contact drying (AE or ADE).

4.3 Intensification by Contact Heating

abscissa of the plot gives the change of gas moisture content in the dryer. The projection of AC expresses the maximal value of this change, that is, the specific moisture uptake capacity of the air (Yas  Yin). With air of given inlet conditions as the only thermal energy carrier, the length of projection AC is unequivocally fixed. However, this projection can be prolonged by _ w , which is noted with the subscript “w” for adding to the dryer a heat flow rate Q “wall” but may be provided from any form of heating element. In a real fluidized _ w will be transferred directly from the heating element of bed, the heat flow rate Q temperature Tw and the surface area Aw to the bed of temperature Tbed with a certain heat transfer coefficient aw, obeying the relationship: _ w ¼ aw Aw ðTw  Tbed Þ: Q

ð4:23Þ

_ w is in reality transferred by contact to the fluidized particles, in order to Though Q easily find the new trajectory of the process it might be imagined as having been _ g and specific heat capacity cP,g, which added to the inlet air of mass flow rate M would then increase its temperature from Tg,in to a higher value of Tg;in
¼ Tg;in þ

_w Q : _ g cp;g M

ð4:24Þ

The new, fictitious condition of inlet air is represented by point D in Fig. 4.20, so that the process would follow a new prolonged fog isotherm towards saturation at point F, with an exit air condition at, perhaps, point E. Alternatively, points A and E may also be connected directly with each other, whereby AE expresses the trajectory of the combined convective and contact drying process. This process is not adiabatic, so that the line AE will not be a line of constant adiabatic saturation temperature. As heat is supplied from the surroundings – in this case from the heating elements – the line AE will be flatter than the line of constant adiabatic saturation temperature through point A, and it may even
be horizontal or upwardly oriented on the plot. In any case, Tas will be higher than
Tas, and the moisture uptake capacity of the air ðYas  Yin Þ will be larger than before (Yas  Yin). Consequently, the process is intensified, which means that more moisture can be removed from the same amount of solids in the same dryer within less time in batch operation, or with less residence time – that is, for a higher product throughput – in continuous operation. Alternatively, smaller equipment may be used at a constant evaporation capacity and batch time or throughput. However, it might be argued that the same benefits could also be obtained by taking, in Fig. 4.20, the path ADE instead of AE. The question is then, why combine warm air with contact heat transfer, and not simply use really hot air? The answer to this point is that combined convective and contact drying can be run at temperatures Tg,in and Tw of similar and moderate magnitude, whereas the temperature Tg,in
has to be significantly larger than both, Tg,in and Tw, in case of purely convective drying. Consequently, combined convective and contact drying can be carried out with a much lower value of the maximal possible temperature in the system than can pure convective drying. As the maximal possible system

113

114

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

temperature is usually limited by specifications for products such as foods or polymers, which may suffer undesirable physical or chemical changes at too-high temperatures, drying in fluidized beds with additional heating elements has a substantial advantage for these types of thermally sensitive materials. A good example of this situation is the drying of the commodity polymer polyvinyl chloride (PVC), using extremely large-scale, cracker-close production units. Keeping Tg,in lower than the grade-dependent threshold temperature, so as to guarantee the avoidance of thermal damage, would correspond to process AB in Fig. 4.20. As the respective water uptake capacity of the air is low, fluidized beds with a huge cross-sectional area would be required for the large-scale production of this material, which would reduce the economic competitiveness and viability of the process. An increase in the gas inlet temperature to Tg,in
would be prohibited by specification, because the PVC particles or deposits would deteriorate; likewise, _ g would not be possible because it would greatly an increase in the gas flow rate M enhance the entrainment of fines. However, the placement of steam tubes in the fluidized bed does enable processing within the necessary thermal restriction at an acceptable investment cost, which means employing a still large but reasonably sized cross-sectional area. Consequently, the processing path AE of Fig. 4.20 is viable, and this is in fact one of the two major technologies used worldwide for the drying of PVC; the second method involves the serial combination of a flash dryer with a so-called “cyclone dryer” (Heinze, 1984; Bunyawanichakul et al., 2006). Although the preceding discussion has explained, on a qualitative basis, the principles of fluidized-bed drying intensification by means of indirect heating, a detailed analysis of the process requires more precise investigation, such as the experimental and theoretical investigations conducted by Groenewold and colleagues (Groenewold, 2004; Groenewold and Tsotsas, 2007; see also Groenewold and Tsotsas, 2001). Therefore, selected results from these authors will be detailed in the following subsections. 4.3.2 Main Effects and Influences

As the experimental investigations performed by Groenewold and coworkers were conducted at the laboratory scale, a single cylindrical immersed element was used that was heated by an electric current rather than by steam. This element was made from copper, had a diameter of 25 mm, a heated length of 50 mm, and a total length of 62 mm. In standard configuration the element was placed vertically at the axis of the (also cylindrical) fluidization chamber, which had an inner diameter of 150 mm and a height of 300 mm. The lower end of the heater was covered with a 10 mmlong poly(tetrafluoroethylene) (PTFE) cup and contacted the porous distributor plate directly; the upper end of the element was extended using a PTFE tube. In a fashion that has already been described in Chapter 1, Volume 2 of this series, the inlet and the outlet moisture content of the air, Yin and Yout, respectively, were measured by infrared spectroscopy, so that with known mass flow rate of dry gas, _ g , and hold-up of dry solids, Ms, the overall water balance was used to calculate M

4.3 Intensification by Contact Heating

both, the solids moisture content XðtÞ ¼ X0 

t _g ð M ðYout  Yin Þdt Ms

ð4:25Þ

t¼0

and the drying rate _ ¼ m

_ g rp;dry d M ðYout  Yin Þ; Ms 6

ð4:26Þ

as functions of time. Here, X0 is the initial moisture content of the solids, rp,dry is the dry particle density, and d is the particle diameter. It should be noted that the _ (evaporation flux) refers to the total surface area of the fluidized drying rate m particles. Simultaneously, the heat transfer coefficient aw was calculated from Eq. 4.23. To this purpose, the temperature in the core of the fluidized bed, Tbed, was measured. Heater temperature was monitored by a Pt100 placed as close as possible to the heater surface and kept approximately constant by a PID controller. The temperature of the heater surface, Tw, was obtained by finite element method (FEM) computations of the temperature field in the heater, being always very close _ w was determined from to the measured temperature. Finally, the heat flow rate Q the electrical power consumption of the heater, after minor correction for its thermal inertia upon slight changes during temperature control. Groenewold performed these investigations in 188 experiments, all of which were in batch operation mode, and with 10 different particle fractions. However, at this point only six of these experiments will be used, performed with three particle fractions, to illustrate the influence of contact heating on fluidized bed drying. A summary of relevant particle properties and operating conditions for the data to be discussed is provided in Tabs. 4.9 and 4.10. Each of the regarded three fractions consisted of spherical c-Al2O3-particles with a narrow size distribution, and are denoted by the same abbreviation (charge name) as in the original studies (Groenewold, 2004; Groenewold and Tsotsas, 2007). As shown in Tab. 4.9, the three fractions differ from each other mainly in terms of particle size, which ranges from 1820 mm (G1800b), over 255 mm (NWA) to 49 mm (NG100). Hence, all particle sizes typically present in fluidized bed drying are covered, as are all relevant Geldart groups (D, B, and A, respectively). The operating conditions of the six selected experiments are summarized in Tab. 4.10. Of these experiments, three were conducted in pure convective mode, Tab. 4.9

Materials in the as-discussed, selected experiments from Groenewold and coworkers.

Charge name

d, mm

rp,dry, kg m3

Geldart group

G1800b NWA NG100

1820 255 49

1070 1325 1850

D B A

115

116

4 Intensification of Fluidized-Bed Processes for Drying and Formulation Tab. 4.10

Main parameters of the as-discussed, selected experiments from Groenewold and

coworkers. Charge

Tg,in,  C

Tw ,  C

_ g , kg h1 M

Yin, g kg1

Ms, kg

G1800b G1800b NWA NWA NG100 NG100

50.2 50.2 50.2 50.1 50.2 49.1

— 140 — 100 — 70

65.3 83.8 17.2 16.6 3.38 2.64

7.69 4.06 4.51 4.49 4.32 6.95

0.810 0.836 0.728 0.705 0.818 0.927

with the heater in place but switched off, so that the field for the value of Tw is void, whilst the other three experiments represented the process of combined convective and contact drying. The inlet gas moisture content Yin was that of ambient air, and the moisture to be removed was demineralized water. Although the gas mass flow rate and solids hold-up could not be maintained exactly the same in runs with and without contact heating, the pairs of runs compared very well one with another. This comparison was carried out in terms of drying curves (as resulting from Eqs 4.25 and 4.26), and the outcome illustrated in Fig. 4.21. The data showed that the intensification of the drying process attained by switching on the heater (empty symbols) was for every particle size very substantial with respect to the basic case of pure convective operation (full symbols), despite the relatively small ratio between the heater and bed cross-sectional area of about 2.8%. The use of a larger heating area in the bed would further increase the drying rates, or it may provide the opportunity to decrease the temperature Tw of the heater at a constant drying rate. This was in full agreement with the qualitative discussion of the preceding section. The lower drying rates for the smaller particles in Fig. 4.21 did not necessarily _ to mean a lower drying performance, but rather resulted from the reference of m the total surface area of the solids, as noted above. With regards to the heat transfer coefficients from Eq. 4.23 for the three experiments with a switched-on heater, the data in Fig. 4.22 shows that such coefficients depend on the moisture content of the solids. With the exception of a transient at the start of the experiments, aw is increasing with the increasing solids moisture content X. This means that the heater-to-bed coefficients that would be obtained for a dry material are additionally enhanced for a wet material, which is an added benefit of the combined drying process in terms of intensification. The additional enhancement was moderate for large particles, such as those of fraction G1800b with an average particle diameter of d ¼ 1820 mm, but was stronger for smaller particles (NWA with d ¼ 255 mm) and strongest for the smallest particles. The extent can be most clearly observed in Fig. 4.23, where the relative heat transfer coefficients arel;exp ðXÞ ¼

aw ðXÞ aw ðX ¼ 0Þ

ð4:27Þ

4.3 Intensification by Contact Heating

Fig. 4.21 Fluidized-bed drying curves of the three selected fractions from Tab. 4.9 for the operating conditions of Tab. 4.10, without (full symbols) and with (empty symbols) contact

heating.(a)Largeparticles,Gl800b;(b)Mediumsizedparticles,NWA;(c)Smallparticles,NG100. Reproduced with permission from Groenewold and Tsotsas (2007).

117

118

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

Fig. 4.22 Heater-to-bed heat transfer coefficients measured during the drying of fluidized particles ofdifferentsizein comparisontothe modelof Martin. ReproducedwithpermissionfromGroenewold and Tsotsas (2007).

Fig. 4.23 Influence of moisture content on wall-to-bed heat transfer by means of relative heat transfer coefficients according to Eq. 4.27. Reproduced with permission from Groenewold and Tsotsas (2007).

4.3 Intensification by Contact Heating

defined by reference to the heat transfer coefficient of dry particles, are plotted. Here, it can be observed that heat transfer to the finest fraction NWA with d ¼ 49 mm is enhanced by more than 50% in comparison to the respective dry product. It should be noted that Macchi et al. (1999) reported an increase of heat transfer to wet particles with a diameter of 2000 mm by about 15% in a spouted bed. Brown et al. (1998) used so-called “phase-change materials” (encapsulated n-octadecane) with particle diameters ranging from 142 to 585 mm, and reported an enhancement of wall-to-bed heat transfer by up to 30% in comparison to respective inert particles. Unfortunately, another, opposing observation can also be made for the 49 mm particles (NG100) in Fig. 4.22. To this purpose, the experimentally determined values of aw were compared with values predicted using a model for contact heat transfer to fluidized particles, as devised by Martin (1984). This model (indicated by solid lines on the plot) meets the measured data at low values of particle moisture content, X, very well for the medium and large-sized particles (Gl800b and NWA, respectively), but misses them for the 49 mm particles. The main reason for this deviation is that the fine particles tend to behave cohesively and agglomerate even in the dry state, effectively behaving like larger particles, in similar fashion to particles with a diameter of 100 mm in the case of fraction NG100. It should be noted that an undesired agglomeration of fines in fluidized beds has been frequently reported in the literature (Moseley and O’Brien, 1993; see also Chapter 7 of Volume 3, Modern Drying Technology), and it has also been reported to reduce contact heat transfer (Pakowski and Mujumdar, 1982). In this respect, part of the advantage of the finest particles in heat transfer will be lost, although the remaining part may be sufficiently high for them to have the highest heat transfer coefficients (as shown in Fig. 4.22). Values of the observed heat transfer coefficients ranged from about 100 to 800 W m2 K1, which is typical of contact heat transfer to fluidized particles. The results of Fig. 4.22 can be further analyzed by defining a new dimensionless heat transfer coefficient by reference of the measured to the calculated heat transfer coefficient for both dry and wet particles as: arel;calc ðXÞ ¼

aexp ðXÞ acalc ðX ¼ 0Þ  acalc ðXÞ aexp ðX ¼ 0Þ

ð4:28Þ

and plotting it in Fig. 4.24. The horizontal basis line of this figure at arel;calc ffi 1 means that influences of particle moisture content included in the model of Martin for the contact heat transfer are the only influences present, and thus accurately captured, whereas higher values indicate additional effects which are not included in this model. Figure 4.24 shows that the former applies for the large particles (G1800b), whereas the latter is the case for the smaller particles (NWA, NG100). To better understand the physical background of the mentioned additional influences, the model of Martin should be briefly commented here, though its equations will not be given, because they are standard and well known from primary literature (Martin, 1984), handbooks (Martin, 2010), and also from Groenewold (2004) and Groenewold and Tsotsas (2007). Martin’s model is a

119

120

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

Fig. 4.24 Transformation of the data from Fig. 4.22 with Eq. 4.28 in order to show the influence of evaporation in the particles on heat transfer. Reproduced with permission from Groenewold and Tsotsas (2007).

particle convection model, transferring ideas from the kinetic gas theory to fluidized particles that come into contact with a heating surface. When such particles are large, they remain nearly isothermal during every single contact with the heating surface, so that the only resistance to heat transfer is by conduction in the gas-filled gap between heater and the particle in short contact with it. This is the so-called “conduction-controlled” limiting case of the model. Additionally, some heat is transferred from the heating element directly to the gas, the flow velocity of which is large with large particles. At the conduction-controlled limit, it does not matter whether the particles are dry or wet. The particle moisture content will influence the wall heat transfer coefficient only via the porosity of the expanded bed, which depends on otherwise constant conditions on particle density and, thus, on the mass of water in the particles. This type of influence can be easily incorporated into the model, and it is the only significant influence of moisture content for large particles, so that the data for fraction G1800b and the predictions of the model are in excellent agreement in Fig. 4.22, lying after transformation according to Eq. 4.28 on the basis line of Fig. 4.24. In contrast, very small particles correspond to the “capacity-controlled” limit of Martin’s model. Remembering that this model has been developed for dry particles which are heated, and not for drying wet particles, it is easy to understand that particles can attain a heating surface temperature after just one wall contact when they are sufficiently small, so that the heat transfer coefficient will then result from the amount of thermal energy that can be taken up by one particle and from the frequency of particle–heater collisions. An influence of moisture content can be formally worked into Martin’s model by simply accounting for the moisture in the

4.3 Intensification by Contact Heating

specific heat capacity of the particles, as if the water would be heated up with the solid skeleton, but not evaporated. Along with the above-mentioned moisturedependent particle density, this leads to the solid lines for calculated heat transfer coefficients in Fig. 4.22. As can be seen, these do not meet the measured data for the fractions NWA and NG100 very well, a point that is also illustrated by the deviation of the transformed data from the basis line in Fig. 4.24. The above-mentioned deviation is not surprising, because it would have been unrealistic to expect the mere consideration of the mass and heat capacity of the liquid to describe the data perfectly. However, it does contain an interesting and not immediately obvious aspect, which concerns the transition from the conductioncontrolled region to the capacity-controlled region of the model. This transition takes place at particle sizes of about 40 mm for dry materials, so that the contact heat transfer to the fluidized bed is rather conduction-controlled above, and capacitycontrolled below, this value (Martin, 1984). Drying particles with a latent heat sink in their interior certainly have a higher heat uptake capacity than dry particles, which undergo just an increase in temperature. Consequently, the influence of drying may be expected to shift the transition from the conduction-controlled to the capacity-controlled regime to smaller particle sizes (smaller than 40 mm), but not to larger particles. However, in that case the fractions NWA and NG100 with d ¼ 255 mm and d ¼ 49 mm, respectively, would be large enough for conduction control, so that they should behave in similar fashion to the very large particles of fraction G1800b, which they do not (Fig. 4.24). Hence, the conclusion is that contact heat transfer to drying particles of the above-mentioned size range is influenced by both, heat conduction through the gas-filled gap between particle and heater during a contact, and heat transfer in the interior of the particle under conditions of phase change by evaporation. The second, particle-side part of the influence makes modeling extremely demanding and tricky, because it constitutes an interrelation between contact heat transfer and drying in the sense of a higher-order, two-way coupling. Additionally, the particle model of Martin should be readjusted to longer than the originally proposed contact times between wall and particles during collisions in order to preserve the experimentally demonstrated influence of particle-side transport phenomena, even for medium-sized particles. These points will be briefly addressed in the following section, along with the more general and – from a practical viewpoint – more important task of calculating the combined convective and contact drying in fluidized beds with reasonable accuracy. 4.3.3 Further Remarks on Modeling

Groenewold, in his studies, evaluated 10 different model versions of combined convective and contact fluidized-bed drying by comparing all experimental results (Groenewold, 2004; Groenewold and Tsotsas, 2007). From this extensive evaluation, only three model versions will be briefly discussed here, denoted by the same letters as in the original studies, namely C, J, and G.

121

122

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

Model version C is short-cut and very easy to apply. The drying kinetics of the material is described by the normalized, single-particle drying curve f(W) according to the original proposal by van Meel (1958). As already discussed in Chapter 1, Volume 2 of this series, f is the drying rate, which is made nondimensional by dividing with the constant, first period drying rate, whereas W is the moisture content, normalized by use of Xcr for the transition from the first to the second drying period and Xeq for the adsorption equilibrium moisture content at the end of the process. As the normalization concept is empirical, f(W) has to be found by measurement. The drying kinetics of the single particle is then scaled-up to the fluidized bed by an appropriate model. In conjunction with the function f(W), this model does not need to consider explicitly the energy balances, so that the simple formulation introduced by Groenewold and Tsotsas (1997) is sufficient. In this, the particles are ideally back-mixed, bubbles are considered to rise through the bed in plug flow, exchanging heat and mass with the suspension phase (dense phase). Plug flow is also assumed for the gas flowing through the suspension phase, which is primarily responsible for the drying, but this assumption is corrected by incorporating the influence of suspension gas back-mixing into apparent Sherwood numbers for particle-to-gas mass transfer. The previously stated Eq. 4.19 describes such apparent Sherwood numbers. All model parameters are taken from the literature, so that there is no degree of freedom for fitting. Contact heat transfer is accounted for in model version C by artificially increasing the temperature of the inlet air to Tg,in
according to Eq. 4.24. To obtain the necessary heat flow rate from Eq. 4.23, the wall heat transfer coefficient aw is calculated by using the original model of Martin. Herein, the influence of moisture content on particle density and heat capacity is taken into account, while the previously discussed, “second-order” influences of phase transition in the particles during heat transfer are simply ignored. Consequently, the heat transfer coefficient aw has exactly the values shown by the solid lines in Fig. 4.22. Drying curves computed in this way correspond to the solid lines in Fig. 4.25. As Fig. 4.25a shows, the agreement with the measured data is excellent for fraction G1800b. However, for fraction NWA more deviation occurs between measurement and prediction (Fig. 4.25b), because the contact heat transfer coefficients cannot be described exactly by Martin’s model for these particles (Fig. 4.22), due to the neglected influence of the latent, evaporative heat sink. Especially, the slope of the measured drying rates with decreasing moisture content is not captured exactly; rather, it is underestimated at high values of X, and overestimated at low values. However, the prediction result is still usable from a practical point of view, which means that the intensification caused by contact heat transfer over the entire duration of the process is evaluated quite accurately. Because of the identified effect of agglomeration, no comparison will be made for fraction NG100. The results of model version J are also depicted (as broken lines) in Fig. 4.25. This model version is similarly simple and almost identical to model version C, apart from the fact the heat transfer coefficient from the heating element is now calculated by assuming an infinitely large specific heat capacity of the particles in Martin’s model. This is equivalent to saying that the evaporative sink in the

4.3 Intensification by Contact Heating

Fig. 4.25 Prediction of drying curves by the short-cut model versions C and J for (a) large particles, G1800b, and (b) medium-sized particles, NWA. Reproduced with permission from Groenewold and Tsotsas (2007).

particles is so strong as to prevent any penetration of heat into them, irrespective of the particle size, and removing completely the capacitive limitation from the model of Martin. For coarse particles, the result of model version J is the same as the result of model version C, and can thus not be distinguished graphically in Fig. 4.6a, because heat transfer to such particles is anyway controlled by conduction

123

124

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

Fig. 4.26 Prediction of drying curves by sophisticated model versions, especially model version G, for medium-sized particles, NWA. Reproduced with permission from Groenewold and Tsotsas (2007).

in the gas-filled gap to the heater, and not by the heat uptake capacity. For the medium-sized particles (Fig. 4.6b), the drying rates predicted by model version J are somewhat higher (though not by much), because even such particles are located rather in the conduction-controlled region according to the original version of the model, as previously discussed, so that the removal of the capacity limitation does not play a crucial role. The best prediction of the drying curve of fraction NWA was achieved by Groenewold and colleagues with model version G. The results of this model version correspond to the upper, somewhat shaky curve in Fig. 4.26, and they are in excellent agreement with the experimental results for particle moisture contents larger than about X ¼ 0.04. The price for the better agreement is, though, a much more complex structure of model version G. In this version, single-particle drying is modeled, rather than by normalization, by a numerical solution of the volumeaveraged mass and heat transfer equations, similar to the equations presented in Chapter 1, Volume 1 of this series (see also Whitaker, 1977). Numerical instabilities created the above-mentioned fluctuations of the solution. With regards to the fluidized particle system, an explicit consideration of the energy balances is now necessary, and this corresponds to an upgrade of the approach of Groenewold and Tsotsas (1997) published by Burgschweiger and Tsotsas (2002), which is more sophisticated and profound than the original, but also more difficult to implement. Moreover, in order to be able to harvest the effect of the coupled consideration of heat transfer and drying in the heat transfer coefficient, the model of Martin must be modified to larger particle–wall contact times during collisions (cf. Molerus et al., 1995). In doing so, the good agreement of this model with a large amount of

4.3 Intensification by Contact Heating

heat transfer data for dry materials should not be lost, and a compensatory correction in the gas–gap conduction term is therefore necessary. However, the extent of this compensation has certain limits (Groenewold and Tsotsas, 2007), and its physical justification is not evident, so that the heat transfer part of the model might be considered as overmanipulated. In combination with the high computational demands of model G, it may be concluded that the short-cut versions C and J are more appropriate for practical tasks. It should be noted that the here undiscussed model versions, such as A and D from Fig. 4.26, include some interesting elements. In some cases, the modified normalization of single-particle drying kinetics as proposed by Burgschweiger et al. (1999) was used. This enabled a separation of the influences of water vapor pressure decrease by adsorption and intraparticle transport resistances – which are intermingled in the van Meel approach – from each other. It requires use of the full fluidized-bed drying model of Burgschweiger and Tsotsas (2002), but makes sense for highly hygroscopic materials, such as c-Al2O3. In another version, the particle model of Martin was replaced with the packet model of Kunii and Levenspiel (1991) for calculating the contact heat transfer coefficient aw. The latter is based on the penetration of heat in assemblies of many particles (“packets”) in contact with the heating surface for a certain time, being similar to contact drying models (Tsotsas et al., 2007; Kwapinska et al., 2008). However, none of the additionally investigated versions was significantly more successful in predicting the measured data than those discussed, nor had any grave advantages from a theoretical point of view. With regards to the many additional experiments conducted by Groenewold, a series performed with a horizontal arrangement of the heater should be mentioned. In these runs, the same heater was located in the middle of a diagonal of the bed, with a distance of 70 mm between its axis and the distributor plate. Both heater ends were insulated using PTFE cups. For particles larger than 770 mm, the measured contact heat transfer coefficients and drying curves were approximately the same as in case of the vertically placed heater, whereas somewhat smaller heat transfer coefficients were obtained with the horizontal arrangement for particles smaller than 255 mm. It should be noted that none of the 10 investigated particle fractions had an average diameter in the range of 255–770 mm. Results on the influence of the arrangement of horizontal tubes on bubbling and heat transfer in large fluidized-bed dryers have been reported by Jonassen (1999), though for a limited selection of particulate materials (potato cubes were the only wet material used). In total, the main results of the comprehensive investigations performed by Groenewold can be summarized as follows:  A very significant intensification of fluidized-bed drying can be attained by using immersed heating elements or heated walls.  The drying process can be computed quite easily and accurately for particles larger than 770 mm.  A coupling between contact heat transfer and particle drying occurs for particles smaller than 255 mm. This is additionally beneficial in terms of process

125

126

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

intensification, but difficult to simulate. Relatively simple models still provide acceptable results. Complex models are somewhat more accurate, but cumbersome in their implementation and in need of further development in their theoretical details, perhaps in combination with CFD (cf. Kuipers et al., 1992; Gustavson and Almstedt, 2000; Wang and Rhodes, 2003).  Very small particles of about 50 mm may tend to agglomerate, even in a dry state, which would reduce the beneficial influence of contact heating.

4.4 Further Methods of Intensification

The principle of fluidized-bed drying intensification, as described in Section 4.3, was to combine the energy supply that takes place by means of the fluidization gas with additional energy supplied by contact of the particles with heated surfaces. In a similar way, other methods of supplementary energy input can also be used, such as microwave heating or heating by infrared radiation. Water in the material to be dried can be liquid, or it can be frozen, as in fluidized-bed atmospheric freezedrying. Although a variety of hybrid processes have emerged in this way they will not be discussed at this point; rather, they are described in detail in other chapters of this volume of Modern Drying Technology (notably Chapters 5, 9, and 10). It should be noted that, irrespective of the energy source used, this energy can be directly provided to the fluidized particles to be dried, or it can be supplied first to inert, typically large particles, which transfer then the heat in their turn to the (typically small) product particles. In such a process the large particles take on the role of the immersed heating elements of Section 4.3, the difference being that they are not fixed in space but are mobile. Their own source of energy can derive from the hot fluidization gas (i.e., convective), from hot walls or inlets (by contact), via dissipation of microwaves, or from a radiant emitter. As already mentioned, these pathways are treated either in the present chapter or in chapters elsewhere in this volume; however, one possible and interesting track that is not discussed elsewhere is briefly introduced here, namely heating by induction. The pathway of induction means that electric energy is applied to create a magnetic field in the fluidization chamber; the magnetic field, in turn, creates an electric current in large and inert metallic particles, which are co-fluidized with the product. The energy of this electric current is immediately dissipated as heat. Such heat is transferred to the fluidization gas and then from the gas to the particles of the product to be dried; alternatively, it may be transmitted directly to the product particles by contact (by collisions between inert and product particles). In order for this process to function, the inert metallic particles and wet product particles must fluidize in a rather homogeneous fashion – that is, their fluidization properties must match. The Archimedes number (see Eq. 4.3) shows that fluidization behavior depends on both the diameter and density of particles; consequently, compact metallic particles will never have the same fluidization behavior as would, for example, food particles, unless they are much smaller. However, as it is

4.5 Conclusion

preferable to use larger rather than smaller inert particles, they must have a lower density than the product, but be metallic in nature. This is achieved by using hollow metallic spheres that can be fabricated with different wall strengths so as to enable the above-mentioned matching. The first investigations into induction (Stresing et al., 2011, 2013) showed the process to be extremely efficient in terms of heat transfer, with very high gas temperatures being attained very soon after switching on the induction field. A fast thermal response occurring after an essentially instantaneous change of electrical input can provide new opportunities for the automatic control of drying processes. It has also been shown that the magnetic field can be easily and successfully focused on the fluidization chamber, with very small losses to the environment. Clearly, continuing studies in this area are expected to establish a new and innovative alternative for the enhancement of fluidized-bed drying by induction heating.

4.5 Conclusion

With heat transfer from gas to particles occurring in all fluidized-bed drying processes, additional heat can be supplied in a variety of ways, either directly to the product or via cofluidized inert particles. Moreover, both the fluidization chamber geometry and air inlet design can be varied in many different ways, for example replacing conventional with spouted fluidization. The hybridization of energy supplies and apparatus design can contribute to significant process intensification, which in turn can lead to higher production capacities using constant sizes of equipment and thermal exposition, or to smaller equipment with less thermal exposure of the product at a constant throughput. Combination of these factors can provide rich opportunities for improved and new industrial processes of drying and wet formulation, and also present manifold challenges and opportunities for research and scientific development.

Additional Notation Used in Chapter 4

G x

ratio of air inlet to bed surface area width of air entrance gap

– m

Greek Letters

c w w

bottom half-angle of cone angular position of gas entrance valve half-angle of spray cone

rad rad rad

127

128

4 Intensification of Fluidized-Bed Processes for Drying and Formulation

Subscripts and Superscripts

app elu f st 0


apparent elutriation corresponding to bed weight stable operation initial, superficial, reference, static (packed bed) fictitious gas inlet condition, modified variable

Abbreviations

CFD CSTR DPM FEM NTU PFR PID PTFE PVC

computational fluid mechanics continuous stirred tank reactor discrete particle modeling finite element method number of transfer units plug flow reactor proportional integral derivative poly(tetrafluoroethylene) poly-vinyl-chloride

References Asenjo, J. A., Munoz, R., Pyle, D. L., 1977. On the transition from a fixed to a spouted bed. Chem. Eng. Sci. 32: 109–117. Bialobrzewski, I., Zielinska, M., Mujumdar, A. S., Markowski, M., 2008. Heat and mass transfer during drying of a bed of shrinking particles: Simulation for carrot cubes dried in a spout-fluidised-bed drier. Int. J. Heat Mass Transfer 51: 4704–4716. Brown, R., Rasberry, J., Overmann, S., 1998. Microencapsulated phase-change materials as heat transfer media in gas-fluidized beds. Powder Technol. 98: 217–222. Bunyawanichakul, P., Kirkpatrick, M. P., Sargison, J. E., Walker, G. J., 2006. A threedimensional simulation of a cyclone dryer. Proceedings of 5th International Conference on CFD in the Process Industries, Melbourne, Australia, 13–15 December 2006. Burgschweiger, J., Groenewold, H., Hirschmann, C., Tsotsas, E., 1999. From

hygroscopic single particle to batch fluidized bed drying kinetics. Can. J. Chem. Eng. 77: 333–341. Burgschweiger, J., Tsotsas, E., 2002. Experimental investigation and modelling of continuous fluidized bed drying under steady-state and dynamics conditions. Chem. Eng. Sci. 57: 5021–5038. El-Naas, M. H., Rognon, S., Legros, R., Mayer, R. C., 2000. Hydrodynamics and mass transfer in a spouted bed dryer. Drying Technol. 18: 323–340. Flisjuk, O. M., Rachmatov, A. M., Raschkowskaja, N. B., 1984. Determination of fundamental construction and hydrodynamic parameters of spouted bed equipment with tangential gas supply and vertical partition wall (in Russian). Zh. Prikl. Chimii 57: 954–956. Fries, L., 2012. Discrete particle modelling of a fluidized bed granulator. Diss.,

References Technical University Hamburg-Harburg, Germany. Fries, L., Dosta, M., Antonyuk, S., Heinrich, S., Palzer, S., 2011. Moisture distribution in fluidized beds with liquid injection. Chem. Eng. Technol. 34: 1076–1084. Geldart, D., 1973. Types of gas fluidization. Powder Technol. 7: 285–292. Gnielinski, V., 2010. Fluid-particle heat transfer in flow through packed beds of solids, in Heat atlas, 2nd edn, Springer, Berlin, Germany, Sect. G9, pp. 743–744. Goroshko, V. D., Rozenbaum, R. B., Todes, O. M., 1958. Approximate hydraulic relationships for suspended beds and hindered fall. Izvestiya VUZOV, Neft I Gaz. 1: 125–131. Gorshtein, A. E., Mukhlenov, I. E., 1964. Zh. Prikl. Khim. 37: 1667. Grace, J. R., 1986. Contacting models and behavior classification of gas-solid and other two-phase suspensions. Can. J. Chem. Eng. 64: 353–363. Groenewold, H., 2004. Wirbelschichttrocknung mit indirekter Beheizung. Diss., Otto von Guericke University Magdeburg, Germany. Groenewold, H., Tsotsas, E., 1997. A new model for fluid bed drying. Drying Technol. 15: 1687–1698. Groenewold, H., Tsotsas, E., 1999. Predicting apparent Sherwood numbers in fluidized beds. Drying Technol. 17: 1557–1570. Groenewold, H., Tsotsas, E., 2001. Experimental investigation and modelling of the influence of indirect heating on fluidized bed drying. Drying Technol. 19: 1739–1754. Groenewold, H., Tsotsas, E., 2007. Drying in fluidized beds with immersed heating elements. Chem. Eng. Sci. 62: 481–502. Gryczka, O., 2009. Untersuchung und Modellierung der Fluiddynamik in prismatischen Strahlschichtapparaten. Diss., Technical University Hamburg-Harburg, Germany. Gustavson, M., Almstedt, A., 2000. Numerical simulation of fluid dynamics in fluidized beds with horizontal heat exchanger tubes. Chem. Eng. Sci. 55: 857–866. Heinze, C., 1984. New cyclone dryer for solid particles. Ger. Chem. Eng. 7: 249–279. Hoffmann, T., Hailu Bedane, A., Peglow, M., Tsotsas, E., Jacob, M., 2011. Particle-gas

mass transfer in a spouted bed with adjustable air inlet. Drying Technol. 29: 257–265. Jonassen, O., 1999. Heat transfer to immersed horizontal tubes in gas fluidized bed dryers. Diss., Norwegian University of Science and Technology, Trondheim. Jordanova, E. N., Mitev, D. T., M€orl, L., 2000. Hydrodynamische Kennwerte eines Strahlschichtapparates mit zwei Gaseing€angen. Chem.-Ing.-Tech. 72: 1059–1060. Kmiec, A., 1980. Hydrodynamics of flows and heat transfer in spouted beds. Chem. Eng. J. 19: 189–200. Kmiec, A., Englart, S., Agnieszka, L., 2009. Mass transfer during air humidification in spouted beds. Can. J. Chem. Eng. 87: 163–168. Kuipers, J. A. M., Prins, W., vanSwaaij, W., 1992. Numerical calculation of wall-to-bed heat transfer coefficients in gas-fluidized beds. AIChE J. 38: 1079–1091. Kunii, D., Levenspiel, O., 1991. Fluidization engineering. 2nd edn, Butterworth, Boston, USA. Kwapinska, M., Saage, G., Tsotsas, E., 2008. Continuous versus discrete modelling of heat transfer to agitated beds. Powder Technol. 181: 331–342. Li, Y., Kwauk, M., 1980. The dynamics of fast fluidization, in Fluidisation III, (eds J. R. Grace, J. M. Matsen), Plenum, New York, USA, pp. 537–544. Macchi, A., Bi, H., Legros, R., Chaouki, J., 1999. An investigation of heat transfer from a vertical tube in a spouted bed. Can. J. Chem. Eng. 77: 45–53. Martin, H., 1984. Heat transfer between gas fluidized beds of solid particles and the surface of immersed heat exchanger elements. Chem. Eng. Process. 18: 157–169, 199–223. Martin, H., 2010. Heat transfer in fluidized beds, in Heat atlas, 2nd edn, Springer, Berlin, Germany, Sect. M5, pp. 1301–1311. Mathur, K. B., Epstein, N., 1974. Spouted beds. Academic Press, New York, USA, p. 284. Mathur, K. B., Ghisler, P. E., 1955. A technique for contacting gases with coarse solid particles. AIChE J. 157–164.

129

130

4 Intensification of Fluidized-Bed Processes for Drying and Formulation Mitev, D. T., 1967. Investigation of fluidized bed hydrodynamics in prismatic equipment (in Russian). Diss., LTI Leningrad, UDSSR. Mitev, D. T., 1979. Theoretical and experimental investigation of hydrodynamics, heat transfer and mass transfer in spouted bed equipment (in Russian). Habil., LTI Leningrad, UDSSR. vanMeel, D. A., 1958. Adiabatic convection batch drying with recirculation of air. Chem. Eng. Sci. 9: 36–44. M€ orl, L., Heinrich, S., Kr€ uger, G., Ihlow, M., Jordanova, E., 2000. Steuerbare Gasanstr€omeinrichtung f€ ur Strahlschichtapparate. German Patent No. 10004939. Molerus, O., 1982. Interpretation of Geldart’s type A, B, C and D powders by taking into account interparticle cohesion forces. Powder Technol. 33: 81–87. Molerus, O., Burschka, A., Dietz, S., 1995. Particle migration at solid surfaces and heat transfer in bubbling fluidized beds. Chem. Eng. Sci. 50: 871–877, 879–895. Moseley, J. L., O’Brien, T. J., 1993. A model for agglomeration in a fluidized bed. Chem. Eng. Sci. 48: 3043–3050. Nikolaev, A. M., Golubev, L. G., 1964. Izv. Vyssh. Ucheb. Zaved. Khim. Teknol., 7: 855. Oliveira, W. P., Freire, J. T., Massarani, G., 1998. Analogy between heat and mass transfer in three spouted bed zones during the drying of liquid materials. Drying Technol. 16: 1939–1955. Olazar, M., San Jose, M. J., Aguayo, A. T., Arandes, M. J., Bilbao, J., 1992. Stable operation conditions for gas-solid contact regimes in conical spouted beds. Ind. Eng. Chem. Res. 31: 1784–1792. Olazar, M., San Jose, M.J., Aguayo, A.T., Arandes, M.J., Bilbao, J., 1993, Design factors of conical spouted beds and jet spouted beds. Ind. Eng. Chem. Res. 32: 1245– 1250. Pakowski, Z., Mujumdar, A. S., 1982. Heat transfer from a horizontal cylinder to a vibrated fluid bed of wet particles. Proceedings of 3rd International Drying

Symposium (IDS1982), Birmingham, UK, pp. 149–155. Piskova, E., 2002. Untersuchung der Fluiddynamik eines Strahlschichtapparates mit zwei parallelen Gaseintritten und seine Anwendung auf die Beschichtung feindisperser Feststoffteilchen. Diss., Otto von Guericke University Magdeburg, Germany. Raschkowskaja, N. B., 1969. Investigation of the drying process of granular material, pastes and solutions from the chemical industry in a spouted bed (in Russian). Diss., LTI Leningrad, UDSSR. Romankow, P. G., Raschkovskaja, N. B., 1968. Drying in fluidized beds (in Russian), Chimia, Leningrad, UDSSR, p. 357. Stresing, A., M€orl, L., Neum, J., Jacob, M., Walther, K., 2011. Non-contact energy transfer to a fluidized bed. Proceedings of European Drying Conference (EuroDrying’2011), Palma de Mallorca, 26– 28 October. Stresing, A., M€orl, L., Khaidurova, A., Jacob, M., Walther, K., 2013. Bestimmung des Zeitverhaltens einer induktiv beheizten Wirbelschicht und deren Einflussgr€oen. Chem.-Ing.-Tech. 32: 308–312. Szafran, R. G., Kmiec, A., 2004. CFD modeling of heat and mass transfer in a spouted bed dryer. Ind. Eng. Chem. Res. 43: 1113–1124. Tsotsas, E., Kwapinska, M., Saage, G., 2007. Modeling of contact dryers. Drying Technol. 25: 1377–1391. Uemaki, O., Kugo, M., 1968. Mass transfer in spouted beds. Kagaku Kogaku 32: 89. Wang, X., Rhodes, M., 2003. Determination of particles residence time at the wall of gas fluidized beds by discrete element method simulation. Chem. Eng. Sci. 58: 387–395. Whitaker, S., 1977. Simultaneous heat, mass, and momentum transfer in porous media: A theory of drying. Adv. Heat Transfer 13: 119– 203. Wurster, D. E., 1959. Air suspension technique of coating drug particles. J. Am. Pharm. Assoc. 48: 451–454.

131

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries Roberto Pisano, Davide Fissore, and Antonello A. Barresi 5.1 Introduction

Commonly, process intensification entails the development of new technologies (i.e., methods and/or equipment) that, compared to those traditionally used today, show the way towards a dramatic improvement in manufacturing. Process intensification is therefore a strategy used by companies to enhance production in terms of reductions in equipment size, increases in the production capacity within a given equipment volume, decreases in energy consumption, reductions in wasted energy and also in the costs associated with waste products. The final result of process intensification is therefore the development of smaller, cleaner, and energyefficient technologies (Stankiewicz and Moulijn, 2000). When freeze-drying pharmaceutics and foods, industry strives to develop new technologies in order to reduce drying times and hence energy consumption. Typically, these new technologies are product-specific, as requirements for process intensification may vary with the original physical state of products being freezedried. In this regard, products can be classified into three categories: liquids; individually quick-frozen products; and a combination of these. In general, in the pharmaceutical industry the materials being dried are solutions, whereas the food industry treats both liquids and solids. Process constraints may be different for the pharmaceutical and food industries, as well as for the above types of product. It follows that a process modification can be beneficial or applicable only to specific material types or equipment. For example, in the pharmaceutical industry sterility is often a predominant requirement, while minor attention is given to energy efficiency. Therefore, a new technology that enhances heat and mass transfer, but compromises sterility conditions, is not suitable for the drying of drugs. Both, the freezing and the drying processing steps must be considered for process intensification (Elia and Barresi, 1998). For liquid products, new methods have been introduced to control freezing, as this impacts on product characteristics and, subsequently, the drying phase. In addition, automatic control systems were

Modern Drying Technology Volume 5: Process Intensification, First Edition. Edited by Evangelos Tsotsas and Arun S. Mujumdar. Ó 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

132

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

designed for the drying stage (Pisano et al., 2010; Barresi and Fissore, 2011) in order to maximize the rate of water removal, and to reduce drying times and, in turn, the processing costs. It must also be noted that freezing has its own time demands, which can also be reduced. In this respect, unconventional methods have been investigated in order to intensify the freezing of both solids and liquids (Elia and Barresi, 1998). Evaporative freezing is one of these methods, and when Ghio et al. (2000) compared the evaporative and conventional freezing of vegetables they showed that pressure reduction would allow a rapid freezing of the material being dried, as well as a partial evaporation of water (ca. 20–40% of the total). This phenomenon makes it possible to reduce both the time and the energy consumption for the subsequent drying step. However, the use of this method is limited as it may damage the product due to cracks being formed during the depressurization. While the pharmaceutical industry produces high-value products which can justify the high costs of freeze-drying, the food industry produces lower-value products and has to face large production volumes. For this reason, new methods have been proposed for food manufacturing as a valid alternative to vacuum freezedrying. Such an example is the atmospheric freeze-drying that, as shown in this chapter, allows a dramatic reduction in energy consumption. Of course, any new method implies the development of novel and, perhaps, unconventional equipment, and atmospheric freeze-drying has not escaped this trend. New techniques using alternative energy sources, such as ultrasound and microwave, were employed to further accelerate the rate of drying and improve the energy efficiency of the process. For this purpose, hybrid operations were also proposed that involve more drying techniques and can lead to high-quality product and faster drying (Venkatachalapathy and Raghavan, 1999). Although process intensification mainly focuses on novel methods and equipment, it can also entail the rediscovery of established technologies, which were applied on the laboratory scale or developed for a different application. A typical example is the continuous freeze-drying plant (Rey, 2010), a technology that has been widely used in food industry since the 1960s and has led to larger productivity and more homogeneous products. In the biological and pharmaceutical industries, freeze-drying is still a batch operation and, despite the elaborate equipment design and sophisticated control systems recently introduced, the heterogeneous drying behavior within a batch – as well as from batch to batch – remains a problem of great concern. One potential solution to this problem is to shift from batch to semi-batch or continuous operation, as did the food industry many years ago. In this chapter, attention is focused on the challenge of process intensification in freeze-drying. The problem of process optimization is first addressed, and exergy analysis is used to identify the best operating conditions to be implemented during drying. Some of the methods and technologies proposed to date to reduce drying times and leading to a more sustainable process are then described and discussed. The advantages and disadvantages of the above methods, as well as their practicality for different types of material, are also outlined.

5.2 Exergetic Analysis (and Optimization) of the Freeze-Drying Process

5.2 Exergetic Analysis (and Optimization) of the Freeze-Drying Process

The amount of energy required by a freeze-drying process is very large – in fact, it is almost double that used in a conventional drying process (Flink, 1977). The most critical step is the primary drying, as this accounts for about 45% of the total energy required by the process (Ratti, 2001). In order to save energy and maximize productivity, besides preserving product quality, one possible strategy consists of minimizing the duration of the process: this requires an identification of the values of temperature of the heating shelf (Tshelf) and of pressure in the drying chamber (Pc) that make it possible to minimize the drying time, besides satisfying the constraints on the maximum temperature allowed by the product, and on the maximum sublimation flux allowed by the freeze-dryer. This problem can be solved either in-line or off-line (Pisano et al., 2013). In the first case, it is necessary to use a monitoring and control system (Barresi et al., 2010; Pisano et al., 2010; Barresi and Fissore, 2011): mathematical modeling is required to estimate the state of the system on the basis of the available measurements (Barresi et al., 2009; Fissore et al., 2011b) and to calculate the values of the manipulated variables (with the goal of achieving the aforementioned goals). In the second case, mathematical modeling can be used to calculate off-line the design space of the primary drying, thus determining the optimal values of the operating conditions (Giordano et al., 2011; Fissore et al., 2011a). The exergy analysis of the process can be a powerful tool to identify the best operating conditions. Exergy is defined as the maximum amount of work that can be extracted from a physical system by exchanging matter and energy with large reservoirs at reference states. The exergy is thus a measure of energy quality, and can be associated with the irreversibility occurring during the process. In order to optimize the freeze-drying process from the energetic point of view, it is thus required to minimize the exergy losses; this would improve the economic efficiency (and the sustainability) of freeze-drying by increasing the efficiency of energy utilization in the process. Few reports have been made in the scientific literature regarding the exergy analysis of freeze-drying processes. Liapis and Bruttini (2008) provided a detailed exergy analysis of the various steps of a freezedrying process (freezing, primary and secondary drying), as well as of vacuum pumping and of vapor condensation, by using a detailed two-dimensional model, while Liu et al. (2008) used a one-dimensional model to carry out a similar analysis. In order to combine information obtained through the design space of the process with the exergy analysis, a simplified approach is required: many calculations are necessary, as the exergy analysis needs to be repeated for all the potential values of Tshelf and Pc that could be used to carry out the process, so that computational time becomes an important issue. Fissore et al. (2014) used the simplified one-dimensional model of Velardi and Barresi (2008) to calculate the temperature profiles in the frozen layer and the sublimation flux as a function of the operating conditions, while the temperature profile in the dried layer is calculated solving the Fourier equation written for the dried cake. The approach of

133

134

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

mass flux x=0

drying chamber

dried product

x = Ld

interface of sublimation frozen product

x = L0

heat flux

shelf

Fig. 5.1 Sketch of a freeze-drying process during the primary drying stage.

Liu et al. (2008) can be used to calculate the exergy losses, but some assumptions of these authors have been removed as their validity appears to be questionable; for example, that the vapor pressure at the sublimation interface is equal to the chamber pressure. The exergy losses due to heat transfer during primary drying pd (Ex loss;h ) are calculated from the temperature profile in the dried layer (see Fig. 5.1):   ð td ð Ld T 0 ld dT 2 pd dxdt; ð5:1Þ Exloss;h ¼ A T 2 dx 0 0 where A is the area of the sublimation interface, T0 is a reference temperature, T the temperature of the product at axial position x, ld and Ld are respectively the thermal conductivity and the thickness of the dried cake, and td is the duration of primary drying. The exergy input resulting from heat transfer above the reference temperature pd (Ex in;h ) is given by:    ð td  dT T0 pd ld Exin;h ¼ A dt: ð5:2Þ 1 dx x¼0 T x¼L0 0 Assuming that the reference temperature T0 is well below product temperature during primary drying, then the exergy input due to heat transfer is given by:       Ti T i T freeze T0 pd þ mwet Dhsub 1 þ Ex in;h ¼ T 0 mwet cp;f ln T0 T freeze Ti    
T0 T0 Ti ; þ T 0 mw;in mw;fin cp;d ln Ti T0 ð5:3Þ where mwet is the mass of product, mw,in and mw,fin are respectively the mass of water in the product at the beginning and at the end of primary drying, Tfreeze is the product temperature at the end of the freezing stage, and Ti is the product temperature at the sublimation interface. With respect to mass transfer in the product being dried, the presence of noncondensable gases resulting from leakage and outgassing of the surfaces in the

5.2 Exergetic Analysis (and Optimization) of the Freeze-Drying Process

system, as well as inert gases from the material being dried, can be neglected. pd Thus, the exergy losses due to water vapor flux in the dried layer (Ex loss;m ) are given by:   ~ ð td ð Ld AT 0 R 1 dpw ðxÞ pd Exloss;m ¼ jw dxdt; ð5:4Þ ~w 0 pw ðx Þ dx M 0 where jw is the sublimation flux, and pw is water vapor partial pressure at axial pd position x. The exergy input in the dried layer due to mass transfer (Exin;m ) is given by:     ð td   ~ pw;i Ti Ti R pd 1 ln ln þ jw c p;v Exin;m ¼ AT 0 dt: ð5:5Þ ~ T T P0 M 0 0 0 w The vapor produced in the drying chamber arrives onto the cold surface of the condenser. The change in temperature, as well as in the state, of the vapor in the condenser can be described as follows:  the temperature of the vapor (Tin,cond) is cooled down to the temperature that causes desublimation of vapor (Tdes);  the vapor desublimates at the temperature Tdes;  the temperature of the ice decreases to reach the final value (Tout,cond). The energy input of the vapor condenser is given by: ð td
 Q cond ¼ A jw T in;cond T des cp;v þ Dhs þ T des 0

Then, the exergy input (Ex cond in ) can be calculated as:   T0 cond 1 ; Exin ¼ Q cond T cooling


T out;cond cp;ice dt:

ð5:6Þ

ð5:7Þ

where Tcooling is the temperature of the cooling medium. The exergy losses (Ex cond loss ) can be calculated using the following equation:   
T des T cond;in T des þ Ex cond mw;fin
T 0 cp;v ln loss ¼ mw;in T cooling T cond;in     T out;cond T out;cond T des 1 1 þ T 0 Dhs þ T 0 cp;ice ln T des T cooling T cooling T des : ð5:8Þ The vacuum pump is used to reduce pressure in the drying chamber during the start-up stage of the process, and to evacuate the noncondensable gases resulting from leakage (air), from controlled leakage used for pressure control (nitrogen, in most cases), and from the material being dried. The exergy input of the vacuum system is given by the power input multiplied by the drying time. In order to calculate the power input it is necessary to know the flow rate of gases that are

135

136

Exloss, %

(a)

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

(b) 70

70

60

60

50

50

40

40

30

30

20

20

10

10 0

0 5

10 15 20 chamber pressure, Pa

5

10 15 20 chamber pressure, Pa

Fig. 5.2 Values of the exergy losses (given as percentage of the total exergy losses) for different values of chamber pressure in the primary drying stage in the drying chamber (&), for vapor condensing ( ) and vacuum pumping (&). (a) Tshelf ¼ 20  C, (b) Tshelf ¼ 0  C.

pumped out. The exergy losses due to the compression of a perfect gas can be written as:   ~ R P out pump Exloss ¼ mg T 0 ln ; ð5:9Þ ~g Pcond M where mg is the mass of noncondensable gases evacuated by the pump, given by the product of the gas flow rate due to leakage (that can be determined by means of ad hoc tests for freeze-dryer qualification) and to controlled leakage (that can be measured by means of a mass-flow meter) and drying time. Figure 5.2 shows the values of the exergy losses in the primary drying stage considering the contribution in the drying chamber for vapor condensing and vacuum pumping (given as percentage of the total exergy loss), for different values of chamber pressure. Here, the freeze-drying of coffee extract is considered as this product is the most common freeze-dried liquid in food industry, due to the fact that the process preserves the coffee flavor to a great extent. It emerges that, when the value of the shelf temperature is low, then most of the exergy losses occur in the condenser and the contribution of the other two sources depends on the chamber pressure. However, when the value of the shelf temperature is increased, the contribution of primary drying in the chamber becomes much more important, especially at lower values of chamber pressure, while the exergy losses of the vacuum pump become almost negligible. Figure 5.3 shows the absolute values of the principal exergy losses in the primary drying step as a function of chamber pressure and heating shelf temperature. It appears that, while the shelf temperature seems to have no effect on the exergy losses in the condenser (this is due to the different contributions of the phenomena occurring in the condenser on the exergy losses), its effect is remarkable on the drying chamber losses. In particular, the higher the temperature of the shelf, the higher are the exergy losses. This is due to the fact that a higher shelf temperature causes the temperature gradient in the dried layer to be larger.

5.2 Exergetic Analysis (and Optimization) of the Freeze-Drying Process

(b)

(a) 200

Exloss, kJ/kg

Exloss, kJ/kg

150

100

137

350

300

250

50

0

200 5

10 15 chamber pressure, Pa

20

5

10 15 chamber pressure, Pa

Fig. 5.3 Values of the exergy loss for the primary drying step in (a) the drying chamber and (b) for vapor condensing as a function of chamber pressure and heating shelf temperature (solid line: 30  C; symbols: 0  C).

The chamber pressure affects both exergy losses in the drying chamber and in the condenser, but the effects are different: while an increasing chamber pressure decreases the exergy losses in the chamber, it increases the exergy losses in the condenser. The chamber pressure affects both heat transfer and mass transfer in the dried layer of the material: the temperature gradient can decrease due to the heat transfer enhancement, and this may reduce the exergy losses. On the other hand, when the chamber pressure increases, the pressure gradient – and thus the exergy losses due to mass transfer – are increased. In a heat transfer-controlled process, heat transfer in the product plays a more important role in the exergy losses than mass transfer in the material does, and so the exergy losses decrease when the chamber pressure is increased. In a mass transfer-controlled process, the effect of chamber pressure would be the opposite. As the process investigated is under heat transfer control, exergy losses decrease when increasing the chamber pressure. With respect to the condenser, when the chamber pressure is increased so too is the sublimation temperature, and this is responsible for the increase in exergy losses. The effect of the operating conditions on the cumulative exergy losses of the process is shown in Fig. 5.4. It appears that the exergy losses can be minimized by working at a high shelf temperature and a low chamber pressure. Previous results should be considered when the design space method is used to develop a freeze-drying cycle. Figure 5.5 shows the design space for coffee extract processed on trays in a pilot-scale freeze-dryer. The approach of Fissore et al. (2011a) has been used to calculate the design space. It appears that the design space is affected by the dried layer thickness as it changes the resistance to vapor flow. Consequently, a couple of values of Tshelf and Pc that belongs to the design space at the beginning of primary drying may lie outside the design space when drying continues. If a simple cycle is desired, where the values of the operating conditions are not modified during the primary drying, then it is necessary to consider the

20

138

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

700

700

Exloss, kJ/kg

(b) 800

600 500

600 500 400

400

300

300 -30

-20 -10 shelf temperature, °C

0

5

10 15 chamber pressure, Pa

20

Fig. 5.4 Values of the total exergy loss during the primary drying step as a function of (a) shelf temperature (solid line: 5 Pa; symbols: 20 Pa) and (b) chamber pressure (solid line: 30  C; symbols: 0  C).

curve of the design space calculated for a value of dried layer thickness approaching the total product thickness (Fig. 5.5a). Obviously, there are many values of operating conditions that make it possible to fulfill a certain constraint about the maximum product temperature. In order to choose among these, it might be worthwhile considering the values of the sublimation flux, with the goal of reducing the duration of primary drying. Figure 5.5b shows values of the sublimation flux in a certain time instant during primary drying (the trend of these curves is the same throughout the primary

(a)

(b)

0

0.6

shelf temperature, °C

Exloss, kJ/kg

(a) 800

-5 -10

0.5

0.01

Ld/L0 0.5

0.4

-15

0.99 -1

-20

0.25

-2

jw, kg h m

-25 2

4

6 8 10 12 14 16 18 20 chamber pressure, Pa

Fig. 5.5 (a) Design space for the primary drying stage of a coffee extract freeze-drying process calculated for different values of dried cake thickness; (b) Values of the sublimation flux as a function of the operating conditions when Ld/

2

4

6

8 10 12 14 16 18 20 chamber pressure, Pa

L0 ¼ 0.5 (dashed line identifies the design space). The product is loaded on the shelf using a metallic tray; the thickness of the frozen product is equal to 12.5 mm and the solute mass fraction is 25%.

5.3 Process Intensification in Vacuum Freeze-Drying of Liquids

Fig. 5.6 Values of the exergy yield as a function of the operating conditions (dotted curves). Solid line identifies the design space when Ld/L0 ¼ 0.5.

drying stage): the optimal operating conditions (i.e., those that maximize the sublimation flux) correspond to low values of Pc (e.g., 8 Pa in this case) and high values of Tshelf (0  C in this case). This information about exergy losses can be used to calculate an exergy yield (dividing the exergy losses by the exergy input), and these data can be added to the design space of the product. An example of these calculations is shown in Fig. 5.6 for the case study of the coffee extract freeze-drying process. It emerges that the values of Tshelf and Pc that allow the maximization of exergy yield are, in this case, those corresponding to high values of chamber pressure and low values of shelf temperature. Hence, when designing the freeze-drying cycle a compromise must be achieved, as the goals are to minimize the duration of primary drying (and, thus, to maximize the sublimation flux) and to maximize the exergy yield (i.e., to minimize the exergy losses).

5.3 Process Intensification in Vacuum Freeze-Drying of Liquids

In the previous section, exergy analysis was applied to the freeze-drying process in order to identify the optimal values for processing conditions to be used during primary drying. Here, the problem of process intensification is addressed from a different point of view – that is, to reduce the cycle time by modification of product structure. Fundamentally, if the material being dried is originally a liquid, then all of those methods that can modify the habit and size of solvent crystals can dramatically alter the rate of sublimation, and hence the time necessary for the primary drying stage. The sublimation rate can be expressed as:

139

140

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

jw ¼

pw;i

pw;c Rp

;

ð5:10Þ

where Rp is an overall resistance to vapor flow, which is the sum of the product resistance, the vial and stopper resistances, and the drying chamber resistance. It has been shown that the product resistance is predominant as it typically contributes more than 90% of the overall resistance (Pikal, 1990). As will be discussed, the above-described resistance is dramatically influenced by freezing conditions and product formulation (see also Chapter 3 in Volume 4, Modern Drying Technology). As shown in Eq. 5.10, in order to increase the sublimation rate, lyophilization professionals can act not only on the product resistance but also on the pressure difference. The quantity pw,i in Eq. 5.10 is the vapor pressure at the interface of sublimation, which is a function of product temperature. Therefore, in order to maximize the driving force for mass transfer, freeze-drying cycles are designed in order that product temperature, hence pw,i, is as close as possible to its limit value. A detailed review and comparison between the various methods proposed to date to achieve the above objective is given by Pisano et al. (2013). Rather, in this section attention is focused on the methods proposed to promote the formation of large ice crystals, such as the use of cosolvents or the regulation of nucleation temperature, as these may have a profound impact on product resistance. 5.3.1 Regulation of Nucleation Temperature During Freezing

During freezing, a solvent is separated from the drug and excipients (or nutrients in the case of foodstuffs) by ice formation; the system is then cooled further to allow the transformation of all other components in a frozen state. Various factors affect the freezing characteristics of the solution prior to drying, the most important being the filling volume and the type of container (Hottot et al., 2007), the time–temperature profile (Rey and Bastien, 1961; Searles et al., 2001b), and the supercooling and ice nucleation temperature (Searles et al., 2001a). Control of the above critical variables is essential not only to achieve optimized cycles and to produce uniform batches, but also to simplify cycle scale-up. Lyophilization professionals often focus on primary and secondary drying, but put very little attention to freezing. Yet, freezing is a crucial step for the freezedrying process as it can impact on both the product characteristics and process performance (Patapoff and Overcashier, 2002; Kasper and Friess, 2011). The product characteristics can be modified only if the nucleation process can be controlled. The control of ice nucleation is further desirable as it forces the nucleation of a batch of vials to occur within narrow temperature and time ranges, and thus enables a more uniform drying behavior to be obtained. In fact, the random nature of the nucleation process causes the nucleation temperature to be distributed around an average value and, as a consequence, vials of the same batch can have different product structures and hence show different behaviors during

5.3 Process Intensification in Vacuum Freeze-Drying of Liquids

primary and secondary drying. The stochastic nature of nucleation also makes it difficult to control the temperature within a desired range of supercooling. During the past few decades, however, a variety of methods have been proposed to induce ice nucleation. Inada et al. (2001) suggested the use of ultrasonic vibrations to induce a phase transition in a metastable material. Although the precise mechanism of this phenomenon is unclear (Zhang et al., 2001; Saclier et al., 2010a), various authors have shown that the technique can be used for the effective control of the size of ice crystals and to produce uniform freezing (Saclier et al., 2010b). Unfortunately, however, the ice structure is not uniform within individual vials, and ice crystals tend to be larger at the top surface (Nakagawa et al., 2006). In the “ice fog” technique proposed by Rambhatla et al. (2004), a flow of nitrogen at low temperature and high pressure is introduced inside the drying chamber, when the product has reached the desired value of temperature for nucleation. This operation produces ice crystals that are suspended in the chamber gas: they enter the vials and act as nucleation agents for ice formation. In this way, the nucleation of ice can be induced at a controlled temperature. Unfortunately, the ice structure is usually not uniform in all vials as the time required for nucleation is typically too long. Hence, in order to allow a more rapid and uniform freezing, Patel et al. (2009) proposed that the pressure in the chamber should be reduced when nitrogen is introduced. Currently, the use of this technique is still limited to laboratory-scale freeze-dryers, as large-scale production units may require the use of convection devices to obtain a uniform distribution of the “ice fog” in the drying chamber. Bursac et al. (2009), and more recently Konstantinidis et al. (2011), presented an effective method to control ice nucleation, namely the depressurization method. In this case, vials are loaded onto the freeze-dryer and pressure is increased above atmospheric level by 1.7–2 bar by the bleeding of inert gases at high pressure. The solution is then cooled to a desired temperature and pressure is decreased to induce ice nucleation simultaneously in all vials. Konstantinidis et al. (2011) showed that this method can be applied to both laboratory-scale and productionscale freeze-dryers. If the freeze-dryer can withstand pressures higher than atmospheric value, then the method can easily be implemented; otherwise, it will be necessary to modify the equipment to support overpressure, and this can be expensive. In this situation the application of a different method for ice nucleation control is suggested. “Vacuum-induced surface freezing” is another method of inducing ice nucleation, which can be applied to pre-existing apparatus without any modifications and, thus, may represent a good alternative to the above method. The vials are first loaded onto pre-cooled shelves, after which the pressure is reduced to almost 1 mbar in order to induce water evaporation. As the evaporation is endothermic, this reduces the temperature of the solution at the top surface and promotes its freezing (Kramer et al., 2002). Once nucleation has been induced, the product temperature must be reduced below the melting point and atmospheric pressure restored rapidly in order to prevent the product from puffing up (because of boiling); otherwise, the ice formed at the vial top will melt again (Liu et al., 2005).

141

142

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

Fig. 5.7 Scanning electron microscopy images of mannitol (metallized samples) that have been freeze-dried from a 5% (w/w) water solution with (a) spontaneous nucleation and (b) forced nucleation. Samples processed in research laboratories of Telstar S.A., Terrassa, Spain.

Figure 5.7 compares the internal structure of freeze-dried products obtained by spontaneous nucleation and by forced nucleation. As expected, the latter sample has larger pores than the sample obtained by spontaneous nucleation. As a result, samples obtained by forced nucleation are characterized by a low product resistance to vapor flow and, therefore, by faster drying rates (see Eq. 5.10). Visual observation confirmed that all vials were nucleated within a short time after pressure decrease and, as a result, they were nucleated at very nearly the same temperature (within 0.1  C) (see Fig. 5.8). The above conditions should guarantee uniform freezing, and this hypothesis was confirmed experimentally (see Fig. 5.9b). In fact, using forced nucleation a

Fig. 5.8 Product temperature profiles during cooling for (a) spontaneous nucleation and (b) forced nucleation of 5% mannitol. Data courtesy of M. Galan and R. Bullich, Telstar S.A., Terrassa, Spain.

5.3 Process Intensification in Vacuum Freeze-Drying of Liquids

Fig. 5.9 Comparison of (a) product temperature as estimated by pressure rise test technique and (b) chamber pressure profiles during primary drying for (, solid line) uncontrolled and ( , dashed line) controlled

freeze-drying cycles of 10% mannitol. The vertical line indicates the endpoint time for ice sublimation. Data courtesy of M. Galan and R. Bullich, Telstar S.A., Terrassa, Spain.

sharp change was observed in the Pirani–Baratron pressure ratio curve close to the completion of ice sublimation. In contrast, the cycle with spontaneous sublimation showed a smooth change of the slope for this curve. It should be noted that, although the freezing is uniform, the internal structure of individual vials is heterogeneous, with a compact layer being present at the product top while the underneath part was characterized by much larger pores. A similar structure was observed for samples nucleated by ultrasonic vibrations (Barresi and Fissore, 2011). Figure 5.9 also shows that the drying time in case of forced nucleation was 10 h shorter than the time required when spontaneous nucleation was used, which corresponds to a 40% reduction in drying time. The use of forced nucleation produces large ice crystals (see Fig. 5.7), which leads to a lower product resistance to vapor flow and thus a high rate of sublimation, although drying was carried out at almost the same product temperature (see Fig. 5.9a). As shown in Section 5.5.2, ultrasound technology can also be used to enhance mass transfer during drying. A number of other methods have also been proposed to modify the freezing of liquids in vials. These are aimed at obtaining a uniform freezing, without any concerns regarding the duration of the subsequent drying stages. Although these techniques cannot be used to modify the habit and size of ice crystals, they can induce nucleation in all vials of the batch within a limited temperature range, and this leads to uniform freezing. Searles et al. (2001a) investigated the use of various vial pretreatments (such as roughening, scoring and scratching) as a means of lowering the degree of supercooling, while Petzold and Aguilera (2009) used insoluble impurities as nucleating agents. Unfortunately, none of the above methods allows a precise control of the ice nucleation and, in turn, of the final structure of the freeze-dried product.

143

144

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

5.3.2 Use of Organic Solvents and Cosolvents

The use of a strictly organic or organic/water system is beneficial to both product quality and process optimization. The potential advantages and disadvantages of using organic solvents and cosolvents (mixed with water, which until now has been the most commonly used solvent) in freeze-drying have been widely discussed by Teagarden and Baker (2002). The main advantage of solvent use, in addition to increasing the solubility of the product, is the increase in the rate of sublimation and hence a decrease in drying time. Before a specific solvent is used in the manufacture of a parenteral product, however, lyophilization professionals must carefully weigh up the associated advantages and disadvantages. The use of organic liquids, either as solvent or cosolvent, produces larger ice crystals, and thus increases the average diameter of the pores created during ice sublimation. As a result, the product resistance to vapor flow is decreased and, if the product temperature does not change, the rate of sublimation will be increased (see Eq. 5.10). Kasraian and DeLuca (1995a) showed, for example, that the product resistance for lactose and sucrose formulations significantly decreased, allowing the sublimation rate to increase, when the product was freeze-dried using a 5% (v/v) aqueous solution of tert-butyl alcohol, instead of pure water. In addition to the reduction in product resistance, most cosolvents used for freeze-drying applications increase the rate of sublimation because they have a higher vapor pressure than water, and hence increase the driving force for mass transfer (Wittaya-Areekul, 1999). A further reduction in energy consumption also results from the fact that the sublimation enthalpy of organic solvents is smaller than that of frozen water. Figure 5.10 compares the product structure obtained by the freeze-drying of 5% (w/w) sucrose solution with pure water and with a 5% (w/w) aqueous solution of tert-butanol. These images show clearly that a different structure (with needle-like crystals) is obtained when 5% tert-butanol is used compared to water. Figure 5.11 compares two freeze-drying cycles for a 5% (w/w) aqueous solution of sucrose; the former was carried out using pure water as solvent,

Fig. 5.10 Scanning electron microscopy images of sucrose samples (metallized) that have been freeze-dried from a 5% (w/w) water solution with (a) pure water and (b) 5% (w/w) aqueous solution of tert-butyl alcohol.

5.3 Process Intensification in Vacuum Freeze-Drying of Liquids

Fig. 5.11 Example of freeze-drying cycles for a 5% (w/w) sucrose solution with (solid line) pure water and (dashed line) 5% (w/w) tert-butyl alcohol aqueous solution. Primary drying was carried out at Tshelf ¼ –20  C and Pc ¼ 10 Pa.

Evolution of (a) product temperature at vial bottom as detected by thermocouples and (b) water concentration in chamber gas as measured by a laser detector. The vertical line indicates the endpoint time for ice sublimation.

while in the latter case a 5% (w/w) aqueous solution of tert-butyl alcohol was used. The same processing conditions (i.e., temperature of the heat transfer fluid and chamber pressure) were used for both cycles. The drying rate was significantly increased when tert-butyl alcohol was used as cosolvent, and consequently the drying time was also decreased by approximately 50%. In fact, as shown in Fig. 5.11b, the water concentration in the drying chamber began to decrease when ice sublimation was complete – that is, after 12.5 h for the cycle with pure water and after 23.5 h for 5% (w/w) tert-butyl alcohol solution. The increase in drying rate was most likely due to the formation of needle-shaped ice crystals (see Fig. 5.10b), which dramatically lowered the product resistance to vapor flow. It should also be noted that product temperature was slightly lower than the value observed when sucrose was freeze-dried in the presence of tert-butyl alcohol (see Fig. 5.11a). Similar results were shown by Kasraian and DeLuca (1995b). It follows that the product could be successfully freeze-dried in the presence of tert-butyl alcohol by using a more aggressive heating policy than that used for aqueous solutions. As a result, a further reduction could be obtained for the drying time when a cosolvent system is used. Of course, the above considerations are valid only if the presence of tert-butyl alcohol does not lower the collapse temperature of the formulation. For sucrosebased formulations, Kasraian and DeLuca (1995a) showed that the addition of tertbutyl alcohol does not alter the collapse temperature; however, there are cases where the presence of an organic solvent together with water can reduce the collapse temperature of the formulation. As an example, Liu et al. (2005) showed that the collapse temperature of sulfobutylether-7-b-cyclodextrin formulation was reduced by 2  C when a 5% (w/w) tert-butyl alcohol aqueous solution was used.

145

146

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

Tesconi et al. (1999) proposed the use of organic solvents that are solid at room temperature as a means of freeze-drying pharmaceutical products, without using conventional freeze-drying equipment. These authors investigated a variety of compounds in terms of their ability to solubilize hydrophobic drugs, to increase the chemical stability of both the pre-dried bulk solution and the dried product, to be readily removed by vacuum drying, and to produce elegant cakes which can easily be reconstituted. Ultimately, an eutectic mixture of chlorobutanol hemihydrate and dimethyl sulfone was found to be an optimum candidate, and the use of this solvent made it possible to carry out the freeze-drying process for a drug model without freezing. Furthermore, the process required only modest heating under vacuum. Unfortunately, this method did not find real application in industry due to various limitations; for example, solutions must be kept in the liquid phase during filtration and filling operations, while a high degree of flammability can cause problems of operator safety. Nonetheless, the application of this method to production units would represent a further step towards process intensification for freeze-drying, though a more accurate study to determine its technical feasibility may be necessary.

5.4 Atmospheric Freeze-Drying

Vacuum freeze-drying produces dried products that retain almost all of their original characteristics, such as color, flavor, and taste. Moreover, the high specific surface area generally allows an easy and rapid rehydration. The main drawback of the process relates to the cost of the operation, as fixed costs can be high due to vacuum requirements while energy costs can be significantly higher compared to other drying processes (the specific moisture extraction rate in a vacuum freezedrying process is in the range of 0.4 kg of water per kWh, as reported by Claussen et al., 2007c). In order to reduce the energy consumption of the process, and thus to improve its sustainability, atmospheric freeze-drying (i.e., drying with cold air or nitrogen at normal pressure) has been proposed, in which case it would be possible to achieve a specific moisture extraction rate ranging from 1.5 to 4.6 kg of water per kWh (Claussen et al., 2007c). In fact, Wolff and Gibert (1990a) reported that an up to 35% saving of energy costs was achievable by using atmospheric freeze-drying instead of vacuum freeze-drying for potato slices. Meryman (1959) first showed that the pressure gradient in the dried cake, rather than the absolute pressure in the system, determined the vapor flow from the interface of sublimation to the drying chamber. This permits a product to be freezedried at atmospheric pressure, provided that a vapor pressure gradient is maintained in the dried product. The atmospheric freeze-drying of various products has been investigated previously, examples being fish products and starter feeds for the fish industry, strawberries and potatoes and lactic acid bacteria, among many others (Alves-Filho, 2002; Alves-Filho et al., 2004; Alves-Filho and Roos, 2006; Alves-Filho and Eikevik, 2008). Attention must be paid, however, to

5.4 Atmospheric Freeze-Drying

avoid any possible confusion between atmospheric freeze-drying and the drying of a frozen product at atmospheric pressure. In fact, for some reported applications of atmospheric freeze-drying the processing conditions used are so aggressive that they instill a doubt that the temperature of the initially frozen product will increase up to the melting temperature of ice. But, if this is the case, ice will melt during the process, the product to be dried will no longer be frozen, and the term “freezedrying” would not be appropriate to describe such a drying process. In fact, an atmospheric freeze-drying process consists of the convective drying of a product maintained at a temperature below its freezing point: typically, the temperature ranges from 10 to 3  C, as this appears to be a good compromise between final product quality and costs, as lower temperatures reduce the possibility of removing moisture with air and thus increase the duration of the process. The possibility of reducing drying time by using a variable process temperature has also been investigated. Indeed, it has been noted that an increasing temperature can inhibit solute transport within the product and enhance the heat transport in the dried product, although in some cases a lower temperature and a longer residence time can improve the product quality (Strømmen et al., 2005). The majority of published reports have dealt with atmospheric freeze-drying in fluidized-bed dryers (Wolff and Gibert, 1990a; Lombra~ na and Villaran, 1996) and also in spray freeze-dryers (Mumenthaler and Leuenberger, 1991; Leuenberger, 2002; Rogers et al., 2003; Leuenberger et al., 2006). When atmospheric freezedrying is carried out in a fluidized bed, we can take advantage of the high values of the heat and mass transfer coefficients; notably, the product must be frozen and granulated before drying. One drawback of this process, however, is represented by the size reduction caused by mechanical cracking, and as an alternative it is possible to carry out the process in a tunnel dryer, even if the heat and mass transfer is worse (Claussen et al., 2007b). Spray freeze-drying appears to be a valuable alternative to produce a free-flowing powder, with a high surface area, a porous end product, and good instant characteristics, with enhanced solubility and a uniform and ultrafine particle size. Spray freeze-drying into liquids, gases (e.g., a refrigerated air stream), and into gases over a fluidized bed have been reported. Recent investigations have been focused on the inclusion of a heat pump in the  drying system (Indergard et al., 2001; Alves-Filho, 2002; Chua et al., 2002), as this not only allows significant energy savings but also permits a careful control of the drying temperature and air humidity. A comprehensive treatment of heat pump drying was provided in Chapter 4 of Volume 4, Modern Drying Technology. The kinetics of atmospheric freeze-drying can be investigated by means of mathematical modeling, and the receding ice front model is perhaps the most common type of model used for this purpose. In this case, the product is considered to be divided into two layers, namely the frozen inner core and the dried outer layer. A receding interface is assumed to exist between the frozen and the dried layer. Subsequently, heat transferred from the surface of the product to the interface is used for sublimation of the ice, which diffuses from the interface to the surface and to the drying chamber at atmospheric pressure. In the case of a

147

148

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

Tair

pw,air

drying chamber

dried product

Qɺ

Ti

wet product

Mɺ

pw,i

interface of sublimation

Fig. 5.12 Sketch of an atmospheric freeze-drying process.

layered material with a planar sublimation front (Fig. 5.12), the transport of heat and mass through the dried layer can be modeled by means of the following equations (Wolff and Gibert, 1990b; Claussen et al., 2007a): A
ðT air T i Þ; 1 Ld þ aext ld   A _ ¼ 1
M
pw;i pw;air ; 1 L ~ d RT i þ bext D

Q_ ¼

ð5:11Þ

ð5:12Þ

where aext and bext are, respectively, the heat and the mass transfer coefficient from the external surface of the product to the gaseous stream, and D is the diffusion coefficient. As the heat flux arriving at the interface between the frozen and the dried product is used for ice sublimation, then: _ _ ¼ Q : M Dhs

ð5:13Þ

Equations 5.11–5.13 can be used to calculate the interface temperature, once the operating conditions (temperature and humidity of the air stream) have been selected, and the values of the heat and mass transfer coefficients have been determined (see, among others, Boeh-Ocansej, 1988; Krokida et al., 2002). It is then possible to calculate the sublimation flux and, finally, the variation of the dried layer thickness with time: Ard ðX 0

X fin Þ

dLd _ ¼ M: dt

ð5:14Þ

Figure 5.13 shows the effect of air velocity, temperature, and relative humidity on drying time for the atmospheric freeze-drying of apple samples (Malus domestica) in a tunnel dryer. It appears that, while the drying time is only slightly reduced when the air velocity is strongly increased, the effect of the air temperature and of the relative humidity is much more relevant, and the drying time can be greatly reduced by increasing air temperature and decreasing the relative humidity.

5.4 Atmospheric Freeze-Drying

Fig. 5.13 Effect of (a) air velocity (air relative humidity ¼ 30%) and (b) air relative humidity (air velocity ¼ 3 m s 1) on the drying time of apple (Malus domestica) samples (product

thickness ¼ 5 mm, D ¼ 5  10 5 m2 s 1) in a tunnel dryer for different values of air temperature (& ¼ 0  C, ¼ 5  C, ~ ¼ 10  C, ^ ¼ 15  C).



In case the interface of sublimation is not planar, as for particles in fluidized-bed drying, the heat and mass balance equations must be modified to account for this (Wolff and Gibert, 1990b; Di Matteo et al., 2003). When comparing the calculated drying time for a tunnel freeze-dryer and a fluidized bed, as shown in Fig. 5.14 for the atmospheric freeze-drying of apple samples (M. domestica), it appears that the drying time is highly reduced in the second type of equipment (due mainly to the higher heat and mass transfer coefficients). Lower values of relative humidity and higher values of temperature result in this case in a decreasing drying time.

(a)

(b)

60

50 drying time, h

drying time, h

50 40 30 20 10 0

60

40 30 20 10

1%

10% 30% air relative humidity

Fig. 5.14 Comparison between the drying time calculated for a tunnel dryer (&, air velocity ¼ 3 m s 1, product thickness ¼ 5 mm) and in a fluidized-bed dryer ( ) for apple

0

1%

10% 30% air relative humidity

(Malus domestica) samples as a function of air relative humidity and temperature (a) Tair ¼ 5  C; (b) Tair ¼ 10  C.

149

150

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

5.5 Use of Combined Technologies for Drying Heat-Sensitive Products

As discussed above, atmospheric freeze-drying produces high-quality products but still entails long drying times. In the past, various research groups have striven to decrease energy consumption in atmospheric freeze-drying, for example, by new technologies that use alternative forms and sources of energy for processing. These technologies can be applied to either enhance heat transfer between product and heat source, such as microwave, radiofrequency and infrared radiation, or simply to intensify the rate of dehydration without increasing the amount of heat supplied to the product, for example, by using high-intensity sonic and ultrasonic waves. The advantages and disadvantages of the above techniques are discussed in the following subsections, with emphasis placed on their application to production equipment. 5.5.1 Microwave-Assisted Drying

When microwaves are used as an energy source, heat is generated within the product being dried. This energy is transferred by molecular excitation (i.e., rotation of molecules with permanent dipole moment and ionic motion) caused by the alternating electromagnetic field and, as a result, heat is supplied to the entire mass of the wet product in a rapid and uniform fashion. The magnitude of the above effect is product-specific as it depends on the dielectric properties of materials being heated, and therefore its application in drying must be evaluated for each individual product. At this point, the use of microwaves is first presented as an alternative energy source to be combined with various drying methods, such as hot-air drying, vacuum-drying, and heat-pump drying. Attention is then focused on the application of microwave energy to the freeze-drying process, evidencing its advantages and current limitations. Tulasidas et al. (1995) showed that the application of microwave energy to drying processes constitutes an efficient method of transferring heat to products for water removal. As the use of microwaves alone cannot complete a drying process, this technique commonly is combined with forced air streams and vacuum conditions (Chou and Chua, 2001). A number of studies have shown the benefits of using microwave-assisted drying for vegetables; a list of recent applications is given in Tab. 5.1, where microwave technology is used to support both hot-air drying and vacuum-drying. In particular, various groups have shown that microwave-assisted vacuum drying can produce high-quality dried products, which often can be rated as equal to freeze-dried products. Kwok et al. (2004) compared various drying techniques (including freeze-drying) for preserving the antioxidant properties of Saskatoon berries. The relatively low consumption of energy, together with a significant reduction of drying time, observed for microwave-assisted technology has led to it being competitive with

5.5 Use of Combined Technologies for Drying Heat-Sensitive Products Tab. 5.1

Examples of microwave-assisted drying for foodstuffs.

Material

Microwave-assisted technology

Reference

Banana Carrot

Vacuum drying Convective drying Vacuum drying

Drouzas and Schubert, 1996 Wang and Xi, 2005 Lin et al., 1998 Regier et al., 2005 Orsat et al., 2007

Cabbage Cranberry Garlic

Ginseng Mushroom Pea Raisin Saskatoon berry Strawberry

Convective/vacuum drying/ Vacuum drying Vacuum drying Convective drying Convective/vacuum drying Vacuum drying Vacuum drying Convective drying Atmospheric freeze-drying Convective drying Vacuum drying Vacuum drying

Xu et al., 2004 Yongsawatdigul and Gunasekaran, 1996a, 1996b Sharma and Prasad, 2006 Sunjka et al., 2004 Cui et al., 2003 Popovich et al., 2005 Torringa et al., 2001 Eikevik et al., 2012 Tulasidas et al., 1995 Kwok et al., 2004 Venkatachalapathy and Raghavan, 1999 B€ohm et al., 2006

freeze-drying, even if the former can entail a potential loss of activity. The best retention was obtained by freeze-drying followed by microwave-assisted vacuum drying. Similar results were obtained for the drying of garlic cloves by Cui et al. (2003). Sharma and Prasad (2006) also observed that microwave-assisted drying allowed an 80% reduction in drying time when compared to hot-air drying, while an even larger reduction can be obtained if compared to freeze-drying process. When Lin et al. (1998) compared freeze-drying, hot-air drying, and microwaveassisted vacuum drying for carrot slices, they observed that both freeze-drying and microwave-assisted vacuum drying can produce a high-quality dried product. Later, Regier et al. (2005) showed that microwave-assisted vacuum drying could produce dried carrots with the highest carotenoid retention when compared to both freezedrying and hot-air drying. Popovich et al. (2005) compared the type and content of ginsenosides contained in ginseng that had been dried using various methods, and found freeze-dried ginseng to have the highest contents of Rg1 and Re ginsenosides. In contrast, microwave-assisted vacuum drying produced a dried product with the highest contents of Rb1 and Rd ginsenosides. Finally, Venkatachalapathy and Raghavan (1999) observed that a combination of microwave-assisted vacuum drying and osmotic dehydration (as pretreatment), when applied to the drying of strawberries, created a high-quality dried

151

152

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

product which was comparable with that obtained by freeze-drying, but in a much shorter time. Various groups (e.g., Ma and Peltre, 1975; Ang et al., 1977) have shown that microwave technology can also be applied to the freeze-drying of foodstuffs, with a dramatic increase in sublimation rate. Eikevik et al. (2012) showed that the drying time for green peas could be halved when atmospheric freeze-drying was combined with microwave radiation. However, the rate of drying and product quality were dramatically affected by the process conditions, for example, fixed versus fluid-bed conditions. Arsem and Ma (1990) also investigated the possibility of using microwave combined with radiative heat transfer as a means of further increasing the drying rate. The above studies demonstrate that the microwave-assisted vacuum drying of foodstuffs is an effective alternative to freeze-drying, as it produces high-quality dried products in a shorter time and entails a low energy consumption. By contrast, the application of microwave technology to freeze-drying can be problematic, as corona discharge and nonuniform heating can occur and produce high-energy zones (Pester et al., 1977). This phenomenon represents a strong limitation for the drying of pharmaceutical products, which must be dried at very low temperatures; moreover, even small and brief variations in drying temperature can cause damage to the active pharmaceutical ingredients. However, it has been shown that microwaves can be applied successfully to the freeze-drying of food liquids, such as milk foam (Sochanski et al., 1990), as their critical quality attributes are less stringent than those required by the pharmaceutical industry. In conclusion, the use of microwaves as an additional heat source during freeze-drying may be beneficial to process performance, as it can enhance and make more uniform the heat transfer to the product. However, in order to promote microwave application as a potential and effective technology for intensifying the freeze-drying process, further studies are required to improve the control of the energy supplied by microwaves. 5.5.2 Ultrasound-Assisted Drying

The use of ultrasound to intensify drying processes was first investigated by Fairbank (1975), although others (Gallego-Juarez et al., 1999; Mulet et al., 2003) have more recently applied direct contact ultrasound to the drying of vegetables and shown the technique to allow dramatic reductions in drying times (cf. Chapter 8 of this Volume). Unlike previous technologies, the use of ultrasound does not constitute an alternative source of heat for processing (increase in product temperature is significant only if the material is exposed for a long time to ultrasound) and, therefore, does not introduce concerns about product overheating or thawing. Various groups have shown that power ultrasound increases the rate of water removal based on mechanical actions on both gas–solid interfaces (Bantle and Eikevik, 2011) and product structure (Garcia-Perez et al., 2012b). The application of high-intensity ultrasonic energy produces rapid contractions and

5.5 Use of Combined Technologies for Drying Heat-Sensitive Products

expansions through the product structure, which in turn promotes the formation of microscopic channels for water motion. High-intensity ultrasonic waves also induce cavitation, which enhances the removal of water that is strongly linked to the solid product (Muralidhara et al., 1985). The use of high-intensity ultrasound technology, combined with atmospheric freeze-drying, has recently been investigated by Bantle and Eikevik (2011) for the drying of peas in fluidized beds. In particular, the use of ultrasound was shown to increase the effective diffusivity by up to 15%, and that such an increase was higher at low temperatures (between 6  C and 0  C) but was marginal at room temperature. A similar result was obtained by Garcia-Perez et al. (2012a), who showed that the use of ultrasound in the convective drying (at atmospheric pressure and low temperatures) of vegetables could reduce the drying time by up to 70%. Such a reduction in drying time was associated with a dramatic increase in vapor diffusivity of up to 400%. Ultrasound technology was also shown to speed up the drying process when an organic solvent (i.e., ethanol) was used in place of water. Figure 5.15 compares the structure of albedo cells as obtained from orange peel in the case of a product dried with or without ultrasound technology. It is well known that drying produces high stress on vegetable cells and can lead to a partial collapse of the material structure (Garcia-Perez et al., 2012b). This phenomenon allows a reduction in the product resistance to vapor flow and, hence, speeds up the dehydration process. A similar behavior was also observed when drying was combined with ultrasound technology although, as shown in Fig. 5.15, in this case the product damage that occurred during drying was more marked. Because of this phenomenon, the product resistance to vapor flow can be further reduced and the rate of water removal increased. The results of the above-described studies have demonstrated that ultrasound technology can be effectively coupled with low-temperature drying to speed up water removal and, hence, lead to cost-effective processes. Nevertheless, the use of this technique in production units remains limited, most likely because of practical difficulties in satisfying industrial-scale requirements.

Fig. 5.15 Scanning electron microscopy images of albedo cells obtained from orange peel in case of (a) hot-air drying and (b) ultrasound-assisted hot-air drying. Illustration courtesy of Prof. Antonio Mulet and Juan A. Carcel, Universidad Politecnica de Valencia, Spain.

153

154

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

5.6 Continuous Freeze-Drying

During the 1960s, the food industry was forced to increase its production to cope with increased requests from the market for freeze-dried foodstuffs, such as milk and coffee. Surprisingly, the above-described objective was achieved not by improving preexistent equipment but, quite oppositely, by abandoning it in favor of new technologies. The food industry, in fact, revamped its equipment and moved away from batch to semi-continuous and, as a last resort, to continuous operations. A first attempt at continuous operation was developed during the 1960s by Leybold and Oetjen, who created the first prototype continuous freeze-dryer for milk products (Oetjen and Haseley, 2004; Rey, 2010). The operation encompassed several steps. The product was first frozen to produce granules that were then loaded onto trays which hang on a monorail. In this way, the product could be moved through the freeze-drying tunnel between temperature-controlled shelves, while the tunnel was divided into several sections, separated by sliding gates. This approach allowed a dramatic reduction in the total cycle time, namely hours versus days. In an alternative configuration, the frozen products were loaded tray-by-tray through a small side lock. The above technology was used to freeze-dry not only coffee and milk but also fish fillets, meat, and vegetables. Although the above-described technology allowed a dramatic reduction in energy consumption and processing costs, as well as a desired increase in the rate of production, the process was semi-continuous and could not satisfy the increased requests of the international market for instant coffee. In fact, this request motivated the food industry to develop a fully continuous line, which was realized in the mid-1970s. Continuous plants for freeze-drying are highly sophisticated and were designed by the food industry itself, which kept the development confidential. In one of the most advanced configurations, granules of frozen product were continuously deposited onto long vibrating trays through a rotating lock. The product being dried moved on trays as a fluidized bed, such that the granules were forced to move into parallel channels through vertical ribs. During the plant design, great attention was paid to the problem of dust formation; for this purpose, various solutions were investigated to prevent granule-to-granule attrition, as well as attrition between the granules and tray walls. Although such attrition was limited, dust formation proved to be unavoidable, and semi-cylindrical screens were installed over the vibrating trays in order to trap any dust produced by mechanical shocks; this prevented the dust from being transported to the vacuum unit (i.e., condenser and vacuum pumps) via the vapor stream. In fact, the use of continuous freeze-drying led to an enormous increase in process efficiency to a point where the total cycle time was of the order of minutes (compared to a few hours for a semicontinuous plant, and days for batch units), throughput could reach tens of tons per day, and the dried products were of a constant and uniform quality. Although the freeze-drying process has been greatly intensified in the food industry by switching from batch to semi-continuous and finally to continuous plants, biological and pharmaceutical products are still largely produced in batch

5.7 Conclusions

units. Typically, the low efficiency of these processes was justified by the high value of the end products, but today this situation has changed as market competition continues to grow, mainly for those products which are no longer intellectually protected. In addition, manufacturers still face recurrent problem of heterogeneity in end product quality within a batch. Although various solutions were investigated and successfully applied in order to contain this latter problem, waste due to product heterogeneity remains high. Furthermore, validation tests proposed by regulatory agencies are increasingly stringent. Today, with multi-shelf freeze-dying cabinets capable of being loaded with up to 100 000 vials, it is unthinkable that vials in different positions can be dried in the same way, even if elaborate designs were proposed by the freeze-dryer manufacturers. One possible solution to this problem would be to switch to a semi-continuous process, as has the food industry during the past 50 years (Rey, 2010), whereby the freeze-drying process can be greatly intensified in terms of throughput, energy consumption, and product quality.

5.7 Conclusions

In freeze-drying, process intensification involves the development of new technologies that can reduce consumption costs and lead the way towards uniformity in end-product quality. The former objective can be achieved by reducing the drying time. As primary drying is the longest and the most energy-consuming stage, its optimization is typically the primary objective, while minor attention is dedicated to the optimization of secondary drying. Various solutions have been discussed in this chapter to achieve the above-described objectives. An example is the use of automatic control to optimize processing conditions; that is, to find in-line a combination of pressure and temperature that maximizes the rate of sublimation, while satisfying constraints posed by the equipment, as well as by product quality attributes (e.g., on the maximum value for product temperature). The control of freezing can also benefit process intensification, as this stage can impact the product structure and hence modify the product resistance to mass transfer. Various groups have also claimed that it can improve batch uniformity. In this chapter, the various methods proposed to date to regulate ice nucleation have been briefly described, and an example of application to a real case provided. An alternative approach to enhancing the drying rate is to use organic solvents, the presence of which leads to larger solvent crystals and hence a lower product resistance to vapor flow. In addition, organic solvents are typically characterized by high values of vapor pressure and so can increase the driving force for mass transfer. Both of these phenomena produce high rates of sublimation. Atmospheric freeze-drying represents a good alternative to vacuum freezedrying, as a means of increasing process efficiency. Further improvements can be obtained by combining atmospheric or vacuum freeze-drying with new technologies that can then be used to enhance heat transfer between the heat source and the

155

156

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries

product (microwave-assisted drying), or simply to intensify the rate of dehydration by altering the product structure (e.g., high intensity sonic and ultrasonic waves). The application of microwave technology to pharmaceutical products can be problematic, however, as the latter products must be dried at very low temperatures and the risk of product overheating is high. In contrast, practical difficulties to satisfy industrial-scale requirements limit the application of high-ultrasound techniques. A further step towards process intensification is provided by continuous plants, and such technologies have been used for food manufacture since the 1960s, allowing dramatic increases in throughput and product quality uniformity. In theory, continuous plants may also be developed for the manufacture of pharmaceuticals, although regulatory concerns may still impede this pathway.

Acknowledgments

The authors acknowledge Irene Oddone, Susanna Sanapo, Salvatore Genco and Alessandro Ando for their contributions to this chapter. The contributions of Miquel Galan and Robert Bullich (Telstar s.a., Terrassa, Spain), and of Prof. Antonio Mulet and Juan A. Carcel, (Universidad Politecnica de Valencia, Valencia, Spain) are also gratefully acknowledged.

Additional Notation Used in Chapter 5

cp Ex L Q Rp

specific heat capacity at constant pressure exergy layer thickness energy overall resistance to vapor flow

Subscripts and Superscripts

c cond cooling d des ext fin freeze, f

drying chamber condenser cooling drying, dry desublimation external final product after freezing, frozen

J kg J m J ms

1

1

K

1

References

h i ice in loss m out pd pump s shelf v 0

heat transfer interface of sublimation ice initial, input, inlet value loss mass transfer outlet primary drying pump sublimation heating shelf vapor initial, reference

References Alves-Filho, O., 2002. Combined innovative heat pump drying technologies and new cold extrusion techniques for production of instant foods. Drying Technol. 20(8): 1541–1557. Alves-Filho, O., Eikevik, T. M., 2008. Heat pump drying kinetics of Spanish cheese. Int. J. Food Eng. 4(6): article 6. DOI: 10.2202/ 1556-3758.1481. Alves-Filho, O., Roos, Y. H., 2006. Advances in multi-purpose drying operations with phase and state transitions. Drying Technol. 24(3): 383–396. Alves-Filho, O., Garcia-Pascual, P., Eikevik, T. M., Strømmen, I., 2004. Dehydration of green peas under atmospheric freeze-drying conditions. Proceedings of 14th International Drying Symposium, Sao Paulo, Brazil, August 22–25. Ang, T. K., Ford, J. D., Pei, D. C. T., 1977. Microwave freeze-drying of food: A theoretical investigation. Int. J. Heat Mass Transfer 20(5): 517–526. Arsem, H. B., Ma, Y. H., 1990. Simulation of a combined microwave and radiant freeze dryer. Drying Technol. 8(5): 993–1016. Bantle, M., Eikevik, T. M., 2011. Parametric study of high-intensity ultrasound in the atmospheric freeze-drying of peas. Drying Technol. 29(10): 1230–1239. Barresi, A. A., Fissore, D., 2011. In-line product quality control of pharmaceuticals in freeze-drying processes, in: Modern drying

technology, Vol. 3: Product quality and formulation, (eds. E. Tsotsas, A. S. Mujumdar), Wiley-VCH, Weinheim, Germany, pp. 91–154. Barresi, A. A., Pisano, R., Fissore, D., Rasetto, V., Velardi, S. A., Vallan, A., Parvis, M., Galan, M., 2009. Monitoring of the primary drying of a freeze-dying process in vials. Chem. Eng. Process. 48(1): 408–423. Barresi, A. A., Fissore, D., Marchisio, D. L., 2010. Process analytical technology in industrial freeze-drying, in: Freeze-drying/ Lyophilization of pharmaceuticals and biological products, (eds. L. Rey, J. C. May), Informa Healthcare, New York, USA, pp. 463–496. Boeh-Ocansej, O., 1988. Freeze drying in a fluidized-bed atmospheric dryer and in a vacuum dryer: Evaluation of external transfer coefficients. J. Food Eng. 7(2): 127–146. B€ohm, V., K€ uhnert, S., Rohm, H., Scholze, G., 2006. Improving the nutritional quality of microwave vacuum dried strawberries: A preliminary study. Food Sci. Technol. Int. 12(1): 67–75. Bursac, R., Sever, R., Hunek, B., 2009. A practical method for resolving the nucleation problem in lyophilization. Bioproc. Int. 7(9): 66–72. Chou, S. K., Chua, K. J., 2001. New hybrid drying technologies for heat sensitive foodstuffs. Trends Food Sci. Technol. 12: 359–369.

157

158

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries Chua, K. J., Chou, S. K., Ho, J. C., Hawlader, M. N. A., 2002. Heat pump drying: Recent developments and future trends. Drying Technol. 20(8): 1579–1610. Claussen, I. C., Andresen, T., Eikevik, T. M., Strømmen, I., 2007a. Atmospheric freeze drying – Modeling and simulation of a tunnel dryer. Drying Technol. 25(12): 1959–1965. Claussen, I. C., Hemmingsen, A. K. T., Rustad, T., Strømmen, I., 2007b. Relations between product structure, sorption characteristics and freezing point of atmospheric freeze dried foods. Drying Technol. 25(5): 853–865. Claussen, I. C., Ustad, T. S., Strømmen, I., Walde, P. M., 2007c. Atmospheric freeze drying A review. Drying Technol. 25(6): 957–967. Cui, Z. W., Xu, S. Y., Sun, D. W., 2003. Dehydration of garlic slices by combined microwave vacuum and air drying. Drying Technol. 21(7): 1173–1184. Di Matteo, P., Donsí, G., Ferrari, G., 2003. The role of heat and mass transfer phenomena in atmospheric freeze-drying of foods in a fluidised bed. J. Food Eng. 59 (2–3): 267–275. Drouzas, A. E., Schubert, H., 1996. Microwave application in vacuum drying of fruit. J. Food Eng. 28(2): 203–209. Eikevik, T. M., Alves-Filho, O., Bantle, M., 2012. Microwave-assisted atmospheric freeze drying of green peas: A case study. Drying Technol. 30(14): 1592–1599. Elia, A. M., Barresi, A. A., 1998. Intensification of transfer fluxes and control of product properties in freeze-drying. Chem. Eng. Process. 37(5): 347–358. Fairbank, H. V., 1975. Applying ultrasound to continuous drying process, in: Proceedings of Ultrasonic International 1975, IPC Science and Technology Press, Guildford, UK, 43–45. Fissore, D., Pisano, R., Barresi, A. A., 2011a. Advanced approach to build the design space for the primary drying of a pharmaceutical freeze-drying process. J. Pharm. Sci. 100(11): 4922–4933. Fissore, D., Pisano, R., Barresi, A. A., 2011b. On the methods based on the Pressure Rise Test for monitoring a freeze-drying process. Drying Technol. 29(1): 73–90.

Fissore, D., Pisano, R., Barresi, A.A., 2014. Applying quality-by-design to develop a coffee freeze-drying process. J. Food Eng. 123: 179–187. Flink, J. M., 1977. Energy analysis in dehydratation processes. Food Technol. 31: 77–84. Gallego-Juarez, J. A., Rodriguez-Corral, G., Galvez Moraleda, J. C., Yang, T. S., 1999. A new high-intensity ultrasonic technology for food dehydration. Drying Technol. 17(3): 597–608. Garcia-Perez, J., Carcel, J. A., Riera, E., Rossello, C., Mulet, A., 2012a. Intensification of low temperature drying by using ultrasound. Drying Technol. 30(11–12): 1199–1208. Garcia-Perez, J. V., Ortu~ no, C., Puig, A., Carcel, J. A., Perez-Munuera, I., 2012b. Enhancement of water transport and microstructural changes induced by highintensity ultrasound application on orange peel drying. Food Bioprocess. Technol. 5(6): 2256–2265. Ghio, S., Barresi, A. A., Rovero, G., 2000. A comparison of evaporative and conventional freezing prior to freeze-drying of fruits and vegetables. Food Bioprod. Process. 78(4): 187–192. Giordano, A., Barresi, A. A., Fissore, D., 2011. On the use of mathematical models to build the design space for the primary drying phase of a pharmaceutical lyophilization process. J. Pharm. Sci. 100 (1): 311–324. Hottot, A., Vessot, S., Andrieu, J., 2007. Freeze drying of pharmaceuticals in vials: Influence of freezing protocol and sample configuration on ice morphology and freezedried cake texture. Chem. Eng. Process. 46(7): 666–674. Inada, T., Zhang, X., Yabe, A., Yoshiyuki, K., 2001. Active control of phase change from supercooled water to ice by ultrasonic vibration, Part 1: Control of freezing temperature. Int. J. Heat Mass Transfer 44 (23): 4523–4531.  Indergard, E., Alves-Filho, O., Strømmen, I., Bøhn, T. I., 2001. Quality and properties of heat pump dried instant potato. Proceedings of 1st Nordic Drying Conference, Trondheim, Norway, June 27–29. Kasper, J. C., Friess, W. F., 2011. The freezing step in lyophilization: Physico-

References chemical fundamentals, freezing methods and consequences on process performance and quality attributes of biopharmaceuticals. Eur. J. Pharm. Biopharm. 78(2): 248–63. Kasraian, K., DeLuca, P. P., 1995a. Effect of tert-butyl alcohol on the resistance of the dry product layer during primary drying. Pharm. Res. 12(4): 491–495. Kasraian, K., DeLuca, P. P., 1995b. Thermal analysis of the tertiary butyl alcohol-water system and its implications on freeze-drying. Pharm. Res. 12(4): 484–490. Konstantinidis, A. K., Kuu, W., Otten, L., Nail, S. L., Sever, R. R., 2011. Controlled nucleation in freeze-drying: Effects on pore size in the dried product layer, mass transfer resistance, and primary drying rate. J. Pharm. Sci. 100(8): 3453–3470. Kramer, M., Sennhenn, B., Lee, G., 2002. Freeze-drying using vacuum-induced surface freezing. J. Pharm. Sci. 91(2): 433–443. Krokida, M. K., Maroulis, Z. B., MarinosKouris, D., 2002. Heat and mass transfer coefficients in drying: Compilation of literature data. Drying Technol. 20(1): 1–18. Kwok, B. H. L., Hu, C., Durance, T., Kitts, D. D., 2004. Dehydration techniques affect phytochemical contents and free radical scavenging activities of Saskatoon berries (Amelanchier alnifolia nutt). J. Food Sci. 69(3): S122–S126. Leuenberger, H., 2002. Spray freezedrying The process of choice for low water soluble drugs. J. Nanopart. Res. 4 (1–2): 111–119. Leuenberger, H., Plitzko, M., Puchkov, M., 2006. Spray freeze drying in a fluidized bed at normal and low pressure. Drying Technol. 24(6): 711–719. Liapis, A. I., Bruttini, R., 2008. Exergy analysis of freeze drying of pharmaceuticals in vials on trays. Int. J. Heat Mass Transfer 51(15–16): 3854–3868. Lin, T. M., Durance, T. D., Scaman, C. H., 1998. Characterization of vacuum microwave, air and freeze dried carrot slices. Food Res. Int. 31(2): 111–117. Liu, J., Viverette, T., Virgin, M., Anderson, M., Dalal, P., 2005. A study of the impact of freezing on the lyophilization of a

concentrated formulation with a high fill depth. Pharm. Dev. Technol. 10(2): 261–272. Liu, Y., Zhao, Y., Feng, X., 2008. Exergy analysis for a freeze-drying process. Appl. Therm. Eng. 28(7): 675–690. Lombra~ na, J. I., Villaran, M. C., 1996. Drying rate and shrinkage effect interaction during freeze drying in an adsorbent medium. J. Chem. Eng. Jpn 29(2): 242–250. Ma, Y. H., Peltre, P. R., 1975. Freeze dehydration by microwave energy. AIChE J. 21(2): 335–344. Meryman, H. T., 1959. Sublimation: Freezedrying without vacuum. Science 130(3376): 628–629. Mulet, A., Carcel, J. A., Sanjuan, N., Bon, J., 2003. New food drying technologies – Use of ultrasound. Food Sci. Technol. Int. 9 (3): 215–221. Mumenthaler, M., Leuenberger, H., 1991. Atmospheric spray freeze drying – A suitable alternative in freeze drying technology. Int. J. Pharm. 72(2): 97–110. Muralidhara, H. S., Ensminger, D., Putman, A., 1985. Acoustic dewatering and drying (low and high frequency): State of the art review. Drying Technol. 3(4): 529–566. Nakagawa, K., Hottot, A., Vessot, S., Andrieu, J., 2006. Influence of controlled nucleation by ultrasounds on ice morphology of frozen formulations for pharmaceutical proteins freeze-drying. Chem. Eng. Process. 45(9): 783–791. Oetjen, G. W., Haseley, P., 2004. Freeze-drying. Wiley-VCH, Weinheim, Germany. Orsat, V., Yang, W., Changrue, V., Raghavan, G. S. V., 2007. Microwave-assisted drying of biomaterials. Food Bioprod. Process. 85(C3): 255–263. Patapoff, T. W., Overcashier, D. E., 2002. The importance of freezing on lyophilization cycle development. BioPharm Int. 15(3): 16–21. Patel, S. M., Bhugra, C., Pikal, M. J., 2009. Reduced pressure ice fog technique for controlled ice nucleation during freezedrying. AAPS PharmSciTech 10(4): 1406–1411. Pester, R. P., Arsem, H. B., Ma, Y. H., 1977. Applications of microwave heating to freezedrying: Perspective. AIChE Symp. Ser. No. 163 73: 131–133.

159

160

5 Intensification of Freeze-Drying for the Pharmaceutical and Food Industries Petzold, G., Aguilera, J., 2009. Ice morphology: Fundamentals and technological applications in foods. Food Biophys. 4(4): 378–396. Pikal, M., 1990. Freeze-drying of proteins, Part 1: Process design. Biopharmacology 3: 18–27. Pisano, R., Fissore, D., Velardi, S. A., Barresi, A. A., 2010. In-line optimization and control of an industrial freeze-drying process for pharmaceuticals. J. Pharm. Sci. 99(11): 4691–4709. Pisano, R., Fissore, D., Barresi, A. A., 2013. Inline and off-line optimization of freezedrying cycles for pharmaceutical products. Drying Technol. 31(8): 905–919. Popovich, D. G., Hu, C., Durance, T. D., Kitts, D. D., 2005. Retention of ginsenosides in drying ginseng root: Comparison of drying methods. J. Food Sci. 70(6): S355–S358. Rambhatla, S., Ramot, R., Bhugra, C., Pikal, M. J., 2004. Heat and mass transfer scale-up issues during freeze drying, Part 2: Control and characterization of the degree of supercooling. AAPS PharmSciTech 5(4): 54–62. Ratti, C., 2001. Hot air and freeze-drying of high-value foods: A review. J. Food Eng. 49 (4): 311–319. Regier, M., Mayer-Miebach, E., Behsnilian, D., Neff, E., Schuchmann, H. P., 2005. Influences of drying and storage of lycopenerich carrots on the carotenoid content. Drying Technol. 23(4): 989–998. Rey, L., 2010. Glimpses into the realm of freeze-drying: Classical issues and new ventures, in: Freeze-drying/Lyophilization of pharmaceuticals and biological products, (eds. L. Rey, J. C. May), Informa Healthcare, New York, USA, pp. 1–28. Rey, L. R., Bastien, M. C., 1961. Biophysical aspects of freeze-drying: Importance of the preliminary freezing and sublimation periods, in: Freeze-drying of foods, (ed. F. R. Fisher), National Academy of Sciences, National Research Council, Washington D.C., USA, pp. 25–42. Rogers, T. L., Nelson, A. C., Sarkari, M., Young, T. J., Johnston, K. P., Williams, R. O., 2003. Enhanced aqueous dissolution of a poorly water soluble drug by novel particle engineering technology: Spray-

freezing into liquid with atmospheric freeze-drying. Pharm. Res. 20(3): 485–493. Saclier, M., Peczalski, R., Andrieu, J., 2010a. A theoretical model for ice primary nucleation induced by acoustic cavitation. Ultrason. Sonochem. 17(1): 98–105. Saclier, M., Peczalski, R., Andrieu, J., 2010b. Effect of ultrasonically induced nucleation on ice crystals’ size and shape during freezing in vials. Chem. Eng. Sci. 65(10): 3064–3071. Searles, J. A., Carpenter, J. F., Randolph, T. W., 2001a. The ice nucleation temperature determines the primary drying rate of lyophilization for samples frozen on a temperature-controlled shelf. J. Pham. Sci. 90(7): 860–871. Searles, J. A., Carpenter, J. F., Randolph, T. W., 2001b. Annealing to optimize the primary drying rate, reduce freezing-induced drying rate heterogeneity, and determine Tg0 in pharmaceutical lyophilization. J. Pharm. Sci. 90(7): 872–887. Sharma, G. P., Prasad, S., 2006. Optimization of process parameters for microwave drying of garlic cloves. J. Food Eng. 75(4): 441–446. Sochanski, J. S., Goyette, J., Bose, T. K., Akyel, C., Bosisio, R., 1990. Freeze dehydration of foamed milk by microwaves. Drying Technol. 8(5): 1017–1037. Stankiewicz, A. I., Moulijn, J. A., 2000. Process intensification: Transforming chemical engineering. Chem. Eng. Prog. 96(1): 22–34. Strømmen, I., Alves-Filho, O., Eikevik, T. M., 2005. Atmospheric freeze drying with heat pumps A new alternative for high-quality dried food products. Proceedings of 3rd Nordic Drying Conference, Karlstad, Sweden, June 15–17. Sunjka, P. S., Rennie, T. J., Beaudry, C., Raghavan, G. S. V., 2004. Microwave/ convective and microwave/vacuum drying of cranberries: A comparative study. Drying Technol. 22(5): 1217–1231. Teagarden, D. L., Baker, D. S., 2002. Practical aspects of lyophilization using non-aqueous co-solvent systems. Eur. J. Pharm. Sci. 15(2): 115–133. Tesconi, M. S., Sepassi, K., Yalkowsky, S. H., 1999. Freeze-drying above room temperature. J. Pharm. Sci. 88(5): 501–506.

References Torringa, E., Esveld, E., Scheewe, I., van den Berg, R., Bartles, P., 2001. Osmotic dehydration as a pre-treatment before combined microwave-hot-air drying of mushrooms. J. Food Eng. 49(2–3): 185–191. Tulasidas, T. N., Raghavan, G. S. V., Mujumdar, A. S., 1995. Microwave drying of grapes in a single mode cavity at 2450 MHz, Part 2: Quality and energy aspects. Drying Technol. 13(8–9): 1949–1971. Velardi, S. A., Barresi, A. A., 2008. Development of simplified models for the freeze-drying process and investigation of the optimal operating conditions. Chem. Eng. Res. Des. 86(1): 9–22. Venkatachalapathy, K., Raghavan, G. S. V., 1999. Combined osmotic and microwave drying of strawberry. Drying Technol. 17(4–5): 837–853. Wang, J., Xi, Y. S., 2005. Drying characteristics and drying quality of carrot using a two-stage microwave process. J. Food Eng. 68(4): 505– 511. Wittaya-Areekul, S., 1999. Freeze-drying from nonaqueous solution. J. Pharm. Sci. 26(1–4): 33–43.

Wolff, E., Gibert, H., 1990a. Atmospheric freeze-drying, Part 1: Design, experimental investigation and energy-saving advantages. Drying Technol. 8(2): 385–404. Wolff, E., Gibert, H., 1990b. Atmospheric freeze drying, Part 2: Modelling drying kinetics using adsorption isotherms. Dry. Technol. 8(2): 405–428. Xu, Y., Zhang, M., Mujumdar, A. S., Zhou, L., Sun, J., 2004. Studies on hot air and microwave vacuum drying of wild cabbage. Drying Technol. 22(9): 2201–2209. Yongsawatdigul, J., Gunasekaran, S., 1996a. Microwave-vacuum-drying of cranberries, Part 1: Energy use and efficiency. J. Food Process. Preserv. 20(2): 121–143. Yongsawatdigul, J., Gunasekaran, S., 1996b. Microwave-vacuum-drying of cranberries, Part 2: Quality evaluation. J. Food Process. Preserv. 20(2): 145–156. Zhang, X., Inada, T., Yabe, A., Lu, S., Kozawa, Y., 2001. Active control of phase change from supercooled water to ice by ultrasonic vibration, Part 2: Generation of ice slurries and effect of bubble nuclei. Int. J. Heat Mass Transfer 44(23): 4533–4539.

161

163

6 Drying of Foamed Materials Ireneusz Zbicinski, Julia Rabaeva, and Artur Lewandowski 6.1 Introduction

A large group of thermally sensitive materials exists, such as mixtures of biological substances, proteins, fats, vitamins and mineral salts, which are difficult to dry by using classical techniques. One method by which to overcome this problem is to foam the material before drying, and two basic techniques of foamed material drying have been developed: (i) the drying of dispersed foams (spray drying); and (ii) the drying of bulk foams (foam-mat drying). A conventional spray drying process is modified to foam spray drying by the injection of an inert gas (e.g., N2, NO2, CO2) into the feedstock with the following formation of foamed droplets during atomization. Compared to the standard spray drying process, the addition of a foaming gas increases the dryer throughput, intensifies the thermal process efficiency, reduces the thermal degradation of the product and increases the retention of highly volatile substances (Hanrahan and Webb, 1961; Bell et al., 1963; Abdul-Rahman et al., 1971; Crosby and Weyl, 1977). Investigations carried out over the past few decades have revealed that the foam spray drying process also affects the final product properties, such as particle diameter, porosity, and bulk density (Hanrahan and Webb, 1961; Hanrahan et al., 1962). In the foam-mat drying method, a liquid raw material is first whipped to produce a bulk volume of foam which is then air-, freeze-, vacuum-, or microwave-dried. Selected surfactants relevant to the properties of the product must be added in order to obtain a stable foam. Foam-mat drying, which is mainly applied to dry liquid or semi-liquid foods, offers a reduction in drying time at low drying temperatures due to the increased interfacial area of foamed materials, a high drying efficiency, and a quick rehydration of the product. The main drawbacks of the foam-mat drying process include foam stability problems, low dryer throughput, and a loss of aromatic components (Marques et al., 2006). Numerous reports have also been made describing a process similar to foaming denoted by aerification or aeration, in which air is introduced to the feed in order to increase the oxygen concentration or to change the rheological properties (Zuniga Modern Drying Technology Volume 5: Process Intensification, First Edition. Edited by Evangelos Tsotsas and Arun S. Mujumdar. Ó 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

164

6 Drying of Foamed Materials

and Aguilera, 2008; Campbell and Mougeot, 1999). Aeration, achieved by bubbling air through the liquid, by spraying the liquid into the air, or by agitation of the liquid is widely used in food manufacturing processes (e.g., ice-cream, bread, cakes, aerated water or cola) and in the chemical industry. An overview of the principles and applications of foam-mat and foam spray drying processes is provided in the following subsections.

6.2 Foam Properties

Foams are disperse systems in which the liquid is a dispersing phase and the gas is a dispersed component. In order to obtain a solution of the consistency of foam, a two-phase liquid–gas system with the application of an appropriate surfactant must be formed (Malysa and Lunkenheimer, 2008). Due to the energy supplied to disperse the gas phase, gas bubbles surrounded by a thin liquid film containing surfactant molecules are released. In the case of pure solutions, the process of foam formation is very difficult because gas bubbles contained in the liquid or solution cannot be stabilized without a surfactant (Prud’Homme and Khan, 1996). A characteristic feature of all foams is low density. Energy must be supplied to the system to obtain a material with the consistency of a foam, both in liquid and solid form, irrespective of its composition and use. Foams can be obtained by both condensation and dispersion methods (Ekserova and Kruglyakov, 1998):  Condensation methods include all processes in which gas bubbles are formed due to a decrease in the external pressure (boiling, cavitation, desorption, chemical reaction).  In dispersion methods (sieving, gas admixing, mixing), the process of foam formation is connected with gas dispersion which involves the introduction of gas bubbles into the solution. Bubbling, pneumatic processing or mechanical action (e.g., stirring) are used to produce foam. Usually, N2, CO2 or atmospheric air, all of which are relatively cheap foam-forming gases, are used in the process (Farajzadeh et al., 2011). Foam is modified during foaming, and the geometric shape of the foam cells, in relation to the foam microstructure, depends on the gas-to-liquid volume ratio. The gas volume in single foam cells can reach even 75% of foam volume. At the first stage of foam formation, a single layer of bubbles forms a thin foam film on the liquid surface. An increase of the gas volume in the liquid then causes foam to further build-up until there is a loss of stability due to the excessively increased volume of gas bubbles (Ekserova and Kruglyakov, 1998). One method of improving foam stability is to increase the concentration of surfactants (e.g., soy, whey, milk protein) and the thickness of the films (Philips et al., 1987). The stability of a foam structure depends on the location of cells and their mutual position in space.

6.2 Foam Properties

Fig. 6.1 Different structures of droplets obtained by atomization. (a) Non-foamed; (b) Foamed with one bubble; (c) Foamed with many bubbles.

In spray drying, a gas-admixing technique is used to produce a dispersed foam from the feed, whereas in foam-mat drying mechanical methods (e.g., mixing, whipping) are applied to stabilize a bulk volume of the foamed raw material. In foam spray drying, depending on the drying and spraying process parameters, irregular and complex structures of foamed particles can be obtained. Figure 6.1 shows an example of foamed (Fig. 6.1b, c) and non-foamed (Fig. 6.1a) particles obtained as a result of atomization (Rabaeva, 2012). Spray foaming develops the following forms of droplet/particle structure: droplets with no gas inside (Fig. 6.1a); droplets with gas inside (Fig. 6.1b); and droplets with entrapped gas bubbles with inner bridges of liquid or dried solution (Fig. 6.1c). The drying mechanism of a droplet depends on the droplet structure. AbdulRahman et al. (1971) and Crosby and Weyl (1977) estimated that the heat and mass transfer rates for foamed droplets (Fig. 6.1b and c) were faster than for non-foamed droplets (Fig. 6.1a), due to an accelerated transport of liquid water to the evaporation front (Ratti and Kudra, 2006) and an increase in the droplet surface area caused by gas expansion during drying. During the foam spray drying process, particles with all of the above structures are formed simultaneously (Fig. 6.1a–c). Figure 6.2a and b show the structure of foamed and non-foamed spray-dried maltodextrin powder at drying temperature 200  C, respectively (Rabaeva, 2012). On analyzing Fig. 6.2b, particle structures can be observed with voids and inner bridges, which are similar to the foamed droplet structure shown in Fig. 6.1c. The effect of foaming on dried particle morphology was also analyzed by Frey and King (1986), who found that gas-admixing foam spray drying produced particles with many small internal voids that may result from the entrainment of air during the atomization or desorption of dissolved air (Greenwald and King, 1981). Large single internal voids were observed for particles produced by the gasdesorption method. These observations were in agreement with results produced by Hanrahan et al. (1962) and Abdul-Rahman et al. (1971), who investigated particle morphology for whole-milk and sodium caseinate solutions. In the foam-mat drying method liquid foods are whipped into stable foams by using mainly mechanical methods, such as mixing to produce a bulk volume of

165

166

6 Drying of Foamed Materials

Fig. 6.2 Microscope images of spray-dried powder (maltodextrin, GMS, drying temperature 200  C). (a) Non-foamed; (b) Foamed.

foam. In order to obtain a stable foam it is necessary to add carefully selected surfactants that are relevant to the properties of the product. An example of the bulk volume structure of foam-dried particles (e.g., maltodextrin/sodium caseinate powder) is shown in Fig. 6.3 (Schoonman et al., 2001). Here, the solid matrix, voids, open and closed pores and bubbles, micropores and cracks create a complex structure that affects both heat and mass transfer during drying. Bulk foams processed by the direct foaming of slurries have a tendency to develop defects such as cracks or cell size gradients as a result of a rise in capillary stress due to water evaporation during drying. Today, a vast literature is available on the reduction of microstructural inhomogeneities of foams by means of, mainly, the selection of suitable surfactants, the concentration of surfactants, and the drying method or adjustment of drying process parameters (Whang et al., 1995; Schoonman et al., 2001; Kim et al., 2004; Pisal et al., 2006; Chino and Dunand, 2008; Sundaram and Durance, 2008; Pradhan and Bhargava, 2008; Loa et al., 2011).

Fig. 6.3 Bulk volume of foam-dried particles (maltodextrin/sodium caseinate powder), according to Schoonman et al. (2001); 1: Solid matrix; 2: Voids; 3: Open pores; 4: Closed pores; 5: Cracks; 6: Connected pores.

6.3 Foam Spray Drying

6.3 Foam Spray Drying 6.3.1 Processing Principles

The foam spray drying process was developed for dairy industry at the start of the twentieth century. The spray drying of foamed materials was patented in 1917 (Campbell, 1917) to dry foamed milk and egg albumin (Ratti and Kudra, 2006), but some 40 years later Sinnamon et al. (1957) found that the vacuum drying of concentrated milk foam would produce a whole-milk powder of good flavor and dispersibility. Morgan et al. (1959) showed that the dispersibility of whole-milk powder could also be maintained in foamed products during drying at atmospheric pressure. Although, since that time, very few studies on foam spray drying have been reported, this technology has more recently attracted attention because of the possibilities of processing hard-to-dry food materials and to control the final product properties such as bulk density, porosity, solubility, and wettability. Two main methods have been developed for feed foaming applied to spray drying, namely gas-desorption and gas-admixing. In the gas-desorption method, the gas (CO2) is dissolved in the liquid feed under moderate pressure; subsequently, after atomization, droplets supersaturated with the gas are formed, such that the bubbles become nucleated and grow within the droplets. In the gas-admixing method, foaming gas is directly added to the feed. However, in order to obtain a foamed feed, a precisely determined amount of inert gas (e.g., air, N2, NO2, CO2) must be injected into the feed at a high pressure, between the pump and the atomizer. The most important part of the foaming system is a mixing device; typical gasmixing devices, as applied by Hanrahan and Webb (1961) and Zbicinski and Rabaeva (2010), are shown in Fig. 6.4. A gas-mixing chamber consists of a main pipe through which the feed flows, and a separate channel for the foaming gas injection. After flowing through the gasmixing chamber, a two-phase feed (solution and gas) is delivered to a mechanical homogenizer to produce fine foaming gas bubbles in the liquid, which is then transferred to the nozzle. An example of a typical set-up of the gas-admixing foaming system for spray drying is shown in Fig. 6.5 (Rabaeva, 2012). The gas-admixing foaming system consists of the following elements:      

tank/mixer of the solution (1) mono pump (2) tank with foaming gas, here N2 (3) gas pressure valve (4) gas heater (5) foaming gas flow meter/control (6)

167

168

6 Drying of Foamed Materials

Fig. 6.4 Gas mixing devices. (a) Adapted from Hanrahan and Webb (1961); (b) Adapted from Zbicinski and Rabaeva (2010).

     

foaming (mixing) device (a system for gas dosing into the feed) (7) homogenizer (8) foamed solution flow rate and density meter (9) pressure nozzle (10) pressure sensors (P) temperature sensor (T).

An aqueous solution of the material to be dried with an appropriate amount of surfactant is prepared in the tank (1) with an agitator, and pumped (2) by a thermostated pipeline to the spraying nozzle (10). Foaming gas from a high-

Fig. 6.5 Schematic diagram of a gas-admixing foaming system. See text for details.

6.3 Foam Spray Drying

pressure tank (3) is injected into the feed in the mixing device (7). The amount of supplied foaming gas (e.g., nitrogen) must be measured and controlled by means of, for example, a mass flow meter (6). A mechanical homogenizer (8) should be used to produce a uniform mixture of foaming gas and raw material and to eliminate pulsations of flow. From an economic point of view, foaming reduces the cost of manufacturing instant powders by requiring only minor equipment changes that do not affect the efficiency of a dryer to produce standard types of powder. Abdul-Rahman et al. (1971) conducted fundamental experiments on the drying of a single drop of a 15% w/w aqueous solution of sodium caseinate foamed with N2 and NO2. The suspended foamed drop, with an initial mass of either 1 or 2 mg, was exposed to bone-dry air at 100  C and 200  C, respectively, flowing with a relative velocity of 0.94 m s 1. The results of these experiments showed that the drying and temperature histories of foamed drops were similar to those of non-foamed drops. The foamed drops, which contained approximately 56% of inert gas (volume fraction), exhibited increased drying rates when compared to non-foamed drops. Although foaming did not suppress the temperature level, the reduced exposure time of the drop, which resulted from the accelerated drying rates, indicated that a five- to 10-fold reduction of thermal degradation could be expected for foamed drops. The results of experiments with drops foamed to different volume fractions of nitrogen suggested that the drying time would be shortened with an increased volume fraction of foaming gas. This proposal concurs with a note by Crosby and Weyl (1977), who stated that drying would be promoted when the gas fraction in foam spray-dried feed stocks was relatively high. Subsequent data delivered by Crosby and Weyl (1977) on the moisture change of foamed and non-foamed supported single drops (15% w/w aqueous sodium caseinate solution) showed a quicker dehydration rate for the foamed drops (Fig. 6.6). The effect of foaming on spray parameters and losses of volatile organic components during foam spray drying was investigated by Frey and King (1986). The tests performed indicated that, for the gas-admixing method of foaming, foaming affected the initial atomization parameters due to an earlier break-up of the liquid sheet to droplets (reduction in sheet length) compared to the non-foamed solution. However, this effect was not observed for the gasdesorption method. Frey and King (1986) proved that foaming by gas-admixing enhanced the retention of highly volatile substances (Fig. 6.7) and also improved powder solubility. Additionally, Frey and King (1986) compared the retention of highly volatile substances in gas-admixing and gas-desorption foam spray drying, and concluded that an enhanced retention of volatile substances takes place only in the gas-admixing method, as the result of an earlier break-up of the liquid flow to droplets during atomization. A noticeable decrease in n-propyl acetate retention was found for the gas desorption feed foaming method when compared to nonfoamed feed; this was due to the nucleation of bubbles within the droplets, which promoted the escape of volatile compounds (Fig. 6.8).

169

170

6 Drying of Foamed Materials

Fig. 6.6 Moisture change for foamed and non-foamed supported drops (15% w/w aqueous sodium caseinate solution, drying temperature 100  C). According to Crosby and Weyl (1977).

A systematic analysis of the foam spray drying process, including investigations of spray hydrodynamics (particle size distribution, particle velocity, centricity), heat and mass transfer between the phases, drying kinetics and the effects of feed foaming on final product properties, was carried out by Zbicinski and Rabaeva (2010) and Rabaeva (2012).

Fig. 6.7 Retention of n-propyl acetate in the nozzle region during gas-admixing and gas-desorption foam spray drying of 60% sucrose solution (200  C air temperature, 70 bar atomization pressure). Adapted from Frey and King (1986).

6.3 Foam Spray Drying

Fig. 6.8 Retention of n-propyl acetate during non-foamed and gas-desorption foam spray drying of 60% sucrose solution (200  C, 70 bar atomization pressure). Adapted from Frey and King (1986).

Figures 6.9 and 6.10 show changes in the moisture content of foamed and nonfoamed feed as a function of the distance from the atomizer for two drying air temperatures (175 and 150  C) for the drying of maltodextrin. The drying process proceeded faster for foamed feed (feed rate 9 kg h 1, foaming gas (N2) rate 40 g h 1) at both air temperatures. The foam spray drying process of maltodextrin was complete at about 0.4 m (1.2 m for non-foamed feed) from the nozzle at a

Fig. 6.9 Profile of material moisture content as a function of time for spray drying with and without gas-admixing (maltodextrin, surfactant GMS, foaming gas N2, drying temperature 175  C).

171

172

6 Drying of Foamed Materials

Fig. 6.10 Profile of material moisture content as a function of time for spray drying with and without gas-admixing (maltodextrin, surfactant GMS, foaming gas N2, drying temperature 150  C).

drying temperature of 175  C (Fig. 6.9) and at about 1 m (1.8 m for non-foamed feed) from the nozzle at a drying temperature of 150  C (Fig. 6.10). The faster fall in foamed material moisture content was explained by an increase in the evaporation area per unit of drop/particle mass due to the expansion of gas trapped by the liquid (Hanrahan and Webb, 1961; Crosby and Weyl, 1977). No effect of feed foaming was reported on the spray hydrodynamics for all admixed gas flow rates. 6.3.2 Final Product Properties

All available studies revealed a strong effect of feed foaming on the final product properties. As early as the 1960s, Hanrahan and Webb (1961) had obtained a stable noncaking powder of cottage cheese whey by introducing compressed air into the solution before atomization. The powder obtained as a result of foaming had excellent flow characteristics, whereas the spray-dried powder that had not been subjected to foaming showed a tendency to adhere to the dryer surface during the process. Information on the effect of foaming on the final properties of spray-dried cottage cheese whey is provided in Table 6.1. These authors found that powder moisture content decreases as a result of foaming, preventing the lactose crystallization that is important in food manufacture where lactose and lactic acid are desirable ingredients. The main goal of the experiments performed by Hanrahan et al. (1962) was to obtain a highly soluble powder which could be reconstituted to a beverage that was

174

6 Drying of Foamed Materials

Fig. 6.11 Particle size distribution in spray-dried powders of whole milk produced with and without gas incorporation. Adapted from Hanrahan et al. (1962).

virtually indistinguishable from fresh pasteurized milk. These authors showed that inert gas, when introduced into the droplets, caused an increase in powder particle diameter and porosity, which in turn affected the material bulk density. The effect of foaming on the particle size distribution of whole-milk powder is shown in Fig. 6.11. Changes in the final product properties of whole-milk foamed with nitrogen, as presented by Hanrahan et al. (1962), are listed in Tab. 6.2. An analysis of the data in Tab. 6.2 shows a decrease in the bulk density of the product, increases in the free fat content and dispersibility, and also an increase in the average individual particle diameter at a higher nitrogen injection rate. A method of producing readily dispersible non-fat dry skim milk by injecting compressed air into the feed was compared with a conventional spray drying process by Bell et al. (1963). These authors studied the effect of various drying conditions on the properties of foam spray-dried non-fat milk, and suggested that the foam spray drying method would produce a rapidly dispersible product during drying, without a subsequent agglomerating step. Bell et al. (1963) claimed that, since the injection of gas improves the efficiency of water removal and enables the drying of skim milk concentrates with up to 60% solids, the dryer throughput might be increased. Tamsma et al. (1971) failed to observe any significant variations in physical properties or flavor when using different foaming gases in the spray drying of whole-milk, and concluded that air can be substituted for nitrogen as the foaming agent. Extensive experiments on the effects of feed rate and admixed gas flow rate on the final product properties for skin-forming materials (maltodextrin) and porous

176

6 Drying of Foamed Materials

materials (detergent) were carried out in a concurrent spray drying tower by Rabaeva (2012). The final properties of maltodextrin and detergent powder for different drying conditions are presented in Tabs. 6.3–6.5. The analysis of results for maltodextrin showed that there is a clear trend in changes of the properties of this product with increasing foaming gas flow rates (Tabs. 6.3 and 6.4). Whilst the bulk, tap and apparent density was decreased, the Sauter mean diameter and particle porosity were each increased, mainly as a result of the expansion of gas bubbles that had become trapped in the droplets during foaming. The wettability of maltodextrin was determined according to the Polish Standard (PN-88/A-86030) and expressed as the time after which a powder was submerged in the water. It was found that the wettability of maltodextrin decreases with an increase in foaming gas flow rate, as larger particles produced by foaming require longer submergence times. A foaming agent (glycerol monostearate; GMS), which was added to the raw material to make the foam more stable, can also affect the wettability properties of maltodextrin. On analyzing the wettability of the detergent, which is a porous material, an opposite effect can be observed, namely that the wettability increases as the wetting time decreases at a higher foaming gas flow rate (see Tab. 6.5). Typically, a better wettability of a porous material is the result of capillary forces that make the wetting of such particles potentially easier than the wetting of skin-forming products, where the skin on the particle surface must be dissolved first. A different effect of feed foaming was observed on the apparent particle density of maltodextrin and detergent. For maltodextrin, the apparent density decreased with higher foaming gas flow rates due to the presence of closed internal voids inside the particles as a result of foaming. For the detergent (Tab. 6.5), the effect of foaming on apparent particle density was less pronounced, due to the porous structure of the material. Although the detergent apparent particle density did not change significantly, its bulk density decreased with an increase in the foaming gas flow rate, which could be related to morphological changes that had occurred during drying (particle shape, surface structure, etc.) No significant effect of feed foaming on the solubility of maltodextrin was observed in the tests performed, whereas a lower solubility of the detergent could be explained by its chemical composition with poorly soluble components in the formula. The natural angle of repose, which determined the “looseness” of powders, is not significantly affected by feed foaming. A slight reduction in the natural angle of repose for maltodextrin was observed with an increase of the foaming gas flow rate (Tabs. 6.3 and 6.4), which was caused by the formation of more cohesive particles (Teunou et al., 1995). Figures 6.12–6.15 show example electron microscopy images of foamed and nonfoamed detergent and maltodextrin powders for different drying temperatures. In these images, a significant difference can be observed in the structure of foamed and non-foamed spray-dried products, with foaming leading to an increase in particle diameter and the formation of a “porous” shell. Under certain drying

6.3 Foam Spray Drying

Fig. 6.12 Microscopic image of spray-dried powder (detergent, drying temperature 200  C) (Rabaeva and Zbicinski, 2010). (a) Non-foamed; (b) Foamed.

conditions, some of the particles have bigger holes in the shell (Figs. 6.14b and 6.15b), caused by expansion of the nitrogen. A similar observation was made by Hanrahan et al. (1962) and Hassenkloever and Eggers (2010) when investigating milk and coffee powders that had been foamed with nitrogen. Foamed particles of maltodextrin (Figs. 6.13b–6.15b) have a more uniform structure and spherical shape which affects the bulk and apparent densities. For foamed material, the larger the powder particles the smaller is the bulk density of the product (Tabs. 6.3 and 6.4). For non-foamed material, the voids between large particles in the bed are filled with smaller particles, which causes an increase in the

Fig. 6.13 Microscopic image of spray-dried powder (maltodextrin, surfactant GMS, drying temperature 200  C). (a) Non-foamed; (b) Foamed.

179

180

6 Drying of Foamed Materials

Fig. 6.14 Microscopic image of spray-dried powder (maltodextrin, surfactant GMS, drying temperature 175  C). (a) Non-foamed; (b) Foamed.

bulk density. An increase in the drying air temperature results in an increase of the volume of foaming gas trapped in the particles, which in turn causes a decrease in the apparent particle density (Greenwald and King, 1981; Verhey, 1973). In the case of a detergent, the particles show more damage due to foaming (Fig. 6.12b), and this may result in a better wettability due to an easier water penetration of the burst particles. In powders obtained at the highest foaming gas flow rate there will be a number of fractured particles that would cause a looser packing of the bed of material and a reduction in the bulk density.

Fig. 6.15 Microscopic image of spray-dried powder (maltodextrin, surfactant GMS, drying temperature 150  C). (a) Non-foamed; (b) Foamed.

6.4 Foam-Mat Drying

Extensive experiments were conducted by Lewandowski et al. (2012) to determine the effect of feed foaming on the efficiency of sunflower oil microencapsulation, and on selected properties of the final product. These experiments were performed in a pilot-plant concurrent spray dryer, using the gas-admixing technique for foaming. The microencapsulation efficiency of the foamed feed was shown to decrease by about 6% for an air temperature of 150  C, and by 25% for a temperature of 200  C at the highest foaming gas injection rate in relation to non-foamed feed. Nevertheless, the microencapsulation efficiency of sunflower oil remained at a level of 80% to 60%, which was satisfactory for the spray drying technique. The same group also confirmed a 50% reduction in the apparent density of powders at different foaming rates in comparison to the powder produced from a non-foamed emulsion. It was concluded that foaming of the emulsion would enable the control of selected powder properties while retaining a high efficiency of the spray drying microencapsulation process. To summarize, the properties of dry powders, belonging to different groups of materials, depend on the parameters of foaming, atomization, and drying. All results presented to date have confirmed a strong effect of feed foaming on the final product properties of porous and skin-forming materials. Examples of these include an increase in the Sauter mean diameter and porosity of the products, and a decrease in the bulk and tap densities at higher foaming gas flow rates. Hence, feed foaming can be used to control the physical properties of selected products in industrial spray drying applications.

6.4 Foam-Mat Drying

In foam-mat drying, liquid foods are first whipped into a stable foam that is then air-, freeze-, vacuum-, or microwave-dried. In general, wet foams are dried by convection (e.g., belt, drum, tray, oven dryers) by flowing hot air over or through a relatively thin (3–10 mm) layer of the foamed material. A benefit of foam-mat drying is that it allows the processing of difficult-to-dry materials such as tomato paste, fruit pulps and juices at low temperatures and within short drying times when compared to non-foamed materials dried using the same type of dryer. The main drawback of the foam-mat drying process is the limited throughput due to the small load of material that can be applied per unit dryer surface area, though this might be compensated by a shorter drying time. The classical drying of food by convection is a slow process that can result in thermal degradation of the material’s structure, shrinkage, a loss of color, and a deterioration of the organoleptic properties of the materials being dried. In the case of ceramic or building materials an increase in both transverse and longitudinal stresses is observed, which can lead to fractures in the material structure. The use of a foam-mat drying technique reduces the energy consumption of the process when compared to the traditional drying of non-foamed material, and also

181

182

6 Drying of Foamed Materials

intensifies the mass transfer (Hart et al., 1963). Products dried with the foam-mat drying method are characterized by a porous structure (Kadam et al., 2012), a low density, and a low water activity. Due to its shorter drying times, foam-mat drying can also reduce investment and operational costs (Ratti and Kudra, 2006). In foam-mat drying, solutions of foams are first prepared in tanks to which foaming gas (N2, NO2, CO2 or air) is applied. Subsequently, depending on the amount of foaming gas applied, the quantity and type of surfactant, and also the type of foamed solution, various foams with different structures will be obtained. Generally, a uniform foam consistency is obtained by using high-speed stirrers and homogenizers. Materials with a low moisture content are hydrated to obtain pulp with a 40–60% water content. The process of foam-mat drying requires the use of foam-forming substances that are capable of producing foams of high stability. For food processing, surfactants of biological character belonging to protein groups, such as egg albumin or milk protein, are used. Typically, a mass fraction of 2–10% egg albumin in the material will cause a 40–50% decrease in material density (Rajkumar et al., 2007; Dattatreya et al., 2010; Krasaekoopt and Bhatia, 2012). Subsequently, Borcherding et al. (2008) showed that the addition of 0.5% methylcellulose to mango pulp produced a foam with between 10% and 25% greater stability. Other materials used to create foams to be dried with foam-mat drying include surfactants such as Spam 60, Tween 80 and GMS. These foams have low density, in the range of 200 to 900 kg m 3. One other parameter that can influence the final product quality is the drying temperature. Depending on the type of material to be dried, and also its thickness, the range of temperatures used for foam drying is 45 to 90  C (Krasaekoopt and Bhatia, 2012). In the case of tomato juice or foamed mango and apple juice containing 40–50% moisture, the temperature range was 50 to 90  C (Kadam et al., 2012; Karim and Wai, 1999; Dattatreya et al., 2010), whereas for dairy products such as foamed yogurt the optimum range of drying temperatures was 50 to 70  C. An increase in the drying temperature not only intensifies the process of evaporation but also increases the susceptibility of the material to deformation. In the case of materials with lower moisture contents, such as foamed bananas (Thuwapanichayanan et al., 2008), the optimum temperature was 60  C. In order to improve final product quality, it is important to first select an appropriate thickness of foam solution to be dried. Initially, it was shown that, depending on the foam layer thickness, the density and porosity of the structure, and also its moisture evaporation rate, were changed (Bates, 1964). In studies performed by Rajkumar et al. (2007), Sankat and Castaigne (2004), Kudra and Ratti (2008), the thickness of the material dried by foam-mat drying was found to determine the drying time. Typically, foams of 1 to 5 mm thickness evaporated faster than non-foamed materials; as examples, the drying time of a 79.6% mango solution of density 1020 kg m 3 was 90–180 min, whereas for a foamed solution of the same initial moisture content and density 540 kg m 3 the drying time was 40–80 min. In the case of 10–20 mm-thick foamed banana pulp, the drying time was reduced from 46 h to 10 h (Sankat and Castaigne, 2004, Kudra and Ratti, 2008). Due to the high sugar content of these materials, the use of foams thicker than 20 mm may lead to

6.4 Foam-Mat Drying

physical changes in the final product. However, in order to eliminate any negative effects of the sugars and acids, which cause product caramelization at higher temperatures, the level of such components should be reduced in the material to be dried (e.g., Hart et al., 1963; Sankat and Castaigne, 2004). The drying of non-foamed fruit juice generally results in the production of a high-viscosity liquid syrup, as the high drying temperature employed causes the sugars contained in the material to melt. Foaming facilitates reductions of both the drying agent temperature and drying time, which may have significant effects not only on the costs of dry material production (Ratti and Kudra, 2006) but also the final product quality. The main goal of research into foam-mat drying is first, to determine optimal drying conditions and, second, to select an appropriate surfactant in order to obtain high-quality products (e.g., Vernon-Cartera et al., 2001; Rajkumar et al., 2007; Kadam et al., 2012; Thuwapanichayanan et al., 2012). In general, foamed materials may dry in constant and falling drying rate period, or in falling rate period only. An example of a foam-mat drying process which is controlled by internal diffusion is shown in Fig. 6.16 (Kudra and Ratti, 2006), where no clear period of constant drying rate is apparent on the drying curves for either foamed or non-foamed apple juice when dried as a 19 mm layer (air temperature: 55  C). Similar results were reported by Thuwapanichayanan et al. (2012) for the drying of banana pulp. When the foam drying process is controlled by both convection and diffusion (e.g., Lewicki, 1975, 2006), the drying rate is seen to depend on product characteristics such as the chemical composition, the type of foaming agent, solids content, and the initial moisture content. Drying curves for foamed tomato paste (Lewicki, 1975, 2006) at different foam densities are shown in Figure 6.17, an

Fig. 6.16 Drying curves for foamed and non-foamed apple juice dried in 19 mm layer by air at 55  C. Adapted from Kudra and Ratti (2006).

183

184

6 Drying of Foamed Materials

Fig. 6.17 Drying curves of foamed tomato paste with both constant and falling drying rate periods. Air temperature 60  C; layer thickness 3 mm. Adapted from Lewicki (2006).

analysis of which shows that the lower the density of the foamed product, the higher is the drying rate (in the case analyzed, the rate is increased by about 50%). Higher drying rates obtained at a lower density result from the specific structure of low-density foams, which contain very small bubbles but have a high level of homogeneity. Aside from layer thickness, air temperature and air velocity, the drying rate of foamed materials depends mainly on the foam’s characteristics such as stability, density, and bubble size. The foam-mat drying technique was used successfully to dry exotic fruits such as mango (Cooke et al., 1976; Rajkumar et al., 2007), starfruit (Karim and Wai, 1999), banana (Sankat and Castaigne, 2004; Thuwapanichayanan et al., 2008, 2012), tomato paste (Lewicki, 1975, 2006; Kadam et al., 2012), and yogurt (Krasaekoopt and Bhatia, 2012). Because the density of foamed materials is lower than that of non-foamed materials (typically in the range of 300–600 kg m 3), the mass load of a foam-mat dryer is also lower, which in turn leads to reductions in both dryer throughput and drying time. When Marques et al. (2006) compared different methods for drying tropical fruit pulps, they recommended convective foam-mat drying for liquid foods as well as semi-liquid and thermosensitive foods, and stressed that the main advantages of this method were a reduced drying time, quick rehydration, and low drying temperatures. The main drawbacks of foam-mat drying were nonenzymatic darkening, a loss of aromatic components, and problems with foam stability. Kudra and Ratti (2006) performed an energy and cost analysis of convective drying (belt conveyor and drum dryer) of mats of foamed and non-foamed apple juice. The energy consumption for drying foamed apple juice was estimated to be

6.4 Foam-Mat Drying

Fig. 6.18 Drying efficiency over moisture content for foamed and non-foamed apple juice (layer thickness 19 mm; air temperature 55  C). Adapted from Kudra and Ratti (2006).

20% of that for drying non-foamed juice, and with an 18% higher throughput that would reduce capital costs by about 11% for a belt conveyor dryer, and by 10% for a drum dryer. According to these authors, the main factor responsible for the reduced drying time was the increased interfacial area of the foamed materials. The drying efficiency for foamed and non-foamed apple juice dried in a 19 mm layer at an air temperature of 55  C is illustrated in Figure 6.18. The drying efficiency of non-foamed apple juice was shown to be about 30% (a standard value for convective drying), whilst the efficiency of foam-mat drying of apple juice was well above 60%, and the result of a well-developed heat and mass transfer area. Thuwapanichayanan et al. (2012) determined the drying curves of banana foams produced with aid from one of three foaming agents, namely egg albumin (EA), soy protein isolate (SPI) and whey protein concentrate (WPC) (Fig. 6.19). The results of these studies showed a strong effect of surfactant type on drying rate due to the development of different foam structures, while the exponential decrease in moisture content with time proved that the drying process of banana foams was controlled by internal diffusion. Recently, several investigations of foam-mat freeze-drying have been reported to combine the advantages of foam-mat drying and freeze-drying, notably a relatively short drying time and the production of high-quality products. The main areas of application included mainly the pharmaceutical and medical industries (Whang et al., 1995; Lee et al., 1997; Kim et al., 2004; Chino and Dunand, 2008), as well as the food and ceramics industries (Fukasawa et al., 2001; Marques et al., 2006; Muthukumaran et al., 2008; Raharitsifa and Ratti, 2010). Muthukumaran et al. (2008) studied the foam-mat freeze-drying of egg white with xanthan gum as a stabilizer to estimate the total drying time and quality of the

185

186

6 Drying of Foamed Materials

Fig. 6.19 Drying curves of banana foams with different foaming additives. Egg albumin (EA), soy protein isolate (SPI), and whey protein concentrate (WPC). Adapted from Thuwapanichayanan et al. (2012).

powder (Fig. 6.20). On analyzing the moisture content reduction, it could be seen that almost 15 h was necessary to remove 50% of ice from the non-foamed material, compared to only 6 h with the foamed sample. This was attributed to an increase in the diffusion coefficient due to the development of a porous structure of the foamed sample.

Fig. 6.20 Foam-mat freeze-drying of egg white. Stabilizer, xanthan gum. Adapted from Muthukumaran et al. (2008).

6.5 Summary

Raharitsifa and Ratti (2010) showed that foaming reduced the time of the freezedrying process for the same sample thickness in comparison with drying nonfoamed apple juice. The foam-mat freeze-drying technique was also successfully applied to produce nanoporous cellulose foams (Deng et al. (2009) with a porosity of 99%, and also macroporous solid foams from multiwalled carbon nanotubes (Thongprachan et al., 2008). Vacuum foam-mat drying (Jaya and Das, 2004), or a combination of vacuum and freeze-drying (Moy, 1971) or vacuum and microwave drying (Sundaram and Durance, 2008), represent other options to produce high-quality powders with a reduced oxidative degradation (i.e., browning), a high storage stability, and freeflowing properties.

6.5 Summary

Foam drying techniques enable the processing of difficult-to-dry materials (e.g., fruit pulps, juices and dairy products) at low temperatures and with short drying times. As a consequence, the thermal degradation of dry products can be reduced when compared to the convective drying of non-foamed materials, whilst the desired final product properties can be obtained. The use of a foam-mat drying technique reduces the energy consumption of the process in relation to the traditional drying of non-foamed material, and also intensifies mass transfer. Products dried with a foam-mat drying technique are characterized by a porous structure, a low density and low water activity; yet, due to the shorter drying times required, the method can also reduce investment and operational costs. The lower throughput of the foam-mat drying process, due to the smaller load of material per unit surface area of the dryer, might be compensated by a shorter drying time. Nonenzymatic darkening of the product, a loss of aromatic components and problems relating to foam stability are the main drawbacks of foam-mat drying. In foam spray drying, essentially all studies performed have confirmed a strong effect of feed foaming on the properties of both porous and skinforming material products. These include an increase in Sauter mean diameter and porosity of the products, a decreased bulk, tap and apparent particle density, and an influence on particle morphology. Foam spray drying enhances the retention of highly volatile substances and improves powder solubility (for gas-admixing foaming methods). The higher drying rate observed with foam spray drying can be explained by an increase in the evaporation area per unit of drop/particle mass due to an expansion of foaming gas trapped by the liquid. Feed foaming in spray drying enables selected physical properties of the products to be controlled, under industrial conditions. The foam spray and foam-mat drying technologies represent modern solutions aimed at reducing the environmental impact and economic costs of drying by

187

188

6 Drying of Foamed Materials

enabling the use of noxious materials at decreased energy consumption during the process. Additional Notation Used in Chapter 6

Nil Abbreviations

d.m. EA GMS SPI WB WPC

dry matter egg albumin glycerol monostearate soy protein isolate wet basis whey protein concentrate

References Abdul-Rahman, Y. A., Crosby, E. J., Bradley, R. L., 1971. Drying of single drops of foamed and non-foamed sodium caseinate solutions. J. Dairy Sci. 54(8): 1111–1118. Bates, R. P., 1964. Factors affecting foam production and stabilization of tropical fruit products. Food Technol. 18(1): 93–96. Bell, R. W., Hanrahan, F. P., Web, B. H., 1963. Foam spray drying methods of making readily dispersible nonfat dry milk. J. Dairy Sci. 46: 1352–1356. Borcherding, K., Lorenzen, P. C., Hoffmann, W., Schrader, K., 2008. Effect of foaming temperature and varying time/temperatureconditions of pre-heating on the foaming properties of skimmed milk. Int. Dairy J. 18 (4): 349–358. Campbell, C. H., 1917. Drying milk. Patent No. 1250427, USA. Campbell, G. M., Mougeot, E., 1999. Creation and characterization of aerated food products. Trends Food Sci. Technol. 10: 283–296. Chino, Y., Dunand, D. C., 2008. Directionally freeze-cast titanium foam with aligned, elongated pores. Acta Mater. 56: 105–113. Cooke, R. D., Breag, G. R., Ferber, C. E. M., Best, P. R., Jones, J., 1976. Studies of mango processing, Part 1: The foam-mat drying of

mango (Alphonso cultivar) puree. J. Food Technol. 11: 463–473. Crosby, E. J., Weyl, R. W., 1977. Foam-spray drying: General principles. AIChE Symp. Ser. 73(163): 82–93. Dattatreya, M. K., Wilson, R. A., Kaur, S., 2010. Determination of biochemical properties of foam mat dried mango powder. Int. J. Food Sci. Technol. 45(8): 1626–1632. Deng, M., Qian Zhou, Q., Du, A., vanKasteren, J., Wang, Y., 2009. Preparation of nanoporous cellulose foams from celluloseionic liquid solutions. Mater. Lett. 63: 1851– 1854. Ekserova, D. R., Kruglyakov, P. M., 1998. Foam and foam films: Theory, experiment, application. Elsevier, Amsterdam, The Netherlands. Farajzadeh, R., Muruganathan, R. M., Rossen, W. R., Krastev, R., 2011. Effect of gas type on foam film permeability and its implications for foam flow in porous media. Adv. Colloid Interface Sci. 168: 71–78. Frey, D. D., King, C. J., 1986. Experimental and theoretical investigation of foam-spray drying, Part 2: Experimental investigation of volatiles loss during foam-spray drying. Ind. Eng. Chem. Fundam. 25: 730–735. Fukasawa, T., Deng, Z.-Y., Ando, M., Ohji, T., Goto, Y., 2001. Pore structure of porous ceramics synthesized from water-based

References slurry by freeze-dry process. J. Mater. Sci. 36: 2523–2527. Greenwald, C. G., King, C. J., 1981. The effects of design and operating conditions on particle morphology for spray dried food. J. Food Process Eng. 4(3): 171–187. Hanrahan, F. P., Webb, B. H., 1961. USDA develops foam-spray drying. Food Eng. 31(8): 33–37. Hanrahan, F. P., Tamsma, A., Fox, K. K., Pallansch, M. Y., 1962. Production and properties of spray dried whole milk. J. Dairy Sci. 45: 27–31. Hart, M. R., Graham, R. P., Ginnette, L. E., Morgan, A. I., 1963. Foams for foam-mat drying. Food Technol. 17(10): 90–92. Hassenkloever, E., Eggers, R., 2010. High pressure spray drying of gas loaded suspensions. Proceedings of 17th International Drying Symposium (IDS2010), Magdeburg, Vol. C, 1659–1666. Jaya, S., Das, H., 2004. Effect of maltodextrin, glycerol monostearate and tricalcium phosphate on vacuum dried mango powder properties. J. Food Eng. 63: 125–134. Kadam, D. M., Wilson, R. A., Kaur, S., Manisha, 2012. Influence of foam mat drying on quality of tomato powder. Int. J. Food Prop. 15(1): 211–220. Karim, A. A., Wai, C. C., 1999. Foam-mat drying of starfruit (Averrhoa carambola L.) puree: Stability and air drying characteristics. Food Chem. 64: 337–343. Kim, H.-W., Knowles, J. C., Kim, H.-E., 2004. Hydroxyapatite and gelatin composite foams processed via novel freeze-drying and crosslinking for use as temporary hard tissue scaffolds. J. Biomed. Mater. Res. 72A: 136–145. Krasaekoopt, W., Bhatia, S., 2012. Production of yogurt powder using foam-mat drying. Assumption University J. Technol. 15(3): 166–171. Kudra, T., Ratti, C., 2006. Foam-mat drying: Energy and cost analyses. Can. Biosyst. Eng. 48(3): 27–32. Kudra, T., Ratti, C., 2008. Process and energy optimization in drying of foamed materials. Transactions of the TSTU 14(4): 812–819. Lee, J. P., Lee, K. H., Song, H. K., 1997. Manufacture of biodegradable packaging foams from agar by freeze-drying. J. Mater. Sci. 32: 5825–5832.

Lewandowski, A., Czy_zewski, M., Zbicinski, I., 2012. Morphology and microencapsulation efficiency of foamed spray-dried sunflower oil. Chem. Process Eng. 33(1): 95–102. Lewicki, P. P., 1975. Mechanisms concerned in foam-mat drying of tomato paste. Trans. Agricultural Academy in Warsaw 55: 1–67 (in Polish). Lewicki, P. P., 2006. Design of hot air drying for better foods. Trends Food Sci. Technol. 17: 153–163. Loa, Y. W., Weia, W. C. J., Hsueha, C. H., 2011. Low thermal conductivity of porous Al2O3 foams for SOFC insulation. Mater. Chem. Phys., 129: 326–330. Malysa, K., Lunkenheimer, K., 2008. Foams under dynamic conditions. Curr. Opin. Colloid Interface Sci. 13: 150–162. Marques, L. G., Silveira, A. M., Freire, J. T., 2006. Freeze-drying characteristics of tropical fruits. Drying Technol. 24: 457–463. Morgan, A. I., Ginnette, L. F., Randall, J. M., Graham, R. P., 1959. Technique for improving instants. J. Food Eng. 31: 86–87. Moy, J. H., 1971. Vacuum-puff freeze drying of tropical fruit juices. J. Food Sci. 36: 906–910. Muthukumaran, A., Ratti, C., Raghavan, V. G. S., 2008. Foam-mat freeze drying of egg white – mathematical modeling, Part 2: Freeze drying and modeling. Drying Technol. 26: 513–518. Pisal, S., Wawde, G., Salvankar, S., Lade, S., Kadam, S., 2006. Vacuum foam drying for preservation of LaSota virus: Effect of additives. AAPS PharmSciTech 7(3): Art. 60. Philips, L. G., Haque, Z., Kinsella, J. E., 1987. A method for the measurement of foam formation and stability. J. Food Sci. 521(4): 1074–1077. Pradhan, M., Bhargava, P., 2008. Defect and microstructural evolution during drying of soapnut-based alumina foams. J. Eur. Ceram. Soc. 28: 3049–3057. Prud’Homme, R. K., Khan, S. A., 1996. Foams: Theory, measurements, and applications. Marcel Dekker, New York, USA. Rabaeva, J., Zbicinski, I., 2010. Gas-admixing foam spray drying of skin-forming and porous materials. Chem. Eng. Equip. 49(4): 62–63. Rabaeva, J., 2012. Kinetics of foam-spray drying process. Diss., Technical University of Lodz, Poland.

189

190

6 Drying of Foamed Materials Raharitsifa, N., Ratti, C., 2010. Foam-mat freeze-drying of apple juice, Part 1: Experimental data and ANN simulations. J. Food Process Eng. 33: 268–283. Rajkumar, P., Kailappan, R., Viswanathan, R., Raghavan, G. S. V., Ratti, C., 2007. Foam-mat drying of alphonso mango pulp. Drying Technol. 25(2): 357–365. Ratti, C., Kudra, T., 2006. Drying of foamed biological materials: Opportunities and challenges. Drying Technol. 24: 1101–1108. Sankat, C. K., Castaigne, F., 2004. Foaming and drying behavior of bananas. Lebensm.-Wiss. Technol. 37: 518–525. Schoonman, A., Mayor, G., Dillmann, M.-L., Bisperink, C., Ubbink, J., 2001. The microstructure of foamed maltodextrin/ sodium caseinate powders: A comparative study by microscopy and physical techniques. Food Res. Int. 34: 913–929. Sinnamon, H. I., Aceto, N. C., Eskew, R. K., Schoppet, E. F., 1957. Dry whole milk, Part 1: A new physical form. J. Dairy Sci. 40: 1036. Sundaram, J., Durance, T. D., 2008. Mechanical and structural characteristics of vacuum microwave dried biopolymer foams. Journal IChE C3(85): 264–272. Tamsma, A. E., Kurtz, F., Pallansch, M. J., 1971. Comparison of flavor and physical properties of foam spray-dried whole milks prepared from concentrates foamed with air and with nitrogen. J. Dairy Sci. 56(2): 161–163. Teunou, E., Vasseur, J., Krawczyk, M., 1995. Measurement and interpretation of bulk solids angle of repose for industrial process design. Powder Handling Process. 7(3): 219–227.

Thongprachan, N., Nakagawa, K., Sano, N., Charinpanitkul, T., Tanthapanichakoon, W., 2008. Preparation of macroporous solid foam from multi-walled carbon nanotubes by freeze-drying technique. Mater. Chem. Phys. 112: 262–269. Thuwapanichayanan, R., Prachayawarakorn, S., Soponronnarit, S., 2008. Modeling of diffusion with shrinkage and quality investigation of banana foam-mat drying. Drying Technol. 25(11): 1326–1333. Thuwapanichayanan, R., Prachayawarakorn, S., Soponronnarit, S., 2012. Effects of foaming agents and foam density on drying characteristics and textural property of banana foams. Food Sci. Technol. 47: 348– 357. Verhey, J. G. P., 1973. Vacuole formation in spray powder particles: Atomization and droplet drying. Neth. Milk Dairy J. 27: 3–16. Vernon-Cartera, E. J., Espinosa-Paredesa, G., Beristainb, C. I., Romero-Tehuitzila, H., 2001. Effect of foaming agents on the stability, rheological properties, drying kinetics and flavour retention of tamarind foam-mats. Food Res. Int. 34: 587–598. Whang, K., Thomas, C. H., Healy, K. E., Nuber, G., 1995. A novel method to fabricate bioabsorbable scaffolds. Polymer 36(4): 837–842. Zbicinski, I., Rabaeva, J., 2010. Analysis of gasadmixing foam spray drying process. Drying Technol. 28(1): 103–110. Zuniga, R. N., Aguilera, J. M., 2008. Aerated food gels: Fabrication and potential applications. Trends Food Sci. Technol. 19: 176–187.

191

7 Process-Induced Minimization of Mass Transfer Barriers for Improved Drying Henry J€ager, Katharina Sch€ossler, and Dietrich Knorr 7.1 Introduction

The drying of solid food matrices depends heavily on diffusion and mass transfer through the cells and tissue. The presence of intact cell membranes in the food raw material limits mass transfer processes due to their barrier function. Furthermore, the tissue structure and the network of intercellular air spaces affect mass transfer during drying. The application of an external pulsed electric field (PEF), the induced local structural changes of the cell membrane, and also the increase in permeability due to the appearance of permanent membrane pores were found to positively affect any subsequent drying processes of porous plant materials. PEF provides an alternative to the mechanical, thermal or enzymatic cell disintegration of plant tissue, providing a short-term (milliseconds), low-energy pretreatment (95%) linear relationships between diffusivity or mass transfer coefficient and the ultrasonic power density applied, in both cases achieving correlation coefficients of over 0.98. Such a linear relationship has also been identified for other products, namely potatoes (Ozuna et al., 2011a) and carrots (García-P
erez et al., 2008). Hence, the highest applied ultrasonic power produces the most violent expansions and contractions of the soft solid material and the highest level of internal mechanical stress, thus maximizing the reduction of internal mass transfer resistance. Moreover, microstirring and the generation of microcurrents at the interfaces are more intense and so produce a greater decrease in external resistance. 8.4.1.3 Influence of the Characteristics of the Medium on Ultrasonic Intensity The noncontact applications of ultrasound involve the transmission of acoustic waves across a medium, from the emitter to the product being treated. Due to the fact that ultrasound is composed of mechanical waves, this medium could affect the acoustic field. Media properties such as viscosity, density or surface tension could improve or make it difficult to transmit vibration and, therefore, to extend the ultrasound effects. Figure 8.6 shows the acoustic pressure produced by a sound probe system in saturated brine and in a 30  Brix sucrose solution (C
arcel, 2003). As can be observed, when operating under identical conditions, the same ultrasonic system produces a differing acoustic pressure in brine than in a sucrose solution. This means that, even when the same level of electric power is applied to the transducer, the effectively applied ultrasonic energy is different in brine than in sugar solution and (as shown in the previous section) the magnitude of ultrasonic effects will also be different. This fact confirms the suggestions made in Section 8.4.1.1, namely that taking into account only the electrical power supplied/ consumed by transducers constitutes an incomplete approach to the ultrasonic field applied. However, it is not only the physical properties of the medium that have a significant influence on the ultrasonic field applied, but also the level of agitation. Ultrasonic waves are orderly vibrations that are transmitted by interaction between molecules. In the case of agitated fluids, these molecules are moving randomly due to the action of the global fluid movement. This fact hinders the tidy interaction between molecules that is needed to transmit the ultrasonic wave, and transmission of the acoustic wave is thus disrupted. In this sense, C
arcel (2003) carried out

8.4 Influence of the Main Process Variables on Drying Intensification by Ultrasound

1.1

Acoustic pressure (bar)

1 0.9 0.8 0.7 0.6 0.5

Brine Sucrose solution

0.4 0

20

40

60

80

100

120

Percentage of electric power applied to tranducer Fig. 8.6 Acoustic pressure measurements carried out with a hydrophone in 800 ml of saturated brine and a sucrose solution (30  Brix). Ultrasound was applied with a

probe system (probe diameter 13 mm at 15 mm from the emitter surface) supplying different percentages of the total electrical power of the equipment (100 W).

ultrasonic field measurements in an ultrasonic bath filled with distilled water. As can be seen in Fig. 8.7, from the measurements carried out without agitation of the bath water it was possible both to identify the stationary field, produced inside the bath by the reflection of ultrasonic waves in the water–air interface, and to locate maximum and minimum acoustic pressure zones. When the water was strongly agitated, the acoustic pressure measured was significantly reduced (to less than one-half), and the differences between maximum and minimum acoustic pressure zones were diminished. It was apparent that the turbulence produced by the water agitation had disturbed the ultrasonic field, in turn reducing the effective ultrasonic energy applied and, therefore, its ability to affect mass transfer processes. The reduction in the ultrasonic field produced by the presence of turbulences occurs not only in liquid but also in gas media. In this sense, García-P
erez et al. (2006a) observed that, during the ultrasonic (30 kW m 3) air drying (40  C) of carrot cubes at air velocities in the range 0.6 to 10 m s 1, the effects of ultrasound on drying kinetics depended on the air velocity. At low air velocities (0.6 m s 1), ultrasound application increased the drying rate, but at higher air velocities (>2 m s 1) the effect of ultrasound on drying kinetics was diminished and even became negligible. García-P
erez et al. (2007) used a diffusion model to quantify the influence of air velocity on ultrasound-mediated effects on the drying rates of carrot, persimmon and lemon peel. In the model used, the external resistance was considered negligible, and therefore the identified effective diffusivity was a lumped parameter that included all effects on both the external and internal

259

8 Drying Assisted by Power Ultrasound

20 Without agitation

Distance from bottom (cm)

260

With agitation Water - air interface

15

10

5

0 0

1

2

3

Acoustic pressure (bar) Fig. 8.7 Variation of acoustic pressure inside an ultrasonic bath (Fungsonics mod. 28 L, 20 kHz; Fungilab S.A., Barcelona, Spain) containing distilled water (28 l). Measurements were carried out with and without water agitation at different distances from the bottom of the bath.

resistances (Mulet, 1994). The conventionally air-dried samples showed that the identified effective moisture diffusivities increased with the increasing air flow rate. However, this influence was only observed up to a threshold value close to 5 m s 1; below this level the process was mainly governed by external resistance. The decrease in external resistance, due to higher air velocities, involved an increase in the effective diffusivity. Above the threshold, water removal was controlled by internal resistance and, for this reason, the effective moisture diffusivity was independent of the air velocity and considered to be a material characteristic. The application of ultrasound increased the identified effective moisture diffusivities, but only at low air velocities. As the improvement observed at high air velocities was negligible, it appears that the influence of ultrasound on drying kinetics disappears when high air velocities are applied. C
arcel et al. (2007a) modeled the drying kinetics of persimmon using two diffusion models taking into account (or not) the influence of external resistance to drying. At low air velocities the effective diffusivity and mass transfer coefficient identified from ultrasonically assisted drying were significantly higher (confidence level >95%) than in conventional drying. However, these differences were decreased for both diffusivity and mass transfer coefficient when the air velocity was increased, and became negligible when it was above 4 m s 1. In order to identify the influence of air velocity on the ultrasonic field and, therefore, on the ultrasound effects, the acoustic field inside the drying chamber was measured using a microphone (Riera et al., 2011). When measurements were taken without an air flow, an average sound pressure level of 154.3 dB was obtained for an electrical power applied to the transducer of 75 W (30 kW m 3). However,

8.4 Influence of the Main Process Variables on Drying Intensification by Ultrasound

Fig. 8.8 Influence of air velocity on the sound pressure level measured inside a cylindrical transducer.

when the air velocity was increased the sound pressure level decreased, and remained almost constant at above 8 m s 1 (Fig. 8.8). This confirmed the fact that turbulences generated by high air velocities disturb the ultrasonic field. The reduction in the actual ultrasonic energy applied to the samples was at least partly responsible for the observed decrease in ultrasound effects and, as a consequence, the application of ultrasound is seen to be more efficient at low air velocities. Currently, certain products must be dried using a low air velocity in order to avoid the appearance of various technical problems, such as casehardening. Moreover, most industrial dryers operate at low air velocities of around 2 m s 1; in these cases, the application of ultrasound could represent a very promising technology, by means of which the kinetics can be intensified while the product quality would remain unaffected. Another important aspect to be considered is the relationship between the applied ultrasonic energy and the mass load. Clearly, when the same level of acoustic energy is applied to dry a large amount of product, the acoustic energy available per unit mass is lower than when a small amount of product is dried, and the capacity of ultrasound to induce effects is also lower. Nevertheless, because the observed effects have thresholds, there is no linear relationship. For this reason, it is important to identify the optimal ratio between the level of applied ultrasonic energy and the mass load of the product treated in order to achieve the most efficient ultrasound application – that is, the greatest effect and, therefore, the lowest energy consumption. C
arcel et al. (2011a) studied this issue during the drying of carrot cubes (40  C and 1 m s 1) both with (75 W,

261

262

8 Drying Assisted by Power Ultrasound

Fig. 8.9 Identified mass transfer coefficients for carrot drying with (75 W, 21.7 kHz) and without ultrasound applications at different mass load densities. From C
arcel et al. (2011a).

21.7 kHz) and without power ultrasound application at several mass load densities (12 to 120 kg m 3). The data were modeled using a diffusion model that took the external resistance into account. In experiments performed without ultrasound application, the increase in mass load density did not affect the effective moisture diffusivity, but did lead to a significant reduction in the mass transfer coefficient. The increase in mass load in the drying chamber created preferential pathways of air flow, which increased the external mass transfer resistance. The application of ultrasound reduced the drying time for all of the mass load densities tested, although increases in the mass load density (over the range studied) did not have any significant effect on the identified effective diffusivities; however, the mass transfer coefficient was significantly reduced (Fig. 8.9). At the highest mass load densities tested (108 and 120 kg m 3), the differences between the mass transfer coefficients of ultrasonically assisted and conventional drying disappeared. Apparently, the high mass load density prevents the ultrasound from reaching the entire sample surface at an appropriate intensity. Thus, the ultrasound effects on the solid–air interface did not evenly cover the whole solid surface, and consequently the global average effect was decreased. To summarize, it can be stated that the main aspect for consideration when applying ultrasound to intensify drying is not the ultrasonic energy generated, but rather the ultrasonic energy received by the samples. Hence, the higher the energy received the greater the influence of ultrasound on mass transport kinetics. In

8.4 Influence of the Main Process Variables on Drying Intensification by Ultrasound

order to increase the amount of received energy, ultrasound application systems must achieve a good coupling between emitters and the treated medium. The attenuation of ultrasonic energy produced by the medium must be minimized, for instance by reducing the turbulence. Finally, an optimal ratio of energy received to mass load treated will allow the greatest ultrasound effects to be achieved with a minimum of energy consumption. 8.4.2 Drying Air Temperature

As in most chemical and biochemical reactions, the temperature can affect the drying kinetics by affecting the moisture diffusivity (Simal et al., 2005). This tendency is usually described by using an Arrhenius-type equation: ! ~a E De ¼ D0 exp ; ð8:5Þ ~ RT ~ a is where De is the effective diffusion coefficient, D0 is the Arrhenius prefactor, E ~ the molar activation energy for moisture diffusion, R is the universal gas constant, and T is the temperature. In overall terms, an increase in temperature accelerates the reactions due to a greater amount of energy being available in the medium, which provides the activation energy needed to “extract” a molecule of water from the product matrix. The application of ultrasound during drying implies the introduction of an additional energy source that could also overcome the barrier of activation energy and make it easy to extract water. Thus, it is expected that ultrasound will have a greater influence on drying when less energy is available in the systems; that is, when drying takes place at low drying air temperature. In contrast, when the drying air provides enough energy – that is, when hot drying air is used – the application of ultrasound has less of an effect, or even no effect at all. Gallego-Ju
arez (1998), when studying the airborne ultrasound application (155 dB) with a vibrating plate to dry carrots at different temperatures (60, 90 and 115  C), noted that the effect of the ultrasonic radiation was significant at low air temperatures (60  C) but was diminished as the temperature increased. At the highest temperature tested (115  C), the ultrasonic effect was negligible and there were no appreciable differences in the experimental drying curves. In a similar study, García-P
erez et al. (2006b) dried carrot cubes (1 m s 1) in a cylindrical vibrating chamber (García-P
erez et al., 2006a) at five different temperatures (30, 40, 50, 60, and 70  C), and then used a diffusion model that takes the external resistance to water transport into account to evaluate how temperature can influence the extent of the ultrasound-mediated effects on drying kinetics. The results showed that the effective diffusivity and mass transfer coefficient were both increased as the air temperature rose; this meant that both the internal and external mass transfer resistances had been reduced due to the increase in drying temperature. The influence of power ultrasound (30.8 kW m 3 and 21.7 kHz) on

263

264

8 Drying Assisted by Power Ultrasound

Fig. 8.10 Influence of drying temperature on the identified effective diffusivity for carrot drying carried out with and without ultrasound application. From García-P
erez et al. (2006b).

both kinetic parameters was significant at low temperatures but negligible at the highest temperature tested (70  C). As can be observed in Fig. 8.10, in the case of conventional drying, the dependence of effective diffusivity fitted an Arrhenius-type equation adequately, the activation energy being a characteristic of the product. For ultrasonically assisted dried samples an Arrhenius behavior was observed at low temperatures; however, when the temperature reached 70  C the identified diffusivity was quite similar for both methods of drying, with and without ultrasound application. These facts confirmed that, at low temperatures, ultrasound would provide mechanical energy that complements the thermal energy provided by air temperature and, jointly, these would contribute to the water transport. At high air temperatures, the thermal energy is higher and masks the effect linked to the mechanical energy provided by ultrasound. Therefore, the application of ultrasound as a means of intensifying air drying processes is more efficient when a low temperature is used. Both, the drying time and drying air temperature determine energy consumption, and the application of ultrasound can affect both aspects so as to contribute to a reduction in the energy needs of drying. García-P
erez et al. (2012a) used a digital power meter to measure the total energy consumption during ultrasonically assisted drying experiments on orange peel (40  C; 1 m s 1; 0, 18.5, 36.7 kW m 3). The main elements considered for the quantification of energy consumption were the heating elements, the ventilation system (fan), and the ultrasound generator. By applying ultrasound, the drying time was reduced by approximately 30% when the experiments were carried out using an ultrasonic power density of 18.5 kW m 3,

8.4 Influence of the Main Process Variables on Drying Intensification by Ultrasound

and by 45% for experiments carried out at 36.7 kW m 3. This shortening of drying times led to a fall in the total measured energy consumption of 12% at 18.5 kW m 3, and of 20% at 36.7 kW m 3 as compared to the conventional drying process. It should be noted that, although these measurements were obtained for a specific dryer, they illustrate the energy-saving possibilities of applying power ultrasound in order to intensify a drying process. As ultrasound application is more effective at low temperatures, this could also represent an interesting approach to intensifying other drying processes that occur at temperatures below or close to the freezing point, such as vacuum freeze-drying or atmospheric freeze-drying. For these drying methods, evaporation or sublimation occurs depending on the drying temperature and solvent involved, but long drying times are required in every case. In this context, attempts have been made to apply ultrasound to vacuum freezedrying processes (Moy and DiMarco, 1972), but with poor results. As ultrasonic waves are mechanical and require a material medium for their propagation, their application under vacuum conditions is difficult. Indeed, in a vacuum only contact ultrasound should be considered. When studying atmospheric freeze-drying, Moy and DiMarco (1970) used a steam-jet whistle operating at frequencies between 10.8 and 12.2 kHz, which is within the human hearing range. Experiments conducted with distilled water, coffee and tea extracts at temperatures of 15 and 26  C revealed an average drying rate increase of between 10% and 100%. However, the intense noise generated by this system would likely be a serious hindrance to its industrial application. Further efforts have been made in this field recently, however, with Bantle and Eikevik (2011) using a commercial transducer (20 kHz; DN 20/2000, Sonotronic) to test the influence of process variables such as temperature and applied power. In these studies, it was shown that ultrasonic application would lead to a maximum reduction of about 10% in the drying time for green peas at 3  C, but that such improvement was almost negligible if the natural variability of raw matter was taken into consideration. A new system by means of which ultrasound can be applied in order to intensify mass transport in low-temperature drying processes has recently been developed (C
arcel et al., 2011b). The application of power ultrasound during atmospheric freeze-drying ( 14  C, 2 m s 1, 0 and 19.5 kW m 3, 21.9 kHz) greatly speeded up the drying kinetics of carrot, apple and eggplant (García-P
erez et al., 2012b), and involved an average reduction in drying time of between 65% and 70%. An analysis of experimental data with a diffusion model that takes external resistance into account showed that ultrasound had influenced both the internal and external mass transfer resistances. Thus, the application of power ultrasound had increased the effective diffusivity (Fig. 8.11) in the range of 407% for eggplant and 428% for apple. With regards to the mass transfer coefficient, the increase was 96% for apple, 152% for carrot, and 170% for eggplant. For all of these products such increases in kinetic parameters were significant at a confidence level of 99%. Therefore, the intensifying effect of ultrasound application on effective diffusivity is greater at low temperatures than at high temperatures, above 0  C. In the case of mass transfer coefficients, the increase produced by ultrasound was similar to that

265

8 Drying Assisted by Power Ultrasound 30

10

Eggplant

Carrot

Apple

25

4

De (10-11 m2/s)

De (10-11 m2/s)

5

3 2 1

De (10-11 m2/s)

266

20 15 10

AIR+US

4

0

0 AIR

6

2

5

0

8

AIR

AIR+US

AIR

AIR+US

Fig. 8.11 Identified effective diffusivities of atmospheric freeze drying ( 14  C, 2 m s 1) of carrot, eggplant and apple, with (AIR þ US) and without (AIR) ultrasound application (19.5 kW m 3, 21.9 kHz). The bars indicate intervals with a confidence level of 99%.

reported for conventional hot-air drying (Gallego-Ju
arez et al., 2007; Puig et al., 2012). It is important to note here that ultrasound effects are more pronounced in regions where the resistance to mass transfer is higher. During freeze-drying, the frozen inner core of the product is shrinking, while a sublimation front is moving back; this generates a porous structure zone between the surface of the material and the front where the internal resistance to mass transfer is located. The effects produced by the application of ultrasound in this zone lead to a significant reduction of this resistance, as observed in the marked increase in identified effective diffusivity. The application of ultrasound at low temperatures is useful not only as a means of intensifying the drying process by sublimation, but also for eliminating other solvents by evaporation. In this context, García-P
erez et al. (2012b) monitored the influence of ultrasound in the removal of ethanol from impregnated apple samples at 14  C (previously freeze-dried). The results obtained showed that the time needed to remove all the ethanol fell by 55%, from 150 min for air-drying to 67.5 min for the ultrasonically assisted process. These results demonstrate the potential of this technology for application in the chemical, cosmetic and pharmaceutical industries for the low-temperature removal of organic solvents in order to preserve quality aspects at lower operating and fixed costs than in freeze-drying. 8.4.3 Ultrasound–Sample Interaction

An important aspect to take into consideration when dealing with the application of high-intensity ultrasound as a means of intensifying a drying process is the type of material to be dried and how it might interact with ultrasonic waves. The interaction of ultrasonic waves with different materials could affect the ultrasound efficiency and therefore, a different influence on mass transport might be exerted. Mulet et al. (2003) compared three solid–liquid dehydration treatments (the osmotic dehydration of apple, the brining of cheese, and the brining of pork meat)

8.4 Influence of the Main Process Variables on Drying Intensification by Ultrasound

carried out with ultrasound application under similar conditions using a commercial ultrasonic bath. The intensification of mass transport produced by ultrasound was higher in apple than in cheese curds, but no significant ultrasonic effects were noted for meat brining. These different behaviors may relate to the different structures of the products being examined. For example, in apple the structure allows a good transmission of ultrasonic waves inside the product, while the large number of air-filled pores present allow air to flow more easily from the pores to the surrounding solution, and also for the solution to flow inside the pores due to successive compressions and expansions being provoked by ultrasound. In contrast, the “rubber” structure of pressed cheese curds will absorb the ultrasound energy, limiting the transmission of internal vibration; the liquid-filled pores would also complicate mass transfer with the solution, so that the influence of ultrasound would be less than in apples. Finally, at the power level tested, the fibrous structure of pork loin and the absence of pores within the tissues appeared to prevent some of the ultrasonic effects and their influence on mass transfer. When the influence of the intensity of applied ultrasound on ultrasound effects was addressed (see Section 8.4.1), the existence was proposed of an ultrasonic intensity threshold above which the influence of ultrasound on mass transfer would become significant. However, the value of this threshold would depend on the material involved. Subsequently, García-P
erez et al. (2009) showed that the influence of power ultrasound on the effective moisture diffusivity in drying carrot showed a threshold of about 15 kW m 3. Nevertheless, due mainly to a wide variability in the raw materials, the difference between the means of ultrasonic and non-ultrasonic treatments for the experiments carried out was not significant (level of confidence >95%) until the acoustic power density in the drying chamber exceeded a value of 20.5 kW m 3 (Fig. 8.12a). This means that, in order to draw effective conclusions from these investigations, a sufficiently large number of experiments must be carried out. Under the same conditions, the effects of ultrasound on the moisture effective diffusivity of lemon peel failed to demonstrate a threshold value, though the mean differences became significant at 12 kW m 3 (Fig. 8.12b). Under similar drying conditions, Ozuna et al. (2010) showed the mean identified potato moisture diffusivity to be statistically different from the not ultrasonically assisted drying (level of confidence >95%) when the acoustic power density applied was higher than 25 kW m 3, while García-P
erez et al. (2011) reported a value of 12 kW m 3 for eggplant. The influence of the applied ultrasonic power density on mass transfer coefficient was similar to that observed for effective diffusivity. This means that a minimum level of ultrasonic density must be applied for significant effects to occur, and that this threshold will differ for products with different structures (García-P
erez et al., 2008, 2011). The realization that the minimum level of acoustic energy needed in order to observe significant ultrasonic effects on drying depends on the material implies that, under adverse conditions for the transmission of ultrasound (e.g., high air velocities), the material could be prone, or not, to drying intensification by ultrasound. In this sense, based on the modeling of experimental kinetics for carrot, persimmon and lemon peel drying at different air velocities, García-P
erez

267

268

(a)

8 Drying Assisted by Power Ultrasound

(b)

1.8

13

Carrot

Lemon peel 11 De (10-10 m2/s)

De (10-10 m2/s)

1.6

1.4

1.2

9 7 5 3

1

1 0

20

40

60

80

100

Ultrasonic power density (kW m-3)

0

20

40

60

80

100

Ultrasonic power density (kW m-3)

Fig. 8.12 Influence of applied ultrasonic power density on the identified effective diffusivity for the drying of (a) carrot and (b) lemon peel. The bars indicate intervals with a confidence level of 95% (García-P
erez et al., 2009).

et al. (2007) noted that the effects of ultrasound on the identified effective diffusivity were significant for all three products at air velocities under 6 m s 1. Above this air velocity, the intensification of ultrasound was negligible for carrot and persimmon, but in the case of lemon peel it was significant up to the highest air velocity tested, of 14 m s 1. This means that, in the case of carrot and persimmon, the disturbing effect that a high air velocity exerts over the acoustic field reduced the energy present in the medium to a level that would not affect the mass transfer process. However, in the case of lemon peel, despite a reduction in acoustic energy, the acoustic level was still able to intensify the drying, confirming that the lemon peel is more sensitive to ultrasonic effects. The level of drying intensification produced by ultrasound also depends on the material treated, this being higher in products with a lower acoustic power density threshold. Table 8.1 shows the identified effective diffusivity for different products dried under similar conditions, with and without ultrasound application. The increase in effective diffusivity varies depending on the product being considered; thus, the application of an acoustic power density of 37 kW m 3 will produce a diffusivity increase of about 40% in drying carrot, and about 200% in the case of eggplant. The same behavior was observed in the mass transfer coefficient (Tab. 8.1); for example, the mass transfer coefficient for carrot was 33% higher when ultrasound was applied, whereas for eggplant the increase reached 230%. These results indicate that those products that present a more compact internal structure (e.g., carrot, persimmon, potato) are less sensitive to the effects of ultrasound on the drying rate. In contrast, the more porous products (e.g., eggplant,

8.4 Influence of the Main Process Variables on Drying Intensification by Ultrasound Tab. 8.1 Identified effective diffusivity (De) and mass transfer coefficient (k) of different products dried (40  C and 1 m s 1) with ultrasound (US; 37 kW m 3, 21 kHz) and without ultrasound (WUS) application. Incr. (%) is the increase for ultrasonic-assisted (US) compared to conventional (WUS) drying values.

Product

De
1010 (m2 s 1) WUS

US

Carrot

2.0  0.1

2.8  0.3

Potato

4.6  0.1

Orange peel Eggplant

Incr. (%)

k
104 (kg H2O m

2

s 1)

Incr. (%)

Reference

GarcíaP
erez et al. (2007) Ozuna et al., (2011a) GarcíaP
erez et al. (2012a) GarcíaP
erez et al. (2011)

WUS

US

39

4.6  0.1

6.6  0.5

33.3

7.5  0.3

64

2.0  0.4

3.2  0.4

58.1

40.4  4.4

61.3  5.3

52

11.7  0.5

24.3  0.2

107.7

8.9  1.3

27.9  3.0

212

1.9  0.4

6.2  0.9

229.4

lemon peel) are more prone to mass transfer intensification by ultrasound. Thus, the porosity could be a key parameter with which to define the extent of the effects of ultrasound on mass transport (C
arcel et al., 2011a). If the product is highly porous, the transmission of vibrations through the solid matrix would be improved such that the ultrasound could affect the inner part of the samples. Moreover, this structure would facilitate the “sponge effect” because it exhibits a low structural resistance to the mechanical stress. With regards to the influence on the mass transfer coefficient, an external porous layer is more prone to ultrasound effects; hence, microstirring at interfaces, pressure variations, oscillating velocities or the formation of microcurrents can occur inside the external pores, thereby increasing the effective surface area for mass transfer. Ozuna et al. (2011b) were the first to attempt to relate the internal structure of the product with the extent of ultrasound influence on the drying rate. These authors studied the evolution of the effective moisture diffusivity versus the ultrasonic power applied during drying for five very different products, namely carrot, potato, lemon peel, orange peel, and eggplant (Fig. 8.13). As noted in Section 8.4.1.2, above an ultrasonic intensity threshold there is a linear relationship between the ultrasonic power level and the effective moisture diffusivity. For carrots and potatoes, due to the figure scale, this threshold is not highlighted, as in Fig. 8.12. The slope of the linear relationship may be used to estimate the efficacy of the ultrasonic application in that, the steeper the slope the more intense the mechanical effects in the material. These slopes were correlated with instrumental textural measurements of the studied products, thereby identifying a linear relationship between slope and hardness (Fig. 8.14): the higher the hardness of the tested products, the lower the slope.

269

270

8 Drying Assisted by Power Ultrasound

Fig. 8.13 Linear relationship between effective diffusivity and applied ultrasonic power density for the drying of different products.

Fig. 8.14 Slope of the linear relationship between effective diffusivity and applied ultrasonic power for the drying of different products, versus instrumentally measured hardness.

8.4 Influence of the Main Process Variables on Drying Intensification by Ultrasound

Consequently, the mechanical compressions and expansions (“sponge effect”) produced by ultrasound in these materials are more intense in soft products, and this leads to a more effective water removal. In the case of the “chewiness” parameter, a similar behavior is observed. However, other determined textural parameters (cohesiveness, springiness, adhesiveness and resilience) failed to correlate well with the effect of ultrasound. A different situation occurs when ultrasound is applied to frozen samples, for example to intensify the atmospheric freeze-drying process. In this case, the structure and microstructure of the products is changed by freezing. During drying, the loss of moisture by sublimation generates two distinct layers in the solid, an inner frozen core and a highly porous external layer. In this way, regardless of the raw product porosity, the applied drying technique generates a very porous structure and, as a consequence, the ultrasound effects are not located in the ice core but rather in the dry outer layer in which vapor diffusion occurs. Thus, the alternating compression and expansion cycles produced by ultrasonic waves (sponge effect) will make the diffusion of vapor through this porous solid matrix much easier. The effect of ultrasound on the effective diffusivity is expected to be similar for all products, regardless of their original porous structure. Consequently, when García-P
erez et al. (2012b) studied the application of ultrasound during the atmospheric freeze-drying of apple, carrot and eggplant (three products with very different native structures and porosities), they failed to identify any major differences in the increases of effective diffusivities produced by ultrasound. García-P
erez et al. (2012b) also showed that the effect of ultrasound on the effective moisture diffusivity was greater (Tab. 8.2) than for air-drying at moderate temperatures (Tab. 8.1). The data in Tab. 8.1 show that, in the case of drying at 40  C, the diffusivity increase produced by ultrasound ranged from 39% for carrot to 212% for eggplant. However, in the case of atmospheric freeze-drying at 10  C the increase was within the range 407–425% for all products. It should also be considered that the acoustic power density applied in the latter case (19.5 kW m 3) was lower than that applied at higher temperatures (37 kW m 3). The high-porosity matrix generated during freezedrying in the external layer of the products causes the ultrasound effects to be

Tab. 8.2 Identified effective diffusivity (De) and mass transfer coefficient (k) of different products during atmospheric freeze drying ( 14  C and 2 m s 1) with ultrasound (US; 19.5 kW m 3, 21.9 kHz) and without ultrasound (WUS) application. Incr. (%) is the increase for ultrasonic-assisted (US) compared to conventional (WUS) drying values. Data from García-P
erez et al. (2012b).

Product

Carrot Apple Eggplant

De
1011 (m2 s 1) WUS

US

0.8  0.1 1.4  0.7 4.4  1.7

4.2  0.4 7.4  2.1 22.3  4.7

Incr. (%)

425 428 407

k
105 (kg H2O m

2

s 1)

WUS

US

3.3  1.5 4.8  0.2 23.7  4.3

8.3  2.3 9.4  0.9 64.1  10.4

Incr. (%)

152 96 170

271

272

8 Drying Assisted by Power Ultrasound

more intense than in conventionally hot air-dried products. The high efficiency of ultrasound application to increase diffusion is also linked to the greater acoustic energy absorption in high-porosity products. On the other hand, due to the generation of a porous external layer, the effects of ultrasound on external mass transfer resistance are similar to those reported for the conventional hot-air drying of highly porous materials. As noted above, the effects produced by ultrasound at interfaces may lead to an increase in the effective surface area when they occur inside the external pores of the material.

8.5 Conclusions

The application of ultrasound represents a promising means of intensifying drying systems. The nature of its effects, which mainly are mechanical, could lead to the use of milder drying conditions, thereby preserving product quality. In this respect, the effects of ultrasound are more intense when the drying conditions are more adverse to moisture transport, namely at low temperatures or low air velocities. Under such conditions, the application of ultrasound will increase the drying rate, thereby shortening the processing time. These facts – as well as the relatively small amount of energy needed to generate ultrasound effects – may help to save energy and reduce processing costs. However, in order to achieve a significant influence of ultrasound, it is necessary to generate and transmit the ultrasonic energy to the drying samples in an efficient manner. In addition, as intensification is linked to the acoustic energy that reaches the samples, it is especially important to consider all of the aspects that might reduce the acoustic field, such as the presence of turbulences, the absorption of energy, wave reflections, and mass load distribution. Finally, the internal structure and characteristics of the material to be dried might affect the extent of process intensification by ultrasound. Consequently, porosity would appear to be one of the main factors to be taken into account, as ultrasound is more effective in the most porous products. Clearly, porous structures generated in food products during atmospheric freeze-drying will provide interesting advantages for the successful application of ultrasound as a means of intensifying this process.

Acknowledgments

The authors of this chapter wish to thank the Spanish Ministry of Economy and Competitiveness for their financial support under projects DPI2012-37466-C03-01, DPI2012-37466-C03-02 and DPI2012-37466-C03-03. The authors also acknowledge the contribution of eng. Ram
on Pe~ na in developing the different drying systems of the ASPA group.

References

Additional Notation Used in Chapter 8

d f I k PA Z

distance from emitter surface wave frequency acoustic intensity mass transfer coefficient acoustic pressure impedance

m s 1 kg m4 s 3 (W m 2) kg m 2 s 1 Nm 2 kg m 2 s 1

attenuation coefficient wavelength

m m

Greek Letters

a l

1

Subscripts

a e 0

activation effective emitted, reference

Abbreviations

AFD d.m. VFD

atmospheric freeze-drying dry matter vacuum freeze-drying

References Bantle, M., Eikevik, T. M., 2011. Parametric Advances in sonochemistry, Vol. 4, (ed. T. study of high-intensity ultrasound in the J. Mason), JAI Press, London, UK, atmospheric freeze drying of peas. Drying pp. 1–73. Technol. 29: 1230–1239. C
arcel, J. A., 2003. Influencia de los Benedito, J., Carcel, J.A., Gonzalez, R., Mulet, ultrasonidos de potencia en procesos de A., 2002. Application of low intensity transferencia de materia. Diss., Universitat ultrasonics to cheese manufacturing Politecnica de Valencia, Spain. processes. Ultrasonics 40: 19–23. C
arcel, J. A., García-P
erez, J. V., Riera, Benedito, J., Mulet, A., Clemente, G., GarcíaE., Mulet, A., 2007a. Influence of high Perez, J.V., 2004. Use of ultrasonics for the intensity ultrasound on drying kinetics composition assessment of olive mill of persimmon. Drying Technol. 25: 185– wastewater (alpechin). Food Res. Int. 37: 193. 595–601. C
arcel, J. A., Benedito, J., Rossell
o, C., Berlan, J., Mason, T.J., 1996. Dosimetry for Mulet, A., 2007b. Influence of ultrasound power ultrasound and sonochemistry, in intensity on mass transfer in apple

273

274

8 Drying Assisted by Power Ultrasound immersed in a sucrose solution. J. Food Eng. 78: 472–479. C
arcel, J. A., Benedito, J., Bon, J., Mulet, A., 2007c. High intensity ultrasound effects on meat brining. Meat Sci. 76: 611–619. C
arcel, J. A., Nogueira, R. I., García-P
erez, J. V., Sanju
an, N., Riera, E., 2010. Ultrasound effects on the mass transfer processes during drying kinetic of olive leaves (Olea europea, var. Serrana). Defect Diffus. Forum 297–301: 1083–1090. C
arcel, J. A., Garcia-Perez, J. V., Riera, E., Mulet, A., 2011a. Improvement of convective drying of carrot by applying power ultrasound: Influence of mass load density. Drying Technol. 29: 174–182. C
arcel, J. A., García-P
erez, J. V., Pe~ na, R., Mulet, A., Riera, E., Acosta, V., GallegoJu
arez, J. A., 2011b. Procedimiento y dispositivo para mejorar la transferencia de materia en procesos a baja temperatura mediante el uso de ultrasonidos de elevada intensidad. International patent, Spanish ref. P201131512, internacional PCT ref. 120120283. C
arcel, J. A., García-P
erez, J. V., Benedito, J., Mulet, A., 2012. Food process innovation through new technology: Use of ultrasound. J. Food Eng. 110: 200–207. Carlin, B., 1960. Ultrasonics. McGraw-Hill, New York, USA. Chiralt, A., Fito, P., Barat, J. M., Andr
es, A., Gonz
alez-Martínez, C., Escriche, I., Camacho, M. M., 2001. Use of vacuum impregnation in food salting process. J. Food Eng. 49: 141–151. Chou, S. K., Chua, K. J., 2001. New hybrid drying technologies for heat sensitive foodstuffs. Trends Food Sci. Technol. 12: 359–369. Chua, K. J., Mujumdar, A. S., Chou, S. K., 2003. Intermittent drying of bioproducts: An overview. Bioresour. Technol. 90: 285–295. Da Mota, V. M., Palau, E., 1999. Acoustic drying of onion. Drying Technol. 17: 855–867. De la Fuente, S., Riera, E., Acosta, V. M., Blanco, A., Gallego-Ju
arez, J. A., 2006. Food drying process by power ultrasound. Ultrasonics 44: e523–e527. Fernandes, F. A. N., Rodrigues, S., 2007. Ultrasound as pre-treatment for drying of fruits: Dehydration of banana. J. Food Eng. 82: 261–267.

Fernandes, F. A. N., Linhares, F. E., Rodrigues, S., 2008. Ultrasound as pre-treatment for drying of pineapple. Ultrason. Sonochem. 15: 1049–1054. Gallego-Ju
arez, J. A., 1998. Some applications of air-borne power ultrasound to food processing, in Ultrasound in food processing, (eds. M. J. W. Povey, T. J. Mason), Chapman & Hall, London, UK, pp. 127–143. Gallego-Ju
arez, J. A., Yang, T., V
azquezMartínez, F., G
alvez-Moraleda, J. C., Rogriguez-Corral, G., 2001. Dehydration method and device, US Patent No. 6233844 B1, May 22. Gallego-Ju
arez, J. A., Riera, E., De la Fuente, S., Rodríguez-Corral, G., Acosta-Aparicio, V., Blanco, A., 2007. Application of high-power ultrasound for dehydration of vegetables: Processes and devices. Drying Technol. 25: 1893–1901. Gallego-Ju
arez, J. A., Rodríguez, G., Acosta, V., Riera, E., 2010. Power ultrasonic transducers with extensive radiators for industrial processing. Ultrason. Sonochem. 17(6): 953– 964. García-P
erez, J. V., C
arcel, J. A., De la Fuente, S., Riera, E., 2006aUltrasonic drying of foodstuff in a fluidized bed: Parametric study. Ultrasonics 44: e539–e543. García-P
erez, J. V., Rossell
o, C., C
arcel, J. A., De la Fuente, S., Mulet, A., 2006b. Effect of air temperature on convective drying assisted by high power ultrasound. Defect Diffus. Forum 258–260: 563–574. García-P
erez, J. V., C
arcel, J. A., Benedito, J., Mulet, A., 2007. Power ultrasound mass transfer enhancement in food drying. Food Bioprod. Process. 85: 247–254. García-P
erez, J. V., C
arcel, J. A., Benedito, J., Riera, E., Mulet, A., 2008. Drying of a low porosity product (carrot) as affected by power ultrasound. Defect Diffus. Forum 273–276: 764–769. García-P
erez, J. V., C
arcel, J. A., Riera, E., Mulet, A., 2009. Influence of the applied acoustic energy on the drying of carrots and lemon peel. Drying Technol. 27: 281–287. García-P
erez, J. V., Ozuna, C., Ortu~ no, C., C
arcel, J. A., Mulet, A., 2011. Modeling ultrasonically assisted convective drying of eggplant. Drying Technol. 29: 1499– 1509.

References García-P
erez, J. V., Ortu~ no, C., Puig, A., C
arcel, J. A., P
erez-Munuera, I., 2012a. Enhancement of water transport and microstructural changes induced by highintensity ultrasound application on orange peel drying. Food Bioprocess Technol. 5: 2256– 2265. García-P
erez, J. V., C
arcel, J. A., Riera, E., Rossell
o, C., Mulet, A., 2012a. Intensification of low-temperature drying by using ultrasound. Drying Technol. 30: 1199–1208. Kentish, A., Ashokkumar, M., 2011. The physical and chemical effects of ultrasound, in Ultrasound technologies for food and bioprocessing, (eds. H. Feng, G. V. BarbosaCanovas, J. Weiss), Springer, New York, USA. pp. 1–12. Khmelev, V. N., Shalunov, A. V., Barsukov, R. V., Abramenko, D. S., Lebedev, A. N., 2011. Studies of ultrasonic dehydration efficiency. J. Zhejiang Univ., Science A 2: 247–254. Kudra, T., Mujumdar, A. S., 2009. Atmospheric freeze-drying, in Advanced drying technologies, 2nd edn, CRC Press, Boca Raton, USA, pp. 327–336. Kutruff, H., 1991. Ultrasonics: Fundamentals and applications. Elsevier, London, UK. Lee, J., Tuziuti, T., Yasui, K., Kentish, S., Grieser, F., Ashokkumar, N., Lida, Y., 2007. Influence of surface-active solutes on the coalescence, clustering, and fragmentation of acoustic bubbles confined in a microspace. J. Phys. Chem. C 111: 19015– 19023. Leighton, T. G., 1998. The principles of cavitation, in Ultrasound in food processing, (eds. M. J. W. Povey, T. J. Mason), Chapman & Hall, London, UK, pp. 151–182. Lewicki, P. P., Michaluk, E., 2004. Drying of tomato pretreated with calcium. Drying Technol. 22: 1813–1827. Li, Z., Raghavan, G. S. V., Orsat, V., 2010. Optimal power control strategies in microwave drying. J. Food Eng. 99: 263–268. Lin, S., Zhang, F., 2000. Measurement of ultrasonic power and electro-acoustic efficiency of high power transducers. Ultrasonics 37: 549–554. Lombra~ na, J. I., Rodriguez, R., Ruiz, U., 2010. Microwave-drying of sliced mushroom. Analysis of temperature control and pressure. Innovative Food Sci. Emerg. Technol. 11: 652–660.

Mason, T. J., 1998. Power ultrasound in food processing: The way forward, in Ultrasound in food processing, (eds. M. J. W. Povey, T. J. Mason), Chapman & Hall, London, UK, pp. 105–126. Mason, T. J., Cordemans, E. D., 1996. Ultrasonic intensification of chemical processing and related operations: A review. Trans. Inst. Chem. Eng. 74: 511–516. Mason, T. J., Lorimer, J. P., 2002. Applied sonochemistry: The uses of power ultrasound in chemistry and processing. Wiley-VCH, Weinheim, Germany. McClements, D. J., 1997. Ultrasonic characterization of food and drinks: Principles, methods and applications. Crit. Rev. Food Sci. 37: 1–46. Moy, J. H., DiMarco, G. R., 1970. Exploring airborne sound in a nonvacuum freeze drying process. J. Food Sci. 35: 811–817. Moy, J. H., DiMarco, G. R., 1972. Freezedrying with ultrasound. Trans. ASAE 15: 373–376. Moreno, J., Simpson, R., Estrada, D., Lorenzen, S., Moraga, D., Almonacid, S., 2011. Effect of pulsed-vacuum and ohmic heating on the osmodehydration kinetics, physical properties and microstructure of apples (cv. Granny Smith). Innovative Food Sci. Emerg. Technol. 12: 562–568. Mujumdar, A. S., Devahastin, S., 2000. Fundamental principles of drying, in Mujumdar’s practical guide to industrial drying, Exergex Corp., Montreal, Canada, pp. 1–22. Mulet, A., 1994, Drying modelling and water diffusivity in carrots and potatoes. J. Food Eng. 22: 329–348. Mulet, A., C
arcel, J. A., Sanju
an, N., Bon, J., 2003. New food drying technologies: Use of ultrasound. Food Sci. Technol. Int. 9: 215– 221. Muralidhara, H. S., Ensminger, D., Putnam, A., 1985. Acoustic dewatering and drying (low and high frequency): State of the art review. Drying Technol. 3: 529–566. Ni~ noles, L., Clemente, G., Ventanas, S., Benedito, J., 2007. Quality assessment of Iberian pigs through backfat ultrasound characterization and fatty acid composition. Meat Sci. 76: 102–111. Ni~ noles, L., Mulet, A., Ventanas, S., Benedito, J., 2011. Ultrasonic characterisation of B. femoris from Iberian pigs of different

275

276

8 Drying Assisted by Power Ultrasound genetics and feeding systems. Meat Sci. 89: 174–180. Oliveira, F. I. P., Gall~ao, M. I., Rodrigues, S., Fernandes, F. A. N., 2011. Dehydration of malay apple (Syzygium malaccense L.) using ultrasound as pre-treatment. Food Bioprocess Technol. 4: 610–615. Ozuna, C., C
arcel, P.C., García-P
erez, J.V., Mulet, A., C
arcel, J. A., 2010. Influencia de la aplicaci
on de ultrasonidos de alta intensidad en el transporte de agua y de NaCl durante el salado de carne en salmueras de diferente concentraci
on. Congreso Espa~ nol de Ingeniería de Alimentos, CESIA 2010. Logro~ no, Spain, October 6–10. Ozuna, C., C
arcel, J. A., García-P
erez, J. V., Mulet, A., 2011a. Improvement of water transport mechanisms during potato drying by applying ultrasound. J. Sci. Food Agric. 91(14): 2511–2517. Ozuna, C., C
arcel, J. A., Santacatalina, J. V., Mulet, A., García-P
erez, J. V., 2011b. Textural properties of vegetables: A key parameter on ultrasonic assisted convective drying. Proceedings of the 11th International Congress of Engineering and Food, ICEF11. Athens, Greece, May 22–26. Peshkovsky, A. S., Peshovsky, S. L., 2010. Acoustic cavitation theory and equipment design principles for industrial applications of high-intensity ultrasound. Nova Science Publishers, New York, USA. Pugin, B., Turner, A. T., 1990. Influence of ultrasound on reactions with metals, in Advances of sonochemistry, (ed. T. J. Mason), JAI Press, London, UK, pp. 81–118. Puig, A., Perez-Munuera, I., C
arcel, J. A., Hernando, I., Garcia-Perez, J. V., 2012. Moisture loss kinetics and microstructural changes in eggplant (Solanum melongena L.) during conventional and ultrasonically assisted convective drying. Food Bioprod. Process. 90: 624–632. nas, P., Pag
an, R., Sala, F. J., Raso, J., Ma~ 1999. Influence of different factors on the output power transferred into medium by ultrasound. Ultrason. Sonochem. 5: 157–162. Rastogi, N. K., Eshtiaghi, M. N., Knorr, D., 1999. Accelerated mass transfer during osmotic dehydration of high intensity electrical field pulse pretreated carrots. J. Food Sci. 64: 1020–1023.

Rastogi, N. K., Raghavarao, K. S. M. S., Niranjan, K., Knorr, D., 2002. Recent developments in osmotic dehydration: Methods to enhance mass transfer. Trends Food Sci. Technol. 13: 48–59. Rawson, A., Tiwari, B. K., Tuohy, M. G., O’Donnell, C. P., Brunton, N., 2011. Effect of ultrasound and blanching pretreatments on polyacetylene and carotenoid content of hot air and freeze dried carrot discs. Ultrason. Sonochem. 18: 1172–1179. Riera, E., García-P
erez, J.V., Acosta, V.M., C
arcel, J.A., Gallego-Ju
arez, J. A., 2011. Computational study of ultrasound-assisted drying of food materials, in Innovative food processing technologies: Advances in multiphysics simulation, (eds. K. Knoerzer, P. Juliano, P. Roupas, C. Versteeg), WileyBlackwell, West Sussex, UK, pp. 265–302. Riley, N., 2001. Steady streaming. Annu. Rev. Fluid Mech. 33: 43–65. S
anchez, E. S., Simal, S., Femenia, A., Benedito, J., Rossell
o, C., 1999. Influence of ultrasound on mass transport during cheese brining. Eur. Food Res. Technol. 209: 215–219. Sch€ossler, K., J€ager, H., Knorr, D., 2012. Effect of continuous and intermittent ultrasound on drying time and effective diffusivity during convective drying of apple and red bell pepper. J. Food Eng. 108: 103–110. Simal, S., Benedito, J., S
anchez, E. S., Rossell
o, C., 1998. Use of ultrasound to increase mass transport rates during osmotic dehydration. J. Food Eng. 36: 323–336. Simal, S., Feminia, A., C
arcel, J. A., Rossell
o, C., 2005. Mathematical modelling of the drying curves of kiwi fruits: Influence of the ripening stage. J. Sci. Food Agric. 85: 425– 432. Simal, S., C
arcel, J. A., Bon, J., Castell-Palou, A., Rossell
o, C., 2006. Mass transfer modelling in an acoustic-assisted osmotic process. Defect Diffus. Forum 258–260: 600–609. Sir
o, I., V
en, C., Balla, C., J
on
as, G., Zeke, I., Friedrich, L., 2009. Application of an ultrasonic assisted curing technique for improving the diffusion of sodium chloride in porcine meat. J. Food Eng. 91: 353–362.

References Soria, A. C., Villamiel, M., 2010. Effect of dehydration of onion slices. J. Food Eng. 78: ultrasound on the technological properties 90–97. and bioactivity of food: A review. Trends Food Turner, I. W., Jolly, P., 1991. Combined Sci. Technol. 21: 323–331. microwave and convective drying of a Strommen, I., Kramer, K., 1994. New porous material. Drying Technol. 9: 1209– applications of heat pumps in 1270. drying processes. Drying Technol. 12: 889– Vega-Mercado, H., G
ongora-Nieto, M. M., 901. Barbosa-C
anovas, G. V., 2001. Advances Sutar, P. P., Gupta, D. K., 2007. Mathematical in dehydration of foods. J. Food Eng. 49: modeling of mass transfer in osmotic 271–289.

277

279

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality Yingqiang Wang, Min Zhang, and Arun S. Mujumdar 9.1 Introduction

Dehydration is one of the main methods used to preserve foods, keeping it in a stable and safe condition as dehydration reduces water activity and extends the shelf life of fresh fruits, vegetables and aquatic products by preventing microbial activity (Wang et al., 2011b; Mujumdar and Law, 2010; Zhang et al., 2006). Furthermore, there are some other objectives of dehydration, for example, quality enhancement, improving ease of handling, further processing, sanitation, and the development of new products (Duan et al., 2010a). Drying is the most common, but also the most energy-consuming, food preservation process (Mujumdar, 2006; Ratti, 2001). Many conventional thermal methods, such as hot-air drying, vacuum-drying and freeze-drying result in low drying rates in the falling rate period of drying and often also in the surface drying region, due to the heat sensitivity of most foodstuffs (Clary et al., 2005; Zhang et al., 2003; Zhang et al., 2005). Furthermore, extended drying times at relatively high temperatures during the falling rate period lead to the undesirable thermal degradation of the product (Mousa and Farid, 2002). In addition, convective drying seriously damages the sensory characteristics and nutritional properties of foods. Occasionally, severe shrinkage caused by convective drying can reduce the rehydration capacity of the dehydrated product. Given the large quantities of materials produced from agriculture, horticulture and other commodities that must be dried each year, and the magnitude of associated energy inputs, there is a continuing need to improve the thermal drying systems used. The main areas for improvement are in energy transfer, drying times to reduce equipment size, and attaining better uniformity of drying rates and product quality (Souraki et al., 2009). Microwave (MW) applications offer potential opportunities for enhancing the drying of food. Microwave heating/drying coupled with other drying techniques such as convective, vacuum- and freeze-drying provides various options for microwave-assisted drying techniques with their respective merits; for example, microwave vacuum-drying (MWVD), microwave freeze-drying (MWFD), Modern Drying Technology Volume 5: Process Intensification, First Edition. Edited by Evangelos Tsotsas and Arun S. Mujumdar. Ó 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

280

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Microwave-assisted drying

Microwave emission mode

Pulsed mode

Constant mode

Continuous and variable mode

Type of drying combined with microwaves

Microwave-assisted convective drying Microwave-assisted vacuum drying Microwave-assisted freeze drying

Microwave application time

Microwave inish drying

Complete microwave drying

Combination sequence

Single-stage combined drying

Multistage combined drying

Microwave-assisted spouted bed drying Fig. 9.1 Classification of microwave-assisted drying techniques.

microwave spouted bed drying (MWSD) and microwave convective drying (MWCD) (Wang et al., 2011b). These microwave-assisted drying processes can also be run under many operating conditions and various energy modes (Arikan et al., 2012). Figure 9.1 shows a number of possibilities of microwave-assisted drying systems that can be exploited. Microwave-assisted drying is based on the volumetric heating of a food material, unlike convective, conduction or radiative heat transfer which is used in conventional drying methods. Microwave heating allows a faster heat transfer in the food material, without the need for large thermal gradients that may damage the product. In the microwave drying of foods, a reduction in drying time of up to 25–90% and a four- to eightfold increase in drying rate, when compared to convective drying, have been reported (Feng et al., 2012). Therefore, microwave drying is a rapid method of moisture removal, in which the energy consumption can be decreased by between 32% and 71% as compared to conventional drying (Duan et al., 2010b; Varith et al., 2007; Durance and Wang, 2002; Sharma and Prasad, 2006a). From the viewpoint of either drying time or drying efficiency, microwave-assisted drying has one major advantage over its conventional drying counterpart in that, for most cases, the quality of the products from microwaveassisted drying under optimal conditions is superior or at least similar to the quality of conventionally dried products (Cui et al., 2003; Lin et al., 1999, 1998). Even microwave-assisted drying such as microwave vacuum-drying can provide dehydrated products that have better characteristics than those dried by freezedrying due to the development of a desirable, crispy texture in the dried foods,

9.2 Microwave-Assisted Drying of Foods

often providing an excellent alternative method to produce no-fried snack products (Wang et al., 2013; Bai-Ngew et al., 2011). Microwave-assisted drying has added advantages that include selective heating, a better and more rapid process control, as well as less floor-space requirements due to the shorter residence time needed for drying (Zhang et al., 2010). However, the drying of food materials may be limited by the danger of burning, due either to temperature runaway (especially when a higher microwave power is used in order to remove moisture more rapidly in the hygroscopic region), or to the development of hot spots, possibly induced by the material’s structure (Holtz et al., 2010). Consequently, both the control of microwave-assisted processes and the selection of a suitable drying strategy for a specific food with the least energy consumption represent crucial problems for the drying of food. It has been reported that large quantities of food materials of different origins and with different technological characteristics, including fruits, vegetables, bean, grain, spice plant, meat, aquatic products, processing waste, restructured products, instant soup, honey, sticky Ganoderma lucidum extract, and pasta, can be dried in microwave-assisted drying equipment. Many nonfood products are also dried in the same equipment, including pharmaceuticals and ceramics.

9.2 Microwave-Assisted Drying of Foods 9.2.1 Basic Principles of Microwave-Assisted Drying

Microwaves are electromagnetic radio waves within the frequency band of 300 MHz to 300 GHz, corresponding to wavelengths of 1 mm to 1 m (Shivhare et al., 2010). Although microwaves cover a wide range of frequencies, their use is restricted to specific frequencies owing to the possibility of their interference with radar or other communication devices. The typical frequencies used in microwaves are 2450 MHz for home-type ovens, and 915 MHz for industrial use (Sahin and Sumnu, 2006). Microwave-assisted drying is based on the volumetric heating by microwaves of dielectric materials such as water, fat and protein, which is a different mechanism from the conduction or convection heating when using an electric heater or flue gas. Conventional heating relies on conduction and convection to transport heat from the heating sources to the product, and this requires a relatively long period of time for heat-sensitive materials and poor thermal conductors such as foodstuffs. In contrast, microwave heating has the potential to deliver heat instantly throughout the product, due to volumetric heat generation. The mechanism of microwave drying involves the exposure of a product to microwaves that penetrate directly into the material, and this results in heating from the inside out. Microwaves are not forms of heat, but rather are forms of electromagnetic energy that manifest as heat through their interaction with the material undergoing exposure (Mujumdar, 2006).

281

282

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Driving force for moisture transport

Temperature difference

Concentration difference

Moisture diffusion

Vapor diffusion

Vapor pressure difference

Hydraulic flux

Fig. 9.2 Mechanisms of moisture transport inside unfrozen food materials during microwave heating. Adapted from Cui (2004).

Microwave energy absorption in foods primarily involves two mechanisms: dipolar relaxation; and ionic conduction (Sahin and Sumnu, 2006). In ionic conduction, ions are accelerated by electric fields, causing them to move towards the direction opposite to their own polarity. This movement of ions provokes collisions with the molecules of the material, thereby creating a disordered kinetic energy such that, as a consequence, heat is generated. Polar molecules subjected to microwave radiation at 2450 MHz will rotate 2.45  109 times per second, and the friction between the rapidly rotating molecules causes heat to be generated throughout the material rather than simply be transferred from the surface to the inner part, as is the case for conventional hot-air drying. The water contained in the food serves as the primary dipolar component responsible for dielectric heating (Ratti, 2008), during which different driving forces for moisture transport are developed from the inside to the outside of the heated material. In this way, the water is moved rapidly to the surface by a combination of moisture diffusion, vapor diffusion and hydraulic flow. The mechanisms of moisture transport inside an unfrozen food material during microwave heating are summarized in Fig. 9.2. The moisture, having reached the surface, can be removed by hot air during microwave convective drying, or by a vacuum pump during microwave vacuum-drying or microwave freeze-drying. This restores the driving forces for moisture transport in the heated material. Subsequently, this process is repeated such that the moisture evaporates and leaves the material in continuous fashion, until drying has been completed (Cui, 2004). During this process, a part of the microwave energy is transformed directly to the heat of vaporization. In microwave-assisted freeze drying, the basic principle is similar to that of traditional freeze-drying, but differs from the other microwave-assisted drying processes at the sublimation drying stage. In frozen food materials the moisture is removed by sublimation; consequently, the microwave energy supplies only the heat of vaporization during the sublimation drying stage, but at a faster rate compared to either conventional conduction or radiation (Jangam et al., 2011; Duan et al., 2010b).

9.2 Microwave-Assisted Drying of Foods

9.2.2 Energy Absorption by Products During Dielectric Heating

Generally, the average power density P (volumetric absorption of microwave energy, W m 3) produced in a material when exposed to microwave energy is defined by the following equation (Al-Harahsheha et al., 2009): P ¼ 2pf e0 e00 E 2 ;

ð9:1Þ

where f is the frequency (in Hz), e0 is the permittivity of free space (8.86  10 12 F m 1), e00 is the effective relative dielectric loss factor, and E is the electric field strength inside the material (in V m 1). Dielectric materials, such as food products, convert electric energy at microwave frequencies into heat. The increase in the temperature of a material due to dielectric heating can be calculated as (Regier and Schubert, 2005): rcP

dT ¼ 55:63  10 dt

12

f E 2 e00 ;

ð9:2Þ

where c P is the specific heat of the material, r is its density, and dT/dt is the rate of temperature increase (all in SI units). The microwave heating rate depends on dielectric properties of the material. 9.2.3 Dielectric Properties

Dielectric properties are the main parameters that provide information about how materials interact with electromagnetic energy during dielectric heating (SosaMorales et al., 2010). Dielectric properties are commonly represented by a complex number, the relative complex permittivity e ¼ e0

je00 ;

ð9:3Þ

where the real part e0 is the dielectric constant and the imaginary part e00 is the pffiffiffiffiffiffi ffi dielectric loss factor, j ¼ 1. Here, e0 indicates the material’s ability to store 00 electrical energy, and e is a measure of its ability to dissipate the electrical energy in the form of heat (Regier and Schubert, 2005; Wang et al., 2011a). The larger the loss factor, the more easily can the material absorb the incident microwave energy. A material’s dielectric properties are dependent on its moisture content, composition and temperature, as well as on the frequency of the field. Both, the dielectric constant and the loss factor increase with moisture content, but decrease with frequency. Although the dielectric constant increases with temperature, the temperature dependence of the loss factor is unpredictable, so that the loss factor may either increase or decrease with increasing temperature, depending on the range of frequency and the moisture content. The dielectric activity of free water present in the food products is high, whereas bound water exhibits a relatively low dielectric activity (Shivhare et al., 2010). The dielectric constant falls steadily in a

283

284

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Fig. 9.3 Dielectric constant of various food materials at 25  C. Adapted from Sahin and Sumnu (2006).

linear manner as the NaCl concentration increases, while the loss factor increases significantly with the increase of NaCl concentration (Al-Harahsheha et al., 2009). The dielectric constants and loss factors of various food materials are summarized in Figs 9.3 and 9.4, respectively. As can be seen in these figures, the dielectric properties of cooking oils are very low because of their nonpolar characteristics, whereas the dielectric properties of water and high-moisturecontaining foods such as fruits, vegetables and meat are high because of dipolar

Fig. 9.4 Dielectric loss factor of various food materials at 25  C. Adapted from Sahin and Sumnu (2006).

9.3 Microwave-Assisted Drying Equipment

rotation. The highest loss factor is observed in the case of salt-containing foods such as ham (Sahin and Sumnu, 2006). 9.2.4 Penetration Depth

The penetration depth is an important measure that is often used to assess whether an electromagnetic field at a certain frequency can provide a relatively uniform heating in a given food product (Regier and Schubert, 2005). The penetration depth (dp ) is usually defined as the depth into a sample where the microwave power has fallen to 1/e (e ¼ 2.718) or 36.8% of its transmitted value. The penetration depth is a function of e0 and e00 , and can be calculated as (Mujumdar, 2006): pffiffiffi l0 2 dp ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð9:4Þ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2p e0 ½

1 þ ðe00 =e0 Þ2

1Š

or

dp ¼

pffiffiffiffi l0 e0 : 2pe00

ð9:5Þ

Here, l0 is the free space microwave wavelength, l0 ¼ c 0 =f , and c0 is the speed of light in free space (c ¼ 3  108 m s 1). Common food products have e00 < 25, which implies a dp of 0.6–1.0 cm. The penetration of microwaves at 915 and 2450 MHz in foods with high moisture contents at room temperature is typically between 0.3 and 7 cm, depending on their dielectric properties and the frequency of the microwave field (Sosa-Morales et al., 2010). At a constant frequency, the penetration depth increases with the dielectric constant but decreases with the material’s loss factor. Both, the dielectric constant and the loss factor decrease with a reduction in moisture content, but the penetration depth is more sensitive to the variation in loss factor. Hence, the penetration depth increases as the material’s moisture content decreases during processing (e.g., drying) (Shivhare et al., 2010). The penetration depth of food materials is significantly increased as the temperature decreases, especially at temperatures below the freezing point (Bai-Ngew et al., 2011). This is beneficial for the uniformity of microwave heating and, hence, for the improvement of product quality in microwave-assisted freeze-drying (Wang et al., 2011a). The penetration depth falls dramatically as the concentration of sodium chloride in foods increases (Al-Harahsheha et al., 2009).

9.3 Microwave-Assisted Drying Equipment

A microwave-assisted drying system is generally composed of a magnetron as microwave power source, a waveguide, a drying chamber, a material plate, a

285

286

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Fig. 9.5 A simplified schematic of a pilot-scale microwave-assisted convective dryer.

dehumidification system, a mechanical drive system, safety protection system, and a sensor and control unit (www.vertpedia.com). Based on the combination of microwaves with other drying techniques, there are mainly four types of microwave-assisted dryer, as discussed below. 9.3.1 Microwave-Assisted Convective Drying Equipment

Figure 9.5 shows a simplified schematic of a pilot-scale microwave-assisted convective dryer described by Pereira et al. (2007) and Ahrn
e et al. (2007). The microwave dryer consists of a microwave generator and a forced air system. The microwave generator has a maximal microwave power of 1000 W operating at a frequency of 2450 MHz. The cylindrical cavity of the dryer with a vertically oriented axis has a diameter of 80 cm and a height of 10 cm, and accommodates a Teflon tray and a built-in scale with an accuracy of 2 g. The weight loss, product temperature and microwave power were recorded every second by a data acquisition system. The air flow inlet and outlet are located in the middle of the chamber, and the heated air is distributed and directed through the lid to the samples placed inside the cavity. Microwave power, air temperature, velocity and humidity were adjustable parameters. The product temperature was measured by fiber-optic temperature sensors (ReFLexTM, Neoptix, Canada) placed in four samples located at different positions of the sample plate. The size of the cavity was optimized by modeling the microwave field distribution for 1 kg of fruit pieces distributed as a single layer. In this dryer, the load does not rotate. A laboratory-scale microwave convective dryer with maximum output of 625 W at 2450 MHz and cavity (drying chamber) dimensions of 300 mm (width), 240 mm

9.3 Microwave-Assisted Drying Equipment

(depth) and 210 mm (height) was reported by Uprit and Mishra (2003). The dryer had the capability of online sample weighing and power (wattage) adjustment, and the load could rotate, but the material temperature was not controlled due to the movement of material. There are three methods in which microwave energy may be combined with hotair drying, by applying the energy:  at the beginning of dehydration processes;  when the drying rate begins to fall; or  during the falling rate period(s) or at low moisture content to facilitate finish drying. In some cases, applying microwave drying at the last stage of the dehydration process can also be very efficient for removing bound water from the product (Zhang et al., 2006). In the microwave-convective air-drying systems, microwave energy spurs the water to reach the surface of the product at a high rate, while the convective flow of air helps in removing the moisture from the drying chamber and also contributes to a more homogeneous drying. Thus, combined microwave and convective drying systems increase not only the drying rate of the product but also the quality of the dry product obtained. 9.3.2 Microwave-Assisted Vacuum Drying Equipment

Figure 9.6 shows a laboratory microwave-vacuum dryer on the basis of hexahedron of the cavity with the dimensions of 360 mm  340 mm  260 mm. The designed nominal power output of the magnetron was 750 W, and the actual measured output was 400 W. The turntable and the motor were connected by a shaft which was dynamically sealed. The structure and dimension of the window for the microwaves to pass was carefully designed in order to avoid any electrical discharge due to lower pressure. The door was sealed by a

5 4 3

6 7 8 9

2 1

Fig. 9.6 Schematic of a laboratory microwave-vacuum dryer. 1, Vacuum pump; 2, Rotating tray; 3, Valve; 4, Observation window; 5, Drying chamber; 6, Infrared thermometer; 7, Vacuum gage; 8, Magnetron power control unit; 9, Magnetron.

287

288

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Fig. 9.7 A pilot microwave-vacuum dryer.

convex–concave face with a microwave leakage of less than 5 mW cm 2. The microwave-vacuum dryer was operated at 25 mbar and the rotation speed of the turntable was 5 rpm (Cui et al., 2003). Figure 9.7 presents, schematically, a pilot vacuum microwave dryer (WZD4S-1, Nanjing, China) with a maximum nominal power of 4.2 kW and an operation frequency of 2450 MHz. The operation power is adjustable between 700 W and 4200 W at 700 W increments. The dimensions of the microwave cavity were 1050 mm, 1080 mm, 800 mm (width, depth and height, respectively). The minimal attainable pressure is 5 kPa. Six material plates in the cavity can move around at 1 rpm during operation. Figure 9.8 shows a schematic diagram of a microwave-vacuum rotary drum dryer reported by Kaensup et al. (2002). In this dryer, a cylindrical vacuum chamber was made from stainless steel having diameter 0.33 m and length 0.4 m. A 2450 MHz magnetron having maximum power output of 800 W was located 110 mm aside of the drying chamber. Microwaves were conducted to the vacuum chamber via a stainless steel waveguide having a rectangular crosssection of 40 mm  80 mm. Polypropylene was chosen to make a perforated rotating drum with guide vane, as this material is almost transparent to microwaves. The drum had a diameter of 0.3 m and length of 0.3 m. A novel vacuum seal made from metal shield attached on Teflon bush enabled the rotary drum to be driven in the chamber by means of an external gear motor. The enclosed chamber was almost microwave-proof, with measured leak radiation at 0.5 mW cm 2. A vacuum pump with a capacity of 63 m3 h 1 at 4 mmHg was used to withdraw vapor from the vacuum chamber. The drum rotation speed was controlled using an electronic inverter. Electrical energy

9.3 Microwave-Assisted Drying Equipment

Fig. 9.8 A microwave-vacuum rotary drum dryer. Adopted from Kaensup et al. (2002).

consumed in the drying process was measured by means of a Watt–hour meter. In operation, the guide vanes agitate the product as the drum is rotated in the chamber, which allows for an even exposure of the product to microwave energy and a more uniform mixing. Furthermore, the moisture expelled during the drying process can easily leave the product, resulting in a shortened drying time and no hot spots. The application of an industrial-scale microwave-vacuum dryer (Model MG8KW, Enwave Corporation, Vancouver, Canada) to dry Saskatoon berries was reported by Pranabendu and Venkatesh (2009). The dryer consisted of a cylindrical stainless steel chamber, where the microwave field and vacuum were applied. Inside the chamber, the berries were contained in a basket made from ultrahigh-density polyethylene, sitting on two rollers made from the same material. One of the rollers was connected directly to a variable-speed motor through the chamber wall; the cogs in the wall of the basket mesh matched the cogs in the roller to provide a smooth rotation. The basket rotation speed was set prior to each drying cycle and kept constant at 10 rpm for all of the drying treatments. The vacuum was provided by an external pump, and the vacuum pressure was kept constant at about 100 kPa (Pranabendu and Venkatesh, 2009). The recommended optimum drying conditions for the drying of Saskatoon berries are a microwave power of 5.7–6 kW and a drying time of 51.5–55 min, with a fruit load of 9.75 to 10 kg..

289

290

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

14 15

3

7

8

13

2 6

12 11 10

5 4

9

1

22

17

16 18

19

20

Fig. 9.9 Laboratory-scale microwave-assisted freeze-drying equipment. 1, MWFD chamber; 2, Fiber-optical temperature sensor; 3, Vacuum breakage valve, for MWFD; 4, Sample supporting plate; 5, MWFD sample; 6 and 7, Microwave sources; 8, Pressure senor, for MWFD chamber; 9, FD chamber; 10 and 12,

T

21

Heating plates; 11, Freeze-dried sample; 13, Pressure senor, for freeze-drying chamber; 14, Vacuum breakage valve, for freeze-drying; 15, Temperature sensor; 16, Freeze-dryer vacuum valve; 17, MWFD vacuum valve; 18, Cold trap; 19, Vacuum pump; 20, Draining valve; 21, Refrigeration compressor; 22, Control system.

9.3.3 Microwave-Assisted Freeze-Drying Equipment

A typical schematic diagram of microwave-assisted freeze-drying equipment is shown in Fig. 9.9. During drying, the pressure was maintained at 100 Pa by a vacuum pump, and the temperature of the cold trap ( 40 to 45  C) was sufficiently low to condense the vapor. The microwave frequency was 2450 MHz and the power could be regulated continually from 0 to 2000 W. A microwave field can be used as the heat source to supply the heat of sublimation needed in the freeze-drying process. Volumetric heating of the material in the vacuum environment makes it possible to greatly improve the rate of sublimation. In the equipment, the microwave resonant cavity was designed as an effective multimode resonant cavity, which makes the electric field distribution uniform when the material to be dried is in a stationary state (Duan et al., 2010b; Wang et al., 2011b). It should be noted that subdivision of the drying chamber of Fig. 9.9 to two parts enables microwave-assisted freezedrying and conventional freeze-drying to be conducted in parallel, under otherwise the same conditions.

9.3 Microwave-Assisted Drying Equipment

Fig. 9.10 Schematic diagram of microwaveassisted spouted bed. 1, Air compressor; 2, Blower; 3, Electric heating; 4, Pressure detector; 5, Valve; 6, Nozzle; 7, Peristaltic pump; 8, Magnetron; 9, Observation

window; 10, Operation door; 11, Spouted bed; 12, Air flow detector; 13, Control box; 14, Power regulator; 15, Optical fiber measurement; 16, Equipment support.

9.3.4 Microwave-Assisted Spouted Bed Drying Equipment

Figure 9.10 shows a typical schematic diagram of a microwave-assisted spouted bed. The system consisted of 2450 MHz microwave power sources, a cavity, a hotair source, a spouted bed, and a control system. Microwave power was provided by four magnetrons, each having 1 kW maximum power capacity. The total microwave power could be regulated between 0 and 4 kW. The temperature of the air for the spouted bed could be controlled between 30 and 150  C through a feed-back loop, and the air velocity was maintained with an adjustable fan. A fiber-optic temperature sensor detected the temperature in the drying cavity. The bottom of the cone was made from a plastic screen to hold the particulate samples and provide a pass for hot air. A metal screen with small holes was placed on the observation window to stop microwave leaking. Figure 9.11 shows the scheme of a laboratory-scale microwave-assisted inert medium fluidized-bed drying experimental apparatus. A domestic microwave oven (Type: LG, MC-2003TR(S)) with a frequency of 2450 MHz was used to set up the dryer. By making a hole under the microwave oven and sealing it, a fluidized bed was placed into the microwave oven. A hole was also made on top of the microwave oven so that air could exit. U-shape metallic tubes were used at the inlet and outlet of the drying air stream to prevent microwaves from coming out of the oven. The other parts of the system were the same as those used in a conventional fluidizedbed dryer. The dryer was a cylindrical Pyrex column of 77 mm diameter and 35 cm

291

292

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Microwave oven

Timer Inert particles Power set switch

Rotameter

Temperature Controller

Drying sample Air distributor

Air Flow rate regulating valve

Electrical heater

Fig. 9.11 Scheme of microwave-assisted inert medium fluidized-bed dryer system. Reproduced with permission from Souraki et al. (2009).

height, equipped with a porous plate for air distribution. An overall amount of 400 g of glass beads of diameter 2.7 mm was used as energy carrier. Although the glass beads are transparent against microwaves, they increase the rate of heat and mass transfer, and act similarly to the turntable in conventional domestic microwave ovens. The fluidization of inerts agitates the drying sample (carrot), so that microwave radiation can be received by all of the sample parts. Drying air was supplied from a high-pressure air source, with the pressure being adjusted by a regulator. Air was passed through a rotameter and then heated by a controlled electrical heater. A temperature controller was used to regulate the temperature of drying air within 1 K. Air humidity was determined by measuring the dry and wet bulb temperatures (Souraki et al., 2009). To date, only microwave hot-air dryers used principally in finish drying processes to level-off the moisture content in pasta, crackers or chips, and microwave vacuum dryers custom-made to specific product requirements, are available commercially (Ratti, 2008). Good microwave-assisted drying equipment should have a uniform distribution of the microwave field, provide a movement of the material to be dried, keep the microwave field strength in the drying chamber lower than the dielectric breakdown strength, and possess an economic dehumidification system.

9.4 Microwave-Assisted Drying Process

Moisture and temperature history relate to quality changes during microwaveassisted drying processes and have, therefore, been intensively investigated. Their

9.4 Microwave-Assisted Drying Process

observation and modeling is helpful for understanding and controlling the drying process as well as for process optimization. 9.4.1 Moisture Loss

As drying proceeds, the moisture is removed continuously from the food material. Moisture loss is generally described by the drying rate (in kg of removed water per kilogram of dry material and time), and by the total drying time spent to remove a specific amount of water in a dryer. In general, a complete microwave-assisted drying process displays three drying periods: a transient rising rate period; a rapid rate drying period; and a falling rate drying period. At the start of drying, the drying rate rises rapidly due to fast microwave heating, until the process enters into the rapid rate drying period, when most of the free moisture is removed at a constant rate or a high falling rate. The drying process then continues to its end at low falling rates (see Fig. 9.12; cf. also Wojdyo et al., 2009; Figiel, 2010; Arikan et al., 2012). The drying rate in the rapid drying period is governed mainly by the rate of external heat and mass transfer, since a film of free water is always available at the evaporating surface. The duration of the rapid rate drying period is dependent on the level of microwave power used, the type of material being dried, and the other drying conditions. As the water content is decreased markedly towards the final drying, two parallel types of behavior can be identified: (i) a considerable decrease in the microwave power absorbed by the foodstuff (associated with a smaller e0 and e00 ); and (ii) a higher resistance to water transport as a result of dried cell layers and a reduction in the driving force caused by concentration and pressure gradients. Both effects contribute to further decreasing the drying rate (Pereira et al., 2007). Compared to its conventional counterpart, moisture can be removed at higher rate during microwave-assisted drying, as shown in Fig. 9.12.

Fig. 9.12 A comparison of typical drying curves between microwave-assisted drying and conventional drying.

293

294

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Factors affecting the drying rate include the type of conventional drying combined with microwaves, microwave power, pulsed mode, load, pretreatment, sample thickness and composition, as well as the performance of the dehumidification system. The drying rate significantly increases and the drying time falls as the microwave power, or the sum of microwave power over time in pulsed mode, is increased. The drying time increases with an increase in the load (Hu et al., 2006; Esturk, 2012; Wojdyo et al., 2009; Figiel, 2010; Arikan et al., 2012). Osmotic treatment reduces the initial moisture content and increases moisture mobility within the sample by diminishing shrinkage, thus significantly reducing the subsequent drying time (Wang et al., 2010a; Al-Harahsheha et al., 2009). The addition of salt and sucrose to restructured potato slices and instant vegetable soup led to a significant increase in the drying rate, due to an enhanced ability of microwaves to be absorbed by these materials during microwave-assisted freeze drying. It was noted that a 17–54% reduction in drying time occurred as the salt content was increased from 1% to 7%, while a reduction in drying time of between 11.1% and 33.3% was found as the sucrose content was increased from 3% to 15% (Wang et al., 2011a; Wang et al., 2010b). With regards to the effect of sample thickness on drying time, it has been concluded by some groups that the drying time when using the microwave method becomes longer for thinner slices (in contrast to the convective method). For example, the use of thinner banana and potato slices, as well as the slicing of garlic cloves, prolonged the duration of microwave-assisted drying (Maskan, 2000; Wang et al., 2004; Figiel, 2009). This may be attributed to the smaller thicknesses of material reducing the absorption of microwave energy if the penetration depth is too large (Figiel, 2009). Others have reported the same trend in results as for conventional convective drying method, notably in the cases of microwave-assisted freeze-drying of instant vegetable soup and microwave-assisted vacuum-drying of mushroom (Giri and Prasad, 2007a; Wang et al., 2009). In addition, during microwave-assisted convective drying, air (both temperature and velocity) plays an important role, not only as a carrier of evaporated moisture but also as it contributes to a more homogeneous and faster drying (Pereira et al., 2007; Feng and Tang, 1998; Funebo and Ohlsson, 1998; Dev et al., 2011). The air during the initial drying stage does not have any significant effect on the drying rate of microwave-assisted spouted bed drying, but it does have an important effect on the final stage of drying (Feng and Tang, 1998). It was reported that the increase in air velocity in the microwave convective drying of banana from 0.3 to 0.8 m s 1 caused a 20–25% reduction in the drying time (Ahrn
e et al., 2007). During microwave-assisted vacuum drying, a higher vacuum increases the driving force for mass transfer and also facilitates the evaporation and volatilization of water from the materials, thus shortening the drying time. However, the effect of system pressure on drying time was not as significant as that of microwave power (Hu et al., 2006). Typical drying curves of three different microwave methods are shown in Fig. 9.13. There are insignificant differences in the shape of the drying curves, but the drying time necessary to reach the desirable moisture content differs among

9.4 Microwave-Assisted Drying Process

Fig. 9.13 Drying curves of three microwave-assisted drying methods.

the various microwave-assisted drying methods (Yan et al., 2010). The drying time of MWFD was 3, 4.3, and 6 times longer than in MWVD, MWSD at 2.0 W g 1 and MWSD at 3.5 W g 1, respectively. The low drying rate in MWFD could be explained by the fact that the dielectric constant value of ice is lower than that of liquid water, which results in a lower conversion of microwave energy to thermal energy in MWFD. In addition, the high drying rate in MWSD can be attributed to the constant movement of carrot pieces within the microwave cavity, which allowed the different parts of the carrot pieces to receive a relatively uniform microwave radiation over a period of time. Table 9.1 shows some drying kinetic model equations used to fit microwave drying data. 9.4.2 Temperature Distributions

Temperature was found to be the most significant predictor of drying process during microwave-assisted drying, as it reflects the general drying performance rather well (Duan et al., 2010b). Three different temperature measurement systems are available (Regier and Schubert, 2005):  Infrared temperature measurement  Fiber-optical measurement  Metal-shielded thermocouples. The infrared thermometer can record the surface temperature of the materials, while the metal thermocouples and the fiber-optical temperature

295

296

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality Tab. 9.1 Some drying kinetic models for microwave-assisted drying of food (symbols are explained in the text and chapter notation; all other symbols represent fitting parameters).

Drying method

Model

Sample

MWVD

MR ¼ expð ktn Þ

MWCD

MR ¼ expð ktn Þ

MWCD

MR ¼ exp½ ðaPMW þ bT ÞtŠ

MWCD

MR ¼ expð ktn Þ

MWVD

First period: X ¼ A Bt First period: X ¼ C þ ½D expð Etފ

Mushrooms (Giri and Prasad, 2007a) Moringa oleifera (Dev et al., 2011) Cooked chickpeas (Gowen et al., 2006) Cooked soybeans (Gowen et al., 2008) Strawberry fruits (Wojdyo et al., 2009)

MWVD

X  2 : X ¼ Xi

P abs t Mdry Dhv

 0 < X < 2 : X ¼ 1:3662X 0:5741 X i

MWVD MWCD

b 1 þ exp½ðt cÞ=dŠ h i MR ¼ exp ðt=aÞb

X ¼aþ

Carrot (Cui et al., 2004) P abs t Mdry Dhv

 Garlic (Figiel, 2009) Osmotically treated pineapple (Botha et al., 2012)

sensor can measure the interior temperature. An infrared thermal imaging camera (IRI 4010 Multi-Purpose Imager; IRISYS, UK) was used to determine the temperature distributions of individual thin chips and to assess hot spots. It is reported that for a sample thickness less than 8 mm, the core temperature of the sample was the same as its surface temperature, with uniform temperature distribution within the sample. However, for sample thickness more than 8 mm, a temperature gradient developed along the thickness of the sample. Generally speaking, if foods such as fruits and vegetables are cut into slices with thickness less than 10 mm before drying, then the sample can be considered to be of uniform temperature (Cui et al., 2005). Figure 9.14 presents typical temperature curves of microwave-assisted drying for food. Here, the shape of the temperature curve of microwave-assisted freeze-drying can be seen to differ from that of the three other types of microwave-assisted drying, due to different drying mechanisms. Temperature changes are heavily dependent on the microwave power, microwave emission mode, type of dryer used, and the presence or absence of a temperature control unit. 9.4.2.1 Temperature Variations at Fixed Levels of Microwave Power Three different temperature stages are observed at constant microwave power in most microwave-assisted drying processes (Clary et al., 2005; Feng et al., 1999; Ahrn
e et al., 2007; Duan et al., 2010b):

9.4 Microwave-Assisted Drying Process

Fig. 9.14 Typical temperature curves of microwave-assisted drying for food, where 1 and 2 indicate the temperature changes in the final drying stage. 1, Convective, vacuum or freeze microwave-assisted drying

at constant microwave power; 2, Adjusted power with feedback temperature control, or variable microwave power, or pulsed mode, or microwave-assisted spouted bed drying under constant microwave power.

 A warming-up period in which the product temperature increases linearly with drying time.  A constant temperature period, which corresponds to the aforementioned constant drying rate period in which moisture is removed at the outer surface of the material and the cooling effect of vaporization remains in balance with the microwave heating.  A heating-up period, in which the drying rate decreases and the sample temperature increases rapidly. During the period of relatively constant temperature, the temperature of microwave-assisted freeze-drying is lowest, which should be lower than the comelting temperature of the material, while the temperature of microwaveassisted vacuum-drying is generally between 26  C and 52  C, depending on the operating pressure. Microwave-assisted convective drying can be run at a higher temperature, depending on the microwave power used and other conditions such as air temperature (Pereira et al., 2007). Similar to the duration of the rapid rate drying period, the duration of the constant temperature period depends on the level of microwave power used, the load, the nature of the material, as well

297

298

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

as the other drying conditions. When most of the free water has been removed, the temperature rises sharply (as shown in Fig. 9.14); this is known as temperature overrun or thermal runaway, due to both an increase in the specific microwave energy input (the energy delivered per unit mass of wet product) and a decrease in the product’s specific heat capacity (this is involved in the conductive heat transfer) with progressing drying (Figiel, 2010; Botha et al., 2012). The microwave power level has a decisive effect on the thermal runaway, with a higher microwave power resulting in a more serious thermal runaway. Low microwave power, and air with a low temperature and a high velocity, had a lowering effect on the thermal runaway during microwave-assisted convective drying (Pereira et al., 2007). In the microwave-assisted spouted bed drying of diced apples, no rise in product temperature was observed with a microwave input power of 6.4 W g 1 (dry basis), but the temperature of the diced apples was 12–15  C above the spouted bed air temperature. The reason for this may be the high surface heat and mass transfer rates, facilitated by the intensive pneumatic agitation, which ensure heat removal at a high rate and enough convective cooling to balance the heat generated internally by microwave energy. The MWSD system provided a unique temperature-leveling effect after a short warming period (Feng et al., 1999). The results of Figiel (2009) showed that the temperature change in garlic cloves and slices subjected to MWVD can be approximated with the exponential function: T ¼ 50:6 þ 25:9 expð X =0:167Þ:

ð9:6Þ

According to a model developed by Cui et al. (2005), the temperature of sliced carrot that has been dried in a microwave-vacuum dryer can be predicted by the following equations: Warming-up period; X ¼ 7:68 : Constant-temperature period; 2  X < 7:68 : Heating-up period; 0 < X < 2 :

T ¼ Ti þ

P abs t Mwet cp

T ¼ T T ¼ P 0:0196 abs X

0:0797

T



ð9:7Þ

Here, T is the saturation temperature of water at the prevailing pressure, and Pabs is the absorbed microwave power (in W). The equations were derived for massspecific microwave power densities from 3.5 W g 1 to 18.7 W g 1. 9.4.2.2 Temperature Variations at Variable Microwave Power without Controlling Temperature Variable microwave power means that the microwave energy is emitted in a pulsed mode by turning on-off the microwave power during the drying process, or that the microwave emission is declining continually with time. Compared with the constant power mode, the variable microwave power mode is able to stabilize the temperature of the sample at a value that depends on the microwave power used, the duration of the power-on and power-off periods, and the other operating conditions, which indicates that thermal runaway can be inhibited in the variable microwave power mode.

9.4 Microwave-Assisted Drying Process

It was reported that intermittent microwave-convective drying conducted at a low drying air temperature and microwave power level with a relatively long power-off time, resulted in a more stable and gentle drying process due to decreasing the temperature (Soysal et al., 2009; Changrue et al., 2008). Botha et al. (2012) found, in microwave-assisted convective drying of osmotically treated pineapple, that when a declining mode with time was used the drying temperature could be effectively controlled due to the maintenance of a relatively constant specific microwave energy input (the energy delivered per unit mass of wet product) during drying. High microwave powers need to be reduced quickly, and faster than the decrease in water content, in order to minimize charring. An inlet air temperature of 70  C was found to be excessive when combined with microwave energy, and resulted in a rapid temperature increase. 9.4.2.3 Temperature Change with Time-Adjusted Power in Feedback Temperature Control The best temperature control is achieved with a phase controller, which can continuously and automatically adjust the variable microwave power, based on predefined power profiles (Li et al., 2010). It is worth emphasizing here that maximum power levels must be selected, based on trial and error, for a designated temperature. If the power was too high, the temperature fluctuations were too large, but if the power was too low, then during certain drying stages the temperature could not reach the preset temperature. Li et al. (2010) investigated the microwave-assisted convective drying of apple slices with time-adjusted power in feedback temperature control. Three desirable temperatures of 75, 65, and 55  C were chosen, and the corresponding maximum power requirements were set at 400, 300, and 240 W, respectively. The relationships of power versus time and of power versus moisture content (d.b.) can be described by the following equations, respectively: "   # i¼3 X t bi 2 ; ð9:8Þ ai exp PðtÞ ¼ ci i¼1 PðXÞ ¼

i¼3 X

di exp

i¼1

"  X

ei fi

2 # ;

ð9 : 9 Þ

where ai to fi are temperature-dependent constants with appropriate units. 9.4.3 Energy Consumption

During microwave-assisted drying (Varith et al., 2007; Holtz et al., 2010), the energy consumption comprises of:  the energy used for producing microwaves;  the energy consumed by the dehumidification system, for example, for heating air or maintaining vacuum and refrigeration; and

299

300

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

 the energy consumed by mechanical drive systems for, for example, the fan and turntable. The electrical consumption can be recorded using a Watt–hour meter. The energy consumption of microwave-assisted drying can be evaluated (Varith et al., 2007; Soysal et al., 2006) by two different efficiency indices, namely specific energy consumption (SEC): SECðkJ=kgH2 OÞ ¼

3600E M s ðX i X f Þ

ð9:10Þ

where E is the total electrical energy consumption (in kWh), X i and X f refer to the initial and final moisture contents (d.b.), respectively, and M s is the mass of dry solid (in kg), or energy efficiency:  gð%Þ ¼ Ms X i

Xf

 Dhv  100 3600E

ð9:11Þ

where Dhv is the evaporation enthalpy of water (2257 kJ kg 1, at 100  C). The typical variations of specific energy consumption and energy efficiency with moisture content during microwave-assisted drying are illustrated in Fig. 9.15. At the start of the drying process the specific energy consumption is high and the energy efficiency is low, as the microwave energy applied is used to raise the material’s temperature and very little moisture is evaporated. Subsequently, a decrease in specific energy consumption and a sharp increase in energy efficiency are observed, and both of these maintain a relatively constant level during drying over a wide range of moisture contents. The specific energy consumption is approximately the same for all materials at intermediate to high moisture contents. When drying continues into the hygroscopic region (lower than a critical moisture

Fig. 9.15 Variation of specific energy consumption and energy efficiency with moisture content during microwave-assisted drying.

9.4 Microwave-Assisted Drying Process

content), where water is more tightly bound, the specific energy consumption increases while the energy efficiency reduces (Holtz et al., 2010). At lower moisture levels, the absence of water causes e00 to fall considerably, and the material is now much less inclined to absorb the offered energy, thereby causing the efficiency of the dryer to drop (Kaensup et al., 2002). The energy consumption of microwave-assisted drying for food is affected by drying time, microwave density, microwave mode, load, the drying method combined with microwaves and the type of combination, as well as other drying conditions such as air temperature, air velocity, pressure, and drum rotating speed. The specific energy consumption is significantly decreased with an increase in microwave density (Sharma and Prasad, 2006a). Pulsed or intermittent microwave applications can result in a more energy-efficient drying than that in continuous microwave mode (Clary et al., 2007; Sunjka et al., 2004; Arikan et al., 2012). In both the pulsed and the continuous cases, the drying efficiency was improved when a lower pressure or temperature was applied (Yongsawatdigul and Gunasekaran, 1996). Sunjka et al. (2004) found the microwave-assisted vacuum drying of cranberries to be more energy-efficient than the microwave-assisted convective drying of the same product. The efficiency of microwave-assisted convective drying was increased in microwave mode with a longer power-off time, but was decreased in the corresponding microwave mode of microwave-assisted vacuum drying. A lower material load tended to decrease the drying efficiency and to increase the specific energy consumption. About a 9.5% increase in drying efficiency and about an 18% (0.92 MJ per kg H2O) decrease in specific energy consumption could be obtained by working with a 128.6 g material load instead of 64.3 g. As the intensity of heat generation is proportional to the content of moisture in a dielectrically dried material, a larger amount of water trapped inside the material would provide a higher drying efficiency and lower specific energy consumption values (Soysal et al., 2006). It was reported that varying the drum rotation speed and the vacuum pressure had no effect on specific energy consumption in the high moisture content range. Rather, the specific energy consumption was insignificantly changed when the moisture content in the product was high for all drum rotating speeds and chamber vacuum levels investigated. However, for a low moisture content, lower drum rotating speeds and a lower chamber pressure resulted in less energy consumption (Kaensup et al., 2002). It was observed during the microwave-assisted convective drying of garlic cloves that, as the air velocity was increased from 1.0 to 2.0 m s 1 at a given air temperature of 70  C, the specific energy consumption also increased from 26.3 to 63.0 MJ kg 1. This was because the increase in air velocity resulted in a cooling of the drying product; that is, its temperature was reduced such that the drying time was increased as a consequence (Sharma and Prasad, 2006a). When Yan et al. (2010) compared the energy consumptions of different microwave-assisted drying processes (Fig. 9.16), MWVD and MWSD were found to have about 75% less energy consumption than MWFD. Moreover, there was no significant difference in energy consumption between MWVD and MWSD. The difference from MWFD was attributed to the fact that the drying time for MWFD

301

302

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Fig. 9.16 Comparison of the energy consumption of different drying methods. Note: Different letters indicate a statistically significant difference.

was long (approximately five times longer than for MWVD and MWSD) and more energy was needed to maintain a high vacuum and a low temperature in the cold trap. Table 9.2 presents the SEC and drying efficiency values for different microwaveassisted drying processes of food materials, as determined by various research groups. The minimal theoretical value of SEC for a dryer is 2.3 MJ per kg of evaporated water, based on the maximal theoretical specific moisture extraction ratio (SMER) of 1.55 kg per kWh and the latent heat of water evaporation at 100  C (Holtz et al., 2009). These results are expected to depend heavily on the scale and design of dryers, as well as on the method used to measure energy consumption (Durance and Wang, 2002). 9.4.4 Dielectric Breakdown

Dielectric breakdown occurs when the electric field intensity in the drying chamber is above a threshold value, and leads to burning of the product when food materials are being dried. Ionization of the gases present in the drying chamber, such as vapor and air, leads to the appearance of purple light, which causes burns on the product surface. The occurrence of this phenomenon leads to major energy losses and excessive heating of the dry zone of the material, seriously damaging the final product. The threshold value of the electric field is normally a function of chamber pressure. The dielectric breakdown easily occurs when the pressure is about 133 Pa – a value commonly used in conventional freeze-drying operations – and when the moisture content of the

9.4 Microwave-Assisted Drying Process Tab. 9.2 Specific energy consumption (SEC) and energy efficiency for microwave-assisted drying processes of food materials.

Energy efficiency (%)

Material; Varied parameters

Reference

4.2–5.6 15–18 10–8.3

44.2–53.7 15.3–12.8 23–26

Soysal et al., 2006 Holtz et al., 2010 Sunjka et al., 2004

MWCD

36.4–29.7

7.7–6.3

MWCD

26.3–62.0

8.7–3.7

MWVD

Beginning period: 60–20;

11.5–3.8

Parsley; load Potato, bread Cranberries; Microwave density and mode Longan; Microwave density, air temperature Garlic cloves; Air velocity, microwave density Chili; Pressure, rotational speed of drum

11.5

MWVD

Constant rate period: 20; Final drying period: 20–700; 2.7–4.9

86.5–46.9

Yongsawatdigul and Gunasekaran, 1996

MWVD

8.6–9.5

26.7–24.2

MWVD

3.3–7.7

43–71

MWFD

49.6

4.6

Cranberries; Pressure, microwave power, microwave mode Tomato; Sequence of combination with hot-air drying Cranberries; Microwave density and microwave mode Sea cucumber

Drying process

SEC (MJ kg

MWCD MWCD MWCD

1

H2O)

Varith et al., 2007

Sharma and Prasad, 2006b Kaensup et al., 2002

11.5–0.32

Durance and Wang, 2002 Sunjka et al., 2004

Duan et al., 2010b

material is low (e.g., during the final drying stage). Consequently, MWFD is difficult to use in industrial applications due to the potential for plasma discharge problems, and it becomes necessary to control the process parameters (vacuum pressure and microwave power intensity) in order to avoid the appearance of this phenomenon. Duan et al. (2010b) suggested that pressure control in the range of 50–100 Pa, combined with microwave power decreasing in time, would ensure that no corona discharge occurred during microwave-assisted freeze-drying. It is important to design a chamber with as small a localized concentration of electromagnetic field as possible (Duan et al., 2010a). The lowest reported electric field intensities for dielectric breakdown in vapor and food have been 20 kV m 1 and 90 kV m 1, respectively (Cui, 2004).

303

304

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

9.4.5 Changes in Dielectric Properties

As moisture is removed during the course of microwave-assisted drying, the dielectric constant e0 and the dielectric loss factor e00 normally decrease gradually, which can be seen in the descending parts of the curves of Fig. 9.17. However, during microwave-assisted freeze-drying, when most ice crystals have been removed, a sharp increase in e0 and e00 was observed by Regier and Schubert (2005) (cf. Fig. 9.17). This occurs because most of the water in the material is frozen during the initial period of the drying process, and this results in low dielectric properties (Wang et al., 2011a). In samples having the same moisture content, both,

Fig. 9.17 Variation of dielectric properties with time for restructured potato slices with different salt contents during MWFD. Reproduced with permission from Wang et al. (2011a).

9.4 Microwave-Assisted Drying Process

e0 and e00 are higher in the presence of salt than in its absence, which indicates a stronger microwave absorption capacity in the salty material. 9.4.6 Quality Changes in Food during Microwave-Assisted Drying

Following the water loss and temperature rises that occur during their microwaveassisted drying, food materials undergo complicated changes in their physicochemical properties, such as color, shape, texture, volatile compounds, retention of nutrients and bioactive ingredients, as well as sensory quality. Generally, the color and nutrient compounds are degraded to some degree, while shrinkage or expansion in shape or structure also occur, depending on the type of material being dried. The shrinkage of seeds, grains, fruits and vegetables has been reported extensively. Shrinkage is related to moisture content because, when moisture is removed from the tissues, a pressure imbalance is created between the inside and outside of the tissue, generating compressive stresses that lead to shrinkage. Factors affecting shrinkage include pre-drying (initial moisture content), osmotic dehydration, sample thickness, microwave power, the final moisture content, and the type of microwave-assisted drying method used. It has been reported that osmotic pretreatment improved volume retention from 20% to 50% for strawberries, and from approximately 20% to 60% for apples (based on fresh volume), compared to samples subjected to microwave-assisted vacuum drying without osmotic pre-treatment (Changrue et al., 2008). Beetroot cubes were dehydrated by convective drying in hot air at 60  C and also by a combination of convective predrying (until the moisture contents were 1.6, 0.6 or 0.27 kg kg 1 d.b.) and microwave-assisted vacuum-finish drying at 240, 360, or 480 W. The use of microwaves at the end of the drying period significantly reduced shrinkage compared to the purely convective method. Increasing the microwave wattage and decreasing the time of pre-drying also improved the quality of beetroot cubes dried using the combined method (Figiel, 2010). Figiel (2009) also investigated the changes in the relative volume of garlic of various sizes, and found that drying to a moisture content of about X ¼ 0.6 was accompanied by a decrease in the relative volume of the material. Beyond this value no further change in the volume of garlic slices was observed, although for whole- and half-cloves a marked increase in volume was apparent. The shrinkage was greater at 240 W microwave power and moisture content over X ¼ 0.6, compared to samples dried at 480 and 720 W that had similar values of relative volume. Slices dried at 240 W until X ¼ 0.11 had the smallest relative volume of all samples dried with the vacuum microwave method. However, their relative volume was still greater than that of slices dried using the convective method. In contrast to the commonly encountered shrinkage in materials of plant origin, some gel materials – such as restructured fish muscle, imitation cheese, and restructured blue honeysuckle snacks containing starch – also display expansion. Figure 9.18 shows the typical expansion process of cheese containing starch during

305

306

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

Fig. 9.18 Dynamics of temperature increase ( ), weight loss (&) and degree of expansion (Î) during microwave heating of imitation cheese samples (fat : starch ratio ¼ 0). Dashed vertical lines with arrows

separate the regions of heating up, moisture loss and expansion, and moisture loss and burning. Reproduced with permission from Arimi et al. (2008).

microwave heating and drying. Three major effects were observed to occur: a temperature increase; expansion; and weight loss. The time course of the above processes may be divided into three distinct regions: heating up; moisture loss and expansion; and moisture loss and burning (Arimi et al., 2008). Typically, food products dried by microwave-assisted processes can display three types of structural change:  Shrinkage with porous structures such as most fruits, vegetables, and mushrooms.  A nonuniform pillow-shaped expansion with open structure, such as potato cubes produced by microwave-enhanced spouted bed drying, restructured blue honeysuckle snacks containing starch produced by MWVD, and microwave-expanded imitation cheese (Arimi et al., 2008; Yan, 2011; Liu et al., 2009).  A uniform expansion with porous structures such as restructured fish muscle (Wang et al., 2013). Both, shrinkage and expansion indicate damage of the original structure of the material, which can be undesirable if the dried food is to be consumed after rehydration. For a dried snack, however, expansion may be considered as a desirable indication of good crispness of the product. Several groups have correlated the volume or size changes of food materials during microwave-assisted drying with the moisture content (Tab. 9.3).

9.4 Microwave-Assisted Drying Process Tab. 9.3 Some correlations for the change of volume of food materials during microwaveassisted drying; (see volume and chapter notation for symbols).

CDMWVD MWVD MWVD MWCD MWVD

Sample

Model

Reference

Beetroots

V=V i ¼ aX þ b

Figiel, 2010

Garlic Button mushrooms Imitation cheese

V=V i ¼ a bX þ cX 2 V=V i ¼ a þ bðX=X i Þ

Figiel, 2009 Giri and Prasad, 2006

½ðV V i Þ=V i Š  100 ¼ 70:72 0:012 exp½17:9X Š i¼3 P d=di ¼ ai X i ðradial shrinkageÞ

Arimi et al., 2010

Carrot

i¼1

Nahimana and Zhang, 2011

In terms of end-product quality, foods processed by microwave-assisted drying often rank between air-dried and freeze-dried products (Orsat et al., 2007). Compared to convective drying, products dried via microwave-assisted processes have a reduced shrinkage, a faster rehydration with a higher capacity, a fresher-like color, greater retention of color or aroma, a higher nutritional value, and a softer or crispier texture. Moreover, microwave drying provides an effective bactericidal action in the dried product (Xu et al., 2004). Quality is further improved when vacuum is used, as the thermal and oxidative stresses are reduced. In some cases, MWVD can yield dehydrated products with characteristics that are similar to or even better than those dried by freeze-drying; for example, MWVD potatoes retain more fresh-like color than freeze-dried potatoes. Typically, the use of MWVD results in crispy and porous textures rather than the spongy and soft texture of freeze-dried foods. The impact of each microwave-assisted drying system on product quality is specific and heavily dependent on the drying conditions used. For example, microwave-assisted drying can conserve volatile aroma compounds at a high level when drying materials of plant origin, and can also produce good aroma and flavor in dehydrated fish products due to the development of aromatic aldehydes, including 2-methylpropanal, 3-methyl-butanal, and furfural (Wang et al., 2013). Rehydrated MWVD-restructured fish was inferior to products produced by conventional drying methods in terms of sensory evaluation, whereas rehydrated MWVD shrimp, carrot slices and russet potatoes were equal to or better than rehydrated freeze-dried samples. Microwave vacuumdried cranberries showed a softer texture and were less tough than cranberries processed by microwave-assisted convective drying. MWCD cranberries were better appreciated by a sensory panel than was the MWVD product, although berries that had undergone both microwave drying methods were less desirable than hot-air-dried cranberries (Sunjka et al., 2004). The application of excessive microwave energy frequently leads to serious charring within the materials due to thermal runaway (as noted above). Charring is also closely

307

308

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

related to the material’s properties, with heterogeneous, square, thick and shrinkable materials being susceptible to the development of charring spots due to their nonuniform dielectric properties (Wang et al., 2013; Holtz et al., 2010; Ahrn
e et al., 2007). By using optimal microwave-assisted drying conditions, the high-quality attributes desired by consumers and the excellent functional attributes of the dehydrated products can be provided and relatively easily maintained.

9.5 Microwave-Assisted Drying Process Control and Optimal Operation 9.5.1 Factors Controlling Microwave-Assisted Drying Processes

In order to achieve high-quality products with specific usage and sensory quality and also to save energy, a reasonable microwave-assisted drying process control is required, but this is a complicated and systematic task. The factors to be considered are shown schematically in Fig. 9.19, and have been discussed in Section 9.4. According to the scheme shown in Fig. 9.19, a relatively optimal drying process can be obtained by seeking a suitable dryer and drying method for a specific food product, as well as by selecting appropriate operating parameters. 9.5.2 Optimal Operation Strategy

Generally speaking, it is expected that multiple objectives are accomplished during the drying process, such as the least energy consumption and costs as well as the highest product quality. The quality characteristics and performance measures of the drying systems usually are described by several response variables that are greatly affected by many factors (see Fig. 9.19), and their interactions. Some of these variables should be maximized, while others should be minimized. In many cases, these responses are in competition one with another; that is, to improve one response may have an opposite effect on another response, further complicating the situation. Numerical optimization techniques, graphical optimization techniques and constrained optimization techniques have all been successfully used for process optimization (Shi et al., 2008). For numerical optimization techniques, either an overall desirability function or a synthetic evaluation index should be established. Desirability is an objective function that ranges from zero (least desirable) to one (most desirable) (Giri and Prasad, 2007b). Numerical optimization identifies a point that maximizes the desirability function or the synthetic evaluation index. Reported examples refer to optimization of the drying of edamames by a combination of hot air and vacuum microwave (Hu et al., 2006), optimization of the drying conditions for vacuum-assisted microwave drying of

9.5 Microwave-Assisted Drying Process Control and Optimal Operation Microwave-assisted drying factors

Drying time Energy consumption Product quality

Drying method

Combination type Combination sequence

Operating parameters

Temperature Microwave emission mode Microwave power Pressure level Air velocity, Air humidity Load and conversion point

Material characteristics

Species Variety Maturity State Composition Blanching Osmotic treatment Shape Size Thickness

Process optimum

Optimal drying conditions

Scale-up

Fig. 9.19 Controlling factors of microwave-assisted drying processes, and process intensification strategy.

green peas (Pisum sativum L.) (Chauhan and Srivastava, 2009), and optimization of the process parameters for the microwave vacuum drying of apple slices, using the response surface method (Han et al., 2010). In the graphical optimization technique, contour diagrams of the different response variables are superimposed and the optimal conditions are determined within the superposition region. Process parameters for the microwave freeze-drying of instant vegetable soup were optimized using this technique by Wang et al. (2010b). Sharma and Prasad (2006b) optimized the process parameters for the microwave-convective drying of garlic cloves using a constrained optimization technique (Tukey’s multicomparison pair-wise test), in an effort to identify the best compromise between quality and drying time. The results revealed that a

309

310

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

microwave power of 40 W, an air temperature of 70  C and an air velocity of 1.0 m s 1 produced dehydrated garlic cloves of good quality and involved a low specific energy consumption during the drying process.

9.6 Concluding Remarks

Microwave-assisted drying provides unique opportunities for the development of advanced food-drying technologies. The main advantage of combining microwaves with other drying methods is to sharply reduce the drying time required for a product and, as long as the product temperature is controlled, then newly developed drying methods can be used to improve product quality. Unfortunately, most reports on technologies of microwave-related combination drying have been based on laboratory-scale systems, and there remains a need for further studies to bridge the gap between laboratory-based research and industrial applications. Currently, a large number of processing choices exists for the microwave-assisted dehydration of foods. Typically, each of these drying systems is product-specific in terms of end product quality and the functionality of its components, and consequently the choice of processing method will depend heavily on the intended use of the product and its market value. Yet, any given process can be detrimental to a particular product, with perhaps significant reductions in the retention of bioactive ingredients and/or their activities. Moreover, the mechanisms of product degradation and bioactivity are related to the processing conditions – principally of time and temperature and their control, with temperature-controlled microwave drying having been proven as the best processing method. In regard of materials, most studies on microwave-assisted drying have focused on agricultural products such as potatoes, carrots, and apples, although some have involved fabricated/formulated products. The main advantages of fabricated products, as noted by Gebhardt (1996), include reproducibility, uniformity and a lack of defects when compared to heterogeneous materials such as raw potato. It is clear that Life Cycle Assessments (LCAs) should be made of the various microwaveassisted drying techniques available, in an effort to determine a definitive selection strategy for dryers and drying processes. In addition, as concerns relating to climate change and global warming begin to affect dryer operations, the need for LCA analyses will become even greater in the future.

Additional Notation Used in Chapter 9

c dp E

speed of light penetration depth electric field strength

ms m Vm

1

1

9.6 Concluding Remarks

f MR P Pabs PMW V X

s –

1

frequency moisture ratio, MR ¼ (X Xeq)/ (Xi Xeq) microwave power density absorbed microwave power microwave power volume moisture content

Wm 3 W W m3 kg water (kg dry matter)

permittivity of free space dielectric constant dielectric loss factor relative complex permittivity wavelength

Fm – – – m

Greek Letters

e0 e0 e00 e l

Subscripts

i f 0

initial final freeze space

Abbreviations

CD d.b. FD LCA MW MWCD MWFD MWSD MWVD SEC w.b.

convective drying dry basis freeze-drying life cycle assessment microwaves microwave-assisted convective drying microwave-assisted freeze-drying microwave-assisted spouted bed drying microwave-assisted vacuum-drying specific energy consumption wet basis

1

1

311

312

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality

References Ahrn
e, L., Pereira, N. R., Staack, N., Floberg, P., 2007. Microwave convective drying of plant foods at constant and variable microwave power. Drying Technol. 25: 1149– 1153. Al-Harahsheha, M., Al-Muhtaseb, A. H., Mageec, T. R. A., 2009. Microwave drying kinetics of tomato pomace: Effect of osmotic dehydration. Chem. Eng. Process. 48: 524– 531. Arikan, M. F., Ayhan, Z., Soysal, Y., Esturk, O., 2012. Drying characteristics and quality parameters of microwave-dried grated carrots. Food Bioprocess Technol. 5: 3217–3229. Arimi, J. M., Duggan, E., O’Riordan, E. D., O’Sullivan, M., Lyng, J. G., 2008. Microwave expansion of imitation cheese containing resistant starch. J. Food Eng. 88(2): 254–262. Arimi, J. M., Duggan, E., O’Riordan, E. D., O’Sullivan, M., Lyng, J. G., 2010. Effect of moisture content and water mobility on microwave expansion of imitation cheese. Food Chem. 121: 509–516. Bai-Ngew, S., Therdthai, N., Dhamvithee, P., 2011. Characterization of microwave vacuum-dried durian chips. J. Food Eng. 104: 114–122. Botha, G. E., Oliveira, J. C., Ahrn
e, L., 2012. Microwave assisted air drying of osmotically treated pineapple with variable power programmes. J. Food Eng. 108: 304– 311. Changrue, V., Orsat, V., Raghavan, G. S. V., 2008. Osmotically dehydrated microwave vacuum drying of strawberries. J. Food Process. Preserv. 32(5): 798–816. Chauhan, A. K. S., Srivastava, A. K., 2009. Optimizing drying conditions for vacuumassisted microwave drying of green peas (Pisum sativum L.). Drying Technol. 27(6): 761–769. Clary, C. D., Wang, S. J., Petrucci, V. E., 2005. Fixed and incremental levels of microwave power application on drying grapes under vacuum. J. Food Sci. 70(5): 344–349. Clary, C. D., Wang, S. J., Petrucci, V. E., 2007. Improved grape quality using microwave vacuum drying associated with temperature control. J. Food Sci. 72(1): E23–E28.

Cui, Z. W., Xu, S. Y., Sun, D. W., 2003. Dehydration of garlic slices by combined microwave-vacuum and air drying. Drying Technol. 21: 1173–1184. Cui, Z. W., 2004. Mathematical modeling of microwave-vacuum drying and its application in food processing. Diss., Jiangnan University, Wuxi, VRC. Cui, Z. W., Xu, S. Y., Sun, D. W., 2004. Microwave-vacuum drying kinetics of carrot slices. J. Food Eng. 65: 157–164. Cui, Z. W., Xu, S. Y., Sun, D. W., Chen, W., 2005. Temperature changes during microwave-vacuum drying of sliced carrots. Drying Technol. 23(5): 1057–1074. Dev, S. R. S., Geetha, P., Orsat, V., Gari
epy, Y., Raghavan, G. S. V., 2011. Effects of microwave-assisted hot air drying and conventional hot air drying on the drying kinetics, color, rehydration, and volatiles of Moringa oleifera. Drying Technol. 29(12): 1452–1458. Duan, X., Zhang, M., Mujumdar, A. S., Wang, R., 2010a. Trends in microwave-assisted freeze drying of foods. Drying Technol. 28: 1– 10. Duan, X., Zhang, M., Mujumdar, A. S., Wang, S. J., 2010b. Microwave freeze drying of sea cucumber (Stichopus japonicus). J. Food Eng. 96: 491–497. Durance, T. E., Wang, J. H., 2002. Energy consumption, density, and rehydration rate of vacuum microwave- and hot-air convection-dehydrated tomatoes. J. Food Sci. 67: 2212–2216. Esturk, O., 2012. Intermittent and continuous microwave-convective air-drying characteristics of sage (Salvia officinalis) leaves. Food Bioprocess Technol. 5(5): 1664–1673. Feng, H., Tang, J., 1998. Microwave finish drying of diced apples in spouted bed. J. Food Sci. 63(4): 679–683. Feng, H., Tang, J., Cavalieri, R. P., 1999. Combined microwave and spouted bed drying of diced apples: Effect of drying conditions on drying kinetics and product temperature. Drying Technol. 17(10): 1981–1998. Feng, H., Yin, Y., Tang, J., 2012. Microwave drying of food and agricultural materials:

References Basics and heat and mass transfer modeling. Food Eng. Rev. 4: 89–106. Figiel, A., 2009. Drying kinetics and quality of vacuum-microwave dehydrated garlic cloves and slices. J. Food Eng. 94: 98–104. Figiel, A., 2010. Drying kinetics and quality of beetroots dehydrated by combination of convective and vacuum-microwave methods. J. Food Eng. 98, 461–470. Funebo, T., Ohlsson, T., 1998. Microwaveassisted air dehydration of apple and mushroom. J. Food Eng. 38: 353–367. Gebhardt, B., 1996. Oils and fats in snack foods, in Bailey’s industrial oil and fat products, 5th ed., Vol. 3 (ed. Y. H. Hui). Wiley, New York, USA. Giri, S. K., Prasad, S., 2006. Modelling shrinkage and density changes during microwave-vacuum drying of button mushroom. Int. J. Food Properties 9: 409–419. Giri, S. K., Prasad, S., 2007a. Drying kinetics and rehydration characteristics of microwave-vacuum and convective hotair dried mushrooms. J. Food Eng. 78: 512–521. Giri, S. K., Prasad, S., 2007b. Optimization of microwave-vacuum drying of button mushrooms using response-surface methodology. Drying Technol. 25(5): 901– 911. Gowen, A., Abu-Ghannam, N., Frias, J., Oliveira, J., 2006. Optimisation of dehydration and rehydration properties of cooked chickpeas (Cicer arietinum L.) undergoing microwave–hot air combination drying. Trends Food Sci. Technol. 17: 177–183. Gowen, A., Abu-Ghannam, N., Frias, J., Oliveira, J., 2008. Modeling dehydration and rehydration of cooked soybeans subjected to combined microwave-hot air drying. Innovative Food Sci. Emerg. Technol. 9: 129– 137. Han, Q. H., Yin, L. J., Li, S. J., Yang, B. N., Ma, J. W., 2010. Optimization of process parameters for microwave vacuum drying of apple slices using response surface method. Drying Technol. 28(4): 523–532. Holtz, E., Ahrn
e, L., Karlsson, H. T., Rittenauer, M., Rasmuson, A., 2009. The role of processing parameters on energy efficiency during microwave convective

drying of porous materials. Drying Technol. 27(2): 173–185. Holtz, E., Ahrn
e, L., Rittenauer, M., Rasmuson, A., 2010. Influence of dielectric and sorption properties on drying behaviour and energy efficiency during microwave convective drying of selected food and non-food inorganic materials. J. Food Eng. 97(2): 144– 153. Hu, Q. G., Zhang, M., Mujumdar, A. S., Xiao, G. N., Sun, J. C., 2006. Drying of edamames by hot air and vacuum microwave combination. J. Food Eng. 77: 977–982. Jangam, S. V., Law, C. L., Mujumdar, A. S., 2011. Drying of foods, vegetables and fruits, Vol. 3, ISBN: 978-981-08-9426-9. Kaensup, W., Chutima, S., Wongwises, S., 2002. Experimental study on drying of chilli in a combined microwave-vacuum rotary drum dryer. Drying Technol. 20(10): 2067–2079. Li, Z., Raghavan, G. S. V., Orsat, V., 2010. Temperature and power control in microwave drying. J. Food Eng. 97: 478–483. Lin, T. M., Durance, T. D., Scaman, C. H., 1998. Characterization of vacuum microwave, air and freeze dried carrot slices. Food Res. Int. 31: 111–117. Lin, T. M., Durance, T. D., Scaman, C. H., 1999. Physical and sensory properties of vacuum microwave dehydrated shrimp. J. Aquat. Food Prod. Technol. 8: 41–53. Liu, C. H., Zheng, X. Z., Jia, S. H., Ding, N. Y., Gao, X. C., 2009. Comparative experiment on hot-air and microwave-vacuum drying and puffing of blue honeysuckle snack. Int. J. Food Eng. 5: 1–9. Maskan, M., 2000. Microwave/air and microwave finish drying of banana. J. Food Eng. 44: 71–78. Mousa, N., Farid, M., 2002. Microwave vacuum drying of banana slices. Drying Technol. 20: 2055–2066. Mujumdar, A. S., 2006. Handbook of industrial drying (3rd ed.). CRC Press, Boca Raton, Florida, USA. Mujumdar, A. S., Law, C. L., 2010. Drying technology: Trends and applications in postharvest processing. Food Bioprocess Technol. 3: 843–852. Nahimana, H., Zhang, M., 2011. Shrinkage and color during microwave vacuum drying of carrot. Drying Technol. 29(7): 836–847.

313

314

9 Microwave-Assisted Drying of Foods – Equipment, Process and Product Quality Orsat, V., Yang, W., Changrue, V., Raghavan, G. S. V., 2007. Microwave-assisted drying of biomaterials. Food Bioprod. Process. 85(C3): 255–263. Pereira, N., Marsaioli, A., Ahrn
e, L., 2007. Effect of microwave power, air velocity and temperature on final drying of osmotically dehydrated bananas. J. Food Eng. 81: 79–87. Pranabendu, M., Venkatesh, M., 2009. Optimization of microwave-vacuum drying parameters of Saskatoon berries using response surface methodology. Drying Technol. 27(10): 1089–1096. Ratti, C., 2001. Hot air and freeze-drying of high-value foods: A review. J. Food Eng. 49(4): 311–319. Ratti, C., 2008. Advances in food dehydration. CRC Press, Boca Raton, USA. Regier, M., Schubert, H., 2005. The microwave processing of foods. Woodhead Publishing, Cambridge, UK. Sahin, S., Sumnu, S. G., 2006. Physical properties of foods. Springer, Berlin Germany. Sharma, G. P., Prasad, S., 2006a. Specific energy consumption in microwave drying of garlic cloves. Energy 31: 1921–1926. Sharma, G. P., Prasad, S., 2006b. Optimization of process parameters for microwave drying of garlic cloves. J. Food Eng. 75: 441–446. Shi, Q. L., Xue, C. H., Zhao, Y., Li, Z. J., Wang, X. Y., Luan, D. L., 2008. Optimization of processing parameters of horse mackerel (Trachurus japonicus) dried in a heat pump dehumidifier using response surface methodology. J. Food Eng. 87: 74–81. Shivhare, U. S., Orsat, V., Raghavan, G. S. V., Ribeiro, C. P., Passos, M. L., 2010. Innovation in Food Engineering: New Techniques and Products. CRC Press, Boca Raton, USA. Sosa-Morales, M. E., Valerio-Junco, L., L
opezMalo, A., García, H. S., 2010. Dielectric properties of foods: Reported data in the 21st century and their potential applications. LWT – Food Sci. Technol. 43(8): 1169–1179. Souraki, B. A., Andr
es, A., Mowla, D., 2009. Mathematical modeling of microwaveassisted inert medium fluidized bed drying of cylindrical carrot samples. Chem. Eng. Process. 48: 296–305.

€ € 2006. Soysal, Y., Oztekin, S., Eren, O., Microwave drying of parsley: Modelling, kinetics, and energy aspects. Biosystems Eng. 93(4): 403–413. Soysal, Y., Ayhan, Z., Esturk, O., Arikan, M. F., 2009. Intermittent microwaveconvective drying of red pepper: Drying kinetics, physical (colour and texture) and sensory quality. Biosystems Eng. 103: 455–463. Sunjka, P. S., Rennie, T. J., Beaudry, C., Raghavan, G. S. V., 2004. Microwave– convective and microwave–vacuum drying of cranberries: A comparative study. Drying Technol. 22(5): 1217–1231. Uprit, S., Mishra, H. N., 2003. Microwave convective drying and storage of soy-fortified paneer. Food Bioprod. Process. 81(2): 89–96. Varith, J., Dijkanarukkul, P., Achariyaviriya, A., Achariyaviriya, S., 2007. Combined microwave-hot air drying of peeled longan. J. Food Eng. 81(2): 459–468. Wang, J., Xiong, Y.-S., Yu, Y., 2004. Microwave drying characteristics of potato and the effect of different microwave powers on the dried quality of potato. Eur. Food Res. Technol. 219: 500–506. Wang, R., Zhang, M., Mujumdar, A. S., Sun, J. C., 2009. Microwave freeze-drying characteristics and sensory quality of instant vegetable soup. Drying Technol. 27(9): 962–968. Wang, R., Zhang, M., Mujumdar, A. S., 2010a. Effect of osmotic dehydration on microwave freeze-drying characteristics and quality of potato chips. Drying Technol. 28(6): 798–806. Wang, R., Zhang, M., Mujumdar, A. S., 2010b. Effect of food ingredient on microwave freeze drying of instant vegetable soup. LWT – Food Sci. Technol. 43: 1144–1150. Wang, R., Zhang, M., Mujumdar, A. S., Sun, J. C., Jiang, H., 2011a. Effect of salt and sucrose content on dielectric properties and microwave freeze drying behavior of restructured potato slices. J. Food Eng. 106: 290–297. Wang, Y., Zhang, M., Mujumdar, A. S., 2011b. Trends in processing technologies for dried aquatic products. Drying Technol. 29: 382–394. Wang, Y., Zhang, M., Mujumdar, A. S., Mothibe., K. J., 2013. Quality changes of dehydrated restructured fish product from

References silver carp (Hypophthalmichthys molitrix) as affected by drying methods. Food Bioprocess Technol. 6(7): 1664–1680. Wojdyo, A., Figiel, A., Oszmianski, J., 2009. Effect of drying methods with the application of vacuum microwaves on the bioactive compounds, color, and antioxidant activity of strawberry fruits. J. Agric. Food Chem. 57: 1337–1343. Xu, Y. Y., Zhang, M., Mujumdar, A. S., Sun, J. C., 2004. Studies on hot air and microwave vacuum drying of wild cabbage. Drying Technol. 22(9): 2201–2209. Yan, W. Q., Zhang, M., Huang, L.-L., Tang, J., Mujumdar, A. S., Sun, J. C., 2010. Studies on different combined microwave drying of carrot pieces. Int. J. Food Sci. Technol. 45: 2141–2148. Yan, W. Q., 2011. Studies on drying uniformity and model of microwave spouted bed dried cutting tuber vegetables pieces. Diss., Jiangnan University, Wuxi, VRC.

Yongsawatdigul, J., Gunasekaran, S., 1996. Microwave–vacuum drying of cranberries, Part I: Energy use and efficiency. J. Food Process. Preserv. 20: 121–143. Zhang, M., Li, C. L., Ding, X. L., 2003. Optimization for preservation of selenium in sweet pepper under low-vacuum dehydration. Drying Technol. 21(3): 569–579. Zhang, M., Li, C. L., Ding, X. L., 2005. Effects of heating conditions on the thermal denaturation of white mushroom suitable for dehydration. Drying Technol. 23(5): 1119–1125. Zhang, M., Tang, J., Mujumdar, A. S., Wang, S., 2006. Trends in microwave-related drying of fruits and vegetables. Trends Food Sci. Technol. 17: 524–534. Zhang, M., Jiang, H., Lim, R. X., 2010. Recent developments in microwave-assisted drying of vegetables, fruits, and aquatic products: Drying kinetics and quality considerations. Drying Technol. 28(11): 1307–1316.

315

317

10 Infrared Drying German Efremov 10.1 Introduction

Drying (dehydration) is, by definition, the removal of moisture (as either water or other volatile compounds) from solids, continuous sheets, solutions and pastes, with the purpose of obtaining a solid product that is sufficiently low in water content. Drying is the most diverse among engineering unit operations, and also the most energy-consuming. The consumption of energy for drying of different materials depends on the type of drying process used. The common modes of heat transfer for drying are:    

Convection from a hot gas in contact with the material or evaporation in vacuum. Conduction from a hot, solid surface in contact with the material. Radiation from a hot gas or surface. Heat generation within the material by dielectric, radiofrequency, or microwave heating.

Drying heat may be supplied by using electric or natural gas-fired radiators, or volumetrically in the microwave or radiofrequency range by placing the wet solid in radiation fields. By exposing an object to infrared (IR) radiation (wavelength of 0.78 to 1000 mm), the thermal energy generated can effectively be absorbed by materials. Infrared radiation can be felt as radiant heat when standing in front of a fire. As radiant heat flux can be adjusted over a wide range of wavelengths it is possible to obtain high drying rates for different surface-wet materials. In direct-heat dryers, the transfer of heat by convection from hot gases to the wet material is often supplemented by thermal radiation from hot surfaces that surround the material. This radiant-heat contribution is usually minor, and ignored. The use of thermal radiation as the major source of heat is a proven technology for the drying of certain films, sheets, textiles and coatings (Devahastin, 2000). During the 1950s A.V. Lykow and coworkers reported the results of their theoretical and experimental studies of infrared drying (Ginzburg, 1969).

Modern Drying Technology Volume 5: Process Intensification, First Edition. Edited by Evangelos Tsotsas and Arun S. Mujumdar. Ó 2014 Wiley-VCH Verlag GmbH & Co. KGaA. Published 2014 by Wiley-VCH Verlag GmbH & Co. KGaA.

318

10 Infrared Drying

Infrared drying provides significant advantages over conventional drying, including reduced heating time, uniform surface heating, reduced quality losses, versatile, simple and compact equipment, and significant energy savings. However, the most popular applications involve use of combined drying. Consequently, IR drying can be coupled effectively with vacuum and heat pump heating to permit the removal of evaporated moisture. Moreover, IR dryers may be used either continuously or intermittently to improve product quality. The aim of this chapter is to evaluate existing knowledge in the area of IR heating, to provide insight for the relationship between product properties and drying processes, and to present an up-to-date view on further research. The chapter also encompasses applications of IR heating in drying operations. Many industrial dryers can be modified to accommodate IR heaters; indeed, combined convective, microwave and IR dryers have been shown to be very attractive in this respect. Other combinations can also be used for drying.

10.2 Radiation Heat Transfer 10.2.1 General Principles

Radiation heat transfer refers to electromagnetic radiation, mainly in the IR range. From a practical viewpoint, the sole effect is thermal in this range – that is, heating of the receiving body by radiation from the emitting body. Above the absolute temperature of zero Kelvin ( 273  C), all substances emit electromagnetic radiation with the speed of light in a vacuum (c ¼ 2.998  108 m s 1). The frequency of the wave, f, is related to the wavelength, l, f ¼ cl;

ð10:1Þ

and varies as a function of the body temperature. The energy transmitted by the wave depends on its frequency. In contrast to conduction and convection, heat transfer by radiation does not require the presence of a material medium. Electromagnetic radiation is created by oscillatory electric charges (electrons and ions), and the frequency of oscillation determines the wavelength and type of radiation emitted. Radio, television waves and microwaves have long wavelengths, whereas ultraviolet, X-rays and gamma rays exist at short wavelengths. In between these is a range of wavelengths known as the optical thermal spectrum, with IR and visible light. Thermal radiation is, therefore, of the same nature as any other type of electromagnetic radiation, but has the strongest heating effect. The wavelength range of thermal radiation is 0.1 to 100 mm; IR radiation falls in this spectrum and is conventionally classified as near-infrared (NIR; 0.75–3.00 mm), medium infrared (MIR; 3.00–25 mm), and far infrared (FIR; 25–100 mm) (see Sandu, 1986). As shown in the spectrum of radiation (Fig. 10.1), the electromagnetic wavelength, l, depends on the manner in which it is generated. Heat transfer can

10.2 Radiation Heat Transfer

Fig. 10.1 Spectrum of radiation.

be viewed as the propagation of electromagnetic waves, consisting of electric and magnetic fields that oscillate at right-angles to each other and to the direction in which the radiation travels. So, radiation heat transfer takes place by the transfer of electromagnetic waves from a hot, opaque surface through a nonabsorbing gas or vacuum to the material being dried. When radiant electromagnetic energy impinges on the surface of a body, it may induce changes in the electronic, vibrational and rotational states of atoms and molecules. The type of mechanism leading to energy absorption is determined by the wavelength range of the incident radiative energy and can be categorized as follows:  Changes in the electronic state corresponding to the wavelength range 0.2 to 0.7 mm (ultraviolet and visible rays).  Changes in the vibrational state corresponding to wavelength range 2.5 to 1000 mm.  Changes in the rotational state corresponding to wavelengths above 1000 mm (microwaves) (Decareau, 1985). The broad region occupied by infrared extends from 0.76 mm (i.e., just beyond the red end of the visible end of the spectrum) to 400 mm. However, the radiation used for process heating occurs between wavelengths of 1 and 5 mm in order to obtain adequate source temperatures. The radiation emitted by a body can be determined if the temperature and nature of its surface (emissivity) are known. 10.2.2 Reflection, Absorption, and Transmission

A freely propagating electromagnetic wave transports the same amount of energy per second. When the wave enters a solid the increase in oscillations of the electrons causes the energy density to increase, so that the wave travels more slowly. An electric field is set up by the oscillating electrons, and this causes a part

319

320

10 Infrared Drying

Q

Qρ n

Qα

Qτ Fig. 10.2 Interaction of thermal radiation with material (incident energy flow rate, Q, absorbed energy flow rate, Qa, reflected energy flow rate, Qr, and energy flow rate transmitted through the material, Qt).

of the electromagnetic field to be reflected. Materials made only of atoms with tightly bound electrons absorb very little energy and are good heat insulators. In general, however, radiation is absorbed, reflected or transmitted by solid materials. By referring the absorbed energy flow rate, Qa, which is the energy converted to heat, the energy flow rate reflected from the surface of the material, Qr, and the energy flow rate transmitted through the material, Qt, to the incident energy flow rate, Q, the absorptivity, a ¼ Qa /Q, reflectivity, r ¼ Qr /Q, and transmissivity, t ¼ Qt /Q, can be defined (Fig. 10.2). It is, by definition a þ r þ t ¼ 1:

ð10:2Þ

Materials may be classified based on their absorptivity, reflectivity and transmissivity. A black body absorbs 100% of the radiation and does not transmit or reflect any (t ¼ 0, r ¼ 0, a ¼ 1). A body that transmits all radiation is termed “transparent,” and is characterized by t ¼ 1, r ¼ 0, a ¼ 0. A body that does not allow any radiation to be transmitted through it is termed “opaque,” and is characterized by t ¼ 0, r þ a ¼ 1. Whilst most solids are opaque, liquids and some solids such as rock salt or glass, on the other hand, have a high transmissivity and so are transparent to radiation. Only the absorbed portion of the incident energy causes heating. A black body is a body that emits the maximal possible amount of thermal radiation at a given temperature. The emissive power of a black body is a sole function of its temperature. Figure 10.1 shows that the peak wavelength emitted by a black body decreases as the temperature of the emitting body increases. The reflection may be either regular (also termed specular) or diffuse (scattering), which depends on the surface finish of the material. In the former case, the angle of incidence of the radiation is equal to the angle of reflection due to highly polished or smooth surface. When the surface has roughnesses larger than the wavelength, radiation is reflected diffusely in all directions. Generally, solid bodies absorb all of the radiation in a very narrow layer near the surface, and this is a very important consideration when modeling the heat-transfer process since, mathematically, this concept transforms a term within the energy balance into a boundary

10.2 Radiation Heat Transfer

condition (Mujumdar, 2007). Infrared heat transmission was used incidentally in the past accompanying other types of heat transfer during dehydration, but IR dryers are now designed to utilize radiant heat as the primary source. 10.2.3 Infrared Spectrum

Visible light is characterized by the fact that the radiation can be directed, focused and controlled by mirrors and lenses, and that prisms and gratings can be used for dispensing it into a spectrum. Ordinary sources of radiation in the visible spectrum (optical spectrum), such as tungsten filament lamps, fluorescent lamps and flames, consist of a very great number of molecules which have electric charges that oscillate independently of each other, producing a range of frequencies. The electromagnetic energy that is emitted from the surface of a heated body consists of a continuous spectrum of frequencies extending over a wide range. The spectral distribution and the amount of energy radiated depend on the temperature of the emitting surface. Measurements show that, for a given temperature, there is a definite frequency at which the radiated power is maximal (peak wavelength). The frequency of the peak is found to vary in direct proportion to the absolute temperature (Fig. 10.1). At room temperature, for example, the peak occurs in the far IR region of the spectrum, and there is no perceptible visible radiation emitted. However, at higher temperatures the maximal power is radiated at correspondingly higher frequencies, and at about 800 K a body begins to glow visibly. The properties of monochromatic radiation are termed “spectral,” and of polychromatic radiation “total” (Sandu, 1986). Hydrogen has relatively simple spectra, as the hydrogen atom consists of an electron and a proton. The electron has only 0.0005 the mass of the proton, and is able to inhabit only certain levels around the proton. To move from one level to another, the electron needs to gain or lose an amount of energy, called a quantum. For radio waves, a quantum is about 0.000 004 eV, for IR a quantum is about 0.004 eV, and for X-rays and c-rays it is about 40 000 eV. Electromagnetic radiation can be considered as stream of photons, which are massless particles traveling in a wave-like pattern. Movement between the lowest levels produces a photon of far-ultraviolet radiation, movement between the next lowest levels produces visible light and near-ultraviolet radiation, whilst movement between the middle levels produces IR radiation. The total amount of radiation emitted by a body per unit area and time is called the total emissive power, E, and this depends on the temperature and the surface characteristics of the body. This energy is emitted from a surface in all directions and at all wavelengths. A black body has, by definition, the maximal possible emissive power, Eb, which depends only on temperature. The emissivity of any body, e, is then defined as the ratio of its total emissive power E to that of a black body at the same temperature, e ¼ E=E b :

ð10:3Þ

321

322

10 Infrared Drying

For a black body, the monochromatic emissive power (emissive power at a certain wavelength) is expressed by Planck’s law of radiation: E b;l ¼

C
1  C2 l5 exp lT 

; 1

ð10:4Þ

where C1 ¼ 3.745  10 16 W m2 and C2 ¼ 1.439  10 2 K m. The monochromatic emissivity of any body is defined as el ¼ El/Eb,l. Kirchhoff’s law states that, under thermodynamic equilibrium (which requires all surfaces to be at the same temperature), the monochromatic absorptivity and emissivity of any body are equal. Equation 10.4 has a maximum that is related to the temperature by Wien’s displacement law: lmax T ¼ 2:898
10 3 ðm KÞ:

ð10:5Þ

Planck’s law of radiation is displayed graphically in Fig. 10.3 for different temperatures and shows that, with increasing wavelength, the spectral intensity of the black body, Eb,l, is first increased, then reaches a maximum, and finally decreases. The increase of temperature results in an appreciable increase of Eb,l, and in a shift of the maximum (peak wavelength) to shorter wavelengths according to Wien’s displacement law (dashed line in Fig. 10.3). The black body radiates energy at every wavelength, so that the curves approach the abscissa of Fig. 10.3 asymptotically. At 5000 K (not depicted) the peak wavelength is about 5  10 7 m, which is in the visible light region, in the yellow-green section.

Fig. 10.3 Planck’s law of radiation.

10.3 Classification, Research, and Applications of Radiation Drying Tab. 10.1

Total emissivity of various surfaces. Temperature ( C)

Material

Aluminum Brass Steel Stainless steel Building brick Fireclay brick Carbon black Water

100 100 100 100 20 500–1000 50–1000 0–100

Emissivity, e Nominal or oxidized

Polished

0.15 0.60 0.75 0.85 0.88–0.93 0.8–0.9 0.96 0.95–0.96

0.05 0.09 0.11 0.17 — — — —

Equation 10.4 may be integrated over all wavelengths to obtain the total emissive power of the black body as a function of temperature to Eb ¼

1 ð

E b;l dl ¼ sT 4 ;

ð10:6Þ

0

where s ¼ 5.67  10 8 W m 2 K 4 is the Stefan–Boltzmann constant. According to Eq. 10.6 (Stefan–Boltzmann law), the total energy radiated by a black body is proportional to the fourth power of the absolute temperature. A gray body is defined as having the same emissivity over the entire wavelength spectrum. Thus, Kirchhoff’s law may be applied to gray bodies independently of their temperature. The emissive power of a gray body is given by E g ¼ esT 4 :

ð10:7Þ

As the emissivity of real surfaces depends on the wavelength, average values over the relevant wavelength range are used in calculations. The average emissivity of various surfaces is listed in Tab. 10.1.

10.3 Classification, Research, and Applications of Radiation Drying 10.3.1 Classification

The various sources of electromagnetic radiation that can be used for drying include wavelengths ranging from the microwave to the visible spectrum (Fig. 10.4). Visible wavelengths are about 0.4 mm (violet light) to 0.7 mm (red light). The sun, the surface of which approximates a black body with temperature 5700 K, radiates most strongly in this range. Infrared wavelengths are longer than visible

323

324

10 Infrared Drying

Fig. 10.4 Spectrum of radiation for drying.

and up to about 100 mm. Thermal radiation is in the wavelength range of 0.1 to 1000 mm, and encompasses mainly the range of IR radiation. Infrared radiation can be further classified into the near-, mid- and far-infrared regions, as discussed above (see also Krishnamurthy et al. (2008)). In general, NIR waves are thermal and advantageous for use in drying, but MIR and FIR waves are less hot. For example, these wavelengths are used for the remote control units of televisions. Typically, the warmer an object, the more IR radiation it will emit. Drying heat may be supplied by using electric or natural gas-fired radiators, or volumetrically by placing the wet solid in dielectric fields in the microwave or radiofrequency range. As radiant heat flux can be adjusted locally over a wide range, it is possible to obtain high drying rates for surface-wet materials. Convection (gas flow) or vacuum operation is needed to remove the evaporated moisture. Infrared drying (4–8 mm wavelength) has found important applications in some niche markets, such as the drying of textiles, thick lumber, coated papers or printed sheets and films (Mujumdar, 2007). It is important to note that IR drying heat is absorbed at the surface, and that the application of continuous IR dryers is more common than that of batch IR dryers. In contrast to IR drying, dielectric drying involves the low-frequency, longwavelength end of the electromagnetic spectrum (Fig. 10.1), where radio waves and microwaves reside. With electrically nonconducting materials, heat is not absorbed at the surface but rather is generated throughout the material, thereby reducing the importance of heat conduction within the material. This is making such drying unique, and also enables the rate of energy dissipation in the material to be controlled over a wide range. Other advantages over conventional drying methods include:  Efficiency of energy usage, because the energy dissipation occurs mainly in the moisture rather than in the solid material.  Operation at low temperatures, thus avoiding high material surface temperatures.  More rapid drying. Dielectric drying is particularly useful for preheating materials and for removing the final traces of internal moisture to speed up drying at the tail end of the falling rate period. The main disadvantages of microwave drying include high capital and

10.3 Classification, Research, and Applications of Radiation Drying

operating (energy) costs. Only about 50% of the line power is converted into the electromagnetic field, and only a part of that is actually absorbed by the drying solid. Radiofrequency (RF) drying takes place at frequencies between 1 and 150 MHz (Seader et al., 2010). Although most materials are poor conductors of 50–60 Hz current, the impedance falls dramatically in the RF region. Such radiation can be used to heat the solid volumetrically, thus reducing the internal resistance to heat transfer. In RF drying, the energy is absorbed selectively by the water molecules. 10.3.2 Solar Drying

Solar radiation barely penetrates beyond the skin of the material, which absorbs only a part of the incident radiation depending on its wavelength. The term “solar dryer” is used for direct solar dryers, and is also reserved for a large variety of convective dryers whereby the products are not exposed to sun but are dried indirectly by air heated by solar energy. The convection of heated air in solar dryers may be either natural or forced. As a comprehensive treatment of solar drying has been provided in Chapter 6, Volume 4 of Modern Drying Technology, only a few examples will be described at this point. Sun drying at village level is common in tropical regions. Both, Bansal and Garg (1987) and Ekechukwu and Norton (1999) have referred to the dehydration of foods by direct exposure to radiation from the sun. Important quantities of fruits, vegetables, grains and fish are dried by this method, with sun-dried tomatoes being a specialty product of increasing popularity. A large proportion of the raisins, and practically all of the dried apricots and figs produced in the world, are sun-dried. Field-level experiments on the solar drying of pineapple using a solar tunnel dryer were conducted by Bala et al. (2003) in Bangladesh. The dryer consisted of a transparent plastic-covered flat plate collector and a drying tunnel connected in series to supply hot air directly into the drying tunnel, using two fans operated by a solar module. This dryer had a loading capacity of 120–150 kg of pineapple, and satisfactory drying times could be attained in a total of eight drying runs. The pineapples being dried in the solar tunnel dryer were completely protected from rain, insects and dust. Proximate analyses indicated that the dried product had a good quality for human consumption. In Japan, Basunia and Abe (2001) conducted thin-layer solar-drying experiments with medium-grain rough rice. In this case, the range of average drying air temperature was 22–35  C, the relative humidities were between 34% and 58%, and the initial moisture content of the rice was about 37%, dry-basis. A mixed-mode natural convection solar grain dryer was used, and data relating to sample weight and dry and wet bulb temperatures of the drying air were recorded continuously from morning to evening for each test. Subsequently, a good-quality dried product was obtained. Thin-layer solar drying experiments were also conducted by Yaldiz et al. (2001) for Turkish Sultana grapes, using an indirect forced convection solar dryer consisting of a solar air heater and a drying cabinet. The air velocity was varied,

325

326

10 Infrared Drying

and drying curves were correlated empirically. Experiments with thin layers of Turkish apricots were performed by Togrul and Pehlivan (2002) in an indirect forced convection solar dryer that consisted of a solar air heater with a conical concentrator and a drying cabinet. Air was forced into the solar air heater by a blower, and hot air obtained there was passed over the apricots. Changes in the mass of apricots and principal drying parameters were recorded continuously. 10.3.3 Infrared Drying

Infrared heating is widely used in industry (Mujumdar, 2007) for surface drying or for the dehydration of thin sheets such as textiles, paper, films, paints, coatings on ceramics, glass, and stone plates. In the automotive industry, IR drying for painton-metal applications and for paints on polyurethane-coated steel sheets has been highly successful. Another sector in which IR plays an important role is the pulp and paper industry. An example has been provided in Sweden, where this industrial sector imposes higher demands on energy consumption than any other. Consequently, a relatively new method that not only improved the paper quality but also achieved an energy-efficient drying involved the use of electrically operated IR radiation heating (Hannervall et al., 1992). Currently, IR drying – which employs the most efficient form of electromagnetic radiation for heat transfer – has been attracting interest in the agricultural industry because of its high thermal efficiency, fast heating response time, and direct absorbability by the material. In fact, this drying technology has appeared as a potential alternative to traditional heating methods for obtaining high-quality dried foods. Although the IR drying of thick porous materials has not yet been fully developed, some research groups have shown (mainly by means of experimental results) that one of the applications where long-wave IR heating is most efficient is in the dehydration of foods (Ginzburg, 1969; Kimura et al., 1992). In fact, IR radiation has been investigated for the drying of industrial grape byproducts (Celma et al., 2009b), seedless grapes (Caglar et al., 2009), olive husks (Celma et al., 2008), cashew kernels (Hebbar and Rastogi, 2001), herbs (Paakkonen et al., 1999), barley (Afzal et al., 1999), longan fruit (Nathakaranakule et al., 2010), potato (Afzal and Abe, 1999; Tan et al., 2001), carrot (Kocabiyik and Tezer, 2009), apple slices (Nowak and Lewicki, 2004; Togrul, 2005; Zhu and Pan, 2009), tomato byproducts (Celma et al., 2009a), shrimp (Fu and Lien, 1998), rough rice (Amaratunga et al., 2005; Abe and Afzal, 1997; Pan et al., 2008a; Khir et al., 2011), onion (Gabel et al., 2006; Jain and Pathare, 2004; Sharma et al., 2005), fresh and sugar-infused blueberries (Shi et al., 2008), persimmon (Kim, 1993), Ecklonia cava edible brown alga (Lee et al., 2010), saffron (Akhondi et al., 2011), pineapple (Tan et al., 2001), and red pepper (Koh et al., 1990; Nasiroglu and Kocabiyik, 2009). In general, these reports deal with the application of IR heating to the drying of different foods and analyses of the characteristics and performance of these particular IR dryers. In Japan, the food industry uses this type of heating for drying of, among others, seaweed, curry sauce, carrots and pumpkins (Kimura et al., 1992).

10.3 Classification, Research, and Applications of Radiation Drying

Fig. 10.5 Schematic diagram of laboratory-scale IR drying equipment. Reproduced with permission from Nasiroglu and Kocabiyik (2009).

Infrared drying provides significant potential advantages over conventional drying (Krishnamurthy et al., 2008), including a reduced heating time, uniform heating, reduced quality losses, an absence of solute migration in food material, versatile, simple and compact equipment, and significant energy savings. Infrared heating can be applied to various food processing operations, including drying, baking, roasting, blanching, pasteurization, and sterilization. In various reports, the effect of IR on food quality attributes has been discussed in the context of samples and process parameters. Applications of IR heating in food processing operations and future research potential are also reviewed. Infrared emitters with a maximal intensity wavelength in the range of 2.5 to 7 mm – that is, located in the MIR-to-FIR zone – have been proven suitable for the drying of agricultural products. A schematic diagram of the laboratory-scale IR drying equipment used by Nasiroglu and Kocabiyik (2009) is presented in Fig. 10.5. The drying chamber had a single-door opening at the front and was equipped with two incandescent lamps each with a power of 250 W. The output power of the lamps could be varied by regulating the voltage, and the air velocity could be adjusted by changing the fan revolution frequency. Before drying, samples were uniformly spread on the wire mesh tray in a thin layer. The sample mass and moisture loss were measured using a digital balance and recorded during drying. Infrared drying may be also classified by the type of emitter, whether an electric emitter (metal-sheath rods, quartz tubes, quartz lamps) or a gas-fired infrared (GIR) emitter (with ceramic-enclosed gas burners). Many previous reports have been primarily focused on the IR drying of agricultural materials using a conventional electric emitter (Abe and Afzal, 1997; Das et al., 2004; Jain and Pathare, 2004; Meeso et al., 2004; Sharma et al., 2005; Kumar et al., 2006). Investigations into paddy drying by using an IR emitter, in which heat was generated from natural gas (Amaratunga et al., 2005), involved the use of a gas-fired infrared (GIR) emitter that radiated wavelengths in the range of 2.5 to 3.0 mm. With

327

328

10 Infrared Drying

a maximum operating temperature of up to 900  C, the emitter was used to explore the drying kinetics of long-grain paddy, especially of premium grade rice (Khao Dok Mali 105 or Jasmine rice). The objective was to identify a model to describe the GIR radiation drying characteristics of this paddy type. Gas-fired IR drying of paddy was also considered by Laohavanich and Wongpichet (2008), who dried wet paddy under different process conditions by applying a pilot-scale experimental GIR dryer. The IR radiation was expressed in terms of peak wavelength of the IR emitter, and the initial moisture content of paddy was varied to study the drying behavior. Although, in general, sun drying and hot-air drying are the most common methods used for paddy drying, several groups (Ginzburg, 1969; Abe and Afzal, 1997) have aimed specifically at the IR drying of paddy. The effect of radiation intensity (0.125 to 0.5 W cm 2) and slab thickness (2.5 to 10.5 mm) on the moisture diffusion coefficient of potatoes during FIR drying has been investigated by Afzal and Abe (1999). The study results showed that diffusivity was increased with increasing radiation intensity and slab thickness, whereas the activation energy for moisture desorption was decreased with increasing slab thickness, and this resulted in higher drying rates for slabs of greater thicknesses. Investigations into the IR radiation drying of high-moisture paddy and parboiled rice were continued by Das et al. (2004), who reported that both radiation intensity and bed depth had significant effects on drying performance and rice quality. X-ray microtomography (see Chapter 4, Volume 2, Modern Drying Technology), coupled with image analysis, represents a nondestructive technique that allows an entire sample to be scanned, providing information such as pore size distribution and total pore volume, without the need for serial cuts (as in the case of scanning electron microscopy). The use of X-ray microtomography was demonstrated when investigating the effects of FIR-assisted drying on the microstructure of banana (Leonard et al., 2008). The main advantages of IR dryers include:  They employ an alternative source of energy, and have a higher rate of heating.  They have a high-efficiency conversion of electrical energy into heat for electrical IR.  They heat only the object, and not the surroundings; selective heating is possible; and they can easily be subdivided to create zones for uniform heating of the surface-wet product.  They respond very quickly to changing process conditions, and have quick startup and shut-down procedures.  Due to their smaller size they save on floor space; they are also easy to control and have lower capital and installation costs. The potential limitations of IR dryers include:  They are difficult to operate with temperature-sensitive materials.

10.3 Classification, Research, and Applications of Radiation Drying

 They have a low penetration depth, and cause skin formation and blistering of some coatings.  As IR is basically a surface phenomenon, it is difficult to dry heavier coatings.  Typically, they cannot be scaled-up in a straightforward manner; laboratory and pilot trials are needed to confirm any design changes.

10.3.4 Catalytic Infrared Drying

Recently, a new method of flameless catalytic infrared (CIR) drying has been developed, in which energy is generated by catalyzing natural gas or propane conversion with a platinum catalyst. Relatively few reports are available regarding CIR dehydration. Catalytic combustion occurs by allowing fuel (natural gas or propane) to enter the back of the air-tight heater pan, and then dispersing it through a catalyst pad at the face. At the same time, the air diffuses through the catalyst pad from the front, causing fuel oxidation. Catalytic combustion occurs without a flame, which can damage the product at a temperature below the ignition temperature. The electromagnetic radiant energy from CIR has peak wavelengths in the range of MIR to FIR radiation (3–7 mm) and is used for drying. The peak wavelengths match reasonably well with the three absorption peaks of liquid water, which could result in a rapid moisture removal. Because the CIR directly converts natural gas or propane to radiant energy, almost without losses, it is more energy efficient than the typical IR emitters which use electricity. A comparison of the drying of high-solids onion with CIR heating and forced air convection (FAC) heating was provided by Gabel et al. (2006). For this, sliced onions were dehydrated at 60, 70 and 80  C by means of FAC heating or CIR heating, with and without air recirculation. An onion sample (250 g for CIR, 150 g for FAC) of intact slices was arranged in a single layer on the drying tray at a loading of 2.5 kg m 2. In general, CIR both with and without air recirculation, had higher maximum drying rates, shorter drying times, and greater drying constants than FAC at moisture contents greater than 50% (d.b.). The CIR laboratory dryer arrangement used by Gabel et al. (2006) is shown in Fig. 10.6. This consisted of a drying chamber with a CIR emitter (Catalytic Drying Technologies, Independence, Kansas, USA) mounted from the top of the chamber. The sample was placed in an aluminum wave-guide (48  30 cm2, upper rim; 42  22 cm2, base perimeter), used to achieve a uniform heating of the entire product and placed on a drying tray. A balance (0.1 g accuracy) was placed beneath the drying tray to measure the product weight during drying. An exhaust fan was located on the top of the drying chamber for ventilation. Two fans (Fig. 10.6a) mounted on each side of the dryer were used to recirculate part of the warm air in the drying chamber. These fans pulled air from the top of the drying chamber and fed it back into the chamber through slits running the entire length of the sides of the drying chamber. The CIR emitter was preheated by an embedded electric

329

330

10 Infrared Drying

Fig. 10.6 (a) Catalytic infrared dryer. 1, natural gas; 2, gas flow control; 3, air exhaust; 4, blower; 5, CIR emitter; 6, recirculation of warm air; 7, computer controller; 8, wave guide; 9, thermocouple;

10, onion sample; 11, balance; (b) Construction of catalytic infrared emitter. Reproduced with permission from Gabel et al. (2006).

element. The natural gas intake was regulated by a gas control valve and a computer. Two thermocouples were used to measure the product temperature; each thermocouple was placed inside the flesh of an onion slice, which was placed within the innermost 10 cm2 of the onion bed. The average temperature was used as input to control the product temperature by turning the emitter on or off, which was achieved by opening and closing the gas supply valve of the emitter. The maximum operating temperature of the emitter was 500  C at wavelength 3.3 mm. The FAC dryer used in the tests was an electrically heated column dryer with a diameter of 33 cm. A fan blew heated air through an electric coil heater and then through the column. The product was placed in a circular mesh tray near the bottom of the column and suspended by wires to the balance that recorded product weight changes during drying. The electric heating coils were automatically switched on and off to maintain the set point temperature, in a similar fashion to the CIR dryer. The average air velocity for all of these tests was maintained at 0.5 m s 1. For CIR, the drying tray was placed in the preheated CIR drying chamber and the thermocouples were positioned as described above. The distance between the emitter and drying tray was 15 cm, with a maximal radiation intensity of 4752 W m 2. During CIR drying tests with air recirculation, both recirculation fans were switched on during the entire test. The targeted final moisture content of the dried onion was set at X ¼ 0.1 in this study. The CIR tests showed much higher drying rates than the FAC drying at X > 0.5 (Tab. 10.2). Below this moisture content, the rates of CIR drying were smaller than the FAC dryer rates, which indicated a lower moisture transfer in the onion slices. Raising the drying temperature caused an increase in the drying rate of both the CIR and FAC trials. In the CIR drying tests, the constant rate period was absent or very brief, because onions are hygroscopic and hygroscopic, and colloidal foods tend to quickly enter the falling rate period. Additionally, a more rapid surface drying by CIR may result

10.3 Classification, Research, and Applications of Radiation Drying Tab. 10.2

Summary of onion drying results.

Drying condition 60  C CIR, recirculation 70  C CIR, recirculation 80  C CIR, recirculation 60  C CIR 70  C CIR 80  C CIR 60  C FAC 70  C FAC 80  C FAC

Drying time to reach X ¼ 0.5 (min)

Max. drying rate g(kg initial mass)

54.5  7.8 32.0  1.0a) 24.0  4.6a) 47.3  7.5 36.5  7.8a) 18.0  2.8a) 56.0  8.5 51.0  1.4 33.0  5.7

46.9  8.4a) 59.9  4.2a) 81.5  6.6a) 47.5  10.5a) 67.7  9.2a) 83.4  4.6a) 20.7  3.1 23.9  3.9 43.2  22.1

1


min

1

a) Significantly different when compared to FAC value at same temperature.

in a quicker entrance to the falling rate period due to water diffusing very slowly to the surface of the onion. The dried onion quality, measured as pungency degradation, was similar for both drying methods at 60 and 70  C. The color analysis showed a better product color (whiter and less yellow) at lower temperatures for CIR and at higher temperatures for FAC. The browning that was visible may have been caused by the higher surface heat flux of CIR heating and the longer process times of FAC drying. Aerobic plate counts and coliform counts were not significantly different for the CIR- or FAC-dried products, although samples dried by CIR had significantly lower yeast and mold counts than those dried by FAC. On basis of their results, Gabel et al. (2006) recommended the use of CIR during the early stages of onion drying. The Office of Industrial Technologies (OIT) in the Energy Efficiency and Renewable Energy Division of the US Department of Energy (www.oit.doe.gov) has proposed the use of a CIR drying system, employing IR energy in the 4–7 mm range (range of MIR to FIR radiation), to dry wood products. The catalytic combustion of natural gas creates IR energy without a direct flame. This is an advantage, because direct flame combustion may damage the product at a temperature below the ignition temperature. Reducing the moisture content with IR drying by transferring energy directly to the moisture instead of heating the air and surrounding metal structure of the drying unit requires less energy, reduces air emissions, and allows more product to be processed compared to conventional drying. When coupled with advances in remote sensing technology and programmable logic systems, wood fiber processors will have the ability to more carefully control the drying units. Subsequently, a large prototype unit has been constructed and tested with wood chips, sawdust, and a variety of agricultural products. As the CIR drying system was proven to dehydrate forest and agricultural products efficiently, the current focus is on the conveyance system for distributing the product evenly throughout the dryer. In addition to perfecting the conveyance system, the demonstration project will document the

331

332

10 Infrared Drying

energy, waste and economic benefits resulting from the construction and operation of a full-scale dehydration unit. One company specializing in the adaptation of flameless CIR drying technologies is, as noted above, the Catalytic Drying Technologies Co. (CDT, www .catalyticdrying.com). CDT is a division of the American Catalytic Industrial Group, Inc., a diversified family of companies aiming at the application of IR drying technologies to agricultural businesses and food processing. Since 1953, CDT has developed and patented IR drying systems for food, feed and fiber, and over 100 applications have been tested to demonstrate improvements in the quality, yield and throughput for dehydration, disinfestation, and thermal treatment. In the case of rice, a 2% increase in yield, a 10% increase in throughput and a 5% cost reduction were reported by using CIR drying instead of hot-air drying. The main advantages of CIR drying appear as higher yields, low capital and operating costs, higher throughputs, uniform drying efficiencies, a higher quality of end products, a lower environmental impact, and ease of control.

10.4 Types of Radiators 10.4.1 General Considerations

Sources of IR radiant heat at surface temperatures in the range of 600–2500 K are electrically heated metal-sheath rods, quartz tubes and quartz lamps, as well as ceramic-enclosed gas burners and catalytic IR burners. Infrared heaters are wellsuited to a wide variety of applications, and designed to transmit large amounts of energy quickly and with a high efficiency. The particular wavelength of the IR radiation has a critical effect on the efficacy of the heating process, and IR heaters, which are optimally matched to the materials to be heated, can provide energy savings of up to 50% over other technologies. With a correct selection of the radiation source the IR energy requirements for drying can be reduced by up to 50%. The wavelength, surface temperature and spectral range for the main types of electric IR radiators are listed in Tab. 10.3. The terms “metal rod,” “ceramic,” “quartz or halogen” heating element describe the main components from which the heating element is made. Tab. 10.3

Wavelengths of various IR radiators.

Radiator or heat source

Peak wavelength (mm)

Surface temperature ( C)

Spectral range

Metal rod radiator Infrared lamp Ceramic surface radiator Halogen lamp Quartz rod radiator

2.8–4.4 >1.3 2.8–5.0